ANALYTICAL APPLICATIONS OF GC-MES IN THE ULTRAVIOLET-VISIBLE AND VUV REGIONS OF THE SPECTRUM (VACUUM-ULTRAVIOLET, GAS, CHROMATOGRAPHY, MICROWAVE, EMISSION)код для вставкиСкачать
INFORMATION TO USERS This reproduction was made from a copy of a docum ent sent to us for microfilming. While the most advanced technology has been used to photograph and reproduce this docum ent, the quality of the reproduction is heavily dependent upon the quality o f the material submitted. The following explanation o f techniques is provided to help clarify markings or notations which may appear on this reproduction. 1.The sign or “ target” for pages apparently lacking from the document photographed is “Missing Page(s)” . If it was possible to obtain the missing page(s) or section, they are spliced into the film along w ith adjacent pages. This may have necessitated cutting through an image and duplicating adjacent pages to assure complete continuity. 2. When an image on the film is obliterated with a round black mark, it is an indication of either blurred copy because o f movement during exposure, duplicate copy, or copyrighted materials that should n o t have been filmed. For blurred pages, a good image o f the page can be found in the adjacent frame. If copyrighted materials were deleted, a target note will appear listing the pages in the adjacent frame. 3. When a map, drawing or chart, etc., is part of the material being photographed, a definite m ethod o f “sectioning” the material has been followed. It is customary to begin filming at the upper left hand com er o f a large sheet and to continue from left to right in equal sections with small overlaps. If necessary, sectioning is continued again—beginning below the first row and continuing on until complete. 4. For illustrations that cannot be satisfactorily reproduced by xerographic means, photographic prints can be purchased at additional cost and inserted into your xerographic copy. These prints are available upon request from the Dissertations Customer Services Departm ent. 5. Some pages in any docum ent may have indistinct print. In all cases the best available copy has been filmed. University Micrtirilms International 300 N. Zeeb Road Ann Arbor, Ml 48106 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 8229959 Pierson,Duane Aron ANALYTICAL APPLICATIONS OF GC-MES IN THE UV-VISIBLE AND VUV REGIONS OF THE SPECTRUM PH.D. 1982 The University o f Iowa University Microfilms International 300 N. Zeeb Road, Ann Arbor, M I 48106 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. PLEASE NOTE: In all cases this material has been filmed in the best possible way from the available copy. Problems encountered with this document have been identified here with a check mark V 1. Glossy photographs or pages. 2. Colored illustrations, paper or print_____ 3. Photographs with dark background_____ 4. Illustrations are poor copy______ 5. Pages with black marks, not original copy. 6. Print shows through as there is text on both sides of page. 7. Indistinct, broken or smali print on several pages 8. Print exceeds margin requirements_____ 9. Tightly bound copy with print lost in spine______ 10. Computer printout pages with indistinct print. 11. Page(s) author. 12. Page(s) _ _ _ _ _ _ seem to be missing in numbering only as text follows. 13. Two pages numbered____________. Text follows. 14. Curling and wrinkled pages______ 15. Other____________________ ____________ ____________ lacking when material received, and not available from school or ___________________ University Microfilms International Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ANALYTICAL APPLICATIONS OF GC-MES IN THE UV-VISIBLE AND VUV REGIONS OF THE SPECTRUM by Duane Aron Pierson A thesis submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Chemistry in the Graduate College of The University of Iowa July, 1982 Thesis supervisor: Professor Clyde W. Frank Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Graduate College The University of Iowa Iowa City, Iowa CERTIFICATE OF APPROVAL PH.D. THESIS This is to certify that the Ph.D. thesis of Duane Aron Pierson has been approved by the Examining Committee for the thesis requirement for the Doctor of Philosophy degree in Chemistry at the July, 1982 graduation. Thesis committee:_ Thesis supervisor M JLts Member __ I., Member Memb Member Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ACKNOWLEDGEMENTS This dissertation is the culmination of many long years of hard work. I wish to thank my parents, my family and my friends for their support and encouragement. I would especially like to thank my wife Dee, for her love and support when I needed it most. As my advisor, Dr. Clyde Erank provided the assistance and the opportunity to grow as an analytical chemist. The type of training I have received will be of great value in making the transition from academics to industry. Finally, I would like to thank Dr. H. Bruce Friedrich for assisting me in preparing this document by computer. ii Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. TABLE OF CONTENTS Page LIST OF T A B L E S ............ . . .. ................ . . . . vi LIST OF F I G U R E S .......... vii LIST OF ABBRE V I A T I O N S ................. . . . . . . . . . ix PART A. ANALYTICAL APPLICATIONS OF GC-MES IN THE UVVISIBLE ..................... . . . . . . . . . 1 CHAPTER I.. INTRODUCTION 2 ................... 2 Background . . . . . . . . Microwave Induced Plasmas ................... 3 System Components ........ 3 Discharge Properties ..................... 8 Sample I n t r o d u c t i o n .......... . . . . . . 12 Gas Chromatography-Microwave Emission Spectroscopy . .......... 16 Method Development ................. . . 17 Applications .............................. 25 S u m m a r y .............. 28 Element Selective Detectors for GC . . . . 28 Conclusions ........ .. . . . . . . . . . 3 1 Purpose of S t u d y ......................... 31 Background ......................... 31 Lactose In t o l e r a n c e ............... .. . . . 33 Hydrogen Breath Test ........ .. . . . . 35 Previous Application of Breath Tests . .. 3 7 Determination of Hydrogen in Breath . . . . 38 II. EXPERIMENTAL . ....................... Introduction . . . . ............... . . . . Gas C h r o m a t o g r a p h . Control Panel . . . . . . . . . . . . . . . Injection P o r t ....................... . 4 7 Exit Port ii.i Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 42 42 43 43 50 Page Column . . . . . . . . . ........ . . . . Detection System . . . . . . . . . ........... Plasma Capillary . . . .............. .......................... Cavity Mount Plasma Ignition ............... . . . . . . Light Pipe . . . . . . .. . . . . . . . . ................ Spectrometer Modifications Wavelength Readout ........................ Photomultiplier Tube Mount ................ Optical Adjustment . . . . . . . ......... III. RESULTS AND DISCUSSION . . . ......... . . . . . 50 53 53 53 53 56 56 61 64 65 72 Plasma Characteristics ..................... 72 Plasma Background Spectrum . . . .. . . . 72 Spectroscopic Temperature . . . . . . . . . 7 6 Detection of Hydrogen Gas .. . . . . . . . . 7 7 S e p a r a t i o n ...................... Detector Optimization . .. . . . . . . . . 8 1 Wavelength Calibration . ................ 83 Calibration of R e s p o n s e ........ .. . . . . 88 Evaluation of Hydrogen Detection . . .. . . . 9 3 Sensitivity and Detection Limit . . . . . . 93 Selectivity . . . . . . . . .............. 97 A c c u r a c y ............ 98 Linear R a n g e .............. 99 Reproducibility . . . . . . . . 100 Hydrogen Breath T e s t ............... . . . . 105 Procedure . 105 Breath Collection ......................... 105 Collection Apparatus ................ 107 Sample Storage ....................108 Breath Test A p p l i c a t i o n s ........... ' . . . . 110 Lactose Studies . . . . . . . . . . . . . 110 Ambient Hydrogen Levels . •. . r' . . . . . . 110 Rebreathing vs. Hyperventilation . . . . 112 Lactose Malabsorption in MS Patients . . 115 Lactulose Study ............... 120 Method Validation . . . . ................. 121 C o n c l u s i o n s ...................... iv Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 80 126 Page PART B. ANALYTICAL APPLICATIONS OF GC-MES BELOW 2000 A 129 CHAPTER IV. INTRODUCTION .............. . . . . . . . . . . 130 Recent Analytical Applications . . . . . . . Purpose of S t u d y ............ . . . . . . . 131 133 V................ EXPERIMENTAL ... 134 Purging the Optical System ................. 134 D e t e c t i o n ................................. 135 VI. RESULTS AND DISCUSSION .. 137 Oxygen Removal . . . ............ Detection of Iodine Emission .......... . . . Wavelength Calibration ................. Method Development ........... Detection of Sulfur Emission . . . . . . . . Wavelength Calibration . ............... Method Development ........ Conclusions . .............................. 137 138 138 139 152 152 154 163 APPENDIX A. SPECTROMETER ALIGNMENT PROCEDURE . . . . 165 APPENDIX B. GC-MES ALIGNMENT PROCEDURE . . ........ 167 APPENDIX C. BENZENE - SPECTRUM 1000 - 2500 A . . . . 168 APPENDIX D. NITROETHANE - SPECTRUM 1000 - 2500 A .. 170 APPENDIX E. MALATHION - SPECTRUM 1000 - 1500 A . .. 172 APPENDIX F. BROMOBUTANE - SPECTRUM 1000 - 2500 A .. 174 APPENDIX G. IODOBUTANE - SPECTRUM 1000 - 2500 A .. 176 APPENDIX H. ANALYTICAL APPLICATIONS OF GC-MES . .. 178 LIST OF REFERENCES . . . . . . . . . . . . . . . . . v Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 189 LIST OF TABLES Table 1. Page Experimental Apparatus . . ........ 46 ................. 73 2. Plasma Background Spectrum 3. Operating Conditions for H2 . 4. Comparison Between Analytical Linesfor Hydrogen 5. Storage Stability for Breath Samples 6. Storage Stability vs. Temperature 7. Ambient Hydrogen L e v e l s .......... 113 8. Comparison between Rebreathing and Hyperventilation . . ................... 114 9. ... . . . . . . . . . . . .............. 82 . . .. . 95 . . 109 . Hyperventilation Collection Method . . Ill 116 10. Residual Lung Volume . . . ....................... 118 11. Reproducibility of Hydroden Determinations 12. Validation of Hydrogen Method . . . .- 13. Observed Iodine Transitions ... ................... 142 14. Separation of Organoiodides . ........... 149 15. Observed Sulfur Transitions ..................... 153 16. Sulfur Response vs. Compound Type . . . . 17. Selectivity for Sulfur vs. Slit W i d t h .......... .. . . . . . 119 .. .. ...127 . . 161 vi Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 163 LIST OF FIGURES Figure Page . . .............. 45 1. GC-MES Experimental Set-up 2. Injection Port Design . . . . . . . 3. Exit Port D e s i g n ......................... 52 4. Microwave Cavity Mount 55 5. Adjustable Light Pipe Mount ..................... 6. Light Pipe Support 7. Wavelength Readout Device 8. PMT - Exit Slit Mount . . . . . . . . . . . . . . . 67 9. Slit Width C a l i b r a t i o n .......... 69 10. Spectrometer Mount 71 11. Argon Plasma S p e c t r a ................. 12. Detector Optimization - (4861 A ) ................. 85 13. Detector Optimization - ^ (6720 A) 14. Hydrogen Calibration Peaks . . . . . . . . . . . . 90 15. Hydrogen Calibration Curve ................. 92 16. Hydrogen Linear Range . . . . . . . . . . . . . . 102 17. Reproducibility of Standard Injections 104 18. Breath Sample Chromatogram 19. Lactulose study - HBT 20. Plasma Doping Apparatus . 21. Detector Optimization - Iodine (1830 A) ......... .............. . 4 9 ....................... .58 ............... .............. . . . . .63 . . . . . . 79 87 ........ ........... .............. 60 123 ... . . ............... vii Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 125 141 144 22. Detector Optimization - Iodine (1844 A) . . . . . . 23. Separation of O r g an o i o d i d e s ..................... 24. Detector Optimization - Sulfur (1807 A) . . . . 25. Sulfur Compounds 148 151 . 156 ............................. 159 viii Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. LIST OF ABBREVIATIONS A Angstroms ARL Applied Research Laboratories b y-intercept C Centigrade cal calorie cm centimeter CMP capacitivly coupled plasma dc direct current EC electron capture EGA evolved gas analysis eV electron volt FID flame ionization detector fg femtogram (1x10 "15gram) FPD flame photometric detector g gram GC gas chromatograph h Plank's constant HBT hydrogen breath test HID helium ionization detector ICP inductively coupled plasma ix Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. i.d. inside diameter I.P-. ionization potential K Kelvin kcal kilocalorie (lxl03calorie) kg kilogram (lxl03gram) LTE local thermodynamic equilibrium LTT lactose tolerance test X wavelength m slope MDL minimum detectable limit MES microwave emission spectroscopy MIP microwave induced plasma MHz megahertz mL milliliter (1x10_3liter) mm millimeter (1x10-3meter) MS multiple sclerosis yg microgram (1x10 "'gram) yL microliter (1x10 "'liter) ng nanogram (1x10 "9gram) NPT National V frequency o.d. outside diameter Pipe Thread Ohms pg picogram (1x10 _12gram) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. PMT photomultiplier tube ppm parts per million ppt parts per thousand psig pounds per square inch (gauge pressure) QA quality assurance r correlation coefficient RIR radiative ionization - recombination model RDS relative standard deviation s second S.D. standard deviation TC thermal conductivity UV ultraviolet VUV vacuum ultraviolet V voltage v/v volume/volume W watts w/v weight/volume xi Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. PART A ANALYTICAL APPLICATIONS OF GC-MES IN THE UV-VISIBLE Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 CHAPTER I INTRODUCTION Background A plasma state is established when sufficient energy is transferred to a gas so that ions and electrons dominate the behavior of the system (1). Plasmas produced by the interaction of electromagnetic fields with inert gases such as argon and helium have been developed as spectral excitation sources since the mid-1950's (2). The inductively coupled plasma (ICP) has received an extaordinary amount of attention for the past ten years. The ICP is powered by inductive coupling with the magnetic field of a radio frequency supply. Excellent reviews on the capabilities and instrumentation associated with the ICP have been written by Fassel and Kniseley (3/4), Greenfield et al. (5) and Sytz (6). Plasmas produced by the interaction of microwave electrical fields with gases have also shown considerable potential as spectral excitation sources. exist. Two general types In the capacitively coupled microwave plasma (CMP), microwaves are conducted through a coaxial waveguide to the tip of a conductive electrode. A flamelike plasma is formed Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3 when an inert gas is forced through the center of the electrode. The second type of microwave discharges are electrodeless devices known as microwave induced plasmas (MIP). Energy is coupled to a flowing inert gas contained in a nonconductive quartz tube via an external cavity or antenna. Excellent reviews dealing with the MIP have been presented by Greenfield et al. (7), Skogerboe and Coleman (2), and more recently by Zander and Hieftje (8). A complete list of references can also be obtained in Analytical Chemistry Reviews on Emission Spectrometry (9,10,11). The following overview will deal exclusively with microwave induced plasmas. Microwave Induced Plasmas System Components Microwave Power Supplies The commercially available power units operating at 2450 MHz are commonly in use. The power supplies are normally variable between 0 and 125 watts. Availability of units that operate at other scientific frequencies with the power, stability and operational requirements for spectroscopic applications is still limited (8). West (12) compared a 2450 MHz MIP with an excitation system operated at 30 MHz and reported certain advantages for both. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4 Microwave Cavities A standing electromagnetic wave is established within the resonant cavity. For microwave frequencies, the size of the cavity necessary to form this standing wave is on the order of centimeters. The purpose of the cavity is to transfer power from the microwave source to the inert gas flowing through the plasma capillary. The efficiency of this power transfer depends on the impedance match between the cavity (including the plasma) and the coaxial cable leading to the magnetron. been described (13,14). Impedance matching devices have Maximum power is transferred to the plasma when the amount reflected back to the magnetron is tuned to a minimum. The electromagnetic fields and currents produced within the cavity decrease with penetration into the cavity wall. Energy can be lost as thermal heating of the cavity when penetration is high. The choice of construction material for the cavity usually involves a compromise. The metals which best inhibit penetration are good conductors such as silver, copper, gold and aluminum. Unfortunately, these metals are either difficult to machine or corrode, making them less efficient. Brass is often used and in some cases is coated with silver or gold (8). One of the original cavity designs was the tapered rectangular (TEq ^3 ) type which was introduced by Broida Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5 (15,16). The plasma capillary tube is positioned in a slot at the short-circuited end, in the electric field direction. Since there is no physical contact between the cavity and plasma capillary, the cavity can be moved without disturbing the discharge. McCormack et al. (17) found the tapered cavity to be more sensitive than the 3/4-wave Evenson coaxial cavity, but that it could not accept as much sample. Two types of foreshortened 3/4-wave coaxial cavities have been described (18). As in the case of the tapered cavity, the discharge is viewed perpendicular to the plasma capillary. However, when the plasma is operated at atmospheric pressure, axial viewing is also possible (19,20). Moye (21) reported difficulties such as high noise and low discharge tube life using the Evenson type 3/4-wave cavity and favored the tapered design. The foreshortened 1/4-wave coaxial cavity is a modification which, unlike the 3/4-wave, allows adjustment of the cavity without breaking vacuum lines to the discharge tube. Another modification is the foreshortened 1/4-wave radial cavity (18). This cavity operates well at reduced pressures and is normally viewed axially. The most significant cavity design improvement was reported by Beenakker (22,23). This cavity was designed for efficient operation in either argon or helium at atmospheric pressure. The cavity operates in the mode with the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6 plasma capillary positioned axially. When helium is used at atmospheric pressure, the plasma is self igniting and is highly stable (24). The plasma emission is viewed.axially. Mulligan et al. (25) compared the performance of the Beenakker cavity to several others in the simultaneous determination of As, Ge, Sb and Sn. The Beenakker TM q ^q cavity represents the state of the art in microwave cavity design and is now commercially available (26). Modifications of this cavity to improve coupling efficiency and to permit direct introduction of aqueous samples is a very active area of research (27,28,29). Plasma Capillary The plasma is contained by a nonconductive tube through which the support gas and analyte species flow. Quartz is the material of choice because of its high dielectric character and transparency in the UV-visible region of the spectrum. The inner diameter of the tube is an important parameter which affects plasma stability and emission intensity (17,21,30,31). The method employed in viewing emission from the plasma varies with the type of resonant cavity. As mentioned previously, the plasma may be viewed axially down the length of the plasma or transversely through the capillary wall. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7 Commercial Instrumentation Rapid development and proven utility of the microwave induced plasma have led to its commercial availability. An inventory of the various systems available was recently presented by Broekaert (26). Manufacturers and a description of the components are given. Radiation Exposure The hazards of stray microwave radiation are not completely understood and remain a controversial topic. This is reflected in the large discrepancy between the safety standards set in the U.S. U.S.S.R. (0.01 mW/cm2). (10 mW/cm2) vs. the Stanley et al. (32) measured stray radiation levels for four cavity designs. Microwave power density dropped to 10 mW/cm2 within 18 cm of the tapered cavity. Stray levels were this high 22 cm from an Evenson 1/4-wave cavity, and 50 cm from a Raytheon type "C" cavity. The Broida 3/4-wave cavity exhibited a maximum power density of 1 mW/cm2 at the tuning adjustment. f Power densities were found to fall off rapidly with distance from the cavities. When metal objects were placed close to the cavities, power levels at points of reinforcement were an order of magnitude larger. These authors constructed a shield from 16 guage sheet metal with 2.38 mm holes. Radiation leakage was reduced to less than 0.05 mW/cm2. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 8 Van Dalen et al. (33) measured stray radiation levels for an actual laboratory situation employing a 1/4-wave Evenson type cavity. When reflected powers were adjusted to <1%, a maximum of 2 mW/cm2 was measured. This stray radiation level was reduced to 0.1 mW/cm2 by enclosing the system in an aluminum box. Quimby et al. (34) suggested shielding may be necessary when argon is used as a support gas. Discharge Properties Unequivocal mechanisms for plasma formation and analyte excitation cannot be found in the literature (8). The following discussion is a summary of the processes taking place in microwave induced plasmas that have received general acceptance. The plasma gas, usually argon or helium, flows through a quartz tube which confines the discharge. This tube is positioned along the axis parallel to the electric field inside of the resonance cavity. When a seed electron is introduced by a Tesla coil, it oscillates in the field. As the electron accelerates, it collides with plasma gas atoms as shown in the following equations: Ar + e --- * Ar(+) + e + e (1) Ar( + ) + e — (2) * Arm + hvcont±nuum Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The ionization of an argon atom produces two electrons characterized by different velocities and temperatures (35,36,23). A high-energy, low-density group is responsible for subsequent ionization of neutral argon atoms via equation 1. The second group, which is characterized by low-energy and high-density recombines with an ionized argon atom to form a metastable species. A stable plasma results from the equilibium between these two processes and is p ressure dependent. When atmospheric pressures are employed, higher electric fields or lower frequencies are required. The excitation mechanism in a MIP is believed to involve the metastable species. levels at 11.49 and 11.66 eV. Argon has two metastable Helium has two levels of higher energy at 19.73 and 20.53 eV. Electrons promoted to a metastable state are forbidden by selection rules to release energy by emitting radiation. Instead, they loose their energy through collisions with other atoms. Because only radiationless processes are allowed and the lifetime of a metastable state is relatively long, they are believed to be responsible for spectral excitation (37,23,38). The following mechanism is consistent with the proposed radiative ionization-recombination (RIR) model: Ar m + M M( + ) + e — Ar + M(+) + e » M* M* -- > M + hv characteristic (3) (4) (5) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 10 Equation 3 represents Penning ionization of the analyte species, M, whose ionization potential must be less than the excitation potential of the metastable state. Excess energy is transferred to the electron in the form of kinetic energy. Radiative recombination of the analyte ion with a low-energy electron produces an excited state which spontaneously decays with emission of characteristic radiation. The RIR model is the most commonly accepted mechanism for plasma excitation. Recently, Brassem et al. (39) re-evaluated the model for a low pressure MIP. The RIR model was favorable to a purely thermal excitation model but was still off by a factor of two. Direct excitation of analyte atoms by metastable species or high energy electrons is possible but definitely not the dominant mechanism (37,39,23). Bush and Vickers (37) characterized excitation conditions for the MIP by measuring spectroscopic temperature, electron temperature, electron concentration, relative argon metastable state concentration and argon emission intensity. They found that the parameters which controlled excitation were the concentration and energy of the two different electron groups and the concentration of the metastable argon atoms. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 11 These authors state that it is unnecessary to define microscopic processes which constitute excitation in a plasma at atmospheric pressure. This is due to the fact that local thermodynamic equilibrium (LTE) is often attained by plasmas at atmospheric pressure where particle densities are high (40). Under these conditions, the population of energy levels follows the Boltzmann distribution, and the Saha equation describes the yield of ionization products (37). The state of the plasma' is defined by its temperature and density. Bush and Vickers (37) measured spectroscopic excitation temperature in the MIP by observing relative radiances of the support gas spectral lines. Use of this method required that wavelengths, energy levels and transition probabilities were known quantities. A plot of log (IX/gA) vs. log E for the spectral transitions yields a straight line with a slope of -1/kT where: I = line intensity g = statistical weight-of the upper level A = transition probability X = wavelength of the observed line E = energy of the upper level k = Boltzmann constant T = excitation temperature (Kelvin) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 12 The spectroscopic temperature of an argon MIP at atmospheric pressure has been reported to be approximately 5000 K at 100 watts of applied power (41,42). For a helium plasma at atmospheric pressure and 100 W, a temperature of 7250 K was observed (24). This increase in temperature is due to the higher energy of helium metastable state. Sample Introduction The early applications of the MIP to spectrochemical analysis reflected the fact that only small amounts of sample could be introduced, preferably already in the gas phase. Hence, the MIP was initially used as a detector for gases, gas chromatographic effluents or volatilized compounds. The discussion of analytical applications of the MIP will be limited to those which employ introduction of the sample as a vapor. The introduction of aqueous samples has taken on a whole new direction since the advent of the Beenakker cavity (22). Greenfield presents an excellent review of the literature with respect to pneumatic nebulization, ultrasonic nebulization and sealed tube excitation of aqueous samples (7). A review by Zander and Hieftje (8) contains a comprehensive listing of elements determined by solution analysis in microwave induced plasmas. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 13 Gases The simplest examples of sample introduction involve direct injection of gases. The first successful use of a MIP in spectrochemical analysis was reported by Broida (16) who determined isotopic concentrations in nitrogen and nitric oxide. Taylor et al. (43) determined trace impurities in the argon plasma support gas. Serravallo and Risby (44) found that the presence of air severely limited the application of direct injection techniques for the determination of gaseous air pollutants, since oxygen and nitrogen quenched the emission from atomic species. et al. (45) determined CC^, NO and SO2 Dagnall in air by GC-MES. Heated Filament Vaporization Runnels and Gibson (46,47) evaporated the analyte species onto a platinum filament. After insertion into the plasma gas stream, the filament was electrically heated to vaporize the sample. Aldous et al. (48) used a platinum or tungsten loop to vaporize samples into a plasma supported at the mouth of the quartz capillary. Kawaguchi et al. (30,49) used a tantalum filament and found that the addition of KC1 enhanced spectral emission for several elements as well as suppressing interference effects. A variety of similar applications employing filament vaporization have been reported (50,51,52). The main Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 14 consideration in interfacing these types of sample introduction systems is that the analyte species must not be allowed to plate out on the walls, of the system before entering the plasma (8). Carbon cup (53,54), carbon rod (55) and platinum boat (56) devices have also been used. Watling (57) amalgamated trace amounts of mercury in seawater onto silver wool after reduction by tin (II) chloride. Mercury was vaporized by heating the silver wool in the argon stream of a microwave plasma. Generation of Gaseous Species The most common approach for many elements is to form the volatile hydride. Lichte and Skogerboe (58) designed an arsine generator which was directly attached to an argon MIP. Barret and Copeland (59) compared various hydride generation techniques with MIP excitation with respect to sensitivity and precision. Mulligan et al. (25) compared four different microwave cavity designs for the simultaneous determination of As, Ge, Sb and Sn as volatile hydrides. Other methods of chemical modification of the analyte to produce a volatile species include reduction of mercury compounds (60) or conversion to volatile chlorides (61). In the latter method, plasma gas containing HC1 was passed over the sample after preheating to 850°C. Conversion and determination of Bi, Cd, Ge, Mo, Pb, Sn, Tl and Zn was Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 15 accomplished. Runnels and Gibson (46) introduced Cu, Co, Cr, Fe and Mn as acetylacetonate chelates. These were vaporized from a platinum filament as previously described. Tanabe et al. (62) determined ultratrace levels of ammonium, nitrite and nitrate nitrogens using a gas generation technique coupled with MIP detection. Quimby et al. (34) trapped volatile trihalomethanes on a porous polymer adsorbent and thermally desorbed the analyte into the plasma. A novel method of introducing solid samples is by direct laser vaporization (63). Picogram levels of Al, Cr, Cu, Fe, Mn, Mo, Ni, Pb, Ti and Zn were determined in metals and required no sample preparation (i.e. dissolution). A similar technique was applied to the analysis of aluminum and zinc samples (64). Layman and Hieftje (20) employed a computer controlled, microarc sample atomization system for trace elemental analysis. Bauer and Natusch (14) have utilized MIP emission detection in evolved gas analysis (EGA). heated in the support gas stream. The sample is Compound identification is based on the coincident observation of its metal and nonmetal components at its temperature of vaporization. Coincidence exists for pure halide, sulfide and sulfate salts of Cd, Hg, Pb and Zn. In a subsequent article (65), carbonate compounds were determined in coal fly ash. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Carbon 16 dioxide was monitored as the evolved gas. One obstacle to the method was the possibility of chemical reactions altering compound identity before vaporization. Gas Chromatography-Microwave Emission Spectroscopy The most common-application of the MIP has been as a detector for gas chromatography (GC-MES). The first practical application of this type was reported in 1965 by McCormack, Tong and Cooke (17). The spectra emitted by organic molecules showed that fragmentation had taken place to produce atomic and diatomic species. This emission could be used in the sensitive and selective determination of compounds containing C, Cl, F, I, P and S. Bache and Lisk simultaneously published their work on the determination of organophosphorous pesticides by GC-MES (66). The popularity of GC-MES is evidenced by the volume of work that has been done since its introduction. The basic requirements for this method are as follows: 1. The analyte must be present in the gaseous state following chromatographic separation. 2. The plasma must be able to handle the amount of effluent coming from the gas chromatographic column. 3. The ionization potential of the desired analyte transition must be smaller than the excitation potential of the metastable state. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 17 Taking these limitations into consideration, the detector can be made element selective simply by varying the wavelength of observation. Method Development Wavelength Calibration A variety of methods have been used to optimize the wavelength of observation. Early investigators set the monochromator to the desired setting and optimized response by injecting standards at small wavelength intervals. McCormack et al. (17) passed argon support gas over volatile compounds of interest and recorded the resulting spectra. A similar method was used by van Dalen et al. (33) who controlled diffusion from capillaries by regulating temperature in relation to solute volatility. Estes, Uden and Barnes were able to optimize wavelength settings for Si, C and H by observing background emission resulting from the quartz capillary tubing and low level hydrocarbon impurities in the carrier gas (67). Tanabe et al. (68) recently published wavelengths, transition assignments and relative intensities for nonmetal species observed in an atmospheric pressure helium MIP. Sensitivity The first GC-MES systems employed tapered cavities with argon as a support gas at atmospheric pressure (17,66,69). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 18 The first attempt to improve sensitivity was reported by Bache and Lisk who. used reduced pressure plasmas in argon (70) and helium (71,72). Sensitivity and linear range were improved at low pressures by minimizing collisional deactivation of the analyte species. Plasma temperature, electron concentration and resulting emission intensities at low pressures were very dependent on flow rate (36). The use of helium further improved results due to the higher energy of the helium metastable state. Microwave plasmas in helium could only be sustained at reduced pressures between 5 and 30 Torr when tapered cavities were used. Moye achieved optimum sensitivity for organophosphorous pesticides in a helium plasma containing 15% argon (21). Slit width and He/Ar ratio were found to have a great effect on the signal to noise ratio. Detection limits for certain elements may be limited by background emission due to impurities in the plasma gas. Runnels and Gibson reported background emission from molecular species such as C 2 , CH, N2 , NH, OH and CN (46). Subsequent investigators have further cataloged the species commonly found in MIP background spectra (24,67). Reamer et al. (73) used a wavelength modulation technique for background correction. To decrease background emission, purification of the plasma gas has also been investigated. Parkinson (74) used Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 19 a glass U-tube packed with activated carbon and cooled with liquid nitrogen. This system effectively removed C>2 / Ar and CC> 2 from He at flow rates of 3 L/min. ^z Runnels and Gibson purified argon by passing it over Mg at 550°C (46). Brenner (75) used molecular sieve 5A cooled in a dry ice/acetone bath for the ultrapurification of helium used in a microwave emission detector. Estes et al. (67) found that scrubbing helium with liquid nitrogen cooled molecular sieve 5A prevented emission from NH and OH but not ^ or N 2 (+). Experimental parameters such as microwave power and carrier gas flow rate also have a significant effect on sensitivity. An increase in power can affect emission intensity by influencing excitation conditions such as metastable concentration. Increased analyte emission is often accompanied by larger background levels, however. The flow rate influences excitation conditions and determines analyte residence time in the detector. Emission intensity varies along the length of the plasma when viewed in a transverse configuration (76,77) and spacially in the crosssection of a plasma viewed axially (67). Selectivity When characteristic emission of the analyte species is observed, a great degree of selectivity is achieved. can be an important advantage when chromatographic Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. This 20 separation is incomplete or chromatograms are very complicated, as is often the case when using capillary columns. The MIP can also be used as a general detector for organics by observing carbon emission. I The MIP has been used in conjunction with a general t detector in gas chromatographic applications. McLean et al. (78) split the gas chromatographic effluent between MIP and flame ionization detectors (FID) as did other investigators (79,76). Tanabe et al. (80) combined thermal conductivity (TC) detection with an atmospheric pressure, helium GC-MES system. Dual strip chart recordings were obtained which aided in compound identification. The selectivity ratio is a term which compares analyte emission intensity at its characteristic wavelength to background emission at that same wavelength. Since the background emission is usually due to either atomic or molecular carbon emission (C, C 2 , CH, CN, CO, etc.), the selectivity ratio is calculated with respect to carbon. Early investigators calculated the ratio in terms of the volume of solvent required to give an equivalent signal (17,66). Subsequent articles defined the selectivity ratio in terms of response per gram-atom of the element vs. carbon (79,34). Dagnall et al. (81) based the selectivity ratio on the molar concentration of carbon vs. analyte species resulting Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 21 in equivalent peak areas. This method of calculation seems more appropriate since emission is directly proportional to the specific number of atoms present. Selectivity ratios reported recently in the literature seem to be following this trend in calculation although there is still some inconsistency (23,67,80). Selectivity is mainly a function of spectral interference from carbon band emission. This being the case, the selectivity ratio can be increased by decreasing the amount of band emission. The use'of helium as the plasma support gas leads to better fragmentation efficiency (71). Better selectivity can be achieved in a helium plasma because atomic lines can be observed with a narrow bandpass, and molecular background emission due to impurities is reduced (23). Estes et al. (67) used a quartz refractor plate background corrector to improve selectivity ratios for elements whose emission lines occur in high carbon background regions. Braun et al. (82) observed transitions in the vacuum ultraviolet region and found molecular background emission to be less of a problem. Selectivity is strongly dependent on the resolution of the monochromator employed and slit width. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 22 Response vs. Structure McCormack et al. (17), as well as Bache and Lisk (66), found that response for nonmetals was dependent on.the structural environment in the parent compound. This required individual calibration for each compound of interest. In subsequent studies with organic mercury (83,84,85) and arsenic compounds (86), no such dependence was found. They attributed this phenomenon to the relative bond strengths between the heteroatom and other atoms in the compound. Nonmetal atoms, unlike metals, form strong bonds with atoms such as carbon and oxygen. Dagnall et al. (87) reported that response for sulfur compounds in an argon plasma was dependent on the structural environment in the parent compound. In a subsequent article (88), the detector was improved by using an alternate cavity with helium as the support gas and inserting a platinum wire catalyst into the plasma tube. The platinum partially vaporized and aided in fragmentation. These same investigators eventually returned to an argon plasma at atomospheric pressure because of simplicity (31). With this system, they described the feasibility of determining interelement ratios between Br, Cl, I, P, S and C (81). These ratios were valid for I, P and S compounds since atomic emission was observed. Results were not good for Br and Cl since band emission was monitored and response was flow rate and concentration dependent. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 23 McLean and coworkers used a low pressure helium MIP for determining empirical formulas of compounds (78). The gas chromatographic effluent was split between a MIP and a flame ionization detector for non-specific detection. Elemental ratios were determined for C, H, D, F, Cl, Br, I, S, P, N and 0. A GC-MES instrument based on McLean's work was manufactured commercially by Applied Research Laboratories. The MPD-850 bulletin claims, as do authors (78,89), that sensitivity in the low pressure helium MIP for a particular element is independent of the type of compound introduced and proportional to the number of atoms present. Closer inspection of the published data reveals that this claim is valid only in homologous series with exclusion of the lowest members (33). Better results have been obtained in atmospheric pressure helium plasmas using the Beenakker cavity design (23). Reproducibility Two major problems have plagued GC-MES since its introduction: i) the inability of the low power plasma to accept large amounts of material; and ii) the loss of transparency of the quartz plasma tube when viewing in the transverse configuration. Even though the MIP has a high excitation temperature, it does not have sufficient energy Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 24 to vaporize or atomize materials' introduced in large quantities. To reduce the effect of the solvent peak on both of these factors, many investigators have simply extinguished the plasma during this time (17). This has been effective when the plasma has time enough to stabilize before the elution of analyte species. A second approach to minimize the solvent effect has been to vent the solvent away from the detector (87). Quimby et al. (79,34,90) used a high temperature valve system which allowed the effluent to be directed either to a vent or to the plasma. Estes et al. (91,67) employed a chemically deactivated, low volume, valveless fluidic logic gas switching interface to vent large quantities of solvent. Inertness was demonstrated by determining chemically active and thermally labile trialkyllead chlorides. An atmospheric pressure helium plasma was used which permitted venting. The plasma could, therefore be viewed axially, circumventing any reproducibility problems due to the loss in transparency of the quartz capillary. The loss in transparency is due to the build-up of carbonaceous and metallic oxide deposits on the interior walls of the discharge tube and devitrification of the quartz at high temperatures. McLean, Stanton and Penketh doped a helium plasma with traces of oxygen or nitrogen gas (78). This scavenger gas was found to reduce carbon Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 25 deposits. Serravallo and Risby also studied the effects of doping gases in GC-MES (92,93). Beenakker prevented carbon deposition by adding 0.1% 02 or N2 which led to the formation of volatile carbon oxides or nitrides (23). This technique also increased the linear range by one order of magnitude without affecting the background or sensitivity. Schwartz reported an important dependence on oxygen scavenger concentration in the determination of hydrogen (94). Optimum detection was achieved with 0.4 to 1.5% oxygen. In a subsequent paper, an anomaly in hydrogen response was reported due to a charge transfer reaction with the scavenger gas (95). To avoid viewing the plasma through the discharge tube wall, Dagnall et al. (31) attempted to develop an "open" plasma detector at the tip of the plasma capillary. Unfortunately, the high flow rates required to sustain a stable plasma were not compatible with gas chromatographic requirements. More recently, Gebhart (77) viewed emission in the transverse configuration through a hole drilled into the side of the quartz capillary. Improved long-term reproducibility was reported. Applications One method that has found widespread application is the formation of volatile chelates followed by separation and Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 26 detection by GC-MES. Dagnall et al. (96) determined volatile metal chelates of Al, Cr, Cu, Ga, Fe, Sc and V. Talmi measured trace levels of selenium in environmental samples after conversion to the thermally stable and volatile piaselenol complex (97). As little as 40 pg Se could be detected after extraction of the complex into toluene and concentration. Talmi also reported the determination of arsenic and antimony after conversion to volatile complexes (98). Other applications of metal chelates have been reported for Al, Be (99), and chromium (99,92,100,93). Other methods of chemical modification prior to analysis include reduction of alkylarsenic acids to arsines using NaBH^ (86). Quimby et al. (90) used capillary column GC-MES to determine aqueous chlorination products of humic substances in water supplies. Non-volatile compounds were derivatized with diazomethane before determination. Since the carrier gas for the capillary column was only 4 mL/min, a make-up gas of 50 mL/min. was required. Novel applications of GC-MES have included the use of reflected microwave power as a general detector (101) and the determination of isotopic concentrations (102,38). One of the most comprehensive applications of GC-MES was recently published by Estes, Uden and Barnes (67). A fused silica gas chromatographic capillary column was interfaced Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 27 to an atmosphic pressure helium MIP. Calibration curves, selectivity ratios and detection limits were established for V, Nb, Cr, Mo, W, Mn, Ee, Ru, Os, Co, Ni, Hg, B, Al., C, Si, Ge, Sn, Pb, P, As, S, Se, F, Cl, Br, I, H and D! No attempt will be made to describe all of the published applications utilizing GC-MES. The interested reader is directed to the review articles previously mentioned. However, an overview of the elements determined, observation wavelength, cavity type, support gas and pressure, detection limits and selectivity ratio is presented in Appendix H. Commercial Instrumentation GC-MES instumentation is available commercially. Hobbs et al. (103) reviewed the use of the ARL MPD-850 for environmental pollution studies. The instrument was capable of selectively monitoring up to 12 elements simultaneously over a wide linear range. Computer programs for selectivity and data handling have also been described (104). Hobbs used a pyrolysis furnace in conjunction with the MPD-850 to analyze river water extracts for compounds containing C, H, Cl and S (105). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 28 Summary Element Selective Detectors for GC The development of element selective detectors has received much attention. Selective detectors simplify chromatograms so that less efficient separation is needed, which often leads to reduced analysis time. When highly efficient capillary columns are used, selective detectors simplify the resulting chromatogram. Element selective detectors often provide qualitative information not available with general detectors. Electron Capture Detector The electron capture (EC) detector is sensitive primarily to halogenated compounds (F<Cl<Br<I). Mulligan et al. (106) found the EC detector to be more sensitive than microwave emission spectroscopy (MES) for the determination of polybrominated biphenyls. however. The MIP was more selective, Other disadvantages of EC detection include susceptibility to contamination, small linear range (102 10*) and unpredictability of response (107). Coulometric and Conductometric Detectors It is usually necessary to oxidize or reduce organic compounds eluting from the gas chromatograph to simple, inorganic gases before detection. For the Ag(+) detector cell, sulfur compounds are reduced with hydrogen to H^S. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 29 Sulfur compounds are oxidized to SC^/SO^ with oxygen at 750"C before entering the I(-) cell. Absolute response to ng levels with selectivities as high as 106" are possible (107). Since the oxidation/reduction furnaces add dead volume to the detector, Fredericks et al. (108) suggested that these electrochemical detectors were best suited for packed columns and relatively simple mixtures. A new conductivity detector design for S, Cl and Ng was developed by Hall (109). Sub-nanogram sensitivity, a linear range of 10s and selectivities as high as 10s were reported. Flame Detectors Gutsche and Herrmann (110) obtained characteristic emission for iodine, bromine and chlorine compounds eluting from a gas chromatograph. The burner contained indium which produced band spectra with halides. Nowak et al. (Ill) developed a flame detector based on sodium emission from a sodium sulfate-coated wire in the presence of halogens. The addition of sodium compounds to a flame ionization detector (FID) to enhance response for phosphorous and halogen compounds was first described by Karmen et al. (112). The most popular flame detector is the flame photometric detector (FPD) used primarily in the sulfur or phosphorous mode (113). When the 3939 A S£ band is observed, high sensitivity can be attained, but response is Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 30 proportional to the square of sulfur concentration. -Several authors have reported that response is dependent on the environment of the sulfur atom in the molecule (114-116). Beroza and Bowmann reported sulfur to hydrocarbon selectivities of 10“ (117). However, when sulfur compounds were not completely separated from the hydrocarbon, quenching of the chemiluminescent emission was possible (118). Golovnya et al. (119) suggested that the FPD should not be used for quantitative gas-liquid chromatography below 0.1 mg/mL because of signal drifting. Plasma Detectors Ellebracht et al. have recently reported the use of a dc discharge plasma as a gas chromatographic detector for sulfur compounds in the vacuum ultraviolet region (120-122). This high temperature, high energy plasma minimized quenching effects and could handle as much as lOOyL of solvent without being extinguished. The relative standard deviation for replicate injections of 259 ppm carbon disulfide in hexane was 4.5%. The minimum detectable level obtained for sulfur at 1807 A with this VUV-PAES system ranged from 100 to 300 pg S/second. The linear dynamic range covered three orders of magnitude with a selectivity ratio of 1000 for a variety of organic solvents. A similar dc discharge for detection of organics and halogens in the UV-visible was reported by Braman and Dynako (123). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 31 Windsor and Denton determined halogens, C, Fe, H, Pb, Si and Sn by GC-ICP (124). Detection limits (low ng range), linear dynamic range (>103) and selectivities were considered favorable to both the FPD and MIP. Unfortunately, intensities for B, F and S were low while N and O lines were completely absent. Conclusions The microwave induced plasma has made an important contribution to selective gas chromatographic detection. The "state of the art" GC-MES system is capable of determining most of the elements in the periodic table with picogram sensitivity in many cases. The detector exhibits a weak, and relatively- simple, background spectrum. It is inexpensive to operate with low power requirements and low gas consumption. The major problem which still requires attention is the sample load capacity of the MIP. The combination of separate evaporator/atomizer units with the MIP is promising. In GC-MES, the trend seems to be moving towards the use of capillary columns. Purpose of Study Background A study was proposed by three faculty members at the University of Iowa, College of Nursing, to investigate Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 32 possible correlations between multiple sclerosis (MS) and a condition known as lactose malabsorption (125). Recent findings indicated a connection between the two and warranted further investigation: 1. In the summer of 1976, a group of 45 MS patients were questioned about their dairy product consumption. Approximately 25% of this group indicated that they no longer drank milk because jof the discomfort that resulted. 2. The Eastern Iowa MS Clinic at the University of Iowa Hospital reported that approximately 50% of their patients refused milk. 3. This aversion to milk and milk products was commonly spoken of during conversations with MS patients at regional MS Society meetings. Fatigue has been described as one of the initial symptoms of MS. This fatigue is the primary factor which limits a patient's activity (126). McAlpine et al. (127) observed that factors such as infections, trauma, fatigue, physical exertion, emotional disturbance and preventative inoculations all appear to be capable of influencing the onset and course of MS. Lactose malabsorption is one possible explanation for this fatigue. Fatigue resulting from food allergies may be due to the excess energy required to support a Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 33 hypersensitive response (125). Foods which are most commonly associated with allergic reactions include sugar, milk, chocolate, wheat, corn, citrus, yeast, spices and eggs (128). Lactose Intolerance Enzymic digestion in the oral cavity, stomach and small intestine removes most dietary constituents (129). However, fibrous constituents (130) and small molecular weight carbohydrates such as stachyose in beans, raffinose in cottonseed meal, and the artificial sugar lactulose are not attacked by mammalian enzymes (131). Instead, they are degraded by the microbial community of the large intestine. Fermentation products in the large intestine include volatile fatty acids, methane, hydrogen and carbon dioxide. Lactose, 0-(S-D-galactopyranosyl-(l-*-4)-0-Dglucopyranose, is a carbohydrate found in milk but otherwise does not occur in nature (132). Individuals who are unable to completely digest lactose are said to be lactose- or "milk"-intolerant. People have been known to suffer from this condition for centuries; however, its cause was not identified until just recently (133). The enzyme lactase, a p-galactosidase, hydrolyzes lactose into its monosaccharide components of glucose and galactose. These can then be absorbed into the small Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 34 intestine. Individuals with decreased lactase activity have been found to absorb as little as 25 to 58% of the lactose they ingest (134). If milk products are consumed, non-absorbed lactose is transported down the digestive tract. As the lactose reaches the colon, it is fermented by intestinal bacteria into lactic acid and hydrogen gas (135,136). The resulting symptoms are watery and acid diarrhea, flatulence, cramps and abdominal bloating (137). The definitive test for lactose malabsorption requires a biopsy of the small intestine and determination of lactase activity (138). One common indirect diagnostic method is the lactose tolerance test (LTT). Blood glucose levels are monitored following an oral dosage of lactose. Lactose intolerant subjects show very small increases in glucose because of decreased absorption. Unfortunately, this test is influenced by factors such as gastric emptying and intermediary glucose metabolism, leading to possible errors (139). In 1968, Levitt et al. reported that breath hydrogen excretion was a useful indicator of carbohydrate malabsorption (140). Diagnosis was based on the principle that nonabsorbed lactose passes through to the colon where hydrogen gas is formed. Hydrogen production in normal subjects was found to average 0.24 mL/min. in the fasting state, increasing to 1.6 mL/min., after intestinal Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 35 instillation of lactose. Approximately 14% of the hydrogen produced in the colon is absorbed into the bloodstream and excreted via the lungs (136). Bacterial degradation of nutrients is the only source of hydrogen in the body (138). Hence, breath measurements are directly proportional to the amount of lactose not absorbed in the small intestine (134). Hydrogen Breath Test Normal subjects exhale little or no hydrogen after fasting overnight. When a dose of lactose is administered orally, only those subjects with lactase deficiencies show an increase in breath hydrogen levels over time. Hydrogen excretion begins to increase an hour after the dose, and normally reaches a maximum at approximately 1.5 to 2 hours (141,142). Measurement of the relative increase in hydrogen concentration over baseline levels is the basis for diagnosis. The use of breath hydrogen excretion as an indicator of carbohydrate malabsorption requires that: i) no appreciable hydrogen is produced from carbohydrates in the small intestine; and ii) delivery of carbohydrate to the colon results in a detectable increase in breath hydrogen excretion (139). An increase of 20 ppm over baseline hydrogen levels has been suggested as a positive indicator for lactose malabsorption when 50 g of lactose is administered (143). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 36 The hydrogen breath test (HBT) has been described as the most suitable test for the screening of lactase deficiency (138). Bond and Levitt (134) found that the quantity of lactose absorbed, as determined by breath hydrogen excretion, correlates very closely to the quantity of unabsorbed lactose aspirated directly from the 'ileum. The HBT was also found to agree with the results of standard lactose tolerance tests (144). The advantage of the HBT over tolerance tests is that it reflects the quantity of sugar not absorbed, so much smaller doses can be given. It is sufficiently sensitive to detect malabsorption pf as little as 5 to 10 grams of carbohydrate (139). Tolerance tests require doses of 50 to 100 grams of carbohydrate. HBT Errors Possible errors associated with the HBT were summarized by Solomons (144). In some cases, elevated baseline hydrogen levels were encountered. This is often the case after a period of sleep due to hypoventilation and decreased passage of flatus. Since interpretation of the HBT is based on peak hydrogen concentrations relative to baseline levels, initially elevated levels may invalidate the test. The simplest solution to this problem is to delay the test diet until baseline hydrogen levels equilibrate. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 37 Delayed gastric emptying or retarded intestinal transit can also obscure HBT results. This problem has been encountered in lactose tolerance tests. The carbohydrate could be delivered to the small intestine by direct instillation, but this defeats the non-invasive nature of the test. Alternatively, the time of observation could be extended in order to detect delayed hydrogen excretion. The production of hydrogen in the colon depends on the availability of bacterial flora and their ability to produce hydrogen. Colonic flora not capable of producing hydrogen from carbohydrate has been reported but is believed to be rare (139). Lack of knowledge about the bacteria responsible for hydrogen formation makes it difficult to predict what factors influence this mechanism. The use of broad spectrum antibiotics has been shown to inhibit hydrogen production by colonic bacteria (144). One way to determine a subject's hydrogen production capability is to administer a non-absorbable carbohydrate such as lactulose. » Previous Application of Breath Tests The technique of measuring exhaled breath constituents with respect to ingested food has been used in a variety of studies. Calloway and co-workers evaluated the suitability of certain diets for space (145) and studied the effects of antibiotics (146). Breath analysis has also been used in Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 38 the study' of intestinal absorption, gross parasitism, liver function and effects of gastro-intestinal bacteria on endogenous and exogenous materials (147). Determination of Hydrogen in Breath Chromatography The determination of trace amounts of hydrogen gas has been reported previously in the literature. These methods normally employ gas-solid chromatographic separation on molecular sieve columns. Separation of hydrogen in breath analysis studies have almost exclusively employed either 5A or 13X molecular sieve columns. Other adsorbent phases such as PorapakfSQ, activated alumina and activated silica have also been used successfully. Separation of hydrogen from other permanent gases is complete at ambient or sub-ambient temperatures. Hydrogen conveniently elutes from the column before other breath constituents. Detectors The first quantitative determinations of respiratory gases employed gas chromatography with thermal conductivity (TC) detection. Carle (148) developed a microbead thermistor detector for sensitive determination of microbial respiratory gases including hydrogen. The use of this detector resulted in a negative peak anomaly for hydrogen at certain concentrations. Levitt and coworkers (136,139,149) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 39 used TC detection in developing the hydrogen breath t6st for lactose intolerance. They reported a detection limit of 10 ppm using 2 mL breath samples. In order to detect hydrogen levels down to a few parts per million (ppm) using TC, Solomons (142) increased the sample loop size to 16 mL. More recently, Niu et al. (150) used TC to detect as little as 2 ppm hydrogen in breath for a 1 mL sample injection. Gearhart et al. (141) employed a helium ionization detector (HID) to determine lactose malabsorption by breath analysis with gas chromatography. The lactose test diet was reduced to 0.25 g per kg body weight because detection limits for hydrogen were improved to sub-ppm levels. However, breath levels below 10 ppm required extensive purification of the helium carrier gas. scrubbed with Ba( 0 H ) 2 Breath samples were and anhydrous CaSO^ before injection. There have been numerous reports about problems associated with the helium ionization detector. Parkinson suggested that these detectors were useful over a very limited concentration range due to anomalous response for hydrogen (151). When extremely pure helium carrier gas was used, a negative peak was observed for traces of hydrogen. As the hydrogen level was increased, the response became progressively less negative and eventually positive! Payne-Bose and coworkers (152) compared the HID to thermal conductivity detection. They found the HID to be variable Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 40 in performance and difficult to maintain. Similar comments were obtained through personal communication with representatives from Varian Instrument Company who formerly marketed the HID. The use of argon ionization detectors has also been reported. Gawlowski et al. (153) found that at flow rates of 25 to 75 mL/min., the detector is mass-flow sensitive. At higher flow rates, sensitivity decreased sharply. The detection limit for hydrogen'was 100 pg/second with a linear dynamic range of 100. The operational mechanism for the argon ionization detector was discussed in a subsequent article by the same authors (154). The development of electrochemical detectors for hydrogen and other permanent gases has also received some attention. Bergman et al. (155) used a metallized-membrane electrode to measure hydrogen production of bacteria in the human gut. This detector was used in a simple portable gas chromatograph. The system was capable of determining less than 10 ppm hydrogen with a precision of ±0.1 ppm. Guglya developed a platinum-coated lithium niobate plate as an electrochemical sensor for as little as 0.2 ppm hydrogen (156). Taylor and coworkers (43) determined trace impurities in argon carrier gas using a microwave induced plasma (MIP). By observing atomic emission from the plasma, element Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 41 specific detection of the impurities, including hydrogen, was obtained. The detection limit for hydrogen at 4861 A was 1.8 ppm by weight. This study was performed on a constant flow of sample (the plasma gas) which improved sensitivity. Lefebure used a 4 MHz discharge in helium at atmospheric pressure for the determination of ppm levels of H, N, 0 and CH^ (157). When both N and H were present in the plasma, a "line" at 3360 A characteristic of the NH radical was observed. Atomic hydrogen emission in the Balmer Series was found to decrease in intensity with an increase in hydrogen concentration! Recently, Schwarz characterized the emission from atomic hydrogen in a MIP (95). Helium was used as the plasma gas at reduced pressure. An anomaly in hydrogen response was observed in the presence of either oxygen or argon. This was believed to be due to a charge transfer reaction between O 2 and H(+) which competed with excitation by metastable helium atoms. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 42 CHAPTER II EXPERIMENTAL Introduction The first objective of this work was to assemble a GC-MES system from individual components developed previously in our laboratory. A Nester Faust, Model 750, gas chromatograph was completely modified. The vacuum ultraviolet spectrograph built by Dreher (158) was converted into a scanning spectrometer. The plasma observation work reported by Gebhart (77) was the basis for the optical interface between the gas chromatograph and the spectrometer. The wavelength readout system was a modification of the digital integration circuit developed by Cox (159). A block diagram of the experimental set-up is given in Figure 1. A quartz capillary tube extended from the end of the chromatographic column. A microwave cavity, or antenna, was positioned so that microwave energy was focused on a small portion of the quartz capillary. The microwave power could be adjusted between zero and 120 watts at a fixed frequency of 2450 MHz. After initiating the plasma with a Tesla coil, it was self-sustained inside the quartz Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 43 capillary when argon was used as the carrier gas. This plasma served as an excitation source for species eluted from the GC. As an analyte species passed through the plasma, excitation occurred with subsequent emission of radiation. The wavelength of the emitted radiation was specific for the species being excited. Emission was observed through a one millimeter hole drilled into the side of the quartz capillary. Light passing through this observation port propagated through a light pipe, based on the principle of total internal reflection, to the entrance slit of the spectrometer. The emission signal was detected by a photomultiplier tube and associated electronics. A summary of the instrumental components is presented in Table 1. Gas Chromatograph Control Panel A control panel was set up for all GC functions and mounted on an instrument rack. Controls for the oven, injection and exit port heaters, oven blower fan, and oven vent were included. Iron-constantan thermocouples were used to monitor column, injection, and exit temperatures. Additional thermostat adjustments were located on the rear of the GC. The zero control was used to set the low temperature limit to ambient conditions. The damping Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 44 Figure 1 GC-MES Experimental Set-up 1) 2) 3) 4) 5) 6) 7) 8) 9) 10) 11) 12) 13) 14) Gas chromatograph Argon carrier gas Drying column Rotameter Microwave power supply Microwave cavity Plasma capillary Light pipe Monochromator Photomultiplier tube HV power supply Electrometer Electronic filter Recorder Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. m Os Figure CN 00 CN Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 46 Table 1 Experimental Apparatus Excitation Microwave generator Raytheon, Model P6M-10X1 Microwave cavity ...... Tapered Monochromator Mount .................. Czerny-Turner modification Focal length ........... 1.075 m Gratings .... . 7500 A blaze, 1180 grooves/mm 4000 A blaze, 1180 grooves/mm Signal measurement HV power supply ....... EMI Gencom, Model 3000R Photomultiplier tube ... Hamamatsu R212 Electrometer ..... .... Keithly, Model 610CR Electronic filter ..... Spectrum, Model 921 Recorder ......... .... Esterline Angus, Model MS401B Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 47 control was set at its maximum clockwise position to minimize temperature fluctuations. The calibration control could be adjusted so that the potentiometer dial setting closely approximated actual temperature. Injection Port The injection port was designed to prevent the sample from coming in contact with any metallic surfaces. The body was fabricated out of brass with a quartz insert as shown in Figure 2. Carrier gas entered the injection port through a 1/8" Swagelock® to 1/8" NPT connection and flowed through to the column via the 1 mm hole in the side of the quartz insert. A constriction in the quartz insert minimized sample migration back into the dead volume adjacent to the septum. One end of the quartz insert was sealed with a septum (Supelco 2-0404) arrangement consisting of a 1/4" Swagelock® to 1/4" NPT connection. Soldered to the septum connection was a heat sink consisting of a circular metal disk and coil of 1/4" o.d. copper tubing. In high temperature GC applications/ water could be flushed through the tubing to protect the septum from deterioration. The opposite end of the injection port was sealed by modifying a 6 mm Swagelock® union. A union connection was required that would hold the quartz insert tightly against the septum in addition to Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 2 Injection Port Design (drawn to scale) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 49 o u u u G <U > o o u fl G CN Figure o -U rH 3 W •H H G u o 0000 OOOO 4J O. <D c/i Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 50 providing a gas tight connection to the column. To accomplish this, a hole was drilled in the block concentric with the quartz insert (0.499" diameter x 13 mm). .A 1/4" Swagelock® nut was turned down until its outside diameter was approximately 0.001" larger than the hole into which it fit. After the nut was cooled in liquid nitrogen, it was rapidly forced into the hole before warming. This resulted in a rigid, leak-free bond which required neither adhesives nor solder which could fail at high temperatures. A 6 mm Swagelock® union sealed with 6 mm graphite ferrules (Supelco 2-2493) completed the flow system to the column. Exit Port The exit port provided the connection between the GC column and the plasma capillary as shown in Figure 3. Graphite ferrules were used in all connections to quartz tubing and provided a leak-free seal at one quarter turn past finger tight. Under these conditions, the plasma capillary could still be removed with careful twisting. This removal was convenient for cleaning, optical adjustment, or replacement purposes since disassembly of the exit port was not required. Column Quartz tubing ( 6 m m o . d . chromatographic column. x4mmi.d.) was used for the Variable lengths could be used depending on separation requirements. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 3 Exit Port Design (drawn to scale) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 52 Figure 3 Plasma Capillary I Observation Hole \ \ \ V \ I Asbestos Fiberglass GC Insulation GC Column Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 53 Detection System Plasma Capillary The specifications for the plasma capillary tube were optimized in previous work done by Gebhart (77). Most important in these considerations were the 1 mm i.d. and. the 1 mm observation port. with a 6 mm o.d. The capillary was 25 cm in length The observation port was placed as close to the exit port as physically possible to minimize cooling and possible condensation of the analyte species. The resulting distance between the column exit and the observation port was 11.5 cm. Cavity Mount A tapered microwave cavity was mounted in a device which allowed vertical adjustment along the length of the plasma capillary. The cavity mount was attached to the GC as illustrated in Figure 4. By raising and lowering the cavity, the portion of the plasma observed could be varied. The observation height was defined as the distance between the observation port and the center of the cavity. Plasma Ignition The plasma was initiated with a Tesla coil mounted at the rear of the GC. A push-button switch on the instrument rack initiated the Tesla coil which provided seed electrons to the plasma along a wire between the coil and the plasma capillary. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 4 Microwave Cavity Mount (drawn to scale) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 55 Figure 4 v Microwave Cavity JZJ HF Asbestos Fiberglass GC Insulation Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 56 Light Pipe Optimal dimensions for the light pipe were determined previously in this laboratory (160). Light pipes were purchased from Wilmad Glass Co. with dimensions of 50 14 mm o.d. x 5 mm i.d. These pipes were cm x specified to be optically flat to 1/3 of a wavelength at 3000 A. The light pipe was supported at the mount illustrated in Figure 5. plasma end by the Adjustment of the set screws provided for alignment with the plasma observation port. The opposite end of the light pipe was held in position by an iris mechanism centered in relation to the entrance slit. An illustration of the iris mechanism and its mount to the spectrometer is given in Figure 6. A 1/4" Swagelock® union was soldered to this mount which provided a means of flushing the light pipe with an inert gas. Spectrometer Modifications The original spectrograph was not designed for scanning purposes. The conversion of the instrument to a scanning spectrometer was accomplished by modifying the grating rotation and scan mechanism from a Perkin Elmer 303 Atomic Absorption Spectrometer. The grating rotation unit was fastened to the optical shelf inside the spectrometer by three pairs of opposing set screws. an adjustment for grating tilt. These screws provided The rotation shaft extended Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 5 Adjustable Light Pipe Mount (drawn to scale) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 58 Figure 5 Top View Asbestos Fiberglass O Gas Chromatograph O Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 6 Light Pipe Support (drawn to scale) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 60 CD •H > +> a o M h +» •H r—I CO View O Eh u Side <D o m Q. CO vo □ □ o ■p Xt <D (7> 0 . ■H ‘H J 0. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 61 through the bottom wall of the spectrometer and was supported by a bearing assembly attached to an exterior shelf. This design provided true rotation with very little friction. The sine bar was attached to the rotation shaft below the exterior shelf. The gear assembly which moved the sine bar and a motorized scan mechanism were mounted to the bottom of the shelf. A 30 rpm motor provided forward and reverse wavelength scanning at approximately 1100 A/minute. A second 1/2 rpm motor could only be used to scan in the forward direction at approximately 15.6 A/minute.. Microswitches were positioned on either side of the sine bar to automatically shut off the drive mechanism at the scan limits. When either of these switches shut off the drive, the scan direction had to be be reversed and started manually. Wavelength could also be scanned manually by depressing the pin which disengaged the 1/2 rpm motor. Wavelength Readout The wavelength readout system was based on a voltage to frequency conversion circuit previously developed in our laboratory for peak integration (159). A block diagram for the readout device is illustrated in Figure 7. The objective was to convert a signal representing grating angle to a digital signal which could be calibrated to display in Angstrom units of wavelength. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 7 Wavelength Readout Device Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 63 Figure 7 •GND f out Ve R>' Rx $<■ wyw AD537 V/F CONVERTER in r\ — VNAAAA + 20 V MP 1029A FREQUENCY METER Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 64 A 25 turn, 0-100 k£2 potentiometer was mechanically attached to the grating rotation gear mechanism. Rotation of the grating resulted in a linear change in resistance (R^). When the proper voltage (V^) was applied between the stationary terminals of the potentiometer, the voltage drop to the slider terminal was linearly proportional to the grating angle. After this point the readout system was very similar to the integration system previously developed. An Analog Devices AD537 voltage to frequency converter integrated circuit chip was employed. Adjustment of controlled the slope of the curve relating readout to actual wavelength. Before attaching the potentiometer to the drive mechanism, R^ had to be adjusted to approximate the correct wavelength. Fine adjustment (±40 A) was made with R^' , a 0-500 k£2 potentiometer. R^1 could be set to give the correct readout for a standard wavelength in the region of interest. Calibration was then accurate to within a few Angstroms over a 500-1000 A range. Photomultiplier Tube Mount The instrument was originally built as a spectrograph with photographic recording and no exit slit. A mount was designed and fabricated from brass to hold a slit mechanism and either side mount or end-on PMT housings. A Hilger slit mechanism was placed into the brass mount illustrated in Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 65 Figure 8. The slit adjustment screw was lengthened so that it extended through the top of the mount. calibration is illustrated in Figure 9. Slit width A side mount PMT housing could be attached directly to the slit mechanism. Holes were included for mounting an end-on PMT housing. Optical Adjustment The gas chromatograph was supported by an adjustable mount. Half-inch bolts were secured at the corners of a base slightly larger than the bottom of the gas chromatograph. These adjustment bolts extended through a rectangular arrangement of one inch angle iron attached to the bottom of the GC. A pair of half-inch nuts on either side of the angle iron support provided independent vertical adjustment at the four corners. This arrangement permitted the entire GC to be moved up, down, or even tilted to align the plasma observation port with the optical path. Another modification of the spectrometer included adjustable supports to aid in alignment with the plasma. * Each end of the spectrometer was supported on an adjustable, vibration-absorbing mount as illustrated in Figure 10. The adjustment screws were positioned in a triangle, two for side-to-side tilt and a third for vertical displacement. Alternating layers of cork and rubber padding were used in the base to minimize vibrations. A summary of the alignment procedures are presented in Appendices A and B. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 8 PMT - Exit Slit Mount (drawn to scale) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 8 Slit Width Indicator (# turns open) Mounting Holes for End-on PMT (3) Slit Adjustment Extension J Purge/Vacuum Port Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 9 Slit Width Calibration Key: O Entrance slit • Exit slit Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 9 Slit Width, mm .50 .25 2.0 Number of Turns Open Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 10 Spectrometer Mount (scale 10mm = 1 inch) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 71 Figure 10 Monochromator I Adjustment 1 Screws r 1 Cork ^ Rubber -,L \ V k..k. \ >>■ \ N-.k. k ' -1 Wood Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 72 CHAPTER III RESULTS AND DISCUSSION Plasma Characteristics Plasma Background Spectrum The plasma background spectrum for argon carrier gas and argon with 5% helium is presented in Table 2. The second positive system of nitrogen was the most prevalent band system in the background spectra. The presence of significant amounts of nitrogen in argon plasmas has been reported (67). This nitrogen system is composed of close, triple-headed bands which are degraded towards shorter wavelengths. Several bands were also observed due to N£(+). These bands are single-headed and degraded towards shorter wavelengths. A final band system centered at 3360 A originates from the NH radical. Bands are degraded in either direction with the Q-branches forming a strong central maximum. Under low dispersion, the Q-maxima of the 0,0 and 1,1 bands are often mistaken for an atomic doublet (161). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 73 Table 2 Plasma Background Spectrum (3000 - 5000 A) X, A Assignment* 3063.6 OH 3090.4 °3 (0,0) 3097.5 3106.1 3136.0 °3 I Argon IAr+5%He ^'theor 19 - 10 69 5 8 34 - 4 24 - 5. N2 (2,1) 10 11 8 N2 (1/0) 19 47 9 3360.0 NH (0,0) 60 29 9 3370.0 NH (1,1) 90 >160 3159.3 3371.3 3469.0 3500.5 3536.7 3548.9 3563.9 3576.9 3582.1 3606.5 3641.7 N2 10 10 (0,0) (3,4) - 1 0 (2,3) - 3 4 N2 (1,2) 6 19 8 V V (3,2) 5 4 3 (2,1) 5 6 4 N2 (0,1) 42 127 10 N2+ (1,0) - 6 4 3 4 1000 3 3 3 ^2 N2 Ar N2 (4,6) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 2 (cont'd.). 3649.8 Ar 3671.9 3710.5 3755.4 3370.4 3834.7 3857.9 3884.3 3894.6 3914.8 3943.0 3947.5 3949.0 3998.4 - <1 800 N2 (3/5) - 1 6 N2 (2/4) 1 4 8 • N2 d/3) 4 12 10 - 1 400 14 ' 38 10 4 3 800 Ar 3804.9 - N2 Ar (0/2) N2 (4,7) - <1 5 V d/1) 2 3 3 N2 (3/6) - 1 7 N2 + (0,0) 2 12 6 N2 (2,5) - 1 8 Ar [ 9 L Ar 1000 » ' 9 N2 Ar d/4) (0,3) 4158.6 N2 Ar 4164.2 2000 2 4 9 21 25 1200 4 7 8 >58 120 1200 Ar 19 13 1000 4181.9 Ar 18 21 1000 4190.7 Ar 47 48 4044.4 4059.4 4191.0 Ar 600 1200 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 75 Table 2 (cont'd.). 4198.3 Ar 4200.7 Ar 4251.2 50 . 126 1200 >56 126 1200 Ar 7 8 800- 4259.4 Ar 40 43 1200 4266.3 Ar 28 29 1200 4272.2 Ar 42 42 1200 4300.1 Ar 33 34 1200 4333.6 Ar 34 34 4335.3 Ar 14 34 4343.6 N, (0,4) ' 800 - 12 1000 1 12 4 4345.2 Ar 4510.7 Ar 14 15 1000 4596.1 Ar 5 4 1000 4628.4 Ar 4 3 1000 4702.3 Ar 5 5 1200 J . 1000 * Assignments in brackets were unresolved. ** Molecular intensity values were obtained from reference 161. Atomic intensity values were obtained from reference 162. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 76 Spectroscopic Temperature Spectroscopic temperatures were based on the relative intensities of argon lines at 4259 A, 4345 A and 4511 A. These lines were reported to be free of self-absorption by Adcock and Plumtree (163). This method utilizes the fact that a plot of log (IX/gA) vs. E should be a straight line that is inversely proportional to temperature. Spectroscopic temperatures were determined for a variety of carrier gas compositions: Argon 4300 K Argon + 5% Helium 4500 K Argon + 9% Helium 6800 K Argon + 12% Helium 6100 K These results were obtained at an observation height of 12 mm below the center of the microwave cavity at a total carrier gas flow rate of 120 mL/min. The increase in temperature with helium content was expected due to the higher excitation potential of the helium metastable state (19.8 eV). A decrease in temperature was observed for helium concentrations greater than approximately 10%. This probably represents a limit for temperature enhancement with helium at 60 watts of applied power. When higher power levels were applied to the 12% He plasma, the plasma capillary began to glow due to the extreme heat. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 77 Detection of Hydrogen Gas In order to determine the feasibility of observing hydrogen emission in an argon microwave plasma, the background emission spectrum was recorded. The argon carrier gas was then doped with a trace amount of hydrogen gas. Figure 11 is a comparison between background argon plasma emission with and without traces of hydrogen. The main features of the plasma spectrum were atomic argon emission lines in the 4000 A to 4500 A region, atomic hydrogen emission at 4861 A and 6563 A, and second order molecular NH emission at 6720 A. The atomic hydrogen emission was from the Balmer series and has been used previously for spectrochemical analysis (43,94). The NH band emission at 3360 A has not been used previously for the analytical determination of hydrogen. This molecular emission has been used for the determination of NH^ in argon, however (43). This emission was investigated in the interest of improving sensitivity for hydrogen. Weak NH emission was observed in the background argon spectrum as well. By observing the plasma through the capillary wall, emission due to hydrogen and nitrogen impurities in the carrier gas were selectively observed. Since argon purity was specified to be 99.998%, NH emission was negligible. The majority of the background NH emission observed at the observation port was assumed to be due to Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 11 Argon Plasma Spectra a) Argon carrier gas b) Argon + trace hydrogen gas Grating Slits Ar flow Observation ht. Power 4000 A 0.5 mm 120 mL/min. -12 mm 70 watts Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 79 Figure 11 a) 4000 5000 6000 7000 o CM 4000 5000 6000 7000 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 80 atmospheric hydrogen and nitrogen. This contribution led to a small, positive background interference which was found to be very constant. It was believed that hydrogen passing through the chromatographic column must combine with atmospheric nitrogen at the observation port. This was confirmed by injecting trace quantities of hydrogen gas and comparing emission through the wall and through the observation port. A much larger chromatographic peak corresponding to hydrogen was observed through the port. On this basis, observation of t h e N H band was investigated for the analytical determination of hydrogen gas. Separation Even though the MES system is element specific, baseline emission is affected by large amounts of any species passing through the plasma. plasma is actually extinguished. In extreme cases, the Hydrogen, therefore, had to be separated from the rest of the constituents of the breath sample. Narrow peaks and a minimum separation time were desired. Molecular sieves have been used extensively in the separation of permanent gases. Types 5A and 13X have both been used since separation is not actually based on a sieving mechanism but on electrostatic interactions (164). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 81 A molecular sieve 13X column will separate, in order, hydrogen, oxygen, nitrogen and methane, with carbon dioxide being completely adsorbed (165). Hydrogen was very efficiently separated from the rest of the gases. A 60/80 mesh size was chosen since flow rates required for maintenance of the plasma were.not possible with smaller sized particles. This problem was assumed to be due to fines which were difficult to remove. Separation was accomplished at ambient temperatures. Flow rate was optimized with respect to peak height, shape and time of analysis. separation parameters. Table 3 is a summary of the optimized Under these conditions, the retention times for hydrogen, oxygen and nitrogen were 24, 42 and 72 seconds, respectively. Detector Optimization To maximize sensitivity, microwave power as well as observation height required optimization. These two parameters were important because of their relation to excitation temperature. Since plasma length increases with power, the optimum height of observation also changes. In order to maximize both power and observation height, these parameters were investigated simultaneously. Detector response to a standard injection of hydrogen was observed at various positions along the length of the plasma for Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 82 Table 3 Operating Conditions for GC column dimensions ..... 6 ft. x 6 mm o.d., 4 mm i.d. GC column packing mesh molecular sieve 13X .... 60/80 Carrier gas ..... ........ Argon Column temperature ...... . Plasma c a p i l l a r y (99.998%), 120 mL/min. Ambient, 20 - 25°C 26 cm x 1/4 in. o.d., 1 mm i.d. Plasma observation hole .. 1 mm diameter, 12 cm from column Microwave power output ... 60W forward, <1W reflected power Observation height ...... -12 mm from cavity center Light pipe dimensions .... 50 cm x 14 mm o.d., 5 mm i.d. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. different power settings. Results for atomic hydrogen emission at 4861 A and second order NH emission at 6720 A are illustrated in Figures 12 and 13, respectively. Atomic emission appears to increase with applied microwave power with maximum emission at 80 watts and a distance of 12 mm from the center of the cavity. A large amount of energy is required to excite atomic hydrogen emission. The profiles for NH emission are similar except at high power levels, where emission falls off drastically toward the center of the cavity. This may be due to the fact that conditions are too energetic for the formation of the NH radical. Wavelength Calibration Wavelength calibration for the determination of metal species is very straightforward due to the availability of hollow cathode lamps. Obtaining wavelength standards for nonmetals or gases is a different problem. As mentioned previously, hydrogen emission lines could be obtained by doping the argon carrier gas with hydrogen. Unfortunately, this approach resulted in contamination of the chromatographic column. After wavelength calibration, it took two to three hours for the hydrogen signal to return to a normal baseline. An alternate approach was devised that by-passed the chromatographic column completely. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 12 Detector Optimization - (4861 A) Grating 4000 A blaze Slits 0.2 mm Ar flow 120 mL/min. Injections 0.2 mL x 16 ppt Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 85 Figure 12 30 25 15 20W Peak Height, mm 20 40 W 60 W 80 W 10 5 50 Distance from Cavity Center, mm Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 13 Detector Optimization. - H 2 (6720 A) Grating Slits . Ar flow Injections 7500 A blaze 0.2 mm 120 mL/min. 0.2 mL x 98 ppm H 2 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 87 Figure 13 30 20 20 W 60 W 80 W Peak Height, mm 25 IQ- 20 30 40 Distance from Cavity Center, mm Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 50 88 An argon/trace hydrogen mixture was added through the top of the plasma capillary tube. The total flow rate through the top closely matched the column flow rate. This method was possible only because of the observation port in the plasma capillary. Once the optimum wavelength was found, the connection to the top of the plasma was removed and analysis could begin immediately. Calibration of Response A calibration gas mixture was purchased from Matheson (H77-252). This standard was specified to be 98 ± 2 ppm hydrogen gas in nitrogen. The disposable cylinder was equipped with a syringe adapter (H77-901) so that samples could be transferred directly to the injection port. To calibrate the detector, the 98 ppm hydrogen calibration gas was injected in amounts varying from 0 to 0.2 mL. Sharp, reproducible peaks were obtained, as shown in Figure 14. Breath concentrations based on the calibration curve ranged from 0 to 10 ppm hydrogen for a 2 mL injection. A representative calibration curve for hydrogen at 3360 A is illustrated in Figure 15. Peak height was used as a measure of response. The use of variable injection volumes and peak height measurement was considered valid because of minimal peak broadening. This assumption was tested by preparing a range of dilutions Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 14 Hydrogen Calibration Peaks Wavelength Concentration Slits Ar flow Observation ht. Power 6720 A 98 ppm H, 0.2 mm 120 mL/min. -12 mm 60 watts Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 14 Figure Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 15 Hydrogen Calibration Curve Wavelength 6720 A Correlation, r .9997 Slope, m 2.8 Intercept, b -2.2 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 92 Figure 15 50 40“ -c O) *« 30- 20- 10- 0 2 6 4 M , 8 ppm by volume Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 10 93 of the 98 ppm hydrogen standard. The corresponding calibration using variable concentrations was found to be identical to using variable injection volumes. The only difference seemed to be the larger error associated with preparing dilutions. Evaluation of Hydrogen Detection The primary requirement for the detection of hydrogen eluting from the chromatographic column was adequate sensitivity. Since breath hydrogen concentrations range from 1 to 200 ppm, sensitivity and linearity over this range were essential. Selective detection for hydrogen was actually unnecessary because it was separated chromatographically. Increased selectivity for hydrogen over other breath components would simplify the chromatogram, however, and ease separation requirements. Since changes in hydrogen excretion were measured relative to baseline levels in the hydrogen breath test, precision was more important than absolute accuracy. Finally, a rapid detection method was desired for fast sample turn-around. Sensitivity and Detection Limit One of the main purposes of developing the GC-MES method for the determination of hydrogen was to improve on the sensitivity of existing methods. The parameters that Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 94 had the greatest effect on sensitivity were wavelength setting accuracy and observation height in the plasma. Other parameters which improved sensitivity to a smaller extent included decreasing carrier gas flow rate, increasing microwave power, and opening the slits. Slit width was limited by the amount of background plasma emission. One of the major reasons for observing NH emission was to improve sensitivity. Sensitivity for this molecular emission was found to be approximately one order of magnitude better than atomic emission at 4861 A. Table 4 is a comparison between analytical results for hydrogen in breath at 4861 A vs. 6720 A. Note that the slope (m) of the calibration curve is much greater for the molecular NH emission. Results for the two wavelengths are identical within experimental error. atomic hydrogen is 13.6 eV. The ionization potential for This is greater than the argon metastable levels at 11.49 and 11.66 eV. Due to the fact that atomic hydrogen emission was observed in the argon plasma at all, a mechanism other than the one involving ionization by argon metastables must be possible. However, the high ionization potential could explain why atomic hydrogen emission was not as intense as the NH emission. Increasing the slit widths to improve sensitivity was much more beneficial when observing NH. Because of the inherent band structure associated with molecular emission, slits could be opened up to increase the bandpass. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 95 Table 4 Comparison Between Analytical Lines for Hydrogen Calibration data: 4861 A 6720 A .9924 .9969 slope, m 0.5 2.9 intercept, b 1.5 -0.6 correlation, r Breath sample concentration (ppm) Sample 1 5 ± 4 7.3 ± .5 2 10 ± 4 11.0 ± .5 3 14 ± 4 13.8 ± .5 4 18 ± 4 17.5 ± .5 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 96 Another method of increasing detection was to increase the volume of sample injected. The major limitation to the injection volume size was peak broadening. This would have been a problem since quantitation was based on peak height. Small amounts of helium were added to the argon carrier gas in another attempt to increase sensitivity. For argon containing approximately 10% helium, large, positive interferences were observed when air, nitrogen or argon samples were injected. In these cases, the plasma length increased at the hydrogen retention time. The signal, therefore, was probably due to an increase in background as the plasma expanded and contracted. The reason for this behavior specifically at the hydrogen retention time could not be explained. Since this phenomenon was observed only when helium was added, pure argon carrier gas was used thereafter. Detector sensitivity was specified by determining the minimum detectable level (MDL). This is the "level" of sample in the detector at the peak maximum, when the signal to noise ratio is 2 (166). For the determination of hydrogen at 3360 A, the MDL was found to be 15 pg hydrogen/second. This specification can be used for comparison to other GC detectors because it is independent of column parameters and sample size. The absolute detection limit for hydrogen was found to be 0.1 ng. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 97 Hydrogen concentrations in this text are given as-ppm hydrogen by volume in air (v/v). These concentrations would be approximately one order of magnitude lower if presented as the weight of hydrogen per volume of air (w/v). Selectivity Observation of characteristic emission results in excellent selectivity. ’When molecular NH emission was monitored, however, a positive response was observed for nitrogen as well as for hydrogen. This "spectral interference" was due to the presence of an intense molecular nitrogen bandhead at 3371 A. Better selectivity was possible when observing second order NH emission at 6720 A due to increased dispersion. The response to nitrogen did not result in a positive interference for hydrogen determination because of the chromatographic separation. Hydrogen response was observed and recorded before* the nitrogen peak eluted from the column. The only consequences of this response to nitrogen were increased analysis time and slightly elevated background emission. Since breath samples are mainly composed of nitrogen, a very large peak was observed. It took approximately five minutes for the signal to return to a normal baseline. At this time another sample could be injected. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 98 Greater selectivity was observed at the 4861 A atomic hydrogen line. t This was due to the absence of nitrogen band emission in this region. The nitrogen did, however, extinguish the plasma for a short period of time. After approximately two to three minutes, the plasma was re-ignited and the next sample could be run. Hence, by observing atomic hydrogen, analysis time could be cut in half at the expense of sensitivity. Accuracy An analytical accuracy of 10% was desired for this method. Absolute measurements such as ambient hydrogen levels or baseline human breath levels could then be compared with literature values. To monitor the accuracy of the hydrogen determinations, a quality assurance (QA) standard was run periodically as an unknown. The QA sample was prepared in a 2 liter volumetric flask. A rubber stopper was fit with a septum and length of Tygon® tubing which could be clamped. The total volume of the system with the stopper and clamp in place was calculated to be 2.06 liters. The flask was evacuated and flushed with nitrogen three times. Atmospheric pressure was approximated by opening the clamp and slowly releasing nitrogen until leakage stopped. Injecting 0.1 mL of hydrogen gas into the flask resulted in a calculated Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 99 concentration of 49 ± 1 ppm hydrogen. Error propagation revealed that uncertainty in this calculation was almost entirely due to the inaccuracy of the hydrogen injection volume. The QA sample was treated as an unknown and was analyzed along with each sample set. Using injections of the 98 ppm standard for calibration, an average concentration of 50 ppm was determined for the QA sample. This represented a relative deviation of 2.5% from the calculated value. Linear Range To determine the linear range of detector response, hydrogen standards of various concentrations were prepared. To achieve a complete range of absolute levels, injection volumes were also varied. Since the range of hydrogen levels extended over four orders of magnitude, electrometer and recorder settings had to be adjusted to keep the response on scale. Correction factors were required to account for nonlinearity in these settings. The linearity of hydrogen determination at the second order NH emission band is illustrated in Figure 16. The linear range extended over approximately two orders of magnitude. Using 2 mL injection volumes, this range represented breath hydrogen concentrations from 0.5 to 50 ppm. The linear range for Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 100 atomic hydrogen emission at 4861 A ranged from 10 to 1000 ppm. Reproducibility The precision of the analytical determination of hydrogen was actually more important than accuracy. By definition, the hydrogen breath test monitors the increase in concentration above baseline, a relative measurement. A reproducibility of 10% was desired in these studies. The short term precision was determined by injecting replicate amounts of the 98 ppm calibration standard. chromatograms are illustrated in Figure 17. These Plasma instability and electronic noise contributed to the 3% relative standard deviation (RSD) of the analytical method. Results were not expected to be as precise for larger sample injections and smaller concentrations. Long term or day-to-day reproducibility was determined by comparing results for the quality assurance standard. The average concentration was found to be 50 ± 3 ppm hydrogen, which represented a relative error of 6%. Short-term reproducibility for the quality assurance samples averaged 1.5% RSD. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 101 Figure 16 Hydrogen Linear Range Wavelength Slits Ar flow Observation ht. Power 6720 A 0.2 nun 120 mL/min. -12 mm 60 watts Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 102 r -rr vo Log » CN H 2 Concentration, nl ■ cn CN uiui 'esuodssy 601 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 103 Figure 17 Reproducibility of Standard Injections Wavelength Slits Ar flow Observation ht. Power 6720 A 0.2 mm 120 mL/min. -12 mm 60 watts Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ZF A 7 CN Figure 17 Nf CO a* CN NT CO CO O. y O' J Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 105 Hydrogen Breath Test Procedure Subjects were required to fast overnight before the test. Breath samples were taken initially to determine baseline hydrogen levels. Lactose was then administered as a suspension in 100 mL of water. 0.6 g lactose per kg body weight. The dose was measured at Duplicate breath samples were taken at half-hour intervals for 3 to 4 hours. Samples were analyzed the same day. Breath Collection The method used in collecting a breath sample depends on the sensitivity of the analytical method. Collection methods were also evaluated with respect to reproducibility and degree of difficulty for the subject. Previous investigators used a "rebreathing" technique which concentrated hydrogen levels in the breath sample (136). This method involves breathing the same volume of air for a period of 2 to 4 minutes. All of the hydrogen excreted during this time is concentrated in a volume of air equal to the sum of the subject's lung volume and the volume of the collection system. Rebreathing for 2 minutes resulted in consistent breath hydrogen levels but required an extreme effort on the part of the subject. It was decided that many multiple sclerosis Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 106 (MS) patients would not be able to undergo this type of testing because of their condition. As detection limits for the determination of hydrogen were improved, concentration of the breath hydrogen was unnecessary. In a second collection method, the subject was instructed'to breathe deeply four times to clear any residual hydrogen from the lungs.' After the fourth breath was used to flush out the collection bag, the sample was obtained. Using this hyperventilation technique, a group of MS patients was tested and showed very little response to the hydrogen breath test (HBT). In a third breathing method, the lungs were never cleared out by hyperventilation. The subject exhaled normally, after which the residual lung volume was collected with a maximum exhalation effort. This breath collection technique has been described in the literature as an end-expiratory sample(152). Hydrogen concentrations were expected to be higher in the residual lung volume. The final breath collection mebhod was similar to the interval sampling techniques described in the literature (144). A normal exhalation was collected without rebreathing, hyperventilating or clearing the residual lung volume. The obvious advantage of this method was that it required no special effort on the part of the subject. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 107 Collection Apparatus Three liter breath collection bags were constructed of Mylar®-coated foil. Samples were collected by breathing through a short length of Tygon® tubing attached to the bag. These bags were designed and supplied by Dr. D. H. Calloway, Department of Nutrition, University of California at Berkeley. Samples containing trace amounts of hydrogen were shown to be stable in these bags for 47 days (142). During the HBT, approximately twenty samples were collected from each subject. For convenience in transport and storage, samples were transferred to 50 mL syringes. These syringes were lubricated with ethylene glycol, rendering them air-tight. The sample was drawn into the syringe through a septum arrangement after the Tygon® tubing had been clamped off. Between sample collections, the bags were flushed twice with dry nitrogen to remove residual hydrogen and moisture. The efficiency of this flushing procedure was determined for a bag which originally contained 80 ppm hydrogen. The second nitrogen flush was stored in the collection bag for 10 minutes. The resulting hydrogen concentration was below the 0.5 ppm detection limit. Because of the large moisture content of human breath, its effect on hydrogen concetration was studied. A small filter containing absorbent paper was used in this study. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 108 The filter (DuPont P101) was originally used to protect personal air sampling pumps from dust and moisture. After the absorbent paper was saturated with moisture, the 98 ppm calibration standard was drawn through the filter and injected on the column several times. Response was compared to samples taken directly from the calibration gas cylinder. Response for the "dry" samples was found to be 43 ± 1 mm while "moist" samples averaged 43 ± 3 mm. equivalent within experimental error. Results were The greater standard deviation observed for the "moist" samples was possibly due to the effect of moisture on the chromatographic column. Sample Storage Sample storage in the 50 mL gas-tight syringes was found to be both convenient and stable. When the syringes were stored in a rack, tip down, a positive pressure was placed on the sample. out of the syringe. Any sample leakage would therefore be The ethylene glycol lubricant did not interfere with either the storage or determination step. The stability of breath samples stored in the syringes over a period of 18 days is presented in Table 5. When breath sampling was performed during the hot summer months, samples were transported under high temperature conditions. The effect of high temperature on sample storage stability was investigated. Dilute hydrogen Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 109 Table 5 Storage Stability for Breath Samples Storage Hydrogen concentration (ppm) Breath sample 1 (days) Breath sample 2 0.3 4.7 ± 0.8 5.0 ± 0.8 1.0 5.9 ± 0.8 5.6 ± 0.6 1.4 5.1 ± 0 5.0 ± 0.9 6.0 5.8 ± 0 5.8 ± 0 11.0 6.0 ± 0.5 5.6 + 0.7 14.0 5.7 ± 0.5 5.5 ± 0.5 18.0 4.3 ± 0.3 5.1 ± 0 Average 5.4 ± 0.7 ppm 5.4 ± 0.3 ppm RSD 6% 13% Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 110 samples were prepared and distributed between two sets of syringes. One set was placed in an oven at 42°C. The remaining set, initially identical in concentration, was stored at a room temperature of approximately 27 °C. The results of this study for a one week storage period are presented in Table 6. No significant differences were found between the two sets of samples as long as they were allowed to cool to room temperature before injection. Breath Test Applications Lactose Studies .The hydrogen breath test (HBT) was used to diagnose lactose malabsorption in a group of subjects with multiple sclerosis (MS). This portion of the study was initiated and set up by three faculty members from the University of Iowa, College of Nursing. Toni Tripp Reimer, Laura Hart and Mildred Freel authored the initial proposal to the Johnson County MS Society (125). These nurses took responsibility for background research, recruiting and scheduling subjects, as well as collecting the samples. Breath collection and analytical methodologies were developed as part of this doctoral research. Ambient Hydrogen Levels Each time a new set of tests were run, air was collected from the room to determine background hydrogen Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Ill Table 6 Storage Stability vs. Temperature Time 27°C 42 °C Sample 2 27°C 42 °C Sample 3 27°C 42 °C 0 1.1 1.1 2.6 2.6 5. 9 5.9 6 1 .6 1.5 2.4 2.8 6.5 6.6 20 1.3 1.3 to • o (hours) Sample 1 2.8 6.1 6.0 168 1.7 1.1 2.8 2.1 5.4 5.7 AVE ± SD 1.2 ± 0.3 2.5 ± 0.3 6.0 ± 0.4 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 112 levels. The results for room hydrogen concentrations are presented in Table 7. to be 1.8 ± .5 ppm. The average concentration was found These hydrogen levels varied less than 0.4 ppm during a single test period. Outdoor air was found to contain less hydrogen. average concentration of 0.8 ppm was measured. An This value compares favorably with values reported in the literature for atmospheric hydrogen concentration (167,168). The difference between indoor and outdoor hydrogen levels could have been due to experimental error. Rebreathing vs. Hyperventilation Table 8 is a comparison of HBT results (ppm hydrogen) between rebreathing and hyperventilation breath collection methods. Subject 1, the normal absorber, was actually classified as a malabsorber when rebreathing was used. Since breath hydrogen was concentrated by this method, the 20 ppm increase criterion for positive diagnosis was invalid. Hyperventilation resulted in breath hydrogen * levels proportional to rebreathing, but much lower. Notice that peak hydrogen excretion occurred between 1.5 and 2 hours for subject 1. Again, levels obtained with rebreathing were much higher. There were two differences between the HBT's for subjects 1 and 2. The malabsorber excreted larger amounts Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 113 Table 7 Ambient Hydrogen Levels Date Room Hydrogen Concentration (ppm) 6/24 2.5 3.8 1.8 1.5 1.8 6/25 1.7 1.6 1.-9 1.8 1.8 2.3 6/26 2.3 1.7 1.3 1.3 1.5 1.7 6/27 1.7 2.9 1.8 1.5 1.5 6/30 1.5 1.5 1.8 1.8 7/7 2.2 2.0 2.3 2.7 10/2 1.0 1.2 1.9 1.2 10/21 1.4 1.3 1.4 1.4 10/22 1.6 1.7 1.5 1.4 11/20 1.2 1.2 2.3 1.1 2.3 2.5 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 114 Table 8 Comparison between Rebreathing and Hyperventi1ati on Hyperventilation Subject 1* Subject 2** 47 26 2 7 0.5 139 20 10 5 1.0 170 - 15 4 1.5 171 43 16 13 to • o (hours) Rebreathing 170 130 18 29 to U1 Time 102 125 10 39 3.0 59 290 7 67 B1 Subject 1 Subject 2 * Subject 1 was a normal absorber ♦♦Subject 2 was known to have lactose malabsorption and MS Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 115 of hydrogen, as expected, and complained of some discomfort. The unexpected observation involved the peak time of hydrogen excretion. The breath hydrogen concentration for subject 2 was still increasing after 3 hours instead of reaching a maximum at 1.5 to 2 hours. Lactose Malabsorption in MS Patients It was decided that the hyperventilation technique could be used to distinguish between lactose absorbers and nonabsorbers. tested. An initial group of ten MS patients was HBT results are presented in Table 9. subjects were shown to be lactose malabsorbers. None of the It was also significant that none of the subjects complained of discomfort during the testing period. Residual Lung Volume At this time the residual lung volume collection technique was developed to concentrate the breath hydrogen somewhat. Results from Table 10 show that this technique is also capable of distinguishing between lactose absorbers and nonabsorbers. However, there were still no conclusive results for MS subjects. The hyperventilation and residual lung volume techniques could be compared since subject 2 in Tables 8 and 9 was the same person. Breath hydrogen concentration in the residual lung volume was significantly higher than for hyperventilation. It is also important to Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 116 Table 9 Hyperventilation Collection Method Subjects* 4 5 1 2 3 7 8 9 10 B1 4 3 5 4 2 2 5 10 6 2 B2 4 4 5 3 2 3 4 8 4 2 0.5 3 6 '5 4 2 2 4 6 4 2 1.0 2 5 4 5 2 2 5 6 4 2 1.5 3 4 4 5 2 2 6 5 3 3 2.0 4 5 4 6 2 2 6 4 3 3 2.5 7 4 4 6 1 2 5 3 5 3 12 4 4 6 2 2 3 3 5 2 3.5 - - 4 5 2 -.. - 4 - 3 5 5 2 CO 6 o Time (hours) 4.0 * All subjects were MS patients Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3 117 note that the hydrogen excretion did not follow exactly the same pattern for subject 2 in these two separate tests. In both cases, however, hydrogen levels were still increasing at 3 hours. Reproducibility of Breath Sampling Breath sample reproducibility was monitored for the tests using the hyperventilation collection method. The relative standard deviation (RSD) between duplicate determinations of the same sample averaged 7.8%. These results are compared to errors associated with the analytical method in Table 11. The increased error associated with the breath samples was due mainly to the increased injection volume required for low concentrations. It was much more difficult to consistently inject a volume as large as 2 mL. Since peak height was used for quantitation, slight peak broadening might have contributed to the loss in precision. Finally, the moisture present in breath samples also could have led to some instability, with respect to column performance. Relative standard deviation between different breath samples collected within minutes of each other was 8.4%. This additional loss of precision compared to the duplicate determinations was minimal. This seemed to indicate that the hyperventilation technique consistently collected representative samples. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 118 Table 10 Residual Lung Volume Time (hours) Bl 2 Subjects* 3 4 5 6 16 3 2 4 2 2 - 2 4 - - - 0.5 11 4 2 6 3 2 1.0 3 21 2 6 2 2 1.5 5 22 3 6 2 2 to • o B2 1 5 22 2 4 2 2 2.5 .4 28 2 4 2 2 3.0 3 37 2 3 2 2 ♦Subject 1 was a normal absorber Subject 2 was known to have lactose malabsorption and MS Subjects 3-6 were MS patients Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 119 Table 11 Reproducibility of Hydroden Determinations Standard hydrogen mixture (98 ppm) Injection volume 0.2 mL Number of determinations 10 Relative standard deviation (RSD) 2.8% Breath samples (average 4 ppm) Injection volume 2.0 mL Number of samples 240 RSD (same sample)* 7.8% RSD (duplicate samples)** 8.4% * duplicate determinations of the same sample ** deviation between duplicate breath samples Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 120 Lactulose Study Lactulose, 4-0-g-D-galactopyranosyl-D-fructofuranose, is an isomerization product of lactose. Humans do.not have an enzyme system capable of hydrolyzing lactulose into its monosaccharide components (131). Since it is very poorly absorbed, some investigators have used it as a standard to which partial absorption of other sugars is compared (134,149). In this investigation, lactulose was administered to a normal subject. Results of the HBT simulated the effect that lactose would have on a milk intolerant subject(130). The object of this study was to test the performance of two breath collection procedures as well as the analytical methodology. Lactulose was administered orally to a normal absorber. Cephulac® is a commercial laxative containing 10 g lactulose per 15 mL. The dose was 0.3 g lactulose per kg body weight. After a period of approximately one hour, the laxative produced belching and flatulence in the subject. Abdominal cramps were'absent, however. An example of a "typical” breath chromatogram is illustrated in Figure 18. The chromatogram is typical in the sense that retention times are shown, with oxygen extinguishing the plasma, and nitrogen response eventually returning to the baseline after re-ignition. The concentration of hydrogen in this chromatogram was very Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 121 high. Since such'a small injection volume could be used, the nitrogen peak returns to the baseline fairly quickly. For a 2 mL breath injection, the nitrogen peak would have been much larger. Figure 19 illustrates the HBT results for the lactulose study. Note that baseline hydrogen levels were- approximately 3 ppm. Maximum hydrogen excretion occurred 1.5 to 2 hours after the lactulose was administered and then fell off gradually. Hydrogen concentrations were larger in the residual lung volume. If the results of the HBT were interpreted in terms of a relative increase in hydrogen concentration, these two breath sampling procedures could give very different results. Another important difference found between the two breath collection techniques was the reproducibility. Collection of the residual lung volume seemed to be difficult for the subject to reproduce. Any change in the exhalation effort would result in a different breath hydrogen concentration. In contrast, collecting a reproducible normal interval breath sample was very easy. Method Validation To insure that results were consistent with those obtained by a universally accepted method, a collaborative Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 122 Figure 18 Breath Sample Chromatogram Wavelength Slits Ar flow Observation ht. Power 6720 A 0.2 mm 120 mL/min. -12 mm 60 watts Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 18 CO vo 3 minutes Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 124 Figure 19 Lactulose Study - HBT (dose = 0 . 3 g/kg body weight) Wavelength Slits Ar flow Observation ht. Power Key 6720 A 0.2 nun 120 mL/min. -12 mm 60 watts • Residual lung volume collection O Normal interval collection Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 125 Figure 19 70 60 <D E “I> 50 x _Q E 40 CL CL M 20 10- 2 0 3 Hours Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4 126 study was set up. A hydrogen breath test was run and duplicate samples were divided into separate sets. As shown in the section on reproducibility, very little error was associated with the breath collection procedure. Because of this collection consistency, hydrogen concentrations should have been nearly identical in the two sets of samples. One set was analyzed by GC-MES using the same conditions as previously described. The second set was sent to Mr. Kabir Younoszai at the University of Iowa Hospital, Department of Pediatrics, for analysis by GC with thermal conductivity detection. compared in Table 12. Results for the two methods are Note that measured concentrations were identical for the two methods when experimental error is considered. The average relative error between the two methods was only 2%. Conclusions The MES system proved to be a sensitive method for the determination of sub-ppm levels of hydrogen in breath. » Improved detection limits were obtained by observing molecular emission from the NH radical. The linear range was limited to about two orders of magnitude but was well suited for breath hydrogen concentrations ranging from 1 to 100 ppm. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 12 Validation of Hydrogen Method GC - TC GC - MES Sample [H2 1,ppm Sample [h 2 ],ppm Mean ± S.D. Rel. Error 1A 10 IB 9 ± 1 10 ± 1 7% 2A 7 2B 7 ± 1 7 ± 1 0% 3A 14 3B 14 ± 1 14 ± 0 0% 4A 14 4B 14 ± 1 14 ± 0 0% 5A 11 5B 11 ± 1 11 ± 0 0% 6A 17 6B 17 ± 1 17 + 0 0% 7A 25 7B 26 ± 1 26 ± 1 3% 8A 17 8B 18 ± 1 18 ± 1 4% AVERAGE 2% Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 128 The determination method was applied to hydrogen breath tests for lactose malabsorption. This study failed to show any correlation between lactose malabsorption and multiple sclerosis as proposed. The most significant finding was that a portion of the multiple sclerosis patients showed a delay in peak hydrogen excretion. This could explain the negative test results since breath hydrogen was not monitored for more than four hours. Further investigations on the digestive transit time of multiple sclerosis patients is called for. Hydrogen levels should be monitored until peak excretion is is observed. Even though the hydrogen breath test is simple and non-invasive, longer studies would probably require in-patient clinical studies. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. PART B ANALYTICAL APPLICATIONS OF GC-MES BELOW 2000 A Preliminary Investigations Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 130 CHAPTER IV INTRODUCTION The second part of this dissertation is devoted to spectrochemical analysis in the vacuum ultraviolet (VUV) region of the spectrum. This region was not used successfully for quantitative analysis until 1955 (169). Relatively few investigations have been conducted below 2000 A, primarily due to the difficulties encountered in observing emission in the VUV region (170). The main reason for limited use of the VUV region has been light absorption by atmospheric oxygen below 2000 A. To circumvent this problem, oxygen is removed from the optical path of. the spectrometer by evacuation or purging with a nonabsorbing gas. The reduced transmission of quartz below 2000 A requires the use of alternate optical materials such as magnesium flouride. Finally, t h e 'availability of detectors for VUV radiation is much more limited than in the UV-visible region. Several factors suggest that utilization of the VUV for spectrochemical analysis would be very advantageous. The resonance lines of numerous elements (20%) are found in the VUV region of the spectrum. Greater sensitivity is often Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 131 possible when observing these transitions as opposed to lines occurring at longer wavelengths (171). Determination of elements such as Br, Cl, N, 0 and S in the UV-visible region is actually very limited. The absence of molecular emission from atmospheric gases and carbon species in the VUV greatly reduces background emission from excitation sources. Recent Analytical Applications. Analytical investigations have been concerned mainly with the development and characterization of new excitation sources. Ellenbracht et al. (120) observed sulfur resonance lines from a dc arc for aqueous solution analysis. The 0.5 ppm detection limit did not vary for different sulfur species. An argon purge system was used to reduce light absorption by atmospheric oxygen., In a subsequent paper, these authors used a dc discharge plasma as a sulfur specific detector for gas chromatography (122). This high-temperature, high-energy plasma minimized solvent quenching effects and was capable of handling large injections without extinction. A detection limit of 300 pg/second and a selectivity ratio of 103 were reported. Dreher and Frank employed an inductively-coupled radio-frequency emission source for VUV spectroscopy (172). Kirkbright et al. (173,174) have investigated ICP excitation Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 132 of S # P, I, As, Se and Hg at wavelengths down to 1700 A. Detection limits for aqueous solutions were in the ppm range. Heine et al. (170) reported preliminary work on the use of the VUV region for GC-ICP. This system was capable of sensitive detection of oxygen and nitrogen emission, unlike previous investigations in the UV-visible region (124). The only VUV observation involving GC-MES was reported by Braun et al. (82). Their goal was to obtain total atomic content of organic compounds by measuring resonance and non-resonance atomic transitions. Since essentially complete fragmentation could be realized in the reduced pressure helium plasma, response was linear with atom content in the molecules. With this is mind, the authors proposed that response calibration curves for a particular element could be obtained for a stable and easily stored compound and would be valid for any compound containing that element. Braun also reported that the choice of emission lines could affect the linearity of the signal. Emission from lines terminating in the ground state could be strongly reversed due to self-absorption by ground state atoms. This problem, however, could have been due to the fact that the discharge was 10 mm in diameter and was viewed in an end-on configuration. Resonance lines for oxygen (1303 A) and Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 33 hydrogen (1216 A) could not be observed because of the background emission due to impurities in the carrier gas. Purpose of Study The purpose of the following investigation was to capitalize on the sensitivity and selectivity possible by observing the VUV region of the spectrum. In combination with the GC-MES system previously described in this dissertation, increased. sensitivity and selectivity could be further Previous investigators have not attempted to optimize MIP detection in this region. In this study, detection has been optimized for both iodine and sulfur emission from organic compounds. Resonance transitions for these two lines have been investigated. In addition to this preliminary work, qualitative results for bromine, carbon, iodine, nitrogen, phosphorus and sulfur were obtained. These results are presented as emission spectra in Appendices C - G. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 134 CHAPTER V EXPERIMENTAL The vacuum spectrometer described in this work had not been used for approximately eight years. Attempts to produce a hard vacuum were unsuccessful. Since the leaks could not be identified, an alternate approach to removing oxygen from the system was investigated. Purging the Optical System Connections were made to the spectrometer to purge the optical system with an inert gas. The port at the top of the spectrometer, originally used for the cold-cathode vacuum gauge, served as an inlet for the purge gas. The diffusion pump was removed and replaced with an aluminum plate. A 1/4" Swagelock® connection attached to this plate was used as an outlet for constant flow purging. Gas flow both in and out of the spectrometer were controlled with rotameters. In order to remove oxygen from the photomultiplier tube (PMT) mount, the exit window was removed from the spectrometer. Vacuum tubing was clamped to the port extending from the bottom of the PMT mount. The Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 135 spectrometer was evacated by attaching this tubing to a mechanical vacuum pump. When not in use, the vacuum tubing was plugged with a rubber stopper. This stopper served as a safety release when gauge pressures exceeded approximately ten pounds per square inch (psig). Spectrometer purge gas pressure was measured by. inserting a gas chromatographic inlet pressure gauge directly into the vacuum tubing. Oxygen also had to be removed from the light pipe optical path. Tubing was attached to the mount positioning the light ..pipe at the entrance slit. Gas flow through the light pipe was controlled with a rotameter. Detection An EMI Gencom (G-26E315) solar blind PMT and its side-on housing (B-215FV) were attached to the exit slit mount. The maximum response of the Csl photocathode at 1300 A falls off sharply at 1100 A. Toward longer wavelengths, response decreases by one order of magnitude at 1800 A and two orders at 2000 A. High voltages of 2 to 3 kV were required for operation. Under certain vacuum conditions, an arc discharge between the photocathode and the nearest conductor outside the tube was possible. This is not a problem when the tube is used under hard vacuum conditions as designed. The breakdown voltage where arcing occurs depends on Paschen's Law. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. There 136 is a range in the pressure-distance curve where breakdown voltage is at a minimum. This minimum was found to correspond to pressures of approximately 0.01 mm to 1000 mm of Hg outside the tube. When attempting to operate the PMT under mechanical pump vacuum, violent arcing was observed. In the purging studies approximating atmospheric pressure, small amplitude "spikes" were recorded.' Since these small discharges occurred up to ten times a minute, the use of a sensitive v electrometer setting was not possible. The arcing problem was solved by pressurizing the spectrometer with the purge gas. At pressures greater than five psig, discharging was not observed. A flow rate of approximately 100 mL/min He into the spectrometer was necessary to maintain this pressure.. The pressure did not build up due to leaks in the system. It.was possible to reduce dark current noise by cooling the PMT housing. A cold-finger type arrangement was set up by wrapping copper foil around the housing. The end of the copper foil extended down into a Dewar of liquid nitrogen. By insulating the copper foil, the temperature of the PMT housing could be reduced to approximately 0°C. A substantial decrease in noise was observed at an electrometer setting of 1x10 "9 Amperes. Cooling was unnecessary for attenuations of 1x10 "• amperes and above. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 137 CHAPTER VI RESULTS AND DISCUSSION Oxygen Removal Two gases were studied with respect to purging efficiency. Nitrogen was thought to be transparent down to approximately 1500 A. The spectrometer was evacuated and then filled with nitrogen. When monitoring the 1849 A line from a mercury source, response did increase while purging with nitrogen. The spectrometer was pressurized with nitrogen up to a few psig and response at 1849 A decreased. This was probably due to oxygen impurities in the nitrogen. Argon background emission was scanned between 0 and 2000 A. The only emission observed was a weak band extending from approximately 1850 to 1900 A. Response was not observed at wavelengths below 1850 A. Better results were obtained when purging the optical system with helium gas. The spectrometer was first evacuated, then pressurized with helium. The process of releasing the helium and subsequent pressurization was repeated several times. Additional evacuation would allow room air to leak into the system. Pressurization of the spectrometer with helium did not seem to reduce transmission Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 138 of the mercury 1849 A line. This was probably due to the low oxygen content of the helium, specified to be less than 1 ppm. The argon background emission spectrum was observed below 2000 A. The band emission observed previously between 1850 and 1900 A was absent. Atomic emission lines were observed at 1657 A, 1740 A and 1931 A. The 1657 and 1931 A lines were due to atomic carbon emission, probably from carbon deposits on the inner wall of the plasma capillary. Carbon emission could also have been due to slight bleeding of the organic stationary phase. The plasma emission line at 1740 A was due to several atomic nitrogen lines between 1740 and 1745 A. Apparently, the plasma was able to fragment molecular nitrogen impurities in the argon carrier gas. This was not expected due to the high bond dissociation energy for nitrogen (226 kcal/mole). A positive response was observed for small injections of atmospheric nitrogen at this wavelength. Detection of Iodine Emission Wavelength Calibration In order to calibrate the wavelength setting for atomic iodine emission, 1-iodobutane was added to argon gas flowing through the top of the plasma capillary. The apparatus used for this purpose is illustrated in Figure 20. This method Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 139 for doping the plasma was useful for volatile compounds. The amount diffusing into the plasma could be controlled by varying the argon flow or the temperature of the sample tube. Screw top test tubes were used to store a variety of sample compounds. These could be attached for wavelength calibration and then easily removed. The most prominent iodine emission line in the VUV region was the atomic resonance line at 1830 A. Many other intense iodine lines were observed in the region between 1600 A and 2100 A. Wavelengths and relative intensities of the observed iodine transitions are listed in Table 13. Method Development Detector Optimization Detector response at 1830 A for 2 yL injections of 500 ppm (w/v) 1-iodopentane was observed at various positions along the length of the plasma. Results for this observation height optimization at various power settings is illustrated in Figure 21. Response for this ground state transition fell off drastically toward the center of the cavity at 80 watts. This was probably due to the increased population of upper level excited states not involved in this transition. The most interesting result when optimizing height was the appearance of spikes in the response curves when Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 140 Figure 20 Plasma Doping Apparatus Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. •141 Figure 20 .Argon Sample Tube Heating Tape Plasma Observation Port Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 142 Table 13 Observed Iodine Transitions X, A Levels, 1/cm I * theoretical 1observed 1640.8 7603 - 68550 2500 1642.1 0 - 60896 2000 1702.1 7603 - 66355 15000 7 1782.8 0 - 56093 12000 >100 1799.1 7603 - 63187 5000 '>100 1830.4 0 - 54633 75000 >100 1844.4 7603 - 61820 15000 91 1876.4 7603 - 60896 2000 38 2062.4 7603 56093 M L J >100 * Intensity values were obtained from reference 175. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 143 Figure 21 Detector Optimization - Iodine (1830 A) Grating Slits Ar flow Injections 1900 A blaze 0.2 nun 100 mL/min. 2yL x 500 ppm 1-iodopentane Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 144 Figure 21 20n 60 w 40 w 15- E E 20 W 05 ’a> X 10- o 4) Q_ 80 W 5- i 10 20 30 40 Distance from C avity Center, mm Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. “ 1 50 145 observing the bottom tip of the plasma. This effect was very similar to the increased hydrogen response when helium was added to argon carrier gas. As a foreign species eluted from the column, the plasma contracted in length. When observing the lower portion of the plasma, it contacted to a point where the tip was even with the observation port. In this case, the plasma formed an "L" shape with the tip extending out of the observation port. Characteristic emission increased somewhat, but the majority of the emission was due to increased background. To support this theory, response was measured at a wavelength setting of 1820 A. The only response for 1-iodopentane was observed at the bottom tip of the plasma. In this region, response was observed as the plasma contracted. The background effect was much greater when helium was flowing through the light- pipe. The plasma plume normally extending from the observation port was minimized by large helium flow rates. As a result, when the plasma contracted to form the "L" shape into the port, the observation volume increased by a factor of 3. Observation \. of this increased emission at the lower end of the plasma was not analytically useful. Power and height optimization results for iodine emission at 1844 A are illustrated in Figure 22. Emission intensity increased towards the center of the plasma. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 146 Response also increased directly with applied power. Comparison of the power and height curves for iodine emission at 1830 A and 1844 A illustrates why these parameters must be optimized for each specific wavelength. The upper level, 1844 A iodine transition requires a higher excitation energy than the ground state transition. Separation of Organoiodine Compounds The chromatographic column dimensions were the same as those used in the hydrogen studies(6' x 6 mm o.d.). The column was packed with 5% OV-1 methyl silicone on 80/100 mesh acid washed Chromosorb W. Table 14 is a summary of the optimized conditions used for the separation and detection of organoiodides. Figure 23 is a chromatogram showing the separation of iodomethane, 1-iodobutane, 1-iodopentane and l-iodo-3-methylbutane. The mixture was prepared by sampling equal volumes of the vapor above the pure compounds. final injection volume was 0.2 mL. The Retention times were directly proportional to boiling point. Sensitivity and Detection Limit Sensitivity for iodine at 1830 A was determined by preparing a dilute standard of 1-iodopentane in petroleum ether. This standard contained 1 ng C ^ H ^ I per yL solvent or 1 ppm (w/v). The analytical conditions were identical to those previously specified. The minimum detectable limit Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 147 Figure 22 Detector Optimization - Iodine (1844 A) Grating Slits Ar flow Injections 1900 A blaze 0.2 mm 100 mL/min. 2 yl x 500 ppm 1-iodopentane Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 148 Figure 22 8n 80 w 60 W Peak Height mm 6-1 5-1 4H 20 W 3H 21 o 10 20 T " T " 30 40 Distance from Cavity Center, mm Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. — I 50 149 Table 14 Separation of Organoiodides GC Conditions Column dimensions 6' x 6mm o.d. Column packing 5% OV-1 on 80/100 Chromosorb W (AW) Argon flow rate 100 ml/min. Column temperature 130°C Inlet temperature 165°C Detector temperature 195°C Detector Conditions Observation height -14 mm Applied power 60 watts Slit width 0.2 mm Wavelength 1830 A Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 150 Figure 23 Separation of- .Organo iodides Key: a) b) c) d) Iodomethane 1-Iodobutane 1-Iodopentane l-Iodo-3-methylbutane Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 151 Response Figure 23 Minutes Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 52 was found to be 170 pg I/second for 1-iodopentane. It was not determined whether response varied between different organoiodides. The absolute detection limit for 1-iodopentane was 2 ng I . Selectivity Selectivity for iodine emission was measured by comparing detector response for 1-iodopentane vs. carbon tetrachloride. Selectivity is defined as the ratio of the molar amounts of iodine and carbon required to give an equivalent response. This selectivity ratio was found to be 7000 at 1830 A. Detection of Sulfur Emission Wavelength Calibration The same wavelength calibration method was used as described previously for organic iodine compounds. Carbon disulfide was doped into the plasma to produce characteristic sulfur emission. Because of the exteme volatility of carbon disulfide, the argon flow through the top of the plasma capillary was actually turned off. Diffusion of CSj into the plasma was sufficient at room temperature. Wavelengths and relative intensities of the observed sulfur transitions are listed in Table 15. The most intense emission was found for the resonance line at 1807 A. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 153 Table 15 Observed Sulfur Transitions X, A Levels, 1/cm I * theoretical 1666.7 9239 - 69239 1 25 52 1782.3 22181 - 78290 12 10 1807.3 0 - 55331 25 >100 1820.4 397 - 55331 25 >100 1826.3 574 - 55331 25 >100 1900.3 0 - 52624 20 25 1914.7 397 - 52624 15 10 1observed * Intensity values were obtained from reference 175. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 154 Method Development Detector Optimization The results obtained for hydrogen and iodine studies revealed that the most useful power range was from 50 to 70 watts. In studying detector response for sulfur, an applied power of 70 watts was chosen. Plasma observation height profiles were compared between argon and argon/helium carrier gas systems. Results for this study at 1807 A are illustrated in Figure 24. The most obvious result of adding helium to the argon carrier gas was the shortened plasma length. content was approximately 10 to 14%. The helium At 70 W and an observation height of -14 mm, response was twice as high when helium was added to the carrier gas. This increased response was again assumed to be due to background emission when the plasma contracted. When observing emission at a point near the lower tip, the plasma shrank just enough so that it extended into the observation port. Another interesting observation was the fact that the background emission profile, measured just before injection, closely paralleled the response to sulfur. There was a large increase in background corresponding to maximum sulfur response when helium was added. This being the case, one could simply set the observation height for maximum background emission to- approximate the region of greatest response for sulfur. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 155 Figure 24 Detector Optimization - Sulfur (1807 A) Grating Slits Ar flow Injections Key: 1900 A blaze 0.2 mm 100 mL/min. 2 y L x 50 ppm Methyl parathion Argon plasma background ▲ Argon plasma, sulfur emission O Argon + 10% He background • Argon + 10% He, sulfur emission A Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 156 Figure 24 60 Height, 40 Peak mm 50 30- 10 - o -o -o ' 20 Distance from 40 C avity Center, mm Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 50 157 Because of the relatively small region of maximum response, reproducibility was a problem. Any small changes in the plasma such as flow rate or helium content led to a large change in response. The large error bars for the data points between 10 and 16 mm from the cavity center illustrate this point. This effect probably could have been minimized by reducing the helium content in the plasma. The relatively "flat" response profile observed in the argon plasma led to more consistent results. Note that background emission again closely paralleled the sulfur response profile. As a compromise between sensitivity and reproducibility, an observation height of 14 mm below the cavity center was chosen for further studies using argon as the carrier gas. Separation Names and structural formulas for the organic sulfur compounds used in this study are presented in Figure 25. / These compounds are usually referred to as organophosphorous pesticides. The same OV-1 column used for the iodine compounds was used in separating these pesticides. Column temperature was set at 200°C. At an argon carrier gas flow rate of 110 mL/min., retention times were as follows: diazinon, 195 sec.; methyl parathion, 270 sec.; malathion, 337 sec.; ethyl parathion, Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 25 Sulfur Compounds Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 25 Methyl Parathion S (CH30)2-P-0 N02 Ethyl Parathion S (C2H 50)2-P-0 N02 Diazinon Malathion Ethion Sf3 s. N CH S OC-Hp 'OH-C. C-O-P 2 S CH' 'N' s0C2H 5 CH, § o (CH30)2-P-S-CH-CH2-C-OC2H5 I C-OC-Hc II 2 5 O C-H-O S S 0CoH c 2 5 P-S-CH9-S-P 2 5 C2H50/ X°C2H5 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 160 348 sec.; and ethion 818 sec. Column efficiency was measured with respect to the total number of theoretical plates. A theoretical plate height of 6 mm was measured, which corresponded to a total of 300 theoretical plates. Sensitivity and Detection Limit Optimum instrumental conditions were used to determine sensitivity for methyl parathion. Seventy watts of microwave power were applied and observation height was set at 14 mm below the cavity center. Slits were set at 0.5 mm. The minimum detectable limit (MDL) for methyl parathion was 4 pg S/second. This respresented an absolute detection limit of 120 pg sulfur. Response vs. Compound Type Sulfur emission response for the different pesticide compounds was compared on an absolute basis. summary of these results. Table 16 is ..a Response was not found to be directly related to the number of sulfur atoms in the pesticide compounds. This discrepancy may have been due either to the inability of the argon plasma to fragment certain sulfur bonds or to differences associated with the separation or determination of the compounds. The difference in response between methyl and ethyl parathion could not be due to differences in fragmentation efficiency. Diazinon is also similar in structure but showed greatly reduced response. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 161 Table 16 Sulfur Response vs. Compound Type Compound Injection Response (mm2/compd.) Response (mmz/ng/s) Methyl Parathion 0.1016 yg 903 7.4 Ethyl Parathion 0.0768 yg 1302 12.0 Diazinon 0.0500 yg 126 1.2 Malathion 0.0984 yg 789 4.1 Ethion 0.0904 yg 2268 6.8 Selectivity Selectivity for the sulfur resonance line, at 1807 A was found to be dependent on slit width. Increasing the slit width led to increased sensitivity for sulfur. However, background emission due to carbon increased to an even greater extent. The argon carrier gas contained approximately 10% helium in this study. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 162 Sulfur response was measured for 2 yL injections of methyl parathion at 50.8 ng/yL. This represented an absolute injection of 12.2 ng sulfur. was used to measure carbon response. Carbon tetrachloride Column temperature was reduced to 40°C so that the plasma would not be extinguished by 0.9 yL injections of CC14 . This represented an absolute injection of 112 yg carbon. A comparison of selectivity ratios at different slit widths is presented in Table 17. a slit width of 0.5 mm. Selectivity was highest at Increasing slit widths led to greater sensitivity but lower.selectivity. This effect was due to the line width of the sulfur emission. Increasing the bandpass increased response for sulfur until it became larger than the line width. Background response to carbon emission increased exponentially with slit width as shown by the normalized response factors given in Table 17. Sensitivity and' selectivity suffered at slit widths less than 0.5 mm. Selectivity was also determined with respect to observation height. widths of 0.5 mm. Pure argon carrier gas was used at slit The selectivity ratio for sulfur vs. carbon response was greatest (2x10**) at an observation height of -18 mm. Selectivity did not vary greatly along the length of the plasma. Results of this study did show that the selectivity ratio dropped by more than an order of Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 163 T a b le 17 Selectivity for Sulfur vs. Slit Width Slit Width (mm) Detector Response (cm2/mole) Sulfur Carbon Selectivity Ratio 0.25 2.4x10s ( D * 4.2x10® (1) 5.7x10“ 0.5 1.3x10® (6) 1.6x10“ (4) 8.1x10“ 0.75 2.2x10® (9) 4.2x10“ (10) 5.2x10“ 1.0 2.3x10® (0) 8.2x10“ (19) 2.8x10“ * Normalized response values are in parentheses. magnitude when the 10% helium was removed from the argon carrier gas. Conclusions Purging the spectrometer with helium proved to be an effective alternative to evacuation. The lowest useful wavelength with helium purging was probably about 1600 A in this study. Purification of the helium to remove traces of oxygen should lower this wavelength limit even further, considering the optical path of the monochromator is nearly 4 meters long. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 164 Observation of VUV emission from iodine compounds did not represent an improvement over sensitivities reported elsewhere in the literature for longer wavelengths.. The selectivity ratio determined for iodine in this study was favorable, however. Much better results were obtained for resonance sulfur emission lines. Sensitivity and selectivity was improved by one order of magnitude over previous GC-MES’investigations at UV-visible wavelengths. Monitoring resonance transitions in quantitative work can lead, to problems when self-absorption of the analytical source is significant. This effect was minimized by using small plasma capillary diameters and transverse viewing to provide an optically thin source. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 165 APPENDIX A SPECTROMETER ALIGNMENT PROCEDURE 1. The table which supported the spectrometer was leveled. 2. Precise leveling was accomplished by adjusting the screws on the vibrationless mounts at each end of the spectrometer. 3. - A laser was set up approximately twenty feet from the entrance slit of the instrument. The light beam from this laser defined the optical path of the instrument. The first requirements were that: i) the beam was level; ii) the beam entered the spectrometer at the exact center of the slit; and iii) the beam hit the horizontal center of the collimating mirror. At this point, the beam was found to be 6.3 cm above the inside shelf of the spectrometer. 4. The collimating mirror height was adjusted so that its vertical center was 6.3 cm above the shelf. 5. The collimating mirror tilt was adjusted so that the incoming beam hit the center of the grating at a height of 6.3 cm from the shelf. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 166 6. The grating tilt was adjusted so that various orders of the diffracted beam could be reflected back to the center of the collimating mirror, and back to the laser. 7. The "camera” mirror height was adjusted so that its center was 6.3 cm above the shelf. 8. The grating was rotated until one of the orders hit the center of the "camera" mirror. 9. The "camera" mirror tilt was adjusted so that the laser beam was reflected towards the center of the exit slit. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 167 APPENDIX B GC-MES ALIGNMENT PROCEDURE 1. The light pipe mount (Figure 5) was adjusted so that emission from the observation hole passed through the center of the light pipe. Since the position of the observation hole with respect to the GC was fixed, this adjustment should not change. 2. After removing the microwave cavity and plasma capillary, the GC mount was adjusted so that the laser beam entered the center of the light pipe. After replacing the plasma capillary, the observation hole should still line up with the light pipe. There should only be a space of approximately 2 mm between the plasma capillary and the end of the light pipe when the other end is touching the entrance slit. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 168 A P P E N D IX C BENZENE - SPECTRUM 1 0 0 0 - Grating Slits Ar flow Observation ht. Power 2500 A 1900 A blaze 0.25 100 mL/min. -9mm 80 watts Assignments of Intense Transitions X, A Assignment I observed 1657.0 Carbon (I) 1742.7 Nitrogen (I) >100 1930.9 Carbon (I) >100 2478.6 Carbon (I) >100 75 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. — , 1000 A 169 - I _ _f t v A VnA v A 2000 A _ 1500 A I Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 170 A P P E N D IX D NITRO ETHANE - SPECTRUM 1 0 0 0 - Grating Slits Ar flow Observation ht. Power 2500 A 1900 A blaze 0.25 mm 100 mL/min. -9 mm 70 watts Assignments for Intense Transitions X, A Assignment 1657.0 Carbon (I ) 1742.7 Nitrogen (I ) >100 1930.9 Carbon (I ) >100 2478.6 Carbon (I) 28 observed 95 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 171 < © o o 7\— /A. < o o in H < o o o CM Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 172 APPENDIX E MALATHION - SPECTRUM 1000 - 1500 A Grating Slits Ar flow Observation ht. Power 1900 A 0.25 mm 100 mL/min -9 mm 70 watts Assignments for Intense Transitions X, A Assignments 1657.0 Carbon (I) 57 1666.7 Sulfur (I) 52 1742.7 Nitrogen (I) 1782.3 Sulfur (I) 10 1807.3 Sulfur (I) >100 1820.4 Sulfur (I) >100 1826.3 Sulfur (I) >100 1900.3 Sulfur (I ) 25 1914.7 Sulfur (I) 10 1930.9 Carbon (I ) >100 observed >100 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 173 < o o m iH ILJ^V^ W aJUJLvV I I I o o (N Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. I 174 APPENDIX ] BROMOBUTANE - SPECTRUM 1000 - 2500 A Note: Grating Slits Ar flow Observation ht. Power Attenuation 1900 A blaze 0.25 mm 100 mL/min. -9 mm 70 watts 1x10_9 (1000-1500 A) 1x1 0 '8 (1500-1500 A) Assignments of Intense Transitions X, A Assignment *observed 1574.8 1576.4 Bromine (I ) 5 1633.4 Bromine (I ) 15 1657.0 Carbon (I ) >100 1742.7 Nitrogen (I ) >100 1930.9 Carbon (I) >100 2478.6 Carbon (I ) >100 1582.3 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1000 A 175 JIaAJ 2000 A 1500 A Ju Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. A 176 APPENDIX G IODOBUTANE - SPECTRUM 1000 - 2500 A Grating Slits Ar flow Observation ht. Power 1900 A blaze 0.25 mm 100 mL/min. -9 mm 80 watts Assignments of Intense Transitions A Assignments *observed 1640.8 Iodine (I) j 1642.1 Iodine (I) ) 1657.0 Carbon (I ) .>100 1702.1 Iodine (I) 7 1742.7 Nitrogen (I) 1782.8 Iodine (I) 97 1799.1 Iodine (I) 52 1830.4 Iodine (I) >100 1844.5 Iodine (I) 10 1930.9 Carbon (I ) 91 2062.4 Iodine (I) 8 2478.6 Carbon (I ) 11 31 >100 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1500 A 1000 A 177 /\w 2000 A u JL JL Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 178 APPENDIX H ANALYTICAL APPLICATIONS OF GC-MES This information is included to review the following features of GC-MES: 1. Elements that can be determined and the wavelength of observation. 2. The influence of cavity t y p e , gas type, gas pressure and analytical wavelength on sensitivity and ' selectivity. 3. Additional references related to the analyte species of interest. Sensitivities are presented in terms of the mass flow rate required to give an analyte signal equal to a certain multiple of the noise. This multiple is generally two but, in some cases, multiples of one or three times the noise have beexi chosen to define "detectability.” No attempt has been made to normalize sensitivity data. As mentioned previously, there are also inconsistencies in the method used for calculating selectivity ratio. this data is included for approximate comparisons only, values were left as found in the literature. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Since 179 KEY X, A Plasma Gas Cavity Sensitivity Selectivity Ref. Aluminum (see also: reference 99) 3962 Ar(atm) 1/4-wave 3962 Ar(atm) tapered 3952 He (atm) ™ 0 io 20 pg/s 0.5 ng 5 pg/s 990 96 - 176 39°? 67 Antimony 2598 Ar(.atm) tapered 50 pg 2000 98 pg 2x10“ 86 6 pg/s 5x10“ 67 Arsensic (see also: references 98,103) 2288 2288 Ar(atm) He (atm) tapered ™ 0io 20 Beryllium 2349 Ar(atm) tapered 10 pg - Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 99 180 Boron (see also: reference 177) 2498 He (atm) ™ qio 4 pg/s 9259 67 Bromine (see also:' references 81,89,34,33,103) 2985 Ar(atm) tapered 200 ng/s 10 17 4786 He(red) tapered 20 pg/s 38 71 4705 Ar(atm) 3/4-wave - 31 4705 He(red) 1/4-wave 90 pg/s 1300 78 4705 He(atm) ™010 5 pg/ s 220 23 4705 He (atm) ™ qio 10 pg/ s 1400 79 ™010 S ’p'g/s 530 80 4705 He(atm) 159 ng 4705 He (atm) ™ 0io 67 pg/s 1060 67 4786 H e (atm) ™010 34 pg/ s 599 67 Carbon (see also: references 17,31,81,89,33,103) 2479 Ar(atm) 3/4-wave 20 pg/s 1 31 2479 He(red) 1/4-wave 80 pg/s 1 78 1931 1931 He (atm) He(atm) ™oio ™010 400 fg/ s 9 pg/s 1 1 23 80 2479 He (atm) ™ qio 3 pg/s 1 67 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 181 Chlorine (see also: references 17,21,31,81/89,33,103) 4795 He(red) tapered 60 pg/s 4795 He(red) 1/4-wave 60 pg/s 510 78 4795 He(atm) ™010 7 pg/s 200 23 4810 He (atm) ™010 16 pg/s 2400 79 4810 He(atm) 16 pg/s 730 34 4795 He (atm) 5 pg/s 1000 80 4795 He (atm) 43 pg/s 610 67 3930 96 ™010 ™010 ™010 44 . 71 Chromium (see also: references 92,100) 3579 Ar(atm) 1/4-wave 3 pg/s 4254 Ar(atm) tapered 1 ng 3579 He(red) 1/4-wave 3579 Ar(atm) 1/4-wave 2677 He(atm) ™010 15 pg/s 900 fg 7000 - 99 93 100 7 pg/s 1x10s 67 6 pg/s 2x10s 67 Cobalt 2407 He(atm) ™010 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 182 Copper 3248 3247 Ar(atm) Ar(atm) tapered 1/4-wave 1 ng 8 pg/s 2250 103 96 Deuterium (see also: references 102,89,103) 6561 He(red) 1/4-wave 6561 He (atm) ™010 90 pg/s 880 78 7 pg/ s 194 67 3 pg/s 10 17 60 pg/s 2300 78 Fluorine (see also: references 89,33,103) 5166 Ar(atm) tapered 6856 He(red) 1/4-wave 6856 He (atm) ™010 8 pg/ s 3500 79 6856 He (atm) ™010 2 pg/ s 820 80 6856 He (atm) ™ 010 180 pg/s 1x10* 67 3 pg/s 1170 96 Galium 2944 Ar(atm) 1/4-wave Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 183 Germanium 2651 He(atm) ™010 ^ pg/ s 8x10“ 67 Hydrogen (see also: references 89,33,94,103) 4861 He(red) 1/4-wave 4861 He (atm) ™010 6563 He (atm) 30 pg/s - 78 2 P9/s “ 23 ™010 ® P9/s 80 4681 He(atm) . TMQ10 16 pg/s - 67 6563 He (atm) ™ 0io 7 pg/s " 67 Iodine (see also: references 69,70,21,89,31,33,103) 2062 Ar(atm) tapered 70 fg/s 1x10“ 17 5338 He(red) tapered. 50 pg/s 38 71 2062 Ar(atm) 3/4-wave 100 pg/s 1000 81 5161 He(red) 1/4-wave 50 pg/s 400 78 5161 2062 He (atm) He (atm) ™ 010 ™ 010 3 pg/s 31 pg/s 130 1100 23 79 2062 He (atm) ™ 0io 31 pg/s 140 34 2062 H e (atm) ™ 0io 7 P^/s 530 80 2062 He (atm) ™ 0io 21 pg/s 5010 67 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 184 Iron 2599 3441 He (atm) Ar(atm) ™ qio 1/4-wave 300 pg/s 13 pg/s 3x10s 1610 67 96 Lead (see also: references 103,91) ™ 0io 2833 4058 He (atm) Ar(atm) 500 fg/s 6 pg 1x10s 8x10'* 79 73 3/4-wave 4058 He (atm) ™oio 2 pg/ s 2x10s 67 2833 He (atm) ™ 010 200 fg/s 2xl°S 67 ™ 0io ™oio 250 fg/s 2x10s 79 2 pg/s lxl°5 67 Manganese 2576 He (atm) 2576 He (atm) Mercury (see also: references 178,179,84,57,85,103) 2537 Ar(atm) tapered 100 pg lxlO4 83 2537 He(red) 1/4-wave 50 fg - 38 ™ 0io ™ qio 1 pg/s 600 fg/ s 9x104 8x10* 79 67 2537 2537 He (atm) He (atm) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 185 Molybdenum. 2816 He (atm) ™oiO 5 pg/s ™ 0 10 3 pg/s 6470 67 ™ 010 69 pg/s 3x10“ 67 2x10“ . 67 Nickel 2316 He (atm) Niobium 2883 He (atm) Nitrogen (see also: references 17,31,33,89,103) 7469 He(red) * 1/4-wave 3 ng/s - 78 Osmium 2256 He (atm) Oxygen (see 7772 He(red) ™ qio 6 pg/s 5x10“ 67 also: references 89,33,103) 1/4-wave 3 ng/s - Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 78 186 Phosphorous (see also: references 17,21, 81,89,103) 2536 Ar(atm) tapered 5 pg/s 100 66 2536 Ar(red) tapered 600 fg/s 1000 70 2536 He(red) tapered 9 pg/s 1000 71 2536 He(atm) ™010 2 pg/s 3x10“ 79 2536 He(atm) ™010 3 pg/s 1x10“ 67 ™010 8 pg/s 1x10* 67 1/4-wave 3 pg/s 1620 96 Ruthenium 2403 He(atm) Scandium 3614 Ar(atm) Selenium (see also: reference 103) 2040 Ar(atm) tapered 40 pg 1x10“ 97 2040 He (atm) ™ qio 5 lxl0'‘ 67 pg/s Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 87 Silicon (see also: reference 76) 2616 He(atm) ™ qio 29 pg/ s 2516 He(atm) ™ qio 9 pg/ s 3900. 79 1580 67 Sulfur (see also: references 17,87,47,88,89,33,103) 5454 He(red) tapered 50 pg/s 22 71 1820 Ar(atm) 3/4-wave 40 pg/s 460 81 5454 He(red) 1/4-wave 90 pg/s 390 78 5454 He(atm) TM ±ra010 25 pg/s 200 23 5454 He(atm) ™010 63 pg/s 250 79 5454 He(atm) ™010 39 pg/s 70 80 5454 He(atm) ™010 52 pg/s 4590 67 ™oiO 2 pg/ s 4x10s 67 Tin 2840 He (atm) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 188 Tungsten 2555 He (atm) ™oio 51 pg/s 5450 67 ™oio 1/4-wave 10 8 6x10“ 1400 67 96 Vanadium 2688 3184 He (atm) Ar(atm) pg/s pg/s Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 189 L I S T OF REFERENCES 1. Marr, G. W . , Plasma Spectroscopy, Elsevier Publishing Co., New York (1968), p. 2. 2. Skogerboe, R. K. and Coleman, G. N., Anal. Chem., 48, 611A (1976). 3. Fassel, V. A. and Kniseley, R. N . , ibid., 46, 1110A (1974) 4. Fassel, V. A. and Kniseley, R. N., ibid., 46, 1155A (1974). 5. Greenfield, S., McGeachin, H. and Smith, P. B.,' Talanta, 23, 1 (1976). 6. Sytz, A., CRC Critical Reviews in Anal. Chem., 4, 216 (1974). 7. Greenfield S., McGeachin, H. and Smith, P. B., Talanta, 22, 553 (1975). 8. Zander, A. T. and Hieftje, G. M . , Appl. Spectrosc., 35, 357, (1981). 9. Barnes, R . , Anal. Chem., 50, 100R, (1978). 10. Boyko, W. J., Keliher, P. N. and Malloy, J. M . , ibid., 52, 53R (1980). 11. Boyko, W. J., Keliher, P. N. and Patterson, J. M., ibid, 54, 188R (1982). 12. West, C. D., ibid, 42, 811 (1970). 13. Bache, C. A. and Lisk, D. J., ibid., 40, 2224 (1968). 14. Bauer, C. F. and Natusch, D. F. S., ibid., 53, 2020 (1981). 15. Kostkowski, H. J. and Broida, H. P., J. Opt. Soc. Am., 46, 246 (1956). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 190 16. Broida, H. P. and Chapman, M. W., Anal. Chem., 30, 2049 (1958). 17. McCormack, A. J., Tong, S. C. and Cooke, W. D., ibid., 37, 1470 (1965). 18. Fehsenfeld, F. C., Evenson, K. M. and Broida, H. P., R e v . Sci. Instrum., 36, 294 (1965). 19. Lichte, F. E. and Skogerboe, R. K . , Anal. Chem., 45, 399 (1973). 20. Layman, L. R. and Hieftje, G. M . , ibid., 47, 194 (1975). 21. Moye, H. A., ibid., 39, 1441 (1967). 22. Beenakker, C. I. M . , Spectrochim. Acta, 3 IB, (1976). 23. Beenakker, C. I. M., ibid., 32B, 173 (1977). 24. Zander, A. T. and Hieftge, G. M . , Anal. Chem.,-50, 1257 (1978). 25. Mulligan, K. K . , Hahn, M. H., Caruso, J. A. and Fricke, F. L., ibid., 51, 1935 (1979). 26. Broekaert, J. A. C., Spectrochim. Acta, 36B, 563 (1981). 27. Matus, L. G. and Boss, C. B., Abstracts, 33rd Pittsburgh Conference on Analytical Chemistry and Applied Spectroscopy, Atlantic City, NJ, March, 1982, Paper No. 443. 28. Bollo-Kamara, A. and Codding, E. G., Abstracts, 32nd Pittsburgh Conference on Analytical Chemistry and Applied Spectroscopy, Atlantic City, NJ, March 1981, Paper No. 512. 29. Matus, L. G. and Boss, C. B., ibid., Paper No. 515. 30. Kawaguchi, H. and Vallee, B. L . , Anal. Chem., 47, 1029 (1975). 31. Dagnall, R. M . , West, T. S. and Whitehead, P., Anal. Chim. Acta, 60, 25, (1972). 32. Stanley, J. L . , Bentley, H. W. and Denton, M. B., Appl. Spectrosc., 27, 265 (1973). 483 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 191 33. van Dalen, J. P. L . , de Lezenne Coulander, P. A. and de Galan, L., Anal. Chim. Acta, 94 , 1 (1977). 34. Quimby, B. D., Delaney, M.' F., Uden, P. C., and Barnes, R. M . , Anal. Chem., 51, 875 (1979). 35. Brassem, P. and Maessen, F. J. M. J., Spectrochim. Acta, 29B, 203 (1974). 36. Brassem, P. and Maessen, F. J. M. J . , ibid., 30B, 547 (1975). 37. Busch, K. W. and Vickers, T. J., ibid., 28B, 85 (1973). 38. Houpt, P. M . , Ana l . Chim. Acta, 86, 129 (1976). 39. Brassem, P., Maessen, F. J. M. J., and de Galan, L., Spectrochim. Acta, 33B, 753 (1978). 40. Venugopalan, H., Reactions Under Plasma Conditions, Wiley-Interscience, New York (1971). 41. Fallgatter, K . , Svoboda, V. and Winefordner, J* D., Appl. Spectrosc., 2 5 , 347 (1971). 42. Zander, A. T., Williams, R. K. and Hieftje, G. M . , Anal. Chem., 49, 2372 (1977). 43. Taylor, H. E., Gibson, J. H. and Skogerboe, R. K., ibid., 42, 876 (1970). 44. Serravello, F. A. and Risby, T. H., ibid., 48, 673 (1976). 45. Dagnall, R. M . , Johnson, D. J. and West, T. S., Spectrosc. Letters, 6, 87 (1973). 46. Runnels, J. H. and Gibson, J. H., Anal. Chem., 39, 1398 (1967). 47. Taylor, H. E., Gibson, J.H. and Skogerboe, ibid., 42, 1569 (1970). 48. Aldous, K. M., Dagnall, R. M . , Sharp, B. T. S., Anal. Chim. Acta, 54, 233 (1971). 49. Kawaguchi, H . , Atsuya, I. Chem., 49, 266, (1977). R. K . , L. and West, and Vallee, B.L . , Anal. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 192 50. Fricke, F. L., Rose, O. R. and Caruso, J. A., Talanta, 23, 317 (1976). 51. Atsuya, I., Kawaguchi, H. and Vallee, B. L., Anal. Biochem., 77, 208 (1977). 52. Atsuya, I., Alter, G. M., Veillon, C. and Vallee, B. L., ibid., 79, 202 (1977). 53. Fricke, F. L., Rose, 0., Jr., and Caruso, J. A., Anal. Chem., 47, 2018 (1975). 54. Rose, O., Mincey, D. W . , Yacynych, A. M . , Heineman, W. R. and Caruso, J. A . , Analyst, 101, 753 (1976). 55. Alder, J. F. and DaCunha, M. T. C . , Can. J. Spectrosc., 2 5 , 32 (1980). 56. Mitchell, D. G . , Aldous, K. M. and Canelli, E., Anal. Chem., 49, 1235 (1977). 57. Watling, R. J., Anal. Chim. Acta, 75, 281 (1975). 58. Lichte, F. E. and Skogerboe, Anal. Chem., 44, 1480 (1972). 59. Barrett, P. and Copeland, T. R . , Chem. Abstracts, 92, 226029c. 60. Lichte, F. E. and Skogerboe, R. K . , Anal Chem., 4 4 , 1321 (1972). 61. Skogerboe, R. K., Dick, D. L., Pavlica, D. A. and Lichte, F. E., ibid., 47. 568 (1975). 62. Tanabe, K . , Matsumoto, K . , Haraguchi, H. and Fuwa, K . , ibid., 52, 2361 (1980). 63. Ishizuka, T. and Uwamino, Y., Chem. Abstracts, 92, 51315p (1980). 64. Leis, F. and Laqua, K . , Spectrochim. Acta, 34B, 307 (1979). 65. Bauer, C. F. and Natusch, D. F. S., Environ. Sci. Techno1 ., 15, 783 (1981). 66. Bache, C. A. and Lisk, D. J., Anal. Chem., 37, 1477 (1965). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 193 67. Estes, S. A., Uden, P. C. and Barnes, R. M., ibid., 53, 1829 (1981). 68. Tanabe, K . , Haraguchi, H. and Fuwa, K . , Spectrochim. Acta, 36B, 119 (1981). 69. Bache, C. A. and Lisk, D. J., Anal. Chem., 38 , 783 (1966). 70. Bache, C. A. and Lisk, D. J., ibid., 38 , 1757 (1966). 71. Bache, C. A. and Lisk, D. J., ibid., 3 9 , 786 (1967). 72. Bache, C. A. 301 (1968). and Lisk, D. J., J. Gas Chromatogr., 6, 73. Reamer, D. C., Zoller, W. H. and O'Haver, T. C., Anal. Chem., 50, 1449 (1978). 74. -Parkinson, R. T., Talanta, 14, 1037 (1967). 75. Brenner, K. S., J. Chromatogr, 167, 365 (1978). 76. ’ Bostick, D. T. and Talmi, Y., J. Chromatogr. Sci., 15, 164 (1977). 77. Gebhart, J. E., Ph.D. Thesis, The University of Iowa, Iowa City, (1978). 78. McLean, W. R . , Stanton, D. L. and Penketh, G. E., Analyst, 98, 432 (1973). 79. Quimby, B. D., Uden, P. C. and Barnes, R. M . , Ana l . Chem., 50, 2112 (1978). 80. Tanabe, K. Haraguchi, H. and Euwa, K . , Spectrochim. Acta, 36B, 633 (1981). 81. Dagnall, R. M . , West, T. S. and Whitehead, P., An a l . Chem., 44, 2074 (1972). 82. Braun, W., Peterson, N. C., Bass, A. M. and Kurylo, M. J., J. Chromatogr., 55, 237 (1971). 83. Bache, C. A. and Lisk, D. J., Anal. Chem., 43, 950 (1971). 84. Talmi, Y . , 85. Talmi, Y. Anal. Chim. Acta, 74, 107 (1975). and Norvell, V. E . , ibid.,85, 203 (1976). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 194 86. Talmi, Y. and Bostick, D. T., Anal. Chem., 47, 2145 (1975). 87. Dagnall, R. M . , Pratt, S. J., West, T. S. and Deans, D. R., Talanta, 16, 797 (1969). 88. Dagnall, R. M . , Pratt, S. J., West, T. S. and Deans, D. R., ibid., 17, 1009 (1970). 89. Lowings, B. J . , Analusis, 1, 510 (1972). 90. Quimby, B. D., Delaney, M. E., Uden, P. C. and Barnes, R. M . , Anal. Chem., 52, 259 (1980). 91. Estes, S. A., Uden, P. C. and Barnes, R. M., ibid., • 53, 1336 (1981). 92. Serravallo, F. A. and Risby, T. H . , ibid., 47, 2141 (1975). 93. Serravallo, F. A. and Risby, T. H., J. Chromatogr. Sci., 12, 585 (1974). 94. Schwartz, F. P., Anal. Chem., 50, 1006 (1978). 95. Schwartz, F. P., ibid., 51, 1508 (1979). 96. Dagnall, R. M . , West, T. S. and Whitehead, P., Analyst, 98, 647 (1973). 97. Talmi, Y. and Andren, A. W . , Anal. Chem., 46, 2122 (1974). 98. Talmi, Y. and Norvell, V. E., ibid., 47, 1510 (1975). 99. Kawaguchi, H . , Sakamoto, T. and Mizuike, A., Talanta, 20, 321 (1973). 100. Black, M. S. and Sievers, R. E., Anal. Chem., 48, 1872 (1976). 101. Dagnall, R. M . , Silvester, M. D., West, T. S. and Whitehead, P., Talanta, 19, 1226 (1972). 102. Luippold, D. A. and Beauchamp, J. L., Anal. Chem., 42, 1374 (1970). 103. Hobbs, J. S., Mannan, C. A. and Beeston, B. E. P., Am. Lab., 11, 43 (1979). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 195 104. Bonnekessel, J. and Klier, M . , Anal. Chim. Acta, 103, 29 (1978). 105. Hobbs, J., Chem. Abstracts, 92, 68924j . 106. Mulligan, K. J., Caruso, J. A. and Fricke, F. Analyst. 105, 1060 (1980). L., 107. Adlard, E. R . , CRC Critical Reviews in Anal. Chem., 5, 13 (1975). 108. Fredericks, E. M. and Harlow, G. A., Anal. Chem., 35, 263 (1964). 109. Hall, R. C., J. Chromatogr. Sci., 12, 152 (1974). 110. Gutsche, B. and Herrmann, R., Z. Anal. Chem., (1969). 245, 274 111. Nowak, A. V. and Malmstadt, H. V., Anal. Chem., 40, 1108 (1968) 112. Karmen, A. and Guiffrida, L., Nature, 201, 1204 (1964). 113. Brody, S. S. and Chaney, J. E., J. Gas Chromatogr., 4, 42 (1966). 114. Mizany, A. I., J. Chromatogr. Sci., 8, 151 (1970). 115. Bowman, M. C. and Beroza, M., Anal. Chem., 40, 1448 (1968). 116. Dagnall, R. M . , Thompson, K. C. and West, T. S., Analyst, 92, 506 (1967). 117. Beroza, M. and Bowman, M. C., Environ. Sci. Technol., 2, 450 (1968). 118. Perry, S. The Inst, G. and Carter, F. W. G . , Gas Chromatography, of Petroleum, London (1971). 119. Golovnya, R. V., Aerov, A. F. and Garbuzov, Anal. Khim., 36, 364 (1981). V. G., Z. 120. Ellebracht, S. R . , Fairless, C. M. and Manahan, S. E., Anal. Chem., 50, 1649 (1978). 121. Swaim, P. D. and Ellebracht, S. R., ibid., 51, 1605 (1979). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 196 122. Treybig, D. S. and Ellebracht, S. R . , ibid., 52, 1633 (1980). 123. Braman, R. S. and Dynako, A., ibid., 40, 95 (1968). 124. Windsor, D. L. and Denton, M. B., J. Chromatogr. Sci., 17, 492 (1979). 125. Freel, M. I., Hart, L. K. and Reimer, T. T., "A Study of the Lactose Intolerance of Multiple Sclerosis Patients," Grant Application to Johnson County Branch Chapter of the Eastern Iowa Multiple Sclerosis Society (1980). 126. McMorrow, K. J., General Practitioner, 18, 126 (1958). 127. McAlpine, D., Lumsden, C. E. and Acheson, E. D., Multiple Sclerosis, A Reappraisal, Livingston, London (1965). 128. Wunderlich, R. C., Fatigue; What Causes It, What It Does To You, What You Can D o •About It, Johnny Reads, Inc., St. Petersburg (1976). 129. Davenport, H. W., Physiology of the Digestive Tract, Chicago (1977). 130. Wolin, M. J., Science, 213, 1463 (1981). 131. Holloway, W. D., Tasman-Jones, C. and Lee S. P., A m . J. Clin. Nutr., 31, 927 (1978). 132. Lehninger, A. L., Biochemistry, Worth Publishers, Inc., New York (1975), p. 2 6 1 . ’ 133. Newcomer, A. D. and McGill, D. B., Gastroenterology, 51, 481 (1966). 134. Bond, J. H. and Levitt, 135. Calloway, D. G . , Colasito, D. J. and Mathews, R. D., Nature, 212, 1238 (1966). 136. Levitt, M. D., N. Engl. 137. Johnson, J., Kretchmer, N. and Simoons, F., Adv. Pediat., 21, 206 (1974). 138. M. D., ibid., 70, 1058 (1976). J. Med., 281, 122 (1969). Newcomer, A. D., McGill, D. B., Thomas, P. J. and Hofmann, A. F . , N. Engl. J. Med., 293, 1232 (1975). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 197 139. Levitt/ M. D. and Donaldson, R. M . , J. Lab. Clin. Med., 75, 937 (1970). 140. Levitt, M. D. and Donaldson, R. M . , Clin. R e s ., 16, 287 (1968). 141. Gearhart, H. L., Bose, D. P., Morrison, R. D., Welsh, J. D. and Smalley, T. K . , Anal. Chem., 48, 393 (1976). 142. Solomons, N. W., Viteri, F. E. and Hamilton, L. H . , N. Engl. J. Med., 90, 856 (1977). 143. Metz, G . , Peters, T. J., Jenkins, D. J. A., Newman, A. and Blendis, L. M . , Lancet, 1, 1155 (1975). 144. Solomons, N. W., Viteri, F. E. and Rosenberg, I. H . , Pediat. Res., 12, 816 (1978). 145. Calloway, D. H. and Murphy, E. L., NASA Contract No. NAS 9-3966, University of California, Berkeley, (1966), pp. 47-52. 146. Murphy, E. L. and Calloway, D. H . , Am. J. D i g . P i s ., 17, 639 (1972). 147. Hepner, G. W . , Gastroenterology, 67, 1250 (1974). 148. Carle, G. C., J. Chromatogr. Sci., 8, 550 (1970). 149. Bond, J.H. and Levitt, M. D., J. Clin. Invest., 51, . 1219 (1972). 150. Niu, H., Schoeller, D. A. and Klein, P. D., J .Lab. Clin. Med., 94, 755 (1979). 151. Parkinson, R. T. and Wilson, R. E., J. Chromatogr., 24, 412 (1966). 152. Payne-Bose, D., Tsegaye, A., Morrison, R. D. and Waller, G. R . , Anal. Biochem., 88, 659 (1978). 153. Gawlowski, J., Niedzielski, J. and Wieckowski, A., J. Chromatogr., 151, 370 (1978). 154. Gawlowski, J., Maurin, J. and Niedzielski, J., ibid, 168, 1 (1979). 155. Bergman, I., Coleman, J. E. and Evans, D., Chromatographia, 8, 581 (1975). 156. Guglya, V. G., Z. Anal. Khim., 34, 405 (1979). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 198 157. Ricard, A. and Lefebvre, J., Analusis, 6, 299 (1978). 158. Dreher, G. D., Ph.D. Thesis, The University of Iowa, Iowa City (1971). 159. Cox, R. D., Ph.D. Thesis, The University of Iowa, Iowa City (1980). 160. Dreher, G. B., Trischan, G. M. and Frank, C. W . , Unpublished Work. 161. Pearse, R. W. B. and Gaydon, A. G., The Identification of Molecular Spectra, John Wiley & Sons, Inc., New York (1941). 162. Striganov, A. R. and Sventitskii, N. S., Tables of Spectral Lines of Neutral and Ionized Atoms, IFI/Plenum Press, New York (1968). 163. Adcock, B. D. and Plumtree, W. E. G., J. Quant. Spectrosc. Radiat. Transfer, 4, 29 (1964). 164. King, J., Jr. and Benson, S. W . , Anal. Chem., 38, 261 (1966). 165. Mindrup, 166. Grob, R. L.,Modern Practice of Gas Chromatography, John Wiley & Sons, Inc., New York (1977), p. 219. 167. King, W. 168. Schmidt, U. and Seiler, W . , J. Geophysical R e s ., 7 5 , 1713 (1970). 169. Romand, J. and Balloffet, G., J. Phys. R a d ., 16, 489 (1955). 170. Heine, D. R . , BAbis, J. S., Denton, M. B., A p p l . Spectrosc., 34, 595 (1980). 171. Zaidel, A. N. and Shreider, E. Y., Vacuum Ultraviolet Spectroscopy, Humphrey Science Publishers, Ann Arbor (1970), p. 370. 172. Dreher, G. B. and Frank, C. W . , Appl. Spectrosc., 28 , 191 (1974). 173. Kirkbright, G. F., Ward, A. Chim. Acta, 62 241 (1972). R . ,J. Chromatogr. Sci., 16, 380 (1978). H . ,Environ. Sci. Techno1 ., 4, 1136 (1970). F. and West, T. S., Anal. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 199 174. Kirkbright, G. F., Ward, A. F. and West, T. S., ibid., 64, 353 (1973). 175. Kelly, R. L., Vacuum Ultraviolet Emission Lines, University of California, Lawrence Radiation . Laboratory, Livermore, CA (1959). 176. Sakamoto, T., Kawaguchi, H. and Mizuike, A., J. Chromatogr., 121, 383 (1976). 177. Sarto, L. G., Estes, S. A., Uden, P. C., Siggia, S. and Barnes, R. M . , Anal. Letters, 14, 205 (1981). 178. Grossman, W. E. L., Eng, J. and Tong, Y. C., Anal. Chim. Acta, 60, 447 (1972). 179. Houpt, P. M. and Compaan, H . , Analusis, 1, 27 (1972). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.