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 M icrofilms International A Bell & H owell Information C o m p a n y 3 0 0 North Z e e b R oad . Ann Arbor, Ml 4 8 1 0 6 * 1 3 4 6 U SA 3 1 3 /7 6 1 -4 7 0 0 8 0 0 /5 2 1 - 0 6 0 0 Order Number 9311750 M icrowave joining of materials Yiin, Tzu-Yuan, Ph.D . The Pennsylvania State University, 1992 UMI 300 N. Zeeb Rd. Ann Arbor, MI 48106 The Pennsylvania State University The Graduate School Department of Engineering Science and Mechanics M ICROW AVE JO IN IN G O F M A TER IA LS A Thesis in Engineering Science and Mechanics by Tzu-Yuan Yiin Submitted in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy December 1992 We approve the thesis of Tzu-Yuan Yiin. Date of Signature Vhsundara V. Varadan Distinguished Alumni Professor of Engineering Science and Mechanics and Electrical Engineering Thesis Advisor Chair of Committee l/lj/Vu/ Vijcfy K Varadan Distinguished Alumni Professor of Engineering Science and Mechanics and Electrical Engineering f . yL Jgsepl/C. Conway, Jr. / ' Professor of Engineering Mechanics Bernhard R. Tittmann Kunkle Professor of Engineering Science and Mechanics 10 1 16 I 7 2Sridhar Komameni Professor of Clay Mineralogy _____ IchardlfrM cN itt' Professor of Engineering Mechanics Head of the Department of Engineering Science and Mechanics A B ST R A C T This thesis studies and characterizes the joining of similar and dissimilar materials using microwave energy; 2.45 GHz microwave frequency single-mode and multi-mode cavities were designed to join the materials. Loss tangents of soda-lime-silicate glass and sintered a silicon carbide (SA-SiC) were measured as a function of temperature, using HP8510B Network Analyzer at microwave frequencies. Loss tangent increased rapidly with temperature, indicating the possibility of using the microwave energy to heat and join the materials. Joining of dissimilar ceramics (composite/reaction bonded silicon carbide (Si-SiC), glass-ceramic/Si-SiC, and Al/sintered a silicon carbide (SA-SiC)) with and without applied pressure were achieved and required only several minutes. Scanning Electronic Microscopy (SEM) showed uniform joint interfaces. Similar ceramics joining with interlayers (SiSiC/glass/Si-SiC, Si-SiC/Al/Si-SiC, and SA-SiC/Al/SA-SiC) were also joined with microwave power. The Si-SiC/glass/Si-SiC interface had no bubbles (CO 2 and CO) produced at the interface, which is the main problem for conventionally joining ceramics with glass as an interlayer. Reaction bonded Silicon Carbide (Si-SiC) joined to itself with Aluminum as an interlayer was perform ed in a multi-mode cavity. Destructive and non-destructive evaluations on the joined specimens showed that the joined specimens reached the quality of the original SiC. Joining of sintered a Silicon Carbide (SA-SiC) to Aluminum was also achieved in this study. Due to the high thermal conductivity of SA-SiC the joining was done in a very short time (1 minute). Sintered a silicon carbide (SA-SiC) was also joined to itself with A1 as the interlayer. A uniform joint was obtained; however, only 43% of original SA-SiC strength was reached. Electron microprobe examination on the interface showed the image of A1 diffusion paths at the interlayer (Si-SiC/Al/Si-SiC) and the chemical reaction between SA-SiC and A1 (SA-SiC/Al/SA-SiC). A state-of-the-art technology, Scanning Acoustic M icroscopy (SAM), was used to examine the joined interface beneath the surface. Since the detection was beneath the surface, more reliable information about the joints was obtained. For the specimens, Si-SiC/glass/Si-SiC, Si-SiC/Al/Si-SiC, SA-SiC/Al, and SA-SiC/Al/SA-SiC, near the surface of the joints the interfaces were very uniform, no defects were detected by SAM; therefore, a sound joint was achieved with microwave heating. Three- and four-point bend tests were performed to find the fracture strength of the joints. From the fracture flaw observation of three point bending test, the Si-SiC/Al/Si-SiC had very high joined strength (219.4 MPa), and the joined specimens were broken away from the joined interface. W ith the two-parameter Weibull approach, the joined SiSiC/Al/Si-SiC reached the same parameters as the original material. Characterization of material properties was also performed on sintered Barium Strontium Titanate (BST) as an example and is discussed in Appendix A. Mechanical and thermal properties of BST were evaluated by different test methods based on the American Standard Test Methods (ASTM). The coverage of the test methods serves to give some insight into the determination of property information in order to provide comment on the quality of ceramics. V TABLE OF CONTENTS LIST OF T A B L E S ....................................................................................................... L IS T OF Page ix FIG U R E S.............................................................................................. x A C K N O W L E D G E M E N T S..................................................................................... xiii Chapter 1 IN T R O D U C T IO N .................................................................. 1 1.1 Scope o f T h esis.................................................................... 1 1.2 E ngineering C eram ics........................................................... 1 1.3 Im portance o f Jo in in g .......................................................... 5 1.4 Ceramic-Ceramic and Ceramic-Metal Joining...................... 5 1.5 Mechanisms and Parameters of Interface............................. 13 1.6 Literature Review of Silicon Carbide Joining with C onventional H eating.......................................................... 13 MICROWAVE HEATING.......................................................... 18 2.1 M icro w av e s.............................................................................. 18 2.2 M icrow ave-M aterial Interactions........................................ 19 2.2.1 Electrical Phenomena of Dielectric Materials 19 2.2.2 Polarization o f M aterials....................................... 21 2.2.3 Complex Dielectric C onstant................................. 24 2.2.4 Power Dissipated in M aterials.............................. 26 2.2.5 Thermal Runaway of M aterials............................ 27 2.3 Loss Tangent Measurement of M aterials............................. 28 2.4 Applications of Microwave Joining.................................... 31 Chapter 2 2.4.1 2.4.2 Multi-Mode Microwave Cavity Joining by Meek and B lake................................................ Single-Mode Cavity Joining by Palaith et al 31 32 vi 2,4.3 Single-Mode Microwave Cavity (6 GHz) Joining by Fukushim a et al................................................. 35 Advantages of M icrowave H eating................................... 38 MICROWAVE JOINING OF M ATERIA LS.......................... 41 Joining of Low Loss Tangent Materials (First T ype)............ 41 Use of High Loss Tangent Materials as Interlayers (Second T y p e)........................................................................... 43 Joining of High Loss Tangent Materials (Third T ype)........... 44 Experim ental P rocedure...................................................... 45 3.4.1 Materials Characterization for Microwave Joining.... 45 3.4.2 Sam ple P rep aratio n ................................................. 48 3.4.3 M icrowave Joining System s.................................. 51 18.104.22.168 Multi-Mode Cavity....................................... 51 22.214.171.124 Single-Mode Cavity....................................... 53 EVALUATION OF JOINED SPECIM ENS........................ 56 Optical Microscopy..................................................................... 56 Scanning Electron Microscopy................................................. 56 Electron M icroprobe A nalysis........................................... 57 Proof T e stin g ............................................................................ 57 4.4.1 Three-Point Bending T est............................................ 58 4.4.2 Four-Point Bending T est.............................................. 59 4.4.3 Two-Parameter W eibull A pproach.......................... 60 Ultrasonic Non-Destructive Method (Scanning Acoustic M icroscopy)....................................... 62 RESULTS AND D ISC U SSIO N ........................................ 64 Optical Micrographs of Microwave Joined Specimens........ 64 vii 5.2 Scanning Electron Microscope and Electron Microprobe A n a ly s is ............................................................................. 69 5.2.1 S i-SiC /A l/S i-SiC .................................................... 69 5.2.2 SA -SiC /A I/SA -SiC ................................................... 70 5.3 Three Point Bending Test R esu lts.......................................... 71 5.4 Four Point Bending Test R esu lts.......................................... 72 5.5 Statistical Analysis of Fracture Strength of Si-SiC and S A -S iC ....................................................................................... 76 Scanning Acoustic Microscopy................................................... 79 C O N C L U S IO N S ....................................................................... 84 5.6 Chapter 6 R E F E R E N C E S ........................................... 87 AppendixSPECIFICATIONS FOR ENGINEERING CERAMICS USING BARIUM STRONTIUM TITANATE (BST) AS AN E X A M P L E ............................................................................................. 91 A.1 G rain A.2 Dynamic Elastic Modulus Measurements........................... A.3 A.4 S iz e ................................................................................ 91 93 A.2.1 O bjective and Background..................................... 93 A.2.2 Velocity of Ultrasonic Wave P u lses.......................... 93 Water Absorption, Bulk Density, Apparent Porosity and Apparent Specific Gravity of Ceramic............................... 97 A.3.1 S ig n ific a n c e ................................................................. 97 A.3.2 Procedure and R esults............................................ 97 Hardness Test o f C eram ic..................................................... 99 A.4,1 W hat Is H ardness ? ................................................. 99 A.4.2 R ockw ell H ard n ess.................................................. 99 A.4.3 General Description and Test Procedure for R ockw ell H ard n ess.................................................. 100 A.4.4 Remarks of the Rockwell Hardness T est.................... 101 A.4.5 Vickers Hardness N um ber (HV)............................ 101 viii A.5 A .6 A.7 A .8 A.9 A.4.6 Conversion of Vickers Hardness to Rockwell H a rd n e s s...................................................................... 103 Flexural Strength Test of C eram ic...................................... 104 A.5.1 O b je c tiv e ..................................................................... 104 A.5.2 D e sc rip tio n .................................................................. 104 A .5.3 R e s u lts .......................................................................... 108 Therm al Shock R esistance................................................. 109 A.6.1 O b je c tiv e ..................................................................... 109 A.6.2 B ac k g ro u n d ................................................................. 109 A . 6 .3 Procedure and R esults............................................ Ill Thermal Conductivity Measurement of Ceramic................. 112 A.7.1 Objective and Background..................................... 112 A.7.2 Procedure and A pparatus........................................ 113 A .7.3 R e s u lts .......................................................................... 115 T herm al E x p an sio n ............................................................... 115 A.8.1 Objective and Background..................................... 115 A .8.2 Procedure and R esults............................................ 116 Summary...................................................................................... 118 ix LIST OF TABLES Table Page 1.1 Typical C eram ic P ro p erties....................................................................... 4 1.2 Examples of Ceramic Joining Classified by T ype.......................................... 10 1.3 Silicon Carbide Joined with Different Ceramic Powders......................... 16 1.4 Comparison o f Silicon Carbide Joining M ethods................................... 17 2.1 Summary of M icrowave Joining of Ceram ics......................................... 37 3.1 M aterials C h aracteristics............................................................................ 47 3.2 Dim ensions of M icrowave Joining M aterials.......................................... 50 5.1 Summary of Materials Joined with Multi-Mode Microwave Cavity. 65 5.2 Comparison of Original and Joined Si-SiC Mechanical Properties 77 5.3 Joined SA -SiC M echanical Properties.................................................... 78 A. 1 Measured Elastic Constants of 7 BST Specimens and the A verage 96 A.2 Water Absorption, Bulk Density, Apparent Porosity and Apparent Specific Gravity of BST Averaged for 10 Specimens ............................. 98 A.3 Vickers Hardness o f 7 BST Specim ens.................................................. 103 A.4 Vickers Hardness Conversion to the Rockwell and R ockw ell Superficial H ardness............................................................... 104 A.5 Weibull Approach of BST.................................................................................. 109 A .6 Results of BST Subjected to Thermal Shock Water Q uenching................ 112 A.7 Thermal Properties Results of BST.................................................................. 115 A .8 Mechanical and Thermal Properties of BST.................................................... 118 X LIST O F FIG U R ES Figure 1.1 Page Conventional heating for joining ceramic. The direction of heat flow is from the surface to the interior...................... 6 1.2 Conventional heating cycle for joining ceramic to metal.......................... 7 1.3 Compliant joint design (edge brazing joining)........................................ 12 1.4 Com pression joint design (ceram ic-m etal)............................................... 12 2.1 Schematic behavior of charge buildup and current flow in (a) an ideal dielectric and (b) a real dielectric material............................ 21 Schematic representation of polarization by dipole chains and bound ch arg es..................................................................................... 22 2.3 Schematic representation of different mechanisms of polarization 23 2.4 Frequency dependence of polarizability of dielectric materials 2.5 Loss tangent as a function of temperature for (a) glass (b) S iC ................ 29 2.6 Schematic diagram and flow chart of Meek and Blake’s microwave jo in in g .............................................................................................................. 33 2.7 Palaith 34 2.8 Fukushima et al.  single-mode set-up.......................................... 36 3.1 First type of microwave joining (arrows indicate the directions of heat flow)......................................... 42 Second type of microwave joining (arrows indicate the directions of heat flow )......................................... 44 Third type of microwave joining (A rrow s indicate the heat flow )............................................................. 46 Three groups of microwave joining (a) rectangular block ceramic joined to rectangular block ceramic with or without interlayer (b) silicon carbide tube joined to silicon carbide tube with interlayer (c) silicon carbide tube joined to metal tube without interlayer.................... 49 M ulti-mode cavity for microwave 54 2.2 3.2 3.3 3.4 3.5 ........ et al.  single mode set-up................................................ joining............................................. 24 xi 3.6 Schematic diagram of standing waves (TE iq j ) in a waveguide cavity where a and b are the broad and narrow sides of the waveguide cross-section and dc is along the z direction and represents the length of the cavity..................................................................................... 53 3.7 Schematic diagram of single-mode cavity........................................................ 54 3.8 Temperature profile of SiC placed in microwave cavities: (a) Single-m ode (b) M ulti-m ode............................................................. 55 Three types of fracture paths (A) high strength joints (B) medium strength jo in ts, and (C) low strength jo in ts......................................................... 58 4.2 Schematic diagram of the four point test................................................. 59 4.3 Schematic of Weibull strength distribution plot........................................ 61 4.4 Reflection and transmission of sound waves. (a) Transmission and reflection of sound waves at an interface (b) Reflection of sound waves at a defect............................................. 62 4.5 Scanning A coustic M icroscope (SAM )................................................... 63 5.1 Optical micrographs of polished (A) Composite / Si-SiC and (B) Glass-ceramic / Si-SiC interfaces showing uniform joining. A rrow s indicate the joined interfaces.................................................... 66 Optical micrographs of polished (C)Si-SiC/glass / Si-SiC and (D) Si-SiC/Al / Si-SiC interfaces showing uniform joining. A rrows indicate the joined interfaces.................................................... 67 Optical micrographs of polished (E) SA-SiC/Al/S A-SiC and (F) SA-SiC/Al interfaces showing uniform joining. A rrow s indicate the joined interfaces.................................................... 68 Photographs of Si-SiC/Al/Si-SiC interface on the same position by (a) SEM (dark phases shows the SiC grains), (b) electron microprobe image of Al on the interface (Al diffused into the Si but not the SiC grains). Arrow indicates the joined interface......................................... 69 Photographs of SA-SiC/Al/SA-SiC interface on the same position by (a) SEM (b) electron microprobe image of Al on the interface (A layer of AI4 C 3 compound). Arrow indicates the joined interface........................ 70 Fracture paths of three-point bend test of Si-SiC. A rrow s indicate the joined interface...................................................... 71 Schematic diagram of SiC four-point bending test (a) Si-SiC (b) S A -S iC ............................................................................. 73 4.1 5.1 5.1 5.2 5.3 5.4 5.5 xii 5.6 Fracture path o f four-point bending fractured Si-SiC specimen 74 5.7 Fracture path of four-point bending fractured SA-SiC specimen 75 5.8 Weibull plots of four-point bend strength of original and joined Si-SiC frac tu re stre n g th s........................................................................................ 77 Weibull plots of four-point bend strength of joined SA-SiC fracture s tre n g th s ........................................................................................................... 78 (A) Scanning Acoustic Micrograph of the Si-SiC/glass/Si-SiC joined interface. Arrows indicate the joined interface....................................... 80 (B) Scanning Acoustic Micrograph of the Si-SiC/Al/Si-SiC joined interface. Arrows indicate the joined interface....................................... 81 (C) Scanning Acoustic Micrograph of the SA-SiC/Al/SA-SiC joined interface. Arrows indicate the joined interface....................................... 82 (D) Scanning Acoustic Micrograph of the SA-SiC/Al joined interface. A rrow s indicate the joined interface....................................................... 83 A. 1 Scanning Electron Micrograph showing grain size of BST........................... 92 A.2 Ultrasonic velocity measurement system with direct contact o f transducer on specim en....................................................................... 95 A.3 Schematic diagram of Rockwell Hardness and test steps......................... 100 A.4 Schematic diagram of the Vickers hardness test........................................ 102 A.5 Schematic diagram of the four point test................................................ 105 A .6 Schematic diagram of biaxial flexure test............................................... 106 A.7 Weibull approach for the four point bending of BST....................................... 108 A .8 Schematic diagram of apparatus for thermal shock resistance parameter............................................................... Ill Schematic diagram for thermal diffusivity measurement......................... 113 A. 10 Temperature curves measured on BST for therm al conductivity m easurem ent......................................................... 114 A. 11 Thermal expansion behavior of BST as function of temperature 117 5.9 5.10 5.10 5.10 5.10 A.9 ACKNOW LEDGEM ENTS I wish to express my sincere gratitude to my advisor, Professor Vasundara V. Varadan, for giving me the chance to work in the Center for the Engineering of Electronic and Acoustic Materials. I also extend my appreciation to my doctoral committee members ~ Professors Vijay K. Varadan, Joseph C. Conway, Bernhard R. Tittmann, and Sridhar Komameni for their advice and comments. My thanks go to Dr. Fathi Selmi for teaching me in the laboratory and supplying the Barium Strontium Titanate (BST) specimens and also to Mr. Richard Hollinger for the thesis correction. I extend additional thanks to my parents and my w ife’s whole family for their continued support and encouragement. Finally, I would like to thank my wife, Chin-Fen, for all her patience and understanding during this period. 1 Chapter 1 IN T R O D U C T IO N 1.1 Scope of Thesis The scope of this study is to characterize the microwave joining technique and apply it to join similar and dissimilar materials. Conventional joining methods are reviewed in this chapter. Procedures for conventional heating and joining and the associated difficulties are evaluated. Due to the absorption of electromagnetic energy by materials, ceramics can be heated internally and possibly joined together. Microwave interaction with materials and the microwave heating technique will be reviewed specialized about joining. A bibliographical review of conventional and microwave joining and the state of the art equipment for joining materials will be discussed in Chapter 2 and Chapter 3. The dielectric loss tangent as a function o f temperature will also be discussed. Heat flow during microwave joining due to the different material types will be classified in Chapter 3. Single-mode and multimode cavities designed in the lab have been used to join engineering ceramics. Joined specimen characterization has been performed with optical microscope, scanning electron microscope (SEM), electron microprobe analysis, and scanning acoustic microscope. Fourpoint and three-point bending tests are used to check the joined fracture strengths. Statistic evaluation of the fracture strengths is performed by the two parameter Weibull approach. 1.2 Engineering Ceramics Ceramics are useful materials for many engineering and structural applications. It is well known that ceramics and metals exhibit wide differences, but in general, ceramics are stronger, more corrosion resistant, and less thermally expansive or conductive than metals. Ceramics have been used as thermal insulators, abrasion resistors, and widely used as tile, 2 bricks and earthenware. In these usages, only some of the desirable properties of ceramics have been applied. Conventional ceramics are made from the original raw mineral. The raw mineral contains too many impurities which make the processing of high quality ceramics difficult and time consuming. To avoid those defects, some ceramics are prepared using extensive and careful processing that produces materials with properties well suited for the most demanding applications. Such ceramics are named "Engineering Ceramics" to denote their superior and highly reproducible properties necessary for com plex device and structural applications. The technology of formulating engineering ceram ics has been developed tremendously in the last decade in response to the requirements o f industry. M ost engineering ceramics are fabricated by consolidation of a compact of pure, finely divided oxide or nonoxide powders in a furnace at high temperature, except for room temperature chemically bonded cement and concrete. The materials that are involved in the engineering ceramics are crystalline ceramics produced by synthetic processing. These include extremely pure, ultrafine particles of alumina, titanates, carbides, borides, feldspar, and other minerals and chemicals that are carefully processed to the desired forms and shapes. The starting powder is very important in the control of microstructure and the resulting physical properties. This dictates that the controlling variables for the nucleation and growth of ceramic powders are very important. New ceramic processing techniques have also been developed, resulting in a new generation of high performance engineering ceramics. These processes include hot isostatic pressing, sol-gel process, microwave sintering, etc. The most important advantage sought in many of these developments is resistance to increased operating temperatures. Engineering ceramics can provide strength and corrosion resistance at high temperature. Operating at high temperature can have high thermodynamic efficiency in energy conversion devices and also efficiencies associated with decreased needs for cooling. A further advantage of engineering ceramics is the greater acceleration allowed by decreased mass. In addition to the technical advantages of 3 engineering ceramics, the potential low cost is particularly attractive. Low cost is possible because of the abundance of raw materials, especially silicon carbide, alumina, and silicon nitride. Table 1.1 shows some typical ceramic properties [Ceramic Source, 1991]. Those applications of engineering ceramics include heat exchangers, automotive engine components such as turbocharger rotors and roller cam followers, power generation components, cutting tools, biomedical implants, and processing equipm ent used for fabricating a variety of polymer, metal and ceramic parts [Cranmer, 1991]. These applications rely on combinations of properties which can be summarized as: • Retention of mechanical properties at high temperature • Low coefficient of friction • Low coefficient of expansion • Corrosion resistance • Thermal insulation • Electrical insulation * Low density 4 Table 1.1 Typical Ceramic Properties. Property Range of Values Specific Examples Vickers Hardness 5 - 44 GPa glass AI2 O3 SiC Failure Strength 7 0 -1 2 0 0 MPa 70 SiC 200-800 0.75 2.7-4.2 5-7 0.75-18MPa m 1/2 glass AI2 O3 SiC Thermal Expansion 20-100 (x 10-7 /°CT) glass 0.2-2000 W/m°K 30 glass M 2 O3 Fracture Toughness K IC Thermal Conductivity 5 20 200-1000 70 -1 0 0 (x l0 '4) AI2 O3 70-90(xl0‘4) SiC 3 7 (x l0 '4) glass AI2 O3 SiC 2-10 50 500 Dielectric Constant at 1 MHz glass AI2 O3 SiC 2-5 9 40 Dielectric Loss at 1 MHz glass l-5 0 0 (x l0 '4) AI2 O 3 2x1 O' 4 SiC 500 xlO ' 4 5 1,3 Importance of Joining Although ceramics can be processed to obtain some useful properties, it is difficult and expensive to make large or complex shapes from conventional processing. The potential use of engineering ceramics is limited by their lack of formability and their brittle nature. The poor formability leads to significant difficulties in manufacturing structural components at reasonable costs and the brittle nature makes it difficult to accurately predict their service lifetimes. Also, use of ceramics for many structural and electronic applications requires that they be connected to other materials at some point. For example, it is popular to make use o f the high temperature properties of ceramics in conjunction with the formability and toughness of metals by combining the two in the form of joints to form more complex parts. The most reliable method for these applications is to join simple pieces together. Each simple piece can be made easily and also tested before joining to eliminate possible flaws. Joining ceramics to each other or to m etals involves considerable difficulties due to the differences in their physical and chemical natures. Many processing techniques have been developed to tackle this problem and this has resulted in a wide variety of joining methods. Whatever method is used, the formation o f successful joints depends on the achievement of intimate contact between the workpieces, the conversion of these contacting surfaces into an atomically bonded surfaces, and the ability of this interface to accommodate thermal expansion mismatch stresses generated either during cooling after fabrication or by temperature changes in operating conditions. 1A £sramic--Ceramig -and.Ceramic-Metal Joining For conventional joining the pieces to be joined are first heated to the desired temperature with heating elements surrounding the pieces. The heat is generated by the heating elements first and then radiates to the surfaces o f the ceramics or metals. The heat can conduct into the interior, By the thermal conductivity of the ceramics or metals. Figure 1.1 shows the schematic diagram of a conventional heating system. Since the thermal 6 conductivities of ceramics and metals are quite different, a very complex heating cycle is needed. Figure 1.2 is a typical heating process for joining a ceramic and a metal. There are four different steps of heating. The first step (a) is to heat at a very low heating rate which will make the joining pieces have very small temperature gradients. Step (b) is to let the joining pieces reach uniform temperatures throughout. Once the pieces have uniform temperatures, a high heating rate is applied (c) to obtain the desired joining temperature, and then the temperature is held (d) to produce reaction or diffusion. The heating rate cycle is not only time consuming but also costly. Joining techniques which are being actively considered for high strength applications can all be classified as reaction bonding and diffusion joining. This covers the class of joints in which the materials involved react during the process of joint formation, either chemically by forming reaction products at the interfaces, or physically by diffusion. Joining pieces Heating element D rectionofheat Figure 1.1 Conventional heating for joining ceramic. The direction of heat flow is from the surface to the interior. 7 1200 Temperature (°C) 1000 Reaction or diffusion 3°C/min 800 600 400 200 4 6 Time (Hour) Figure 1.2 Conventional heating cycle for joining ceramic to metal. 8 For example, fusion welding, which involves localized melting o f the materials to be joined, can produce joints having high strength and good refractoriness. It is limited in that the materials to be joined should have a good match for their thermal expansion coefficients and melting points. Also, the joint formation requires very high localized fabrication temperatures. Rice  showed that some ceramics, like silicon carbide, cannot be joined by fusion, even when using electron beam methods (high local heating), because of the high melting temperature of silicon carbide (2815.5°C). Diffusion bonding is similar to fusion welding, with sim ilar properties and limitations, though it relies more on diffusion at the interface rather than actual melting of the materials involved [Akselsen, 1992], Consequently, it requires lower temperatures to form the joint compared to fusion welding but requires longer fabrication times and high pressure during the joining process which limits the ability to join complex shapes. The temperatures required are of the order of 1500°C and the pressure is around 15 MPa and higher which are quite high compared to brazed joints. The brazing joining method involves putting a thin active metal between the two ceramic pieces on ceramic and metal pieces to be joined. There are many methods to make the interlayer metals [Mizuhara et al., 1989]: (1) a sheet of active metal (2) a mixture sheet of active metal with joining metal or ceramic (3) commercially available brazing alloy (copper-silver-titanium). This process has several advantages which include: obviating the need for wettable ceramics, lower fabrication temperatures (thus decreasing thermal contraction mismatch) and the reduced rate of degrading chemical interactions. The joints are also compliant because of the ductile interlayer. However, the process has significant disadvantages such as long fabrication times and application temperatures limited by the low-melting soft interlayers. Also, the process requires application of pressure which in turn limits the geometries of the possible joints and requires complicated jigs and fixtures. Glazing involves formation o f joints using glass interlayers. This technology has been commercialized and is primarily used to provide hermetic sealing for arc enclosures or other similar applications. The brittleness of the glass phase, stringent requirements for thermal expansion matching and consequently the complications involved in designing 9 these joints puts them at a significant disadvantage [Zdaniewski et al., 1987]. Another common conventional joining method is the moly-manganese process (MoMn brazing). It is a two-step process: metallizing the ceramic and then joining to ceramic or metal [Mizuhara et al.,1989]. The metallization process is coating a thin layer of metal on the ceramic surface, which is performed at high temperature. It is a very time-consuming and expensive process. It seems that reaction bonding and diffusion joining have excellent potential for ceramic-ceramic and ceramic-metal joining because they can supply good strength. However, one must also note that a severe stress concentration sometimes accompanies joining. The problems of grain growth and formation of pores, in addition to thermal expansion mismatch, then become serious. Especially for the ceramic-metal joining, high thermal stress can be created at the joint because of the thermal expansion mismatch around the interfaces. Therefore, the thermal stress problems must be overcome to obtain reliable joints. The best way to achieve a joint and also overcome the thermal expansion mismatch is by engineering design. Different designs have been developed for different applications. There are two popular designs discussed as follows: (1) Compliant joint design (Figure 1.3): the thermal expansion mismatch problem is reduced by the concentric distortion of the metal cylinder. Also the ductile metal interlayer will deform plastically to distribute the thermal stress on the surface of the ceramic. (2) Compression joint design (Figure 1.4): In this design, the thermal expansion difference between metal and ceramic is used as an advantage to join them. For example, by putting a ceramic part into a metal part at high temperature and then cooling down to room temperature, the outside metal will shrink more than the inner ceramic for a metal having higher thermal expansion coefficient. This will produce a compressive force from the metal to the ceramic to produce a high joint strength. Overall, the conventional methods for joining ceramic-ceramic and ceramic-metal can be summarized as in Table 1.2. 10 Table 1.2 Examples of Ceramic Joining Classified by Type [Loehman, 1990]. CERAMIC-CERAMIC JOINING No Filler Material Name Description Fusion welding Interfacial fusion using localized heat source Diffusion bonding Direct contact with heat and pressure With Filler Material Glazing Inter-facial composition melts, wets, with non-metallic liquid and then bonds (bonding with molten glass is common) Mo-Mn brazing Metallized ceramic with Mo-Mn and then metal braze Active metal brazing Molten alloy containing a reactive metal wets, reacts and bonds to ceramic 11 Table 1. 2 (continued). CERAMIC-METAL JOINING No Filler Material Name Description Glass-metal bonding Molten glass wets metal and reacts when cool Diffusion bonding Direct contact under heat and pressure With Filler Material Reactive, non-metallic Reactive composition wets and reacts liquid bonding while molten and then joined when cool Mo-Mn brazing then metal braze Metallized with Mo-Mn process and Active metal brazing Molten alloy containing a reactive metal wets, reacts and bonds 12 Metal c=i I = Ductile metal interlayer | Ceramic cylinder Figure 1.3 Compliant joint design (edge brazing joining). Ceramic 6 87 511529 Metal Figure 1.4 Compression joint design (ceramic-metal). 13 1.5 Mechanisms and Parameters of Interface rSchwartz.19921 The mechanism of joined interfaces has been investigated for many years. These studies have scrutinized the chemistry of the interfacial reactions and how it affects the bond strength of the joint. The most common approach is the sessile drop techniques. Such experiments provide the means of studying wetting and spreading of a liquid on a rigid solid and changes in the surface and interfacial energies. The technique also provides a convenient method of studying the reactions that occur at the interface by characterizing cross-sections obtained by cutting a frozen sessile drop assembly perpendicularly to the interface. 1.6 Literature Review of Silicon Carbide Joining with Conventional Heating Silicon carbide is a widely used ceramic characterized by high hardness and good strength retention at high temperatures. It has good oxidation resistance due to a protective SiC>2 surface layer. Its thermal shock resistance is considered good for a ceramic due to its relatively low thermal expansion and high thermal conductivity. Silicon carbide exists in a cubic form termed beta silicon carbide and in noncubic forms, which are collectively termed alpha silicon carbide. The alpha form is more stable at high temperature [Srinivasan et al., 1989; Coppola et al.,1982]. Silicon carbide can retain high strength and rigidity to about 1600°C. This coupled with high thermal shock resistance and high thermal conductivity has made it become the candidate material for many high technology applications. The joining of silicon carbides has attracted a lot of researchers. The following are some important examples from the literature covering 1980 through the present. Iseki et al.  joined two blocks of sintered silicon carbide with silicon carbide powder as the interlayer by holding the joining specimens at 1650 °C for 30 minutes in Ar atmosphere with a pressure of 100 MPa. The interlayer silicon carbide powder was prepared by heating a mixture of SiC>2 and carbon black with metallic aluminum. Al played 14 a significant role at the interface. In Iseki’s later paper [Iseki et al., 1984], it was proved that Al has a chemical reaction with silicon carbide. The joined strength was around 83% of the original silicon carbide strength. Iseki et al.  later joined two pieces of sintered silicon carbide with Germanium (Ge) metal powder as the interlayer. The joining was done in a graphite susceptor with induction heating in vacuum holding the specimens at 1180°C for 10 minutes. The Ge was detected to have diffusion into and reaction with the Si phase in silicon carbide. The joined strength was around 50 % of the original silicon carbide strength. Yajima et al.  reported the use of polymers of borodiphenylsiloxane (PB)h mixed with silicon carbide powder as the interlayer to join reaction bonded silicon carbide. The joining specimens were heated at 1500°C for one hour in N2 atmosphere. The (PB)^ heated up to 1500°C in N 2 produces silicon carbide and graphite. The maximum joint strength was around 54 % of the original strength. Iseki et al.  used 0.5 mm thick aluminum (Al) metal as an interlayer to join reaction bonded silicon carbide and sintered silicon carbide. The joining temperature was 1000°C in vacuum for an hour by induction heating. There was no reaction layer at the reaction bonded silicon carbide and Al interface. In this joint, aluminum and free silicon in the silicon carbide formed a continuous liquid phase at high temperature, allowing the Al to penetrate into the silicon carbide. However, cracks were formed because of the thermal expansion mismatch between Al and silicon carbide. There was a reaction at the interface of sintered silicon carbide and Al. It has been reported by Clinton [Iseki et al., 1984] that silicon carbide reacts with Al at 800°C to produce AI4 C 3 . The AI4 C 3 phase was formed at the SA-SiC/Al interface. It is proved later by Yano et al.  that the origin of this difference between reaction bonded silicon carbide and sintered silicon carbide is due to the free silicon in the reaction bonded silicon carbide. Since there are some cracks at the reaction bonded silicon a carbide and Al interface due to the thermal expansion mismatch. The strength is only half of the strength of joined sintered silicon carbide. Interface 15 examination of Sintered a silicon carbide (SA-SiC) /Al/Sintered a silicon carbide (SA-SiC) using high resolution Transmission Electron Microscopy (TEM) was investigated by Yano et al. . A reaction layer (AI4 C 3 ) was produced during the high temperature (1000°C) joining processing. The misfit between the silicon carbide and AI4 C 3 lattice was accommodated by interfacial dislocations and the atomic arrangement of the transition phase. The strength o f the SA-SiC/Al/SA-SiC joint is attributable to the transition layers between both the SA-SiC/Al4 C 3 and the AI4 C 3/AI interfaces, which shows a relatively high coherency of atomic configuration. Moore  has also joined commercial sintered silicon carbide (SA-SiC) by three different methods: hot pressing, diffusion welding and brazing with different ceramic submicron powders as the interlayer. For hot press joining, two silicon carbide blocks were put in a graphite susceptor inside an induction heating furnace with an A r atmosphere. Holding at 1950°C for 3 hours with 13.8 MPa applied pressure, joining was completed with gross plastic deformation. Grain growth across the joining interface gave evidence of a good joint. For the diffusion and brazing joining, six different powders were used as the interlayers. Hot isostatic pressing (HIP) was applied under the conditions of 1950°C and 138 MPa argon pressure applied for two hours. A diffusion joint was achieved and the results are summarized in Table 1.3. It can be learned from this paper that a homogeneous joint can be made by the use of an improved method involving application of a uniform, ultra-thin interlayer which will react completely with the silicon carbide base material, or the use of an interlayer consisting of a mixture of the interlayer powder with submicron silicon carbide powder. 16 Table 1.3 Silicon Carbide Joined with Different Ceramic Powders. Interlayer voids at the interface cracks at the interface A1B No Yes b 4c 3 Yes Yes ZtB 2 Yes No SiB 4 No No AI4 C 3 Yes No MoSi2 No No Boadi et al.  have reported a successfully joined silicon carbide with an AgCu-Ti alloy foil interlayer. The use of foils was thought much simpler and easier to handle and to control interlayer thickness. The interlayer was prepared carefully by using an ultrasonic vibrator and dipping in acid solution to minimize oxidation or contamination of the surfaces. The sandwich specimens (SA-SiC/foil/SA-SiC) were placed in a graphite susceptor and heated in vacuum using an induction heating furnace. Joining temperature was held at 850 to 950°C for 30 minutes. The holding time and temperature decide the joint quality. An original joined silicon carbide strength was achieved by holding at 950°C 30 minutes. Table 1.4 summarizes the joining silicon carbide using the conventional methods. As we can see from the results, the joining time is very long; also the temperature and pressure are very high which consume energy and require high cost facilities. From this review it appears that the best method for conventionally joining "Engineering Ceramics,” like silicon carbide, is to use active metal brazing to produce a reaction interlayer which can give high bond strength. Table 1.4 Compression o f Silicon Carbide Joining Methods. IN T E R LAYER J O IN IN G M ETHOD M A X IM U M TEM P. (°C ) H O L D IN G T IM E ( h r .) A P P L IE D PRESSU RE (M P a ) FRA CTU RE STRENGTH ( % O R IG IN A L ) 100 83 1/6 Vacuum 50 1500 1 Vacuum 54 Brazing 1000 1 Vacuum 50 Hot-Pressure W elding 1950 3 13.8 A1B, % Q , ZrB 2 powder Diffusion 1950 2 138 SiB 4 A h Q M o S i2 powder Brazing 1950 2 Ag-Cu-Ti alloy Brazing 950 0.5 SiC powder Diffusion 1650 0.5 Ge Diffusion 1180 (PB ^ polymer Diffusion Al No 138 Vacuum ? ? ? 100 18 Chapter 2 MICROWAVE H EA TIN G Microwaves are electromagnetic waves which are frequently used as heating sources. Many types of devices have been developed to produce microwaves; in the United States, a magnetron operating at 2.45 GHz is the most popular for industrial applications. There are several potential benefits of using microwave energy to join materials because microwave power causes the material to heat itself. This leads to a rapid and uniform temperature rise, ensuring high thermal efficiency. 2,1 Microwaves The microwave frequency portion of the electromagnetic spectrum has been traditionally used for the transmission and reception of information in communication and radar. Microwaves were first used for heating in the 1940's following the invention of the magnetron during the World War II. Engineers and scientists were then presented with a unique challenge of putting such a device for generating microwaves to peaceful and profitable use [Metaxas, 1983]. In the broadest sense, the electromagnetic spectrum covers the range of frequencies from dc to visible light to X-rays and above. The microwave portion of the spectrum lies in the frequency band between 300 MHz to 300 GHz with wavelengths of 1 m to 1 mm. The other characteristics of electromagnetic waves, besides frequency and wavelength, are magnitude, phase, and the ability to propagate, i.e., transfer energy from one point to another. It is the exploitation of these properties that allows the heating of materials and the determination of their electrical characteristics. With the advent of low cost microwave sources and consumer items has also come a resurgence of interest in using the energy and properties of these frequencies for other important applications [Bruce, 1988]. There are important reasons for using microwave 19 heating over conventional processing methods for certain applications. In the microwave process, heat is generated internally within the ceramic instead of originating from outside heating sources. As a result of this internal and volumetric heating, the direction of heat flow and thermal gradients are the reverse of the conventional heating process [Sutton, 1989]. 2,2 Microwave-Material Interactions Materials are composed of arrangements of atoms as ions or molecules, each made of positively and negatively charged particles having various configurations in empty space and varying states of relative motion. What is of primary interest here is the interaction of electromagnetic waves with materials. Because of the polarization properties of a material, the wavelength of the electromagnetic wave becomes shorter and the magnitude of the electric field is attenuated in the material. This attenuation of the field results in the conversion of electromagnetic energy into thermal energy within the material, i.e., heating. The origin o f the heating lies in the ability of the electric field to polarize the charges in the material and the inability of this polarization to follow extremely rapid reversals o f the electric field. Therefore, the polarization lags the applied electric Held ensuring that the resulting current has a component in phase with the applied electric field resulting in the dissipation of power within the insulting material. 2.2.1 Electrical Phenomena of Dielectric Materials First of all, the electrical properties of a dielectric material can be discussed in relation to the capacitor. A capacitor consists of two conductors separated by air or dielectric material. It is used to store or release an electrical charge Q. The charge on a capacitor is Q = CV (2 . 1) 20 where V is the applied voltage and C is the capacitance. The voltage is proportional to the amount of stored charge, and the current passing through the capacitor is V=qyc= J l d t / C ( 2 .2 ) I = C ( dV / dt) (2.3) V = V0 exp (jcot) (2.4) Ic =j(oCV (2.5) with a sinusoidal voltage a charging current results which is 90° advanced in phase in relation to the applied voltage. However, there is an inertia to charge movement that shows up as a relaxation time for charge transport in a real dielectric material as shown in Figure 2 . 1. In equations (2.4 & 2.5), j equals V-l, and to equals 2 jrf, where f is the frequency in cycles per second. The capacitance C contains both a geometrical and a material factor. For a large plate capacitor of area A and thickness d the geometrical capacitance in vacuum is given by Co = A£()/d ( 2 .6 ) where £q is the permittivity of free space (Eq = 8.86 x 1 0 " ^ F/m). If a ceramic material of dielectric constant £ is inserted between the capacitor plates, 21 c =c0(e / £o) =Cq£ (2.7) where £ ’ is the relative dielectric constant Charge Voltage Current I Q Time- Time- Time(a) I Q Time- Time Time—► (b) Figure 2.1 Schematic behavior of charge buildup and current flow in (a) an ideal dielectric and (b) a real dielectric material. 2.2.2 Polarization of Materials A dielectric material reacts to an electric field differently from free space because it contains charge carriers that can be displaced, and charge displacements within the » dielectric can be neutralized a part of the applied field. Since V = Q / C and C = Cq £ , V = Q /(C o £ ) (2-8) 22 That is, only a fraction of the total charge, the free charge, sets up an electric field; the remainder, the bound charge, polarizes and neutralizes part of the electric field. Figure 2.2 shows the total electric flux density D as the sum of the electric field E and polarization field P: D =£q E + P = EE (2.9) where the polarization is the surface charge density of the bound charge, equal to the dipole moment per unit volume of material: (2 . 10) P = N |X N is the number of dipoles per unit volume and |i is the average dipole moment. Thus polarization can equivalently designate either the bound charge density or the dipole moment per unit volume. EE © Dipole ffl ® EB - \- F—) Free charge |—i—i |—h | Bound charge Figure 2.2 Schematic representation of polarization by dipole chains and bound charges. 23 O f all dipoles or induced dipoles in the material, there are various possible mechanisms for polarization in a dielectric material [Von Hippel, 1954]: (a) Electronic polarization: space charges arising from localized electrical conduction, or the shift of the negative electron cloud in relation to the positive atom nucleus in an electric field (Figure 2.3(a)). (b) ionic polarization (atomic polarization): the displacement of positive and negative ions in relation to one another (Figure 2.3(b)), (c) orientational polarization: occurring in materials possessing permanent electric dipoles randomly oriented in the absence of an external field, but undergoing an orientation toward the applied electric field (Figure 2.3(c)). No field Field applied *4 E (a) Electronic polarization (b) Ionic polarization (c) Orientational polarization Figure 2.3 Schematic representation of different mechanisms of polarization. 24 The polarizability of a dielectric material is dependent on frequency as shown in Figure 2.4. The electronic polarization process will rapidly follow alternative fields in the visible part ( E H z ) of the spectrum. Ionic polarization process is able to follow an applied high frequency field and contribute to the dielectric constant at frequencies to the infrared region (THz) of the spectrum. Orientation polarization has relaxation times corresponding at GHz frequencies. .o dipolar electronic 6 8 12 14 16 Frequency Figure 2.4 Frequency dependence of polarizability of dielectric materials. 2.2.3 Complex Dielectric Constant Maxwell modified Ampere’s law for static fields by including a displacement current density term caused by the rate of change of the total electric flux and derived the following expression for the total current density in a material [Johnk, 1975] V x ( B / p ) = J c + 3 (e E ) / 3t ( 2 . 11) 25 It states that the total current density (curl of B/(i) at any point in a region is the sum o f the electric current density J c (= oE) and the displacement current density 9( e E )/9t at that point. B is the magnetic flux density, ji is the permeability and o is the conductivity of the material. As for sinusoidal electric field variations, E = Eg e -i®1 , equation (2.11) attains the following form for the total current density: J = (J Eq + jco£ Eq (2.12) Re-arranging equation (2.12) yields J = j (0 Eg (£’ -jO / lOEfl) Eq (2.13) However, in free space, where a = 0, Maxwell’s circuital law given by (2.13) becomes J = J g) £ q £ ’ Eq (2.14) Comparison of equations (2.13) and (2.14), we can get an effective dielectric constant of the material where conduction effects dominate: £*c = £ ’ -j(j / w£g = E’ - j £”c (2.15) It is shown that, in order to satisfy Maxwell circuital law in a real dielectric material, the dielectric constant must attain a complex form to account for any loss mechanisms. By 26 analogy any other form of loss, such as polarization, we can define an effective loss factor £”e f f as £”e f f = e ”+ a / ( o e O (2.16) The complex dielectric constant can then be given by e* = e’-je”eff (2.17) The ratio o f the effective loss factor to that of the dielectric constant tan 8 = £”ef f / £ ’ (2.18) is called the loss tangent of material. 2.2.4 Power Dissipated in Materials Microwave heating involves the conversion o f electromagnetic energy into h eat Energy is transported through dielectric material by means of electromagnetic waves. The power flow through a closed surface can be calculated from the integration of the Poynting vector [Johnk, 1976] P = E x H (W/m2) (2.19) Therefore the power dissipated in a dielectric material can be calculated by the integration of equation (2,19) over volume V: Jv V- ( E x H ) d V = J s’ ( E x H)- d S ’ (2.20) 27 using the divergence theorem: V • (E x H ) = (V x E> H - ( V x H > E (2.21) By definition the average power is Pav = - 0/2) I s ’ Real <E x H > dS ’ (2.22) The average power absorbed per unit volume P(W/m^), provides the following basis for heating: Pav = <oe0 e"(.ff | E0 R (2.23) where E q (V/m) is the magnitude of the internal field [Metaxas et al., 1983]. Equation (2.23) shows that the power absorbed varies linearly with the frequency, the dielectric loss factor, and varies with the square of the electric field. 2.2,5 Thermal Runaway of Materials As was mentioned earlier, the principle of microwave heating is based on the fact that the dielectric loss tangent of the materials increases rapidly with temperature. In general, the rate of rise of temperature of a dielectric material, dT/dt, is proportional to its heat input, w £q e"eff I E q P , but heat is conducted away from the material at a rate proportional to a t v 2 t , where a t is the theimal diffusivity. Tan 8 initially will rise slowly with increasing temperature, and then some critical point (Tcrjt) is reached. At temperatures above Tcrjt, tan 5 rises very rapidly, which causes a condition of thermal runaway in a 28 microwave heated material. As tan 8 begins to increase rapidly, the material begins to absorb microwave energy more efficiently, which also raises the temperature. This causes tan 5 to rise even faster. The net result is an exponential increase (runaway) in temperature. Thermal runaway is an important aspect of microwave heating. It can cause undesirable hot spots within a material; it can also be used to heat materials at rapid rates. 2.3 Loss Tangent Measurement of Materials To investigate the thermal runaway phenomena of glass and silicon carbide, the loss tangent was measured as a function of temperature in the frequency range of 8 to 13 GHz. A 15 cm square plate was placed vertically inside a resistively heated furnace. The furnace has two microwave transparent but thermally insulating windows allowing the microwave beam from two spot focusing antennas to pass through the furnace and heated sample. The antennas are connected to a HP 8510B vector network analyzer, TRL calibration and time dom ain gating are used to reduce system errors [Hollinger et al., 1991; Varadan et al.,1991b]. The complex transmission coefficient is measured from which the dielectric constant and loss tangent are computed. Figure 2.5 shows the rapid increase o f the dielectric loss tangent of soda-lime-glass and silicon carbide at microwave frequencies. For the soda-lime-glass, the high loss tangent above 300°C is due to the sodium ionic conduction. This agrees with Meek and Blake's prediction  and also explains how a material with a low loss tangent at room temperature can be easily joined using microwave power. The loss tangent of silicon carbide also increases with temperature. Therefore, ceramics with both low and high loss tangents at room temperature can be heated up with microwave power to high temperature. 29 0. 15- 0. 05- o.oo. 0 100 200 300 400 500 600 temperature ( °C ) (a) Glass Figure 2.5 Loss tangent as a function of temperature for (a) glass (b) SiC 30 loss tangent Q 8 -T - Q2- - Q0- i i i i j i i i i 1 i i i i [ i i r 'r ri i i i I is SO 75 160 E5 ISO temperature (°C ) (b) SiC Figure 2.5 (Cont.) r i >r l 175 ZD 31 2.4 Applications of Microwave Joining In the 1980's, microwave energy has been successfully used to join similar ceramics with or without an adhesive ceramic as summarized in Table 2.1. Microwave heating has the potential for uniform, rapid heating since the energy is absorbed directly inside the heated object, rather than being conducted in from the outside. The possibility for this kind o f joining is to heat up the low dielectric loss ceramic to its thermal runaway temperature region. As the ceramic reaches this region, the loss tangent will increase exponentially with temperature. Because the energy absorbed by ceramic is proportional to the loss tangent, it can be heated up to very high temperatures and joined in a short time if the thermal runaway temperature is reached. 2.4.1 Multi-Mode Microwave Cavity Joining bv Meek and Blake H9861 Meek and Blake  first bonded two plates of AI2 O 3 and A ^C ^-to-K ovar by microwave heating using a mixture of glass sealing powders applied as a thin layer between the plates. The process is shown in Figure 2.6. The powders contained a microwave coupling material (watch oil) which is dependent upon the frequency, power and duration of the microwave energy to be absorbed so the interlayer could be selectively heated and rapidly fused. The joining specimens were placed in a conventional home-type (700W) microwave oven with fibrous alumina insulation which provides good thermal insulation while being “invisible” to microwave energy to be absorbed by the coupling agent. At full power, the microwave energy is coupled to the organic material of the coupling agent and to easily polarizable lead atoms present in the glass sealing material. As the watch oil slowly rises in temperature, it causes the glass to also rise in temperature by convective heat transfer from the microwave heated organic oil coupling agent. Eventually, the oil-glass combination reaches the decomposition temperature (about 320°C) at which point glass liquid phase occurs thus initiating microwave coupling directly to the glass sealing material. Also at this temperature, higher frequency relaxation mechanisms predominate which 32 couple to the microwave radiation. The glass sealing material then reaches the temperatures of 700°C to 800°C which causes chemical reaction at the interface and bonding is achieved in 100 minutes. By using the x-ray analysis, it was discussed that the reaction o f the glass sealing material forms a chemically different seal than that which would be formed by conventional heating because it is formed by diffusion rather than by wetting of the reactants. Also, the microwave process proves to be 8 times faster than conventional joining processes and uses about 5 % of the energy. 2.4.2 Single-Mode Cavity Joining A single-mode cavity as shown in Figure 2.7 can resonate with the desired electromagnetic wave pattern (TE jq ^) in a rectangular cavity composed of a variable iris combined with a movable end wall to compensate for the shift in the cavity resonant frequency resulting from changes in the properties of the materials. The power is supplied by a 2.45 GHz magnetron. The microwaves are directed through standard waveguides to a microwave reflection bridge composed of a ferrite circulator, which sends any power reflected from the cavity to a voltage sampling probe and power head for measurement. High energy concentration on the joining area results in high efficiency for microwave heating and joining. The single-mode cavity was first used to join A ljO j-glass-A ^C ^ in place of Meek’s multi-mode microwave oven . By placing the samples to be joined in a region of high electric Held, joints equivalent to Meek’s A ^O j-g lass-A ljO j  were made with much less power (75W) and in much less time (10 minutes) [Palaith et al.,1988]. Mullite was joined by the same single-mode cavity with microwave heating of the ends of two rods butted together. The ceramic samples protrude through the broad walls of the cavity at its geometrical center. Pressure is applied to the ends protruding outside the cavity with a strain gauge measuring the pressure. Two Mullite rods were joined together without any sintering aids at 1300°C in about 5 minutes [Palaith et al.,1988]. 33 CERAMIC SEALING MATERIAL COUPLING AGENT 1 COMBINED INSULATIVE SURROUNDING TEMPERATURE AND TIME CONTROLS W MICROWAVE FINISHED PRODUCT Figure 2. 6 Schematic diagram and flow chart of Meek and Blake’s  microwave joining. Compressor Acoustic probe Magnetron Circuta,<,r i .C avity probe circulator Pow er heads Power meters Voltage sampling probe Vacuum system Strip chart reader IR pyrometer Figure 2.7 Palaith et al.  single mode set-up. co -p* 35 From the characterization of the joined specimens, some useful features of microwave joining of ceramics can be noted [Palaith et al.,1988]: (a) Power consumption is minimal (100-300 watts) and time required is very short; (b) Microscopic homogeneity is retained in the vicinity of the joint interface on the scale of one micron; (c) Tlie joined specimens exhibit greater strength than the as-received material; (d) Test specimens do not fail at or near the joint, and the joint region is harder and tougher than the balance of the specimen. 2.4.3 Single-Mode Microwave Cavity (6 GHzi Joining Fukushima et al.  used a computer-controlled single mode rectangular cavity, shown in Figure 2.8, to self-join rods of A I 2 O 3 and Si 3 N 4 . The heating source is a microwave oscillator with a klystron capable of amplifying the microwave power up to 3 kW. The cavity was designed to resonate the T E j q 3 mode when microwave power is applied. A directional coupler was used for detecting the reflected power from the cavity; also a Pt-Rh type thermocouple was used to calibrate the temperature. An adjustable iris and plunger were used to tune the cavity to resonate with the joining specimens inside during the process. By using the computer controlled technique, the heating rate of the specimen could be set to a constant which produced a sound joint. Alumina rods of 92% and 96% purity could be directly bonded without any sintering aids. Maximum joint strengths, which were equal to the strength of the original materials, were achieved at temperatures of 1750°C and above in 3 minutes. The bending strength of joined 92% alumina rods was 420 MPa, which was retained at temperatures of up to 800°C. The joining of high purity AI2 O 3 or Si3 N 4 rods required a thin layer of a lower purity sheet placed in the joint prior to microwave heating. The silicon nitride specimens required a nitrogen atmosphere for strong joints to occur. From the microscopic observation, there was little difference in microstructure before and after joining. The sintering aids in grain boundary phases were preferentially heated and melted or diffused. Pyrometer Microwave oscillator Sample Waveguide Attenuator Klystron 0 Plunger Directional coupler Pulse motor Reflection Ins control Controller Plunger control nanpmmre omputer Figure 2.8 Fukushima et al.  single-mode set-up. Table 2.1 Summary o f M icrowave Joining o f Ceramics. M A X IM U M P R O C E S S IN G TEM P. T IM E (m in.) (° C ) A P P L IE D FRA CTU RE P R E S SU R E STREN G TH (M P a ) (% O R IG IN A L ) M A T E R IA L IN T E R JO IN IN G LA Y ER M ETHOD Alumina (96%) G lass Diffusion 1450 100 No Alumina (92%) No Diffusion 1850 3 0 .6 100 Diffusion 1720 3 1.8 90 No Diffusion 1300 3 1.2 100 No Diffusion 1850 20 10 SN501 Diffusion 1720 6 6 .2 74 No Diffusion 1300 9 iqq Alumina (99%) Silicon Nitride (SN501) Silicon Nitride (SN220) Silicon Nitride (SN220) Mullite Alumina 4~5 ? 20 38 2.5 Advantages of Microwave Heating There are several potential benefits of using microwave energy to join materials, all of which derive from the fact that the electromagnetic field penetrates the material and heats internally. This leads to a rapid temperature rise, since one does not have to wait for the heat to be conducted from the surface to the interior. It is quite clear that high quality joined specimens of both oxide and non-oxide ceramics can be made in very short times using modest microwave power and compression, with very little surface preparation. The increased hardness and the scanning electron microscope (SEM) studies in the joint region of the materials showing the microscopic homogeneity of the joint interface suggest a joining mechanism: melting and diffusion of grain boundary phases. The following characteristics and advantages of the microwave heating process bode well for its eventual adoption in future research [Sutton, 1989]: (a) Direct coupling (absorbing) of microwave creates volumetric heating • Potential to heat large sections uniformly • Reversed thermal gradients: surface cooler than interior • Process materials at lower surface temperatures • Rapid removal of water, binders, and gases without rupture or cracking. • Internal stresses reduced by lower thermal gradients • Heat in clean (pure) environment; air, controlled atmospheres, vacuum, or pressure • Control partial pressure of reactive gases for selective oxidadon/reduction of certain phases • Improvement of product quality, uniformity, and yields • Instantaneous response to microwave power changes • Lower thermal mass; precise and automated temperature control 39 (b) Dielectric losses (and heating) accelerate very rapidly with increasing temperature above Tcrjt • Ability to heat "transparent" materials above Tcrjt • Very rapid processing (2 to 50 times faster than conventional) • Densify materials rapidly with minimum grain growth (accelerated sintering) • Reduce process costs (time, energy, and labor) • Ability to heat ceramics well above 2000°C (in air, vacuum, or controlled atmospheres) (c) Microwaves are polarized and coherent; location of maximum electric and magnetic fields can be controlled • Capability of high energy concentration in short times and in selected regions • Frequency and power level optimization for given material, size, and shape • Potential for process automation, flexibility, efficiency, and energy savings • Precise heating of selected region, i e.,brazing or sealing of joints, fiber drawing and plasma generation • Acceleration of sintering and diffusion due to high electric fields; thus densificadon at lower temperatures (d) Differential microwave coupling of phases, additives, and constituents leads to selective heating • Synthesis of new materials and microstructures • Heating of selected zones (brazing and sealing) • Enhanced coupling of microwave transparent materials • Use of fugitive coupling materials for preheating of otherwise transparent materials 40 • Use of microwave coupling materials as shapes or containers to heat the more transparent materials • Superior control over state of oxidation through selective heating of phases and control over oxygen partial pressure 41 Chapter 3 M ICROW AVE JO IN IN G O F M A TER IA LS The use of microwave heating for joining ceramics can, generally, be divided into three types depending on the properties o f materials. Low loss tangent materials, like alumina, cannot absorb microwave energy very well at room temperature. Also, the thermal runaway effect does not dominate until high temperatures. This kind of material can be bonded together without an interlayer and is the first type of microwave joining. In the second type of microwave joining high loss tangent materials, which are easily susceptible to the thermal runaway effect are joined, or a conducting polymer composite is used as an interlayer, to absorb the microwave energy and heat up fast. This is a localized heating process at the interlayer. The third type is joining high loss tangent materials, like silicon carbide, with a low melting temperature material as the interlayer. This type o f joining applies to the microwave heating of high loss tangent ceramics with a reactive interlayer. This solves the most difficult problem encountered in conventional joining, namely, thermal expansion mismatch at the joining interface. What is emphasized here again is that, in all three types, the ceramics are heated from the interior, not from the outer surface as with conventional heating. 3.1 Joining of Low Loss Tangent Materials (First Type! The first type involves the joining of two similar, low room temperature loss tangent ceramics together as shown in Figure 3.1. In this type, the ceramic samples are heated to a high temperature (~1600°C) when they are joined together. The joining process is like melting the joining interfaces. The fracture strength of the joint can be as high as the original ceramic fracture strength according to published results [Palaith, 1989 & Fukushima, 1989]. The big problem with this method is the difficulty to reach and then 42 control the thermal runaway effect because of the steep slope of loss tangent vs. temperature. Another problem is heating the ceramics with microwaves from room temperature to the thermal runaway region. Because of the low loss tangent this may take several minutes. As we know, the thermal runaway region may induce thermal microcracks inside the material and should be avoided. This method can be improved by surrounding the materials to be joined with a high loss tangent material, like silicon carbide. The SiC will heat up fast and then radiate the heat to the joining pieces. The joining pieces can then reach the thermal runaway temperature easily and join. Direction of heat Working region Temperature Force Figure 3.1 First type of microwave joining (arrows indicate the directions of heat flow). 43 12_U se o f High Loss Tangent Materials as Interlavers (Second Type) The second type of microwave joining is to join two, low room temperature loss tangent ceramics by using a high-room-temperature loss tangent ceramic as an interlayer which can reach the thermal runaway region easily with microwave heating as shown in Figure 3.2. The microwave energy is absorbed easily by the interlayer and the heat is conducted to the ceramic. With the localized heating on the interlayer, the thickness of interlayer is not easy to optimize. If a thinner interlayer is used, it can not absorb heat rapidly, w hile thicker interlayer will induce the same problem as the first type (microcracks). Therefore the joined strength never reaches the original strength. To solve the problem of this type, a high loss tangent, with low melting point material, like soda-lime-silicate glass, can be used as the interlayer. The loss tangent of soda-lime-silicate glass starts to increase exponentially around 300°C, and the glass becomes soft around 500°C. The localized heating idea with microwave heating has been successfully applied to join thermoplastics [Varadan et al.,1991a]. There are two kinds of interlayers used for joining thermoplastics. The first kind of interlayer is a conductive polymer composite. The conductive polymer composite will heat up fast and melt, the thermoplastic around the interfaces will also melt when heated by the conducting polymer. At that moment, if pressure is applied, the melting interlayer mixed with melting thermoplastics will be squeezed out of the interface. After cooling down, the thermoplastics are joined without any interlayer remaining. Another kind of interlayer is using a high microwave coupling liquid sprayed on the interfaces. The coupling liquid heats up the interfaces, reaches its boiling temperature and then evaporates out. Since the boiling temperature of many solvents is only around 100°C, the application can only be applied to low melting point material joining, like thermoplastics. With careful design of the joining parameters (time, pressure, power), there will not be much thermal expansion mismatch due to the interlayer. 44 Base material f 1 .5 o c Interlayer t/3 Working region of interlayer Base material Temperature Force Figure 3.2 Second type of microwave joining (arrows indicate the directions of heat flow). 3.3 Joining of High Loss Tangent Materials (Third Type") The third type is to join high loss tangent ceramics together at a lower temperature (~1200°C). Some ceramics, like silicon carbide, have a high loss tangent even at room temperature and a high thermal conductivity. Therefore, they are very easy to heat with microwave power to high temperatures without damaging the samples. This is because the high thermal conductivity of the ceramics conducts the heat to the surfaces where it can escape and thus eliminates the temperature gradient. These properties o f ceramics can be used to join them together with active metal as an adhesive. A1 metal has a low thermal expansion coefficient and has been chosen as a candidate material to join engineering ceramics because of its low melting temperature and possible chemical reaction with ceramics in a conventional furnace [Iseki et al., 1984]. In the microwave heating, although metals cannot absorb microwave energy due to their high conductivity, the heat generated from the ceramics could heat up the metal by conduction. We can call this type a “ Selective Heating” method. It can also be applied to join two totally different materials, such as 45 ceramic to metal, with microwave heating. As mentioned earlier, the loss tangent of silicon carbide has a tendency to increase with temperature (thermal runaway effect). It can be heated up internally to over 1000°C in a very short time as shown in Figure 2.2 (b) and has been widely used as a susceptor to preheat microwave sintering specimens [Yu et al.,1991; Janney et al.,1992]. Putting SiC with metal together in the microwave field, the metal part will reflect the electromagnetic waves and is not heated by the microwave field, only the SiC heats up. If the metal is in contact with the SiC, the heat will conduct to it, and only the joining area on the metal is heated up to make the bond. As mentioned in Chapter 1, the conventional joining method has difficulty in making the temperatures of the joining pieces uniform, especially for joining ceramic to metal which has a much higher thermal conductivity. Slowing down the heating rate can reduce the temperature difference, but high cost plus a complex heating cycle will come with it. This type of microwave joining can easily solve the conventional joining problems. This avoids the problem on thermal expansion mismatch because the bulk metal is not heated in the short time needed. Figure 3.3 shows the schematic diagram and the directions of heat flow to metal with microwave power. 3.4 Experimental Procedure 3.4.1 Materials Characterization for Microwave Joining The compositions of the materials used for microwave joining are shown in Table 3.1. The aluminum used as the adhesive contains 98.5 vol. % Al, 0.5 vol.% Fe and 1.0 vol.% Si. The surface roughness o f Al foil for interlayer was around 1 fim and the thickness was 16 |xm before joining. To minimize the oxidation on the Al surface and avoid contamination on all the surfaces, the Al was used immediately after cleaning in the ultrasonic cleaner. 46 Force Force i ♦ SiC Direction of heat SiC Force T Force Figure 3.3 Third type of microwave joining (Arrows indicate the heat flow). 47 Table 3.1 Materials Characteristics. Aluminum: 98.5 vol.% Al, 0.5 vol.% Fe, 1.0 vol.% Si Composites: 20 % carbon fiber reinforced glass matrix composite Glass: Soda-Lime-Silicate microscope slide glass Glass-ceramic: LAS (L i20A l203-Si02) Si-SiC: Siliconized silicon carbide Bi-modal SiC grains with 11 vol.% free silicon metal surrounded (Norton Co.,) SA-SiC: Hexoloy® alpha sintered silicon carbide Submicron SiC powders mixed with 0.1-5 wt.% boron and 0.1-2 wt. % carbon (Carborundum Co.,) Commercial soda-lime-silicate glass and LAS (Li2 O-Al2 O 3 -Si0 2 ) glass-ceramic were used in the study. The glass-ceramic process, discovered by Stookey and Armistead [McMillan, 1964], is a technique by which many small crystals are grown in a glass in order to transform the glass article to a fine-grained, fully dense polycrystalline ceramic. Such ceramics generally have mechanical properties that are superior to the glasses from which they were formed and they are much more stable at high temperatures. The advantage of ceramics formed by this process is that they contain no porosity as is normally encountered in conventionally processed polycrystalline ceramics. The process consists of melting a glass and forming the desired shape, which is followed by crystallization using a multi-temperature heat treatment. Carbon fiber reinforced glass matrix composite is made by mixing 20% carbon fiber with silicate glass as the matrix. The materials were combined in a twin screw extruder. In this process, the glass matrix is first melted, then specific weight fraction of the fiber and the glass matrix are mixed in the extruder. During the extrusion process, the two materials are thoroughly mixed together and pushed towards a circular die. In the die, the material cools and hardens into a small diameter cylinder. The cylinder is then chopped into 48 small pellets. The composite pellets are then melted for injection molding. During injection molding, the melted material is injected into a mold with high pressure. The high pressure helps to restrict the volume of voids in the final product. The composite then cools and solidifies in the mold [Lubin, 1982], Commercial, reaction bonded silicon carbide (Si-SiC) samples were made by Norton Co., Worcester, Mass. The Si-SiC exhibits a bi-modal microstructure o f SiC grains, the small grains range from 2 to 5 Jim and large grains range 100 to 150 pm. The Si-SiC is manufactured by a slip casting process. The raw materials are milled and prepared into slurry for slip casting. The slip is then cast into molds and removed for drying. The green components undergoes the siliconization process where they are fired at temperatures exceeding 1600°C to facilitate both the densification of the SiC and the infiltration of the polycrystalline silicon metal. This material contains 11% by volume free silicon, and the bulk density was 3.1 g/cm^. Si-SiC appears to contain no porosity, indicating a perfect bonding and infiltration between the silicon and the SiC grains. The sintered alpha silicon carbide (SA-SiC) is made by an extrusion process by Carborundum. Hexoloy® alpha silicon carbide is produced by pressureless sintering the ultra-pure submicron powder derived from the original Acheson process. The powder is mixed with boron (~0.1-5.0 % by weight) and carbon (~0.1-2.0 % by weight) sintering aids, then formed into complex shapes by a variety o f methods and consolidated by sintering at a temperature above 2000°C. The material has a uniform microstructure with 25 pm sized grains and contains approximately 2-3 % porosity. 3.4.2 Sample Preparation The dimension of the materials to be joined are summarized as shown in Figure 3.4 and Table 3.2. They can be divided into three groups: (a) rectangular block ceramic joined to rectangular block ceramic with or without interlayer (b) silicon carbide tube joined to 49 silicon carbide tube with interlayer (c) silicon carbide tube joined to metal tube without interlayer. All the joining interfaces were ground first with a 15 Jim grit diamond wheel and then polished down to 1 Jim with diamond pastes. The surfaces were prepared parallel to each other to avoid any mismatch between them. To minimize the oxidation on the metals and avoid contam ination on all the surfaces, the samples thus prepared were used immediately after cleaning in the ultrasonic cleaner. After microwave joining, the specimens were cut perpendicular to the joined interfaces with 320 grit diamond saw and gradually polished to 0.25 mm with diamond pastes. After ultrasonic cleaning, the specimens were put in a desiccator and made ready for the non-destructive evaluation. Tf"H ^ j Ceramic p Ceramic Zeramir Interlayer Ceramic (a) I ■ Ceramic (b) Metal (Al) (c) Figure 3.4 Three groups of microwave joining (a) rectangular block ceramic joined to rectangular block ceramic with or without interlayer (b) silicon carbide tube joined to silicon carbide tube with interlayer (c) silicon carbide tube joined to metal tube without interlayer. 50 Table 3.2 Dimensions of Microwave Joining Materials. A B C t (mm) (mm) (mm) (mm) Joining material Group Si-SiC (a) 30 20 6 SA-SiC (a) 35 25 6 SA-SiC (b) SA-SiC (c) Composite (a) 10 2 10 Glass-ceramic (a) 10 2 10 Aluminum (interlayer) (a) 30 Aluminum (interlayer) (b) Aluminum (0 Glass (a) D1 (mm) 45 D2 El (mm) (mm) E2 40 12 6 0.016 0.016 45 40 20 30 6 0.15 E3 (mm) (mir 12 9 51 3.4,3 Microwave Joining Systems 3,4.3.1 Multi-Mode Cavity A multi-mode microwave power supply was used first for this application to find the best technique for microwave joining. The multi-mode cavity used for joining samples was a Kenmore 700 watt house-hold microwave oven as shown in Figure 3.5. The oven contains a pulse power supply magnetron and a 12”x l 2 ” cavity. The rotating plate and plastic parts in the cavity were removed. One small hole drilled from the top of the cavity is designed to put a quartz tube and thermocouple into the cavity. A microwave leakage detector was turned on whenever power was on to make sure the system was safe. The oven was placed on two multi-jacks which can be adjusted to apply pressure on the specimens. The pressure was added by increasing the height of the multi-jacks. The force on the specimens was read on the weight scale below the multijacks. The temperature was recorded by putting a thermocouple on the joining interface during heating. Although the multi-mode microwave oven has non-uniform wave patterns in the cavity, SiC has a high dielectric loss factor at room temperature and behaves like a sink to absorb almost all the energy from the magnetron. Also, the samples are small compared to the wavelength, so the field variations in the sample are small. For the low loss tangent materials joining, the specimens can be surrounded by the SiC bars or put into a SiC cylinder. The loss tangent of SiC increases with temperature (thermal runaway effect), and the heat radiates to the joining specimens, once the thermal runaway region of the low loss tangent specimens have been reached, then the specimens could heat up and join. Force _L .cavity c sramii: Magnetron (power supply) c sramii: Weight Scale Figure 3.5 Multi-mode cavity for microwave joining. 53 3.4.3,2 Single-Mode Cavity In preparation for single-mode cavity microwave joining, the specimens were placed inside the center of a single mode copper cavity with microwave power supplied as shown in Figure 3.7. The power was coupled to the cavity through a variable iris and tuned to resonate at 2.45 GHz by an adjustable short. The electromagnetic wave pattern in the rectangular cavity is TE jq j mode as shown schematically in Figure 3.6. Pressure was applied to the specimen from outside the cavity through tubes on the broad sides of the cavity. These tubes are designed to provide access to the cavity and are designed to have a cut off frequency much higher than 2.45 GHz so there is no microwave leakage. The whole process was conducted in air atmosphere during heating. When joining was done, the power was turned off and the samples were cooled down to room temperature. The heating history of SiC recorded on the joining interface by a pyrometer when SiC was placed in single-mode cavity and by a thermocouple when placed in a multi-mode cavity is shown in Figure 3.8. b x Figure 3.6 Schematic diagram of standing waves (TE 1Q3 ) in a waveguide cavity where a and b are the broad and narrow sides of the waveguide cross-section and dc is along the z direction and represents the length of the cavity. POWER SOURCE CONTROL UN IT FORCE ■ CAVITY APPLICATOR MAGNETRON HEAD 0TO 3K W 2.45 G H Z WAVEGUIDE COUPLER PYROMETER 71 VARIABLE O CERAMIC SAMPLE FORCE Figure 3.7 Schematic diagram of single-mode cavity. ADJUSTABLE WAVEGUIDE SHORT 55 Temperature (°C) 1200 - 800(a) Single-mode (b) Multi-mode 400- 0 180 360 540 720 900 Time(Seconds) Figure 3.8 Temperature profile of SiC placed in microwave cavities: (a) Single-mode (b) Multi-mode. 56 Chapter 4 EVALUATION O F JO IN E D SP E C IM E N S This chapter is concerned with the testing, characterization, and non-destructive evaluation of the microwave joined specimens. The purpose of the testing is to provide a basis to evaluate the joined specimen properties. 4.1 Optical Microscopy Optical Microscopy is a primary analytical technique used to examine microstructure with reflected light from the specimen surface. Optical microscopy is often well suited to the characterization of grain structure, porosity, and identifying flaws and defects in material surfaces [Bloss,1961]. 4.2 Scanning Electron Microscopy For the surface examination of the joined interfaces, scanning electron microscopy (SEM) and electron microprobe techniques were performed. The SEM has a Robinson backscatter detector and can image surface features in the pm range. In this instrument the electron beam is brought to a fine focus on a very small region of the specimen and the scattered electrons are collected on the fixed detector, the signal of which is fed to a cathode-ray tube (CRT). The beam is swept across the surface of the specimen and the CRT, which is synchronized with the beam, and traces an image of the specimen [Reimer, 1972]. 57 4.3 Electron Microprobe analysis Electron microprobe analysis provides the means o f determining the chemical composition o f very small volumes at the joined interface. The name derives from the essential feature of a fine electron beam which is directed at the point to be analyzed.The xrays generated by the incident beam are characteristic of the elements, and the intensity of these x-rays is an approximately linear funcdon of concentration. The electron microprobe consists of an electrooptical system which features an electron beam that is focused onto an area about 1 |im in diameter on the surface of the specimen, a detection system which measures the intensity of the characteristic radiation of the elements to be determined. Scanning the beam over the specimen allows the distribution of chosen elements on the surface of the specimen to be displayed as an enlarged image on the screen o f a CRT. If the electron beam is swept across the surface while the x-ray detector is set to record a wavelength characteristic of a particular element, a plot can be obtained on an oscilloscope; brightness in the oscilloscope image corresponds to points in the specimen where the elem ent is found in abundance. In addition, the slowing of the electrons produces a continuous spectrum which constitutes a background upon which the characteristic lines are superimposed [Mckinley, 1972]. 4.4 Proof Testing Joint strength measurements are needed to monitor the development of a joining process. Conventional testing methods can be classified into tensile, bend and shear tests. The strength is calculated by dividing the fracture load by the area of the joined surface. The tensile and shear tests may not produce pure tensile or shear at the joint because of the hold-down forces or thermal contraction in different material configurations, both of which can affect the joined strength. Furthermore, it is very difficult and expensive to machine specimens and grip then without causing damage. For those reasons, strengths of joined 58 specimens are often determined by bend tests. In bend tests, the stress is not uniform but one considers the strength in terms o f the maximum tensile stress being applied. This maximum stress occurs in some region along one of the outer surfaces of the specimen. In this way, fracture is generally biased to flaws at or near the outer surface. Three and four point bend tests are two common methods to determine the joint strengths. 4.4.1 Three Point Bending Test Most of the three-point bending test fracture specimens can be classified into three types according to fracture paths as shown in Figure 4.1 which can approximately tell the joined strength [Suganuma et al., 1986]. The first type of fracture path consists of a ductile fracture plane within the interlayer (metal) and a brittle fracture plane within the ceramic. Figure 4.1(A) shows this type fracture path. The joint strength o f this type could have at least 80 % of the original ceramic strength. Type B and Type C are similar to each other but the strength o f type B is higher than that of type C. Differing from type A, the fracture paths of both types are limited within a narrow region near the interface. Ceramic Ceramic (A) (B) (C) Figure 4.1 Three types of fracture paths (A) high strength joints (B) medium strength joints, and (C) low strength joints. 59 4.4.2 Four Point Bending Test Up to now there are no national or international standards specifying loading conditions, shapes and dimensions of the specimens, but symmetrical specimens of rectangular cross sections are usually preferred. Using elastic beam theory [Beer & Johnston,1976], the flexural joined strength of sandwich specimens of rectangular cross section as shown in Figure 4.2 can be determined from the fracture load F and the specimen dimensions from the equation: Gf = 3FD/B2W (4.1) Where Gf (MPa) is fracture strength, F (N) is the fracture load, W (mm.) and B (mm.) are the width and height of the specimen, and D (mm.) is the difference between inner span and outer span. This formula which is used often assumes the material to be elastic (the load-displacement behavior needs to be linear). F Joined interface Figure 4.2 Schematic diagram of the four point test. 60 4.4.3 Two Parameter Weibull Approach The strength distribution of ceramics can often have a large standard deviation and thus it is often important to know the form of this distribution. An empirical distribution that has become very popular is the two parameter Weibull distribution and is based on a specimen at its weakest link. This distribution can be written in the form ln(l/(l-F))=K V ( Gf/o 0 )m (4.2) where m is the Weibull modulus, Oq is die characteristic strength, Of is the fracture strength o f specimen, V is the factor of volumetric flaws, K is a geometry factor, and F is the failure probability which can be determined by different ranking procedures based on the fracture stress. The failure probability is determined by F = (n -0 .5 )/N (4 .3 ) where n is the rank number o f the specimen from lowest to highest stress and N is the total number of specimens tested. By calculating F and measuring the maximum stress at fracture (i.e., a,^), the Weibull modulus can be determined from the slope of a best fit linear regression o f a lnln (1/1-F) vs. In o m plot as seen in Figure 4.3. One can also determine the characteristic strength ( oq) by setting the failure probability (F) equal to 63.2% . K is a factor that depends on the loading geometry in the strength test. For the case of uniform tension, the entire volume of the specimen is loaded in tension; thus, K = l. If a specimen is loaded in any other manner besides uniform tension, K will be less than unity. Oq is a normalizing constant called the characteristic strength which represents the strength 61 where the cumulative failure distribution equals zero. This always occurs at a probability of failure equal to 63.2%. The parameter m is generally called the W eibull modulus which represents the degree of variability in the flaw population. Values of m for ceramics are generally in the range 5 to 20. The width of the strength distribution decreases as m increases. As we see from equation (4.2), the strength of a specimen failing at a particular failure probability depends of the loading and the specimen size. The equation assumes failure can occur from anywhere in the body (volumetric flaws). If failure occurs only from surface flaws, V in equation (4.2) is replaced by A, an area term. - - .999 - .632 | ce © L. &© U 3 •M 3 fc - ■ .007 LN (Maximum Strength) 12 Figure 4.3 Schematic of Weibull strength distribution plot. 4.5 Ultrasonic Non-Destructive Method (Scanning Acoustic Microscopy) W hen the properties of a material are sensitive to overall structural features, measurements of sound velocity or ultrasonic attenuation can be used for NDE. The ultrasonic detection and characterization of individual interior defects, such as voids, cracks, and inclusions in the size range of 20 to 100 pm, in ceramics usually require the use of high frequency, high resolution techniques. These techniques are generally referred to as acoustic microscopy. Most conventional ultrasonic NDE is based on the transmission or back-reflection of ultrasonic waves to generate an acoustic image of a defect or second phase, with the contrast depending on the acoustic mismatch between the different media (Figure 4.4). Reflected ^ Incident ^ ~ ~ Transmitted * medium 1 medium 2 (a) medium 1 \ ] X medium 2 ► I defect back wall (b) Figure 4.4 Reflection and transmission of sound waves, (a) Transmission and reflection of sound waves at an interface (b) reflection of sound waves at a defect. 63 The need to detect defects on or near the surface has attracted methods based on the recently developed acoustic microscope in which the images are formed by the interaction of sound waves with the specimen. Acoustic Microscopy is the general term applied to high resolution, high frequency ultrasonic inspection techniques which produce images of features beneath the surface of a specimen. Because ultrasonic energy requires continuity of materials to propagate, internal defects such as voids, inclusions, delamination and cracks interfere with the transmission and (or) reflection of ultrasonic signals. The image contrast results from the differences in the elastic properties o f the various constituents in the specimen and can thus yield quite different information from the conventional optical microscope [Gilmore et al., 1987]. In the scanning acoustic microscope (Figure 4.5) the ultrasonic beam is focused on the surface by refraction through the coupling media and either the lens or specimen can be scanned to build up an image from the reflection of longitudinal waves back into the coupling media. Contrast is brought about through variations in the reflection coefficients across the sample. This method is receiving increasing attention in the NDE field and shows great promise because it has advantage over optical microscope o f penetrating slightly into the material. Imaging system High-frequency pulse Transducer Coupling med (water) Specimen Figure 4.5 Scanning Acoustic Microscope (SAM) 64 Chapter 5 R E SU L T S AND D ISC U SSIO N Microwave power was successfully applied to join materials as discussed in chapter 3. Joined silicon carbide joining specimens were then chosen for the destructive and non destructive evaluations to check the joined specimens qualities. Optical and Scanning Electron Microscopy (SEM) were performed to check the uniformity of interfaces on the surfaces. Chemical reaction and element diffusion along the interfaces were detected by the electron microprobe technique which could image the elements at the joined interface. Fracture strength and fracture flaws along the joined interfaces were also tested by four point and three point bend tests. Statistical analysis of joined SA-SiC and joined Si-SiC fracture strengths were performed with the two parameter Weibull approach. Scanning Acoustic Microscopy (SAM) examination beneath the joined interface can provided a more complete evaluation of the microwave joined silicon carbides. 5.1 Optical Micrographs of Microwave Joined Specimens Table 5.1 summarizes the joining conditions for silicon carbide joined to itself and other materials. Figure 5.1 (A) to Figure 5.1 (F) are the typical optical micrographs taken at the joined interfaces. Uniform interfaces without any cracks were observed for these specimens. 65 Table 5.1 Summary of Materials Joined with Multi-Mode Microwave Cavity. MATERIAL INTER LAYER PROCESSING TIME (Min.) APPLIED PRESSURE (MPa) APPLIED POWER (Watt) Composite/Si-SiC No 10 0.034 450 Glass-ceramic/Si-SiC No 20 2.0 450 Si-SiC/Si-SiC Glass 2 1.0 450 Si-SiC/Si-SiC A1 5 1.2 450 SA-SiC/SA-SiC A1 10 1.24 700 SA-SiC/Al No 1 No 700 (B) Figure 5.1 Optical micrographs of polished (A) Composite / Si-SiC and (B) Glass-ceramic / Si-SiC interfaces showing uniform joining. Arrows indicate the joined interfaces. (D) Figure 5.1 Optical micrographs of polished (C) Si-SiC/glass / Si-SiC and (D) Si-SiC/Al /Si-SiC interfaces showing uniform joining. Arrows indicate the joined interfaces. 68 (F) Figure 5.1 Optical micrographs of polished (E) S A-SiC/Al / SA-SiC and (F) SA-SiC/Al interfaces showing uniform joining. Arrows indicate the joined interfaces. 69 5.2 Scanning Electron Microscope and Electron Microprobe Analysis 5,2.1 Si-SiC/Al/Si-SiC Figure 5.2 shows the SEM and electron microprobe photographs of Si-SiC/Al/SiSiC. From the variation of A 1 composition across Si-SiC/Al/Si-SiC detected by electron microprobe, no reaction layer between silicon carbide grain and aluminum could be detected. While aluminum and the free silicon phase in Si-SiC formed a continuous liquid phase at high temperature, in addition to the microwave heating enhanced diffusion, A1 penetrates (diffuses) into the Si phase deeply but not into SiC grains. No cracks were formed at the interface, which could be attributed to the short heating dme. (a) (b) Figure 5.2 Photographs of Si-SiC/Al/Si-SiC interface on the same position by (a) SEM (dark phases shows the SiC grains), (b) electron microprobe image of A 1 at the interface (A1 diffused into the Si but not the SiC grains). Arrow indicate the joined interface. 70 5.2.2 SA-SiC/Al/SA-SiC Figure 5.3 shows the SEM and electron microprobe photographs of SA-SiC/Al/SASiC interface. Since the SA-SiC contains only SiC, no silicon phase as in Si-SiC, a layer of chemical reaction product A ^ C j ^ j (3SiC(g) + 4A 1^ -» Al^Cj^g^ +3Si(sp was formed at the interface. Although the joint appears good, the weak fracture strength (152 MPa) and high tendency to absorb water of the AI4 C 3 layer [Iseki et al., 1983] at room temperature make the SA-SiC/Al/SA-SiC joint strength very low (195 MPa). This has been shown by the four point bending test, only 50% o f original strength was reached by joined SASiC/Al/SA-SiC. (a) (b) Figure 5.3 Photographs of SA-SiC/Al/SA-SiC interface on the same position by (a) SEM (b) electron microprobe image of Al at the interface (A layer of AI4 C 3 compound). Arrow indicates the joined interface 71 5,3 Three Point Bending Test Results Joined Si-SiC's specimens were indented on the interfaces for three-point bend test to estimate strength from the direction of the fracture paths. The typical three-point bending Si-SiC fracture specimen shows in Figure 5.4 that the joint has the same fracture path as type (A) of Figure 4.1, the fracture path still flowed into the base material which indicates at least 80 % of the original strength. 2 0 0 >’ 200 Figure 5.4 Fracture path for the three-point bend test of Si-SiC. Arrows indicate the joined interface. 72 5.4 Four Point Bending Test Results Original Si-SiC's without microwave joining and joined Si-SiC's were cut into rectangular bars of dimension 1.5 x 2.0 x 30 mm 3 as shown in Figure 5.5 (a) for the fourpoint bending tests. For SA-SiC’s joined, the original and sandwich joined specimens were cut into 5 x 5 x 50 mm^ bars for the four point bending tests as shown in Figure 5.5 (b). All the tensile surfaces for testing were polished down to 1 |im and edges were beveled to remove any surface crack flaws caused by machining of the testing bars. Ten original Si-SiC and SA-SiC bars and 10 joined Si-SiC and SA-SiC bars were fractured separately, using an Instron test machine in air atmosphere at a cross head speed of 0.5 mm/min. Typical fracture path of the Si-SiC four-point bend test is shown in Figure 5.6. For the four point-bend tests, all the joined specimens fractured away from the interface but still in the inner span. The Si-SiC/Al/Si-SiC interface has a deep diffusion o f Al into the Si phase to make continuously mixed phases; however, the SA-SiC contains a chemical reaction layer of A ^ C j ^ at the interface, and as was pointed out earlier the compound AI4 C 3 has a tendency to be corroded by water and weak fracture strength (152 MPa) at room temperature [Iseki et al., 1983]. The fracture path o f the SA-SiC was along the AI4 C 3 layer as shown in Figure 5.7, the joined strength compared to the original material strength is quite different. 73 1.5 mm Si-SiC Si-SiC 1.5 mm 25 mm (a) Si-SiC 4.5 mm SA-SiC SA-SiC 4.0 mm 13 mm 39 mm (b) SA-SiC Figure 5.5 Schematic diagram of SiC four-point bending test (a) Si-SiC (b) SA-SiC. 74 200 ^im Figure 5.6 Fracture path of four-point bending fracture Si-SiC specimen. Figure 5.7 Fracture path of four-point bending fracture SA-SiC specimen. 76 Microwave heating enhanced the chemical reaction between Al and SA-SiC and the resulting compound AI4 C 3 layer has weaken the joined strength. Suitable interlayers which can react with SA-SiC and produce a strong compound should be chosen to solve the problem. 5.5 Statistical Analysis of the Fracture Strength of Si-SiC and SA-SiC The two parameter Weibull modulus plot was used for statistically analyzing the fracture strengths of joined Si-SiC's and SA-SiC’s. Results from the two parameter Weibull statistical approach are shown in Figure 5.8 and Tabe 5.2 for Si-SiC's. The joined specimen reached 100% of original Si-SiC strength due to the deep diffusion of Al metal. This data proves that the interface has a higher strength than the Si-SiC. The average fracture strength o f SA-SiC’s is 194.4 MPa which is only 50% of original SA-SiC strength as shown in Figure 5.9 and Table 5.3. This is due to the chemical reaction between Al and SiC; a weak layer containing AI4 C 3 was formed at the interface. 77 Fracture Strength (MPa) 200.3 - i. » 244.7 270.4 original joined £ Ln Strength (MPa) Figure 5.8 Weibull plots of four-point bend strength of original and joined Si-SiC fracture strengths. Table 5.2 Comparison of Original and Joined Si-SiC Mechanical Properties. Specimen Configuration Average Weibull Strength (MPa) Modulus (m) Standard Deviation Characteristic Strength (Oq) Original 215.4 15.1 17 2 2 2 .2 Joined 219.4 14.2 18 225.6 78 Fracture Strength (MPa) 164 r 200.3 244.7 El Joined SA-SiC El □ IQ □ □ i - 2 □ ' -3- IS □ 4.9 5.1 5.3 5.5 5.7 Ln Strength (MPa) Figure 5.9 Weibull plots of four-point bend strength of joined SA-SiC fracture strengths. Table 5.3 Joined SA-SiC Mechanical Properties Specimen Configuration Weibull Average Strength (MPa) Modulus (m) Standard Deviation Characteristic Strength ( o q ) Joined 194.4 46 213 5.1 79 5.6 Scanning Acoustic Microscopy Due to the specimen preparation, damage always happens on the surface. So, Scanning Acoustic Microscopy (SAM) was used to examine the microwave joints beneath the surface. A high frequency transducer (400 MHz) was applied. Figure 5.10 shows the joined interface beneath the surface. The colors on the micrographs tell the relative amplitude of reflected waves; for example, if black is shown on the image micrograph, it tells that a defect exists. Also the resolving power of the transducer is 2.5 |Lim, any defects or cracks larger than 2.5 (im can be detected and shown on the image micrograph. There were no defects or cracks detected in the microwave joints; sound joints beneath the surface were achieved with microwave heating. 80 Figure 5.10 (A) Scanning Acoustic Micrograph of the Si-SiC/glass/Si-SiC joined interface. Arrows indicate the joined interface. Figure 5.10 (B) Scanning Acoustic Micrograph of the Si-SiC/Al/Si-SiC joined interface. Arrows indicate the joined interface. 82 Figure 5.10 (C) Scanning Acoustic Micrograph of the SA-SiC/Al/SA-SiC joined interface. Arrows indicate the joined interface. Figure 5,10 (D) Scanning Acoustic Micrograph of the SA-SiC/Al joined interface. Areows indicate the joined interface. 84 Chapter 6 C O N C L U S IO N S The main objective of this investigation was to apply microwave heating to join materials. 2.45 GHz microwave frequency single-mode and multi-mode cavities were designed to join the materials. Measurements of the loss tangents of soda-lime-silicate glass and sintered a silicon carbide (SA-SiC) were performed with HP8510B Network Analyzer to prove the loss tangent of ceramics can increase rapidly with temperature above some temperatures - thermal runaway effect, which is the key for microwave heating. Three types o f microwave heating to join materials have been proposed. Combined with the idea for conventional joining of silicon carbide, the third type of microwave joining has been pointed out as the best way to join silicon carbide. Dissimilar material joints (Composite/Si-SiC, Glass-ceramic/Si-SiC, and Al/SASiC) with or without applied pressure have been achieved in minutes. Optical micrographs and SEM show uniform joining interfaces. Similar ceramics joining with interlayers (Si-SiC/glass/Si-SiC, Si-SiC/Al/Si-SiC, and SA-SiC/Al/SA-SiC) were also joined with microwave power. The Si-SiC/glass/Si-SiC interface has no bubbles (CO 2 , CO) produced at the interface which cannot be achieved by conventional heating with glass as interlayer. Optical micrograph also shows the sound joint. Reaction bonded silicon carbide (Si-SiC) joined to itself with aluminum as an interlayer was done in a multi-mode cavity. Three and four point bending tests were performed to test the fracture behavior of the joints. From the fracture flaw observation of three point bending test, the Si-SiC/Al/Si-SiC has very high joint strength. Also the four point bending fracture test shown that the Si-SiC/Al/Si-SiC fracture specimens were broken away from the joined interface; interface strength equal to the original base material is 85 proved. With the two parameter Weibull approach, the joined Si-SiC/Al/Si-SiC has reached the same parameters as the original material. High joined strength is due to the diffusion of Al metal into the Si phase of Si-SiC. Electron microprobe analysis has proved this phenomenon at the joined interface. Sintered a silicon carbide (SA-SiC) was also joined to itself with Al as the interlayer. Although a uniform joint was made, the chemical reaction between Al and SiC produced an AI4 C 3 compound which is much weaker than the SA-SiC, and only 43% of original SA-SiC strength was reached. Electron microprobe examination on the interface has shown the image of chemical reaction layer between SA-SiC and Al. Sintered a silicon carbide (SA-SiC) joined to Al metal was also achieved in this study. Due to the high thermal conductivity of SA-SiC, the joining can be done in a very short time. The most modem technology, Scanning Acoustic Microscope (SAM), was used to examine the joined interface beneath the surface. Since the detection is beneath the surface, more confident conclusions about the joints can be made. Si-SiC/glass/Si-SiC, SiSiC/Al/Si-SiC, SA-SiC/Al, and SA-SiC/Al/SA-SiC near the surface of the joints have very uniform interfaces, no defects can be detected by SAM, a sound joint is achieved with microwave heating. The main conclusions of the investigation can be summarized as follows: • The problems with conventional heating to join material can be solved with microwave heating. • Ceramics can be heated up in a short time with minimal microwave power supplied. • High loss tangent ceramics, such as silicon carbide, have high heating rates when placed in a microwave cavity. 86 * Microwave joining of silicon carbide to itself or other materials can be achieved. • The Si-SiC/Al/Si-SiC joint has reached the original Si-SiC strength due to the “microwave enhanced diffusion” effect. •The SA-SiC/Al/SA-SiC joint strength is only 43% of original SA-SiC strength due to the “microwave enhanced chemical reaction” which produces a weak layer compound. 87 REFEREN CES A k se ls e n , O.M ., “Review Diffusion Bonding of Ceram ics,” Journal of Materials Science. 27, pp. 569-579 (1992). A S T M -E 1 1 2, "Standard Test Methods for Determining Average Grain Size,” Annual Book of ASTM Standards. Vol. 02.01, pp. 893-918 (1991). A S T M -C 7 6 9 , "Standard Test Method for Sonic Velocity in Manufactured Carbon and Graphite Materials for Use in Obtaining an Approximate Young’s Modulus,” Annual Book o f ASTM Standards. Vol. 15.01, pp. 228-230 (1991). A S T M -E 1 8 , "Standard Test Methods for Rockwell Hardness and Rockwell Superficial Hardness of Metallic Materials." Annual Book of ASTM Standards. Vol. 02.03, pp. 561573 (1991). A S T M -E 1 4 0 , "Standard Hardness Conversion Tables for M etals,” Annual Book of ASTM Standards. Vol. 02.01, pp. 888-902 (1991). A S T M -C 2 0 , "Standard Test Methods for Apparent Porosity, Water Absorption, Apparent Specific Gravity, and Bulk Density of Burned Refractory Brick and Shapes by Boiling Water.” Annual Book of ASTM Standards. Vol. 15.01, pp. 5-7 (1991). A S T M -F 3 9 4 , "Standard Test Method for Biaxial Flexure Strength (Modulus o f Rupture) o f Ceramic Substrates,” Annual Book of ASTM Standards. Vol. 10.04, pp. 313-317 (1991). A ST M -C 554,"Standard Test Method for Fired Glazed Ceramic Whitewares by a Thermal Shock M ethod,” Annual Book of ASTM Standards. Vol. 15.02, pp. 153-154 (1991). A S T M -C 4 0 8 "Standard Test Method for Thermal Conductivity of Whiteware Ceramics,” Annual Book of ASTM Standards. Vol. 15.02, pp. 114-117 (1991). A ST M -C 372,"Standard Test Method for Linear Thermal Expansion of Porcelain Enamel and Glaze Frits and Fired Ceramic Whiteware Products by the Dilatometer M ethod," Annual Book of ASTM Standards. Vol. 02.01, pp. 108-111 (1991). B ee r, F. P. and Johnston, E. R., Mechanics of Materials, pp. 150-59, 351-55 (1981). B illin g to n , N. S., J. Sci. Inst.. 2 6 ,2 0 (1949). B lo s s , F.D., An Introduction to the Methods of Optical Crystallography, Holt Rinehart and Winston (1961). B o ad i, J. K., Yano, T., Iseki, T., "Brazing o f Pressureless-sintered SiC using Ag-Cu-Ti Alloy" pp. 2431-2434, J. Mater. Sci. (1987). 88 B ru c e, R. W.,"New Frontiers in Use of Microwave Energy: Power and Metrology," pp. 3-15 in Microwave Processing of Materials. Symposium Processing, Vol. 124. Edited by W. H. Sutton, M. H. Brooks, and I. J. Chabinsky. Materials Research Society, Pittsburgh, PA (1988). Ceramic Source, Vol. 5, Am. Ceram. Soc.. Columbus, OH (1991). Coppola, J.A., Hailey, L.H., and Mcmurty, C.H., “Sintered Silicon Carbide Body,” U.S. Patent 4.312.954. Jan. 26,(1982). Cranmer, D.C., “ Overview of Technical, Engineering, and Advanced Ceram ics,” Engineered Materials handbook. Vol. 4, pp. 16-20, ASM international (1991). Davidge, R. W., Mechanical Behavior of Ceramics. Cambridge University Press (1979). Fukushima, H., Yamanaka, T.,and Matsui, M., "Microwave Heating of Ceramics and Its Application to Joining," pp. 267-72 in Microwave Processing of Materials. Symposium Processing, Vol. 124. Edited by W. H. Sutton, M. H. Brooks, and I. J. Chabinsky. Materials Research Society, Pittsburgh, PA (1988). Gilmore, R. S., Johnson, R. E., T rzaskos, C.R., and Young, J.D .,” Acoustic Microscopy: Materials Art and Materials Science,” Review of Progress in Quantitative Nondestructive Evaluation. Vol. 6 A, Plenum Press, pp. 553 (1987). Hayden, H. W., Moffatt, W. G. and Wulff, J.. The Structure and Properties of Materials. \blum e III: Mechanical Behavior, John Wiley and Sons, New York, NY (1965). Himsworth, F. R.,”Determination of the Thermal Conductivity of Insulating Bricks,” Trans. Brit. Ceram. Soc.. 56, 345-55 (1957). Hollinger, R. D., Varadan, V.V., Varadan, V.K., and Ghodgaonkar, D. K., “Free-Space M easurem ents o f High-Temperature, Complex Dielectric Properties at M icrowave Frequencies,” Ceramic Transactions. Vol.21, pp. 243-250 (1991). Iseki, T., Arakawa, K., Suzuki, H., "Joining of Dense Silicon Carbide by Hotpressing,"pp. 1049-1050. J. Mater. Sci. Lett. 15 (1980). Iseki, T., Kameda, T., and Maruyama, T., “ Some Properties of Sintered A14C3,” L. Mater. Sci. Lett. 2. pp. 675-676(1983). Iseki, T., Kameda, T., Maruyama, T„ "Interfacial reactions between SiC and aluminum during joining," pp. 1692-1698.J. Mater. Sci. 19(1984). Janney, M.A., Calhoun, C.L., Kimrey, H.D.,’’Microwave Sintering of Solid Oxide Fuel Cell Materials: I, Zirconia -8 Mol.% Yttria,” J.Am.Ceram. Soc.. 75 , pp. 341-46 (1992). Johnk, C.A., Engineering Electromagnetic Fields and Waves. John W iley & Sons, New York, NY (1976). 89 K atay am a, K ., W atanabe, T., Matoba, M., Katoh, N ."D evelopm ent o f Brazing Technology for Ceramic Turbocharger Rotors" (1986). SAE Tech. Pap. Ser. 861128. K in g e ry , W.D., Bowen, H.K. and Uhlman, D. R .. Introduction to Ceram ics. 2nd edition, John Wiley & Sons, New York, NY (1976). L o e h m a n , R.E.,"Joining Engineering Ceramics," pp. 83-98 in Advances in Joining Newer Structural Materials. Symposium Processing of the International Conference held in Montreal, Canada, 23-25 July (1990). L u b in , G ., Handbook o f Composites. Van Nostrand Reinhold Company, New York (1982). M a rsh a ll, C. W. and Rudnik, A., Fracture Mechanics of Ceramics, pp. 69-92,Plenum Press, NY (1974). M ck in ley , T.P., Ed., The Electron Microprobe. John Wiley & Sons (1966). M cM illan, P.W., Glass-Ceramics. Academic Press, London (1964). M eek, T. T., Blake, R. D., "Ceramic-ceramic Seals by Microwave Heating," pp. 270274, J. Mater. Sci. Lett. 5, 270 (1986) M etax as, A. C., M eredith, R. J., Industrial Microwave Heating. Peter Peregrinus, London (1983). M izu h ara, H., Huebel, E. and Oyama, T., “High-Reliability Joining of Ceramic to Metal” Am. Ceram. Soc. Bull. . Vol. 68(9 ), 1989. M oore, T. J., "Feasibility Study of the Welding of SiC," J. Am. Ceram. Soc.. 6 8 ( 6 ) C151-153 (1985). N e w n h a m , R. E., Jang, S. J., Xu, M .and Jones, F., “Fundamental Interaction Mechanisms between Microwaves and Matter,” Ceramic Transactions. Vol. 21, pp 51-67 (1991). P a la ith , D., Silberglitt, R., Wu, C. C., Kleiner, R. and Libelo, E., L., "Microwave joining of ceramics," pp. 255-66 in Microwave Processing of Materials. Symposium Processing, Vol. 124. Edited by W. H. Sutton, M. H. Brooks, and I. J. Chabinsky. Materials Research Society, Pittsburgh, PA (1988). P a la ith , D., Silberglitt, R., \b l. 68(9) (1989). "Microwave joining of ceramics," Am. Ceram. Soc. B ull.. R eim er, J., Scanning Electron Microscopy. Springer Verlag (1972). R ice, R. W., “Joining of Ceramics,” in NRL-7085, Naval Research Laboratory, 23, 1970. 90 Schwartz, M. M., Ceramic Joining. ASM INTERNATIONAL, Materials Park, Ohio (1990). Schwartz, M. M., Handbook of Structural Ceramics. McGraw-Hill, Inc., New York (1992). Srinivasan, M .,”The silicon C arbide Fam ily o f S tru ctu ral C eram ics,” in Watchman,J.B.Jr.,(Ed.) Structural ceramics, vol. 29 o f Treatise on Materials Science and Technology. Academic Press, pp. 99-159(1989). Sutton, W. H., “Microwave Processing of Ceramic Materials,” Am. Ceram. Soc. B ull.. \b l. 6 8 , No.2, pp. 376-86 (1989). Touloukian, v. S ., Therm ophvsical Properties of H igh Temperatures Solid Materials.MacMillan. New York, N Y (1967-1992). Von H ippel, A.R., Dielectrics and Waves. John Wiley & Sons, New York (1954). Varadan, V.K., Varadan, V.V., “Microwave Joining and Repair of Composite Materials,” Polymer Engineering and Science. Mid-April. Vol.31. No.7 (1991a). Varadan, V.V., Hollinger, R.D., Ghodgaonkar, D.K., and Varadan, V.K., “ Free-Space, Broadband M easurem ents of High-Temperature, Complex Dielectric Properties at Microwave Frequencies.” IEEE Trans. Instrumen. Meas. Vol. 40, No. 5 (1991b). Yajima, S., Okamura, K., Shishido, T., Hasegawa, Y., Matsuzawa, T.,"Joining of SiC to SiC using Polyborosiloxane," pp. 253, J. Am. Ceram. Soc.. Vol. 60, No 2(1981) Yiin, T. Y„ Varadan, V.V., Varadan, V. K. and Conway, J. C., “ Microwave Joining of Si-SiC/Al/Si-SiC." Ceramic Transactions. Vol. 21, pp 507-514 (1991). Wolfenden, A., Harmouche, M. R., “Dynamic Young’s Modulus Measurements in Metallic Materials: Results of an Interlaboratory Testing Program.” Journal of Testing and Evaluation. JTEVA. Vol. 17, No. 1, Jan. (1989), pp. 2-13. Yu, X. D., Varadan, V. V. and Varadan, V. K., “Application of Microwave Processing to Simultaneous Sintering and Joining of Ceramics,” Ceramic Transactions. Vol. 21, pp 497506 (1991). Zdaniewski, W.A., Conway, J.C., and Kirchner, H.P., “Effect of Joint Thickness and Residual Stresses on the Properties of Ceramic Adhesive Joints,” J. Am. Ceram. Soc. , 70, pp. 104-118 (1987). 91 Appendix S P E C IF IC A T IO N S F O R E N G IN E E R IN G C E R A M IC S U SIN G B A RIU M ST R O N T IU M TITA N ATE (BST) AS AN E X A M PL E Knowledge of accurate values for the properties of ceramics becomes indispensable for their utilization in engineering design. The designer often needs to know or be warned that the quoted values are valid over specified ranges o f temperature, pressure, mechanical stress (including cycling), and chemical environment. Mechanical and thermal properties of Barium Strontium Titanate (BST) were evaluated by different test methods as a specific example. We tried to apply or develop reliable tests which sim ulate the working environmental conditions of the actual component where strength predictions are to be made. This can be an almost impossible task since com ponents used in industrial applications often experience very complex states which cannot be accurately simulated in laboratory experiments. Therefore, American Standard Test M ethods (ASTM ) are performed to provide a data base by which to assess the strength of Barium Strontium Titanate under known conditions. The coverage of the test methods and results in this chapter serves to give some insight into the determination of property information in order to provide comment on the quality of Barium Strontium Titanate (BST). A .l Grain Size fASTM El 121 The sintered ceramic specimen surface was first ground to 50 pm surface roughness and followed by automatic polishing down to 0.5 |im surface roughness using SiC sand papers or diamond pastes. After the polishing treatments, the specimen was flat, smooth, and free from physical imperfections. In order to fully reveal the microstructure of the specimen, a further treatment of thermal etching was required. The bonding of atoms, 92 ions, or molecules is weaker at exposed positions such as grain boundaries and imperfections; therefore etching first initiates at these locations yielding a delineation of the microstructure. Thermal etching was performed at 50°C below the ceramic sintering temperature for 10 minutes. After cooling down to room temperature, ceramic specimen was ultrasonically cleaned in acetone and gold coated to prevent charging during the surface examination. The gold coating was done by sputtering in a low vacuum chamber. The SEM examination has a Robinson backscatter detector and can imagine the surface up to |am range. Figure A .l shows the typical SEM picture of BST which has grain size around 7.4 [im. Figure A .l Scanning Electron Micrograph showing the grain size of BST. 93 A.2 Dynamic Elastic Modulus Measurements fASTM-C7691 A.2.1 Objective and Background The objective is to measure the Young’s elastic modulus (E), shear modulus (|i), Bulk modulus (B) and Poisson’s ratio (v) of ceramic using the ultrasonic method [ASTMC769]. In engineering and science the elastic modulus is of fundamental and technological importance. It has applications in areas such as load-deflection, buckling, thermoelastic stresses, fracture mechanics, creep, interatomic potentials, lattice defects, thermodynamic equations o f state, free energy and thermal expansion. When a stress (force/aiea) is applied to a material, it gives rise to a strain (change in volume or shape). The strain is actually caused by stretching of the atomic bonds in the solid and in some materials the atoms may move relative to each other. For solids, there is usually some range of stresses over which the deformation of the materials is elastic, i.e. the deformation is reversible. In a linear elastic material the stress (a) is proportional to the strain ( e), g =E e, where E is the elastic constant known as Young’s modulus. The properties of a material under applied shear and pressure are described by the shear modulus (|i) and the bulk modulus (B). The Poisson’s ratio (v) is defined as the negative ratio of the transverse strain to axial strain [Kingery et al., 1976]. For an isotropic material, we only need two elastic constants to describe their behavior. The magnitude of the elastic constants depends on the strength and density of the atomic bonds in a material. A.2.2 Velocity of Ultrasonic Wave Pulses For ultrasonic wavelengths less than the dimensions of the specimens, two normal modes of wave propagation in isotropic media prevail. They are the longitudinal and shear 94 modes, with respective velocities and Vs. Longitudinal waves, sometimes referred to as com pressional waves, alternately compress and dilate the material lattice (generate compressive and tensile strains) as they pass by. The resulting particle motion o f the material is parallel to the direction of wave propagation. Shear waves, on the other hand, generate particle displacements perpendicular to the propagation direction, causing the material lattice to shear as the waves pass by. The wave speed V in the specimen is determined by measuring the transit time t of an ultrasonic pulse over a known path L in the specimen, and by calculating V=L/t. The ultrasonic velocity measurements are made at a frequency of 5 MHz for longitudinal and 10 MHz for transverse waves, with higher frequency being used for shorter path length. All measurements are made at a room tem perature of 21°C [W olfenden et al., 1989], Figure A.2 illustrates the principal components used for a majority of the measurements. A pulser/receiver unit transmits a very short spike voltage to a traducer, generating a broad-band ultrasonic wave of short duration. The transducer converts the excitation into a mechanical oscillation or sound wave which is coupled into the specimen to propagate at the sound velocity. The coupling is significantly enhanced by using a thin layer (much less than an ultrasonic wavelength) of liquid. Commercially available transducers have a casing that houses a piezoelectric element and often an impedance matching circuit, which are designed for convenient electrical attachment to the pulser/receiver via coaxial cable and standard connectors. After entering the specim en, the ultrasonic pulse echoes back and forth between the faces o f the specimens, while constantly decaying in amplitude due to scattering, absorption, and boundary interface losses. Each time the wave pulse is incident at the traducer/specimen interface, a portion of the elastic wave energy is converted into a electrical signal by the transducer. The received signal is then amplified and displayed on an oscilloscope so that the transit time measurements can be made. Electronic techniques exist to automate the measurements with direct computer control. The accuracy of these measurements depends on the dimensions of the specimen (path length, end-face parallelism, etc.), the particular ultrasonic coupling technique, and the signal-to-noise ratio. Typically, the accuracy for these transit time measurements is ± 0.1 % or better. The densities of the specimens for this 95 study were determined by the Archimedes method, by which the density of the specimen is measured relative to that of distilled water. From these two wave speeds and the density p, all the elastic parameters of the material can be calculated: Young's modulus = p Vs2 (3V ^ 2 - 4 Vs2) / ( - Vs2 ) (A. 1) Bulk modulus = p ( V L2 -(4/3) Vs2) (A.2) Shear modulus= p Vs2 (A.3) Poisson'sratio= (V L 2 -2V S2 ) / ( 2V L 2 -2V S2 ) (A.4) where = Longitudinal wave speed in the specimen and Vs = Shear wave speed in the specimen. The elastic constants of seven BST specimens were determined as shown in Table A .l. PULSER/RECEVER SCOPE TRANSIT TIME TESTING SAMPLE Figure A.2 Ultrasonic velocity measurement system with direct contact of transducer on specimen. Table A .l Measured Elastic Constants of 7 BST Specimens and the Average. Density vs Vl Young Bulk Shear (g/cm3) (m/s) (m/s) (GPa) (GPa) (GPa) 5.14 4342 7076 232.14 128.13 96.89 0 .2 0 5.20 4796 7534 277.06 135.55 119.49 0.16 5.01 3808 6527 180.57 116.64 72.69 0.24 5.08 3966 6599 194.59 114.76 79.92 0 .2 2 5.10 3947 6818 198.27 131.11 79.44 0.25 5.13 3769 6529 182.25 121.59 72.89 0.25 5.16 3700 6068 170.11 95.81 70.64 0 .2 0 Poisson BST Elastic Constant Averages With Standard Deviation Density gm/cm3 5.12±0.06 Elastic modulus GPa 185.2±37.4 Shear modulus GPa 75.1±17.8 Bulk modulus GPa 116.0±17.8 Poission’s ratio 0.23±0.03 97 A.3 Water Absorption. Bulk Density. Apparent Porosity and Apparent Specific Gravity of Ceramic TASTM-C201 A.3.1 Significance Measurement of density, porosity, and specific gravity is a tool for determining the degree of maturation of ceramic. These properties are widely used in the evaluation and comparison of product quality and as part of the criteria for selection, or for determining structural properties that may be required for a given application. The test method is a standard method suitable for use in quality control, research and development, establishing criteria for and evaluating compliance with specifications, and providing data for design purposes. A.3,2 Procedure and Results First of all, 10 ceramic specimens are dried to constant weight by heating them to 105~110°C in a oven, followed by cooling in a desiccator. Then the dry mass (D) is determined with a weight scale balance. The next step is to place the specimens in a pan of distilled water and boil for 5 hours, making sure the specimens are covered with water at all times and not in contact with the heated bottom of the container. After the boiling period, the test specimens are cooled to room temperature while still completely covered with water for a minimum o f 24 hours. For determining the suspended weight (S), the specimens are placed in a wire loop that is suspended from one arm o f the balance. The balance is previously counter-balanced with the wire in place and immersed in water to the same depth as is used when the specimens are in place. After determining the suspended weight, each specimen is lightly blotted with a cotton cloths to remove all drops of water from the surface and then the saturated weight (W) is determined. The average results of water absorption, bulk density, apparent porosity and apparent specific gravity of BST were calculated for ten BST specimens based on the 98 definitions and the formula as shown in Table A.2. Table A.2 Water Absorption, Bulk Density, Apparent Porosity and Apparent Specific Gravity of BST Averaged for 10 Specimens. Name Unit Formula Result Dry Weight (D) g D 0.7455 Suspended Weight (S) g S 0.6083 Saturated Weight (W) g W 0.7469 Exterior Volume (V) cm ^ w-s 0.14 \folume of open pores cm-* W-D 1.4x10-3 \blum e of impervious portion cm-* D-S 0.14 Apparent Porosity (P) % (W-D)xl00/V 1 .0 1 ± 0 .0 2 Water Absorption (A) % (W-D)x 100/D 0.18±0.01 Bulk density (B) g/cm^ D/V 5.38±0.06 D / (D-S) 5.43±0.05 Apparent Specific Gravity CD Theoretical Density g/cm^ 5.6 99 A,4 Hardness Test of Ceramic TASTM-E18& ASTM-E1401 A.4.1 W hat is Hardness ? Hardness is defined in the conventional sense as a means o f specifying the resistance of a material to deformation, scratching, and erosion. Hardness tests are based on indenting the sample with a hard indenter, which may be spherical, conical, or pyramidal. Common techniques for measuring hardness in ceramics are known as Vickers (HV), Rockwell superficial (HR), and Knoop (HK). A A 2 Rockwell Hardness A) Rockwell Hardness Test: An Indentation hardness test using a verified machine to force a hard steel ball indenter under specified conditions into the surface of the material under test in two operations, and to measure the difference in depth of the indentation under the specified conditions of preliminary and total test forces (minor and major loads, respectively). B) Rockwell Hardness Number, HR: A number derived from the net increase in the depth of indentation as the force on an indenter is increased from a specified preliminary test force to a specified total test force and then returned to the preliminary test force. For example 64 HRC = Rockwell Hardness Number of 64 on Rockwell C Scale. 81 HR30N = Rockwell Superficial Hardness Number of 81 on Rockwell 30N Scale. A.4.3 General Description and Test Procedure for Rockwell Hardness Figure A. 3 shows how the Rockwell Hardness was investigated and the steps to make the indentation. Po Po + PI Po I t j : 300 ::0 A B i ± T : iQQ. 1 Scale Scale A=Depth of penetration under test force before application of additional load. B=Increase in depth of penetration under additional load. C=Permanent increase in depth of penetration under preliminary test force after removal of additional force. HR=Rockwell hardness number = 100 - C. Figure A. 3 Schematic diagram of Rockwell Hardness and test steps. 101 A.4.4 Remarks of the Rockwell Hardness Test • There is no Rockwell hardness value designated by a number alone, because it is necessary to indicate which indenter and force have been employed in making the test • The difference in depth is normally measured by an electronic device or by a dial indicator, • The hardness value, the higher the number, the harder the material. A.4.5 Vickers Hardness Number (HV1 An indentation hardness test which forces a square-based pyramidal diamond indenter having specified face angles (136°) into the surface of the material under test and the the diagonals of the resulting impression after removal of the load are measured as shown in Figure A.4 The V ickers Hardness number is obtained by dividing the applied load in kilograms-force by the surface area of the indentation in square millimetres computed from the mean o f the measured diagonals of the indentation. Formula used to compute Vickers hardness are given below: HV = P / As = 2 P sin (a/2) / d 2 = 1.8544 P / d 2 (A.5) where P is the applied force in Kgf, A s is the surface area of indentation in mm., a is the face angle of indenter (136°), and d is the mean diagonal length in mm. 102 I I—(?) <x=136 0 between opposite faces Testing specimen Number Symbol 1 a Angle at the vertex of the pyramidal indenter (136°) 2 P Test load in kilograms-force 3 d Arithmetic mean of the two diagonals d j and d 2 Designation Figure A.4 Schematic diagram of the Vickers hardness test. 103 A.4.6 Conversion of Vickers Hardness to Rockwell Hardness Sometimes it is impossible to test the material under the conditions specified, so standard Hardness Conversion Tables (ASTM E140) can be applied to solve this problem. ASTM-E140 gives the relationships between Vickers Hardness, Rockwell Hardness, Rockwell Superficial Hardness and Knoop Hardness; many of those conversion values were obtained from computer-generated curves o f actual test data, ceramic was first indented by Vickers Hardness set-up and the results are shown in Table A.3. By using the conversion table, the Vickers Hardness can be converted to the Rockwell Hardness as we want and is shown in Table A.4. Table A.3 Vickers Hardness of 7 BST specimens. Specimen d(mm) P(Kgf) HV A 0.153 5 396 B 0.158 5 371 C 0.154 5 391 D 0.154 5 391 E 0.150 5 412 F 0.158 5 371 G 0.148 5 423 Average =392±19 1 04 Table A.4 Vickers Hardness Conversion to the Rockwell and Rockwell Superficial Hardness. Vickers 392 Rockwell Superficial Rockwell A scale D scale C scale (60Kgf) (lOOKgf) (90Kgf) 15N 30N 45N 70.4 80.4 59.5 43.1 55.4 40 A.5 Flexural Strength Test of Ceramic 1ASTM-F3941. A.5,1 Objective The objective here is to determine the flexural strength and fracture toughness of ceramic using four point bending and biaxial flexure strength (modulus o f rupture) methods. A.5.2 Description In bend tests, the stress is not uniform but one considers the strength in terms of the maximum tensile stress being applied. This maximum stress occurs in some region along one o f the outer surfaces of the specimen. In this way, fracture is generally biased to flaws at or near the outer surface. Up to now there are no national or international standards specifying loading conditions, shapes and dimensions of the specimens, but symmetrical specimens o f rectangular cross sections are usually preferred as shown in Figure A.5. By using elastic beam theory, the flexural strength of ceramic specimens of rectangular cross section can be determined from the fracture load F and the specimen dimensions from the equation [Beer & Johnston, 1981]: 105 o bf = 3FD/B2W (A.6 ) This formula used often assumes the material to be elastic (the load-displacement behavior needs to be linear). F B _L Figure A.5 Schematic diagram of the four point test The strength of the ceramic was also measured by another method, biaxial flexure method, using a mechanical testing machine. In this test, the center of the ceramic is m arked as the intersection of three lines drawn from diametrically opposite points approximately 60° to each other, ceramic pellets were placed on three small steel balls which form an equilateral triangle on the circular bottom fixture as shown in Figure A.6 . The circle that circumscribed these three steel balls has a diameter of about 1 inch. A steel hemispherical ball was placed with its flat surface on the top of the other fixture, and through a piston on the other side of the fixture, the sample is fractured into two or three parts in the biaxial flexure mode. The biaxial flexure strength is calculated using the equations given below: a f = -3P / 4 P d 2 (X-Y) (A.7) 106 X = (1+u) ln(B/C )2 + [(1-tj)/2] (B/C )2 <A -8) Y=(l+t))[l+ln(A/C)2] + (1 -d) (A/C )2 <A '9> where d is the Poisson's ratio, B is the radius o f the loaded area, A is the radius of the support circle, C is the radius o f the sample, P is the applied load at failure, and d is the thickness of the sample at the point of failure. P Steel balls BST 2B » i |4 _ L BST T Figure A .6 Schematic diagram of biaxial flexure test. 1 07 Fracture toughness is a measure of the energy consumed during differential extension of flaws or cracks determined at the critical stress. This is expressed by different numbers depending on the test method of loading a crack tip. K jq is the m ost common number for fracture toughness and it represents the critical fracture toughness in mode I loading, which is tensile loading. Fracture toughness can be regarded as a material constant. It is a function of the stress and the critical flaw size [Marshall et al., 1974]. In this study, fracture toughness was measured by fractography observation. The four point bend fracture surface was examined with optical stereomicroscope to find the surface failure origin length. Fracture toughness then can be calculated using the formula KIC = Y o Y V c (A. 10) where Y is the surface crack factor (=1.24), trY is the fracture strength from four point bend test, and C is the surface failure origin length. The strength distribution of ceramics can often have a large standard deviation and thus it is often important to know the form of this distribution. An empirical distribution that has become very popular is the two parameter Weibull distribution and is based on a specimen at its weakest link. This distribution can be written in the form ln(l/(l-F))=K V ( <yf/a 0 )m (A. 11) where F is the failure probability, K is a factor that depends on the loading geometry in the strength test and CTq and m are constants that determine the magnitude and width of the strength distribution respectively. The parameter m is generally called the Weibull modulus and <Jq is a normalizing stress. The width of the strength distribution decreases as m increases. As we see from the equation, the strength of a specimen failing at a particular 108 failure probability depends on the loading and the specimen size. The equation assumes failure can occur from anywhere in the body (volumetric flaws). If failure occurs only from surface flaws, V in equation is replaced by A, an area term. A.5.3 Results The two parameter Weibull modulus plot was used for statistically analyzing the four-point bending fracture strengths of ceram ic. Results from the two parameter Weibull statistical approach are shown in Figure A.7 and Table A.5. The fracture toughness of BST is 1.28 MPaVm. Fracture Strength (MPa) 100 120 2 1 0 £i -1 a □ Q -4 4.0 4.2 4.4 4.6 4.8 5.0 Ln (Failure Stress) MPa Figure A.7 Weibull approach for the four point bending test of BST. 109 Table A.5 Weibull Approach of BST Average strength (MPa) Biaxial flexure Average Standard strength (MPa) deviation Four point bend Weibull modulus (m) Characteristic strength 122 105 33 10 8.5 W A .6 Thermal Shock Resistance 1ASTM-C5541. A.6.1 Objective The purpose of this test is to determine the thermal shock resistance parameter, R, of ceramic to the sudden exposure to extreme changes in temperature. These conditions may be encountered in equipment operated intermittently in low temperature areas. Effects of thermal shock include cracking and delamination of substrates or wafers, opening of terminal seals and case seams, and changes in electrical characteristics due to moisture effects or to mechanical displacement of conductors or insulating materials. A.6.2 Background When ceramic materials are subjected to a rapid change in temperature, there is often a variation in temperature through the body. The temperature variation leads to thermal stresses, as one part of the body 'wants' to be a different size and shape than some other part and substantial stresses develop in the material. Resistance to weakening or fracture under these conditions is called thermal shock resistance. Generally, ceramics have relatively poor thermal shock resistance due to their low strength, high thermal expansion and low thermal conductivities. Nonetheless, ceramics 1 10 are often used under conditions where their thermal shock resistance is critical. An analysis of the stresses caused by thermal gradients yields the relation [Kingery e ta l., 1976]: a = E a (Ta-T)/(l-v) (A. 12) where o = stress at the surface, E = elastic modulus, a = thermal expansion coefficient, v = Poisson's ratio, Ta = 'average' temperature of the sample, and T = surface temperature. For a water quench, the rate of surface heat transfer can be fast enough that Ta remains at approximately the initial sample temperature and T at the surface is the temperature o f the bath. As the AT of the quench increases, eventually o will reach Gf and the sample fails, thus defining a critical ATC, the maximum temperature difference in the body under steady-state heat flow conditions or the maximum allowable temperature difference the body can be subjected to in a convective environment (applicable under severe heat transfer conditions). Rearranging the equation gives g = E a AT/(l-v) (A. 13) which for the fracture condition becomes ATC = Gf (l-v)/E a (A. 14) The right hand side is often denoted by R, which is a thermal shock resistance parameter. It is clear that good thermal shock resistance to avoid fracture initiation is associated with high values of fracture stress and thermal conductivity, and low values of 111 Young’s modulus of elasticity and the coefficient of thermal expansion. A.6.3 Procedure and Results Temperature-controlled baths containing water were used for the temperature treatment. Ten ceramic specimens were first immersed in bath A at high temperatures as shown in Figure A .8 for a minimum 5 minutes. Immediately after the preconditioning time, the specimens were transferred to B which had water kept at the temperature of 0°C. The specimens remained at the low temperature for a minimum of 5 minutes and then were transferred to C which had the same condition as A for at least of 5 minutes; this was so called one cycle. The transfer time from high temperature to low temperature (A to B) and from low temperature to high temperature (B to C) was less than 10 seconds. If there were no specimens broken after 15 complete cycles, the high temperature bath was increased 5°C. If 2 specimens broke during the specific temperature cycle, the temperature difference in this cycle was considered as the thermal shock resistance parameter (ATC). The results are shown in Table A.6 . The ATC of BST is equal to 95 °C. B Oven Thermometer Test specimen • for one cycle: A (5 mins) => B (5 mins) =* C (5 mins) • A and C have the same condition. • B was kept at 0 °C. Figure A .8 Schematic diagram of apparatus for thermal shock resistance parameter. 112 Table A .6 Results of BST Subjected to Thermal Shock Water Quenching. A T(°Q TEST SPECIMENS FRACTURE SPECIMENS 70 10 0 75 10 0 80 10 0 85 10 0 90 10 0 95 10 2 ATC = 95 °C A.7 Thermal Conductivity Measurement of Ceramic 1ASTM-C4Q81. A.7.1 Objective and Background The objective is to determine the temperature dependence of thermal conductivity of ceramic. Thermal conductivity is defined as the proportionality constant k, which relates the amount of heat (dQ) flowing during a time interval dt across an area A under a steady state thermal gradient dT/dx [Kingery et ah,1976]. dQ/dt = -kA (dT/dx) (A. 15) However, many ceramics have unsteady-state thermal conditions (heating and cooling) in actual application. Under these conditions, temperatures at any point within the material change with time and can be described by 113 dT/dt = D (d2 T/dx2) (A. 16) where t = time, T = temperature, x = distance and D = thermal diffusivity. Thermal diffusivity (D) is related to thermal conductivity (k) as follows: D = k/pCp (A. 17) where p = density and C p = specific heat. Therefore, thermal conductivity can be indirectly determined from thermal diffusivity measured under transient heating conditions. A.7.2 Procedure and Apparatus fHimsworth. 19571 A long bar of ceramic was placed in contact with in an alumina substrate. The apparatus consists of a one-revolution per twenty minutes motor, a microswitch, a relay, and a hot-plate. The motor turns a cam which is cut to open or close the microswitch every ten minutes; this microswitch in turn controls the relay which turns the hot-plate on and off. The alternate heating and cooling of the hot-plate generates a thermal wave in the ceramic which causes the temperature at any given point in the material to vaiy with time as shown in Figure A.9. TEMPERATURE READOUT #2 TIMER COMPUTER Figure A.9 Schematic diagram for thermal diffusivity measurement. 114 The temperature was measured by thermocouples at two suitable distances on the ceramic and read every 100 seconds. It was found that the temperature curves vary periodically as shown in Figure A. 10, giving an approximate sine curve. The curve for the point more remote from the heated end has a smaller amplitude than the other. From the amplitude and phase shift of the two curves the thermal diffusivity can be calculated by Fourier analysis [Billington, 1949]: D=L2/2 [I/LNCAj A ^)] [1/time lag] (A. 18) where D is the thermal diffusivity of ceram ic, L is the distance between thermocouples, A j and A 2 are the amplitudes of first and second curves, time lag is the time by which second curve lags behind first. 200 190 * 160 - Point #1 on B.S.T. Point #2 on B.S.T. ODD ■■ 150 ' 140 130 0 5 10 15 20 25 30 Time interval (minutes) Figure A. 10 Temperature curves measured on BST for thermal conductivity measurement. 115 A.7.3 Results The density was measured by Archimedes principle. Specific heat of ceramic was obtained from Touloukian . Table A.7 shows the results of thermal diffusivity and thermal conductivity measurement of ten BST specimens. Table A.7 Thermal Properties Results of BST. Property Units Results Density gm/cm^ 5 .4 ± 0 .1 Specific heat Ws/g°K 0.512 Thermal diffusivity m^/s (x 10 *^) 75.3±15 Thermal conductivity cal-cm/cm^-°C-sec 0.075±0.005 A .8 Thermal Expansion TASTM-C3721. A.8.1 Objective and Background The objective is to measure the thermal expansion coefficient of ceramic with Harrop dilatometer analyzer. Ceramic undergoes a reversible change in dimensions when heated. The magnitude and temperature range over which such changes occur determine the thermal expansion coefficient (a ). Generally, materials increase in volume with increased amplitude about their mean lattice positions. One method for representing thermal expansion is via a linear thermal expansion coefficient (aj) which relates the instantaneous change in dimension to the temperature 116 interval: otj = dl /I dT (A. 19) where 1 is a linear dimension. It is important to recognize that the thermal expansion coefficient can vary markedly over a given application temperature range. Many ceramics exhibit more than one phase, and the difference in the thermal expansion behavior between phase transitions must be accounted for to understand the thermal expansion behavior. Several other features that can affect the thermal expansion behavior are the presence of microstresses. One important consequence of microstresses is the formation of microcracks in the body, which in turn can result in significant hysteresis in the thermal expansion behavior during a heating/cooling cycle [Hayden et al., 1965]. A.8,2 Procedure and Results Thermal expansion was measured with Harrop dilatometer analyzer at a heating rate o f 3°C/min. over the temperature range of 1150°C. A linearly variable differential transformer and data acquisition processor were used to record the incremental length change as a function of temperature as shown in Figure A. 11. Due to the BST phase transition around 1000°C and probably microstresses effects, the thermal expansion coefficient of heating is 7.65 ppm/°C, while the thermal expansion coefficient o f cooling is 10.37 ppm/°C. BST 1.5 p h /^ < ^ Percent Tine «= 7. 653p €05 (nin) Expansion _ s=10JS7ppli/*C -1 0 250 500 250 1000 1250 Tenperature (°C) Figure A .l 1 Thermal expansion behavior o f BST as function o f temperature. 1500 118 A.9 Summary Table A.8 is the summary of the mechanical and thermal property testing results of the BST. Table A.8 Mechanical and Thermal Properties of BST. PROPERTY UNITS TEST METHODS RESULTS GRAIN SIZE pm SEM 7.4 THEORETICAL DENSITY g/cm3 BULK DENSITY g/cm3 5.6 ASTM-C623 5.4±0.06 ELASTIC MODULUS GPa ASTM-C623 185±37.4 SHEAR MODULUS GPa ASTM-C623 75±17.8 BULK MODULUS GPa ASTM-C623 116± 18 POIS SON’S RATIO ASTM-C623 0.23±0.03 WATER ABSORPTION ASTM-C20 0.18±0.01 POROSITY ASTM-C20 1.01 ±0.02 119 PROPERTY UNITS SPECIFIC GRAVITY TEST METHODS RESULTS ASTM-C20 5.4310.05 HARDNESS Rockwell (15N scale) ASTM-E18 ASTM-E140 80.4±4 FLEXURAL STRENGTH MPa (21 °C) ASTM-F394 122±8.5 FRACTURE TOUGHNESS MPaVm ASTM-F394 1.2810.05 THERMAL SHOCK *C ASTM-C554 92 THERMAL CONDUCTIVITY cal-cm/cm2-°C-Sec ASTM-C408 0.07510.005 SPECIFIC HEAT Ws/g°K THERMAL DIFFUSIVITY m2/s (x 10'8) ASTM-C408 75.33115 EXPANSION COEFFICIENT ppm/°C ASTM-C372 7.6 0.512 VITA Tzu-Yuan Yiin was bom in Kaohsiung, Taiwan, R.O.C., on November 3, 1961. He received his B.S. in Naval Architecture Engineering from the National Taiwan University in June 1985. After graduation, he served in the Chinese Navy for two years. In the fall of 1988 he entered the Department of Engineering Science and Mechanics at The Pennsylvania State University for Ph.D. study. In January 1989, he joined the Center for the Electronic and Acoustic Materials as a research assistant. He is a member of American Ceramic Society and American Welding Society.