INFORMATION TO USERS This manuscript has been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer. The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction. In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion. Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand corner and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back of the book. Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6" x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order. University Microfilms International A Bell & Howell Information Company 300 North Zeeb Road. Ann Arbor. Ml 48106-1346 USA 313/761-4700 800/521-0600 R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout p erm ission . Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Order N um ber 9518741 U tilization o f microwaves in ceramic processing Fang, Yi, P h.D . The Pennsylvania State University, 1994 UMI 300 N. ZeebRd. Ann Arbor, MI 48106 R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The Pennsylvania State University The Graduate School U T IL IZ A T IO N O F M ICROW AVES IN C E R A M IC PR O C E S S IN G A Thesis in Solid State Science by Yi Fang Submitted in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy December, 1994 R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. We approve the thesis of Yi Fang. Date of Signature Della M. Roy y Chair of Committee Professor o f Materials Science, Emerita Thesis Co-Advisor JU. Dinesh K^Agrawal Associate Professor o f Materials Thesis Co-Advisor Rustum Roy Evan Pugh Professor of the S^lid State "Ttcsis-Co-Advisor Paul W. Brown Professor of Ceramic Science and Engineering David J. Green Professor of Ceramic Science and Engineering <— ^ d .~ r - S’, Robert N. Pangbom Professor of Engineering Mechanics In Charge o f Graduate Program in Solid State Science R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. in ABSTRACT The studies described in this thesis are focused mainly on microwave processing of various ceramic materials. The material systems selected for the studies include some traditional ceramic materials such as alumina (AI2 Q 3) and mullite (3 A l 2 ( V 2 Si0 2 ), new materials such as biomedical materials-hydroxyapatite [HAp, Ca 10 (PO4 )6 (OH)2 ] ceramics, newly developed low-thermal-expansion m aterials based on NZP (NaZr 2 P 3 0 12 ), and various zirconia-containing ceramic composites. M icrowave processing studies were carried out on the selected m aterials in a m ultim ode m icrowave cavity at 2.45 GHz. Conventional sintering experiments were conducted for comparison. In the HAp system, dense, porous, and transparent HAp ceramics were successfully fabricated by microwave processing at ambient pressure. A kinetics study on the densification shows that the regularly dense HAp ceramics could be fabricated by microwave processing at temperatures 100°C lower than that by conventional method. The apparent activation energy for the densification of HAp was found to be 112 kJ/mole and 181 kJ/mol. in microwave and conventional sintering, respectively. M icrow ave sintering studies o f m ullite ceramics were carried out using both crystalline and amorphous (gel) precursors. In the temperature range up to 1500°C, microwave enhancing effect on densification o f crystalline mullite and monophasic mullite gel was not significant. The densification o f diphasic mullite xerogel was significantly accelerated in the temperature range between 1100-1250°C. Microwave irradiation for 5-10 minutes lowered the mullitization temperature of the diphasic gel by 75-100°C. In addition, microwave heating promotes seeding effect o f the diphasic gel. By microwave processing, 0.5% seeding (monophasic) gel significantly accelerated mullite transformation o f the diphasic gel, while by conventional heating, seeding effect did not show up with less than 10% seeding gel. A kinetics study found that the apparent activation energy for the densification o f the diphasic mullite xerogel in the microwave field is 396 kJ/mole, while that under conventional conditions is 702 kJ/mole. Using the diphasic mullite aerogel as starting material, transparent mullite ceramics R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission. were fabricated under am bient pressure by microwave and conventional sintering, respectively. It is suggested that the microwave enhanced densification, mullite transformation, and epitaxy of the diphasic mullite gel are mainly attributed to the decrease in viscosity of the amorphous silica in the gel, which effectively promotes viscous flow o f the gel and accelerates the diffusion of the dissolved alumina. In contrast, the low dielectric loss and low solid-state diffusivity of the ions are responsible for the sluggish sintering kinetics of crystalline mullite and monophasic gel which converts to mullite below 1000°C. Microwave processing enhanced the densification of the amorphous alumina by 2649% over conventional sintering, but only 2.3% for the crystalline precursor, indicating that the amorphous alumina absorbs microwaves more efficiently than the crystalline one. Significant microwave enhancing effect on densification was also verified with two materials of [NZP] structure and various zirconia (PSZ)-containing composites. In the case o f the composites, the microwave enhancement decreases with increasing the PSZ content due to the better sinterability of PSZ than the matrix materials. Based on the observations in the current work, it is suggested that, besides the intrinsic dielectric properties, free energy potential, or activity, which is related to various higher-energy defects and surface area, is an important factor responsible for microwave absorption o f materials irradiated. Specific surface area is an indicator of free energy. Also, a higher surface area also implies low er bulk thermal conductivity and higher scattering effect to microwaves. Therefore, the higher the surface area, the stronger the microwave absorption. It is also suggested that the microwave enhancing effect on densification and related processes is due to an athermal effect which promotes the chemical potential as the driving force for the thermal processes. In a thermal process of a specified material, the degree of microwave enhancing effect depends on the relative magnitude of the athermal microwave effect and the total thermal energy supplied to the material. Therefore, the microwave enhancing effect is shielded when the processing temperature is too high. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. V TABLE OF CONTENTS Page A BSTRA CT ................................................................................................................. iii L IS T O F FIG U R E S ..................................................................................................... xi LIST OF TABLES ........................................................................................................... xviii PRINCIPAL SYMBOLS AND ABBREVIATIONS ....................................................... xx A C K N O W L E D G M E N T S ............................................................................................... xxi Chapter 1. IN T R O D U C T IO N ....................................................................................................... 1 1.1. 1.2. G eneral S ta te m e n t .......................................................................................... S tatem ent o f P roblem ................................................................................... 1 2 1.3. O bjectives and Scope o f Study .................................................................. 5 2. LITERATURE R E V IE W .......................................................................................... 2.1. M icro w av es ...................................................................................................... 2.2. M icrowave Processing o f M aterials ............................................................ 2.2.1. M icrow ave-M aterial Interactions .................................... 7 7 10 10 2.2.2. Dielectric Properties o f Ceramic M aterials .................................... 2.2.3. C haracteristics of M icrowave H eating ........................................... 12 15 2.2.4. H istorical Review o f M icrowave Sintering ..................................... 22 2.2.5. Current Status of Microwave Sintering Technology .......................... 24 2 .2.5.1. 2 .2.5.2. M icrow ave A pplicators .................................................... 24 N ew D evelopm ents ........................................................... 25 3. EXPERIMENTAL PR O C E D U R E S 3.1. 3.2. ....................................................................... 3.2.1. M icrow ave F u rn ace ........................................................................... 3.2.2. Temperature M easurement in Microwave P ro c essin g ...................... 3.2.3. C on v en tio n al F u rn ace ........................................................................ 3.3. S in terin g P ro c ess ........................................................................................... 3.4. 28 S tartin g M aterials .......................................................................................... 28 Experim ental Setups for Sintering ............................................................. 28 28 31 32 33 C h a ra c te riz a tio n ............................................................................................... 33 3.4.1. P o w d er C h a ra cte riz atio n ...................................................................... 34 3.4.2. T herm ogravim etric A nalysis ............................................................... 34 R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. TABLE OF CONTENTS (continued) Chapter Pages 3.4.3. D ifferential Therm al A nalysis ......................................................... 34 3.4.4. D ila to m e try ........................................................................................... 35 3.4.5. Pow der X -R ay D iffraction .............................................................. 3.4.6. L attice Param eter Refinem ent ......................................................... 35 35 3.4.7. D ensity M easurem ent ....................................................................... 3.4.8. Mercury Intrusion Porosimetry................................................................. 35 36 3.4.9. Diametral Tensile Strength M easurem ent ........................................ 3.4.10. Fracture Toughness M easurem ent .................................................. 3.4.11. Scanning E lectron M icroscopy .................................................... 3.4.12. Transm ission Electron M icroscopy ............................................. 36 37 37 39 3.4.13. In frared Spectrom etry .................................................................... 39 4. MICROWAVE SINTERING OF HADROXYAPATITE ......................................... 40 4.1. In tro d u c tio n ....................................................................................................... 40 4.1.1. G e n e ral .................................................................................................. 40 4.1.2. Application o f HAp as a Biom aterial ............................................. 41 4.2. Synthesis and Thermal Stability of Hydroxyapadte (H A p )........................ 44 4.2.1. In tro d u c tio n ........................................................................................ 44 4.2.2. E xperim ental Procedure 4 .2 .2 .1 . H A p Synthesis .................................................................. 47 ........................................................... 47 4 .2 .2 .2 . C haracterization ..................................................................49 4.2 .2 .3 . T herm al Stability ............................................................. 49 4.2.3. R esults and D iscussion ................................................................... 50 4.2 .3 .1 . M orphology .........................................................................50 4.2 .3 .2 . 184.108.40.206. C ry sta llin ity S to ichiom etry ....................................................................... 54 .................................................................... 54 4.2 .3 .4 . TG A R esults............... ............................................................. 57 220.127.116.11. L attice Param eters .......................................................... 62 4 .2 .3 . 6 . T herm al Stability ............................................................. 62 4.2.4. S um m ary ............................................................................................... 65 4.3. Fabrication of Regularly Dense HAp Ceramics ......................................... 71 4.3.1. In tro d u c tio n .......................................................................................... 71 4.3.2. E xperim ental Procedure .................................................................. 72 R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. vii TABLE OF CONTENTS (continued) Chapter Page. 4 .3 .2 .1 . 4 .3 .2 .2 . M a te ria l................................................................................... 72 S in terin g Setup ................................................................. 72 4 .3 .2 .3 . 4 .3 .2 .4 . Tem perature M easurem ent .............................................. 73 S in te rin g .............................................................................. 76 4 .3 .2 .5 . 4 .3.3. R esults C h a ra cte riz atio n and D iscussion ................................................................ ................................................................ 76 77 4 .3 .3 .1 . 4 .3 .3 .2 . Tem perature M easurem ent .............................................. 77 S intering Process ............................................................. 77 4 .3 .3 .3 . 4 .3 .3 .4 . U niform ness in Sintering .............................................. 81 S in tered D ensity ............................................................... 81 4 .3 .3 .5 . 4 .3 .4 .6 . 4 .3 .3 .7 . 4 .3 .3 . 8 . G rain Size and M icrostructure ....................................... Phases ............................................................................... M echanical Strength ...................................................... E nergy Consum ption .................................................... 85 85 89 89 ............................................................................................ 4.3.4. S um m ary 4.4. M icrowave Processing o f Porous HAp C eram ics........................................ 92 92 .................................................................... 4.4.1. In tro d u c tio n 4.4.2. E xperim ental Procedure .................................................................... 92 93 4 .4 .2 .1 . 4 .4 .2 .2 . H A p Precursor Pow ders ............................................... Pow der C onsolidation .................................................... 93 93 4 .4 .2 .3 . 4 .4 .2 .4 . ........................................................................... S in te rin g C h aracterizatio n ............................................................... 95 95 4.4.3. R esults 4 .4 .3 .1 . 4 .4 .3 .2 . 4 .4 .3 .3 . and D iscussion ................................................................... 95 P o ro sity ............................................................................ 97 P h a se s ................................................................................. M icrostructure ................................................................. 97 99 4 .4 .3 .4 . S tren g th ............................................................................. 4.4.4. S um m ary ................................................................................................ 99 99 4.5. Fabrication o f Transparent HAp Ceramics ................................................ 102 4.5.1. In tro d u c tio n .......................... .................................................................... 102 4 .5 .2 . E xperim ental Procedure .................................................................... 103 4.5.3. R esults and D iscussion .....................................................................103 4.5.4. S um m ary ................................................................................................ 112 R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout perm ission. viii TABLE OF CONTENTS (continued) Chapter 4 .6 . Page, D ensification K inetics o f HA p ..................................................................... 112 4.6.1. In tro d u c tio n ........................................................................................... 112 4.6.2. E xperim ental Procedure ................................................................. 113 4.6.3. R esults and D iscussion ...................................................................... 113 4.6.4. S u m m ary ................................................................................................ 115 5. MICROWAVE SINTERING O F M ULLITE CERAMICS ............................... 118 5.1. In tro d u c tio n ...................................................................................................... 118 5.2. Synthesis o f M ullite P recursors .............................................................. 120 5.2.1. M o n o p h asic G el ................................................................................... 120 5.2.2. D ip h a sic G el ....................................................................................... 120 5.2.3. Seeded D iphasic G els ...................................................................... 121 5.2.4. D iphasic M ullite A erogels ............................................................... 122 5.3. M ullite Transform ation o f D iphasic Gels ................................................... 123 5.3.1. In tro d u c tio n ........................................................................................... 123 5.3.2. E xperim ental Procedure .................................................................. 123 5.3.3. R esults and D iscussion .................................................................... 124 5 .3 .3 .1 . C rystallization o f Gels ................................................... 124 5.4. 5 .3 .3 .2 . Seeding Effect o f D iphasic Gels .................................... 124 Microwave Sintering of M ullite Ceramics (Regular and Transparent) 132 5.4.1. In tro d u c tio n ....................................................................................... 132 5.4.2. E xperim ental P rocedure ................................................................... 132 5.4.3. R esults and D iscussion .................................................................. 133 5 .5. D ensification Kinetics o f D iphasic Gel .................................... 146 5.5.1. In tro d u c tio n ........................................................................................... 146 5.5.2. E xperim ental Procedure ....................................................................146 5.5.3. R esu lts and D iscu ssio n ...................................................................... 147 5.6. Summary on M icrowave Processing o f M ullite Materials ........................ 151 6 . MICROWAVE SINTERING OF OTHER CERAM ICS ...................................... 153 6.1. M icrow ave Sintering o f A lum ina ............................................................ 153 6.1.1. In tro d u c tio n ........................................................................................... 153 6.1.2. E xperim ental Procedure .................................................................... 153 R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission. ix TABLE OF CONTENTS (continued) Chapter 6.2. Page. 6.1.3. R esults and D iscussion .................................................................... 154 M icrow ave Sintering o f [NZP] Ceramics ................................................ 157 6.2.1. In tro d u c tio n ............................................................................................ 157 6.2.2. E xperim ental P rocedure ................................................................... 157 6.2.3. R esults and D iscussion ......................................................................158 6.2.4. S um m ary ..................................................................................................164 7. MICROWAVE SINTERING OF CERAM IC COMPOSITES .............................. 165 7 .1 . In tro d u c tio n ......................................................................................................... 165 7 .2 . M icrowave Sintering of Hydroxyapatite-Based C om posites..................... 166 7 .2 .1 . In tr o d u c tio n ................................................................................................. 166 7.2.2. E xperim ental P rocedure ..................................................................... 167 18.104.22.168. 22.214.171.124. 1.2.23. 126.96.36.199. S tarting M aterials ......................................................... 167 C om paction .......................................................................168 S in te rin g .......................................................................... 168 C haracterizatio n ............................................................. 170 7.2.3. Results and D iscussion .................................................................. 170 7 .2.3.1. H e a tin g .............................................................................. 170 1.23.2. 1.233. Phase C om position ....................................................... 170 S intered D ensity ........................................................... 172 188.8.131.52. 184.108.40.206. ....................................................... 174 Fracture Toughness M ic ro stru c tu re .....................................................................175 7 .2.4. Sum m ary .......................................................................................... 7.3. Microwave Sintering o f Various Zirconia-Containing C o m p o sites 175 178 7.3.1. In tro d u c tio n ....................................................................................... 178 7.3.2. E xperim ental Procedure ................................................................. 179 7 .3.2.1. 13.2.2. 1 3 .2 3 . Starting M aterials ......................................................... 179 S in te rin g ............................................................................. 180 C haracterizatio n ............................................................. 181 7.3.3. Results and D iscussion .................................................................. 181 7.3.4. S um m ary .................................................................................................. 196 R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout p erm ission . x TABLE OF CONTENTS (continued) Chapter 8 Page . DISCUSSION ON MICROWAVE ENHANCING EFFECTS ...............................197 8.1. M icrowave Phenom ena o f M icrowave Effects ........................................... 197 8.2. Driving Force for Densification in Solid-State S in te rin g ............................ 198 8.3. Loss M echanism s in M icrow ave H eating ................................................... 202 8.4. Interactions o f M aterial Surfaces with M icrowaves .....................................203 8.4.1. D ie le c tric L osses ................................................................................. 203 8.4.2. T herm al C onductivity .........................................................................204 8.4.3. S urface S cattering ............................................................................ 204 8.5. Atherm al E ffect in M icrowave Processing ..................................................204 8 .6 . M icrow ave E nhanced D iffusion .................................................................. 206 8.7. M icrowave Effects on Diphasic M ullite Gels ..............................................207 8.7.1. Observed M icrowave Effects on Mullite M a te ria ls...........................207 8.7.2. C haracteristics o f M ullite G els .........................................................208 8.7.3. Microwave Effect on Densification of Diphasic Mullite X erogel 209 8.7.4. M icrowave Effect on Crystallization and E p ita x y .......................... 212 8 . 8 . M icrowave Effect on Sintering of Other Materials .................................... 213 8.9. Influence of Processing Temperature on M icrowave E f f e c t.......................... 214 8.10.G rain G row th during M icrowave Processing ............................................ 216 9. C O N C L U S IO N S ...........................................:.............................................................. 217 9.1. Sum m ary and C onclusions ......................................................................... 217 9.2. Suggestions fo r F uture S tudies................................................................... 222 9.2.1. Dielectric Properties o f Amorphous Materials .............................. 222 9.2.2. M icrowave Sintering o f HAp in Moist E nvironm ent.................... 222 9.2.3. Better Transparent Mullite Ceramics by Process M odification R E FE R E N C E S 223 ............................................................................................................... 224 R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. xi LIST OF FIGURES Page Chapter 2 Figure 2.1 Electromagnetic spectrum showing the band o f microwaves ............... 8 Figure 2.2. Microwave frequencies and wavelengths used in materials processing.... 9 Figure 2.3. Classification of m icrowave-m aterial interactions................................... 11 Figure 2.4. Relative dielectric constant (8-10 GHz) versus temperature..................... 16 Figure 2.5. Loss tangent (8-10 GHz) versus tem perature.......................................... 17 Figure 2.6. Heating patterns in conventional and microwave furnaces...................... 18 Figure 2.7. Therm al runaway during m icrowave heating........................................ 20 Schematic display of the sintering packet in the microwave processing. 1.Turning table. 2. Zirconia cylinder. 3. Samples. 4. Thermocouple. 5. Fibermax insulator. 6 . MoSi 2 rods. 7. M icrowave port................. 30 Schematic showing of the cracks in Vickers and Knoop indentations 38 Figure 4.1. Solubility of calcium phosphates at 25°C............................................ 42 Figure 4.2. Influence of moisture on the decomposition of H A p............................. 46 Figure 4.3. Micrographic morphology of HAp synthesized by (a) hydrolysis of a-TCP, (b) hydrolysis of brushite (DCPD) plus ripening treatment, (c) precipitation, and (d) hydrothermal m ethod............................................. 51 A micrograph showing that the HAp crystals were damaged by electron beam bombardment during TEM observation at 120 kV. ........................... 52 Morphology of HAp produced by the hydration of a-C a 3P 0 4 at 25°C fo r 7 days .................................................................................................. 53 Chapter 3 Figure 3.1. Figure 3.2. Chapter 4 Figure 4.4. Figure 4.5. Figure 4.6. XRD patterns of the as-synthesized HAp prepared by (a) hydrolysis of DCPD, (b) hydrolysis of DCPD plus triple ripening treatment, (c) precipitation, (d) hydrothermal method, and (e) hydrolysis of a-T C P ....... 55 R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. LIST OF FIGURES (continued) Page Figure 4.7. XRD patterns showing that the crystallinity of the precipitated HAp increased with heating temperature. All samples were fired for 1 h. The marked (•) peaks belong to TCP; all unmarked belong to HAp 56 Figure 4.8. IR spectra of HAp prepared by various m ethods................................ 60 Figure 4.9. TG thermograms of HAp synthesized by (a) hydrolysis of a-TCP, (b) hydrothermal method, (c) DCPD hydrolysis, ripened twice, (d) DCPD hydrolysis, ripened thrice, (e) DCPD hydrolysis without ripening, (f) DCPD hydrolysis (ripened once),and (g) precipitation................... 63 Figure 4.10. XRD patterns showing that the hydrothermally (200 °C, 1.5 MPa) synthesized HAp fired at 1370 °C in air of 50% RH (11.88 mm Hg) rem ains undecom posed ......................................................................... 66 Figure 4.11. XRD patterns o f the as-synthesized HAp by DCPD hydrolysis, heated at various temperatures. It is seen that the decomposition started at about 700 °C. 714 °C, p-TCP + HAp; 7 3 3 ,1100°C, p-TCP. 1300°C; a-T C P + (3-TCP. 67 Figure 4.12. XRD patterns showing the influence of repeated ripening on the thermal stability of HAp obtained by DCPD hydrolysis. Samples were heated 1 h at 1100 °C in air of 50% RH. (a) Without ripening: (3-TCP. (b) Ripened once: HAp+ P-TCP. (c) Ripened twice: HAp+P-TCP (tr.). (d) Ripened thrice: HAp only, no deco m p o sitio n........................................................ 68 Figure 4.13. XRD patterns showing the phase composition of SOR (strengthened one-step ripening) heated 1 h at 1350 °C and 1370 °C, respectively. The ripening after the hydrolysis of DCPD was carried out at 75 °C and (a) 20% CCD, 70 min., initial pH 8.87, (b) 60% CCD, 24 h, initial pH=8.87, (c) 20% CCD, 70 min., initial pH=11.0, and (d) 60% CCD, 24 h, initial pH =l 1. The marked peaks (•) belong to a-TCP. .................. 69 Figure 4.14. Schematic showing of the arrangements in the microwave sintering cavity. The sintering package includes (1) thermocouple, (2) zirconia cylinder, (3) pack of samples, (4) thermal insulating fibers, (5) power level control, and (6 ) tim e control....................................................... 74 Figure 4.15. Schematic showing of relative positions of thermocouple and p-TCP pellets during the temperature measurement calibration. Labeled are (1) thermocouple, (2) porous zirconia, and (3) TCP p e lle ts ................................................................................................................ 75 R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. LIST OF FIGURES (continued) Page Figure. 4.16. XRD patterns (Cu Ka) o f /J-TCP samples used to calibrate temperature m easurement (a) Conventionally sintered for 15 min. at temperatures indicated, (b) Microwave sintered at 1125±10 °C for 15 min. (pellet positions are indicated). (•): a-TCP. Unmarked peaks: /J-TCP. ............ 78 Figure 4.17. Temperature-time curves in the microwave sintering of HAp at (a) 1200 °C and (b) 1300 °C. Sintering time: (A) 5 min., (-----) 10 min., (••••) 2 0 m in ............................................................................................... 80 Figure 4.18. Photograph showing the uniform shrinkage of microwave sintered HAp pellets. The dime-pattems on the green pellets (large) have been uniformly reduced to the sintered pellets (small) by microwave sintering at 1200 °C for 7 m in............................................................................... 82 Figure 4.19. Relative density (% theoretical) of HAp ceramics sintered by (1) microwave/ 1300 °C/10 min., (2) microwave/1200 °C/10 min., (3) conventional/1300 °C/2 h, and (4) conventional/1200 °C/2 h................ 83 Figure 4.20. SEM micrographs showing (a) as-sintered surface and (b) fracture surface of HAp conventionally sintered at 1200 °C for 120 m in.............. 86 Figure 4.21. SEM micrographs of (a) as-sintered surface and (b) fracture surface of HAp sintered by microwave processing at 1200 °C for 10 min........... 87 Figure 4.22. Diametral tensile strength of HAp ceramics (a) versus sintering density: (O) microwave sintered, and (A) conventionally sintered; (b) versus green density (1) microwave/20 min.; (2) microwave/10 min., and (3) conventional/2 h .......................................................................................... 90 Figure 4.23. Comparison of temperature-time profiles in the sintering of HAp by (1) microwave and (2 ) conventional m ethods.............................................. 91 Figure 4.24. Morphology of the starting HAp powders composed of (A) whiskers synthesized by the hydrolysis of a-T C P and (B) agglomeration of fine single crystals synthesized by the hydrolysis of brushite followed by ripening treatm ent with calcium chloride................................................ 94 Figure 4.25. Microstructures o f the microwave-processed porous HAp ceramics (fracture surface), (a, b) Starting with fine HAp powder: (a) 1150°C, 3 min., porosity 73.3%. (b) 1200°C, 5 min., porosity 24.5%. (c, d) Starting with HAp whiskers, pressed to different green densities (c: 43.58%, d: 54.29%), both fired at 1150°C for 1 min. Final porosity (c) 47% ; (d) 2 8 % .................................................................................. 100 Figure 4.26. Diametral tensile strength of microwave-processed HAp ceramics 101 R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. xiv LIST OF FIGURES (continued) Page Figure 4.27. Morphology o f fine hydroxyapatite powders synthesized by (a) hydrothermal treatment and (b) hydrolysis of brushite followed by "ripening" treatm ent ............................................................................... 104 Figure 4.28. Heating curves of sintering of hydroxyapatite ceramics by (a) microwave processing and (b) conventional sintering............................................... 105 Figure 4.29. Dilatometry curve of the hydrothermally synthesized HAp heating at 5°C per m in u te .......................................................................................... 106 Figure 4.30. Transparent HAp ceramics fabricated at ambient pressure in air (a-c) by microwave processing for 5 min. at (a) 1150°C, (b) 1125°C, (c) 1100°C, totally irradiated for 18 min., and (d) by conventional sintering at 1150 °C for 5 min. after heating to 1150°C at 5 °C/min., totally heated fo r 230 m in ................................................................................................. 109 Figure 4.31. Micrographs of the as-sintered surface of the transparent hydroxyapatite ceramics sintered by (a) microwave processing and (b) conventional m e th o d ............................................................................................................. 110 Figure 4.32. Powder XRD pattern of the transparent HAp ceramics by microwave processing at 1150 °C for 5 min., showing that the ceramics are highly crystalline single phase H A p................................................................. Ill Figure 4.33. Sintered density as a function o f time of HAp sintered by (a) microwave and (b) conventional m ethods................................................................ 114 Figure 4.34. Arrhenius plot of the densification of HAp under (a) microwave and (b) conventional conditions........................................................................... 117 Chapter 5 Figure 5.1. DTA curve o f monophasic aluminosilicate gel, showing that crystal lization of the gel takes place at around 970°C. Heating rate was 10°C per m in u te ......................................................................................... 125 Figure 5.2. DTA curve o f diphasic gel. The exothermal peak at 1328°C indicates the occurrence o f m ullitization............................................................... 126 Figure 5.3. XRD patterns showing that in conventional firing, mullitization of the diphasic aluminosilicate gel has not started at 1300°C. Mullite shows up in the sample fired at 1325°C........................................................ 127 XRD patterns showing that at 1250°C, crystallization of the diphasic aluminosilicate gel starts when subjected to microwave irradiation 128 Figure 5.4. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. xv LIST OF FIGURES (continued) Page Figure 5.5. Dilatometry curves of (a) monophasic and (b) diphasic mullite xerogels heated at 3°C /m in...................................................................................... 139 Figure 5.6. Microstructure of mullite ceramics step-sintered at 1200°C for 30 min. followed by 1500°C for 20 min. by (a) microwave and (b) conventional 141 m ethod. ............................................................................................... Figure 5.7. Microstructure of (a) as-sintered surface and (b) fracture microstructure of a diphasic mullite sample microwave sintered for 20 min. at 1300°C to relative density o f 94% ................................................................... 142 Microstructure of conventionally sintered (1250°Cx30 min. + 1500°C x30 min.) specimen of (a) pure and (b) seeded (5%) diphasic mullite g e ls ................................................................................................................... 143 Figure 5.8. Figure 5.9. Mullite ceramics sintered by (a) microwave (1300°Cxl0 min.) and (b, c, d) conventional (b, c: 1300°Cxl0 min.; d: 5°C/min. to 1320°C with no holding) methods. Sample (c) was mode of the diphasic xerogel and opaque, while the rest were made of aerogel and transparent.......... 144 Figure 5.10. Dilatometry curve o f diphasic mullite aerogel heated at 3°C/min........... 145 Figure 5.11. Densification curves o f diphasic mullite gel under (a) microwave and (b) conventional conditions................................................................... 148 Figure 5.12. Comparison in density o f diphasic mullite gel sintered by conventional and microwave processing at various tem peratures............................. 149 Figure 5.13. Arrehenius curve o f diphasic mullite gels sintered by (a) microwave and (b) conventional m ethods................................................................ 150 Chapter 6 Figure 6.1. Morphology o f the alumina sintered by microwaves at 1500°C for 20 m in ......................................................................................................... 156 Figure 6.2. A typical heating curve of CSZP ceramic in a microwave field............ 159 Figure 6.3. XRD patterns of CSZP (a) before and (b) after microwave sintering at 1300°C fo r 30 m in .................................................................................. 160 Figure 6.4. Influence o f compaction pressure (green density), sintering temperature and time on the sintered density o f CSZP. ............................................ 161 Comparison of the sintered density of CSZP ceramics processed by microwave and conventional methods at 1300°C................................ 162 Figure 6.5. R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission. xvi LIST OF FIGURES (continued) Page Figure 6 .6 . Micrographs o f CSZP sintered at 1300°C by (a) microwave sintering for 30 min. and (b) conventional sintering for 24 h............................ 163 Figure 7.1. The arrangem ent inside the m icrowave oven..................................... 169 Figure 7.2. A typical heating curve for microwave processing of HAp (HAp-F.) and HAp based composites, where porous zirconia was used as accelerator. ....................................................................................... 171 The microstructure of microwave sintered HAp/PSZ composite, showing the simultaneous sintering o f HAp and PSZ particles 176 Chapter 7 Figure 7.3. Figure 7.4. The microstructures of the fracture surfaces o f the HAp/SiCw composites sintered by (A) microwave processing and (B) hot pressing (perpendicular to the compaction direction)......................................... 177 Figure 7.5. Heating temperature as a function of time for the zirconia-containing composites based on (a) alumina and (b) mullite, (c) CSZP, and (d) B Z P S ...................................................................................................... 182 Figure 7.6. Densities of AI2 Q 3/PSZ composites at various mixing proportions. The alumina powder is (a) amorphous (RP-alumina) and (b) highly crystalline, high purity, crystalline alum ina............................................... 184 Figure 7.7. Sintered density of mullite/PSZ composites. Diphasic gel (a-c) and monophasic gel was used as mullite precursor, respectively. ................ 185 Figure 7.8. Sintered density of [NZPJ/PSZ composites using (a) CSZP and (b) BZPS as the [NZP] component, respectively. .......................................... 187 Figure 7.9. Microstructure of A1203/PSZ (30%) composite sintered by microwave processing at 1500°C for 20 m in............................................................ 189 Figure 7.10. Microstructure of mullite/PSZ (30%) composite sintered by microwave processing at 1200°C for 30 min. followed by 1500°C for 20 m in 190 Figure 7.11. Microstructure of CSZP/PSZ (10%) composite conventionally sintered at 1500°C fo r 20 m in .............................................................................. 191 Figure 7.12. Microstructure of BZPS/PSZ (20%) composite sintered by microwave processing at 1500°C for 20 m in.......................................................... 192 R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission. xvii LIST OF FIGURES (continued) Chapter 8 Page Figure 8.1. Alternate paths for matter transport in the initial stages of sintering 199 Figure 8.2. Schematic showing the energy barrier in the sintering process. As the free energy (chemical potential) o f starting material increases from A to A l, A 2 ,..., the driving force for the process increases from AG to A G 1,..., and the activation energy required for the process correspondingly decreases from Ea to E a l, E a 2 , ..................................201 Figure 8.3. Schematic showing of an amorphous structure..................................... 205 Figure 8.4. A model of the diphasic mullite gel system, showing that 8-A1203 particles are homogeneously dispersed in the silica gel........................ 210 Figure 8.5. Schematic presentation of (a) ordered crystalline silica and (b) random netw ork glassy form o f the same com position..................................... 211 Figure 8 .6 . Arrhenius plot o f the rates of densification and grain growth................. 216 R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. xviii L IS T O F T A B L E S Chapter 2 Page Some key characteristics and features of microwave processing.............. 21 Ca/P ratios of HAp samples determined by atomic absorption spectroscopy. ................................................................................................. 58 Table 4.2. Spectrochemical analysis of chloride ions in HAp samples (in ppm ) 58 Table 4.3. Lattice parameters of the HAp synthesized in the current study, A 64 Table 4.4. The main phase composition appearing in the HAp samples heated in air of 50% relative humidity at various temperatures........................... 70 Table 4.5. Intensity ratio of a-T C P (d = 2.909A) to p-TCP (d = 2.880A) in Table 2.1. Chapter 4 Table 4.1. the p-TCP after firing under different conditions for temperature m easurem ent c a lib ra tio n ............................................................................... 79 Table 4.6. Grain size of HAp pellets sintered under different conditions, p m .............. Table 4.7. Experimental Conditions of the Microwave Processing to Fabricate P orous HA p ceram ics................................................................................... 96 Table 4.8. Relative Density of Microwave-Processed Porous HAp Ceramics, %....... 98 Table 4.9. Processing conditions and results for microwave and conventional sin terin g o f H A p ............................................................................................ 108 Table 4.10. The InK values from the densification curves for Arrhenius plot at relative density 60% ................................................................................... 88 116 Chapter 5 Table 5.1. Mullite contents (%) in the conventionally heated diphasic mullite gel 129 Table 5.2. Mullite contents (%) in the microwave heated diphasic mullite gel 130 Table 5.3. Relative density (%) of monophasic mullite gel sintered at 1200°C x 20 m in ...........................................................................................................135 R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission. xix LIST OF TABLES (continued) Page Table 5.4. Comparison of sintered density of different starting materials (microwave sintering: 1300°C x 10 min. + 1600°C x 10 m in.).................................. 135 Table 5.5. Density of mullite conventionally sintered (1550°C x 20 min., 3h total h e a tin g ).............................................................................................................. 136 Table 5.6. Relative density of the seeded diphasic mullite gels after sintering (%)..... 136 Table 5.7. Sintered density of seeded diaphasic mullite gel (microwave sintering)...... 137 Table 5.8. Relative density of diphasic mullite gel after microwave step sintering (1200°C x 60 min. + 1500°C x 20 m in.)............................................... 137 Table 5.9. Density of unseeded diphasic mullite gel under various microwave sintering sch ed u les..........................................................................................138 Table 5.10. Kinetics values for activation energy calculation for diphasic gel 150 Chapter 6 Table 6.1. Density data of RP-alumina before and after sintering (1500°C)............. 155 Chapter 7 Table 7.1. Table 7.2. Summary of the processing conditions and some results measured on the sintered HAp and HA p-based composites........................................ 173 Microwave enhancement in densities over the conventionally sintered sam ples, %...................................................................................................... 191 Chapter 8 Table 8.1. Alternate paths for matter transport during the initial stages of sintering.... 199 Table 8.2. Influence of processing temperature on microwave effect...................... 222 R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. xx PRINCIPAL SYMBOLS AND ABBREVIATIONS HAp = hydroxyapatite, Ca 10 (PO4 )6 (OH )2 CCD = calcium chloride dihydrate, CaCl2' 2H20 DCPD = dicalcium phosphate dihydrate, CaH P0 4 2H20 TCP = tricalcium phosphate, Ca 3 (P 0 4 ) 2 TeCP = tetracalcium phosphate, Ca4 (P 0 4)20 CSZP = calcium strontium zirconium phosphate BZPS = barium zirconium phosphosilicate PSZ = partially stabilized zirconia (tetragonal) XRD = x-ray diffraction SEM = scanning electronic microscopy TGA = thermal gravimetric analysis DTA = differential thermal analysis £* = € - ye” = £o(s. r - the relative complex permittivity £q = 8 .8 6 x 10- 12 F/m, the permittivity of free space e’r = the relative dielectric constant €’ef f = th e effective relative dielectric loss factor a = the total effective conductivity (S/m) 8 = dielectric loss factor / = frequency P = power absorbed per unit volume E = internal electric field X0= firee-space wave length D = the depth of microwave penetration H = permeativity jl0 = permeativity of ffeespace KIC = fracture toughness (type I load) R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. xxi ACKNOWLEDGMENTS The author would like to express his sincere gratitude and respect to his advisors Dr. Della M. Roy, Dr. Rustum Roy, and Dr. Dinesh K. Agrawal, for their encourage, patience, guidance, suggestions, and various helps throughout the course of this work. The author wishes to thank Dr. Paul W. Brown, and Dr. David J. Green, for their comments and suggestions for the revision of this thesis. The author also wishes to thank Dr. Jiping Cheng and Dr. Choon-Keun Park for their assistance in experimental work. The author is indebted to all the M ends at the Materials Research Laboratory and at the University who directly or indirectly helped the author with the work toward this thesis. The financial support from the National Science Foundation under research contract No. 8812824, which made the research work of this thesis possible, is greatly appreciated. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 1 Chapter 1 INTRODUCTION 1.1. General Statement Recent advances in microwave processing of materials has opened a completely new area of research in the materials science and engineering. W hile the early study started in the late forties by Sutherland, the eminent British spectroscopist who devoted considerable energy to the study of drying o f large ceramic bodies using W W II radar (Roy, 1988), the study got attracted more attention in the seventies and rapidly developed in the recent years. The first symposium on microwave processing of materials held by the Materials Research Society in 1988 marks the beginning of the worldwide professional study in the microwave processing of materials as a new area. Since then, several annual symposia have been held by the M aterials Research Society and the American Ceramic Society alternatively. Hundreds of papers have been published or presented in this area and more and more attention is drown. Variety of materials have been studied by microwave processing. The research areas include the basic theory of the interaction between microwave and materials, microwave sintering, joining, drying, and synthesis, microwave remediation of nuclear and hazardous waste, m odeling o f m icrow ave processing, system design, dielectric property measurement, and so on. The focus on the microwave processing is how to utilize the microwave energy as an alternative in materials processing to substantially save time and energy consumption, thus save the capital of the unit product, and to significantly improve certain properties of the products. M icrowave processing of materials is fundamentally different from conventional processing in its heating mechanism. In microwave heating, heat is generated internally by interaction of microwave and the material, while in conventional heating, the heat is generated out of the work piece by heating elements or or by combustion of fuels then R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 2 transferred to the work piece. Thus the temperature gradients in the microwave heating is just opposite that in conventional case. In addition, the heat generated in the microwave heating depends on the microwave absorbing efficiency o f the material, which is then dependent on the dielectric loss factor of the material. So that the heating efficiency in a microwave field is m aterial-dependent. Furtherm ore, microwave can penetrate most ceramic materials to certain depth, so that microwave heating is a volumetric heating process. These features m ake microwave processing to certain ceram ic materials potentially advantageous, such as high heating rate, short processing time, thus low processing cost; uniform volumetric heating, low grain growth to densification ratio, thus uniform microstructure and finer grain size, resulting in improved mechanical properties. 1.2. Statement of Problem Although microwave processing of various ceramic materials have been studied, the unexplored important materials are still numerous, and the research on the new materials or the improvement o f the existing materials is endless. For example, there is either no or very little reported data on microwave processing o f NZP (NaZr2 P 3 0 12) materials, HAp [Ca 10 (PO4 )6 (OH)2] and m ullite ceramics. On the other hand, although microwave enhanced sintering and densification is experimentally evident, the mechanism behind the observation is not at all clear. Thus both intensive and extensive studies on microwave processing o f materials are significant. About ten years ago, Roy (1985) and coworkers observed that silica and alumina gels can efficiently absorb m icrowave energy and get m elted even in a regular kitchen microwave oven. That was the first discovery that these ceramic materials can be quickly heated in microwave field. Microwave processing of numerous ceramic materials have been studied since then, but these are mostly crystalline materials. In contrast, research on microwave processing of noncrystalline or amorphous inorganic solids, such as gels, and the solids o f low crystallinity is very limited. By heating the wet m ullite gels in an insulated cavity of 2.45 GHz, 700-W ordinary microwave oven, Komarneni et al. (1988) obtained crystalline mullite from the single phase and diphasic mullite precursor gels after R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission. 3 microwave heating for 20-25 min. However, the mechanism behind the rapid heating of the aluminosilicate gels in the microwave field as observed by Roy and coworkers is still not clear till now. Mullite (3 Al20 3 -2 Si0 2) is one of the most commonly used refractory materials. It is suitable for high temperature structure applications because of its high melting temperature (1820±10°C), low therm al expansion coefficient, low thermal conductivity, excellent thermal shock resistance, and good creep resistance at high temperatures. Mullite-based glass-ceram ics have now em erged as candidates for high-perform ance packaging applications (Aksay, 1991). Low dielectric constant and optical transparency o f fine grained polycrystalline mullite has led to interest in potential applications as a host material as a solid-state laser activator (Aksay, 1991). High melting point and low solid-state ion diffusivity require the sintering of mullite to be conducted at very high temperature for a longtim e. Practical sintering of mullite from crystalline powder needs 1600-1700°C for at least a few hours. Reaction sintering from crystalline alumina and silica also requires similar high temperature. By sol-gel routes, mullite can be synthesized and sintered at lower temperatures. Due to its low dielectric constant (e=6.7), crystalline mullite is almost transparent to m icrowaves, so that it is difficult to sinter by m icrowave processing. However, a diphasic mullite gel is a system totally different from crystalline mullite before mullitization (to crystallize to mullite). Therefore, the behavior of the diphasic mullite gel during heating, sintering, and mullitization in the microwave field is different, and thus it seems possible to obtain desired mullite ceramics by m icrowave processing. For this reason, mullite sintering was conducted in the microwave field mainly using mullite gels as the precursors. HAp is a very important biomaterial showing the best compatibility, known thus far, with human tissues, both soft and hard. Since the 70’s, the research on the theory and engineering design of the bioceramics based on HAp has been widely conducted. HAp ceramics are now comm ercially available for non-load-bearing biomedical use. For example, Interpore200™ porous HAp granules for alveolar ridge reconstruction R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission. is 4 commercially available from Interpore International in the United States, while in Europe, OSPROVIT 0.8 is supplied in granular form for repairing bone defects, particularly in periodontal surgery. HAp-coated metallic prosthetic implants have been used, the clinical application of pure or composite HAp ceramics in load-bearing implantation, however, has not yet been achieved, basically due to the intrinsic brittleness of this material in ceramic form. Some investigators attempted the transformation toughening o f HAp ceramics by using partially stabilized zirconia (tetragonal, PSZ) (Wu, 1988; Costa, 1993) but failed to achieve high density by conventional sintering, simply because of the differential sintering behavior o f HAp and zirconia. A case in which dense HAp/PSZ composite was achieved by hot isostatic pressing (HIP) was reported by Yoshimura et al. (1989). Since microwave processing enhanced densification of several ceramics (Janney and Kimrey, 1991), it is worthwhile to see its feasibility for the fabrication of HAp/PSZ composite, as well as various HAp ceramics at ambient pressure. [NZP] is a large structural family of low thermal expansion materials named after the parent composition of sodium zirconium phosphate, NaZr 2 P 3 0 12 (Alamo and Roy, 1984; Roy et al., 1984; Agrawal and Stubican, 1985). These materials are expected to be used for superionic conductors, radwaste m anagem ent, and for low therm al expansion applications. The sintering of these materials requires very long time at high temperatures. No study has been carried on microwave sintering o f NZP materials thus far. Alumina is one of the most important high temperature ceramics and has been used as model material for microwave processing by many researchers (Sutton, 1989). Usually crystalline alumina was used for sintering study. There is no report on the microwave sintering of alumina with amorphous precursors, which may behave quite differently from the crystalline alumina powder in the microwave field. As the development o f science and technology put more and higher demands to materials science and engineering. Just like a more advanced scientific research program usually needs the cooperation of more people, the achievement o f high performance of materials for advanced applications often requires combination of the properties of more than one component. And the higher the performance, the more complex the composition. In this sense, the route of composite is a general trend in the development of new materials. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 5 A composite is fabricated either for the desired mechanical or functional purpose. The im proved properties can be usually achieved by tailoring the com ponents through com prom ise. The combination sometimes can even create new properties. In the composites for improved mechanical properties, “brittle A (matrix) + tough B (additive) = moderately tough C” is an example of compromise, while “brittle A + brittle B = tough C” is a example for the case of creation. Ceramics are intrinsically brittle, thus composite is a new approach to make them tougher. Sintering is a necessary step in the fabrication of ceramics and ceramic composites. Different materials often behavior differently during the sintering. The existence of significant differentiate sinterabilities will certainly hinder the achievement of successful sintering of the composites. This proves to be a problem in the fabrication o f composites by conventional sintering. Fortunately, this may be the right place for microwave processing to show its advantages, since microwave heating has a selective heating effect to different components in micro-scale. For this reason, microwave processing of composites was conducted in this study. Metastable zirconia in the tetragonal phase, also known as partially stabilized zirconia (PSZ), has been investigated for transformation toughening (Lange, 1982; Green et al, 1989) in ceramics fabrication. Since zirconia is a good microwave absorber at elevated temperatures, the zirconia-containing ceramic composites were chosen as a part of this microwave processing study. 1.3. Objectives and Scope of Study The main objectives of this work are to evaluate and understand the microwave effect on the sintering, densification and other related phenomena during microwave processing of the selected ceramic materials, and to explore the possibility to fabricate the ceramics with improved properties and/or energy and time saving by microwave processing. The research involves the synthesis, processing, and characterization of various ceramic materials of general interest in ceramic industry. Synthesis techniques include solid state reaction, wet chemical reaction, sol-gel, hydrothermal, and ultrasonic methods. P rocessing includes fabrication o f various ceram ics and ceram ic com posites by conventional and microwave methods. Characterization covers various testing techniques R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 6 on composition, crystal structure, density, porosity, morphology, microstructure, thermal properties, m echanical properties, etc. T he m aterials studied include HAp, [Ca 2Q(P0 4 )6 (0 H) 2 ], mullite (3 A l 20 3 .2 Si0 2), alumina (A12 0 3), zirconia (Z r0 2), and two low-thermal-expansion [NZP] materials, i.e., calcium strontium zirconium phosphate (CSZP) and barium zirconium phosphosilicate (BZPS). Microwave effect on the sintering, densification, and crystallization of these materials was evaluated. The processing parameters that influence the microwave effects, and the mechanisms related to the microwave enhanced sintering or densification of the all selected materials have been discussed. The emphasis o f the study was on the synthesis and microwave effects on the selected m aterials during sintering. M echanical property m easurem ents were only evaluated in a few lim ited cases. Dielectric properties are important in microwave processing study. The measurements of dielectric properties at high temperatures constitute another branch in the research on microwave processing of materials, thus they are not included in the current study. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 7 Chapter 2 LITERATURE REVIEW 2.1. Microwaves Microwaves are electromagnetic waves ranging from 1 m to 1 mm in wavelength at frequencies from 0.3 to 300 GHz. The most commonly used frequency for processing foods and ceramics is 2.45 GHz with a corresponding wavelength of 122 mm. Figure 2.1 shows the microwave band in the electromagnetic spectrum. M ost of the microwave frequency band is used for communications and radar. The Federal Communications Commission (FCC) has allocated 0.915 MHz, 2.45, 5.85, and 20.2-21.2 GHz for industrial, scientific, and medical use. Presently, only 915 MHz and 2.45 GHz microwaves are widely used in industrial and medial application since these uses are based on water heating. A few other frequencies such as 28, 60, 140 GHz, and 500 M Hz are available on a lim ited basis, because they are used as power sources in accelerators plasma fusion devices, and commercial broadcasting (Katz, 1992). Figure 2.2 shows some m icrowave frequencies and wavelengths used in processing of materials (Sutton, 1993). Microwaves possess certain useful characteristics, one of the most important being that microwave wavelengths are of the same size as any structure used to guide or enclose them (Fuller, 1990). M icrowave pulses can be very short so that they can be used for distance or tim e m easurem ent. com puters. Also this makes them compatible with high speed The high frequency of m icrowaves offers very large bandwidths for communication links. Microwave radiation penetrates fog and clouds, travels in straight lines and gives distinct shadows and reflections enabling it to be used for distance and direction m easurem ent and in radar systems. Use o f m icrowaves is necessary for communication with satellites because they can pass through the ionosphere which reflects lower frequency radio waves. Microwave energy is absorbed very efficiently by water or any material containing water so that microwaves can be used for heating and drying. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. o Microwave ■o TJ CO ■H Frequencies used In ceramics processing n if) (0 o ■D CO O i_ (So O-D 12 n CO I - T3 n 0.915 |2i45l 28 53 60 84 100 - GHz 330 a>£go,f> CQ 122 10.7 5.7 5.0 3.6 3.0 • mm U .0 0 -OIL I g o z > ll£ > 1 1 (f) = 10 MHz 100 MHz 1 GHz MF HF VHF UHF 1000 m 100 m m T " 1 MHz 10 m 1 m — i— 100 GHz 10121014 EHF SHF 10 cm 300 GHz 1 cm 1 mm 1 p.m Figure 2.1. Electromagnetic spectrum showing the band of microwaves. (Source: Sutton) 00 9 1000 (GHz/ mm) (0.915/328) A Industrial (Ukraine) ■ Industrial (U.S.) + RSD • RSD (2.45/122) 100 Wavelength (k.mm) (6 /5 0 ) (28/11) » (37/8) 0.1 0.1 0.3 1 3.0 10 30 100 < ------------------- Micro waves 300 1000 10000 W Frequency (GHz) Figure 2.2. Microwave frequencies and wavelengths used in materials processing. (Source: Sutton) R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 10 M icrowaves are becoming m ore and m ore widely used. The applications include broadcasting, communication, radar, heating, moisture measurement, power transmission. Microwaves are also potentially hazardous due to their internal heating effect (Bruce, 1988; Fuller, 1990). The effect may not be felt until damage has already been dome because the heating may be internal whereas our body is designed more to warn us about externally applied heat. From heat balance consideration of standard man in standard conditions, 100 W/m 2 (10 mW /cm2) is considered to be the safe upper limit even during infinite exposure because thermoregulatory systems compensate for any power absorption. A power level of 10 W /m 2 is considered to give no heating effect even under adverse conditions of ambient temperature and humidity (Fuller, 1990). There is also some evidence of a athermal effect through the nervous system, although the effect is hard to prove and controversy still surrounds it. It is claimed (Fuller, 1990) that exposure over a period of years to pow er levels greater than 2 W /m 2 can lead to nervous system disturbances, although occupational exposure of healthy adults to this power level seems to have no adverse effects. However the probable safe level for continuous exposure of the general population ought to be even lower. Different countries have different recommended limits, ranging from 100 W/m 2 to 0.01 W/m2. 2.2. Microwave Processing of Materials 2.2.1. Microwave-Material Interactions In contrast with visible waves, except for lasers, microwaves are coherent and polarized. Microwaves also obey the laws of optics and can be transmitted, absorbed, or reflected , depending on the material type. As shown in Fig. 2.3. (Sutton, 1989), metals are opaque to microwaves and thus are good reflectors. Ceramic dielectric materials (e.g., AI2 Q 3, MgO, S i0 2, most glasses) are transparent to microwaves at ambient temperature, however, when heated to above a critical temperature, these materials begin to absorb R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 11 M aterial ty p e P e n e tra tio n TRANSPARENT (Low lo ss insulator) T otal OPAQUE (C onductor) None (R eflected ) ABSORBER (Lossy in su lato r) P artial to T otal ABSORBER (Mixed) (a) Matrix = low lo s s in su lato r (b) F ib e r /p a rtic le s /a d d itiv e s = (ab so rb in g m a te ria ls) P artial to T otal A/VWVW /VA/Wwv*- Figure 2.3. Classification of microwave-material interactions. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 12 microwave energy more efficiently. Other ceramic materials (e.g., Co 2 0}, M n 0 2, NiO, and CuO) absorb microwaves efficiently at room temperature. Basic microwave-material interaction mechanisms depend strongly on the dielectric and magnetic properties o f the material. This causes a strong dependence o f microwave absorption on frequency, material particle size, shape, temperature, and density (Tinga, 1988). The interactions between microwaves and m atter are many. Among them, four important polarization mechanisms are in solids, and three of them lead to losses in the microwave region: space charges arising from localized electrical conduction, rotating electric dipoles, and ionic polarization associated with far-infrared vibrations (Newnham, 1991). Conduction losses from localized space charge effects are responsible for microwave loss in transition metal oxides with mixed valence states. Alkali ions cause microwave loss in glasses and other silicates, especially when loosely coordinated in a m anner allowing considerable rattling room. Oscillating Na ions can be pictured either as localized conduction or as reorientable dipoles. Organic compounds and polymers are useful in ceramics processing. Dipole relaxation spectra are common in water, alcohol and other polar liquids, leading to intense energy absorption in the microwave regions. Polymers also absorb in this region of the electromagnetic spectrum with smaller segments o f the polym er chain resonating at higher frequencies and continuing to low er temperatures. Localized resonances caused by piezoelectric electromechanical coupling can also cause microwave loss. And, not all microwave losses take place through the electric vector of the wave. The magnetic field vector can initiate energy transfer through magnetic resonance effects in which the unpaired electron spins precess about internal fields. Kenkre (1991) developed a theory based on Debye relaxation o f interstitials, vacancies and bivacancies to describe the absorption o f microwave energy by ceramic materials. 2.2.2. Dielectric Properties of Ceramic Materials The the absorption of microwaves by a dielectric material depends on the m aterial’s R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission. 13 complex permittivity, £* (F/m), which is composed o f a real part ( e \ dielectric constant) and an imaginary part (e”, dielectric loss factor) by e* = e f- / e" = eo ( £ r - // j £e ff) where j = (-1)1/2, £o is the permittivity o f free space (£^=8.86xl0 ' 12 F/m), e’r is the relative dielectric constant, and e”eJyis the effective relative dielectric loss factor. W hen microwaves penetrate and propagate through a dielectric material, the internal electric fields generated within the affected volume induce translational motions of free or bound charges (e.g., electrons or ions) and rotate charge complexes such as dipoles. The resistance o f these induced motions due to inertial, elastic, and frictional forces , which are frequency dependent, causes losses, and attenuates the electric field. As a consequence of these losses, volum etric heating results (Sutton, 1989). The loss tangent (tan 8 ) is commonly used to describe these losses as // ,an 5 = 5 # = — S _ er 2jtfeoer where a is the total effective conductivity (S/m) caused by conduction and displacement currents a n d /is the frequency (GHz). The power absorbed per unit volume P (W/m3) is expressed as: P = c r |£ |2 = 2jif so e ’rtan8l£l2 where E (V/m) is the magnitude of the internal field. This equation shows that the power absorbed varies linearly with the frequency, the relative dielectric constant and tanS, and varies with the square of the electric field. When the microwaves penetrate and propagate through and absorbing material, the electric fields are attenuated. The penetration depth, at which the incident power is reduced by one half, is described by: R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 14 D = ---------- 3A° 8.6867rtanW e^/eo where A0 is the incident or free-space wavelength. It is evident that for greater wavelengths (and lower frequencies), the penetration depth is also greater. For example, the penetration depth of water is 7.6 cm at 0.915 GHz (X0 = 32.8 cm) but 1.3 cm at 2.45 (Aq = 12.3 cm) (Sutton, 1989). GH z Although low frequencies result in greater penetration depths, the heating does not necessarily increase, since the internal field E could be low, depending on the properties of the material. When the wavelength (frequency) is specified, the higher the tan 8 (the more lossy the material), the smaller the penetration depth. The penetrating ability of microwaves in the material is also expressed by the skin depth (SD) which is defined as the depth at which the electric field falls to 1/e or 37% of its initial value (Newnham, 1991; Katz, 1992): SD = l/( 7t/|icr ) 1/2 where / is frequency, p. is magnetic permeability, and a is conductivity. For copper, a = S .S x lO ^ ^ m -1, |X=(J.o, where |i 0 is the permeability of free space, and the skin depth is less than 1 pm at microwave frequencies. Since field penetration is proportional to <r1/2, microwave energy heats semimetals and semiconductors better than copper (Newnham et al., 1991). The relative dielectric constant (e?r) and the loss tangent (tan 8 ) are the two most widely used and measured parameters that describe the behavior of a dielectric material under the influence of a microwave field. They both affect the power absorbed and the half-power depth, and thus, they influence the volumetric heating behavior of a given material. The value of efr is a measure of the polarizability of a material in an electric field, whereas the value o f tan 8 is a measure o f the loss (or absorption) of the microwave energy within the material. The values of s ’r and tan 8 increase with temperature during heating. The increase in R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 15 e?r with tem perature is due to an increase in the polarizability caused by volumetric expansion. (Ho, 1988). For example, A12 C^ has a significantly greater coefficient of thermal expansion, which leads to a greater increase in its e’r . In addition, the composition and the density have a m ajor influence on the value o f € r , higher density tends to give a greater € r . Figure 2.4. shows the influence of temperature on dielectric constant (Sutton, 1989). In contrast to ef r , tan 8 is far more affected by temperature. In general, it rises slowly with increasing temperature, until a critical point is reached, beyond which it rises very rapidly. Compositional additives and impurities usually account for the rapid rise in tanS. Figure 2.5. shows the influence of temperature on dielectric loss tangent. It was reported that the rapid increase in tan5 in polycrystalline ceramics is associated with the softening of intergranular and amorphous phases, which caused an increase in the local conductivity, a 2.2.3. Characteristics of Microwave Heating Microwave heating is fundamentally different from conventional heating in its heating mechanism. Due to the penetrating nature of microwaves radiation, the microwave heating is an internal dielectric heating. In microwave heating, heat is generated by microwavematerial interaction within the material instead o f being transferred from the surrounding heating elements as in the conventional case. Because o f the penetrating nature o f the microwaves, the heating is a bulk (volumetric) effect. That is, the whole volume o f the work piece within the limit of microwave penetration is heating simultaneously. W hen the work piece is large in size, the thermal gradients of the material being microwave heated are reversed that of conventional heating. In other words, the center of the work piece being microwave irradiated is hotter than the surface before thermal equilibrium is achieved. Figure 2.6. shows the heating pattern in conventional and m icrowave heating. The penetrating nature o f the microwave energy also implies a quick response between the input power and the temperature of the work piece thus the thermal inertia is much lower than in R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission. 16 Temperature (°C) 200 12 600 “i i 1000 1400 r » w 11 pure co 10 I - AIjO, ** m c 9 o C o U O © © © 8 7 „ ^ Pyrolytic Si,N4 j~ 3-15 9/cm3 6 - .> 5 © GC 97% pure AljOj •“ 9606 (Glass-Ceramic) mA>c b n 0 *-45 ^ cm i ot-pressed 4 Fused silica * 3 1000 g/cmJ 2.06 g/cm* 1.94 a/cm* Si*N* i 20% ((porosity) slip cast 200 0 3000 Temperature (°F) Figure 2.4. Relative dielectric constant (8-10 GHz) versus temperature (Source: Walton). R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 17 Temperature (°C) 200 ^ *0 1000 1400 0 .0 1 0 97% pure AljOj Pyrolytic SijN« c £ 600 0.008 SiO; c g) 0.006 C Z i/i 3 9606 (G lass-ceram ic) 99 + % pure 'A I 2O a 0.004 0 .0 0 2 Pyrolytic BN 0 500 1000 1500 200 0 2500 Temperature (°F) Figure 2.5. Loss tangent (8-10 GHz) versus temperature (Walton, 1970). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 18 C o n v e n tio n a l Heating element M icrow ave Microwave port Sample ulatlon Insulation Furnace Metal shell Cavity Figure 2.6. Heating pattern under conventional and microwave heating conditions. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 19 conventional processing. In other words, there is no thermal lag during microwave heating. This makes it possible to heat the work piece at much higher heating rates (10lOOx) than conventional process. The rapid sintering leads to limited grain growth and fine microstructure, consequently improved mechanical properties. In contrast to the heating process, cooling after microwave processing is basically the same as in conventional case. Since microwave heating is a dielectric heating, microwave absorbing efficiency, and thus the heating rate, is dependent on the dielectric losses of the material. For a dielectric material, the dielectric constant and loss tangent are both a function of temperature. The increase of the dielectric loss tangent is usually steeper than that of dielectric constant. The phenomenon o f thermal runaway is a typical characteristic frequently observed in microwave heating of materials. It is characterized by a steep temperature rise in the materials when heated in a microwave field. As the microwave power absorption is directly proportional to the relative dielectric constant £?n and the dielectric loss tangent tanS, an increase in these two parameters will lead to the proportional increase in the temperature of the work piece being processed. The higher temperature further increases the microwave absorption and, thus the net effect is an exponential increase in temperature. This “snow-ball” effect in temperature increase is termed as “thermal runaway”. Figure 2.7. shows a temperature-time curve with thermal runaway during microwave heating of zirconia in this study. The shape of the heating curves with thermal runaway may vary with the type of material (Kenkre, et al., 1991). Thermal runaway may benefit materials processing in some cases (Fang et al. 1991), but in some cases, it may cause thermal gradients which may lead to cracks and inhomogeneities. Therefore, precautions should be taken to control and monitor the thermal runaway. Research on microwave processing of materials to date has involved drying, curing, binder burnout, calcining, sintering, melting, joining, surface modification and waste remediation, etc. Some o f the key characteristics of microwave processing are listed in Table 2.1. (Sutton, 1993). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 20 1200 900 CJ a>" = 4-> ro 600 c_ OJ a. e CD h— 300 30 0 Time, min Figure 2.7. Thermal runaway during microwave heating. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 2.1. Some key characteristics and features of microwave processing (Sutton, 1993). Characteristic 1. Penetrating radiation, direct bulk healing Feature Benefits (over conventional heating) • Materials heat internally • Reversed thermal gradtents (AT) • Lower surface temperatures • • • • • • Potential to heat large sections uniformly AT favors CVI; matrix infiltration Reduced skin effect on drying Removal of binders a gases without cracking Improved product qualify and yields Materials &composite synthesis Automation, precise temp, control Rapid response to power level; pulsed power Challenqes • M/W transparent mat'ls difficult to heal • Hot spots, cracking • Large AT in low thermal condudivily mal'is, and nonunilorm heating • • Instantaneous power/lemp. response Low thermal m ass • • • Applicator can be remote from power source • Heat in d e an environment • Heat under conlrdled vacuum or pressure and gas composition • Materials synthesis • S ee differential coupfing • Arcing, plasmas • Require new equipt. designs special readion vessels • Controlling internal temp. 2. Field cSstributions can be controlled • High energy concentration • Optimize power level vs. time • mm-waves can b e focused or defocused, rastered a s desired • Precise heating ol irregular sections • Precise healing ol selected regions (brazing, welding, plasma generation, Ib er drawing) • Process automation, flexbitty, energy saving • Synthesis ol mat'ls, composites, powders, coatings • Equpment more cosily and complex • Requires specialized equipt. 3. Dielectric losses accelerate rapidly above TCI| • Very rapid healing • • • • • Hot spots, ardng • Nonunilorm temperature • Control of thermal runaway 4. Differential coupling ol materials • Selective heating of internal or surface phases, additives or constituents • Healing of MIW transparent mat'ls via additives, fugitive phases, etc. • Hybrid heating (adive containers) • Materials synthesis • Selective zone heating (joining, brazing, sealing) • Controlled chemical readions, oxidation, redudion; use ol M/W transparent containers • Drying, curing, annealing; matrix inlillralion • Readions with unwanted impurities • Contamination with insulation or other phases 5. Self-limiting  • Selective healing ceases (self regulating) after certain processes have been completed • Below critical femp., drying & curing are sell-regulating • Completion ol certain phase changes is self-regulating • Undesired decoupling during healing in certain produds • Difficult to maintain temp. Rapid processing (2 - 1000x lactor) Heat mal'is > 2000°C Density materials with min. grain growth Capable lo heat MfW transparent materials > Ta , 22 2.2.4. Historical Review of Microwave Sintering of Materials Research on microwave processing of ceramics materials has been carried out for 40 years since Von Hippel (1954) began the examination o f the interaction between microwaves and oxide materials, and how the loss characteristics of the materials vary as a function o f microwave frequency. M uch o f his theory and some data on microwave material interactions provided a valuable background for microwave heating. Soon after, Feiker (1959) reported that dielectric materials could be heated rapidly by microwaves at 915 MHz. Ford (1967) conducted high-temperature chemical processing via microwave irradiation. Among the pioneers, Tinga (1968,1970,1973) and coworkers studied the interaction of microwave and materials, the thermal processing o f oxides and composite materials using m icrowaves of 915 M H z and 2.45 GHz, and carried out dielectric property measurements in the microwave field, and developed reaction cavities for a uniform microwave heating. It has been reported that microwave processing of ceramic materials takes place at accelerated rates or lower temperatures than in conventional processing. For example, the examination of the rates o f sintering for microwave and conventional heating by Janney (1991) and coworkers reveals that the apparent activation energy for sintering of MgOdoped alumina in a 28-GHz microwave furnace was less than one third that observed for conventional sintering in a conventional furnace (160 vs. 575 kJ/mol.). They also found that the grain growth in dense, hot-pressed alumina is structurally similar for microwave and conventional annealing conditions. In both cases, normal grain growth is observed and the m icrostructural param eters such as grain shape, anisotropy, and grain size distribution are similar and cubic grain growth kinetics are followed. However, annealing in a 28-GHz m icrowave furnace accelerates the growth kinetics significantly: e.g., m icrowave grain growth kinetics at 1500°C are the same as conventional grain growth kinetics at 1700°C and nearly two orders of magnitude higher than conventional grain growth kinetics at 1500°C. The activation energy for grain growth was about 20% lower for microwave annealing as compared with conventional annealing (480 vs. 590 kJ/mol.) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 23 which probably represents an acceleration of diffusional processes. They showed that a “microwave effect” can manifest itself in a dense ceramic body and that no free pore-solid interface is necessary. M eek (1987) has proposed a model in which the surfaces of the powder particles and pores are heated preferentially in the sintering compact. And Holcombe etal. (1988) carry the idea of local preferential heating to the extreme in suggesting that the surfaces of pores might be heated to the melting point by microwave excitation. Johnson (1991) carried out the heat flow calculations of microwave heated grain boundaries in ceramics and concluded that the temperature of any given grain boundary cannot be significantly different from the local average temperature. So that some mechanism other than hot grain boundaries m ust be sought to explain the observed enhanced sintering. In contrast, Tinga’s model (1993) indicates that very large differential heating rates at the microstructural level are physically possible, although the implications of such rates are not at all clear. Booske et al. (1991,1992) have proposed that the observation of enhanced diffusion and sintering rates in microwave (versus conventional) sinterings are due to a athermal energy (or phonon) distribution induced by the intense microwave radiation field. This is most likely the case if the microwave radiation resonantly couples to specific elastic waves or normal modes of oscillation in the crystal lattice. The possibilities for coupling to normal mode oscillations appear greater in polycrystalline media (versus single crystal) due to additional effects associated with surfaces, grain boundaries, and concentrations of point defects. Alternatively, microscopic spatial nonuniformity in microwave heating may yield several populations of ions, each characterized by a different “temperature”. Averaged over a macroscopic region, this would appear as athermal energy distribution. The exact forms of the lattice phonon and ion energy distributions during intense microwave heating of ionic solids, and the exact m echanism s for m icrowave-to-phonon energy transfer during microwave sintering are unknown at this time. However, it is clear that point defect mobility (and thus sintering rates) is m ost sensitive to the high energy tail of the distribution, while m easurem ents with standard tem perature diagnostics (such as Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 24 thermocouples) are sensitive to “bulk” features of the distribution. It was shown that a small, 30% difference in the mean energy between the thermal and athermal distributions leads to orders o f magnitude difference in the probability of point defect motion. This is due to the distribution o f an excess energy in the athermal case in the high energy tail, where the probability is most sensitively affected. However, refined calculations (Freeman et al., 1993) indicate that the effect of athermal distributions of lattice ion energies caused by microwave fields is too small to explain the observed data unless the microwave strength approaches the lattice field strength. M ost microwave sintering experimental configurations have microwave field strengths less than or equal to 105 to 106 V/m, whereas lattice field strengths are ~10 1 1 V/m. More recently, the same researchers (Freeman et al., 1993) proposed that a quasi static polarization o f the lattice near a point defect as a possible mechanism that lowers the activation energy barrier for diffusion of that defect. The same researchers (Freeman et al., 1993) further indicated that microwave fields do not change microscopic activation energies for ionic mobility. Nevertheless, they have identified the existence o f a microwave effect on ion transport in crystalline solids, which seems to occur when there is a preexisting chemical potential gradient in the crystal. In summary, microwave enhanced sintering of various ceram ic m aterials are experimentally identified, but the mechanisms for explanation o f the observed microwave effect are still at a developmental stage. 2.2.5. Current Status of Microwave Sintering Technology 220.127.116.11. Microwave Applicators Microwave processing systems consist of a microwave source, an applicator, and the systems to control the heating. An applicator is a device (cavity) that provides a means for applying the microwave energy from the generator to the work piece to be processed. The principal purpose of the applicator is to focus the microwave energy to the material load, Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 25 and to provide impedance matching in order to efficiently transfer the microwaves into the load. Microwave cavities can be either single-mode or multimode. A multimode applicator is a cavity, typically a enclosed metal box as home microwave oven, in which several fundam ental standing waves or modes are superimposed to produce a standing wave pattern. The multi-mode applicator is m ost commonly used in microwave processing. Its main drawback is that the non-uniform electric field distribution may cause inhomogeneous heating. By using mode stirrer or moving the work piece through the field via a turntable, a more uniform heating pattern can be obtained. When microwaves propagate into a cavity, they establish a standing wave pattern. By properly choosing the dimensions of the cavity, a fundamental standing wave pattern can be produced. Such a cavity is said to be singlemoded. The single-mode applicator is used for precise and predictable control over the fields within the work piece. Such applicators maximize the field strength at the work piece location so that rapid and uniform heating can be achieved. However, the heating volume is limited, and homogeneous heating is difficult to achieve with large samples. The design of the applicator can also be changed to provide localized heating in specific regions of the work piece, which is desired in the case of joining o f materials, drawing glass fibers, or creating plasmas to rapidly sinter various ceramic materials. Single and multimode cavities may both be tuned to produce a desired mode pattern. Tuning is usually accomplished with the aid of a sliding shorting plunger, which changes the dimensions of the cavity. 2.2.52. New Developments Microwave processing of ceramic materials has received considerable attention over the past several years and its unique properties has opened up some new approaches to the fabrication of new ceramics with improved properties and/or lower cost. On the one hand, microwave study is being extended to more and more new material systems, on the other hand, scale-up of the microwave sintering techniques is just starting. However, due to some inherent limitations of the microwave process, most work has been confined only to Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 26 small laboratory samples. M icrowave hydrothermal process has been carried out by Li and Kom am eni (Li, 1993). Various fine ceramic powders, including unary, binary, and ternary oxides have been successfully synthesized by microwave-hydrothermal processing. The experimental results showed that the interaction o f microwaves with materials (solids, liquids, and interface) can lead to different morphologies and interface effect between solids and solutions. It was also shown that microwave-hydrothermal synthesis enhances the kinetics of crystallization of ceramic powders by one or two orders of magnitude compared to conventional-hydrothermal process. Rapid selective annealing of Cu thin films on silicon using microwaves was recently reported by Brain et al. (1994). Selective annealing of metallic thin films is possible in the lim it where the film thickness is much less than the skin depth of the m aterial at microwaves frequencies, with the additional advantage that the relatively low absorption of silicon deters substrate heating. It was found that as a silicon wafer is heated it becomes less microwave absorbing, independent of the doping concentration. M icrowave-assisted firing was reported by W roe (1993). In microwave-assisted firing the volumetric heating provided by the microwaves heats the components, while the mere conventional radiative heating provided by gas flame or electric resistance heating elements minimizes heat loss from the surface of the components, by providing heat to the surface and its surrounding. This prevents the creation of the temperature profiles/gradients associated with both conventional and microwave-only firing. Thus microwave assisted firing leads to producing uniform microstructures and low thermal stress. Blake, and Katz (1993) carried out microwave sintering of large alumina bodies. By using carbon as a setter material and a microwave susceptor, and low density aluminasilicate as insulation material, hexagonal tiles weighing 420 g were sintered to relative density of 93% by microwave processing. Based on the characteristics of volumetric heating of the work piece, high heating rates, less insulation, high efficiency of the heating process, possibly as high as 80-90% (Katz, 1992), lower processing costs are anticipated for microwave microwave sintering besides the improved product quality. The actual data for power consumption and cost for Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 27 ceramics for microwave sintering are very limited. Reported cases are by Katz and Blake (1991), and Patterson et al. (1991) for alumina ceramics, and Das and Curlee (1987) for common ceramics. If using fossil fuel or natural gas for conventional sintering, microwave processing w ould not be advantageous in cost, because o f the total efficiency of the conversion from fossil fuel to electricity then to microwaves, and finally to heat, is lower than that directly convert fuel to heat. This comparison seems not so valuable. A better comparison would be that between two techniques both using electricity. Compared with conventional electric furnace sintering, microwave sintering results in an energy saving of as much as 90% (Petterson et al., 1991). Apparently, more work should be done for an economic evaluation in microwave processing. As a new direction, continuous microwave sintering o f industrial sized work pieces has been developed by Cheng (1994). In this technique, the centrosymmetric work-piece is continuously fed in a high-power microwave cavity to subject a dynamic sintering process through simultaneous rotation and translation movements. M ullite and alumina rods and rollers o f <£40x2400 mm have been fabricated by the continuous microwave sintering. High quality, short processing time, and low rejection rate are typical features of this process. Although the transferring of microwave sintering technique is fraught with difficulty and commercialization o f this technology is still far off, and yet there is no sign that microwave sintering will replace all conventional sintering techniques, however, it is doubtless that the unique features o f microwave processing will find its applications in the areas where microwave processing shows definite advantages over conventional processes. It can be expected that the ongoing extensive and intensive studies on microwave sintering will promote both the understanding of the mechanisms behind the observed microwave enhanced sintering effects and the development of the microwave sintering techniques. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 28 Chapter 3 EXPERIMENTAL PROCEDURES 3.1. Starting Materials The materials used in this study include hydroxyapatite [HAp, [C a 10 (PO 4 )6 (OH)2], mullite (3Al 2 0 3 -2Si02), alumina (A12 0 3), calcium-strontium-zirconium phosphate (CSZP, CaQ 5 Sr 0 5Zr4 P 6 0 24), barium-zirconium phosphosilicate (BZPS, B aj 2 5 Zr4 P 5 5 Si0 5 0 24), and partially stabilized zirconia (PSZ, Z r0 2 -Y). Among these, HAp was synthesized by hydrolysis, precipitation, and hydrothermal methods; mullite precursors were synthesized by sol-gel methods; and the other materials were acquired from commercial vendors. The precursor powders were consolidated into pellets o f 0.5 or 0.25 in. diameter by uniaxial press for sintering. Polyvinyl alcohol (PVA) solution was used as the binder in some cases. The details of synthesis, sources, and preparation of the test specimens of various materials are given in the respective chapters. 3.2. Experimental Setup for Sintering The experimental setups for sintering include both microwave and conventional furnaces for comparison purpose. 3.2.1. Microwave Furnace M ost microwave sintering experiments were carried out in a multimode, 900 W (magnetron output power), 2.45 GHz microwave oven (Panasonic). The microwave cavity is about one cubic foot (14” x 14.75” x 8.25”) in volume. There is a glass-ceramic turntable at the bottom of the cavity. The microwave port is located on a side wall of the cavity. To make this kitchen microwave oven suitable for high-temperature sintering Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 29 experiments, following modifications were made: 1. To control the heating rate, a voltage regulating system consisting of two variacs was added to the power supply to the magnetron. 2. To efficiently cool the magnetron, the power supply to the fan was separated from that to the magnetron. This makes it possible to continuously cool the magnetron even if the power to the magnetron is off. 3. To maintain the cavity at low temperature, so that there is no excess heat transferred from the cavity to the magnetron, a hosepipe of copper was soldered onto the outside surface (under the outer shell) of the top of the cavity for water cooling. 4. An aluminum plate was placed on the turntable to prevent the glass-ceramic turntable from being overheated. 5. A sintering packet was made for efficient microwave heating. The sintering packet (Fig.3.1) consists of a crucible, zirconia cylinder (Zircar Products, Inc., Florida, NY), m olybdenum disilicide (M oSi2) rods, and therm al insulation D uraboards™ 3000 (Carborundum Co., Niagara Falls, NY). The crucible is also made of the same board. The Duraboard was made from a blend of Fiberfrax bulk ceramic fibers and Fibermax polycrystalline high alumina fibers. The typical chemical composition of the board is 70% A12 0 3 and 26% S i0 2 and 4-6% weight loss on ignition. As claimed by the manufacturer, Duraboards™3000 offers low thermal conductivity, high temperature stability and excellent resistance to thermal shock and chemical attack. It can withstand temperatures up to 1650°C (3000°F) for continuous use. The zirconia cylinder vertically surrounding the samples functions both as a microwave susceptor (microwave-absorbing material) and as a thermal insulator. As the microwave susceptor, it absorbs and converts microwaves to heat, thus preheats the samples if the samples do not absorb microwaves well at low temperatures. As an insulator, it prevents heat of the samples from dissipation. Molybdenum disilicide (M oSi2) is an excellent heating elem ent material used in electric resistance furnace, with maximum operating temperature of about 1870°C. MoSi 2 is resistant to thermal shock and can be subjected to extremely rapid heat-cool cycling for Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 30 Ka Figure 3.1. Schematic display of the sintering packet in the m icrowave processing. l.Tumtable. 2. Zirconia cylinder. 3. Samples. 4. Thermocouple. 5. Fibermax insulator. 6. MoSi2 rods. 7. Microwave port. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 31 prolonged periods of time with no adverse effects. Also, the resistance of these high temperature heating elements does not change with use, which allows for replacement of individual elements in the electric furnace at any time without having to match resistance values of new elements to the older. Although there is no previous report on the use of MoSi2 as heating elements in microwave processing, it has been found to be an excellent candidate materials for microwave heating. Eight molybdenum disilicide rods vertically inserted around the zirconia cylinder can trigger microwave plasma at lower temperatures and help homogenize temperature distribution around the samples, and lower temperature gradient in the vertical direction inside the zirconia cylinder. The rotation of the turntable during sintering ensures uniform heating conditions in the microwave field. The heating rate can be controlled fairly well by an external variac voltage-regulating system (course and fine control). In the sintering of HAp and CSZP ceramics, a different microwave furnace (500 W, JE43, GE) was used, in which there was not a turntable, instead, there was a mode stirrer for homogeneous field distribution. The configuration and the principle were the same. 3.2.2. Temperature Measurement in Microwave Processing Temperature measurement in microwave sintering heating is one area which is very critical and matter of concern for its accuracy and reliability. This is so because microwave enhancement in ceramics sintering is usually reported as the lowering of processing temperatures, or at the same temperature but improved properties, while only precise measurement can make the comparison between microwave and conventional processing methods reliable. Typical methods for measuring temperature in high temperature furnaces are radiation pyrometers, thermocouples, and optical fiber probes. Fiber optics and pyrometry methods are based on the spectra o f infrared radiation at high temperatures and are not interfered with microwaves. However, in the use of pyrometer, the emissivity value of the material is critical for an accurate temperature measurement, and fiber optics needs the use of a standard material. The sources for the data of emissivity are very limited, and the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 32 emissivity of a material depends on surface roughness, density, temperature, etc., hence pyrom etry is not a simple and direct method. Thermocouples are widely used in conventional high-temperature furnaces and have excellent reliability and accuracy. However, in a microwave field, the interference of microwaves with the electromotive force of the thermocouple makes the temperature readings random or erroneous before the temperature rises to such a level that microwaves are efficiently absorbed by the workpiece. When properly shielded and grounded, microwave interference can be totally prevented and thermocouples can provide reliable temperature readings. In this study, temperature was measured by a platinum-shielded S-type (Pt-PtlORh) thermocouple. The shielding Pt foil on the thermocouple is properly grounded to the m etallic wall of the microwave cavity, so that the microwave interference to the electromotive force can be completely avoided during the sintering. Since the samples are under consistent rotation mode, and the thermocouple is stationary, it is impractical to allow the tip of the thermocouple in contact with the samples. To solve this problem, a small piece of alumina or that o f the material being processed, is attached with the shielded tip of the thermocouple, so that the temperature measured by the thermocouple is the temperature of the material and not that of the air in the vicinity of the specimen. Furthermore, since the tip of the thermocouple is very close to the surface of sample, the measured temperature is identical or very close to that of the specimen. 3.2.3. Conventional Furnace An electric furnace, exclusively devoted for this study, was built with molybdenum disilicide as heating elements. The furnace is capable of high operating temperature (up to 1650°C), high heating rate, and low thermal inertia. The heating rates were comparable to that in microwave sintering. The furnace has the capacity of providing heating rates up to 300°C per min. below 1000°C, and can attain a temperature of 1500°C in 13 min. The temperature in the conventional sintering was measured with the same S-type thermocouple as in the microwave sintering, but without platinum shielding. The samples were placed in a platinum crucible which was in contact with the tip o f the thermocouple. The heating Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 33 rates and soaking temperatures were also controlled by regulating the power input with a variac. In the earlier experiments of this study on HAp, conventional sintering was carried out in a 5 kW programmable box furnace (Model 51524, Lindberg, Watertown, WI). Although temperature as high as 1400°C could be achieved, the heating speed in this furnace was lower than in the microwave processing. In the case this furnace was used, it will be specified. 3.3. Sintering Process M ulti-pellet runs were conducted in both microwave and conventional sintering experiments. In microwave sintering, the pellets, usually in two (0.25 inch pellets) to three (0.5 inch pellets) layers, were placed at the center of the sintering packet. Each layer has 6 7 pellets of the 0.25 inch diameter, but only one pellet of the 0.5 diameter. Heating rate in both microwave and conventional sintering was about 100°C per minute. Since heating and cooling rates are comparable, for the duplicate sample of the same material, when sintering time and temperature are identical in both case, the difference in the sintered samples produced by two methods should be attributed to the microwave effect. Because the sintering characteristics of different materials are different, the sintering temperature and heating schedule may vary from material to material. The details of the sintering conditions for each material will be given in the respective section/chapter. 3.4. Characterization A variety o f characterization tools were employed for the characterization o f the starting materials and the sintered specimens. The main techniques are described as the following. For those that are routinely used, the description and principle of the techniques are omitted for clarity. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 34 3.4.1. Powder Characterizations Specific surface area, particle size, and morphology o f the starting powders were determined by standard techniques. The specific surface area of the starting materials before com paction was m easured by BET m ethod using nitrogen absorption on an Autosorb-1 Automatic Volumetric Sorption Analyzer (Quanta Chrome, Syosset, NY). Particle size and m orphology o f the powders were exam ined by either scanning or transmission electronic microscopy. 3.4.2. Thermogravimetric Analysis Thermogravimetric analysis (TGA) was carried out on the synthetic HAp powders and the aluminosilicate gels to study the weight loss of these materials as a function of temperature (thermogram profile). Any reaction, decomposition, or phase transformation involving weight change will be recorded as a function of tem perature. The TGA experiments were performed on a Perkin-Elmer DELTA/TGA 7 system. The system is automatically controlled by a computer program. The heating rate for all the materials used in this study was kept at 10°C per minute. 3.4.3. Differential Thermal Analysis Differential thermal analysis (DTA) was carried out on the aluminosilicate (mullite) gels to determine the mullite transformation (crystallization) temperatures. The DTA used in this study was Perkin-Elmer DTA, which is also computer programmed. The heating rate used was usually 10°C per minute. The maximum operation temperature is 1500°C. The phase transformation process is accompanied by heat absorption or heat evolution, corresponding to endothermal or exothermal peaks on the thermograms. For example, the process of crystallization from amorphous state is an exothermal process, while the process o f melting is an endothermal process. From the peaks on the thermograms, the phase transformation temperatures can be identified. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 35 3.4.4. Dilatometry The thermal mechanical analysis (TMA) was carried out on selected specimens to study the dimensional change during heating. The TMA setup is a computer-driven system (Innovative Thermal Systems, NY) in which desired heating and cooling rates can be programmed. The maximum temperature is 1500°C. The heating rate of 5°C/min. was used in this study. The dimensional change as the percentage o f the original size is automatically recorded as a function of temperature, or alternatively, as a function of time. The range of sintering temperature can be conveniently determined from the TM A curve. 3.4.5. Powder X-Ray Diffraction X-ray diffraction was carried on the powdered materials for phase identification. The equipment used in this study is a computerized Scintag Diffractometer, Model V (Scintag Inc., Sunnyvale, CA). For qualitative analysis, the samples were ground to fine powders and placed on glass slides with acetone. For quantitative analysis, the samples were pressed in the cavity of the glass slide. Internal reference material such as quartz or corundum was used in the quantitative analysis. The continuous scan at 2° (two theta)/min. was used for both qualitative and quantitative analyses. 3.4.6. Lattice Parameter Refinement In the study of the thermal stability of HAp, crystal lattice parameter refinement was perform ed by XRD analysis on the Scintag D iffractom eter in conjunction with a computerized calculation program developed at the Intercollege M aterials Research Laboratory at The Pennsylvania State University, University Park, PA. 3.4.7. Density Measurement Since the specimens were in simple circular shape, green and sintered density of the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 36 pellets were m easured by w eight and geometrical m easurem ents. The weight was measured by an electronic balance with precision of ±0.0001 gram. The dimension was measured to an accuracy of ±0.0001 inch with a micrometer. In a few cases when the samples are irregular in shape, the density was measured using the wax-coated samples by Archimedes principle using kerosene or water as immersing liquid. 3.4.8. Mercury Intrusion Porosimetry The porosities o f selected sintered porous HAp ceramic samples were determined by a Quantachrome Scanning M ercury Porosimeter. The evacuated porous material to be detected is immersed in mercury. The mercury is then forced into the pores of the material under pressure. The pore size is inversely proportional to the pressure applied. Under the operation pressures up to 386.2 M Pa (56000 psi), the corresponding pore sizes are in the range from 7.5 (im to 1.8 nm. During scanning, the porosity, or volume of the intruded mercury, is recorded as a function of applied pressure. 3.4.9. Diametral Tensile Strength Measurement The mechanical strength of sintered regular and porous HAp ceramics were measured by diam etral m ethod (Rudanick et al., 1963). The pellets used for the strength measurement were sintered samples with green diameter of one half inch in diameter. Ordinary blotter paper was used as a padding material between the sample and each of the platens. Except where otherwise indicated, the crosshead speed was set at 0.002 inch/min. The maximum applied load at failure of the sample was recorded on a X-Y chart recorder. The diametral tensile strength (a), in MPa, was calculated by the following formula: o = 2 L 1%Dt where L is the applied load (N), D is the sample diameter (m), and t is the thickness of the pellet (m). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 37 3.4.10. Fracture Toughness Measurements Toughness is a m easure of the resistance o f a m aterial to fracture. Fracture toughness, also known as critical stress intensity factor, of a brittle material was calculated from the formula below, K c = JcPC3/2 where % is a calibration factor, for Vickers indentation, % = 0.016CE/H)1/2, P is the indentation load, and c is the size of the crack causing failure (Fig. 3.2.), E is Young’s modulus, and H is the hardness. The ratio E/H was determined on Knoop indentation using the formula (Marshall, et al., 1982, 1988; Chantikul et a l ., 1981): E/H = a (b/a - b’/a ) " 1 where a is a gradient (= 0.45) between indentation dimensions and the ratio E/H, a and b are the length of the major and minor axes of the indentation diagonal, respectively, in the loaded state, b ’ is the distortion length o f the residual impression of m inor axis after releasing the load. The indentation was conducted on the polished mirror planes of the selected sintered pellets to determine K jo 3.4.11. Scanning Electron Microscopy The morphology of the starting materials, the microstructures of the as-sintered surfaces as well as the fracture surfaces o f the sintered samples were observed with a scanning electron microscope (SEM, ISI-DS 130, Akashi Beam Technology Corp.). The average grain sizes of HAp ceramics were also measured from the scanning electron micrographs. Before the microscope observation, samples were dried and coated with gold by sputtering method to avoid electric charging. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 38 cracks i 2c Vickers indentation cracks T I JL ■H Knoop indentation Figure 3.2. Schematic showing of Vickers and Knoop indentations. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 39 3.4.12. Transmission Electron Microscopy D ue to the lim itation in resolution, the m orphology and particle size o f the hydrothermally synthesized nanocrystalline HAp could not be resolved by SEM, so that transmission electron microscopy (TEM) was performed on a Philips EM420 transmission electron microscope at an acceleration voltage of 120 kV. The high acceleration voltage substantially increased the resolution o f the image, which made the study o f the unique morphology o f the hydrothermally synthesized HAp successful. 3.4.13. Infrared Spectrometry To detect the existence o f the functional ions, especially hydroxyls, in the synthetic HAp samples, infrared (IR) spectrometry was carried out on the samples in KBr pellets in a Perkin-Elmer 283-B dipersive infrared spectrometer. The IR absorption o f the selected specimens was measured in the wave-number range from 200 to 4000 cm"1. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 40 Chapter 4 MICROWAVE SINTERING OF HYDROXYAPATTTE 4.1. Introduction 4.1.1. General Hydroxyapatite (also known as hydroxylapatite, HAp) is one of the minerals in the apatite [Ca 5 (P0 4 ) 3 (F, Cl, OH)] family. The name “apatite” is derived from a Greek apatao, deceive, since this mineral was long confused with other minerals (Milovsky, 1982). Apatite represents the most abundant o f the naturally occurring phosphatic minerals (McConnell, 1973), serving as the source for development of many phosphoms containing compounds and the production o f phosphatic fertilizers. During the past few decades, the use of apatite crystalline phases for fluorescent lamp phosphors (Johnson, 1962) has been a major development, and other solid state applications have been explored, such as use for single-crystal laser host materials. The third and perhaps the m ost important field for human kind are the biological apatites which in polycrystalline or semicrystalline form constitute the major inorganic matter of bones and teeth o f vertebrates. The above define principally the family o f phosphatic minerals and related synthetic compounds. In addition, apatite also is used to describe a structure type, the specific crystal structure having the space group symmetry P63lm possessed by the compound Ca 10 (PO4 )6F 2 , fluorapatite, or slight distortions thereof. Nearly half the periodic table may be substituted into the position o f one of the four ions Ca2+, P 5+, 0 2‘, or F", singly or in various combinations to yield m odified compounds or entirely different isostructural chemical species (Roy, 1978). The stoichiometric chemical formula o f HAp is C a 10 (PO4 )6 (OH)2. The crystal structure of HAp belongs to hexagonal system; theoretical density is 3.16 g/cm 3; M ohr’s hardness is 5. HAp crystals occur in variety o f morphologies: prismatic, tabular, granular, Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 41 reniform, stalactitic, fibrous, etc. HAp is brittle, fracture is uneven to conchoidal. 4.1.2. Application of HAp as a Biomedical Material In all the phosphate com pounds, HAp is the only one that has excellent biocompatibility without biodegradation. Actually, HAp shows the best biocompatibility over any other biomaterials used for implants so far. This brings about a great potential for HAp in the applications in dental prostheses and cements, and orthopedic surgery, and so on. A few examples are given below. a. Combinations of calcium phosphates as remineralizer and dental cements The solubility study in the system C aO -H gP C ^-^O at ambient temperature (Fig. 4.1) shows that in the region which is not too acidic nor too basic, HAp is the least soluble in all the listed phosphates. In other words, a solution which is saturated with respect to the calcium phosphates represented by the isotherms over that of HAp is supersaturated with respect to HAp already. Any intercept in the diagram is a singular point at which two calcium phosphates are in equilibrium at the corresponding pH. When the solids of the two compounds exist in the solution in excess, the pH will rem ain stable and HAp will precipitate from the solution with continuous dissolution o f the two reacting salts. If the pH drifts away from the value of intercept, say, the solution becomes more acidic, the salt which is more basic will dissolve and drag the pH o f the solution back, and vice versa. So that HAp can keep crystallizing at a stable pH regime. Based on these principles, a series of dental remineralizers and dental cements were invented by Brown and Chow (1986) by various combinations from the phosphate compounds. The main advantages o f these materials include long active time, stable pH, adjustable remineralization rate, and suitable for dental cements. b. HAp ceramics used for implantation Aging, cosmetic corrections, injuries due to sports, traffic and other accidents, and medically-related surgery, etc., are the main reasons for repair and replacem ent o f body parts. Compounded by increased life expectancy, frequent traveling and complex life styles, demand for dental and orthopedic surgical procedures has risen steadily in the recent Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. • moles per /iter 42 7or<3/ Calcium T \ CaHP04 2HjO CaHPQ, Solubility Isotheim s at 25° C pH Figure 4.1. Solubility of calcium phosphates at 25‘C. (Source: Brown et al.) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 43 decades. The materials used for tissue reconstruction have been changing with the development of science and technology. Historically, ivory, gold, brass, silver, platinum, tantalum, stainless steel, polymethylmethacrylate, cobalt alloy, titanium, alumina ceramics, graphite, tricalcium phosphates, and HAp ceramics, and so on, have been used for tissue reconstruction. The use of phosphate ceramics for implants was suggested in the 1970’s. Since then a lot of research work has been done in this area. Some cases o f using HAp ceramics for loaded (e.g. tooth implants) and unloaded (e.g. middle ear canal wall prostheses) implantation have been reported. In most cases, HAp was used only for some so-called “small” surgeries, such as oral surgery, ear, nose, and throat surgeries, where the implants are not subjected to mechanical loads other than compressive forces, because of the brittleness o f HAp ceramics. Before the mechanical property, especially the toughness of the HAp ceramics is solved, some alternative approaches were taken for applications. HAp coating on certain metallic implanting materials is a way to combine the excellent biocompatibility and the excellent mechanical properties of the metals, thus load-bearing implants can be made. It was found that after implantation, a dense HAp surface will lose up to 15 pm thickness due to dissolution processes, so that a coating should have a minimal thickness sufficiently exceeding this lower limit. On the other hand, the thicker the coating, the more it will behave like a brittle ceramic. Therefore, a coating with a thickness of 30-50 pm is thought to be suitable. HAp coating can be obtained electrolytically, but the coating is loose. De Groot reported the spray technique for HAp coatings for implants in surgery. Flame spraying or plasma spraying technique has been used. The fatigue tests and tensile bond testing as compared with uncoated or titanium coated substrates for two different Ti alloys shows that the HAp coated surfaces are at least equal to the other surfaces, both in fatigue and tensile bonding. It is therefore concluded that HAp coatings, as obtained by plasma-spraying on titanium alloys as substrates, are suitable for loadbearing high strength bioactive implants (de Groot, 1987). Hydroxyapatite, with the excellent biocompatibility, is a very new and promising candidate for hard tissue implantation in biomedical field. Some new dental cements with HAp as hydration product, has been patented (Brown and Chow, 1985, 1986). HAp Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 44 ceramics are expected to be used as implantation. However, a lot of problems remain to be solved before their wide practical uses, especially, the m echanical properties and biodegradation. On the other hand, in order for the pure HAp ceramic implants to be practically useful, the mechanical strengths should be comparable to those of natural bones, i.e., compressive strength greater than or equal to 4000 kg/cm 2 (~400 MPa), bending strength greater than or equal to 1500 kg/cm 2 (-150 MPa). Because of the generally relation between porosity and strength, the m ost promising approach to achieve these is to make HAp into dense ceramics or ceramic composites. In addition, the durability of HAp ceramics must be ensured. It is known that at high temperature, HAp may decompose into (3-triclalcium phosphate (P-TCP) or a-tricalcium phosphate (a-T C P ), depending on the heating temperature and duration. W hen the temperature is higher than 1150°C which is the transformation temperature between a - and P-TCP, a-T C P will be obtained. This decomposition causes a few problems. First, the decomposition may result in cracks in the sintered body during sintering process due to the volum e change. Second, if a -T C P which is reactive with water, is produced, biodegradation will be induced and thus result in the damage of the ceramic body. Also, it is noticed that the decomposition of HAp may take place at high temperatures partially or completely, depending on the stoichiometry of the sample. So that it is important to investigate the factors affecting the thermal stability of HAp and offer ways to enhance its thermal stability. 4.2. Synthesis and Thermal Stability of HAp 4.2.1. Introduction Synthetic hydroxy apatite (H A p), w ith a stoichiom etric com position of Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 45 Ca 10 (PO4 ) 6 (OH)2, has excellent biocompatibility with human tooth and bone, and thus making it very attractive for biomedical applications. The studies o f HAp cements (Brown & Chow, 1985; Brown & Chow, 1986; Brown & Fulmer 1991), ceramics (Heise, et al., 1990; Roy, 1975; Fang et al. 1991; Fang et al. 1992; Agrawal et al. 1992), and composites (Ioku et al. 1989; W u & Yeh, 1988; Yoshimura et al., 1989; Fang et al. 1993), for biomedical applications, have been well documented in the recent years (Ducheyne & Lemons, 1988). As a biomedical material, HAp is mainly used in its ceramic forms fabricated by sintering the compacted powder of synthetic HAp at temperatures between 1000 and 1350 °C (Willmann, 1990). The standards for HAp used for bone replacement (ASTM F 1185-88, 1988; Jarcho, 1976) specify that the material must be in the ceramic state and exclude non-ceramic hydroxyapatite since non-ceramic HAp does not exhibit the biological/medical properties as ceramic HAp does. Because of this, high thermal stability o f starting HAp powder is very important for fabricating good HAp ceramics. HAp can be prepared by various synthesis methods, including solid-state reaction (dry process) and wet-chemical methods (Kanazawa, 1989; Jarcho et al., 1976; Monma & Kamiya, 1987; Monma, et al. 1988; Monma et al. 1981; Hattori et al., 1988). In general, the wet processes which have been widely used to provide finer, more homogeneous and more reactive powders than in dry methods. Decomposition o f HAp at high temperatures is frequently encountered during the fabrication of HAp ceramics. This inhibits the sinterability of the apatite ceramics and transforms HAp partially or completely into a mixture of tetracalcium phosphate (TetCP), Ca4 (P0 4 )2 0 , and tricalcium phosphate (TCP), Ca3 (P 0 4)2. TetCP and the a-form of TCP readily react with water. If such reactions occur in the bioceramics under the physiological conditions, the mechanical integrity of the ceramics would be deteriorated. To obtain ceramics of phase pure HAp, the decomposition o f HAp during ceramic fabrication should be prevented. The dehydroxylation and decomposition o f HAp is a process accompanied by the release of hydroxyls or molecular water from HAp, therefore the partial pressure o f water vapor is an important factor affecting the thermal stability o f HAp. Riboud (1968) showed the dependence of the decomposition temperature o f HAp on the water vapor partial pressure (Fig. 4.2.). Thus it Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 46 3CaOPoOc + 4CaO'P9Oi Temperature, 1500 1450 1400 Hydroxyapatite 1350 1300 100 P m o nimHg Figure 4.2 Influence of moisture on the decomposition of HAp. (Source: Riboud) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Al is possible that the dehydroxlation and decomposition o f HAp can be suppressed by sintering it in the atmosphere o f higher partial pressure o f water vapor. This has been demonstrated by Wang & Chaki (1994) recently. The objective here is to study various aspects related to the thermal stability (in air) o f synthetic HAp powders prepared by wetchemical methods, and investigate the influence o f synthesis conditions on the thermal stability of HAp, at temperatures above 700°C. 4.2.2. Experimental Procedures 18.104.22.168. HAp Synthesis A. By hydrolysis of a-T C P [a-C a 3 (P 04)2] First, a-T C P was synthesized by a solid-state reaction method in the following way. Stoichiometric amounts o f fine powders o f C a C 0 3 and N H 4 H 2 PO 4 (both were reagent grade) were hand-mixed in acetone to obtain a homogeneous mixture. This mixture was air-dried and then heated at 300 °C for 16 hours to remove NH 3 and H 2 0 , and then at 900 °C for 24 hours to remove C 0 2- The calcined product was once again ground to improve homogeneity, and finally fired at 1200 °C for 16 hours to form a-TCP. 50 grams o f a-T C P powder (-325 mesh) in 2000 ml deionized water was used for hydrolysis at 60 °C to form HAp. The pH of the system was adjusted in the range of 7.68.5 with NH 4 OH and maintained relatively unchanged during the hydrolysis by using more NH 4 OH as desired. The suspension was subjected to continuous stirring condition throughout the entire hydrolysis process which lasted for about two hours. The product was then repeatedly washed three times with deionized water and dried at 90 °C in a ventilated oven. Synthesis o f HAp was also tried in the early stage by the hydration o f a-TCP. In this method, pellets o f 0.5 inch diam eter were pressed with the fine a-T C P pow der as Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 48 synthesized above. The pellets were then cured in deionized water at pH 8 at 38°C for the tim e periods up to 28 days. Phase and morphology studies were carried out on the resulting materials. B. By hydrolysis of brushite plus ripening Brushite ( C a H P O ^ ^ O ) , also called dibasic calcium phosphate dihydrates (DCPD) (J.T. Baker Chemical Co., Philipsburg, NJ) was used as a precursor to prepare HAp by hydrolysis under the same conditions as employed for the a-T C P hydrolysis mentioned above. The powder thus obtained was washed with deionized water and dried, and then ripened. The ripening treatment was conducted under varying conditions as described below: a. Adding 20% (by weight) calcium chloride dihydrate (CCD), CaCl 2 , 2 H 2 0 , under the same conditions as in the hydrolysis. The ripening was carried out three times repeatedly. The product was washed with deionized water and dried at 90 °C after each ripening. b. Instead of the above three-step treatment, ripening was alternatively carried out at 75±5 °C using 60% CCD, with the initial pH at 8.87 and 11.00, respectively. This treatment lasted 24 h. The ripening of 20% CCD was also carried out at the same pH for 70 min. for comparison. After the ripening treatment, the powders were washed with deionized water and dried at 100 °C. This strengthened (high CCD content, high pH, and long treating time) one-step ripening will be referred as “SOR” hereafter. C. By precipitation A precipitate was prepared by mixing stoichiometric amounts (Ca/P Mol. ratio =1.67) of Ca(N 0 3 )2 *4 H 2 0 and NH 4 H 2 P 0 4 solutions (both o f 0.5 M) at room temperature and pH 11 (adjusted w ith NH 4 O H ). NH 4 H 2 P 0 4 solution was slow ly added to the C a(N 0 3 )2 ,4H 20 solution over an hour, and the mixture was subjected to constant stirring for 24 hours. The resultant white precipitate was centrifuged to a slurry. The slurry thus Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 49 obtained was then dried at 60 °C. D. By hydrothermal method A portion of the precipitate obtained in (C) above, was hydrothermally treated at 200 °C in a Parr bomb under the vapor pressure of about 1.5 MPa for 24 h. The product obtained by the hydrothermal treatment was washed with deionized water and dried at 110 °C for further experiments. 22.214.171.124. Characterization The morphology of the HAp powders prepared by various methods was examined by electron microscopy. The phase composition was determined by powder X-ray diffraction (XRD) using a Scintag diffractometer (Scintag Inc., Sunnyvale, CA) and CuK a radiation. Lattice parameter refinement was conducted by XRD in conjunction with a computerized calculation program. Thermogravimetric analysis (TGA) was carried out on HAp powders at a heating rate of 10 °C per min. The chemical analysis was performed on selected HAp samples using spectrochemical methods. 126.96.36.199. Thermal Stability The thermal stability of the HAp samples prepared in Sec. 188.8.131.52. was investigated by XRD analysis of the air-quenched samples after isothermal heating. The samples were heated at various temperatures in air with 50% relative humidity. The firing time for all samples at all temperatures was 1 h. The heated samples were ground to fine powders and spread on the slides with acetone for qualitative XRD analysis. The thermal stability was characterized by the decomposition temperature of the specific HAp sample. Since the decomposition of HAp is always accompanied by the appearance of TCP (either (3- or De form), the minimum temperature at which the HAp sample was heated and TCP was detected by XRD from the sample was recorded as the decomposition temperature of that Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 50 sample. A high decomposition temperature of the HAp implies good thermal stability, and low decomposition temperature, poor thermal stability. The appearance of diffraction peak at d=2.905A (20=30.8°) was used as the evidence of a-T C P and ^=2.88A (20=31°) for (3TCP. 4.2.3. Results and Discussion 184.108.40.206. Morphology Figure 4.3 shows the morphologies of the as-synthesized HAp powders as seen under SEM or TEM. As can be seen, hydrolysis o f a-T C P produces highly crystalline HAp whiskers (Fig.4 3A). Typical grain size of these single crystals was about 0.2 x 5 (im. HAp obtained from the hydrolysis of DCPD was a mixture o f finer grains and some non-uniform needle-shaped crystallites (Fig.4.3B). The morphology of the HAp from precipitation was agglomerates of tiny particles (Fig.4.3C). The hydrothermal method produces very fine HAp powder. The single crystal particles are uniform in shape, with an average particle size of 0.025 x 0.1 |J.m (Fig.4.3D). It is obvious from this study that the method o f preparation and the nature of precursors strongly influences the morphology of HAp powders. It was also noticed that in vacuum, the electron bombardment (at 120 kV accelerating voltage) made the hydrothermally synthesized HAp particles decompose when the same observed area was irradiated continuously for more than five minutes. Figure 4.4 shows the electron-beam-damaged HAp crystals during a few minutes delayed TEM observation. The hydration of a-T C P at 25°C for 7 days at pH 8 produced HAp crystals o f stalactitic morphology (Fig. 4.5.). XRD analysis indicated that most of the a-T C P had not hydrated at this stage, so that the pure HAp was not obtained by the hydration of a-TCP. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 51 Figure 4.3. M orphology o f HAp synthesized by (a) hydrolysis o f a-TCP, (b) hydrolysis of brushite (DCPD) plus ripening treatm ent, (c) precipitation, and (d) hydrothermal method. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 52 ! 0.1 |Jim Figure 4.4. Micrographs showing (a) the undecomposed crystals and (b) decomposed crystals of HAp by electron beam bombardment during TEM observation at 120 kV. The HAp powder was hydrothermally synthesized at 200*C and 1.5 MPa. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 53 Figure 4.5. M orphology o f HAp produced by the hydration o f a-C a 3 P 0 4 at 25°C for 7 days. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 54 220.127.116.11. Crystallinity The crystallinity of the synthetic HAp can be evaluated from either electron micrographs or and/or X-ray diffraction patterns. An XRD pattern containing strong and sharp peaks corresponds to high crystallinity, and that o f broad and low intensity diffraction peaks corresponds to poor crystallinity. Figure 4.6 indicates that the crystallinity of the as-synthesized HAp prepared by the hydrolysis of a-T C P was very high; that from hydrothermal method was high; that from DCPD was poor; and HAp prepared by precipitation method was almost X-ray amorphous (Fig.4.6A). Comparison of the crystallinity of the precipitate and the hydrothermally synthesized HAp samples indicated that the amorphous precipitate had undergone crystallization under the hydrothermal conditions. This made the hydrothermally synthesized HAp different from that directly from the the precipitation. It was also noticed that ripening treatment did not significantly change the crystallinity of the product. Heating did promote the crystallinity of the HAp. For example, crystallinity of the precipitated HAp increased with the firing temperature (Fig. 4.7.). Similar effect was also observed in the HAp prepared by the hydrolysis of DCPD plus ripening. 18.104.22.168. Stoichiometry The Ca/P m ole ratios o f the HAp samples determ ined by atomic absorption spectroscopy technique are presented in Table 4.1. According to the chemical formula, stoichiometric HAp should have Ca/P molar ratio of 1.67, therefore the HAp with Ca/P lower than 1.67 is characterized as calcium deficient, or nonstoichiometric. Although nonstoichiometry of HAp may also include hydroxyl-deficiency, the latter will not be considered in this study. It was noticed that the HAp samples prepared by hydrolysis of a TCP and DCPD (unripened) were both calcium -deficient, while the samples from precipitation and hydrothermal treatment were stoichiometric. The ripening treatment increased the Ca/P ratio of the treated sample from 1.52 to 1.68, thus stoichiometric HAp Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 4.6. XRD patterns of the as-synthesized HAp powders prepared by (a) hydrolysis of C aH P 0 4 *2H2 0 , (b) hydrolysis o f C aH P 0 4 2H20 plus triple ripening treatment, (c) precipitation, (d) hydrothermal method, and (e) hydrolysis o f a-C a 3 P 0 4. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 56 a s - s y n th e s iz e d T w o th eta , d e g re e Figure 4.7. XRD patterns showing that the crystallinity o f the precipitated HAp increased with heating temperature. All samples were heated for 1 h. The marked (•) peaks belong to TCP, and all unmarked peaks belong to HAp. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 57 was obtained through ripening. In order to examine whether the material obtained after repeated ripening with calcium chloride was chloroapatite Ca5(P 04)3Cl or HAp, special chemical analysis of the powders was carried out by atomic absorption spectroscopic technique on the ripened samples, using an unripened sample as reference. The results are given in Table 4.2. It should be noted that N 0 3' and S 0 42' sources were never used nor introduced during the synthesis. Thus the fact that the amounts of C l' were at the same level as that of N 0 3' and S 0 42', could be considered either as the impurities present in the material, or as the signal of “instrument noise”. The exceptionally high count of Cl" in sample “C” might be due to incomplete washing after ripening, otherwise the C l' content of sample “D” , which was ripened one more time than sample “C”, would be even higher. A simple calculation shows that if 1% O H ' in the HAp is substituted by Cl", the content o f C l' would be about 700 ppm. If complete substitution occurred, the content of Cl" should be 6 . 8 %, or 68000 ppm. Furthermore, the incorporation of C l' in HAp would make O H ' stretching band in IR absorption pattern shift from 3573 to 3495 cm '^ M o n m a et al., 1981), but the IR spectra of the samples ripened with CaCl 2 showed O H ' stretching at 3560 cm'-*- which was very close to 3573 cm ' 1 (Fig. 4.8.). Apparently the substitution of Cl" for O H ' to form chloroapatite did not occur in the ripening process under the current processing conditions. 22.214.171.124. TGA Results It was reported (Yamashita & Kanazawa, 1989) that nonstoichiom etric HAp decomposes via dehydration according to the following reactions: <700 °C Caio-x(HP 0 4 )x(p 0 4 )6 _x(OH)2 .xnH 2 0 ------ >C a 10.x(P2 O7 )x(PO4 )6 .2 x(OH) 2 +nH 2 O(g) (a) 700 - 800 °C ---------------> (l-x)C a 10 (PO4 )6 (OH)2 +3Ca 3 (PO4 )2 +xH2O (g) (b) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 58 Table 4.1 Ca/P ratios of HAp samples determined by atomic absorption spectroscopy. Method Ca/P Synthesis conditions Hydrolysis of a-T C P 1.65 60 °C, 2 h, pH 7.6 - 8.5 Hydrolysis of DCPD 1.52 As hydrolyzed, 60°C, 2 h, pH 7.6 - 8.5 Hydrolysis of DCPD 1 .6 8 Ripened once, 60°C, 2 h, 20% CCD Precipitation 1.67 25°C, 24 h, pH 11 Hydrothermal 1.67 200°C, 1.5 MPa, 12 h, pH 11 Table 4.2. Spectrochemical analysis of chloride ions in HAp samples (in ppm). cr n o 3- S 0 42’ A 80 90 110 B 130 70 40 Ripened once C 460 200 140 Ripened twice D 130 70 50 Ripened thrice Sample Remarks Before ripening Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 59 The resultants after therm al treatm ent are stoichiom etric HAp and TCP. At temperatures higher than 900 °C, further partial dehydration of HAp may take place as follows. >900 °C Ca 10 (PO4 )6 (OH ) 2 ----------- > C a 10 (PO4 )6 (OH)2 _2 yOy + yH20 (g) (c) According to above reactions one would observe a gradual weight loss before 700 °C, a sharp loss in 700-800 °C region, and a slight and gradual loss above 900 °C in the TGA thermogram of a nonstoichiometric HAp sample. There should not be any sharp loss in w eight for stoichiom etric HAp in its TG A curve. Thus stoichiom etric and nonstoichiometric HAp samples can be easily distinguished by TGA. The TGA thermograms of HAp prepared by different methods are shown in Fig. 4.9. Among these only two samples, i.e., HAp prepared by the hydrolysis of a-T C P and DCPD (without ripening) showed sharp deflection (Fig. 4.9. A, E) between 700-800 °C in the thermograms, indicating they were both nonstoichiometric. The rest of the samples did not show such a sharp loss, and therefore they were considered to be stoichiometric (Fig. 4.9.B, C, D, F, G). The difference in the wt. loss below 300 °C was due to the degree of drying before the TGA experiment. The substantial wt. loss o f the precipitated HAp below 200 °C indicated that this sample contained large amount of entrapped water. The DCPDderived samples also showed moderate weight loss, at low temperature range. The HAp from a-T C P hydrolysis showed the least weight loss in the low temperature range, because this sample was composed of relatively large crystals and undergone near complete drying prior to TGA experiment. The whiskered HAp had good crystallinity but still showed apparent wt. loss between 700 and 800 °C (Fig. 4.9., A) and is believed to be nonstoichiometric. This conclusion is in good agreement with what reported by Monma et al. (1981) These results indicate that good crystallinity does not necessarily imply good stoichiometry, and vice versa. The precipitated HAp had poor crystallinity (broad XRD peaks, Fig. 4.6C) but did not show any sharp weight loss in the 700-800 °C region in the TGA curve (Fig. 4.9.F). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 60 co|- 4000 3500 3000 200 0 1800 1600 1400 Wave number, cm ' 1200 1000 800 T ift JT ft 7T5*o 1 Figure 4.8. (a) IR spectra o f HAp prepared by (1) hydrolysis of DCPD, (2-4) hydrolysis o f DCPD followed by ripening treatment for 1,2, and 3 times with CaCl2, (5) hydrolysis o f a-TC P, and (6 ) precipitation. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 61 48 •8 ■uS H u ■S I•C .o' •S 0) £ c C /3 <o o fi o <D a* C /3 g 00 Tf B s W) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 62 126.96.36.199. Lattice Parameters The lattice parameters of the HAp samples determined by XRD are listed in Table 4.3. The reported values from the JCPDS file (9-432) was considered as references for stoichiometric HAp. Compared with the as-synthesized HAp by DCPD hydrolysis (sample A) it is seen that ripening reduced (sample C) both a and c slightly, as the Ca/P ratio was increased (Table 4.3.). Compared with the values of the precipitate (sample D), it is obvious that hydrothermal treatment reduced both a and c through crystallization thus making the resultant HAp denser in crystal structure. 188.8.131.52. Thermal Stability The thermal stability is discussed with respect to the decomposition o f the specific HAp sample. The precipitated HAp, which was chemically stoichiometric as observed in the previous section, was found to be stable up to about 900 °C. The highest thermal stability up to 1370 °C was found in the hydrothermally synthesized HAp (Fig. 4.10.). It is evident that the crystallization of the precipitate under the hydrothermal conditions is a very effective way to drive the amorphous precipitate of exact stoichiometry towards the highly thermally stable crystalline HAp. Although crystallization also occurred when heating the precipitate in air, the pressure of water vapor in air was low. Under the hydrothermal condition of high pH region, the atmosphere is saturated and the pressure is high, thus favoring the formation of fine, uniform, stoichiometric, and structurally dense crystalline HAp. For the DCPD-derived HAp samples, ripening treatment really made difference. The as-synthesized HAp by the hydrolysis of DCPD was nonstoichiom etric which had completely transformed to P-TCP at 733 °C (Fig. 4.11.), but the sample ripened once with 20% CCD remained stable up to 1000 °C. Repeated ripening treatment (three times) further enhanced the thermal stability up to about 1200 °C. Figure 4.12. illustrates the effect of ripening on the thermal stability of HAp. The increase of Ca/P ratio of the HAp from 1.52 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 63 o 5 oo 10 15 0 500 1000 T e m p e ra tu re , *C Figure 4.9. TG thermograms o f HAp synthesized by (a) hydrolysis o f a-TCP, (b) hydrothermal method, (c) DCPD hydrolysis, ripened twice, (d) DCPD hydrolysis, ripened thrice, (e) DCPD hydrolysis without ripening, (f) DCPD hydrolysis (ripened once), and (g) precipitation. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 64 Table 4.3. Lattice parameters o f the HAp samples synthesized in the current study, A. Sample No. a c Synthesis method Reference 9.418 6.884 JCPDS file (No.9-432) A 9.432 6.890 Hydrolysis of DCPD B 9.430 6.878 Hydrolysis of a-T C P C 9.428 6.875 “A” plus ripening D 9.427 6.893 Precipitation E 9.418 6.884 Hydrothermal Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 65 to 1.67 or 1.68 (Table 4.3.) clearly proves that the stoichiometry is one of the key factors that control the thermal stability of HAp. However, since both temperature and pressure in the ripening process were much lower than in the hydrothermal process, the repeated ripening did not produce results as good as in the hydrothermal method. Also, increasing treating time, increasing amount of CCD, and higher pH, all favored the transformation of nonstoichiometric HAp into stoichiometry. This is quite evident in the SOR samples. For example, the HAp ripened with 60% CCD at initial pH of 11 and treated for 24 h was stable up to 1370 °C. The samples used for comparison, i.e., those treated at lower pH, less CCD, or shorter period of soaking time, were stable up to 1300-1350 °C. Fig. 4.13 shows the XRD patterns of the SOR treated HAp sample, heating at 1350 °C and 1370 °C for 1 h. Table 4.4 summarizes the phase composition results of the HAp samples heated at various temperatures. It can be seen that the thermal stability of the synthetic HAp depends upon both the nature of the precursor and the synthesis method. The hydrothermal treatment substantially enhanced the thermal stability. The hydrolysis of DCPD followed by the strengthened one-step ripening (SOR) offered good thermal stability. The as- synthesized HAp by hydrolysis of a-T C P began to decompose at 800 °C and remained as a mixture of HAp and TCP over a large temperature range. For the SOR samples, only thermal stability was studied. Comprehensive explanation on the behavior o f these samples needs further study. It should be indicated that during heating in air, HAp might have at least partially transformed to the so-called oxy-hydroxyapatite (OHA) by dehydroxylation. If this is the case, then the HAp detected in the samples heated at high temperatures should be more accurately called OHA. In order to completely prevent the appearance of OHA, sintering of HAp should be conducted in moisture (Wang & Chaki, 1993). 4.2.4. Summary Thermal stability of several synthetic hydroxyapatite samples prepared by wetchemical methods was studied. The HAp powder synthesized by the hydrothermal method Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 66 . . A_A I 80 as 90 99 CALCIUM PHOSPHATE HYDROXIDE / HYOROXYLAPATITE ..................... 1 , . l.l I................. . 40 1 *9 ,J. CAS ( P 0 4 ) 3 90 ( 0 H I Two theta, degree Figure 4.10. XRD patterns of the hydrothermally (200 °C, 1.5 MPa) synthesized HAp fired at 1370 °C in air of 50% RH (11.88 mm Hg) remains undecomposed. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 67 693 C , 1 h r 733 C , I h r UOO C, 1 h r . i r -■»* 20 Ji M i ^ T 30 i^ i I ~~*1 ~T 40 1100c> 1hr ~*'V ** r‘ ~i ^ tf “ * |‘ 50 60 26, degrees Figure 4.11, XRD patterns of the as-synthesized HAp by DCPD hydrolysis, heated at various temperatures. It is seen that the decomposition started at about 700 °C. 714 °C, p-TCP + HAp; 7 3 3 ,1 100°C, p-TCP. 1300°C; a-T C P + P-TCP. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 68 I 20 T'A' A"1 ■ ! 25 I" J L m I ' ’ H * ^ ^ 1 '1*1 30 35 . . a . *8 w . AO *4^ A5 T w o th eta , d e g re e Figure 4.12. XRD patterns showing the influence of repeated ripening on the thermal stability of HAp obtained by DCPD hydrolysis. Samples were heated 1 h at 1100 °C in air o f 50% RH. (a) W ithout ripening: p-TCP. (b) Ripened once: HAp+ P-TCP. (c) Ripened twice: HAp+P-TCP (tr.). (d) Ripened thrice: HAp only, no decomposition. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 69 1 3 5 0 * C » : a . 1 i . a . L 1 i i a U . 1 r t 20 r l i i i l : . A : . J l ^ . 1 . ; . a ' j A i» 1 A l " i 2j 30 J - ■ ^ ■ * i -1 - r 1 1‘ 35 1 1 1* 40 3 t 1*1 l i ^ i " 45 50 1370 *C . J L !_ i a -1 . . a Av.I . .1. aA 1. a I Ai i : .l . A » A: : *A_ ... .A . A a ^ l\ A .. . B . I\. J ■A -+ T -,4 --r-A -i 1 . i l. . f„,». , ■y.L-.-.A-. T\vo theta, degree Figure 4.13. XRD patterns showing the phase composition of SOR (strengthened one-step ripening) heated 1 h at 1350 °C and 1370 °C, respectively. The ripening after the hydrolysis o f DCPD was carried out at 75 °C and (A) 20% CCD, 70 min., initial pH 8.87, (B) 60% CCD, 24 h, initial pH=8.87, (C) 20% CCD, 70 min., initial pH=11.0, and (D) 60% CCD, 24 h, initial pH = l 1. The marked peaks (•) belong to a-TCP. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 70 Table 4.4. The main phase composition appearing in the HAp samples heated in air of 50% relative humidity at various temperatures. T CC) Hvdrolvsis of DCPD Hydrolysis of a-TCP PPT. Hydrothermal Ripened As is i n m SOR 700 H H H H H H H H 800 H + p(tr.) H H H H H 900 H+P H H H H H+a(tr.) H 1000 H+P H+p H+P(tr.) H H H+a(tr.) H 1100 H+P P P P P H H+P H+P(tr.) H H H+a H 1200 H+a a - H+P H+a(tr.) H H+a H 1300 H+a a H+a H +a H+a H H+a H a ,p —a - or P-TCP. H—HAp. PPT.—precipitated. I, II, and III indicate the number of ripening time. SOR—strengthened one-step ripening, (tr.)— trace amount, defined arbitrarily by relative intensity of XRD patterns. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 71 was uniformly nanocrystalline. This material showed very high thermal stability up to 1370 °C in air of 50% relative humidity. The as-synthesized HAp by the hydrolysis of brushite showed poor thermal stability, starting decomposing at about 700 °C. Repeated ripening with addition o f calcium chloride enhanced the thermal stability up to about 1200 °C. The strengthened one-step ripening treatment of this DCPD-derived HAp had increased the decomposition temperature o f the HAp up to 1370 °C. The HAp obtained by the hydrolysis o f a -T C P was highly crystalline but nonstoichiom etric and partially decomposes at about 800 °C. The as-synthesized HAp obtained directly by precipitation was almost X-ray amorphous. The crystallinity of this HAp increased with temperature till about 900 °C at which it decomposes. The results in this study indicate that the thermal stability of the synthetic HAp does not depend upon the crystallinity but the stoichiometry and method of synthesis. In air of 50% relative humidity, the chemically stoichiometric and structurally dense HAp showed the best thermal stability. To avoid dehydroxylation and decomposition of HAp during ceramics preparation, either humid atmosphere or HAp powder of high thermal stability should be used. However, the combination of both should enable to completely prevent the decomposition of HAp in the general sintering temperature range. 4.3. Fabrication of Regularly Dense HAp Ceramics 4.3.1. Introduction Microwave processing technique is still in an exploratory stage of development. Although it has been investigated for many ceramic materials, no microwave sintering of HAp prior to this study was reported. In the current study, regular dense HAp ceramics have been fabricated by microwave sintering in a 500 W microwave furnace. Circular-plate specimens of various green densities were sintered in the furnace at 1200 and 1300 °C, for 5, 10, and 20 minutes, respectively. Ceramics with density up to 97% of the theoretical Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 72 were obtained. The ceramics sintered by microwave and by conventional methods were characterized for density, grain size, microstructure, and strength, and the results were compared. 4.3.2. Experimental Procedure 184.108.40.206. Material HAp powder prepared by the hydrolysis of brushite (C aH P 0 4 -2H 2 0 ) , followed by ripening with 60% wt. CaCl2 2H20 was used in this study. The ripened HAp was dried at 90 °C in a ventilated oven and then ground to -325 mesh followed by heating at 500 °C for 24 h. The BET surface area of this powder was found to be 47.4 m 2 /g. Powder XRD confirmed that the m aterial was pure HAp with low crystallinity. W ithout using any binder, pellets of 12.7 mm (0.5 inch) diameter were made at uniaxial pressures from 35 to 351 MPa to get various green densities. Two 19.05 mm (3/4 in.) diameter pellets, on which the patterns of a dime were pressed, were used to test the uniformity of shrinkage during microwave sintering. 220.127.116.11. Sintering Setup A 500 W, 2.45 GHz commercial microwave oven (Model JE43, GE Co.) was modified for the sintering experiments. The arrangement inside the oven is schematically shown in Fig. 4.14. In this case there was no turntable, but there was a mode stirrer above the cavity. Samples were stacked in layers at the center o f a vertically placed porous zirconia cylinder (Zircar Products, Inc., Florida, NY) (28 mm ID, 51 mm OD and 30 mm in height), which was placed on a thick Fiberfrax (Fiberfrax, Harbinson-Carborundum Corp., Niagara Falls, NY) cushion, and then covered all around with a layer of Fiberfrax thermal insulation fibers. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 73 18.104.22.168. Temperature Measurement An alumina-sheathed and Pt foil-shielded S-type thermocouple was inserted from the left wall o f the oven to m onitor the temperature of the samples during sintering. The thermocouple was so arranged that the tip was only about 2 mm away from the surface of a sample (Fig. 4.15). The microwave interference with the tem perature m easurem ent is a common observation at low temperatures, but once coupling between microwaves and the material is initiated (i.e. when zirconia and specimens begin to absorb microwave effectively) the tem perature readings get stabilized. The lower temperature readings are recorded by m omentarily switching off the power supply to the m icrowave oven. The response time was less than 3 seconds. Tricalcium phosphate (TCP), C a 3 (P04)2, was used to calibrate and determine the accuracy o f the therm ocouple-temperature-measurem ent m ethod used in the current m icrowave experiments. TCP exists in two crystalline forms: a-TC P (high-temperature form) and j3-TCP (low-temperature form). The transition of /i-TCP to a-T C P starts at about 1125 °C and the process is sluggish. P-TCP will transform completely to a-T C P if temperature is sufficiently greater than 1125°C or if kept above the transition temperature for a sufficiently long time. Although a-T C P is a high-temperature form, it can also m etastably exist at room temperature even in regularly cooled samples. Therefore by comparing the relative amounts of two TCP phases in the sample of /J-TCP heat-treated by conventional and microwave methods at temperatures around the transition temperature, an accurate temperature of microwave heated sample can be determined indirectly. For this purpose, the following experiments were conducted. Three pellets of labsynthesized P-TCP were stacked one over another inside the porous zirconia cylinder in the m icrowave set-up, with the thermocouple located at a distance o f about 2 mm from the pellet in the middle (Fig. 4.15). The rest of the experimental setup was identical to that Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 74 Figure 4.14. Schematic showing of the arrangements in the microwave sintering cavity. The sintering package includes (1) thermocouple, (2) zirconia cylinder, (3) pack of samples, (4) thermal insulating fibers, (5) power level control, and (6 ) time control. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 75 Figure 4.15. Schematic showing of relative positions of thermocouple and (3-TCP pellets during the temperature measurement calibration. Labeled are (1) thermocouple, (2) porous zirconia, and (3) TCP pellets. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 76 used for HAp sintering (Fig. 4.14). The samples were heated in the microwave field at 1125±10 °C for 15 min., then cooled to room temperature. Another set of three pellets was sintered (one pellet per run) in a conventional furnace for 15 min. at 1100,1125, and 1150 °C, respectively, then air-quenched to room temperature. XRD was carried out on all samples sintered for phase identification. 22.214.171.124. Sintering Six multi-sample runs (six HAp pellets with green densities ranging from 32.6 to 51.4%, stacked in two layers in each run) of microwave processing were carried out for 5, 10 and 20 minutes at 1200 and 1300 °C, respectively. The two pellets with the dimepatterns were microwave sintered at 1200 °C for 7 min. to check the uniformness in shrinkage. During the sintering, the microwave oven was set at the full power level (500 W output), and the temperature was controlled manually by on-and-off operations of the pow er supply. After each run, the system was allowed to cool naturally to room temperature. For comparison, conventional sintering of the same material was also carried out at the same temperatures (2 h soaking at each sintering temperature, with heating and cooling rates o f 5 °C per min.) in a 5 kW programmable box furnace (Model 51524, Lindberg, Watertown, W I ). 126.96.36.199. Characterization Density, grain size, microstructure, and tensile strength of the sintered pellets were measured. Density was determined by direct measurement of the geometric size and weight of the specimens. Grain size was measured from the electron micrographs o f the assintered surfaces. M icrostructure was examined on the fracture surfaces with the SEM. Phase compositions were identified by XRD. The tensile strength was m easured by diametral-compression test (Rudanick et al.) in a MTS810 (Material Test System, MTS System Corp., Minneapolis, MN) with a crosshead speed of 0.051 mm (0.002 in) per minute. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 77 4.3.3. Results and Discussion 188.8.131.52. Temperature Measurement Figure 4.16. shows the XRD patterns of six TCP samples processed in this study. As can be seen, conventionally, the sintered sample at 1100 °C remained pure P-TCP [Fig. 16, (a) 1]; a-T C P appeared as a minor phase at 1125 °C [(a) 2] and as a major phase at 1150 °C [(a) 3]. These results indicate that in conventional heating, the phase transition from p- to a-T C P started at about 1125 °C when soaked for 15 min., and substantially transformed at 1150 °C. In contrast, in case of microwave processing at 1125+10 °C, the pellet located at the bottom of the stack remained p-TCP [(b) 1], the middle pellet contained only trace amount o f a-TCP, and p-TCP was still the major phase in the top pellet [(b)3j. Table 4.5. lists the alp intensity ratio (using intensity of d = 2.909A for a-TCP, and d = 2.880A for p-TCP) of the characteristic reflections in the XRD patterns. As can be seen, the middle pellet sintered by microwave at 1125110 °C is comparable in the a//3 intensity ratio to that of the conventionally sintered sample at 1125 °C, while the top pellet shows higher a/p ratio than the one conventionally sintered at 1125 °C, but obviously much lower than that sintered at 1150 °C. The bottom pellet was the same as conventionally sintered at 1100 °C (no trace o f a-TCP). Considering the temperature deviation of +10 °C in the microwave sintering, the measured temperature (on the middle pellet) was in excellent agreement with the expected value. Thus the validity and accuracy of the temperature measurement with the thermocouple in the current microwave processing is confirmed. 184.108.40.206. Sintering Process Figure 4.17. shows the temperature-time curves of the six sets of HAp specimens sintered by microwave processing. The run at 1200 °C for 5 min. was irregular, therefore the results of this run are not included here. As can be seen, each curve has three parts, corresponding to three stages: heating, soaking, and cooling. The sharp temperature increase in the heating curve is characterized by the phenomenon of thermal runaway. It is Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 78 3. top 3. 1150°C 1 A— 2 . middle 2.1125 °C J aJLa^ l-J 1.1100 °C 20 1. 30 29, degrees 40 20 30 bottom 40 20, degrees Figure. 4.16. XRD patterns (Cu K a ) o f p-TCP samples used to calibrate temperature measurement, (a) Conventionally sintered for 15 min. at temperatures indicated, (b) Microwave sintered at 1125±10 °C for 15 min. (pellet positions are indicated). (•): a-TCP. Unmarked peaks: /J-TCP. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 79 Table 4.5. Intensity ratio of a-T C P id = 2.909A) to /J-TCP (d = 2.880A) in the /3-TCP after firing under different conditions for temperature measurement calibration. Conventional, 15 min. 1150°C 1125°C 1.769 0.172 Microwave, 1125±10 °C, 15 min. 1100°C 0 top middle bottom 0.282 0.111 0 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 80 1400 1400 1200’C 1200 1300‘C 1200 1000 1000 800 BOO O o I I* £ 8, 600 400 400 200 200 0 0 0 10 20 30 40 60 20 30 40 50 60 Time, min Time, min (a) (b) Figure 4.17. Temperature-time curves in the microwave sintering o f HAp at (a) 1200 °C and (b) 1300 °C. Sintering time: (A) 5 min., (-----) 10 min., (• • •) 20 min. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 81 well known that zirconia is a microwave absorber, so that the zirconia cylinder plays a dual role during the m icrowave heating: (i) a thermal insulator, and (ii) a preheater. Furthermore, since the amount o f HAp (less than 6 g per run) used in the sintering was much less than the amount of zirconia (about 2 0 g), it is proposed that the zirconia cylinder triggered thermal runaway. To confirm this, and to determ ine the role of Fiberfrax insulation, two blank tests were carried out: (a) with Fiberfrax only (no sample and no zirconia cylinder), temperature rose only to 369 °C after a 60-min irradiation; (b) with Fiberfrax and the zirconia cylinder, thermal runaway was noticed after 21 min. This confirms that the zirconia cylinder played a key role in triggering thermal runaway in the current microwave sintering. At low temperature, HAp absorbed little microwave and was heated up by the zirconia cylinder which got heated first by absorbing microwave energy. When temperature was sufficiently high, both zirconia and HAp would absorb microwave more efficiently, thus the sintering o f the HAp samples was a complex result of both direct and indirect microwave heating. 220.127.116.11. Uniformness in Sintering Although having experienced high-rate heating (thermal runaway) and cooling (in the first few min. of cooling the rate was high), the HAp pellets did not show any cracks after microwave sintering. As shown in Fig. 4.18., all the details o f the coin-pattems have been uniformly reduced and preserved on the sintered HAp pellets after 7 min. microwave sintering at 1200 °C. This indicates that under proper conditions, microwave sintering is quite uniform and homogenous. The uniform shrinkage should be attributed to the volumetric heating, good thermal insulation, neglected thermal gradients, and the small size of the specimens. 18.104.22.168. Sintered Density Figure 4.19. shows the relative densities of the microwave sintered pellets versus the corresponding green densities. The values of the conventionally sintered pellets were also Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 82 Figure 4.18. Photograph showing the uniform shrinkage o f microwave sintered HAp pellets. The dim e-pattem s on the green pellets (large) have been uniformly reduced and preserved on the sintered pellets (small) by microwave sintering at 1200 °C for 7 m in .. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 83 100 30 35 40 45 50 55 Green density, % Figure 4.19. Relative density (% theoretical) o f HAp ceram ics sintered by (1) microwave/1300 °C/10 min., (2) microwave/1200 °C/10 min., (3) conventional/1300 °C/2 h, and (4) conventional/1200 °C/2 h. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 84 plotted in the same graph for comparison. It can be seen that at the same temperature, higher densities were achieved in the microwave sintered pellets, and this is more obvious when green density was high (typically over 45%). Generally, the densification of ceramics in a solid-state sintering is a process of shrinkage of the compact or removal of porosity. Although quite a few mass transfer processes take place simultaneously during sintering, only transfer o f matter from the particle volume or from the grain boundary between particles causes shrinkage and pore elimination (Kingery et al., 1976, p.475). Therefore, it can be suggested that grain boundary diffusion between the HAp particles was substantially enhanced during the microwave sintering. The grain boundary diffusion takes place only after the particles are necked (in contact) and at a suitable temperature (to overcome the energy barrier). In the pellets of low green density, the number of contact points is limited. As the green density increases, the particles-in-contact correspondingly increases, leading to increased sintered density in both microwave and conventional sintering (Fig. 4.19.). However, the increase in the microwave sintered samples was more than that in the conventionally sintered samples. Also, the microwave sintering took much shorter time (< 1 0 % of the time of the conventional sintering), it can be further suggested that the densification rate (through boundary diffusion) in the microwave sintering process was much higher than in the conventional sintering. The volumetric heating in the microwave sintering is a consequence of a series of energy losses. The principle of microwave sintering has been reviewed by Sutton (1989). When microwaves penetrate and propagate through a dielectric material, the internal fields generated within the affected volume induce translational motions of free and bound charges and rotate charge complexes such as dipoles. The resistance of these induced motions due to inertial, elastic, and frictional forces, causes losses and attenuate the electric field. As a consequence of these losses, volumetric heating results. This principle should be applicable to the microwave sintering of HAp also. It is also suggested that the hydroxyl groups in the HAp structure might have played an important role in causing losses and enhancing densification in the microwave sintering of HAp. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 85 22.214.171.124. Grain Size and Microstructure Figures 4.20. and 4.21 show the SEM photographs of the HAp specimens sintered at 1200°C by microwave and conventional processing, respectively. The differences in morphology o f these samples can be clearly seen. Compared with the conventionally sintered sample (Fig. 4.20.a) in which the entire as-sintered surface is rough and uneven, and the grains are angular with boundaries shown by the straight lines, the microwave sintered sample (Fig. 4.21.a) shows unique features: the whole area looks flat and smooth, and the grains are tightly bounded. Since the driving force for the densification of solid-phase sintering is the decrease in surface area and the lowering of the surface free energy by the elimination of solid-vapor interfaces (Kingery, et al., 1976, p.469), the rough and uneven as-sintered surfaces correspond to a higher residual surface area and imply a lesser degree of sintering. By comparing the SEM photographs of the fracture surfaces, it can be seen that the microwave sintered HAp (Fig. 4.21.b) has a lower total porosity and finer pore size than the conventionally sintered specimen (Fig. 4.20.b). Table 4.6. shows a comparison between the average grain size of the microwave and the conventionally sintered HAp samples. The grain size of the conventionally sintered HAp (1200°C) is close to that of microwave sintered one at the same temperature. For the pellets sintered at 1300 °C, the grain size of both microwave and conventionally sintered samples had increased substantially, but the grains in conventionally sintered sample were much larger. In addition, the abnormal grain growth was pronounced in the conventionally sintered pellet, but was less obvious in the microwave sintered one. The relatively lower grain growth during microwave sintering is attributed to the shorter sintering time and a more effective densification process. 4.3.3. 6 . Phases XRD analysis on the sintered HAp samples detected a-T C P in trace, but still, HAp was the major phase. This indicates that a partial decomposition of HAp took place during Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 86 8 . 02KX 5U 0 0 1 7 MW11 8F Figure 4.20. SEM micrographs showing (a) as-sintered surface and (b) fracture surface of HAp sintered by conventional processing at 1200 °C for 120 min. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 87 20KU 3 . 8 7 KX Figure 4.21. SEM micrographs o f (a) as-sintered surface and (b) fracture surface of HAp sintered by microwave processing at 1200 °C for 10 min.sintering. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 88 sintering. To prevent the decomposition, either lower processing temperatures should be used, or some special precautions, such as special atmosphere or using HAp of high thermal stability, should be taken. Table 4.6. Grain size of HAp pellets sintered under different conditions, |im. Microwave 5 min. 10 min. Conventional 20 min. 120 min. 1200 °C ... 1.05 1.48 1 .1 2 1300°C 2.06 3.16 3.39 6.52 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 89 126.96.36.199. Mechanical Strength Figure 4.22 (a) shows the diametral tensile strength as a function of the sintered density of the HAp ceramics, regardless of the sintering time and temperature. The data for both the microwave and conventionally sintered HAp seem to follow the same densitystrength relation. The diametral tensile strength of the pellets sintered at 1200 °C has been plotted in Fig. 4.22 (b). In this figure, it can be seen that starting with the same green density, a microwave sintered pellet exhibited higher strength than the conventionally sintered HAp: 10 min. microwave sintering achieved higher strength [Fig.4.22 (b), 2] than the 2 h conventional sintering [Fig.4.22 (b), 3], and the pellets of 20 min. microwave sintering demonstrated even higher strength [Fig.4.22 (b), 1]. Comparing with Fig.4.22 (a), it is clear that the enhanced mechanical strength in the microwave sintered HAp ceramics should be attributed to their higher densities. 4.3.3. 8 . Energy Consumption Besides the denser microstructure, finer microstructure, smaller grain size, and the improved mechanical strength, the economic issues of microwave sintering in terms of both time and energy are very important. Fig. 4.23 compares the temperature-time profiles of microwave and conventional sintering processes. Taking a typical run at 1200 °C for 10 min. with a 500 W microwave oven for example, the microwave sintering process consumed energy only in the heating and soaking periods, which was about 0.5 h in total, while the conventional sintering process required 10 h (it consumed energy in the cooling cycle, also) and a furnace with a minimum of 1 kW (although a 5 kW furnace was used in the current study). Thus microwave sintering for the laboratory scale specimens requires only about 5% of the time, and even a smaller percentage of energy, of the conventional process. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 90 60 •B 60 S 40 £ia JD "S c 3 e aj 20 0 BO 70 80 90 100 Sintered density, % (a) 30 35 40 45 50 Green density, % (b) Figure 4.22. Diametral tensile strength of HAp ceramics (a) versus sintered density: (o) microwave sintered, (A) conventionally sintered; (b) versus initial green density: (1) microwave/20 min., (2) microwave/10 min., and (3) conventionally/2 h. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 91 1400 1200 U 1000 800 600 .<o 400 200 0 100 200 300 400 500 600 Time, min Figure 4.23. Com parison o f tem perature-time profiles in the sintering o f HAp by (1) microwave and ( 2 ) conventional methods. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 92 4.3.4. Summary HAp ceramics have been sintered by microwave heating. The results show that HAp samples with varying green density can be sintered by microwave irradiation to high densities within a short period o f time. In contrast to conventional method, microwave sintering of hydroxyapatite consum ed m uch less energy, offered apparently denser m icrostructure, finer grain size, and consequently, im proved m echanical strength. Technically, if a microwave oven is adjustable in both power and frequency, with a combined feedback temperature controlling system, the efficiency o f the process could be even higher. 4.4. Microwave Processing of Porous Hydroxyapatite Ceramics 4.4.1. Introduction Porous hydroxyapatite ceramics may be used for highly effectively restoring bony mass. For example, porous HAp granules can be used to restore the deficient alveolar ridge. The pioneering work to develop porous HAp ceram ics by the route of the replamineform technique was done by Weber etal. (Weber etal., 1969; R.A. W hite etal., 1972; E.W. W hite et al., 1975; D.M. Roy, 1975). Based on the study o f the microstructure of sea-urchins and corals by Weber et al. (1969), the first and possibly only commercially available biomimetic material even twenty years later was developed by W hite, W eber and W hite (R.A. W hite, 1972; E.W. W hite, 1975) who devised the replamine process, and D. Roy (Roy, 1975; Roy and Linnehan, 1974; Eysel and Roy, 1975) who learned to convert the C aC 0 3 to HAP. These porous HAp ceramics are available commercially from Interpore International for dental and bone transplants. HAp is the obvious material expected to be widely used for biomedical applications in humans. For implants, HAp ceramics, either dense or porous, can be used for different purposes. As shown in previous section, regularly dense HAp ceramics have been successfully fabricated by m icrowave processing. The results show that, in addition to lowering Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 93 sintering temperature and enhancing mechanical strength, one can sinter HAp ceramics in a few minutes, consuming less than one tenth o f the energy or time than required for the conventional process. Because o f the usefulness o f porous HAp ceramics in biomedical applications, fabrication o f porous HAp ceramics by m icrow ave processing was investigated in the current study. 4.4.2. Experimental Procedure 188.8.131.52. HAp Precursor Powders Two kinds of HAp-precursor powders were used. A. W hiskers o f pure HAp, with typical size of 0.2 x 5 (J.m, BET surface area 10.8 m 2/g, synthesized by the hydrolysis o f tricalcium phosphate [a-C a 3 (P0 4 )2 , a-TCP)] at 60°C, pH 8±0.5, dried at 165°C (Fang, et al. 1994). B. Pure HAp, agglomerations of very fine needle-like crystals, with typical size of single crystal size at 0.05 x 0.5 pm (estimated from microgragh), BET surface area 47.4 m 2/g, prepared by the hydrolysis of brushite (C aH P0 4 .2H 2 0 ) followed by "ripening" treatment with CaCl2 (Fang et al., 1994; M onma and Kamiya, 1987), sieved to -325 mesh, and dried at 500°C. Figure 4.24. shows the morphologies of these powders. 184.108.40.206. Pow der Consolidation Three groups of circular pellets of 12.7 mm (0.5 in) diameter were pressed with (1) powder A and (2) powder B, both at 35 - 140 MPa obtain various green densities, and (3) mixture (will be designated as C) o f powder B with various amounts of ammonium carbonate [(NH 4 )2 C 0 3 'H 2 0 ], pressed at 140 MPa. Ammonium carbonate was used to introduce additional porosity in the specimens. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 94 Figure 4.24. M orphology o f the starting HAp pow ders composed of (a) whiskers synthesized by the hydrolysis o f a-T C P and (b) agglomeration of fine single crystals synthesized by the hydrolysis o f brushite followed by ripening treatment with calcium chloride. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 95 4A.2.3. Sintering Sintering was carried out in air in a 2.45 GHz, 500 W commercial microwave oven. The experimental setup was described earlier in Sec. 4.3. The processing conditions are listed in Table 4.7. For pure HAp (A and B), the specimens (batches I to HI) were heated directly from room temperature to the final sintering temperature, and soaked at the peak tem perature for only 1 min., then cooled to room temperature. In the case of HAp containing ammonium carbonate (C), one batch (TV) was heated at 120°C for 12 h and then 400°C for 20 min. in an electric furnace to remove ammonium carbonate (decomposition temperature 58°C), and then transferred to the microwave oven, processed at 1150°C for 3 min. Another batch (V) was first heated in the microwave oven at the lower power level to remove ammonium carbonate, then heated at the full power input of the oven to 1200°C, and allowed to cool after soaking at 1200°C for 5 min. 220.127.116.11. Characterization Porosity and density of the sintered specimens were determined by geometrical measurements. M ercury intrusion porosim etry experiments were carried out on three representative specimens to determine open porosity. The maximum pressure for intrusion was 362 MPa (52,500 p.s.i.). Microstructures of the processed specimens were examined by SEM. XRD studies were carried out to identify the phase composition of the sintered samples. Tensile strength was determined by means of diametral-compression (Brazilian) tests (Rudanick etal.), with a crosshead speed of 0.051 mm (0.002 in)/min. 4.4.3. Results and Discussion The products were porous HAp ceramics with porosities from 10 to 75%. All the processed specimens were perfect in shape without any deformation or visual microcracks. The details are given as the following. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 96 Table 4.7. Microwave processing conditions in fabricating porous HAp ceramics. Batch I II in IV V Starting material A B B c* c* Wt. of HAp, g 1.5 2 .8 3.0 2 .8 2 .8 1200 1150 1200 1 3 5 Sintering temp., °C 1150 Sintering time, min. 1 1150 1 * Powder B plus ammonium carbonate. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 97 18.104.22.168. Porosity Apparent densities of the sintered ceramics vs. the green densities are listed in Table 4.8. Within the same batch, a linear relation between final and green densities is observed. Porosity can be controlled by adjusting the starting material, green density, or sintering time and/or temperature. The starting materials influence the final porosity because of the difference in sinterability and microstructure. The higher the specific surface area, the higher the sinterability. The specific surface area of powder B (finer powder) is much higher than that of A (whiskers), so that at the same green density, under the same sintering conditions, the pellets made o f powder B (batch n , Tables 4.7 and 4.8) achieved higher density than those of powder A (batch I). In addition, increasing processing temperature results in higher density, which can be seen by comparing batch II with III. Increasing sintering duration also has a similar effect (II and IV). The mercury intrusion porosimetry detects the open porosity of a sample. The porosimetry experiment was carried out on three representative samples (apparent density around 50%). Among them, the specimen made o f whiskers showed total open porosity, and those o f powders B and C had 82% and 84% open porosity, respectively. The difference is again due to the morphology and sinterability of the starting materials. 22.214.171.124. Phases XRD analysis made on the specimens after microwave processing shows that the samples of powder B (the ripened HAp) contain pure HAp, but those made o f the whiskers (powder A) show small amount of a-TCP, which is attributed to partial decomposition of the starting HAp during sintering. The use of ammonium carbonate, during compaction, effectively increased porosity (batch IV and V, Table II) without introducing any apparent contamination. W hen sugar was tried for the same purpose, however, some problems were encountered. First, sugar fuses during heating, thus making the samples deform or collapse, especially when the sugar/powder ratio is high. Secondly, because of the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 98 Table 4.8. Relative density (%) of microwave-processed porous HAp ceramics. Density n I Batch D j# d 2# III IV V Dl °2 Dl d2 Dl d2 Dl d2 43.6 52.9 34.2 51.4 34.1 58.0 16.3 26.7 15.5 30.3 49.3 55.8 39.0 65.8 39.0 77.1 25.5 43.3 23.2 44.5 51.8 68.8 44.5 68.7 44.5 81.9 27.3 51.2 27.3 52.5 72.0 48.3 82.4 48.7 91.2 32.2 56.6 30.4 58.4 37.6 65.3 36.3 69.6 40.4 69.0 38.8 75.5 54.3 # D j, D2, relative density before and after microwave processing. For IV and V, D j's are the values after removal of ammonium carbonate. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 99 insufficient oxygen and very rapid sintering, some carbon is trapped in the ceramics before it is com pletely oxidized, resulting in C-contam ination. Both the collapse and C- contamination (darkening) of the specimens were observed in the samples containing sugar. Obviously, ammonium carbonate is much more suitable for the purpose of introducing high porosity in the rapid sintering of HAp by microwave processing. 126.96.36.199. M icrostructure The m icrostructures of a few representative porous HAp ceramics are shown in Figure 4.25. It can be seen that the microstructures o f these ceramics are statistically homogeneous. In the whisker-made samples, some whiskers are still visible inside the cavity. Such a microstructure is expected to help the ingrowth of the tissue when used in biomedical applications. However, compared to the size (>100 p.m) required for ingrowth of bone tissue (Lavernia and Schoenung, 1991), the pore size should be much larger. Nevertheless, this should not to be difficult to achieve by m odifying the processing conditions described above. 188.8.131.52. Strength The tensile strength measured by the diametral-compression method is plotted as a function of density in Fig. 4.26. It can be seen that most porous HAp ceramics fabricated in the current research have shown "reasonable" strength values, while, of course, there is a trade-off between the porosity and the strength o f the HAp ceramics. 4.4.4. Summary on Microwave Processing of Porous HAp Ceramics Porous HAp ceramics with porosity up to 73% have been fabricated by microwave processing at 1150 to 1200°C for 1 to 5 min. The sintering can be accomplished within a few minutes under optimum conditions. For the specimens of about 50% porosity, mercury Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 100 1 9KU 19KU 803X 2 . 96KX 10U Figure 4.25. Microstructures o f the microwave-processed porous HAp ceramics (fracture surface), (a, b) Starting with fine HAp powder: (a) 1150°C, 3 min., porosity 73.3%. (b) 1200°C, 5 m in., porosity 24.5%. (c, d) Starting with HAp whiskers, pressed to different green densities (c: 43.58%, d: 54.29%), both fired at 1150°C for 1 min. Final porosity (c) 47%; (d) 28%. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 101 porosimetry shows that 82-100% of the pores are connected (open). The porosity of the ceramics can be controlled by adjusting the starting material, green density, sintering time or temperature. Pore size can also be adjusted. The addition o f ammonium carbonate in the green specimens helps in adjusting porosity. The microstructure o f the microwave sintered porous ceramics is homogeneous and the tensile strengths are reasonably good. 50 40 cS g 30 to G £ GO 20 ’S3 G £ 10 0 40 60 80 100 Density, % Figure 4.26. D iam etral tensile strength vs. density o f m icrow ave-processed HAp ceramics. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 102 4.5. Fabrication of Transparent HAp Ceramics For the first time, transparent hydroxyapatite ceramics have been fabricated at ambient pressure in this study by microwave and conventional sintering m ethods. This was achieved essentially using a fine crystalline hydroxyapatite pow der synthesized hydrothermally as a starting material. The sintered hydroxyapatite ceramic was phase pure and the average grain size was around 0 .2 (am. 4.5.1. Introduction Aoki et al. (1987) used a HAp percutaneous device for continuous and long-term blood pressure and deep body temperature measurement without introducing any infection. In fact, transparent HAp can be a better candidate for such an application, since it can work as a window to observe the changes inside. The com bination o f the excellent biocompatability and transparency makes the transparent HAp ceramics unique. It is expected that such ceramics will find good application in the biomedical area, and probably in some other areas as well. Some work on the fabrication of transparent HAp ceramics by hot isostatic pressing (HIP) has been reported (Yoshimura et al., 1989; Uem atru et al., 1989; Li and Hermansson, 1990). In these methods, either filter cake, hydrothermally synthesized ultrafine powder, or commercial HAp powder was used as starting material and the HIP processing conditions were 800 - 1275°C at the pressures o f 100 - 200 M Pa for 1 - 2 h. U nder the HIP conditions nearly complete densification with limited grain growth can be achieved at relatively low processing temperatures, so that transparent HAp ceramics could be obtained. However, the procedure of the HIP process and conditions are very complex and the experiment takes almost a whole day for completion. In the current study, a different approach, i.e., the ambient-pressure sintering, or, in other words, pressureless sintering, was adopted to fabricate transparent H A p ceramics. The experiments were carried out in air following conventional and microwave heating. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 103 4.5.2. Experimental Procedure The HAp powders synthesized in (Sec. 184.108.40.206.) by hydrothermal method (Powder I) and by the hydrolysis of brushite (C aH P0 4 -2H 20 ) followed by a "ripening" treatment (Powder II) were used in this study. Figure 4.27. shows the morphology of these starting powders . Before compaction, the powder was sieved to -325 mesh and heated at 500 °C. Circular pellets o f 12.7 mm diam eter and about 1 mm thick were uniaxially com pacted at pressure up to 350 MPa. As-pressed thin pellets from pow der I were translucent, and those of powder II were all opaque. For the conventional sintering, thin pellets of 6.35 mm diameter were pressed at 350 MPa. The microwave sintering was carried out in a modified 500-W microwave oven. The experimental details and temperature measurement have been described in Section 4.4. The samples were heated in the microwave oven in the air of 50% relative humidity directly from room temperature and sintered for 5 min. at 1150 °C, then allowed to air cool naturally. The conventional sintering was carried out in a dilatometer furnace in which heating rates could be well controlled through a computer program. This sample was heated at 5 °C per minute to 1150 °C and held at this temperature for 5 min., then allowed to cool down to room temperature, by turning off the power. 4.5.3. Results and Discussion Figure 4.28. shows the heating curves of the ambient-pressure sintering o f HAp ceramics. In the microwave sintering, the sintering temperature of 1150 °C had reached in 13 min. In the conventional sintering, it took 225 min. to reach the same temperature. A dilatometry study (Fig. 4.29.) showed that the sintering of HAp begins at around 670 °C. At the heating rate of 5 °C/min., it took 96 min. from 670 °C to 1150 °C. Thus there were actually 101 min. over 670 °C for sintering under the conventional condition. This was over ten times longer than that in the microwave processing (in which the sintering above 670 °C was only 8 min.). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 104 Figure 4.27. Morphology o f the HAp powders synthesized by (a) hydrothermal method, and (b) hydrolysis of C aH P04-2H20 plus ripening treatment. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 105 1200 000 u u <u 3 nia> a E <u H 800 600 400 Microwave Conventional 200 0 50 150 1 00 200 25 0 3 00 Tim e, m in. Figure 4.28. Heating curves o f sintering of hydroxyapatite ceramics by (a) microwave processing and (b) conventional sintering. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 106 0 300 - 2 .5 250 Percent Expansion 6 S 8 .1 - C -5 .0 200 i -7 .5 - 150 10.0 -12.5 8 8 3 .7 -C j * 00 -15.0 50 -17.5 - 20.0 250 500 I _L 750 1000 1250 ^tem perature ,°C Figure 4.29. Dilatometry curve of the hydrothermally synthesized HAp heating at 5 °C per min. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 107 The processing conditions are listed in Table 4.9. All the pellets sintered by both the m icrowave and the conventional processes were perfect in shape and pure in phase, but only the pellets prepared from Powder I turned transparent (Fig. 4.30). The specimens made from Powder II were all opaque. This indicates that the nature of starting material is crucial for achieving transparency in the HAp ceramics. Density measurement showed that the sintered specimens made from Powder I were nearly 100% dense (+99%), while those from Pow der II were only 93-95% of the theoretical density. The SEM micrographs of the as-sintered surface of the transparent HAp ceramics (Fig. 4.31) show that the specimens were well sintered, and the grain growth was very limited, with average grain size of 0.2 - 0.3 |xm. XRD pattern (Fig. 4.32.) indicates that the transparent HAp ceramics are highly crystalline, pure HAp. Transparency is an optical property of materials. For a specific m aterial, to be transparent it must not absorb light. The nature o f material is of course the most important factor that affects transparency. For example, it is impossible to achieve transparency in metals under normal conditions, since the numerous free electrons in metals will absorb the photons when visible light strike them. A glass is transparent because it has short-range ordered structure only, and also it is optically isotropic. There is no grain boundary in a glass, and hence very little scattering or absorption of light. Ceramics are generally polycrystalline. The grain boundaries in ceramics strongly scatter light. However, if the grain size is smaller than the wavelength of the visible light (0.4 - 0.7 p.m), light can transmit through the ceramic just like it travels through a grating. Owing to the difference in light absorption, the impurity phases in a ceramic will certainly affect transparency, usually decreasing transparency by scattering or absorption. Porosity also influences transparency in the same manner. In short, density, purity, and grain size are the key factors that influence the transparency of a ceramic. To achieve transparency in a ceramic, efforts should be made to eliminate or minimize scattering or absorption of light. In this Study the transparency was achieved by using the hydrothermally synthesized HAp powder (Powder I) which has the following characteristics: high purity, high thermal stability, fine and uniform particles, good crystallinity, excellent sinterablity, etc. These Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 108 Table 4.9. Processing conditions and results for microwave and conventional sintering of HAp. Powder Compactioni Process MPa Sintering Hold Total Processing temp., °C min. min. Sintered pellets I 350 microwave 1150 5 20 transparent n 350 microwave 1150 5 20 opaque i 350 conventional 1150 5 230 transparent n 350 conventional 1150 5 230 opaque Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 4.30. Transparent HAp ceramics fabricated at ambient pressure in air (a-c) by microwave processing for 5 min. at (a) 1150°C, (b) 1125'C, (c) 1100°C, totally irradiated for 18 min., and (d) by conventional sintering at 1150 "C for 5 min. after heating to 1150°C at 5 °C/min., totally heated for 230 min. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 110 111 0 0 1 9 MW2.4SB Figure 4.31. M icrographs o f the as-sintered surface o f the transparent hydroxyapatite ceramics sintered by (a) microwave processing and (b) conventional method. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Ill ■tftf 20 30 40 50 Two theta, degree Figure 4.32. Pow der XRD pattern of the transparent HAp ceramics by microwave processing at 1150 °C for 5 min., showing that the ceramics are highly crystalline single phase HAp. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 112 unique properties, as well as the relatively high green density (60%), and sintering without substantial grain growth, made it possible to realize transparency in the resultant HAp ceramics. By contrast, Powder II was composed of nonuniform and relatively large particles, with some large needle-like single crystals of aspect ratio o f showed lower packing efficiency and poorer sinterability. 10, so that this powder It was noticed that, at the compaction pressure of 350 MPa, the green density of the pellets made from Powder I reached 59.9% but that of Powder II was 54.9%. It is more difficult to fully densify the pellets made of the powder with the large needle-shaped particles, there were always some residual porosity in the pellets of Powder II after sintering. Besides, the hydrolysisderived HAp was probably less pure than the hydrothermally synthesized one, thus making it impossible to achieve transparency in the specimens made of Powder II. 4.5.4. Summary Transparent HAp ceramics were successfully fabricated at ambient pressure in air by m icrow ave as well as by conventional sintering o f the pow der-com pacts o f the hydrothermally synthesized HAp. The microwave sintering was accomplished in 5 min. at 1150 °C, and the total processing time was only about 20 min., while the conventional sintering took about 4 h. High quality of the starting HAp powder is a key factor to achieve transparency. Specifically, the fine nanocrystalline HAp prepared by the hydrothermal method has high purity, high thermal stability, high sinterability, which are essential to fabricate transparent HAp ceramics. 4.6. Densification Kinetics of HAp 4.6.1. Introduction It has been observed in the previous sections that by microwave processing, one can achieve significant enhancement in sintering or densification of HAp. In the case o f the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 113 sintering of regularly dense HAp, the microwave effect was observed as the enhanced densification; in the fabrication o f porous HAp ceramics, the microwave effect was mainly noticed as enhancement in sintering, since in the latter case, connection between particles to produce certain mechanical strength instead of densification was the main purpose of the processing. To evaluate the microwave effect on the densification of HAp, kinetics study was carried out. 4.6.2. Experimental Procedures Densification kinetics of HAp was studied under both microwave and conventional conditions. U sing the DCPD hydrolysis derived HAp (triply ripened with calcium chloride) as the starting material, pellets of 0.25 inch diameter were uniaxially pressed at 350 MPa, with appropriate PVA (2% solution) as the binder. The samples were sintered by microwave and conventional methods in a temperature range of 950 to 1200°C for 5 to 60 min. The microwave sintering setup was the modified 900-W furnace with a turntable (Sec. 3.2.). The conventional sintering setup was the one with comparable heating rates as the microwave furnace. The heating rate in both microwave and conventional cases were about 100°C per minute. The sintering time and temperature were the same in both cases. The relative densities after sintering were m easured by weight and dim ensional measurements to determine the kinetics of densification. Activation energy of densification under both microwave and conventional conditions were derived from Arrehnius formula. 4.6.3. Results and Discussion Densities of HAp sintered under microwave and conventional conditions in the temperature range of 950 to 1200°C and soaking time for up to 60 min. show that density increases as increasing sintering temperature or time. Figure 4.33 shows the sintered densities as a function of sintering time at various temperatures. In general, the influence of temperature on densification rate can be expressed by the Arrhenius Equation (Ye et al., 1986): Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 114 % 1 0 0 -i Relative den sity, 90 80 - mw1200°C m w 1100°C mw1000°C mw 950°C 60 - ◦ 0 10 20 30 40 50 60 Sintering time, min d en sity, % 100 80 cv1100°C cv1000°C cv 950°C Relative 70 - b 50 0 10 20 30 40 50 60 Sintering time, min. Figure 4.33. Sintered density as a function o f sintering time o f HAp sintered by (a) microwave and (b) conventional methods. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 115 K = A expC-Eg/RT) where K is the densification rate, A is a constant with the same dimension as K, E a is the apparent activation energy of densification, R the gas constant, and T the absolute temperature in Kelvin. Taking logarithm gives lnK = ln A - E a(RT )-1 Plotting InK versus 1/T gives Fig. 4.34. The tangent of the line in the plot is -Eg/R, from which the apparent activation energy Ea can be calculated. Using densification rates (Table 4.10) at 60% relative density, plot o f logarithm of densification rate (InK) versus reciprocal absolute temperature (1/T) was made (Fig. 4.34) It is noticed that the apparent activation energy for the densification o f HAp under the microwave processing condition is about 40% lower than that for conventional sintering. Since microwave sintering effect o f any material directly depend on the dielectric properties of the material. The lower activation energy of HAp may indicate a relatively high dielectric loss factor of this material, which might be attributed to the dipoles of hydroxyl ions in the crystal structure of HAp. The tangent values for the lines for conventional and microwave sintering are -2.1737 and -1.3452, respectively. Accordingly, the apparent activation energies for conventional and m icrowave sintering are 180.7 and 111.8 kJ/m ole, respectively. 4.6.4. Summary on the Densification Kinetics of HAp Densification kinetics study of HAp in temperature range from 950 to 1200°C under microwave and conventional conditions shows that microwave processing apparently enhances densification of HAp. A t relative density of 60%, the apparent activation energy for microwave sintering o f HAp is 112 kJ/mole, which is about 40% lower than the conventional process (181 kJ/mole). Since the sintering temperature, time, and heating rate Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 116 in microwave and conventional processes were comparable, the lower activation energy indirectly indicates that HAp has fair dielectric loss properties to effectively couple with microwaves to enhance its densification. Table 4.10. The InA values from the densification curves for Arrhenius plot at relative density 60%. Microwave Conventional 1/T K (l/m in.) InK K (l/m in.) 950°C (1223K) 8 .1 7 7 x l0 '4 0.0241 -3.7232 0.0160 -4.1382 1000°C (1273K) 7 .85 5 x l0 ’4 0.0588 -2.8336 0.0417 -3.1780 1100°C (1373K) 7.283x1 O' 4 0.1163 -2.1514 1200°C (1473K) 6.789xl0-4 Temperature 0.1795 InK -1.7176 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 117 y = 13.737-2.1737x RA2 10000(1 fT), 1/K' Figure 4.34. Arrhenius plot o f the densification o f HAp under (a) microwave and (b) conventional conditions. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 118 Chapter 5 MICROWAVE PROCESSING OF MULLITE CERAMICS 5.1. Introduction Mullite, 3 A l 20 3 -2 Si0 2 , is one of the m ost commonly used ceramic refractory materials. It has been considered an important material for high temperature structure applications. The favored role of mullite for this is mainly due to its high melting point (1828±10°C, M ah et al., 1983), low thermal expansion coefficient, low therm al conductivity, excellent resistance against therm al-shock, and good high-tem perature mechanical properties (e.g., high strength and good creep resistance). Mullite is also characterized by its low dielectric constant (e = 6.7, compared to 9.8 for AI2 Q 3) and thus it is a promising candidate for substrates in electronic packaging (Aksay et al., 1991). Moreover, mullite has a good transmission at high temperature in middle IR band, so that it can be used as windows in middle IR spectrum, especially at high temperatures. Low dielectric constant and optical transmittance of fine-grained polycrystalline mullite has also triggered interest in its potential applications as a solid-state laser activator (Aksay et al., 1991). However, it is not easy to obtain dense mullite because the temperature required for the sintering is very high, usually in the range o f 1600 - 1700 °C under conventional conditions. High purity mullite powders are not easily found in the market compared to silicon oxide and aluminum oxide powders. The use of reaction sintering of mullite is an attractive approach (Rodrigo and Boch, 1985) to prepare dense mullite ceramics. Reaction sintering of mullite may be carried out with fine crystalline alumina and silica powders as starting material, but the sintering temperatures in such a case are very high, because both the solidstate reaction to form mullite and the subsequent sintering are sluggish at relatively low tem peratures. The sol-gel m ethod is a popular approach to obtain fine, pure, homogeneous, and reactive mullite precursor powders. Both monophasic (polymeric) and Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 119 diphasic (colloidal) mullite gels are used in ceramic fabrication. Monophasic is a mullite precursor mixed at molecular scale. Complete conversion of this kind of gel to crystalline mullite can be achieved below 1000 °C (at about 970 °C). Correspondingly, the activation energy of nucleation for this conversion was found to be 293-362 kJ/m ol (Li and Thomson, 1991), but its sintering after crystallization requires temperatures around 16001700 °C or hot pressing (Huling and Messing, 1989). Diphasic mullite gel is a mullite precursor m ixed at nanometer scale. The crystallization temperature o f a colloidal boehmite-silica diphasic gel was observed as high as 1364 °C (Huling and Messing, 1989). Wei and Halloran (1988) calculated the activation energy of nucleation in this case to be 987±63 kJ/mol. The higher activation energy slows the kinetics of diphasic mullite gel and allows it to be densified before crystallization to high density by viscous deformation of amorphous silica (Wei and Halloran, 1988). In this way, it is possible to obtain sintered mullite at lower temperatures with diphasic gel. In order to fabricate high-performance mullite ceramics, it is necessary to tailor their microstructures. Interlocked acicular grain structure, for example, is necessary when creep resistance is desired, while fine and equiaxed grained mullite is suitable for IR transparence or room temperature strength (Mroz andLaughner, 1989). It is well known that by adding seeds to a precursor, crystallization direction can be preferentially controlled, or crystallization temperatures can be lowered, since the energy barrier for nucleation can be overcome in this way. The crystallization temperature of monophasic mullite gel is significantly lower than that of diphasic mullite gel. When appropriate amount of momophasic mullite gel is mixed with diphasic mullite gel and heated, the crystallization o f monophasic gel at lower temperature will provide diphasic mullite gel with many mullite crystallites which would then work as in-situ seeds and promote the crystallization of diphasic mullite gel. Since the mixing is conducted in the wet state, obviously it is easier to obtain best homogeneity and finest seeds. Huling and Messing (1989) have reported their results on nucleation of such a gel mixture knows as hybrid gel. They found that seeding diphasic gel with monophasic gel can effectively decreased the grain size of the sintered mullite and remove intragranular pores usually trapped in the sintered body, although the decrease in the m ullite Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 120 transformation temperature of diphasic mullite is limited (about 30 °C). Microwave sintering usually offers a controllable grain-growth and more uniform and finer grain sizes compared to the conventional sintering. The objective of this investigation was to study the sintering behavior of m ullite in m icrowave processing, including densification and mullite transformation of seeded and unseeded mullite gels. 5.2. Synthesis of Mullite Precursors Sol-gel derived precursors are characterized by high purity, high surface area, high activity, and high homogeneity. Mullite precursors used in this study were synthesized by sol-gel methods. The mullite gels are the aluminosilicate gels with mullite stoichiometry (3Al2 0 3 -2Si02), including monophasic, diphasic, and seeded gels. By different drying methods, xerogels and aerogels were obtained. Xerogels were obtained by drying the gels in air, while aerogels were obtained by supercritical drying. The details will be described in the following sections. 5.2.1. Monophasic Gel Monophasic mullite gel was prepared by mixing stoichiometric amounts (mullite, 3 Al20 3 -2 Si0 2 ) of tetraethyl orthosilicate (TEOS) and Al ( N 0 3 )2 -9H20 dissolved in absolute ethanol at room temperature, gelling at 60 °C, drying at 120 °C, and denitrating for 4 h at 465 °C. The xerogel thus obtained was ground with an agate mortar and sieved to 325 mesh (45 pm opening). The bulk density of the dried powder is 2.03 g/cm3. 5.2.2. Diphasic Gel W ith addition of concentrated nitrate acid (H N 0 3) to get pH 2.0 (at room temperature) in water, boehmite (AlOOH) was allowed to hydrolyze in the acidic water overnight at 95 °C under reflux condition. The solid/water ratio was about 1:50. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 121 Translucent boehmite sol was obtained at the end of the hydrolysis. The boehmite sol was mixed by vigorous stirring with an ammonium-stabilized colloidal silica sol (Ludox, DuPont Chemicals, Wilmington, DE) at room temperature. The A l 2 0^/SiO 2 molar ratio was 3:2. The mixed sol was heated at 60 °C with continuous magnetic stirring to remove excessive water, and this finally gelled at the same temperature by prolonged sitting. The gel was dried at 110 °C then denitrated at 550 °C, and then ground and sieved to fine powder o f -325 mesh. The bulk density of the powder was 2.94 g/cm3. The heating process at 550°C was necessary to transform boehmite to y-Al2 0} and prevent cracking of the pellets. In the earlier trials, it was found that all the pellets pressed from the gel powder dehydrated at 400 °C cracked after microwave sintering, but those made the powder calcined at 700 °C were intact after the same process. It is believed that the cracking of the pellets of the 400 °C-heated powder occurred due to the transition of boehmite into yA12 Q 3. Further experiment showed that the minimum temperature for the denitration was 500°C. Therefore, in all the subsequent experiments, the powders were calcined at 500 550 °C. The heating time and temperature were limited to a minimum value to preserve the highest possible specific surface area, ensuring the best reactivity of the resulting powder. 5.2.3. Seeded Diphasic Gels In order to study the influence of seeding on the crystallization and densification of diphasic m ullite gel in m icrowave processing, diphasic gels were seeded with (a) monophasic gel and (b) crystalline mullite as seeds. For gel seeding, the seeding (monophasic) gel as prepared in Sec. 5.2.1. (0.5 to 30% by weight of dry material) was homogeneously mixed with the diphasic gel matrix by magnet stirring when the gel was wet. The seeded gel was dried at 60°C followed by 120°C to get xerogel. The xerogel was further dehydrated and denitrated at 550°C after grinding into fine powder. For crystalline mullite seeding, submicron mullite seeds were prepared by grinding crystalline mullite and separating appropriate size by decantation in deionized water. The Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 122 fine mullite particles in the supernatant were collected by drying as seeds. The seeds of 2 or 5% by weight were added to the diphasic sols as prepared in Sec. 5.2.2. before gelation. The drying and powder-making processes were identical to that in monophasic gel seeding. 5.2.4. Diphasic Mullite Aerogel Aerogels are porous materials produced by a sol-gel process followed by supercritical drying o f the wet gel in an autoclave (Fricke & Emmerling). Aerogels have a spongelike, open-pore structure with a large inner surface. The pore size o f aerogels is in the range between 1-100 nm. Monolithic aerogels are used in Cerenkov detectors in high-energy physics; translucent granular aerogels can be used as insulator for the purpose of passive solar energy usage; opacified aerogel powders are being tested as substitutes for chlorofluorocarbon-blown polyurethane foams and as thermal super insulations in heatstorage systems; experiments have been performed with aerogels as catalytic substrates, acoustic impedance matching layers, containment for fusion fuels, and gas filters; and aerogels are also used in radioluminescent light and energy sources. Since aerogels free from agglomeration, and still possess high surface area which means higher activity, the sinterability of mullite aerogel is very high. Rahaman et al. (1988) sintered the mechanically compacted aerogels prepared by supercritical extraction with C 0 2 to nearly theoretical density below 1200°C. In this study, mullite aerogel was prepared by supercritical extraction with methanol (CH 3 OH). In the preparation, the same wet diphasic gel as prepared for the diphasic xerogel (Sec. 5.2.2.) was introduced in glass cells and immersed in a methanol (CH3 OH) bath at room temperature to exchange the water trapped in the gels. After 24 h, the immersing liquid was replaced by fresh absolute methanol to allow further exchange to take place. The exchange was repeatedly carried out three times. After the exchange, the cells were placed in an autoclave with 17-20 mL methanol in it fo r supercritical drying. The supercritical drying was conducted at 260°C,1200 p.s.i., which was slightly above the supercritical conditions o f methanol, 240°C, 1155 p.s.i.. The system was heated from room temperature to 260°C in 3 h, and soaked at 260°C for 30 min. Then the methanol was extracted under the supercritical Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 123 conditions under flowing argon. Since the methanol is removed under the supercritical conditions, at which there is no distinction between liquid and gaseous phases, and absence of surface tension, practically no shrinkage, thus highly porous and translucent aerogels are resulted. 5.3. Mullite Transformation o f Diphasic Gels 5.3.1. Introduction Since the particle size of the diphasic mullite gels is in nanometer scale, so that the mullitization temperature of diphasic gel is significantly higher than that of monophasic gel. Now that microwave processing can enhance sintering of many ceramic materials, it may also affect the mullitization of diphasic mullite gel. In the current study, microwave effect on the crystallization of seeded and unseeded diphasic gels was investigated. 5.3.2. Experimental Procedure The microwave effect on the crystallization of diphasic aluminosilicate gel was studied using the diphasic xerogel. Influence of seeding on the crystallization was also studied. The hybrid gels as described in Sec. 5.2.3. containing up to 30% monophasic gel were used for this study. For convenience, the starting powders were pressed into pellets of 0.25 in. diameter. Samples were heated in a microwave field as well as in the conventional furnace at various temperatures between 1250 to 1400°C for 10 min. Both qualitative and quantitative XRD analyses were performed on the samples heated under various conditions. For the quantitative analysis, a-alum ina was used as the internal standard. The XRD integrated intensities o f mullite at (210) and that of corundum at (113) were used to evaluate the degree o f conversion of the diphasic gel to the crystalline mullite. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 124 5. 3.3. Results and Discussion 220.127.116.11. Crystallization of the Gels Crystallization o f an amorphous phase is a process toward a state of low er free energy, so that it is an exothermic process. Therefore, the exothermic peaks on the DTA curves correspond to the crystallization temperature. DTA study shows that monophasic mullite gel crystallizes at 970°C (Fig. 5.1.), and diphasic gel crystallizes at 1328°C (Fig. 5.2.). The difference in the crystallization temperature can be explained by the degree of mixing scale of the compositions. Monophasic gel is an amorphous precursor with atomicscale arrangement of -Al-O-Si- groupings very similar to that of mullite (Sunderesan & Aksay, 1991), so that it readily crystallizes to m ullite at low er tem perature. The homogeneity between alumina and silica sources in diphasic gel is at nanometer scale, thus the crystallization is controlled by the reaction between the two phases and requires a higher temperature. Qualitative XRD studies show that under conventional conditions, diphasic mullite gel transforms to crystalline mullite at around 1325°C by conventional heating (Fig. 5.3), whereas by microwave heating, crystalline mullite was detected in the sample heated at 1250°C (Fig. 5.4.). This indicates that m icrow ave heating accelerates m ullite transformation o f diphasic gel. The appearance o f phases as sequence of temperature, however, was the same in both conventional and microwave processing. Silica remains amorphous before mullitization, while alumina exists as 8-Al2 0 3 before mullitization. Therefore mullite transformation of this diphasic gel is the reaction between 5-Al2 0 3 and amorphous silica. This result is in agreement with Wei and Halloran (1988). 18.104.22.168. Seeding Effect of Diphasic Gels The results of quantitative XRD o f the conventionally and microwave fired diphasic mullite gels are listed in Tables 5.1. and 5.2., respectively. Seeding with 5% or less mullite Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 125 a. qo a a z DEGREES UJ V 4.00' a oo 60.00 190.00 320.00 450.00 5B0.00 710.00 TE M PER A TU R E F igure 5,1. 970.00 CC) DTA therm ogram o f m onophasic alum inosilicate gel, show ing that crystallization of the gel takes place at around 970°C. Heating rate was 10°C per minute. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 126 100 1328*C LU UJ oz (J UJ □ acoi5un 22aao 34aoo uaoo aaoo m oo oka loeam ukoo wdaoo TEMPERATURE <C) Figure 5.2. DTA thermogram o f diphasic gel. The exothermal peak at 1328'C indicates the occurrence of mullitization. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 127 soo c 1300X5 ****"* »", 1325X 5 ... . T w o theta, d e g re e Figure 5.3. XRD patterns o f diphasic mullite gels heated conventionally, indicating that mullitization did not occur below about 1325*C. Phases are (a) amorphous, (b) 8-Al2 0 3, (c) mullite + 8 -AI2 O3 , and (d) mullite. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 128 500 DRIED 1250X5 MW 1300C 20 10 30 40 00 00 70 Two theta, degree Figure 5.4. XRD patterns showing that crystallization o f the diphasic mullite gel starts at 1250°C when subjected to microwave irradiation. Detected phases are (a) amorphous, (b) 5-Al2 0 3, (c) mullite + 5-Al2 0 3, and (d) mullite. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 129 Table 5.1. Mullite contents (%) in the conventionally heated diphasic mullite gel. seeds% 0 0.5 1 5 10 20 1250°C 0 0 0 0 0 0 — 1275°C 0 0 0 0 0 0 14.33 1300°C 0 0 0 0 25.73 32.23 2 1 .0 0 1325°C 75.1 81.85 67.81 54.12 74.01 56.54 31.85 1350°C 77.24 83.66 66.36 55.72 55.26 26.54 1400°C 91.74 83.98 65.78 53.53 27.07 85.79 30 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 130 Table 5.2. Mullite contents (%) in the microwave heated diphasic mullite gel. seeds% 0 0.5 1 5 10 20 30 1250°C 0 34.11 31.81 37.16 52.00 19.11 1275°C 55.93 79.94 74.50 40.51 71.27 45.96 21.57 1300°C 79.90 77.66 69.58 48.69 82.29 49.94 25.41 1325°C 84.75 72.67 67.85 54.17 85.05 52.15 29.11 1350°C 87.41 90.77 71.41 55.95 98.06 57.28 25.48 1400°C 91.86 84.25 93.36 69.42 90.53 64.87 34.48 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 131 seeds did not show any improvement in crystallization. W eak XRD intensities of mullite show up in the samples with 10% or more seeds, when fired at 1300°C. At or above 1325°C, the XRD intensities become stronger. At the same processing temperature, the rate of mullite conversion decreases as the content of seeds increases. For the same amount of seeds, conversion rate increases with increasing temperature. Under microwave conditions, crystalline mullite phase was detected on the sample heated at 1250°C (Fig. 5.3). W ithout seeds, the conversion rate was about 60% at 1275°C, and about 92% at 1400°C. In the seeded samples, even with only 0.5% seeds, 34% of the gel got converted to mullite at 1250°C. At the same temperature, for the sample containing same amount of seeds, microwave processed samples always show higher conversion rate. Therefore, microwave irradiation promotes seeding effect on diphasic gel. The conversion rate increases as increasing amount o f seeds up to 10%, and then decreases. This can be explained from the point o f view of nutrient (gel) available to the seeds for the growth of the crystallites. The process o f mullitization of the gel includes nucleation followed by crystal growth. In the seeded gel, mullite crystallites converted from monophasic gel at lower temperatures function as seeds in the diphasic gel, so that at higher temperatures, the diphasic gel around each seed will supply crystal nutrient to the seed with A12 0 3 to S i0 2 ratio of mullite stoichiometry to form crystalline mullite. This is an epitaxy process. Within the limitation o f the crystal nutrient available to each seed to grow to a mature crystal, increase seeds will promote mullite transformation, when the numerical concentration o f seeds is too high, the nutrient available to each seed decreases to such a level that no well defined crystals will be obtained. This explains why in microwave processing o f seeded diphasic mullite gel at 1250°C, XRD intensity increases with seeds to 10%, then decreases with further increase in seeds. With 30% seeds, mullite could not even be detected by XRD. Compared with the case of microwave processing in which mullite was detected at 1250°C, it is seen that the m icrowave enhancem ent in crystallization o f diphasic aluminosilicate gels o f mullite stoichiometry is about 75-100°C. Although it is well known that the detecting limitation of XRD for a crystalline phase is about 5%, XRD did not reveal the existence of the seeds converted from the seeding Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 132 (monophasic) gel in the diphasic matrix, even in amount up to 30%, when conventionally fired at temperatures below 1325°C. Since the monophasic gel crystallizes at 970°C, there is no reason to doubt about the existence of crystalline mullite from the monophasic phase after firing at such high temperatures. There are at least two reasons to explain this phenomenon. First, the crystallites were too small to induce X-ray diffraction but variable amount of line broadening would occur. Second, the microabsorption effect (Cullity, 1978) o f the untransformed diphasic gels would further weaken the diffraction. The final result is that the XRD intensities totally disappeared. 5.4. Microwave Sintering o f Mullite Ceramics (Regular and Transparent) 5.4.1. Introduction Due to the low diffusion rate of the ions in mullite ions, the sintering of mullite ceramics generally requires very high temperature (1600-1700°C) and long sintering times (at least a few hours). Reaction sintering may apparently lower the sintering temperature for mullite. Crystalline alumina and silica powders may be used for the reaction sintering to obtain mullite, but high temperature is still required in this case since both components are in crystalline state and the homogeneity between the reactants is very limited due to the existence o f high activation energy barrier. Adopting an approach of diphasic gel, however, the mullite precursor can be prepared with high reactivity and high homogeneity to favor reaction sintering. In this study, microwave sintering o f mullite ceramics was conducted using diphasic gels as the starting material, and also using monophasic gel and crystalline materials for comparison. Transparent mullite ceramics were sintered using the mechanically compacted diphasic aerogel, with the xerogel o f the same composition as comparison. 5.4.2. Experimental Procedure The monophasic and diphasic gels as well as the seeded diphasic gel as described in Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 133 Sec. 5.2. were used for sintering experiments. For comparison, fine powder o f crystalline mullite (from Chichibu Cement Co., Ltd., Japan) was also used in one experiment. The m icrowave sintering was carried out in the m odified 900-W, 2.45 GHz m ultim ode microwave furnace. Conventional sintering was conducted in a high temperature, high heating rate electric furnace. Both sintering setups have been described in detail in Sec. 3.2 and 3.3. For the best densification effect, step sintering was adopted for samples of diphasic gels, in which samples were processed below m ullitization tem perature for densification by viscous flow and then crystallized at elevated temperatures. Various sintering schedules were used. Appropriate amount of PVA (2% solution) was used as a binder during compaction. For transparent mullite ceramics, the diphasic aerogel as described in Sec. 5.2.5 was used as starting material. Thin circular pellets of 0.25 or 0.5 inch diameter were compacted by uniaxial pressing at 280-350 MPa. To avoid contamination, no binder was used. The as-compacted green pellets were translucent. Microwave sintering o f transparent mullite was carried out at 1300°C for 10 min. Conventional sintering was carried out in a dilatometer furnace heating at 5°C/min. to 1320°C. Once the specimen temperature reached 1320°C, the power was shut o ff to let the furnace cool down naturally. The peak temperature was so chosen to provide best densification with least crystallization, because the diphasic gel transforms to mullite at 1330°C. Thermal mechanical behavior of the monophasic gel, diphasic gel, and diphasic aerogel, was studied by dilatometry. Fine powders of the above three gels were compacted into pellets of 0.25 in. diameter and then subjected to heating in a dilatometry furnace (Sec. 3.4.4.) at 3°C/min. from room temperature to 1500°C. During heating, the change in thickness of the specimen as a percentage of the original dimension was continuously monitored and recorded by the computer of the system. The thermal mechanical analysis (TMA) curve was then plotted. 5.4.3. Results and Discussion Tables 5.3. through 5.9 list the density results of the mullite samples sintered under Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 134 various conditions, including influence of starting material, green density, and sintering schedule on densification o f mullite. It is seen that the crystalline mullite and monophasic mullite gel are difficult to sinter at temperatures up to 1600°C, while diphasic mullite gel can be sintered to around 95% relative density. The sinterability of crystalline mullite precursor is the lowest; that o f m onophasic gel is also poor, because it converts to crystalline form below the sintering temperature; the sinterability of diphasic gel is quite high, since it densities before crystallization; the sinterability of the mixture of diphasic gel and monophasic gel of equal amounts is in between that of diphasic and monophasic gels. The sinterability difference between monophasic and diphasic gels is more clear from the TMA curves (Fig. 5.5.). The first sharp shrinkage of monophasic gel takes place at 970°C, corresponding to the crystallization of the gel (Fig. 5.1.). After this, the shrinkage with temperature slows down. The initial shrinkage of diphasic xerogel starts at about 700°C through the viscous flow of silica gel. Then obvious sintering takes place at about 1050°C when the viscosity of silica gel further lowers at elevated temperatures. This process continues till around 1300°C, presumably as a result of nucleation of mullite crystals. As mullite transformation takes place at 1328°C (Fig. 5.2.), the viscous flow mechanism slows down, and so does the sintering. The shrinkage of diphasic gel is much greater and takes place at higher temperature than that of monophasic gel, indicating that the diphasic gel favors sintering of mullite ceramics. It is also found that seeding does not enhance densification of the diphasic mullite gel. This is due to the fact that seeding lowers m ullitization tem perature and enhances conversion of diphasic gel to crystalline mullite which has low diffusion rates than the gel itself. As expected, increasing compaction pressure or green density increases sintered density. Step sintering works well for diphasic mullite gel, since it allows the diphasic gel to densify before crystallization and the densification of the gel by viscous deformation is more efficient than the solid state diffusion of the crystalline mullite. Figures 5.6-8 show the microstructures of sintered mullite samples of diphasic gel. The samples in Fig. 5.6 were polished and thermally etched at 1400°C for 30 min. The microstructure of the microwave sintered specimen (Fig. 5.6.a) is more dense than the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 135 Table 5.3. Relative density (%) of monophasic mullite gel sintered at 1200°C x 20 min. compaction, M Pa microwave conventional 40 140 280 420 560 700 840 50.04 49.52 51.72 54.07 52.78 56.76 53.44 57.05 56.16 46.16 46.01 46.75 Table 5.4. Comparison of sintered density of different starting materials (microwave sintering: 1300°C x 10 min. + 1600°Cx 10 min.). green density, % sintered density, % crystalline monophasic 59.58 40.15 66.99 67.80 diphasic 44.66 91.40 starting material Table 5.5. Density of mullite conventionally sintered (1550°C x 20 min., 3h total heating). starting materials green density, % sintered density, % diphasic gels (D) monophasic (M) 47.14 40.15 90.60 67.48 D : M = 50 : 50 47.64 74.93 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 136 Table 5.6. Relative density (%) of the seeded diphasic mullite gels after sintering. ________ 1200°C___________________ 1200x20* 1200x40 10 min. +1400x5 +1400x5 +1400x5 +1500x20 +1500x20 20 min. 40 min. 60 min. 1200x60 1200x20 1200x40 m icrow ave no seed 76.94 79.05 81.14 86.09 89.64 90.58 88.91 89.41 90.72 2.5% 74.61 78.93 82.40 83.32 84.76 84.79 85.22 85.73 85.36 5.0% 73.76 81.21 81.35 83.24 84.85 85.59 85.82 85.27 85.79 D:M=1 63.73 65.25 66.52 67.64 67.21 67.17 67.79 71.57 72.49 87.41 92.04 91.22 conventional no seed 61.79 66.53 70.09 70.35 90.51 2.5% 55.98 62.39 65.43 68.34 84.97 — 85.68 87.61 87.88 5.0% 56.67 64.74 65.45 71.34 86.13 — 83.29 87.02 86.49 D:M=1 52.65 55.25 57.46 60.27 67.71 — 67.38 70.78 70.76 * Temperature (°C) x time (min.). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 137 Table 5.7. Sintered density of seeded diphasic mullite gel (microwave sintering). densitv. % Seed% green no seeds 2 .0 % 5.0% sintered green sintered 1300°C x 20 min. 90.06 45.53 1450°C x 2 0 min. 48.20 94.10 44.87 44.89 47.07 47.22 87.91 87.84 92.00 91.00 Table 5.8. Relative density of diphasic mullite gel after microwave step sintering (1200°C x 60 min. + 1500°C x 20 min.). comp, pressure (MPa) sintered density (%) 140 210 280 420 560 82.93 86.46 89.08 89.93 91.15 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 138 Table 5.9. Density of unseeded diphasic mullite gel under various microwave sintering schedules. Sintering schedule (°C x min. + °C x min.) green density, % sintered density, % 1300 x 15 + 1400 x 5 46.24 87.22 1300 x 15 + 1500 x 20 44.98 87.56 1300 x 30 + 1475 x 20 44.48 86.73 1300 x 30 + 1550 x 20 47.47 94.63 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 139 10 5 200 Tine Percent (nin) Expansion 400 -1 5 0 250 500 750 1000 1250 1500 Tenperature (°C) o 500 400 Percent Tine 1 2 9 2 . »C Cnin) Expansion -20 B Tenperature (°C) Fig. 5.5. Dilatometry curves of (a) monophasic and (b) diphasic mullite xerogels heated at 3°C/min. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 140 conventional sample (Fig.5.6.b). A lamellar structure in the microwave sintered sample is obvious, while the sample conventionally sintered does not show the same ordered structure. Figure 5.7. shows the as-sintered surface (a) and fracture microstructure (b) of a diphasic mullite sample sintered by microwave processing at 1300°C for 20 min. The relative density of this sample was 94%. The grain size ranges between 1-5 |im. The two specimens shown in Fig. 5.8. were step-sintered conventionally (1250°C x 30 min. followed by 1500°C x 30 min.). Sample (a) was pure diphasic gel, while (b) was seeded by 5% mullite seeds. It can be seen that the average grain size of the seeded sample was significantly smaller. This is attributed to the increased nucleation frequency. The observation is in good agreement with what reported by Huling and Messing (1989). The residual porosity of about 5% may be attributed to the intragranular pores. Figure 5.9. shows the mullite ceramic pellets sintered by (a) microwave and (b-d) conventional methods. Among them, only the samples made from aerogel (a, b, and d) are transparent. Sample (c), made from diphasic xerogel, is opaque, although it was sintered under identical conditions (conventional, 5°C/min. to 1320°C, no holding) as (b) and (d). Dilatometry study (Fig. 5.10.) shows that except for an additional shrinkage below 479°C, the thermal mechanical behavior of the diphasic aerogel is basically the same as the xerogel (Fig. 5.5.b). The agglom eration in the powders decreases hom ogeneity and hinders the achievement of full densification. It introduces additional intergranular porosity which may partially rem ain in the sintered body as residual porosity. The large porosity and inhomogeneity of microstructure strongly scatter visible light, so that transparent mullite was not achieved with the diphasic xerogel. The aerogel preparation is a key for the fabrication of the transparent mullite ceramics, since it prevents the gel from agglomeration. Taking 3.56, 2.2, and 3.17 as the densities of 8-Al2 0 3, amorphous silica, and crystalline mullite, respectively, calculation shows that the mullite transformation will leave about 3% porosity in the mullite ceramics made from pure diphasic mullite gel. These residual pores will decrease transparency in the case of transparent mullite ceramics. This problem may be solved by seeding. Huling and Messing (1989) reported that sintering the unmodified Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 141 Figures 5.6 M icrostructures of diphasic m ullite gel step-sintered (1200°C/30 min. + 1500'Cx20 min.) by (a) microwave and (b) conventional method. The samples were polished and thermally etched at 1400*C for 30 min. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 142 Figure 5.7. Microstrucure of (a) as-sintered surface and (b) fracture surface o f a diphasic mullite sample microwave sintered for 20 min. at 1300*C to relative density o f 94%. i Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 143 2 5 KU 1 0 . 5KX 1U 00 0 5 YF230D Fig. 5.8. Microstructure of conventionally sintered (1250°Cx30 min. + 1500°Cx30 min.) specimen of (a) pure and (b) seeded (5%) diphasic mullite gels. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 144 Transparent Mullite Ceramics Ceramics 4 . Figure 5.9. M ullite ceramics sintered by (a) microwave (1300*Cxl0 min.) and (b, c, d) conventional (b, c: 1300’CxlO min.; d: 5'C/m in. to 1320*C without holding) methods. Sample (c), made from diphasic xerogel is opaque, while the rest made from aerogel show transparency. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 145 o Tine (n ln ) sag a 258 508 250 to w 1258 1588 Tenperature (°C) Figure 5.10. Dilatometry curve o f diphasic mullite aerogel heated at 3 °C/min. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 146 colloidal (diphasic) gel for 2 h at 1550°C leaves many faceted intragranular pores that are removed only by increasing the sintering time or temperature. In contrast, they found that the increased nucleation frequency and resulting increased grain-boundary area in the hybrid gels (diphasic gel seeded by monophasic gel) allows short-range vacancy diffusion to promote pore removal at lower temperatures or shorter times. Thus by using hybrid gel, the transparency of transparent mullite ceramics could be further improved by effectively remove the intragranular pores. XRD study indicated that the transparent sample sintered by conventional method was not mullitized, while that sintered by the microwave method was crystalline. This is in good agreement with the results on the crystallization of diphasic gels, i.e., mullitization takes place at low er temperatures in the m icrowave processing than in conventional sintering. However, it is interesting that mullitization in the microwave processed sample did not reduce the transparency, showing that transparent mullite ceramics can be fabricated directly by sintering. 5.5. Densification Kinetics of Diphasic Mullite Gel 5.5.1. Introduction It has been seen that the densification o f diphasic alum inosilicate gels under m icrowave conditions is substantially enhanced. To better understand and evaluate the microwave effect, a kinetic study was carried out. Since the densification of the diphasic gels is mainly attributed to the viscous sintering of the gels, the firing process in the kinetics study was limited below the mullitization temperature. This means that there is no crystallization occurring during the firing so that the kinetics applies only to a single viscous densification process. 5.5.2. Experimental Procedure Pellets of quarter inch diameter were pressed at 350 M Pa unaxially, then sintered at Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 147 1100, 1150, 1200, 1250, and 1300°C for various durations, under microwave and conventional conditions, respectively. Same sintering setups as described in Sec. 5.3. and 5.4. were used. Densities after sintering were measured and densification (density versus time) curves were plotted. By plotting the natural logarithm of densification rate versus the inverse absolute temperature (1/T), the apparent activation energies for densification were derived from Arrhenius equation. 5.5.3. Results and Discussion The densification curves o f diphasic m ullite gel are shown in Fig. 5.11. The densities achieved by microwave processing are always higher than that by conventional sintering (Fig.5.12). For example, at 1100°C, the density (48.54%) after 5 min. of microwave processing is equal to that by conventional sintering for 60 min.; at 1200°C, 60 min. of conventional sintering gave 67.54% relative density, while microwave processing for 10 min. achieved 68.52%, and 78.16% at 60 min. Similarly, the density achieved at 1250°C by 60 min. of conventional sintering is 81.74%, which could be obtained within 10 min. by microwave processing. At 1300°C, the difference is less obvious. In microwave processing, the densification in the temperature range from 1150°C to 1250°C is very significant. Beyond this range, the density increase with temperature is much smaller because of mullite transformation. Under conventional conditions, however, the density change between 1250°C and 1300°C is still large, since mullitization has not started in this case. By the same treatment as in Sec. 4.6.3., the densification rates K as the tangents of the densification curves (Fig.5.11.) at 50% relative density at various temperatures. The values are listed in Table 5.10. Plotting InK versus 1/T, two lines are obtained (Fig. 5.13.). For microwave processing, InK = 29.635 - 4.7675 T 'l Accordingly, the apparent activation energy Ea = 4.7675 x 104 x 8.314 = 396.4 (kJ/mole) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 148 «niiM‘,*l,,,1Bt IIIIIIIIIH I« ',,,1 m ,,,I ,,,,,I I ,I I A o ,,IIMI — O — (A C mnHij^iiiini •a —O " a> 0) > ro a) DC •IIIW#""'*1 mw1300°C mw1250°C mw1200°C mw1150°C mw1100°C Sintering time, min. Q uihh111111" ■' 0 ■ ■mitiQiiiiii (0 c a> •a > *-> cv1300°C cv1250°C ■"A cv1200°C IIIIWI0 IIIIIH cv1100°C llllllllllll■••|■■l■■•■ln" ," ,l,,,,,,,l,,ll^ ll," , ................. • <u DC 20 40 Sintering time, min. Figure 5.11. Densification curves of diphasic m ullite gel under (a) microwave and (b) conventional conditions. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 149 54 80 52 jfi 50 70 5k «C *ao © *0 46 60 o 44 1 100°C 40 20 Sintering 30 40 tim e, 80- SO 1200°C DC microwave •hmohm> conventional "m ^ ^ m 42 «*•*#«*« 40 60 0 10 m in . 20 S in te rin g 30 microwave conventional 40 tim e, 50 60 min. " ^ 5k C SO 1300°C 1250°C microwave conventional o 10 20 S i n t e ri n g 30 tim e, 40 50 60 m in . 0 10 20 S in te rin g 30 tim e, 40 50 60 min. Figure 5.12. Comparison in densities of diphasic mullite xerogel sintered by conventional and microwave processing at various temperatures. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 150 Table 5.10. Kinetics values for activation energy calculation of diphasic mullite gel. microwave m conventional K (l/m in.) InK 1150°C (1423K) 7.027xl0-4 1.9385 -3.9433 1200°C (1473K) 6.789xl0-4 7.5927 1250°C (1523K) 6.566xl0-4 17.400 temperature K (l/m in.) InK -2.5780 I.5466 -4.1691 -1.7487 II.0 0 4 -2.2069 59.395 -0.5210 1300°C (1573K) 6.357x10-4 y = 29.635 - 4.7675X R A2 = 0.985 - 2- c • O -3 - microwave conventional -4 - y = 53.21 4 - 8.4486X RA2 = 0.999 6.2 6 .4 6.6 6.8 7 .0 7 .2 10000(1/T), 1/K° Figure 5.13. Arrhenius curve of diphasic mullite gels sintered by (a) microwave and (b) conventional methods. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 151 Similarly, for conventional sintering, InK = 53.214 - 8.449 T ' 1 and Ea = 8.449 x 8.314 x 104 = 702.4 (kJ/mole). It is seen that under the microwave processing conditions, the apparent activation energy for the densification of diphasic mullite gel is significantly (about 44% ) smaller than that under conventional sintering conditions. The smaller activation energy in the microwave processing indicates that the diphasic gel couples well with microwaves, which in turn implies that the dielectric loss of the diaphasic gel at high temperature is fairly high. This should be attributed to the special microstructure and the particle size o f the diphasic gel. In the diphasic gel, the alumina particles exist in 8-Al2 0 3 of low crystallinity, and the silica exists in amorphous state. Additionally, both phases are highly divided (high specific surface area, or very fine particle size). The combination of these factors makes the diphasic gel system different from the crystalline materials of the same components (crystalline alumina and silica), as well as the crystalline mullite, creating unique conditions that may favor the microwave coupling. 5.6. Summary on Microwave Processing of Mullite Materials Monophasic, diphasic, hybrid, and seeded aluminosilicate gels were synthesized by sol-gel methods. Xerogel was obtained by drying these materials in air. Aerogel was prepared by drying the diphasic gel under supercritical conditions. M ullite ceramics were sintered using the above materials. It was found that in the temperature range between 1200 to 1500°C, for the time period of an hour, diphasic gels sinters well. The sinterability of the crystalline mullite is poor, the monophasic gel crystallizes before temperature reaching this range thus behaves like the crystalline mullite, the sinterability of the seeded gels is in between the diphasic gel and the crystalline mullite. M icrowave processing accelerates the mullite transformation o f the diphasic gel, Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 152 lowering transformation temperature by about 75-100°C. Microwave processing also enhances densification of the diphasic aluminosilicate gels. The study on the densification kinetics of the diphasic gel shows that the apparent activation energy of this material under microwave processing conditions is 396 kJ/mole, while that in the conventional case is 702 kJ/mol. The significantly lower activation energy indicates that the diphasic aluminosilicate gels couple well with the microwaves. It also implies that the dielectric loss factor of the diphasic gels at high temperatures is fairly high. The observed high microwave coupling phenomenon of the diphasic gel is attributed to the unique characteristics of this gel, including the high free energy state of the two components, particle size, high porosity, high surface area, and high homogeneity, etc. The interaction of microwaves with mullite gels will be discussed further in Chapter 8 . Regular m ullite ceramics with relative density up to 95% were fabricated by microwave sintering using the diphasic gel as starting material. Step sintering works well for the diphasic gels in both microwave and conventional processes. Transparent mullite ceramics were fabricated at ambient pressure by microwave at 1300°C and conventional processes at 1320°C, respectively, using the diphasic aerogel. The transparency of the obtained samples is good, but might be further improved by using hybrid gel and/or optimizing sintering schedule. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 153 Chapter 6 MICROWAVE SINTERING OF OTHER CERAMICS Besides HAp and mullite, alum ina and [NZP] ceramics were also sintered by microwave processing. The precursors of these two materials were acquired directly from commercial sources. 6.1. Microwave Sintering of Alumina 6.1.1. Introduction Alumina is a very important material and exists in various polymorphs. a-A l 2 0 3 , corundum, a crystal form of alumina, is widely used in ceramic industry for numerous applications. The melting temperature of a-Al 2 0 3 is over 2000°C, so that the sintering of alumina requires very high temperatures. Microwave sintering of alumina has been carried out (Cheng et al., Patil et al.) by many investigators who used crystalline alumina powder as starting material. In this study, an amorphous alumina was used as starting material to investigate the sintering of A12 0 3 in a microwave field. 6.1.2. Experimental Procedure Two starting materials were used in this study: one is crystalline and the other is amorphous. The crystalline powder is a fine a-alum ina generally used in polishing (Buehler). The average particle size is 0.05 |0.m. The amorphous powder is a commercial material (Rhone-Poulenc, France). The powder was compacted into circular pellets of a half inch diameter. No sintering aids were used. Microwave sintering was carried out at 1500°C for 20 min. in the modified 900-W m icrowave furnace with the turntable. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 154 Conventional sintering was carried out in the high heating rate electric furnace. Heating rates in both cases were about 100°C/min. Green and sintered densities were measured. Microstructure of the sintered sample o f amorphous alumina was examined by SEM. 6.1.3. Results and Discussion Crystalline alum ina powder with the green density of 52%, when sintered by microwaves at 1500°C for 20 min., produced a relative density of 78%, compared with 76% achieved by conventional sintering. In this case, microwave enhancem ent in densification is very little. The density data for the amorphous starting material are listed in Table 6.1. Although the samples were not sintered to high densities, microwave effect is clearly seen. At 1500°C, the microwave sintered density is 26% higher when sintered for 20 min., and 31% higher when sintered for 60 min., than the conventionally sintered samples. SEM micrograph in Fig. 6.1. shows that the porosity of the sintered alum ina is consisted o f two types o f pores: intergranular and intragranular. The primary particles are only about 1 -2 jxm, while the agglomerates of tens or hundreds of micro are composed of numerous primary particles. Under the experimental conditions, sintering mainly occurred between the tiny primary particles. Further densification could be achieved by increasing temperature, time, or improving the morphology of starting material. The microwave effect on the densification of amorphous alumina was found to be one order of magnitude higher than that on crystalline alumina, although the starting material of the latter was very fine, and the green density o f the compacts was high. This again indicates that the amorphous alumina couples with microwaves more efficiently than the crystalline alumina. It is believed that the high free energy state of the amorphous alumina and the scattering effect to the microwaves made the difference. The related mechanisms will be discussed in Chapter 8 . Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 155 Table 6.1. Density data of RP-alumina before and after sintering (1500°C). microwave 20 min, conventional 60 min. 20 min. 60 min. Green, % 38.70 39.54 39.70 39.76 Sintered, % 59.97 64.34 47.60 48.97 microwave effect 20 min. 26% 60 min. 31.4% Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 6 . 1 . Microstructure of the alumina sintered by microwaves at 1500°C for 20 min.using from amorphous as starting material. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 157 6.2. Microwave Sintering of [NZP] Ceramics 6.2.1. Introduction C a 1.xSrxZr4 P 6 0 24 , (0<x<1.0), or CSZP, is an important member of a large structural family o f low thermal expansion materials known as [NZP] system after the parent composition of sodium zirconium phosphate, NaZr2 P 3 0 12 (Alamo and Roy, Roy et al., Agrawal and Stubican). The potential applications in superionic conductors, radwaste management and low thermal expansion ceramics, have made the study o f these materials very attractive. Synthesis and conventional sintering, as well as characterization of CSZP, have been carried out recently by Limaye et al. Microwave sintering o f these materials, however, has not yet been reported. It is the objective o f the present study to show the possibility of sintering CSZP ceramics by microwave processing, and the influence of microwave interactions on the resultant ceramics. 6.2.2. Experimental Procedure The precursor powder of C a 0 -5 Sr0 5 Zr4 P 6 0 24 was synthesized by a solution sol-gel method, as previously described in the literature (Limaye et al.). Using fine powders of -325 mesh, cylindrical pellets of 12.7 mm (0.5 in) in diameter were uniaxially pressed at various pressures up to 840 MPa. An appropriate amount of 2% aqueous solution of polyvinyl alcohol (PVA) was used as a binder during compaction. Microwave sintering was carried out in a 500 W, 2.45 GHz commercial microwave oven (GE, model J43) with suitable modifications as described in Sec. 3.2. Multipellet sintering was conducted. The pellets were placed on a Fiberfrax™ insulation layer and then shielded with a porous zirconia cylinder which was then surrounded by Fiberfrax insulating material. The specimens were directly heated from room temperature to the sintering temperatures for various time periods, then allowed to cool naturally by turning off the power. Temperature was controlled by manually pulsing the microwave power Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 158 input. For comparison, conventional sintering was also carried out at 1300°C for 30 min. and 24 h, respectively, in a programmable furnace. 6.2.3. Results and Discussion A typical heating curve o f the microwave processed CSZP ceramics is shown in Fig. 6.2. Thermal runaw ay was noticed in the current sintering after about 40 min. of microwave irradiation. XRD study (Fig. 6.3.) showed that the sintered CSZP ceramic was highly crystalline compared with the starting material. Figure 6.4. shows the sintered densities of the microwave processed samples as a function of compaction pressure or green density. The sintered density was obviously dependent on sintering temperature, time, and green density. As can be expected, with increasing green density, sintering tem perature, or sintering time, sintered density increased. Figure 6.5 provides a comparison o f sintered densities o f CSZP at 1300°C by microwave processing and the conventional method. As can be seen from the figure, at the same sintering temperature (1300°C), a 30-min. microwave processing provided the sintered densities of the CSZP ceramics 13-17% higher than the conventional sintering of 30 min., and 5-8% higher than those for 24 h . The fact that the 24 h conventionally sintered CSZP was still less than 90% dense indicates that the diffusion o f the CSZP material under the conventional processing conditions was low. The significantly higher sintered densities achieved after a short period of microwave processing indicates that the densification rate o f CSZP in the microwave field was much greater than in the conventional oven. Since boundary diffusion is the primary mechanism for densification of ceramics during sintering, the boundary diffusion of the CSZP material must have been enhanced during the microwave processing. The m icrostructures of the as-sintered surfaces o f both m icrowave and the conventionally sintered CSZP are shown in Fig. 6 .6 . The grain size of the CSZP ceramics sintered both by microwave and by conventional heating did not show much difference. Compared with the higher sintered density achieved by microwave processing within a short processing time, the rate o f grain growth in the microwave processed sample was Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 159 Temperature, 2000 1000 0 20 40 60 80 100 Time, min. Figure 6.2. A typical heating curve o f CSZP ceramic in a microwave field. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 160 AS R E C E IV E D CS2P2 10 20 30 40 50 TWo theta, degree Figure 6.3. XRD patterns of CSZP (a) before and (b) after microwave sintering at 1300°C for 30 min. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 161 100 >* "<7> c o ■o 73 1450°C, 45 min. 1450°C, 30 min. 1300°C, 30 min. 1450°C, 20 min. 0) »_ 0) c W 20 0 400 600 800 1000 Compaction pressure, MPa Figure 6.4. Influence of compaction pressure (green density), sintering temperature and time on the sintered density of CSZP. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 162 80 Sintered density, % 100 M/W 30 min. CN 24 h C/V 30 min. 200 400 600 800 1000 Compaction pressure, MPa Figure 6.5. Comparison of the sintered density of CSZP ceramics processed by microwave and conventional methods at 1300*C. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 163 A' ■ 4 . 0 7 KX Figure 6 .6 . 5U 0007 YF i 8 2 E 5 M icrographs o f CSZP sintered at 1300°C by (a) microwave method for 30 min. and (b) conventional method for 24 h. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 16 4 relatively suppressed. Higher sintered density could be achieved by increasing sintering temperatures and/or time. 6.2.4. Summary on Microwave Sintering of CSZP Microwave sintering o f CSZP ceramics was conducted at 2.45 GHz at 1300 and 1450°C, respectively. The results show that the sintered density of CSZP increases with increasing green density, sintering temperature and sintering time. Compared with the samples conventionally sintered at 1300°C for 30 min. and 24 h, a higher sintered density was achieved in microwave sintering for 30 min. This indicates that the densification rate of CSZP was substantially enhanced during the microwave processing. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 165 Chapter 7 MICROWAVE SINTERING OF VARIOUS CERAMIC COMPOSITES 7.1. Introduction A composite consists o f two or more physically distinct and mechanically separable phases. The importance o f a composite lies in the fact by modifying the amounts of components one can control and tailor the properties to desired values. To meet the increasing demands in high-performance applications, often a composite approach is adopted to improve the functional and mechanical properties. In a composite, one can not only tailor properties o f the components through compromise (e.g. brittle + tough —> moderately tough), but also “create” new properties. Two different brittle com ponents can lead to a tough com posite. For exam ple, mullite/SiC(whiskers) is such a brittle-brittle composite in which mullite is the matrix and silicon carbide whiskers work as a reinforcing or toughening component. M any composites have been developed using whiskers or particulate materials as reinforcing ag e n t. As composite m aterials become a trend in material development, processing of composites becomes more and more important. In the preparation o f ceramic composites, sintering is an important step. Different components in a composite may have different sinterability, which makes sintering of the composite difficult in conventional process. Since microwave processing is fundamentally different from conventional processing in heating mechanism, it may be unique, or an alternative to conventional method. In this study, various zirconia-containing ceramic composites based on HAp, mullite, alumina, CSZP, and BZPS, were sintered by microwave processing. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 166 7.2. Microwave Sintering of HAp-Based Composites 7.2.1. Introduction In the last two decades, m uch work has been done to develop a fundamental understanding of various aspects of HAp. HAp ceramics have now been used clinically for nonload-bearing applications. However, pure HAp ceramic is intrinsically brittle. The fracture toughness o f dense HAp ceramics is about 1 MPaVm which is only a quarter to a half that of bones. Because of this, HAp ceramics could not be used for any load-bearing implantation unless its toughness is substantially improved. Therefore, toughening o f HAp is an important subject. Unfortunately, research work carried out in this area is very limited. The reported cases include HAp dispersed with Z r 0 2 particles and a-S i 3 N4 whiskers. W hen preparing H A p/Z r0 2 composites, sintering has to be carried out at high temperatures which could lead to decomposition of HAp (Ioku et al., W u and Yeh). Wu and Yeh studied conventional sintering of H A p/Z r0 2 composites containing 20% by weight of Z r0 2 (about 11.5% by volume). They found that a 3-h sintering at 1100-1400°C led to relative densities o f about 50 to 60% in one case, and 75 to 80% in another. The decomposition o f HAp and formation of C aZ r0 3 were observed in their study. Fang et al. (1989) also tried to sinter the HAp/PSZ composite by conventional sintering alone, or conventional sintering follow ed by hot isostatic pressing (HIP). H ow ever, a 2-h conventional sintering at 1250°C led to relative densities of only 64-80%, while the post sintering HIP at 100 MPa improved the density only by 1-2%. Compared to the pure HAp which can be easily sintered to around or over 95% by conventional sintering, it is obvious that the addition o f Z r 0 2 to HAp has a deteriorating effect on the sinterability of the mixture. It is impossible to achieve a dense H A p/Z r0 2 composite by conventional sintering at the ambient pressure. By hot pressing at 1050°C and 30 M Pa for 1 h, then post-sintered at the same tem perature at 200 M Pa in Ar for 1 h, Yoshimura et al. (Yoshimura et al., 1989) could make a dense H A p/Zr0 2 composite without decomposition Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 167 of HAp. They also prepared H A p/a-Si 3 N 4 by hot pressing at 30 MPa and obtained a dense composite, but they observed a reaction between HAp and a-S i 3 N4. Through the dispersion of Z i0 2 particles or a -S i 3 N 4 whiskers in HAp, the fracture toughness, K IC, had been improved by these investigators to around 3 MPaVm. As described in Chapter 4, fabrication of regular, porous, and transparent HAp ceramics has been achieved. Microwave sintering has a heating mechanism totally different from that of the conventional method. The sintering efficiency in microwave processing is material dependent. Materials that have a high dielectric loss absorb microwave energy efficiently and are easy to heat in a microwave field, while those of low dielectric loss absorb microwave energy less efficiently and are hard to heat. Because of this, selective sintering/heating is possible when processing a multi-phase system. The success in microwave sintering of HAp indicates that the microwave absorption of HAp is good enough at least at elevated temperatures, zirconia also absorbs microwaves very efficiently One can expect good sinterability o f HAp/PSZ composites by microwave processing. Silicon carbide is also a good m icrowave absorber, therefore one can expect good sinterability on HAp/SiC composites in a microwave field. In the current study, microwave sintering of HAp-based composites, i.e., HAp/PSZ and HAp/SiC, was attempted. 7.2.2. Experimental Procedure 22.214.171.124. Starting Materials Two HAp powders were used in the current study. HAp-A was synthesized by hydrotherm al m ethod, and H A p-B w as prepared by hydrolysis o f brushite (CaHP0 4 -2H 2 0 ) followed by ripening treatment (Section 4.2.). For the HAp/SiC composite, the volume proportion of HAp-A:SiC was 90:10. The SiC whiskers were o f an average thickness 0.5 pm with varying lengths up to 25 pm. The whiskers were added to the precipitate slurry of HAp-A by magnetic stirring. The mixture Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 168 of precipitate and whiskers was then subjected to the same hydrothermal, drying, and disagglomeration processes as HAp-A. For the HAp/PSZ composite, the tetragonal phase zirconia stabilized by 3 mol% yttria [Zirconia Sales (America), Inc., Atlanta, GA] of sub-micrometer particle size was used. The volume proportion of HAp-B:PSZ was 90:10. The two powders were dry mixed in a plastic jar mill using zirconia balls as mixing media. The mixture was then sieved to -325 mesh to further improve homogeneity. 1.2.22. Compaction Pellets (12.7 mm diameter x 5 mm thick) of HAp, HAp/PSZ and HAp/SiC powders w ere com pacted uniaxially for both m icrowave and conventional sintering. The compaction pressure for all pellets was 350 MPa, except for the case of HAp/PSZ in which 420 M Pa was used. 1.2.23. Sintering A 500 W, 2.45 GHz commercial microwave oven (GE, Model JE 43) was used for sintering. Only one pellet was used in each run. The arrangement inside the microwave oven is schematically shown in Fig.7.1. The pellet was placed on a board of Fibermax insulation (Carborundum Co., Niagara Falls, NY) at the center o f the cavity o f the microwave oven, then vertically surrounded by two coaxial porous zirconia insulating cylinders o f 30 mm in height, and then further insulated with the Fibermax boards. One end-open Pt-PtlORh thermocouple, sheathed with alumina, was inserted from the bottom of the microwave oven, with the tip touching the sample, to monitor the temperature during sintering. The processing temperature was controlled by pulsing the microwave power. Heating started from the room temperature. After sintering, the system was allowed to cool to room temperature. For comparison, conventional sintering of some pellets with the same green densities was carried out at 1200°C in a 5 kW programmable box furnace, with both heating and Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 169 Mlcrowava ow n / / / / Zirconia cylinder / / / / / / / / / / /V /V ,V ,V /V /V . \ \ \ Kv a VA' \ N \ V xV \ \ \ \ \ \ \ \ \ v a v a v * / \ \ \ \ \ \ \ \ \ S \ \ \ \ \ \ \ / / / S * / 4 / / / / / , \ S\ \ \ A WQ (Sample 1 | | S S S 5 / / / / / ./ ./ /. /. vW W W V o / V/ \ / \/ < »//// / / / / \/ <\ \/ 4 > x/V/ v/ W / /s /W/ V / ■ / / / / / ✓ / / / ✓ / Rofractory wools / . Thermocouple Figure 7.1. The arrangement inside the microwave oven. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 170 cooling rates at 5°C/min. A HAp/SiC sample was uniaxially hot pressed in vacuum at 1200°C and 36.6 M Pa for 15 Min. in a graphite m old of 12.7 mm diameter. 126.96.36.199. Characterization The sintered samples were characterized to determine density, phase composition, microstructure, and fracture toughness. Density was determined by weight and dimension measurements. Phase composition was analyzed by powder X-ray diffraction (XRD) using a Scintag diffractom eter (Scintag Inc., Sunnyvale, CA). M icrostructures were exam ined with a scanning electron microscope (SEM, ISI-DS 130, Akashi Beam Technology Corp., Japan). Fracture toughness was measured by diamond indentation using a Vickers indenter as described in Sec. 3.4.10. 7.2.3. Results and Discussion 188.8.131.52. Heating A typical heating curve for microwave processing in the current study is shown in Fig. 7.2. It usually took 20 to 50 min. to trigger thermal runaway. Because of the high m icrowave absorption and relatively large mass o f the Z r0 2 cylinders over that of the sample, the therm al runaway behavior here had much to do with the porous zirconia cylinders used to vertically shield the sample. 1.23.2. Phase Composition By XRD analysis, it was found that both the microwave and conventionally sintered HAp/PSZ composites were composed of HAp and tetragonal zirconia with trace amounts of a- and /J-Ca3 (P0 4 )2 . No C aZ r0 3 was detected. Trace a- and /?-Ca3 (P 0 4 ) 2 were also detected in the pure HAp pellet after 1 h conventional sintering at 1200°C. This indicates Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 171 1400 1200 1000 *« 800 600 a. 400 200 0 10 20 30 40 50 Time, Min. Figure 7.2. A typical heating curve for microwave processing o f HAp-based composites, where porous zirconia was used as accelerator. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 72 that a partial decomposition o f HAp took place during sintering, but the decomposition was due to the thermal stability o f the starting HAp rather than the microwave processing or any reaction between HAp and Z r0 2. In the case of the HAp/SiC system, the conventional sintering at 1200°C led to a slight oxidation on the surface o f the pellet, while the microwave processing caused very extensive oxidation. A thick layer o f oxidation product was observed on the surface o f the HAp/SiC pellets after microwave sintering even without additional holding. A few minutes of microwave processing at 1200°C would lead to complete oxidation of the SiC in the HAp/SiC mixture. Microwave sintering o f HAp/SiC in Ar gas was also attempted, but it did not help. It is believed that the lack o f moisture in the sintering atmosphere caused HAp to decompose at lower temperatures, while the OH- ion decomposition of HAp made the oxidation of SiC even worse. A complex mixture o f S i0 2 phases was detected in the oxidation product. The hot-pressed HAp/SiC sample was composed of HAp and a large amount of a-C a 3 (P 04)2. The decomposition o f HAp in this case was very serious because the sample was hot-pressed in vacuum. The hot-pressed HAp/SiC composite turned black due to the diffusion of graphite from the mold. 184.108.40.206. Sintered Density For HAp/PSZ, conventional sintering at 1200°C led to a sintered density o f only 75% of the theoretical (Table 7.1), and prolonged sintering tim e from ~30 min. to 4 h did not bring about any significant increase in sintered density. Under microwave processing conditions, however, a 20 min. sintering at 1200°C led to a sintered density o f 92.7% in the pellet o f 59.8% green density, which was 24% m ore dense with respect to the conventionally sintered sample. This substantial enhancement in density is attributed to selective heating in the microwave processing. Zirconia absorbs microwave energy much more effectively than HAp, so that it could promote densification o f a HAp/PSZ composite in microwave field when the amount of Z r 0 2 in the composite is appropriate. This is just the reverse of the case in conventional sintering where the existence of Z r0 2 makes the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 173 Table 7.1. Summary o f the processing conditions and some results m easured on the sintered HAp and HAp-based composites. No. Processing Composition p*g, % p*s, % K ic 1 HAp m/w 1200°C x 10 min. 52.82 93.3 1.56 2 HAp/10%PSZ m/w 1200°C x 20 min. 54.33 87.0 1.18 3 HAp/10%PSZ m/w 1100°C x 30 min. 54.02 89.1 1 .2 0 4 HAp/10%PSZ m/w 1200°C x 20 min. 59.79 92.7 3.88 5 HAp/10%PSZ c/v 1200°C x 24 min. 60.42 74.81 — 6 HAp/10%PSZ c/v 1200°C x 60 min. 59.86 75.08 — 7 HAp/10%PSZ c/v 1200°Cxl20 min. 60.42 75.43 — 8 HAp/10%PSZ c/v 1200°Cx240 min. 60.06 75.2 — 9 HAp/10%SiC HP 1200°Cx 15 min. — 72.5 2.85 10 HAp/10%SiC c/v 1200°Cxl20 min. 60.0 66.83 — 11 HAp/10%SiC m/w 1200°C no holding 58.5 — — *Pg, p s — green density and sintered density, respectively. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 174 sinterability o f HAp/PSZ very poor. This demonstrates the advantage o f microwave processing for such a special system. It should be mentioned that the low er sintered densities of other two microwave sintered HAp/PSZ samples were due to their lower green densities. In the case of HAp/SiC, the conventional sintering at 1200°C for 2 h led to a density of 6 6 . 8 %, and hot pressing, 72.5%. The density of the microwave sintered sample was not measured due to the serious oxidation on the surface after processing, but from SEM micrographs, its density looks comparable to that o f the hot pressed one. The low sintered density o f HA p/SiC was due to the m orphology o f the silicon carbide whiskers. Obviously, it is more difficult to achieve a high sintered density in a whisker dispersed system than in the particle dispersed one. 220.127.116.11. Fracture Toughness K IC of 3.88 M Pa'®- has been achieved in the HAp/10%PSZ composite o f 92.7% density sintered by microwave processing. Compared with 1.5 MPaVm" from the pure HAp o f 93.3% density, a 150% net increase in K IC was realized. The significant im provem ent in KIC was attributed to the transformation toughening m echanism of tetragonal Z r0 2 dispersion. The toughening effect was not observed in the two other HAp/PSZ samples with density less than 90%. This seems to show that the transformation toughening could be realized only when density is over a certain value, most likely greater than 90%. Apparently, fracture toughness, K IC, is a function o f density. It increases with increasing density, but the relation is not linear. Since the densities of the HAp/PSZ samples sintered by the conventional method were well below 90% o f the theoretical, no toughening was expected in these samples, and thus KIC was not measured for them. In the case of the SiC-containing composite, although the density of the hot-pressed sample was only 72.5%, an increase in fracture toughness was still observed. K IC of 2.85 MPaVm was achieved in this sample, compared with the zirconia dispersed sample, where no toughening effect was achieved below 90% density. This is not surprising since the mechanism o f whisker toughening is different from that of transformation toughening. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 175 Another factor which should be considered is that the K IC o f this HAp/SiC sample was measured on the plane perpendicular to the compaction direction and thus more whiskers had contributed to the toughening because o f the preferred orientation o f the whiskers due to the uniaxial compaction. 18.104.22.168. Microstructure Figure 7.3. shows the m icrostructures o f the HAp/PSZ composite sintered by m icrow ave at 1200°C. It can be seen that the distribution o f PSZ particles is homogeneous, and the microstructure is quite dense. The sintering o f some PSZ particles is also evident. The microstructures of HAp/SiC composites, sintered by microwave as well as by hot pressing are shown in Fig. 7.4. Obviously, both samples show very porous microstructure. The bonding between SiC whiskers and HAp matrix, however, looks good. If the oxidation o f SiC could be avoided, yet high densification could be achieved, then improved toughening could be expected. In both the cases of PSZ and SiC dispersed HAp ceramics, no microcracks between the dispersed phase and the matrix were observed, indicating that there was no significant mismatch in the thermal expansion coefficients between the toughening agents and the HAp matrix. 7.2.4. Summary Com posites o f hydroxyapatite/partially stabilized zirconia (HAp/PSZ) and hydroxyapatite/silicon carbide whiskers (HAp/SiC) were sintered at 1100-1200°C by microwave at 2.45 GHz. Characterization o f the sintered composites was carried out by density, microstructure, phase composition, and fracture toughness measurements. The results show that although not yet fully densified, a much higher sintered density in the HAp/PSZ composite was achieved by microwave sintering than by conventional sintering at the same temperature. A relative density o f 93% was achieved by 20 min. microwave processing at 1200°C. Comparatively, 2 h conventional sintering of the same material at 1200°C led to only 75.5% relative density. KIC o f this microwave sintered HAp/PSZ of Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 176 Figure 7.3. The microstructure o f microwave sintered HAp/PSZ composite, showing the simultaneous sintering o f HAp and PSZ particles. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 177 Figure 7.4. The m icrostructures o f the fracture surfaces o f the HAp/SiCw composites sintered by (a) microwave processing and (b) hot pressing (perpendicular to the compaction direction). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 178 93% density was found to be 3.88 MPaVm, which is 2.5 times of the value for pure HAp o f the same density. A further increase in KI(-< could be expected if full or nearly full densification was achieved. Sintering of PSZ particles in the HAp/PSZ composite was also observed in the microwave processed sample. Microwave sintering of HAp/SiC was not successful in the current study due to the oxidation of SiC in air at high temperature. 7.3. Microwave Sintering o f Various Zirconia-Containing Composites 7.3.1. Introduction Ceramics are usually brittle and transformation toughening is adopted to improve fracture toughness o f various ceramics. Partially stabilized tetragonal zirconia is a toughening agent popularly used. Since any toughening effect in such a composite through transformation o f zirconia directly depends on the density of the composite, a high density is critically important to achieve the desired transformation toughening. The addition of zirconia, however, may reduce sinterability of the ceramic and thus high densities become difficult to achieve under conventional sintering conditions. For example, about 95% relative density is easy to achieve in conventional sintering o f HAp, once zirconia is added to HAp, the sinterability o f the composite becomes very poor. In conventional sintering of HAp/Zi0 2 >earlier researchers achieved a relative density of only about 50-78% at 11001400°C for 3 hours (Wu and Yeh, 1988). W hile by microwave processing, a relative density of 93% in HAp/10%ZrO 2 composites (Sec. 7.2.) at 1200°C for only 30 min. Since microwaves have the selective heating effect to the materials of different dielectric properties, microwave processing may be advantageous for sintering some ceramic composites. Stabilized tetragonal zirconia is a transformation toughening agent and also an good microwave absorber. It can be preferentially heated in a microwave field. The objective o f this study was to see how well microwave processing works with zirconia-containing composites. For this purpose, microwave sintering of ceramic composites o f zirconia and four Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 179 matrices, including alumina, mullite, calcium strontium zirconium phosphate (CSZP), Cao.5Sr0-5Zr4P6024, and barium zirconium phosphosilicate (BZPS or B S25), Baj 25 Zr4 P 5 5 Si0.5 Q 24 , was carried out in the current study. Among them, alumina and mullite are important ceramic materials commonly used, CSZP and BS25 belong to a very large new family known as [NZP] and characterized for its low thermal expansion behavior (Limaye et al., 1991; Roy et al., 1989 ). Conventional sintering of the duplicate samples was also carried out in a high capacity, low therm al inertia furnace at the heating rates comparable to those in the m icrowave processing. In this way, any differences in the sintering behavior would be attributable to the effect of microwave irradiation. 7.3.2. Experimental Procedure 7.3.2.I. Starting Materials The starting m aterial o f alum ina was an amorphous alum ina pow der from a commercial source (Rhone-Poulenc, France). The powder was sieved to -325 mesh (45 pm opening) before compaction. For comparison, a crystalline alumina powder (Linda B, high purity a-alum ina) of 0.05 pm average particle size was also used for composites fabrication. The mullite precursor was prepared by a diphasic xerogel method (Sec. 5.2.) by mixing sols o f boehmite (AlOOH) and silica (Ludox) at a m olar ratio of 3:2. A monophasic gel powder was also used for mullite/PSZ composites for comparison. The CSZP powder was also prepared by a sol-gel method (Limaye et al., 1991). The BS25 pow der was synthesized by a solid-state reaction method using dry oxide precursors (Huang et al., 1994). The zirconia powder (HSY-3.0, Zirconia Sales, Inc., Atlanta, GA) used in this study was tetragonal phase partially stabilized by Y 2 Q 3 (5.4 wt.%), with a minor monoclinic phase. The average particle size of this zirconia powder was about 0.1 pm . Each matrix powder was mixed with zirconia in various proportions. An appropriate Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 80 amount o f 2% PVA solution was used as a binder during mixing. The m ixture was hom ogenized by hand m ixing with an agate mortar. The alum ina/zirconia and BS25/zirconia mixtures were compacted into pellets of 0.5 inch diameter, and the mixtures o f mullite/zirconia and CSZP were compacted into pellets o f 0.25 inch diameter. The compaction was uniaxially carried out at 350 MPa. The relative green densities of the ascompacted pellets of various systems were in the range of 39-53% for alumina/zirconia, 46-51% for mullite/zirconia, 53-58% for CSZP/zirconia, and 60-67% for BS25/zirconia, respectively. No sintering additive was used in this study. 13.2.2. Sintering M icrowave sintering was carried out in a 900 W , 2.45 GHz m icrowave oven (Panasonic) with a turntable. The sintering packet has been described and schematically illustrated in Section 3.2. M ulti-pellet run was adopted in the sintering. The pellets, usually in two (0.25 inch pellets) to three (0.5 inch pellets) layers, were placed in the center o f the packet. Each layer containing 6-7 pellets of the 0.25 inch samples, but only one pellet o f the 0.5 inch samples. Conventional sintering was carried out in the high capacity, low thermal inertia, high temperature furnace (see Section 3.2.). The sintering of the composites based on alumina, mullite, and CSZP, was conducted at 1500°C. Composites of BZPS (BS25) were sintered at 1550°C. All samples were soaked at the peak temperature for 20 min. The sintering of the diphasic mullite gel is a reaction sintering process. The earlier results (Sec. 5.3.) showed that the diphasic mullite gel used in this study starts crystallizing (mullitization) at about 1320°C. Since the densification o f gels through viscous flow is far more efficient than that o f the related crystalline materials through ordinary solid diffusion, step sintering was adopted for the composites based on the mullite gel, in which the samples were first heated to 1200°C, then soaked 30 min. for densification by viscous flow, and finally heated to 1500°C for mullitization. Sintering of the composites based on crystalline mullite (Chichibu Cement Co.) was 20 min. at 1500°C. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 181 13 .23 . Characterization Sintered density was m easured by weighing and dimensional measurement of the pellets. Powder x-ray diffraction (CuAT^) was carried out with a Scintag diffractometer on the starting powders and on some selected sintered samples for phase identification. The microstructure of the sintered specimens was examined by SEM. 7.3.3. Results and Discussion The heating curves are plotted in Fig. 7.5. It is seen from the figures that while the soaking times in both microwave and conventional processes are identical, heating time to the desired sintering temperatures for microwave processing was never longer than in conventional process. Actually, conventional heating took a few minutes longer in the cases of alumina, CSZP and BZPS (BS25) composites. Sintered densities as a function of zirconia content are shown in Figures 7.6-8. From Fig. 7.6., it is clearly seen that the microwave enhancing effect on the composites based on the amorphous alumina (a) is much stronger than that based on the crystalline alumina (b). W ith the amorphous alumina as the matrix, microwave effect is stronger when zirconia content is lower, and weaker as zirconia content increases. W hen crystalline alumina (Linda B) was used as the matrix, microwave effect is not as obvious, and not dependent on the zirconia content (b). This clearly indicates that the amorphous alumina precursor couples better with microwaves than the crystalline powder. Figure 7.7. shows sintered density o f mullite/zirconia composites. When using the diphasic mullite gel as the matrix, no significant microwave effect was observed on the samples step-sintered at 1200°C for 30 min. followed by 1500°C for 20 min. (Fig. 7.7.a). Keeping the low er tem perature sintering conditions unchanged but reducing higher temperature (1500°C) sintering time from 30 to 4 min. did not show much microwave effect either (Fig. 7.7.b). W hile when sintered only at 1200°C for 30 min., the microwave effect was very clear (Fig. 1.1.c). Since at 1200°C, the diphasic gel has not transformed to Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 82 1600 • 0 8o0© *00 0000003 1400 - O o 1200- O l_ 1000 3 (0 0) O. V. E a l- cP - ◦ 800 - e 600 400 - 200 alumina/zirconia • microwave O conventional • 0 20 10 T -r T 30 40 50 60 Time, min. 1600 O o 0) L- 3 •• 1400 - 1200- •0*0 * 0 * 0 oOO •O 1000 +* 800- k 0) CL 600- • o 400 K> (0 E 0) H CTOOOOO • O mullite/zirconia 200 H O i 0 i__ microwave conventional —r20 40 60 80 Time, min. Figure 7.5. H eating tem perature as a function of tim e for the zirconia-containing composites based on (a) alumina, (b) mullite. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 183 1600 ^ ••O * 0 * 0 * 0 * 0 0 0 0 0 O • 1200 o o • O Temperature, o 1400 :°° o c o o o * o •• CSZP/zirconia • microwave O conventional • 10 20 30 40 50 Time, min. 1600 • o*o*oc*ooo Temperature, °C 1400 • 1200 o_o p • o o o & o o o 1000 800 600400 •IT BS25/zirconia • microwave O conventional O* 0 200 H ■>. 0 10 T T 20 30 40 50 Time, min. Figure 7.5. (Continued) Heating tem perature as a function o f time for the zirconiacontaining composites based on (c) CSZP, and (d) BZPS. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 184 100 ■ 90 - □ m/w 1500°C x 20 min c/v 1500°Cx20min 80 7 0 -, 60 50 - 20 0 40 60 80 100 Zirconia Content, vol.% 100 • O m/w 1500°C x 20 min c/v 1500°Cx20m in 90 - o <0 o c O a o > O 80 0) cc ooo 70 l b T T T T 20 40 60 80 100 Zirconia Content, vol.% Figure 7.6. Sintered densities of AI2 O3/PSZ composites at various mixing proportions. The alumina powder is (a) amorphous and (b) highly crystalline, high purity alumina. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 185 100 95 - 90 c/v 1200°Cx30' + 1500°Cx30' m/w 1200°Cx30' +1500°Cx30' 85 20 0 40 60 80 100 Zirconia Content, vol.% 100 -I sS ~ w c 0) 95 H ■o 0> > 90- M ■ O O 0) m/w 30 min. m/w 4 min. DC 85 £ —T“ —T“ —r - —r~ 20 40 60 80 —i 100 Zirconia content, vol.% Figure 7.7. Sintered densities of mullite/PSZ composites. Diphasic xerogel (a-c) and monophasic xerogel (d) was used as mullite precursor, respectively. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 186 100 -1 c 90 - >• « c <u ■o Q) > re "aj DC • O m/w 1200Cx30 min C/v 1200 0x30 min • 80- • • 70 bO ' o o o O O O o 60 50 I ■> 20 i r~ 40 1 60 i— 80 100 Zirconia content, vol% 100 -i 90 80 - • 60 O m/w 1500°C x 20 min c/v 1500°C x 20 min 50 0 10 20 30 40 50 60 Zirconia Content, vol.% Figure 7.7. (continued) Sintered densities o f mullite/PSZ composites using (c) dipasic mullite gel and (d) monophasic mullite gel as the matrix, respectively. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 187 mullite, while at 1500°C, the mullite transformation is over. It is clear that the diphasic mullite gel strongly absorbs microwaves before mullitization, while the crystalline mullite does not in the temperature range up to 1500°C. This is evidenced in the result of the composites using monophasic gel as the matrix (Fig.7.7.-d), in which the microwave effect on densification is negligible. Since monophasic gel crystallizes to mullite at 970°C, at higher temperature it behaves just as crystalline mullite. The sintered densities of the composites based on [NZP] materials (CSZP and BZPS) were plotted in Fig. 7.8. Both systems show the same trend of density dependence on the zirconia content. Generally, as zirconia content increases, the relative density increases, then decrease after reaching the maximum. The decrease stops at about 80% PSZ content, then increases again to a higher steeply to the maximum density o f pure PSZ. The microwave enhancing effect on densification in both cases is obvious, but that in the BZPS-based system is more significant when PSZ content is low. The peak density on the density curve indicates the optimum mixing ratio of the two components in the systems, while the “valley” density indicates inefficient sintering at the related proportion range of the two components. The density o f pure PSZ is much higher than that o f the [NZP] materials, indicating that PSZ has better sinterability. Addition of PSZ to the two [NZP] materials promotes sinterability, while the addition of the [NZP] materials to PSZ matrix deteriorates sinterability, but the relationship is not linear. In all the materials studied, density increases as increasing zirconia content in both microwave and conventional sintering, with the mullite system before crystallization as an exception. Higher densities have been achieved in microwave sintering in all cases. Significantly higher sintered density was produced in the microwave sintered specim ens of alum ina/zirconia system. The difference betw een m icrow ave and conventional sintering in the low zirconia content side is more significant. As zirconia content increased, the difference diminished, indicating that the microwave enhancement in sintering o f alum ina is more effective than o f zirconia, which is due to the better sinterability of zirconia over alumina. The relatively low density was due to the coarseness of the starting powder, low green density, low processing temperature, and short processing time. Furthermore, no sintering additive was used. Microwave Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 188 100 • m/w 1500°C x 20 min c/v 1500°Cx20m in O '« c 0> Q 0> 90 - A > JO 0> DC 20 0 40 60 80 100 Zirconia Content, vol.% 100 • < • O O o o • • <n c 0) o d> > 90- • o >• 0 • O O • 9 o 80 B 0° • <D O DC 70 m/w 1500°C x 20 min c/v 1500°C x 20 min I T i 20 40 60 80 100 Zirconia Content, vol.% Figure 7.8. Sintered density o f [NZP]/PSZ composites using (a) CSZP and (b) BZPS as the [NZP] component, respectively. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 189 frequency might also partly responsible for the relatively low density. K im rey and cow orkers (1991) have shown that alum ina/zirconia system couples better with microwaves at 28 GHz than at 2.45 GHz. In the mullite/zirconia system, no significant difference in sintered density was observed after sintering at 1500°C. A comparison of the density after the first-stage sintering (1200°C x 30 min.), however, clearly showed the difference (Fig. 7.7.). The densities after the 30 min. microwave heating at 1200°C were much higher than those conventionally heated samples. Obviously, microwave irradiation is more efficient in the densification of the diphasic mullite gels before mullitization. This is in good agreement with the earlier observation o f Roy et al. (1984, 1985) that the alumina and silica gels absorb microwave energy and get heated very quickly. Once the crystalline mullite is formed, the sintering mechanism changes and there is no more viscous flow. Further densification can only rely on the boundary diffusion and related factors as in ordinary solid state sintering. Since the solid state diffusion of ions in m ullite is very low, conventional sintering of m ullite is usually carried out at 1600-1700°C (Huling and Messing, 1993) for many hours. Due to the limit in processing temperature and time, and more importantly, the transparency of the crystalline mullite to microwaves, the microwave sintering of the m ullite system after m ullitization, under the current experimental conditions, did not show much advantage over conventional method. In the B S25/zirconia system containing up to 30% zirconia, the microwave enhancement was also obvious. The increase in density o f the CSZP/zirconia system, although obvious, was not as significant. This is probably due to the high processing temperature. When processing temperature is sufficiently high, the solid state diffusion of ions under the conventional conditions might also have been substantially accelerated to allow the microwaves to lose their effectiveness. The results of the previous study (Sec. 6.2.) showed that pure CSZP could be sintered to a relative density of 94% by microwave processing at 1300°C for 30 min. which was about 24% higher than that achieved in conventional sintering under the same temperature and same duration. This may explain why the CSZP/zirconia system sintered at 1500°C did not show as obvious microwave enhancement. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 190 Table 7.2. summarizes the enhancement of sintered density of the composites through microwave sintering. The values are obtained from the difference of sintered densities (microwave sintered density minus conventionally sintered density) divided by the conventionally sintered density, and they can be considered as the microwave effect in densification. It can be seen that with increasing zirconia content up to 40 vol.%, microwave enhancement in densification decreases in alumina and mullite systems, slighdy increases in the CSZP system, but the trend is not clear in the BS25 system. Actually, an optimum zirconia content may exist in certain systems for the best densification. The microwave enhancement in the composites o f alumina, mullite, CSZP, and BS25 are 2046%, 24-34%, 1-2%, and 4-13%, respectively. The ceramics without addition o f zirconia showed even higher enhancement, with CSZP as an exception. In both microwave and conventional sintering, there is a general trend of increasing sintered density with zirconia content, in all the systems studied, with the mullite/zirconia before mullitization as an exception. This is due to the fact that all the matrix materials used in this study are hard to sinter, while the sinterability of zirconia is significantly better, so that the addition of zirconia improves the sinterability of the resultant composites. This was also observed by other investigators under both m icrow ave and conventional conditions (Kimrey et al., 1991; M oya and Osendi, 1984). In the mullite/zirconia system, the sintering before mullitization is caused by the viscous deformation o f silica gel, which starts at temperature as low as 700°C (Wei and Halloran, 1988). The addition of zirconia powder separates the gel and thus reduces the viscous deform ation, so that before mullitization, the density of the composites decreases as increasing zirconia content. The trend is reversed after mullitization since at higher temperatures, the sintering mechanism changes to solid state diffusion, so that the presence of the easier-to-sinter zirconia will improve the sinterability of the crystalline mullite/zirconia composites. Powder XRD study showed that the sintered alumina was pure a-A l 2 0 3 phase when there was no zirconia addition, and was a mixture of the alumina (matrix) and tetragonal zirconia in the composites, with monoclinic zirconia in trace amount. Same is true for the mullite and BS25 systems. In the CSZP system, zirconia had completely transformed to Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 191 Table 7.2. M icrowave enhancement in densities (%) over the conventionally sintered samples. N o. Z r 0 2, vol.% A lu m in a M ullite (gel) CSZP 1500°C/20 min BS25 1500°C/20 min 1200°C/30 min 48.69 40.29 1 .0 0 13.36 1550°C/20 min 1 0 2 2 34.46 1 .1 2 7.68 3 5 32.45 0.64 11.16 4 10 46.40 30.52 1.97 13.40 5 20 38.55 26.85 1 .0 0 9.91 6 30 32.41 25.61 1.23 3.86 7 40 19.82 23.75 2.04 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 192 monoclinic phase. Further study is needed to explain this phenomenon. The pure mullite gel remained amorphous after sintering at 1200°C for 30 min., while after the processing at 1500°C, the m aterial was found to be highly crystalline single phase m ullite. No decomposition was found in any of these systems. SEM observations (Figs. 7.9.-12.) indicated that the m icrostructures of the composites were fairly homogeneous. Since the powders were m ixed in dry condition, agglomerates of zirconia could be seen. The homogeneity could be substantially improved by mixing in the wet state, especially during the synthesis of the matrix materials. As expected, the microwave sintered samples showed denser m icrostructure in general because of the increased density. Because duplicate green samples were used, the heating rates in the microwave and conventional sintering were comparable, temperature measurement was the same, and sintering time and temperatures were identical, the enhanced densification in the microwave processed samples could be certainly attributed to the microwave effect. Although some of the composites were not very dense, we believe that it is only a problem of optimization of processing parameters. 7.3.4. Summary on Microwave Sintering of Zirconia-Containing Composites Microwave sintering studies of various ceramic composites containing zirconia were carried out along with parallel conventional sintering studies. The results show that microwave processing significantly enhanced the densification in most cases, but the enhancement decreased with increasing zirconia content. This is due to the relatively higher sinterability o f zirconia over the matrix materials. The maximum enhancement was up to 46%, varying with the system and zirconia content. The influence of zirconia addition on the densification of the composites is determined by the sinterability difference of the components. The addition of zirconia to a harder-to-sinter matrix enhances densification, under both microwave and conventional conditions. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 193 Figure 7.9. M icrostructure o f A12 O j/PSZ (30%) com posite sintered by m icrowave processing at 1500'C for 20 min. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 194 Figure 7.10. Microstructure of mullite/PSZ (30%) composite sintered by microwave processing at 1200*C for 30 min. followed by 1500*C for 20 min. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 195 Figure 7.11. As-sintered surface o f CSZP/PSZ (10%) composite sintered at 1500°C for 20 min. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 196 Figure 7.12. M icrostructure o f BZPS/PSZ (20%) composite sintered by microwave processing at 1500°C for 20 min. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 197 Chapter 8 DISCUSSION ON MICROWAVE ENHANCING EFFECTS 8.1. Observed Phenomena of Microwave Enhancing Effects It has been observed in the experiments described in Chapters 4 to 7 that microwave processing generally enhanced densification of the ceramic materials selected for the current studies, including HAp, diphasic mullite gel, alumina, [NZP] materials, and zirconiacontaining composites. By microwave enhancing effect, or microwave enhancement, it is meant that a process is accelerated by microwave heating under the same heating schedule as in conventional processing, or lower temperature is required to achieve the same results. The process is thermal, including sintering, densification, and cheminal reaction, etc. The observed microwave enhancing phenomena in the current research are basically in two categories: microwave enhanced densification and microwave accelerated mullite transformation. The densification includes sintering of crystalline, semicrystalline, and amorphous ceramic materials. There are two phenomena in this category. First, for the same composition, the microwave enhancement was more significant for the amorphous and semicrystalline forms than for the crystalline m aterial. Second, the microwave enhancement effect diminished at higher processing temperatures. The examples for the first phenomenon are the sintering of mullite and alumina. Under the current experimental conditions, at 1500°C, the microwave enhancement in densification of amorphous alumina was found to be one order of magnitude higher than that of the crystalline alumina. Similarly, the microwave enhancing effect on densification of crystalline mullite was negligible at 1500°C, while 40% higher density was achieved with diphasic mullite gel by sintering at 1200°C for 20 minutes. The apparent activation energies for densification of HAp and diphasic mullite gel in the microwave heating were found to be substantially lower than that in conventional processing. A typical example for the second phenomenon in the first category is the sintering of CSZP, in which the microwave enhancing effect was Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 198 very significant at 1300°C, but negligible at 1500°C. The microwave enhancement of the mullite transformation of the diphasic mullite gel was manifested as the decrease in mullitization temperature and the promotion of seeding effect. In order to prom ote our understanding o f these m icrowave effects, we need to discuss the driving force for densification in general solid-state sintering and the mechanisms of the interactions between microwaves and the specific materials. 8.2. Driving Force for Densification in Solid-State Sintering Solid-state diffusion was the fundam ental m echanism responsible for the densification of all the materials in the current study except the diphasic mullite gel and the composites related to it. It is well known that in solid-state sintering the mechanisms related to material transfer include evaporation and condensation, viscous flow, surface diffusion, grain-boundary or lattice diffusion, and plastic deformation (Kingery et al., p. 470). During the initial stages of sintering, the difference in free energy or chemical potential between the neck area and the surface of the particle provides a driving force which causes the transfer of material by the fastest means available. If the vapor pressure is low, material transfer may occur more readily by solid-state processes, several of which can be imagined. As shown in Figure 8.1. and Table 8.1, in addition to vapor transport (process 3), matter can move from the particle surface, from the particle bulk, or from the grain boundary between particles by surface, lattice, or grain-boundary diffusion. Which one or more of these processes actually contributes significantly to the sintering process in a particular system depends on the relative magnitude of their rate, since each is a parallel method of lowering the free energy of the system. However, there is a most significant difference between these paths for matter transport: the transfer of material from the surface to the neck by surface or lattice diffusion, like vapor transport, does not lead to any decrease in distance between particle centers. In other words, these processes do not result in shrinkage o f the compact and a decrease in porosity. Only transfer of matter from the particle volume or from the grain boundary between particles causes shrinkage and pore Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 199 Grain boundary Figure 8.1. Alternate paths for matter transport in the initial stages of sintering. (Ashby) Table 8.1. Alternate Paths for Matter Transport During the Initial Stages o f Sintering. Mechanism (Fig. 8.1.) Transport Path Source of Matter Sink of Matter 1 Surface diffusion Surface Neck 2 3 Lattice diffusion Vapor transport Surface Surface Neck Neck 4 5 Boundary diffusion Lattice diffusion Grain boundary Grain boundary Neck Neck 6 Lattice diffusion Dislocations Neck Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 200 elim ination (Kingery et al., p. 474). From Fig. 8.1. and Table 8.1., we can see thatprocesses 4-6 are mechanisms possibly responsible for shrinkage. W hether all or which of these mechanisms are enhanced by microwave irradiation needs a more detailed analysis. Sintering is a necessary thermal process of ceramics fabrication in which bonding and densification take place in a particulate material, leading to a solid body with mechanical strength. The driving force for sintering process is the difference in Gibbs free energy of the material before and after sintering (AG). This is schematically shown in Fig. 8.2. The fine ceramic powder before sintering is in a higher free energy state (A) than the sintered body with free energy (B). However, the change from state A to B is not a spontaneous process because there is an energy barrier (Ea) to overcome. For the process from state A to B to occur, a certain amount o f energy must be supplied to the material. The minimum energy required for per m ole o f material to overcome the energy barrier is known as activation energy (Ea) for the process. Heating is a means to supply activation energy to the material to overcome this energy barrier. In a simple sintering case of a crystalline ceramic powder compact during which there is no phase transformation, the free energy change that gives rise to densification is the decrease in surface area and lowering o f the surface free energy by the elimination of solidvapor interfaces. This takes place with the coincidental formation of new but lower-energy solid-solid interfaces (Kingery et al., p.469). Obviously, increasing the free energy of the starting powder would lower the activation energy for sintering. It is seen in Fig. 8.2. that when free energy increases from A to A j, A2, A 3 ..., the activation energy correspondingly decreases from E a to Eal, E ^ , E ^ ... Inversely, the lowering of the apparent activation energy implies an increase in free energy o f the material before sintering. Notice that surface area is a storage of free energy, because on the surfaces there are various kinds of crystal defect, vacancies, lattice distortion, dangling bonds, etc., which are at higher energy state than the bulk o f regular lattice. The higher the surface area (the finer the powder), the higher the free energy, and the more reactive the powder, so that very fine powders may sinter at lower temperatures. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 201 E A3 i Eal i Ea Process coordinate Figure 8.2. Schematic showing the energy barrier in the sintering process. As the free energy (chemical potential) of starting material increases from A to A j, A 2, ..., the driving force for the process increases from AG to AGj, AG2, ..., and the activation energy required for the process correspondingly decreases from Ea to ^ a l’ ^a 2 ’ Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 202 8.3. Loss Mechanisms in Microwave Heating From the point of view of microwave-material interaction, a higher surface area will also benefit microwave absorption. Most ceramic materials are ionic solids and can be considered as dielectrics. The dielectric losses are related to electric polarization. There are various possible mechanisms for polarization in a dielectric m aterial, such as electric polarization, atom ic (ionic) polarization, orientation polarization, and space charge polarization. There is a frequency dependence o f the dielectric constant (Kingery et al., p. 923). Tinga (1988) suggests that ionic conduction and dipole orientation are the major mechanisms causing microwave heating at the lower ISM (Industrial, Science and Medical) frequencies for which commercial power sources are available. Similarly, Katz et al. (1991, 1992) indicate that electromagnetic energy can be dissipated in materials through electronic polarization, ionic polarization, conduction, and interfacial polarization. Among these, electronic polarization and ionic vibration are resonance phenomena and, if operable during microwave heating, have the potential to directly change the ion jum p frequency. However, due to the big difference between the microwave frequencies used in microwave heating (109 to 1010 Hz) and the ion jump frequency (about 1013 Hz), ion jump mechanism is not possible. Electronic polarization occurs at even higher frequencies than ionic vibration, thus it is not thought to be operable at microwave frequencies either. Instead, ion jump relaxation, interfacial polarization, and conduction are possible in microwave heating. Ion jump relaxation in a crystalline ceramic occurs when an aliovalent ion (impurity cation or anion with a valence different from that of its host sublattice) and vacancy form an associated pair. An aliovalent ion-vacancy pair has a dipole moment associated with it that responds to the applied electric field. The vacancy is thought to jum p around the aliovalent ion to align its dipole moment with the electric field. Interfacial polarization occurs at a structural inhomogeneity such as a grain boundary, dislocation, or vacancy cluster. In an ionic lattice there will be a localized disruption in electroneutrality at such a structural inhomogeneity with a net dipole moment that will align itself with the applied field. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 203 Conduction occurs when vacancies are not associated with other defects and hence are not localized. Unassociated vacancies are much more mobile than associated pairs and migrate in response to the electric field. This mechanism is mostly of interest at low frequencies. In addition, as indicated by Meng et al. (1994), photon-phonon interactions are another mechanism for absorption of microwave energy, especially in the high frequency range (millimeter-wave and far-infrared). However, its contribution to the frequency range o f 2-20 GHz cannot be ignored, particularly in relatively pure crystals where it could even be the dominant process. The interaction between phonons and photons resulting from quantization o f the electrom agnetic fields and lattice vibrational dynamics leads to microwave absorption in crystals (Gurevich & Tagantsev, 1991). In summary, dipole orientation through ion jum p relaxation and interfacial polarization, ionic conduction, and possibly also photon-phonon interactions, are some mechanisms responsible for microwave absorption in microwave heating. 8.4. Interactions of Material Surfaces with Microwaves 8.4.1. Dielectric Losses Powders are collection of finely dispersed particles. It is well known that the finer the particles, the more active the powder. The origin of the activity o f the powders is the surface area, lattice distortions, and defects. To certain extent, the surface structure of the fine powder is similar to the that of the amorphous. On the surfaces o f the fine particles, there are numerous defects and lattice distortions, which readily couple with microwaves through the mechanisms described above, thus the microwave absorption of finer powders should be stronger. Inversely, surface area is a measure of defect concentration, activity, free energy state, and thus microwave absorption. The higher the free energy state, the stronger the microwave absorption. Accordingly, grain boundaries will couple with microwave more strongly than the bulk of the grain, and amorphous materials will couple much more strongly with microwaves than the crystalline form of the same composition. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 204 8.4.2. Thermal Conductivity Structural imperfections also affect thermal conductivity. Am ong insulators, amorphous solids have smaller thermal conductivities than crystals because the phonon waves are scattered more often in an aperiodic structure (Newnham, 1975). Chemical impurities also promote phonon scattering. Similarly, the thermal conductivities of grain boundaries should be smaller than that of the bulk due to the concentrated structural defects and impurities along the grain boundaries. This will lead to localized heat dissipation and consequently more microwave absorption during irradiation. 8.4.3. Surface Scattering The scattering of microwaves by the material being processed is another possible m echanism to prom ote m icrowave absorption. A green ceramic com pact contains numerous particles and is porous. The irregular shapes, surface roughness, and various defects on and near the surfaces of the particles will scatter microwaves. Such a scattering effect will increase with increasing specific surface area and the degree of imperfection of the material. The scattering effect will increase microwave absorption of the material. As a consequence, powders o f amorphous m aterials such as gels, should have a strong microwave absorption. Figure 8.3. schematically shows the structure of the amorphous material. 8.5. Athermal Effect in Microwave Processing In Sec. 4.6.3., the apparent activation energy for the densification of HAp was found to be 112 kJ/mole in microwave processing, compared with 181 kJ/mole in conventional processing. In the case of diphasic mullite gel (Sec. 5.5.3.), the apparent activation energy for m icrowave and conventional processing was found to be 396 and 702 kJ/mole, respectively. Similar phenomena were observed by other researchers. For example, Janney and Kimrey (1991) reported that the apparent activation energy for microwave Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 205 Figure 8.3. Schematic showing o f an amorphous structure. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 206 sintering o f alumina at 28 GHz was 160 kJ/m ole, compared with 575 kJ/m ole for conventional sintering. W hat does the lower apparent activation energy mean? From Fig. 8 .2 ., we can see that a lower apparent activation energy implies an increase in free energy o f the material. In other words, microwave irradiation seems to promote the chemical potential of the material besides thermal heating. Since the accelerated densification was observed at the same temperatures, the microwave effect must be an athermal effect, which is an additional energy gain beside the volumetric thermal heating during microwave irradiation. 8 .6 . Microwave Enhanced Diffusion Although no satisfactory model has yet been established, microwave enhanced diffusion is evident in experiments. Freeman (1993, 1994) et al., for example, observed an increase of ionic current in an NaCl crystal due to the application of microwave field. Their experiment was conducted to make the with/without microwave comparison in-situ and simultaneously, so that there were no questions of tem perature m easurem ent inaccuracies. These researchers even measured ionic current without supplying external voltage to the crystal. If this result is confirmed, then the microwave fields m ust be supplying the driving force (chemical potential) for the ionic conduction. W hether the driving force effect observed with the single crystal cancels out in polycrystals is still unknown. Rothman (1994) confirms the the microwave enhanced diffusion in his critical review and indicates that the enhanced diffusion observed by Janney and Kimrey (1991) with sapphire single crystal and Fathi et al. (1992) with silicate glass was volume diffusion, since there was no grain boundary in these cases. For polycrystalline m aterials, the microwave enhancement could not be so simply attributed to volume diffusion because numerous grain boundaries with various defects exist. Ionic mobility along a grain boundary can be 3 to 4 orders of magnitude greater than in the bulk (Freeman et al., 1994), so that grain boundary diffusion should be mainly responsible for the densification in the sintering of ceramics. Also, it has been indicated (Sec. 8.4.1.) that grain boundaries Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 207 couple with microwaves more strongly than the bulk, thus the enhancement in sintering should be attributed to the enhancement in grain boundary diffusion. M eek (1987) and Brooske et al. (1991, 1992) have proposed some models to interpret the microwave enhanced densification, but no satisfactory theory has yet been developed. Obviously, more efforts are needed in this aspect. In summary, the above discussion suggests that free energy, or, indirectly, surface area, is an important factor that influences microwave absorption of the material, beside other loss mechanisms. Generally, a higher free energy state corresponds to higher dielectric loss, lower thermal conductivity, stronger scattering effect, thus better microwave absorption. For the same composition, the microwave effect on sintering should increase in order of single crystals, polycrystals, and the amorphous. It is also suggested that microwave enhanced sintering of polycrystalline ceramic materials is mainly resulted from the enhancement in grain boundary diffusion. Now that the microwave effect exists, then the more microwaves absorbed by the material irradiated, the stronger the microwave effect. This is confirmed by the observation o f Freeman et al. (1994) that the ionic current in the NaCl crystal in the microwave field increases with increasing microwave power. 8.7. Microwave Effects on Diphasic Mullite Gels 8.7.1. Observed Microwave Effects on Mullite Materials It has been shown in Chapter 5 that microwave processing significantly accelerated crystallization and densification of diphasic mullite xerogels. By microwave processing, the mullite transformation temperature o f the diphasic gel was lowered by about 75-100°C. The sintering of mullite and m ullite-based composites was much more efficient than conventional processing when using the diphasic xerogel as the mullite precursor. By utilizing microwave processing, the apparent activation energy for the densification o f the diphasic xerogel was lowered from 702 to 396 kJ/mole. By contrast, in the temperature range up to 1500°C, the microwave effect on the densification of monophasic mullite gel and crystalline mullite was not significant. This clearly indicates that the microwave Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 208 enhancing effect on the densification and mullite transformation is directly related to the unique properties of the diphasic gel. 8.7.2. Characteristics of Mullite Gels The scale of homogeneity of the alumina and silica species in the monophasic mullite gel is at the atomic level, mullite forms in the monophasic gel at temperature as low as 970°C. Below this temperature, the sintering o f the monophasic gel is negligible, while over this temperature, crystalline mullite has formed thus sintering and densification rely on the solid-state diffusion mechanism. Since the crystalline mullite is a low-loss material, it does not couple so efficiently with microwaves in the tem perature range of current experiments. Also, due to the low diffusivity of the crystalline m ullite, the rate of densification o f mullite below 1500°C is limited. In contrast, the two phases, alumina and silica sources, in the diphasic gel are mixed at the scale of the nanometer range (~ 1 -1 0 0 nm), thus mullite formation can be delayed to significantly higher temperatures around 1350°C, depending upon the fineness of the alum ina particles, or the degree of hydrolysis o f alum ina source during the sol-gel preparation. Huling and M essing (1989) observed m ullitization at 1364°C with the colloidal boehmite-silia gel. During mullitization, the formation and growth of mullite from the diphasic gels occurs via a nucleation-and-growth mechanism by the reaction between S-A12 0 3 and amorphous S i0 2. It has been established (Sundaresan and Aksay, 1991) that this reaction takes place within the amorphous silicate-rich matrix rather than at the alumina/silica interfaces. Now that microwave processing substantially enhanced mullite transformation and densification of the diphasic gel, the microwave effect must have functioned through the controlling mechanism(s) of these processes. Then, what is the mechanism in each of these two processes? We will discuss them separately. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 209 8.7.3. Microwave Effect on Densification of Diphasic Mullite Xerogel In the diphasic mullite gel, the alumina particles of nanometer size are mixed with silica gel. It has been identified previously that the mullite transformation o f the current diphasic gel takes place at about 1328°C under conventional condition, and 1250°C in the microwave processing (Sec. 22.214.171.124.). The densification o f the diphasic gel by viscous flow of the silica occurs before the mullite transformation. Figure 8.4. shows the diphasic gel system, in which 8-Al2 0 3 particles are homogeneously dispersed in the silica gel matrix. Taking 3.65 and 2.2 g/cm 3 as the densities o f 8 -AI2 O 3 and the silica gel, respectively, calculation shows that in such a gel system o f m ullite stoichiometry (3Ai2 0 3 -2Si02), the volume percentage of 8-Al2 0 3 and the amorphous silica is 60.6% and 39.4%, respectively. If we suppose that all the alumina particles are uniformly covered by a thin layer of silica gel, then the thickness of the silica layer on an alumina particle of 30 nm diameter, for example, would be 3.52 nm, roughly one tenth of the particle size. Silica gel shows viscous flow starting about 700°C and remains amorphous before the mullitization of the diphasic gel. The viscosity of silica decreases as temperature increases, so that the densification of the diphasic mullite gel is a result of viscous flow of the silica gel. Any factors that lower the viscosity of the silica gel will enhance the densification of the diphasic gel. Thus it is obvious that the observed microwave enhanced densification of the diphasic mullite gel was a result of the decrease in the viscosity o f silica gel in the microwave field. As described previously, the amorphous materials should absorb microwave more efficiently than crystals o f the same composition. For example, it was found that the low density polyethylene has twice the microwave loss level of high density polyethylene, and the higher losses o f low density polyethylene have been attributed to the amorphous regions where potions of polyethylene chains are able to oscillate in response to the microwave field (Newnham, 1991). Similarly, the silica gel is amorphous and has a distorted structure (Fig. 8.5.), thus it should have higher dielectric losses and absorb Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. alumina silica Figure 8.4. Model o f the diphasic mullite gel system, showing that 8 -AI2 O3 particles are homogeneously dispersed in the silica gel. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 8.5. Schematic presentation of (a) ordered crystalline silica and (b) random network glassy form o f the same composition. (Kingery e t al., p. 97) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission 212 microwave energy more efficiently. It can be suggested that the microwave absorption by the gel brings about the volumetric heating as well as athermal effects. The microwave enhancing effect on sintering and densification is a result o f the athermal effect, which might be attributed to the increased kinetic energy o f the silica segments locally without creating thermal effect in the bulk. 8.7.4. Microwave Effect on Crystallization and Epitaxy In the mullite transformation of the diphasic gel, either nucleation or growth controls the total process. If nucleation was the process controlling mechanism, then seeding would significantly accelerate the transformation. However, this was not the case. Both the current study (Sec. 126.96.36.199.) and previous work (Huling and M essing, 1989) show that the m ullite transformation temperature o f the diphasic gel lowered under conventional heating is very limited, only about 30° C. This clearly indicates that it is the growth rather than nucleation that controls the mullite transformation in the diphasic gel. The growth rate of m ullite grains can be controlled either by the dissolution of alumina particles into the amorphous phase, or by the diffusion of the dissolved Al species to the mullite grains. W ei and Holloran (1988) suggested that the mullite forms on the interface and grows toward the diphasic region, and the growth mechanism is controlled by a near-interface diffusion process (about 30 nm scale) or by the interfacial reaction itself. On the other hand, Sundaresan and Aksay (1991) established a model to support the concept that the growth rate of mullite grains in the diphasic gel is controlled by the dissolution of the alumina particles into the amorphous phase. The dissolution of alumina particles in the silica gel can be accelerated by increasing the kinetic energy o f the Al species on the surfaces of 5 -Al2 Q 3 particles, whereas the diffusion rate of the Al species in the silica gel may be accelerated by decreasing the viscosity of the gel. W hen subjected to m icrowave irradiation, the absorption of microwaves by the alumina particles might excite the lattice of the alumina and allow more Al species to be released to the region near the surfaces of alumina particles, leading to a Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 213 greater concentration gradient of the dissolved alumina between the surfaces of alumina particles and the matrix. This would increase the driving force for the diffusion of alumina species. On the other hand, the microwave irradiation also increases the kinetic energy of silica segments of the gel, thus increases the motion of the silica particles and decreases the viscosity of the silica gel. As a result, the mobility of the alumina species in the silica gel is promoted. Although both mechanisms are possible, the latter appears to be dominating. This is so because that it has been concluded that the microwave enhanced densification of the diphasic mullite gel is a result o f lowering viscosity of silica gel, while the enhanced densification was observed at about 1200°C which was below the temperature of mullite transformation of the diphasic gel in the microwave field. In summary, the accelerated mullite transformation of the diphasic gel is mainly attributed to the lowering of the viscosity of the silica gel. Seeding lowers the energy barrier in nucleation and may preferentially guide the direction of the crystallization. Seeding in the diphasic mullite gel increases the nucleation frequency and, at the same time, decreases the diffusion distance of the ions. Increasing nucleation num ber increases the demand of ions to the crystal growing sites, while decreasing diffusion distance shortens the time of ion transport. In conventional heating, if the process is not nucleation-controlled, then the enhancement in grain growth is attributed to the decrease in diffusion distance. In microwave processing, however, the microwave effect lowers the viscosity of the amorphous silica and accelerates the ion transport. This additional microwave effect significantly accelerates the crystallization process. Thus an enhanced epitaxy o f the diphasic mullite gel is a natural result of lowering viscosity of the silica gel. This result, in turn, suggests that the crystallization of the diphasic mullite gel is a diffusion-controlled process. 8 .8 . Microwave Effect on Sintering of Other Materials The microwave enhancement in the sintering o f alumina is very significant when using amorphous alum ina as starting material. For example, 20 min. o f microwave sintering o f amorphous alumina at 1500°C achieved 49% higher densification than that Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 214 achieved by conventional sintering at the same temperature for same duration. When using crystalline alumina as starting material, the microwave effect is neglected at the same temperature. Apparently, the amorphous alumina had a higher microwave absorbing efficiency thus showed stronger microwave effect. The densification of HAp was significantly enhanced by microwave processing. It has been found (Sec. 4.6.) that the apparent activation energy for the densification of HAp in microwave sintering is substantially lower than that in conventional sintering 181 kJ/m ole). (1 1 2 versus In this case, the hydroxyls in the crystal structure of HAp, Ca 10(PO4 )6 (OH) 2 , are thought to play an important role in microwave absorption. The hydroxyls are polar function groups which may orient themselves in response to the microwave fields thus make HAp absorb microwaves efficiently. In addition, high surface area of the HAp must also have made contribution to the the microwave absorption. To the two [NZP] materials, CSZP and BZPS, surface effect may be the main interaction mechanism with microwaves contributing to microwave enhanced densification. 8.9. Influence of Processing Temperature on Microwave Effect As shown in Fig. 8.2., for the process to occur from state A to B, energy greater than Ea has to be supplied to the system for activation. E a is the activation energy for the system with respect to this process. The energy supplied to the system can be either thermal or athermal. The thermal energy increases kinetics, and athermal energy increases the driving force (chemical potential). The relative magnitude of the activation energy and the supplied energy determines the occurrence of the process. It was noted that when sintered at 1300°C for 30 min., the sintered density of CSZP was significantly higher than that achieved by conventional sintering for 24 h (Sec. 6.2.), while when sintered at 1500°C for 20 min., microwave effect was almost negligible (1%). Why the microwave enhancing effect in densification diminishes at elevated temperatures? It can be explained as the following. In m icrowave processing, the sintering or densification is a combined effect of thermal heating plus microwave effect (presumably athermal), while under conventional conditions, the same process simply relies on thermal Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 215 heating. Since microwave effect accelerates the process of sintering or densification, it is an additional driving force. W hen the material is processed at such a temperature that the thermal energy alone is not enough to trigger off sintering, sintering may be already in process in the microwave process due to the additional microwave effect. W hen the material is processed at a temperature that thermal energy alone is just enough to trigger sintering, the microwave process will certainly result in a promoted sintering. In these cases, the net promotion is totally attributed to the microwave effect. When temperature is sufficiently high, the conventional heating alone supplies excess therm al energy to substantially enhance the sintering or densification, then the microwave effect would be negligible. Let E t be the thermal energy, and Emw be the athermal microwave effect, T be temperature (T 0 <T 1<T 2 <T 3 ), and suppose Emw<Ea, then the above discussion can be summarized as Table 8.2. Thus it is clear that the sintering of CSZP at 1300°C is case 3, while that at 1500°C is case 4 in Table 8.2. Table 8.2. Influence of processing temperature on microwave effect. Process Microwave Effect 1. T = T 0 , E l0 < E a, a n d E l0 + E mw < Ea Does not occur Invisible 2. T = T j, E t l < E a, but E tl + Emw > Ea Occurs Yes 3. T = T2, E g > Ea, so E g + Emw » Ea Enhanced Yes 4. T = T 3, E t3 » Ea, then E g + Emw ~ E g Enhanced Negligible Conditions Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 216 8.10. Grain Growth during Microwave Processing Grain growth is a process taking place in the late stage of the sintering process of ceramics. Generally, it is thought that at lower tem peratures, grain growth is more significant but densification is not as significant, while at elevated tem peratures, densification process dominates (Fig. 8 .6 .). When the densification processes to such a stage that the removal of porosity is difficult, the recrystallization takes place and the irregular grain growth will occur. In microwave processing, the heating rate is generally higher, so that the time expended on heating stage is shorter, and grain growth in the low temperature range can be relatively suppressed. Although microwave processing enhances both densification and grain growth, the ratio of grain growth rate to the densification rate is smaller than in conventional sintering. This was identified in microwave sintering of HAp ceramics (Sec. 4.3). sintering 0) ca cn o grain g ro w th _i high T low T 1 /R T Figure 8 .6 . Arrhenius plot of the rates of densification and grain growth. (Katz, 1992). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 217 Chapter 9 CONCLUSIONS 9.1. Summary and Conclusions This thesis has focused on microwave processing of various ceramic materials, including hydroxyapatite, mullite, alumina, NZP, and composites o f these materials and partially stabilized zirconia. M icrowave processing experiments were carried out in a mulitmode microwave cavity of 2.45 GHz. Conventional sintering experiments of the duplicate samples were carried out in a special electric furnace with heating rates comparable to those in microwave processing. Microwave effect on densification of these materials has been evaluated. The mechanisms related to the microwave enhancement in densification o f the selected materials and the crystallization of diphasic mullite gel have been discussed. In the synthesis of HAp, emphasis was put on the achievement o f pure, active HAp powder with high thermal stability. The studies o f thermal stability were based on the consideration that HAp o f poor thermal stability will decompose during high-temperature sintering; only highly thermally stable HAp starting material can ensure the achievement of pure HAp ceramics after sintering in air at high temperatures. HAp powders were synthesized by various w et chemical m ethods, including hydrolysis o f brushite (C aH P 04-2H 2 0 ) in a basic region (~pH 8 ) at temperatures around 70°C; hydrolysis of tricalcium phosphate [a-C a 3 (P 04)2] under the save conditions; precipitation method from solutions; and hydrothermal synthesis. It was found that the activity o f HAp products from these methods are comparable, with that by hydrolysis of tracalcium phosphate as exception (lower specific surface area), while the thermal stability varies one from the other. Hydrolysis of DCPD produces fine HAp powder of poor crystallinity. This HAp is Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 218 nonstoichiometric (calcium deficient) and shows poor thermal stability. It decomposes at about 800°C. Ripening treatment of this HAp with calcium chloride in basic region significantly increased the Ca/P ratio and thus improves the thermal stability. The ripened sample was found to be stable to 1350°C. Hydrolysis o f tricalcium phosphate produced highly crystalline HAp whiskers. However, this HAp is also nonstoichiometric and could not be ripened with calcium chloride as m entioned above. Partial decomposition, or dehydroxylation, was observed when heated to 900°C. Precipitation from solutions of C a(N 0 3) 2 and NH 4 H 2 P 0 4 is amorphous and fine. It becomes crystalline when heated to above 700°C. The crystallinity increases with temperature. However, partial decomposition of this HAp was observed at 900° C. The HAp powder synthesized by hydrothermally treating the precipitate from the above solutions at 200°C and 1.5 MPa for 10 to 24 h results in well defined nanocrystalline HAp. This product is stoichiometric and show excellent thermal stability. It is stable to 1370°C in air of 50% relative humidity. Lattice param eter refinem ent of the above HAp samples reveals that chemical stoichiometry and dense crystal structure are the keys to obtain high thermal stability of synthetic HAp. Microwave sintering of various HAp ceramics has been carried out. Regularly dense HAp ceramics were achieved by microwave sintering of compacts of ripened HAp powder. Relative density up to 97% has been achieved in about 10 min. at 1200°C. Porous HAp ceramics with up to 74% porosity were produced with the ripened HAp powder, as well as the HAp whiskers. In the former case, ammonium carbonate was used to create porosity, while in the latter case, the porosity is created naturally by the entanglement of HAp whiskers. Diametral tensile strength measurement o f the above ceramics show that they were fairly strong. Highly transparent HAp ceramics were achieved by both microwave and conventional sintering in air at ambient pressure. Hydrothermally synthesized HAp was used for the fabrication of the transparent HAp ceramics. The microwave processing of such ceramics takes only 5 min. at 1150°C. A kinetics study shows that the densification of HAp by microwave processing has Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 219 lower apparent activation energy than in conventional process, 112 kJ/mol. vs. 181 kJ/mol. It is believed that the hydroxyl ions in the HAp crystal structure is responsible for the efficient sintering of HAp ceramics in the microwave field. M icrowave sintering o f mullite ceramics was carried out with crystalline mullite, monophasic, diphasic, as well as hybrid aluminosilicate gels as starting materials. Owing to the limitation in the working temperature range o f the refractory materials used for thermal insulation, most sintering experiments were carried out only up to 1500°C. It was observed that diphasic gels can be sintered to relative density near 95% at 1300-1500°C within half an hour. Step sintering process works well for the diphasic gels. In this process, the diphasic gel was first densified at 1200-1300°C, then crystallized to mullite by soaking at elevated temperatures for minutes. The crystalline mullite is difficult to densify under the sam e processing conditions, because m ullite does not absorb microwaves efficiently, and at the processing temperatures, the solid state diffusion rate of m ullite is very low. The monophasic gel transforms to mullite at 970°C. Before the transformation, sintering effect is negligible, and after the transformation, it becomes crystalline and behaves in the same way as the crystalline mullite. The densification behavior o f the seeded gels is in between diphasic and monophasic gels. Seeding the diphasic gel low ers m ullitization tem perature by about 30°C, but also lowers the densification process, because of the slow densification kinetics o f the crystalline mullite. The viscous flow o f amorphous silica in the diphasic gels allows the gels densify much more efficiently than crystalline mullite by solid state diffusion. Transparent mullite ceramics were fabricated in this study by both microwave and conventional sintering in air at ambient pressure. Aerogels m ade from the diphasic aluminosilicate gels by supercritical drying were used as starting material for the transparent mullite. High purity, excellent sinterability, and most importantly, the agglomeration-free microstructure, are the key points leading to the transparency. A kinetics study shows that the diphasic mullite gel densifies efficiently at 11501200°C. The densification rate of the diphasic gel in the microwave field is significantly higher than in conventional case. The apparent activation energy for the densification of the diphasic gel in microwave processing found in this study is 396 kJ/mole, compared with Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 220 702 kJ/mole for conventional sintering. It was found in this study that microwave irradiation not only significantly enhances the densification of the diphasic mullite gel, but also apparently accelerates its mullite transform ation. W ithout seeding, m icrowave processing lowers the transform ation temperature by 75-100°C. In conventional processing, the effect o f seeding on mullite transformation was not detected when the content of seeds (monophasic gel) was less than 10%, but 0.5% seeds shows apparent acceleration in m ullite transform ation in teh microwave irradiated samples. This clearly indicates that microwave processing promotes seeding effect on the diaphasic mullite gel. It is suggested, from the above observations, that the microwave enhancement in the densification, and the mullite transformation, and the effect of epitaxy of the diphasic gels are mainly attributed to the decrease in the viscosity o f the amorphous silica in the diphasic gels by microwave irradiation. It is supposed that the loosely bonded fragments of the distorted silica netw ork orient themselves in response to the m icrowave field thus efficiently obsorb microwaves, lowering the viscosity. The attribution of the microwave enhenced densfication and mullite transformation to the decrease in viscosity implies that the mullite transformation in the diaphsic gel is controlled by the diffusion instead of dissolution of the 8 -alumina in the gel. M icrowave sintering of alum ina was conducted at 1500°C for 20 min. using amorphous and crystalline starting materials, respectively. It was found that under the current experimental conditions, when using the amorphous alumina as starting material, the microwave enhancing effect on densification is very significant, while when using crystalline starting m aterial, the microwave enhancing effect was negligible. This indicates that the amorphous alumina precursor absorbs microwaves more efficiently than the crystalline alumina. M icrowave sintering experiments of CSZP ceramics were conducted at 1300 and 1450°C, respectively. The samples microwave sintered at 1300°C for 30 min. achieved higher density than those conventionally sintered at the same temperature for 24 h, indicating that the densification rate of CSZP was substantially enhanced by microwave Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 221 processing. The average grain size of the microwave sintered sample is about the same as that conventionally sintered for 24h. This implies that microwave irradiation enhances densification as well as grain growth, but the extent of enhancement in grain growth is less than that in the densification. In other words, the grain growth process o f CSZP in the microwave processing process was relatively suppressed. Various ceramic composites containing partcially stabilized zirconia were sintered by microwave processing. The microwave enhancement in densification was also identified with these material systems. For the HAp/PSZ system, the typical density of microwave sintered sam ple was 24% higher than the conventionally sintered sam ples. Correspondingly, significant toughening effect has been achieved. The fracture toughness (KIC) o f 3.88 MPaVm has been m easured in the m icrow ave sintered sample of HAp/10%PSZ, compared with 1.56 MPaVm of the pure HAp ceramic of the same relative density. The microwave enhancing effect on densification of the composites based on the amorphous alumina and the diphasic mullite gel was substantial. That with the composites based on the NZP materials (CSZP and BZPS) was not as significant due to the high processing temperatures. The microwave enhancing effect on densification decreases with increasing zirconia content due to the relatively higher sinterability of PSZ over the matrix materials. From the data obtained in the current studies, it is proposed that the free energy state, or chemical potential, of the powder materials is an important extrinsic cause of microwave absorption due to the defects, lattice distortions, and surface scattering which couple with microwave well. The higher the free energy state, the stronger the microwave enhancing effect. Thus the microwave effect on the materials of the same composition increases in the order of single crystals, polycrystals, semicrystal, and the amorphous. It was also proposed that the microwave enhancement in densification and related phenomena is an athermal effect, although the mechanisms are not clear at present. The obtained data indicates that the influence of the microwave enhancement on a desired process o f the materials also depends on the relative magnitude of the athermal effect and thermal effect, thus on the processing temperatures. When processing temperature is Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 222 substantially higher than that necessary for regular conventional heating, the microwave effect will be shielded. 9.2. Suggestions for Future Studies 9.2.1. Dielectric Properties of Amorphous Materials From the studies carried out in this thesis, microwave enhancement in densification of the selected ceramic materials is evident. The most significant microwave effect was observed on the materials of higher free energy state, such as the diphasic mullite gel, amorphous alumina, and HAp of low crystallinity or high surface area. The higher the free energy, the stronger the microwave enhancing effect. Although the mechanisms of microwave enhancement in the densification of these materials, as well as the acceleration in m ullite transformation of diphasic aluminosilicate gels, has been proposed, a more detailed analysis is definitely necessary to promote the understanding of the microwave effect. Based on the experimental observations and the proposed mechanisms, a model could be developed to describe the relationship of the dielectric properties and of the free energy status of the ceramic materials. This will lead to a better control and process design o f the microwave processing. 9.2.2. Microwave Sintering of HAp in Moist Environment In addition, decomposition temperature of HAp depends on the moisture content, or partial pressure of water vapor, in the sintering atmosphere. All sintering experiments in this thesis were carried out in air, necessitating the use of highly thermally stable HAp to ensure that pure HAp was obtained in the resultant ceramics. As indicated in the previous chapter, only the HAp synthesized by hydrothermal method and by ripening of hydrolysis product o f DCPD meet the requirement in thermal stability for high temperature sintering. However, if sintered in a moist environment, the decomposition of HAp would be largely suppressed. It is suggested that study of microwave sintering is carried out in moisture Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 223 using unripened as-synthesized HAp, or HAp prepared by precipitation. 9.2.3. Better Transparent Mullite Ceramics by Process Modification Transparent HAp and m ullite ceram ic disks were fabricated in this thesis by m icrowave and conventional sintering, respetively. Such transparent ceramics were achieved for the first time by ambient pressure sintering. Although the transparency of HAp is high, there is some room for improvement o f the transparency of the mullite ceramics. Realizing the important potential of transparent mullite ceramics in special applications, efforts should be made to effectively rem ove the residual porosity in the ceramics. This may be accomplished by seeding the diphasic mullite gel, or lowering heating rate to completely rem ove the chem ical w ater (silonal groups). These two approaches may be adopted independendy, or in combination. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 224 REFERENCES Agrawal, D.K., Fang, Y., Roy, D.M ., and Roy, R., Fabrication o f hydroxyapatite ceramics by microwave processing, Mat. Res. Soc. Symp. Proc., Vol. 269, 231-236, MRS, Pittsburgh, PA, 1992. ASTM F I 185-88, Standard specification for composition of ceramic hydroxyapatite for surgical implant, American Society of Testing and Materials, 1988. Blake, R.D., and J.D. Katz, “Microwave sintering of large alumina bodies”, in Ceramic Transactions, Vol. 36, M icrowaves: Theory and Application in M aterials Processing”, edited by D.E. Clark, W.R. Tinga, and J.R. Laia, The Am. Ceram. Soc., Westerville, OH, pp. 459-465, (1993). Brain, R.A., H.A. Atwater, “Rapid selective annealing of Cu thin films on silicon using microwaves”, Mat. Res. Soc. 1994 Spring M eeting Abstracts, pp. 311, 1994. Brooske, J.H., R.F. Cooper, and I. Dobson, and L. McCaughan, “Models o f nonthrmal effects on ionic mobility during microwave processing of crystalline solids”, in Ceramic Transactions, Vol. 21, M icrowaves: Theory and Application in M aterials Processing”, edited by D.E. Clark, F.D. Gac, and W.H. Sutton, Am. Ceram. Soc., W esterville, OH, pp. 185-191, (1991). Brooske, J.H., R.F. Cooper, and I. Dobson, “Mechanisms for nonthermal effects on ionic mobility during microwave processing of crystalline solids”, J. Mater. Res., Vol. 7., No. 2, 495-501, (1992). Brooske, J.H., R.F. Cooper, and I. Dobson, and L. McCaughan, “Models o f nonthrmal effects on ionic mobility during microwave processing of crystalline solids”, in Ceramic Transactions, Vol. 21, Microwaves: Theory and Application in M aterials Processing”, edited by D.E. Clark, F.D. Gac, and W.H. Sutton, Am. Ceram. Soc., W esterville, OH, pp. 185-191, (1991). Brow n, P.W., and M. Fulm er, “K inetics o f hydroxyapatite form ation at low temperature”, / . Am. Ceram. Soc., 74(5), 934-940, 1991. Brown, W .E., L.C. Chow, “Dental Restorative Cement Pastes”, US Pat. No. 4518430, May 21, 1985. Brown, W.E., L.C. Chow, “A new calcium phosphate, water-setting cement”, in Cements Research Progress, edited by P.W. Brown, Cements Division, Am. Ceram. Soc., 1986. Brown, W.E., Chow, L .C ., Combinations o f sparingly soluble calcium phosphates, in slurries and pastes as mineralizers and cements, US Pat. 4612053, Sep. 16,1986. Bruce, R.W., “New frontiers in the use of microwave energy: power and m etrology”, Materials Research Society Symposium Proceedings, Vol. 124, Microwave Processing of Materials, edited by W.H. Sutton, M.H. Brooks, and I J . Chabinsky, Materials Research Society, Pittsburgh, PA., 3-15, 1988. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 225 Cheng, J., Y. Fang, D.K. Agrawal, Z. W an, L. Chen, Y. Zhang, J. Zhou, and X. Dong, “Continuous m icrowave sintering o f ceram ics” , Ceramic Transactions, Microwaves: Theory and Applications in M aterials Processing, Am. Ceram. Soc., Westerville, OH 1994. Cheng, J., J. Qiu, J. Zhou, and N. Ye, “Densification kinetics o f alumina during m icrowave sintering” , in M icrowave Processing of Materials III, M aterials Research Society Symposium Proceedings Vol. 269, Edited by R.L. Beatty, W.H. Sutton, and M.F. Iskander, pp. 323-328, Materials Research Society, Pittsburgh, Pennsyvania (1992). Cheng, J., Y. Fang, D.K. Agrawal, Z. W an, L. Chen, Y. Zhang, J. Zhou, and X. Dong, “ Continuous microwave sintering o f ceram ics”, Ceramic Transactions, Microwaves: Theory and Applications in M aterials Processing, Am. Ceram. Soc., Westerville, OH 1994. Costa, B.J., and J.O. Herrm an, “Preparation and sintering o f an hydroxylapatite composite”, in Sintering and microstructure development symposium (paper No. SXVP-993), the 95th Annual Meeting of Am. Ceram. Soc. Abstracts, pp. 377, (1993). Costa, B.J., and J.O. Herrm an, “Preparation and sintering o f an hydroxylapatite composite”, in Sintering and microstructure development symposium (paper No. SXVP-993), the 95th Annual Meeting of Am. Ceram. Soc. Abstracts, pp. 377, (1993). Brown, W .E., and L.C. Chow, “Dental restorative cement pastes”, U.S. Pat. 4518430, M ay 21, 1985. Cullity, B.D., Elements o f X-Ray Diffraction, 2nd edition, pp.418, Addison-Wesley Publishing Compary, Inc., 1978. Ducheyne, P., and J.E. Lemons, Eds. Bioceramics: Material characteristics versus in vivo behavior, Annals of the New York Academy of Sciences, Vol. 523, The New York Academy of sciences, New York, NY., 1988. Eysel, W., D. M. Roy, "Topotactic Reaction o f Aragonite to Hydroxyapatite", Zeits. Kirst. 141, 11-24(1975). Fang, Y., D.K. Agrawal, D.M. Roy, and R. Roy, Rapid sintering o f hydroxyapatite ceramics by microwave processing, Ceramic Transactions, Vol. 21, edited by D.E. Clark, F.D. Gac, and W.H. Sutton, American Ceramic Society, Westerville, Ohio, 1991. Fang, Y., D.K. Agrawal, D.M. Roy, and R. Roy, “Fabrication of porous hydroxyapatite ceramics by microwave processing”, J. Mater. Res., \b l. 7(2), 490-494, 1992. Fang, Y., D.K. Agrawal, D. M. Roy, and R. Roy, "Rapid Sintering of Hydroxyapatite Ceramics by Microwave Processing", Microwaves: Theory and Application in Microwave Processing, Ceramic Transactions, Vol. 21, pp. 349-356 (1991). Fang, Y, D.K. Agrawal, D.M . Roy, and R. Roy, “Microwave sintering of calcium strontium zirconium phosphate ceramics”, Ceramic Transactions, vol. 36, Microwaves: Theory and application in materials processing n, edited by D.E. Clark, W.R. Tinga, and F.R. Laia, pp. 109-114, Am. Ceram. Soc., Westerville, OH, 1993. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 226 Fang, Y., D.K. Agrawal, R. Roy, and D.M. Roy, "Synthesis, dehydration, and ceramics of hydroxylapatite", 34-T-89, 91st Anuual M eeting o f Am. Ceram. Soc., Indianapolis, April 23-27,1989. Fang, Y, D.M . Roy, J. Cheng, R. Roy, and D.K. Agrawal, “M icrowave sintering of hydroxyapatite-based composites”, Ceramic Transactions, vol. 36, Microwaves: Theory and application in materials processing n, edited by D.E. Clark, W .R. Tinga, and F.R. Laia, pp. 397-407, Am. Ceram. Soc., Westerville, OH, 1993. Fang, Y., D.M. Roy, J. Cheng, R. Roy, and D.K. Agrawal, Microwave sintering of hydroxyapatite-based composites”, in Ceramic Transactions, Microwaves: Theory and Application in Materials Processing II, Am. Ceram. Soc., Inc., Westerville, OH (1993). Fathi, Z., D.E. Clark, and R. Hutcheon, “ Surface m odification o f ceram ics using microwave energy”, in Microwave Processing of M aterials III, M at. Res. Soc. Symp. Proc., Vol. 269, edited by R.L. Beatty, W.H. Sutton, and M.F. Iskander, pp. 347-351, Pittsburgh, PA 1992. Feiker, G.E., and N.C. Gittinger, “Rapid Heating of Dielectric M aterials at 915 Me”, Electrical Engineering, Vol. 78, No. 11, pp. 1089, (1959). Ford, J.D ., “H igh Temperature Chemical Processing via M icrowave Absorption”, J. Microwave Power, Vol. 2-2, pp. 61-64, (1967). Freem an, S.A., J.H. Booske, R.F. Cooper, B. Meng, J. Kieffer, and B J . Reardon, “ Studies o f microwave field effects on ionic transport in ionic crystalline solids”, in Ceramic Transactions, Vol. 36, M icrowaves: Theory and Application in M aterials Processing”, edited by D.E. Clark, W.R. Tinga, and J.R. Laia, The Am. Ceram. Soc., Westerville, OH, pp. 213-220, (1993). Freem an, S.A., J.H. Booske, R.F. Cooper, B. Meng, J. Kieffer, and B J . Reardon, “Effects of high power microwave fields on ionic transport in ceramics and ionic crystalline solids”, Proc. M icrowave Absorbing Materials for Accelerators Conf., Newport News, VA, 1993, to be published. Freem an, S.A., J.H. Booske, R.F. Cooper, B. Meng, “M icrowave radiation effects on ionic current in ionic crystalline solids”, in Microwave Processing of Materials IV, Mat. Res. Soc. Symp. Proc., Vol. 347, edited by M.F. Iskander, R J . Lauf, and W.H. Sutton, pp. 479-485, Pittsburgh, PA, 1994. Fricke, J. and A. Emmerling, “Aerogels”, J. Am. Ceram. Soc., 75 [8 ] 2027-36 (1992). Fuller, A J . Baden, M icrowaves, An introduction to microwave theory and techniques, third edition, Pergamon Press, p p .1-3, 1990. Green, D .J., R .H J.H annink, and M.V.Swain, Transformation Toughening of Ceramics, CRC Press, Boca Raton, Florida, 1989. Gurevich, V.L., A.K. Tagantsev, Adv. in Phys. 40, 719-767 (1991), quoted by Meng et al. (1994). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 227 De Groot, K., “Hydroylapatite coating for implants in surgery”, in Ceramics in Clinical Applications, edited by P. Vincenzini, Elsevier (1987), pp381-386. Hanker, J.S., B.L. Giammara, Biomedical Mat. and Devices, MRS Symp. Proc., Vol. 110, MRS, Pittsburgh, PA, 1989. Hattori, T., Y. Iwadate, T. Kato, Hydrothermal synthesis of hydroxyapatite from calcium acetate and triethyl phosphate, Advanced Cer. Mat. 3 (4), 426-428, 1988. H eise, U., J.F. Osborn, and F. Duwe, Hdroxyapatite ceramic as a bone substitute, International Orthopedics (SICOT) 14:329-338,1990. Hench, L.L., Bioceramics: from concept to clinic, J. Am. Ceram. Soc., 1A (7), 14841510, 1991. Ho, W.W., “Hign-temperature Dielectric Properties of Ploycrystalline Ceramics”, pp. 137148 in M icrowave Processing of M aterials, Vol. 124. Edited by W.H. Sutton, MH. Brooks, and I.J. Chabinsky. Materials Research Society, Pittsburgh, PA, 1988. Holcombe, C.E., T.T. Meek, and N.L. Dykes, “Enhanced Thermal Shock Properties of Y2 Q 3 -2 wt% Zr 0 2 Heated Using 2.45 GHz R adiation”, pp. 227-234 in M icrowave Processing o f M aterials, Vol. 124, Edited by W .H. Sutton, M.H. Brooks, and I.J. Chabinsky. Materials Research Society, Pittsburgh, PA, 1988. Huling, J.C., G.L. Messing, “Hybrid gels for homoepitactic nucleation of mullite”, J. Am. Ceram. Soc., 72(9) 1925-29 (1989). Ioku, I., M. Yoshimura, S. Somiya, Hydrothermal synthesis of hydroxyapatite powders with zirconia dispersion, in Biomedical materials and devices, MRS Sympo. Proc., Vol. 110, 445-450, 1989. Janney, M ark A., Hal D. Kimrey, M onica A.Schmidt, and James O. Kiggans, “Grain Growth in M icrowave-Annealed Alumina”, J. Am. Ceram. Soc., 74  1675-1681 (1991). Janney, M ark A., and Hal D. Kimrey, “Diffusion-controlled processes in microwave-fired oxide ceramics”, Mat. Res. Soc. Symp. Proc. Vol. 189, 215-227 (1991). Jarcho, M ., Bolen, C.H., Thomas, M .B., Bobick, J., Kay, J.F., D orem us, R.H., Hydroxyapatite synthesis and characterization in dense polycrystalline form, J. Mat. Sci. 11, 2027-35, 1976. Johnson, D. Lynn, “Microwave Heating o f Grain Boundaries in Ceramics”, J. Am. Ceram. Soc., 74  849-850 (1991). Kanazawa, T ., Ed., Inorganic Phosphate Materials, Elsevier, 1989. Katz, J.D., “Microwave sintering o f ceramics”, Annu. Rev. Mater. Sci. 1992. 22:153-70. Kenkre, V.M., “Theory of microwave interactions with ceramic materials”, in Ceramic Transactions, Vol. 21, Microwaves: Theory and Application in Materials Processing”, Am. Ceram. Soc., Westerville, OH, pp. 69-80, (1991). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 228 Kenkre, V.M, L.Skala, and M.W. Weiser, “Theory of microwave interactions in ceramic materials: the phenomenon of thermal runaway”, J. Mater. Sci. 26 (1991) 2483-2489. Kimrey, H.D., J.O. Kiggans, M.A. Janney, and R.L. Beatty, “Microwave sintering of zirconia-toughened alumina composites”, Materials Research Society Symposium Proceedings, vol. 189, pp.243-255, MRS, 1991. Kingery, W.D., H.K. Bowen, D.R. Uhlm ann, Introduction to Ceramics, 2nd edition, John W iley & Sons, 1976. Kom am eni, S., and R. Roy, "Anomalous Microwave M elting of Zeolites", Materials Letters, 4  107-110 (1986). Lange, F.F., “Transformation toughening”, J. Mater. Sci., 17 (1982) 225-262. Lavernia, C., and J. M. Schoenung, "Calcium Phosphate Ceramics as Bone Substitutes", Am. Ceram. Soc. Bull., 70, 95-100 (1991). Li, Qing Hua, “Synthesis of ceramic powders by m icrowave-hydrothermal process”, Master thesis, The Pennsylvania State University, 1993. Limaye, S.Y., D.K. Agrawal, R. Roy, and Y. Mehrotro, “Synthesis, sintering and thermal expansion o f C a 1.xSrxZr4 P 6 0 24 -an ultra-low thermal expansion ceramic system”, J. Mater. Sci., Vol. 26, pp. 93-98,(1991). Mah, T.-I., K.S. M azdiyasni, “M echanical properties of m ullite”, J. Am. Ceram. Soc., \b l. 6 6 , No. 10, 699-703 (1983). M ason, T.J., and J.P. Lorim er, Sonochemistry: Theory, A pplicatins and Uses of Ultrasound in Chemistry (Ellis H orw oodLtd., Chichester, U.K., 1988). McConnell, D., Apatite: Its Crystal Chemistry, Mineralogy, Utilization, and Geologic and Biologic Occurrences, Springer-Verlag, New York, Wien, 1973. Meek, T.T. “Proposed Model for the Sintering of a Dielectric in a M icrowave Field”, J. Mater. Sci. Lett., 6 , 638-640 (1987). Meng, B., J.Booke, R.Cooper, andS. Freeman, “Microwave absorption in NaCl crystals with various controlled defect conditions”, in Microwave Processing o f Materials IV, Mat. Res. Soc. Symp. Proc., Vol. 347, edited by M.F. Iskander, R.J. Lauf, and W.H. Sutton, pp. 467-472, Pittsburgh, PA, 1994. Milovsky, A.V., Mineralogy & Petrography, pp.235, Mir Publishers, Moscow, 1982. Monma, H., and T. Kamiya, Preparation of hydroxyapatite by the hydrolysis of brushite, J. Mat. Sci. 22, 4247-50, 1987. M onma, H., Takahashi, T., Ushio, H ., Soeda, S., and Kiyosawa, T., Preparation and properties of porous apatite by the hydration and hardening method, Gypsum & Lime, 212, 25-28, 1988. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 229 M onma, H., S. Ueno, and T. Kanazawa, Properties of hydroxyapatite prepared by the hydrolysis of tricalcium phosphate, J. Chem. Tech. Biotechnol., 31,15-24, 1981. Moya, J.S., M.I. Osendi, “Microstructure and mechanical properties of mullite/Zr0 2 composites”, J. Mat. Sci., 19(1984) 2909-2914. Newnham, R.E., S.J. Jang, M. Xu, and F. Jones, “Fundamental interaction mechanisms between microwaves and matter”, in Ceramic Transactions, Vol. 21, Microwaves: Theory and Application in Materials Processing”, edited by D.E. Clark, F.D. Gac, and W .H. Sutton, Am. Ceram. Soc., Westerville, OH, pp. 51-67, (1991). Newnham, R.E., Structure-Property Relations, Springer-Verlag, pp. 57, 1975. Patil, D.S., B.C. Mutsuddy, J. Gavulic, and M. Dachimene, “M icrowave sintering of alumina ceramics in a single mode applicator”, in Ceramic Transactions, Microwaves: Theory and Application in Materials Procession, edited by D.E. Clark, F.D. Gac, and W.H. Sutton, Am. Ceram. Soc., Westerville, Ohio, pp.301-309 (1991). Raham an, M .N., L.C. De Jonghe, S.L. Shinde, and P.H. Tewari, “Sintering and microstructure of mullite aerogels” , J. Am. Ceram. Soc., 71  C-338-C-341 (1988). Riboud, P.V., Bull. Soc. Chim. Fr., Spec. No. Cpmparaison de la stabilite de I’apatite d’oxydedeferet del’hydroxyapatite a haute temperature, 1701-1703, 1968. Rothman, S.J., “Critical assessment of microwave-enhanced diffusion” , in Microwave Processing of Materials IV, M at. Res. Soc. Symp. Proc., Vol. 347, edited by M.F. Iskander, R.J. Lauf, and W.H. Sutton, pp. 9-18, Pittsburgh, PA, 1994. Roy, R., D.K. Agrawal, and V. Srikanth, “Acoustic wave stimulation of low temperature ceramic reactions”, / . Mater. Res. 6, 2412-2416 (1991). Roy, D.M ., Porous Biomaterials and M ethod o f Making Same, US Pat. No. 3929971, 1975. Roy, D. M., and S. K. Linnehan, "Hydroxyapatite formed from Coral Skeletal Carbonate by Hydrothermal Exchange", Nature, 247, 220-22 (1974). Roy, D.M ., L.E. Drafall, land R. Roy, Crystal Chemistry, Crystal Growth, and Phase Equilibria of Apatites, Academic Press, Inc. (1978) pp.186. Roy, R., L.J. Yang, and S. Komameni, Am. Ceram. Soc. Bull., 63 (1984) 459. Roy, R., S. Komarneni, and L.J. Yang, “Controlled microwave heating and melting of gels”, J. Am. Ceram. Soc., 6 8 (7) 392-395 (1985). R udanick, A., A. R. H unter, and F. C. Holden, "An Analysis of the DiametralCompression Test", Materials Research Standards, pp. 283, April, 1963. Stein, D.F., “Microwave processing — An emerging industrial technology?”, Mat. Res. Soc. 1994 Spring Meeting Abstracts, pp. 299,1994. Suslick, K.S., “Sonochemistry”, Science, Vol. 247, 1439-1445 (1990). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 230 Sundaresan, Sankaran and Ilhan A. Aksay, “Mullitization of Diphasic Alumino-silicate Gels”, J. Am. Ceram. Soc., 74 2388-92 (1991). Sutton, W.H., "Microwave Processing of Ceramic Materials", Am. Ceram. Soc. Bull.,  376-86 (1989). 68 Tinga, W.R., and E.M . Edwards, “Dielectric M easurement Using Sw ept Frequency Techniques”, J. Microwave Power, vol. 3, No. 3, pp. 144-75 (1968). Tinga, W.R., and W.A.G. Voss, M icrowave Power Engineering,Vol. 2, pp.189-194, Academic Press, New York (1968). Tinga, W.R., “Interaction of Microwaves with Materials”, in Proceedings of IMPI, Short Course for Users o f Microwave Power (1970). Tinga, W .R., “Fundamentals of microwave-material interactions and sintering”, Mat. Res. Soc. Symp. Proc. Vol. 124, edited by W.H. Sutton, M.H. Brooks, and I J . Chabinsky, Mat. Res. Soc., pp. 33-43, 1988. Tinga, W .R., “Microwave material interactions and process design modeling”, in Ceramic Transactions, Vol. 36, Microwaves: Theory and Application in Materials Processing n, edited by D.E. Clark, J.R. Laia, Jr., pp. 29-43, Am. Ceram. Soc., Westerville, OH, 1993. Vbn Hipel, A.R., Dielectric Materials and Applications, Technology Press o f M IT and Wiley, New York, (1954). W alton, J.D ., Jr., “Inorganic random es” ; pp. 229-344 in Random e Engineering Handbook. Edited by J.D. Walton, Jr., Marcel Dekker, Inc., New York, 1970. Wang, P.E., T.K. Chaki, Sintering behaviour and mechanical properties o f hydroxyapatite and dicalcium phosphate, J. Mat. Sci.: Mat. in Medicine, 4, 150-158, 1993. Weber, J. N., R. T. Greer, B. T. Voight, E. W. White, and R. Roy, "Unusual Strength of Echinoderm Calcite Related Structure", J. Untrastructure Research, 26, 355-66 (1969). Wei, W en-Cheng, and John W. H alloran, “Phase T ransform ation o f D iphasic Aluminosilicate Gels”, J. Am. Ceram, soc., 71  166-72 (1988). Wei, W en-Cheng, and John W. H alloran, “Transform ation Kinetics o f Diphasic Aluminosilicate Gels”, / . Am. Ceram, soc., 71  581-87 (1988). W hite, E. W., J. N. Weber, D. M. Roy, E. L. Owen, R. T. Chiroff and R. A. White, "Replamineform Porous Biomaterials for Hard Tissue Implant Applications", J. Biomed. Mat. Res. Symp. No. 6 , 23-27 (1975). W hite, R. A., J. N. Weber, and E. W. W hite, "Replamineform: A New Process for Preparing Ceramics, Metal, and Polymer Prosthetic M aterials", Science, 176, 922-24 (1972). W illmann, G., Normbestimmungen fur hydrxyapatit als Knochenersatz (Standards for Hydroxyapatite used for Bone Replacement), Sonderdruck aus Biomedizinische Tecknik, 35(9), 205-207, 1990. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 231 W roe, F.C.R., “Scaling up the microwave firing o f ceramics”, in Ceramic Transactions, \fol. 36, Microwaves: Theory and Application in Materials Processing II, edited by D.E. Clark, W.R. Tinga, and J.R. Laia, Am. Ceram. Soc., Wsterville, OH, 449-458, 1993. W u, J.-M ., T.-S. Yeh, Sintering o f Hydroxyapatite-zirconia composite materials, J. Mat. Sci., 23, 3771-3777, 1988. Yamamuro, T., L.L. Hench, J. Wilson, Eds. Handbook o f Bioactive Ceramics, Vol. II, Calcium Phosphate and Hydroxylapatite Ceramics, CRC Press, Boca Raton, FL 1990. Yamashita, K., Kanazawa, T., “Hydroxyapatite”, in Inorganic Phosphate Materials, edited by Kanazawa, T., Elsevier, 1989, p l8 . Ye, Ruilun, Y.H. Fang, and P.W. Lu, Physical Chemistry of Inorganic Materials, China Building Industry Press, 1986, 337-338. Yoshimura, M., K. Ioku, andS. Somiya, “Apatite ceramics with tailored microstructures from hydrothermally-prepared ultra-fine hydroxyapatite crystals”, Euro-Ceramics, Vol. 3, edited by G. de W ith, R.A.Terpstra, and R. Metselaar, Elsevier, 3.17-3.20, 1989. Zhong, J.P., Z. F athi, G.P. LaTorre, D.C. Folz, and D.E. Clark, “ M icrowave densification of porous silica gel”, in Ceramic Transactions, Vol. 36, Microwaves: Theory and Application in Materials Processing II, edited by D.E. Clark, W.R. Tinga, and J.R. Laia, Am. Ceram. Soc., Wsterville, OH, 157-164, 1993. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. VITA Yi Fang was born in Buchu Milage, Zhaotong, Yunnan Province, China. He attended The First Middle School of Zhaotong Prefecture. He studied at Wuhan Institute of Building Materials (presently Wuhan University o f Technology) and graduated with a B.S. in materials engineering in 1982. He worked at Guangxi University until he joined the Intercollege Materials Research Laboratory, The Pennsylvania State University, University Park, PA, as an exchange visitor in 1987. H e worked at the Intercollege M aterials Research Laboratory as a visiting research assistant untill 1989. In the spring of 1989, he joined the graduate program in Solid State Science at The Pennsylvania State University for his Ph.D. During these years as a research assistant, he worked m ainly in materials synthesis and characterization. He authored and coauthored around 20 technical papers on cementitious materials, bioceramic materials, ultrasonic chemical synthesis, ultrasonic sieving, and microwave sintering of various ceramic materials. He is a member o f The American Ceramic Society and Materials Research Society. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.