UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Critical evaluation of the coated wire glucose sensor Sharareh, Shiva 1991

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-UBC_1991_A1 S52.pdf [ 33.3MB ]
Metadata
JSON: 831-1.0059023.json
JSON-LD: 831-1.0059023-ld.json
RDF/XML (Pretty): 831-1.0059023-rdf.xml
RDF/JSON: 831-1.0059023-rdf.json
Turtle: 831-1.0059023-turtle.txt
N-Triples: 831-1.0059023-rdf-ntriples.txt
Original Record: 831-1.0059023-source.json
Full Text
831-1.0059023-fulltext.txt
Citation
831-1.0059023.ris

Full Text

CRITICAL E V A L U A T I O N OF T H E C O A T E D W I R E G L U C O S E SENSOR By SHIVA S H A R A R E H B.Sc. University of New Mexico 1984 M.Sc. University of New Mexico 1986 A T H E S I S S U B M I T T E D I N P A R T I A L F U L F I L L M E N T O F T H E R E Q U I R E M E N T S F O R T H E D E G R E E O F D O C T O R O F P H I L O S O P H Y in T H E F A C U L T Y O F G R A D U A T E S T U D I E S C H E M I C A L E N G I N E E R I N G We accept this thesis as conforming to the required standard T H E U N I V E R S I T Y O F B R I T I S H C O L U M B I A J U N E 1991 © SHIVA S H A R A R E H , 1991 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of Br i t i sh Columbia , I agree that the L ibrary shall make it freely available for refer-ence and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Chemical Engineering The University of British Columbia 2075 Wesbrook Place Vancouver, Canada V6T 1W5 Date: Abstract Previous literature indicated that a coated wire glucose sensor consisting of a liquid ion exchange material and an ionic glucose complex coated on a platinum wire responded to changes in glucose concentration because of a shift in the equilibrium between the associated and dissociated forms of the B a-glucose complex, which could be monitored electrochemically. The purpose of this work originally was to investigate experimentally the implantation of a coated wire glucose sensor in subcutaneous tissue for monitoring glucose concentration. A glucose sensor is needed to continuously monitor blood glucose levels in conjunction with an insulin delivery system for the control of type one diabetes. From the data gathered in this research, it was observed that glucose addition does not cause any change in the potentiometric response of the coated wire electrodes either transiently or at steady state. Despite the rather poor reproducibility of some of the measurements reported in the literature, this work has been able to explain how the misinterpretation of the reported results could have occurred. The objective of the thesis as research progressed changed to the explanation of the tran-sient behavior of the glucose sensor and to the determination of what is actually causing the potentiometric response. Experiments were conducted to determine the role played by each component which was used in the preparation of the electrode. The effect of each individual component was observed potentiometrically and it was found that by removing the quaternary ammonium salt (Ali-quat) the potentiometric drift stopped. By replicating identical electrodes, variability of the potentiometric signals was recognized. It was suspected that the variability could be caused by reactions at the platinum surface. Potentiometric measurements were made using platinum n wires which had been used in constructing the previous electrodes. The levels and types of con-tamination and oxidation of the surfaces were studied with X-ray photoelectron spectroscopy and the results were compared to their individual potentiometric responses. Irreproducibility of plat inum surfaces was found to be effected by the degree of contamina-tion and the type and amount of oxidation to which the plat inum surfaces were exposed. B y placing a coating around the plat inum wire, the surface saturates wi th oxygen causing blockage of active sites of the plat inum, resulting in a different morphology of the surface which can result i n a different equilibrium potential. It was observed that : 1. Using plat inum wires potentiometrically can cause irreproducible signals. 2. Coated wire sensors are not suitable for glucose sensing. It is suggested that the observed changes of potential which had been reported in the literature were actually due to the exchange of Cl~ and (P0 4) 3~ ions in the ion-exchange material as well as to a plat inum surface potential. This thesis confirms that coated wire glucose sensors as described in the literature do not work according to the proposed mechanism. This electrode is not i n principle a reliable working electrode. The operation of the electrode is independent of any changes in glucose concentration. in Table of Contents Abstract ii List of Figures ix List of Tables xix Dedication xxiii Acknowledgements xxiv 1 I N T R O D U C T I O N 1 2 L I T E R A T U R E R E V I E W 4 2.1 Diabetes 5 2.1.1 Types of Diabetes 5 2.1.2 Clinical Aspects Of Diabetes 6 2.1.3 Treatment of Diabetes 10 2.2 Glucose Sensors : 15 2.2.1 Enzyme Electrode 15 2.3 Optical Electrode 20 2.3.1 Sensors Based on Direct Electron Transfer 21 2.4 Implantation and Needle Type Sensors 22 2.5 Ion Selective Electrode 24 2.5.1 Liquid Ion Exchangers 25 2.5.2 Coated Wire Ion Selective Electrodes 27 2.6 Coated Wire Glucose Sensor 28 iv 2.7 Preparation and Effect of Complexes 37 3 P R E L I M I N A R Y E X P E R I M E N T A L A P P R O A C H 46 3.1 The Overall Objective of the Thesis 46 3.2 Preparation of the Complexes 47 3.3 Complex Analysis 48 3.3.1 Infrared Spectroscopy 48 3.3.2 Nuclear Magnetic Resonance Spectroscopy 49 3.4 IR Results of Selected Complexes 49 3.4.1 Summary and Conclusions 55 3.5 N M R Analysis 56 3.6 Experimental Apparatus and Materials Used for Construct ing Electrodes . . . . 57 3.6.1 Lists of Mater ia l Used 57 3.6.2 Hydrogel Preparation 59 3.6.3 Polyacrylamide G e l Preparat ion 60 3.6.4 Condit ioning and Storage of the Sensors 61 3.6.5 Preparation of Phosphate Buffer 61 3.7 Preparation of the Electrodes 61 3.7.1 Identification of the Compounds Used for Preparat ion of the Electrodes . 62 3.7.2 Coated Wire Electrode 62 3.7.3 Fabrication of Identical Coated Wire Sensors 62 3.8 Electrode Testing Equipment 66 3.8.1 Meters Used 66 3.8.2 Testing Cel l 66 3.8.3 Potentiometric M e t h o d Using Needle Electrodes . . 66 3.8.4 Potentiometric M e t h o d Using Coated W i r e Electrodes 68 3.9 Prehminary Experimental Results and Discussions 68 v 4 P R E P A R A T I O N O F T H E E L E C T R O D E S 7 3 4.1 Miniaturized Electrodes 73 4.1.1 Preparation of the Glass Electrode 73 4.1.2 Potentiometric Results Observed From Glass Electrodes 74 4.1.3 Preparation of Needle Type Sensors 77 4.1.4 Potentiometric Results Observed From Needle Type Sensor 77 4.1.5 Preparation of Membrane Electrodes . . . 82 4.1.6 Potentiometric Results Observed From Membrane Electrodes 85 5 R E P R O D U C I B I L I T Y O F T H E L I T E R A T U R E R E S U L T S 9 2 5.1 Potentiometric Method 92 5.2 Results of the Potentiometric Test 96 6 N E W A P P R O A C H T O T H E P O T E N T I O M E T R I C O B S E R V A T I O N S 1 0 0 6.1 Potentiometric Observation 100 6.1.1 Description of Electrodes No.l-No.18 101 6.1.2 Testing Solutions . . 103 6.1.3 Treatment and Cleaning of the Platinum Wires 103 6.1.4 Results Obtained For Platinum Wires Treated with Acetone 103 6.1.5 Isolation of Each Component of the Electrode .119 6.1.6 Air Dried Electrodes 119 6.1.7 Description of Electrodes No.19-No.71 120 6.1.8 Comparison of Literature Results With the Data Obtained 140 6.1.9 Conditioned Electrodes 144 6.1.10 Description of Electrodes No.72-No.144 144 6.1.11 Potentiometric Results Obtained For Conditioned Electrodes 149 6.1.12 Effect of the Platinum Surface on the Electrode Response 161 6.1.13 Effect of the Alcohol Adsorption to the Platinum Surface 164 vi 7 X - R A Y P H O T O E L E C T R O N S P E C T R O S C O P Y 172 7.1 Electrochemical Studies of P l a t i n u m Surfaces 172 7.2 X-ray PhotoElectron Spectroscopy 175 7.3 Introduction to X P S 175 7.4 Instrumentation 182 7.5 Experimental Section 185 7.5.1 Reagents Used 185 7.5.2 Apparatus 185 7.6 Preparation A n d Modif icat ion O f P l a t i n u m Surfaces 186 7.7 History of P l a t i n u m Wires 188 7.8 Mathematical Background and Procedure 190 7.9 Results and Discussion 199 7.9.1 Electrochemical studies 199 7.9.2 X P S Studies 206 7.9.3 Comparisons of Electrochemical Observations w i t h X P S Results 215 7.9.4 Conclusion 219 8 D I S C U S S I O N O F R E S U L T S 223 9 C O N C L U S I O N S A N D R E C O M M E N D A T I O N S 227 9.1 Conclusions 227 9.2 Recommendation 228 Nomenclature 230 References 232 A Description of Electrodes 247 B S T A T I S T I C A L A N A L Y S I S 253 vii B.0.1 Input Data 253 B.0.2 Type I, II, III, IV SAS Tests 254 B.0.3 Manova Statement 254 B.0.4 Output Results 255 C N M R analysis 297 C l N M R results 297 D Raw Data 3 1 0 E Equipment Specification 404 vin List of Figures 2.1 Hormonal Regulation of Glucose [1] 7 2.2 The Effect of Ingestion of Carbohydrate Foods on the Concentration of B lood Glucose, Insulin Product ion and the Effect of Insulin on Glucose Entry Into Body Cells [7] 8 2.3 Dose-Response Curves for the Different Metabol ic Act ions of Insulin [2] 9 2.4 Schematic Diagram of the A r t i f i c i a l Endocrine Pancreas Outside the Body . . . . 12 2.5 I: Programmable Implantable Medicat ion System, II: The Biostator, III: Exac Tech B lood Glucose Meter , I V : Direct 30/30 13 2.6 I: Principle of Glucose Enzyme Electrodes Operat ing by Detection of Glucose Consumption, II: Schematic Drawing of Diffusion of Glucose and Oxygen . . . . 17 2.7 The Fiber Optic BioafHnity Sensor 20 2.8 The Results of a Mercury Electrode [79] 29 2.9 Continuous Monitoring of Varying Glucose Concentrat ion i n phosphate buffer solution[83] 32 2.10 Proposed Molecular Structure of the Glucose Complex [86] 33 2.11 Site Formation of a Complex on A d d i t i o n of C a l c i u m Ions [90] 40 2.12 R a m a n Spectra of Equi l ibrated 2 .0M D-Glucose i n the Range of (700 - 3 0 0 ) c m - 1 . 41 2.13 Raman Spectra of Equil ibrated 2 . 0 M D Glucose Solution Containing the Salt Indicated 42 2.14 F . t . i r Spectra of D Glucose and Its A l k a l i n e - E a r t h M e t a l Ions i n the Region of (1800 - 500)cm- l .[104] 44 ix 3.1 IR Spectra of D-Glucose and Ca Complexes in the Regions of (500-1800 cm 1), a) Ca(D-Glucose)Cl 2 .4H 20, b) /3D-Glucose, C )QD - Glucose 50 3.2 IR Spectra of Barium Complexes in the Regions of (500-1800 c m - 1 ) , d) Ba-Glucose(Freeze Dried), e) B a C l 2 Compound, f) Ba-Glucose(Recrystahize) . . . . 51 3.3 IR Spectra of Barium Complexes in the Regions of(300-4000 c m - 1 ) , g)Ba-Glucose(Freeze Dried), h) B a C l 2 Compound, i) Ba-Glucose(Recrystallize) 52 3.4 IR Spectra of D-Glucose and Copper Complexes in the Regions of (500-1800 c m - 1 ) j) Cu(Ac)2, k) Glucose Compound, 1) Cu-Glucose(Recrystalize), m ) Cu-Glucose (Slow Dehydration) '. 53 3.5 IR Spectra of D-Glucose and Copper Complexes in the Regions of (300-4000 c m - 1 ) 54 3.6 The Shape of Coated Wire Glucose Sensors 64 3.7 Fabrication of Identical Coated Wire Glucose Sensors 65 3.8 Schematic Drawing of the Electrochemical Cell 67 3.9 Schematic Diagram of the Potentiometric Apparatus 69 3.10 I: The Response of a Coated Wire Electrode, Complex No.l in 40,80,120(mg/dl) Glucose Using Hydrogel Membrane as Outer Layer, II: The Response of a Coat ed Wire Electrode, Complex No.l in 40,80,1 20(mg/dl) Glucose Using Hydrogel Membrane as Outer Layer (Conditioned) 71 3.11 I: The Response of a Coated Wire Electrode, Complex No.6, Using Hydrogel Membrane as an Outside Layer, Measured in Different Concentrations of Glucose II: The Response of Coated Wire Electrode ( Conditioned), Complex no.l, Using Tissue Membrane as an Outside Layer, Measured in Different Concentrations of Glucose 72 4.1 Schematic Drawing of the Liquid Membrane Glucose Sensor . : 75 4.2 Comparisons of Voltage-Time Responses of Conditioned Glass Electrodes, I: Complex No. l , II: Complex No.2, Tested in Deionized Distilled Water, P V C Membr ane, Final Concentration of glucose = (600 mg/dl) 76 x 4.3 Comparisons of Voltage-Time Responses of Condit ioned Glass Electrodes, I: Complex No .3 , II: Complex No.4, Tested in Deionized Disti l led Water, P V C M e m b r ane, F i n a l Concentration of glucose = (600 mg/dl ) 78 4.4 Comparisons of Voltage-Time Responses of Condit ioned Glass Electrodes, I: Complex No.5 , II: Complex No.6, Tested in Deionized Dist i l led Water, P V C M e m b r ane, F i n a l Concentration of glucose = (600 mg/dl ) 79 4.5 Construction of a Needle Type Glucose Sensors 80 4.6 I: The Response of 16 Gauge Needle Electrode, Complex No.6 Using P V C M e m -brane as an Outside Layer, Measured i n Different Concentrations of Glucose, II: The Response of a Needle Type Electrode, C o mplex No.6 , Using Tissue Membrane as an Outside Layer, Measured i n Different Concentrations of Glucose 81 4.7 Voltage-Time Response of Conditioned Needle Electrode, I: Complex N o . l , Tested in Deionized Dist i l led Water , F i n a l concentration of glucose = (300mg/dl) . . . 83 4.8 Comparisons of Voltage-Time Responses of Condit ioned Needle Electrodes, I: Complex No.2 , II: Complex No.6 , Tested i n Deionized Dist i l led Water, F i n a l co ncentration of glucose = (300mg/dl) 84 4.9 Voltage-Time Curve of P V C Membrane Electrode A , Using Cu-glucose Complex, Tested i n 3 Consecutive Days, 100ml pH=7.4 , F i n a l concentration of glucose — ( 300mg/dl) 87 4.10 Comparisons of Volt age-Time Responses of P V C Membrane Electrodes, Using I: BaCl2, II: Ba-Glucose, III: CuCh, F i n a l concentration of glucose = (300mg/dl) 88 4.11 Comparisons of Voltage-Time Responses of Hydroge l Membrane Electrodes, Us-ing Inner Solution of I: Cu(AC)2, II: FeClz , F i n a l concentration of glucose = (300mg/dl) 89 xi 4.12 Comparisons of Voltage-Time Responses of Ba-Glucose complex, Using I: Poly-acrylamite with 3500mv Cellophane, II: Polyacrylamite with lOOOmv Cellophane, III: Polyacrylamite with 0.02/zm Hydro philic membrane , IIH: 0.03/xm Hy-drophobic Membrane, Tested in pH=7.4, Final concentration of glucose = (300mg/dl) 90 5.1 I: Schematic Drawing of the Potentiometric Testing System, II: Continuous Mon-itoring of Varying Glucose Concentration [83] 93 5.2 Continuous Monitoring of Varying Glucose Concentration [83] 94 5.3 Schematic Diagram of the Potentiometric Apparatus 95 5.4 Voltage-Time Curve of I: Ba-Glucose Complex (0.2g) , II: Controlled Glucose (0.6g) Electrodes Conditioned in 18(mg/dl) Glucose in H2O ,Tested in Bu ffer pH=7.4 98 . 5.5 Voltage-Time Curve of I: Ba-Glucose Complex , II: Controlled Glucose Electrodes Conditioned in 18(mg/dl) Glucose in Buffer pH=7.4 ,Tested in Buffer pH=7 .4 . 99 6.1 Comparisons Between the Potentiometric Responses of I: 0.3g Ba-Glucose Com-plex, II: 0.6g Ba-Glucose Complex 105 6.2 Comparisons Between the Steady State Potentials of I: 0.3g Ba-Glucose Compex, II: 0.6g Ba-Glucose Complex After Addition of Glucose and KC1 110 6.3 Comparisons Between the Potentiometric Responses of Electrodes 1,2,3 Testing on 3 Different Days 113 6.4 Comparisons Between the Steady State Potentials of Electrodes 1,2,3 Testing on 3 Different Days After Glucose Addition 114 6.5 Comparisons Between the Potentiometric Responses of I: 0.3g Ba-Glucose Com-plex in Deionized H2O , II: 0.3g Ba-Glucose Complex in Acetate Buffer Sol ution 115 6.6 Comparisons Between the Steady State Potentials of I: 0.3g Ba-Glucose Complex in Deionized H2O , II: 0.3g Ba-Glucose Complex in Acetate Buffer Solu tion After the Addition of Glucose and KCl 116 xii < 6.7 Comparisons Between the Transient Potentiometric Responses of I: 0.3g Ba-Glucose Complex, II: 0.6g Ba-Glucose Complex After Addition of Glucose . . . . 128 6.8 Comparisons Between the Steady State Potentials of O.lg Ba-Glucose Complex I: Transiently, II: After Glucose Addition 129 6.9 Comparisons Between Transient Potentiometric Responses of I: O.Og BaCl2, II: 0.3g BaCl2, HI: 0.8g BaCl2 134 6.10 Comparisons Between Transient Potentiometric Responses of I: 0.8g Glucose, II: 0.8g Ba-Glucose Complex , III: 0.8g BaCl2 135 6.11 Comparisons Between the Potentiometric Responses of I: O.Og Ba-Glucose Com-plex II: O.lg Ba-Glucose Complex, III: 0.8g Ba-Glucose Complex 136 6.12 Observations Between the Steady State Potentials of 0.8g Glucose Compound When Changing the Buffer Solution 138 6.13 Comparisons Between Responses of Coated Wire Glucose Sensors from Literature [83] 142 6.14 Comparisons Between the Potentiometric Responses of Electrode 69 (O.lg Ba-Glucose Complex), Electrode 70 (O.Og Ba-Glucose Complex), Electrode 71 (0.8g Ba-Glucose Complex) 143 6.15 Comparisons Between Potentiometric Responses of I: O.Og Ba-Glucose Complex (Conditioned), II: O.Og Ba-Glucose Complex (Conditioned), Electrode No.84, A ir Dried 150 6.16 Comparisons Between Potentiometric Responses of I: 0.3g Ba-Glucose Complex (Conditioned), II: 0.3g Ba-Glucose Complex (Conditioned), Electrode No.81, A ir Dried 155 6.17 Effect of Voltage-Cone. Curves on Electrode Response [83] 156 6.18 Comparisons Between Potentiometric Responses of Aliquat to Decanol Ratio Without Ba-Glucose Complex 157 xm 6.19 Comparisons Between Potentiometric Responses of 1,11: Decanol Coating (Con-ditioned) 158 6.20 Comparisons Between Potentiometric Responses of I: Decanol Coating (Air Dried), II: Decanol Coating (Conditioned) 159 6.21 Comparisons Between Potentiometric Responses of Bare Platinum (Air dried) . . 163 6.22 Comparisons Between Potentiometric Responses of I: Bare Platinum( Conditioned), Decanol Addition 166 6.23 Comparisons Between Potentiometric Responses of Addition of n-Decyl Alcohol to I: Bare Platinum (New Buffer), II: Bare Platinum (New Buffer) 167 6.24 Comparisons Between Potentiometric Responses of Addition of n-Decyl Alcohol to I: Glass Electrodes (New Buffer), II: Glass electrodes (Air dried) 170 6.25 Comparisons Between Potentiometric Responses of Addition of n-Decyl Alcohol to I: Glass Electrode (New Buffer), II: Glass Electrode (New Buffer) 171 7.1 Diagram of (I) Photoelectric Process, (II) Auger Process 177 7.2 I: Spectrum of a Clean Platinum Surface [157], II: Typical X-ray PhotoElectron Spectrum of a Platinum Electrode With Surface Contamination 180 7.3 The Experimental Apparatus of XPS Surface Analysis System 183 7.4 Data Acquisition and Data Processing Equipment Which Is Connected With XPS Spectrometer 184 7.5 The Shape of Coated Wire Glucose Sensor 187 7.6 Table of a Clean Platinum Surface With Comparison to Contaminated Surfaces of Platinum [157] 191 7.7 Comparison of Clean Surface with Experimental and Calculated Peak for Elec-trode No.3 : 192 7.8 I: Peak With Markers for Background Subtraction (a single marker on the low energy side of the peak, a pair of markers on the high energy side), II: Li terature: D.A. Shirley: Phys. Rev. B5, 4709 (1972) 195 xiv 7.9 Potentiometric Responses of Bare Platinum Electrodes Treated as Indicated in Table 7.2 in Phosphate Buffer pH of 7.4, I: Bare Platinum Surfaces, II: Bar e Platinum Surfaces After Addition of n-De cyl Alcohol 203 7.10 Potentiometric Responses of Bare Platinum Electrodes Retreated and Tested in Phosphate Buffer pH of 7.4, I: Bare Platinum Surfaces, II: Bare Platinum S urfaces After Addition of n-Decyl Alcohol 204 7.11 Potentiometric Responses of Bare Platinum Electrodes Treated and Measured in 200ml pH 7.4 Phosphate Buffer After Addition of Different Amount of I: Glu cose, II: n-Decyl Alcohol 207 7.12 Comparisons Between X-ray PhotoElectron Spectra of Platinum Wire I: No.5, II: No.6, III: No.7 208 7.13 XPS Results of Platinum Wires No.5, No.6, No.7 (Narrow Scan) I: Oxygen level, II: Carbon level, III: Platinum level 210 7.14 Comparisons Between X-ray Electrophoton Spectra of Platinum Wire No.3, No.14 and No.25 211 7.15 XPS Results for Electrode No.14 Before and After Sputtering. Spectra 217 Rep-resents Before Sputtering and Spectra 223 Represents After Sputtering 21 3 7.16 Comparison of XPS Results of Electrode No.4 and Electrode No.4 216 7.17 XPS results for comparing the spectra of electrode No. after exposure to a U H V for 3 weeks, Spectra 212 represents the data 217 7.18 I: Contarnination Level and AVoltage Observed After Addition of Decanol to Treated Platinum Surfaces, II: Oxidation level and AVoltage Observed After Addition of Decanol to Treate d Platinum Surfaces 218 C l N M R Spectra of Glucose 298 C.2 L N M R Spectra of KN03-Glucose Complex, II-.NMR Spectra of Ca(AC)2-Glucose Complex 299 xv C.3 I:NMR Spectra of Fe50 4-Glucose Complex, II:NMR Spectra of KN03-Glucose Complex 300 C.4 I:NMR Spectra of Cu(Ac)2-Glucose Complex, II:NMR Spectra of 5aC7 2-Glucose Complex 301 C.5 I:NMR spectra of £a(Ac) 3 -Glucose Complex, ILNMR Spectra of CuS04-Glucose Complex 302 C.6 N M R Results of 5aC72-Glucose Complex Prepared in 1985 303 C.7 I:NMR Spectra of Aliquat in Decanol with BaCl2 304 C.8 T.NMR Spectra of Aliquat in Decanol with aD-Glucose 305 C. 9 I:NMR Spectra of Aliquat in Decanol with Ba-Glucose Complex 306 CIO I:NMR Spectra of Aliquat in Decanol . , 307 C l l I:NMR Spectra of N-Decyl Alcohol 308 D. l Transient Response and Steady State Response to Glucose and KCl Addition for I: Electrode 1,2,3 II: Electrode 4,5,6 III: Electrode 7,8,9 351 D.2 Transient Response and Steady State Response to Glucose and KCl Addition for I: Electrode 10,11,12 II: Electrode 13,14,15 III: Electrode 16,17,18 35 2 D.3 Transient Response and Steady State Response to Glucose Addition for I: Elec-trode 1,2,3 II: Electrode 4,5,6 III: Electrode 7,8,9, IV: Electrode 10,11,12 , V:Electrode 13,14,15, Day 2 353 D.4 Transient Response and Steady State Response to Glucose Addition for I: Elec-trode 1,2,3 II: Electrode 4,5,6 , Day 3 354 D.5 Transient Response of I: Electrode 7,8,9 II: Electrode 10,11,12, III: Electrode 13,14,15, Day 2 355 D.6 Transient Response of I: Electrode 19,20,21, II: Electrode 22,23,24, III: Electrode 25,26,27, IV: Electrode 28,29,30, V : Electrode 31,32,33, VI: Electro de 34,35,36 . 356 D.7 Transient Response of I: Electrode 28,29,30, II: Electrode 31,32,33, III: Electrode 34,35,36 , Day 3 357 xvi D.8 Transient Response of I: Electrode 40,41,42, II: Electrode 43,44,45, III: Electrode 46,47,48, IV: Electrode 49,50,51, V: Electrode 52,53,54 , VI: Electr ode 55,56,56a, Day 2 358 D.9 Transient Response of I: Electrode 37,38,39, II: Electrode 40,41,42, III: Electrode 43,44,45, IV: Electrode 46,47,48, V: Electrode 49,50,51, VI: Electro de 52,53,54 . 359 D.10 Transient Response of I: Electrode 55,56, II: Electrode 57,58,59, III: Electrode 60,61,62, IV: Electrode 63,64,65, V: Electrode 66,67,68, VI: Electrode 69,70,71 . 360 D . l l Transient Response of I: Electrode 72,73,74, II: Electrode 75,76,77, III: Electrode 78,79,80, IV: Electrode 81,82,83, V: Electrode 84,85,86, VI: Electr ode 84,87,88 . 361 D.12 Transient Response of I: Electrode 90,91,92, II: Electrode 93,94,95, III: Electrode 96,97,98, IV: Electrode 99,100,101, V: Electrode 101,102,103, VI: E lectrode 104,105,106 362 D.13 Transient Response of I: Electrode 60,61,62, II: Electrode 63,64,65, III: Electrode 66,67,68, IV: Electrode 84,85,86, V : Electrode 90,91,92, Day 2 36 3 D.14 Transient Response of I: Electrode 107,108,109, II: Electrode 110,111,112, III: Electrode 113,114,115, IV: Electrode 116,117,118, V: Electrode 119,120, 121, VI: Electrode 122,123 364 D.15 Transient Response of I: Electrode 124,125,126, II: Electrode 127,128,129, III: Electrode 130,131, IV: Electrode 132,133,134, V : Electrode 135,136,137, VI: Elec-trode 138 365 D.16 Transient Response of I: Electrode 139, II: Electrode 140, III: Electrode 141, IV: Electrode 142,143,144, V : Electrode 145, VI: Electrode 146,147 366 D.17 Transient Response of I: Electrode 148, II: Electrode 149,150,151, III: Electrode 152, IV: Electrode 153, V : Electrode 154,155,156, VI: Electrode 157,1 58 367 D.18 Transient Response of I: Electrode 159,160,161 368 D.19 Comparison of Experimental and Calculated Peaks for Electrode No.l and Elec-trode No.3 396 xvii D.20 Comparison of Experimental and Calculated Peaks for Electrode No.2 and Elec-trode No.4 397 D.21 Comparison of Experimental and Calculated Peaks for Electrode No.5 and Elec-trode No.7 398 D.22 Comparison of Experimental and Calculated Peaks for Electrode No.19 and Elec-trode No.4 (Rec) 399 D.23 Comparison of Experimental and Calculated Peaks for Electrode No.14 and Elec-trode No.25 400 D.24 Comparison of I: Initial, II: Steady State, III: Final After Addition of N-Decyl Alcohol Versus Oxidation Level (Pttot/Pt4f) , Retreated Platinum Elect rodes . 401 D.25 Comparison of I: Initial, II: Steady State, III: Final After Addition of N-Decyl Alcohol Versus Oxidation Level (Pt(OH)2/pt4f) , Retreated Platinu m Electrodes402 D.26 Comparison of I: Initial, II: Steady State, III: Final After Addition of N-Decyl Alcohol Versus Oxidation Level (PtO/Pt4f) , Retreated Platinum Electro des . . 403 xviii List of Tables 2.1 Comparative Results of Barium and Chloride leaching test [86] 35 2.2 Comparative Results of Glucose Leaching [86] 35 3.1 N M R Analysis of Glucose and Glucose Salts 57 3.2 List of Materials Used 58 3.3 Hydrogel Formations 59 3.4 Polyacrylamite Formations 61 6.1 Electrode Description 102 6.2 Stability of Coated Wire Sensors Before and After Addition of Glucose. (Pt Cleaned with Acetone) 107 6.3 Effect of Testing Solution on Electrode Response 118 6.4 Description of the Levels Used for Air Dried Electrodes 120 6.5 Description of Air Dried Electrodes 122 6.6 Stability of Coated Wire Sensor After Addition of Glucose. (Platinum Treated with HNO3 Acid and flame treated) 126 6.7 Stability of Coated Wire Sensors (Platinum treated with HNO3 Acid and Ther-mally Flamed 131 6.8 The Effect of Buffer Change on Electrode Response 139 6.9 Conditioned Electrode Description 145 6.10 Description of Conditioned Electrodes 146 6.11 Stability of Air Dried Coated Wire Sensor treated with HNO3 Acid and Ther-mally Flame Treated 153 xix 6.12 Stability of Conditioned Coated Wire Sensor Treated with HN03 Acid and Thermally Flame Treated 153 6.13 Stability of Air dried and Conditioned Coated Wire Sensor Treated with HN03 Acid and Thermally Flame Treated (Decanol Electrodes) 160 6.14 Potentiometric Behavior of Treated Platinum Wires 164 6.15 Effect of Addition of Decanol to Treated Bare Platinums 168 7.1 Atomic Orbital Nomenclature [155] 178 7.2 History of the Platinum Electrodes 189 7.3 Computer Output Results of a Typical Run, a) (Curve Fitting), b) (Integrated Values) 198 7.4 Area Ratio's For Platinum Electrodes (Integrated value) 200 7.5 Oxidation Ratio's For Platinum Electrodes (Curve fit value) 200 7.6 Potentiometric Response of Bare Platinum Electrodes Treated as Indicated and Measured in Air-Saturated pH 7.4 Phosphate Buffer [114] 205 7.7 Comparisons Between Treated and Retreated Platinum Surfaces According to Levels of Contamination and Oxidation 222 A. l Electrode Description 248 B. l The Effect of BaCl-i Concentration on Electrode Response 258 B.2 The Effect of Ba-Glucose Complex on Electrode Response 260 B.3 The Effect of Buffer Change on Electrode Response 264 B.4 The Effect of Life Time on Electrode Response 271 B.5 The Effect of Changing Measuring Solution on Electrode Response 274 B.6 Decanol Addition to Electrode Response 277 B.7 The Effect of the Sensor Intervariabihty 279 B.8 The Effect of Glucose Addition on Electrode Response 281 B.9 The Effect of Conditioned Electrodes on Electrode Response 288 xx B.10 The Effect of Oxides on Electrode Response 293 C l N M R Analysis of Glucose and Glucose Salts 309 D . l Experimental Raw Data . . • 311 D.2 Voltage-Time Responses of P V C Membrane Electrodes 369 D.3 The Voltage-Time Responses of CuCl2(Se2), Ba - Glucose (Se3), 5aC72(Se4), Membrane Electrodes 370 D.4 The Voltage-Time Responses of Cu(Ac)2{Se5), and FeC/ 3(Se6) Membrane Elec-trodes 370 D.5 The Voltage-Time Responses of Polyacrylamide Gel Membrane Electrodes . . . . 371 D.6 The Voltage-Time Response of a Glucose Electrode(Se28) in Comparison to Aliq/Dec Electrode(Se29) and Ba-Glucose Electrode (Se30) 372 D.7 The Voltage-Time Responses of Glucose Electrodes 373 D.8 The Voltage-Time Responses of a Glucose Electrode(Se33) in Comparison to a Ba-Glucose Electrode(Se34) 374 D.9 The Voltage-Time Responses of a Ba-Glucose Electrode(Se35) in Comparison to Glucose Electrode(Se36) and Aliquat to Decanol(Se37) 376 D.10 The Voltage-Time Responses of C a-Glucose (Complex No.2)(Se38) with Ca-Glucose(Se39) in Comparison to K-Glucose(Complex No.6)(Se40) 377 D . l l The Voltage-Time Responses of La-Glucose Electrodes(Complex No.3) 380 D.12 The Voltage-Time Response of Ba-Glucose Electrode(Complex No. l)(Se43) . . . 383 D.13 The Voltage-Time Responses of Cu-Glucose Electrodes(Complex No.4)(Se44,Se45)384 D.14 The Voltage-Time Responses of Fe-Glucose Electrode(complex N o.5)(Se46) in Comparison to Ba-Glucose Electrodes(Se47,Se48) 387 D.15 The Voltage-Time Responses of Electrodes Tested for XPS 388 D.16 The Voltage-Time Responses of the Retreated Surfaces for XPS Studies . . . . . 389 D.17 The XPS Computer Results of Platinum Surfaces Presented in Chapter 6 . . . . 390 xxi E . l Specifications and Sensitivity of I: M o d e l 617 K e i t h l y Electrometer , II: M o d e l 616 K e i t h l y Electrometer 405 E.2 Specifications and Sensitivity of I: M o d e l 640 K e i t h l y Electrometer , II: M o d e l 177 K e i t h l y Electrometer 406 xxi i Dedication I dedicate this thesis to my loving father "Morteza Sharareh" and my beautiful and loving mother "Nahid Sharareh" for giving me the opportunity for bettering myself. xxiii Acknowledgements I would like to take this opportunity to thank my research supervisor Professor K.L.Pinder for his guidance, support and inspiration during this work who, since the beginning, gave generously his time and effort towards the creation and correction of this work. I would like to thank Professor K . L . Mitchell and Dr. Philip Wong for giving us permission for using the XPS equipment and also their expert suggestions in the area of surface science. I am also greatful to Professors J.L.Bert and C.W.Oloman for acting as members of the supervisory committee. Secondly, I would like to thank the chemical engineering workshop, and chemical engineering stores especially Mr. Horace Lam, for their cooperation and kindness during my stay in the department. I would also like to thank Dr. Clive Bereton for his suggestions and being a friend. I would like to thank a few special collegues whose moral support and friendship has helped me concentrate and complete this thesis: Dr. David Taylor, Dr. Elrnarghani Besher, Mr. Clive Chappie, Ms. Jyoti Dhar, and Ms. Linda Hou. I would like to also thank all my other friends for believing in me. Last but not the least, I would like to thank my three wonderful brothers; Shahroek, Shah-poor and Shaliin and my sister Shohreh for their love and being there for me. I would like to mention families and friends which have made my stay pleasant in Vancouver. Mr and Mrs. Hossein Daha, Mr and Mrs. Hossein Pazand and Mr and Mrs. Faraydon Hadad, Kathy and Keyvan Hadad, Mrs. Lily Bozorgmanesh, and Mr. Ah Fathi. Their love and kindness will always be remembered. xxiv C h a p t e r 1 I N T R O D U C T I O N Diabetes is a disease characterized by relatively high blood glucose levels due to insulin defi-ciency. Diabetes mellitus is characterized by a relative or absolute insulin deficiency. It is based on an irregular or insufficient release of insulin by the beta cells causing an excessive blood sugar level [1]. Diabetes inspidus is characterized by the constant excretion of a large volume of highly dilute urine. Therapy of diabetes commonly involves several injections of insulin daily. In this way blood sugar can be matched only incompletely to normal requirements. A certain improvement in the therapy has been achieved by portable units that function as open-loop systems without a glucose measuring device. A prerequisite for further extention of the therapy is the miniaturization of the portable devices to a size which can be implanted and operate at least for one year. Intensive efforts are being devoted to developing a miniaturized long-term stable and selec-tive sensor for continuously monitoring the glucose level in blood or extracellular body fluids. The utilization of a glucose sensor within an implanted closed-loop artificial pancreas neces-sitates the conversion of the glucose concentration into an electrical signal on the basis of an intra-or extravascular measurement. One type of a glucose sensor which has been proposed for this use is based on the enzymatic oxidation of glucose by glucose oxidase to gluconic acid: Glucose + O2 + H20 =^= Gluconic acid + H2O2 " " (1-1) The determination of the glucose concentration is based either indirectly on the formation of hydrogen peroxide or by oxygen consumption or the formation of hydrogen peroxide using a 1 Chapter 1. INTRODUCTION 2 polarographic electrode covered by a membrane composed of the immobilized enzyme G O D [2]. A limitation to long-term application to date is the irreversible loss of enzymatic activity and the stability of the enzyme/membrane system. These effects are very important when recalibrating the electrodes. A new approach for monitoring blood glucose levels has been reported which results in a lin-early decreasing potential with increasing glucose concentration without utilizing an oxidation reduction reaction [3]. The life time of these electrodes was reported to be a few days, a few weeks, and nine weeks according to whether the electrode was coated with no hydrogel, positive or negative hydrogel respectively. By using such membranes it was possible to determine which component has become associated with glucose. It was stated that this type of glucose sensor could provide opportunities for a much greater degree of control of glucose levels in diabet-ics, and could be possibly used for implantation. Therefore, this research was undertaken to gather more experimental data to gain a fuller understanding and to explain the behavior of a rniniaturized coated wire glucose sensor. The purpose of this work was to study the effects of implantation of a needle type electrode in subcutaneous tissue. Needle type glucose sensors were prepared and tested and it was observed that the behavior of the electrode was not due to changing glucose concentration. The research objective was then changed to an investigation of the electrode behavior to understand what was actually causing the electrode potential change. The work described hi this thesis consists of: 1. A literature review on: • Previous studies of glucose sensors and their potentiometric and amperometric re-sponses to changing glucose concentration in either vitro or vivo studies. • Previous studies on ion selective electrodes based on liquid ion exchangers and coated wire ion selective electrodes. • Previous studies on preparation of complexes based on sugar-metal interaction and its application to metal-cation complex formation. Chapter 1. INTRODUCTION 3 • The electrochemical studies of platinum surfaces. • Study of X-ray photoelectron spectroscopy for surface reaction studies. 2. Preliminary experimental approach, which contains the overall objective of the thesis, and describes experimental set up and procedure followed in the preHminary experimental approach. 3. Preparation of the electrodes, which describes different types of electrodes, their prepa-ration and the results obtained in comparison with each other. 4. Attempts to reproduce the literature results which are reported for potentiometric mea-surements. 5. New approach to the potentiometric observation to describe the mechanism of the tran-sient behavior of electrodes. 6. X-ray photoelectron spectroscopy. An introduction to the area, equipment used and description of the results obtained from surface analysis. 7. Conclusions and recommendations for further research. 8. Appendices containing • Description of Electrodes. • Statistical Analysis. • N M R Analysis. • Raw Data. • Specifications and Sensitivity of the Meters Used. Chapter 2 L I T E R A T U R E R E V I E W During the decade from 1974-1984 the expectation that bio- engineering might make an impor-tant contribution to patient care was realized for several diseases, and especially for diabetes mellitus [4]. As with kidney disease, where technology has had its great impact through the development of an "artificial organ", it was in 1974 that two groups (Albisser et al [5],1974,Pfeif-fer et al.[6], 1974) published their early papers on the construction and testing of an "artificial pancreas". When it became possible to continuously monitor blood glucose concentrations and to ex-amine frequent blood samples in insulin dependent diabetics treated by conventional insulin injections, it was realized how difficult it is in most patients to achieve and maintain physio-logical blood glucose levels throughout the entire day. Much effort has been placed on evolving ways for improving diabetic control, in the sense of bringing metabolite levels closer to that of the nondiabetic. The principle contribution of engineering to this has been in the insulin delivery systems [4]. The recent advances in electronic miniaturization and computers were obviously the key to their application to this problem. Attempts have been made to develop an implantable glucose sensor. By coupling such a sensor to an automatic insulin delivery system, an artificial pancreas can be developed [7,8]. By means of a feedback control, the organ could maintain desirable glucose homeostasis in diabetic patients. This sensor could also be used in measuring the glucose concentration in subcutaneous tissues which can be compared to the blood glucose concentration in the entire body [9]. Some clinical aspects of this disease will be briefly described to explain the necessity for an improved insuhn delivery system. 4 Chapter 2. LITERATURE REVIEW 5 2.1 Diabetes 2.1.1 Types of Diabetes Certain observations seemed to support the idea that extrapancreatic influences are the cause or precipitating factors of diabetes. These include various hormones, trauma, emotional stress, obesity, viral infection or a reduced binding ability of insulin at the peripheral target cells due to a decrease in the number of receptors [10]. Other possible extra pancreatic factors that may cause diabetes are heredity, excessive destruction of insulin, and the presence of substances that are either of a protein nature or are associated with protein that bind or hold insulin and thus oppose its action [10] . A simple explanation could be that diabetes is a chronic systemic disease and is a result of insulin deficiency. Long term complications of the disease include retinopathy, nephropathy and neuropathy. There are two types of spontaneous diabetes in humans: • Juvenile-onset diabetes, insulin dependent diabetes mellitus, Type I. • Maturity-onset diabetes, non-insulin dependent diabetes mellitus, Type II. Persons in classification of type I are dependent upon injected insulin to prevent the ac-cumulation of toxic ketonic acids in their blood. In most cases, the bulk of the /3-cells in the pancreas have been destroyed. Many of these patients require daily injections for the mainte-nance of life. Diabetes mellitus is a disease due to metabolic disturbance in which the body tissues are unable to utilize or store glucose (sugar) normally [5]. As a result, greater than normal amounts of unused glucose are found in the blood stream and some body tissues. Even-tually, the level of blood glucose becomes sufficiently high to cause the kidneys to lose glucose in the urine. Traditionally, people with diabetes are treated with one or more daily injections of insulin which are administered into the subcutaneous tissue of the abdomen, arms, legs or buttocks. Diabetes is one of the leading causes of death by disease in the United States. Life span from the time of onset of type II diabetes is reduced by 30% for the average patient. The life span of insulin- dependent, or juvenile onset, diabetes is reduced considerably more. According Chapter 2. LITERATURE REVIEW 6 to the National Commission on diabetes, this disease now affects 5% of the population in the United States (approximately 10 million people), and it is estimated that a similar number of cases remain undiagnosed. 2.1.2 Clinical Aspects Of Diabetes In most patients it is sufficient to prevent the acute symptoms of hyperglycaemia, although excess msulin at certain times during the day can cause, instead, hypoglycaemia. Here the patient experiences the symptoms of low blood glucose levels which might include sweating, pins and needles in the hps and tongue, hunger, headache, double vision, slurring of speech, confusion and eventually coma unless food or glucose is taken [4]. Temporary increase in insulin requirements and thus hyperglycaemia can also occur during intermittent illness, menstruation and stress. What is Insulin Insulin is a protein substance produced in specialized cells clustered within the pancreas (Islets of Langerhans). It is released into the blood stream by these cells (beta cells) when they are stimulated by a rise in blood glucose levels (see Figure 2.1) . It is thought of as a chemical key that unlocks the doors of cells allowing glucose to enter. In the presence of adequate in-sulin, the cell bodies remove glucose from the blood stream and prevent the blood glucose level from becoming excessive (see Figure 2.2). Insulin is actually a particularly difficult drug to administer to patients. Because it has a relatively short intravenous half-life (3-4 min), it must be infused continuously to maintain constant blood levels. Moreover, it has multiple metabolic actions, each operative at a different dose level (Figure 2.3). The steepness of these various dose-response curves means that small changes in insulin concentration produce quite large changes in biological effect. And since blood glucose level is influenced by: 1. Food intake Chapter 2. LITERATURE REVIEW Decrease 1n blood sugar level lncre ase in blood sugar • Is vel Alpha cells secrete glucagon Glycogen tn Hver -* glucose & release 1 Increase In blood sugar level PANCREAS Beta cells secrete-1nsul.1n Decrease in blood sugar level Blood glucose stored in l iver as glycogen 1 Increase glucose consumption by cells for energy Figure 2.1: Hormonal Regulation of Glucose [1] Figure 2.2: The Effect of Ingestion of Carbohydrate Foods on the Concentration of Blood Glucose, Insulin Production and the Effect of Insulin on Glucose Entry Into Body Cells [7] Chapter 2. LITERATURE REVIEW 9 Pet cent m a x i m u m 'csponse Plasma insulin. pU/ml Figure 2.3: Dose-Response Curves for the Different Metabolic Actions of Insulin [2] Chapter 2. LITERATURE REVIEW 10 • Food quality • Irregularity of meals 2. Insulin 3. Exercise 4. Stress • Emotional • Illness accompained by fever • Injury leading to pain it is very important to measure the glucose levels accurately. All of these physiological re-quirements can be met by the mechanical delivery of insulin from an infusion pump, since it is possible to administer a constant but, when necessary, variable and accurately controlled dosage rate. Long term clinical treatment includes dietary restrictions and daily injection of insulin. 2.1.3 Treatment of Diabetes Injections of insulin 2-3 times daily cannot provide the delicate balance needed for body fuel metabolism to reach normal levels. Because of this,the patient might go through alternating phases of hyperglycaemia and hypoglycaemia. It is the irregular control over diabetes that current treatment offers that has prompted extensive research to find an alternate method that will provide the diabetic a normal life. Presently, there are several possibilities that researchers have focused their efforts on: 1. Transplantation of beta cells from a healthy individual into the diabetic. 2. Artificial endocrine pancreas which would mimic a normal organ. 3. Polymeric artificial pancreas or insulin delivery device. Chapter 2. LITERATURE REVIEW 11 4. "Hybrid" artificial pancreas as with (5 cells plasma, a polymeric device. Although attempts have been made to cure rats of diabetes with islet cell transplants, it is a more formidable task in larger animals due to graft rejection. Immunological suppressants may be given to increase the life of the transplant, but this cannot be done without side effects. Chambers have been made from semipermeable membranes which screen the transplanted cells from the immune system while allowing diffusion of glucose and insulin. Unfortunately, the chambers become severely limiting. Under special laboratory conditions that involve multiple feeding and regular injection of insulin, one can greatly improve diabetic control. However, the almost complete restoration of a normal metabolic profile can only be seen in patients who are connected to an artificial pancreas, a closed system for the continuous monitoring of blood sugar levels and delivery of calibrated amounts of insulin to the body throughout the twenty four hour day (Figure 2.4). Closed loop devices for use at the bedside are available at the present time for short term investigations. A device for long-term portable use is not available due to the lack of an adequate glucose sensing device. A continuous glucose sensor is needed to close the loop. Such a sensor should send a continuous electrical signal proportional to the blood glucose level to a controller which could set the delivery rate of insulin to the body. Different types of insulin delivery systems are known but not limited to: 1. The programmable implantable medication system. 2. The biostator. • Glucose analyzer. 3. Exac tech blood glucose meter. 4. Direct 30/30. Figure 2.5 shows an example of such systems. The programmable implantable medication system (telemetry system) is a device which can monitor glucose levels in subcutaneous tissues without blood withdrawal. The clinicians program the patient control unit, using a separate Chapter 2. LITERATURE REVIEW 12 GLUCOSE COMPUTER ANALYZER Figure 2.4: Schematic Diagram of the Artificial Endocrine Pancreas Outside the Body Figure 2.5: I: Programmable Implantable Medication System, II: The Biostator, III: Exac Tech Blood Glucose Meter, IY: Direct 30/30 Chapter 2. LITERATURE REVIEW 14 computer, defining a set of basel profiles with a cycle of from 1 to 60 days with a choice of prandial options. The patient control unit is used not only to communicate with the implant, but to transmit stored data from the implant or data relating to the patient's condition to the clinicians computer; either directly or over the phone. At the present time , the telemetry system operates with a needle type glucose sensor based on hydrogen peroxide measurements. Other types of needle sensors have been developed and their operation will be explained in later sections. The ideal characteristics of an implantable needle type glucose sensor are [11,12]: • It should be biocompatible. • It should be reliable, reproducible. • No or nunimum inhibition or poisoning. • No interference from other body chemicals. • It should be low cost. • It should be easily fabricated. • It should have fast response(approximately 1 minute). • It should measure continuously the tissue glucose level. • It should be rruniaturized. The clinicians are also expecting the new technology of biosensors to produce a perfect sensor which does not have any physiological problems including operating convenience, cost, size, and performance and diversity. The next section describes certain types of glucose sensors which appear to have special medical uses. Most glucose sensors are based on electrochemical and optical measuring tech-niques. Efforts are being devoted to the task of developing a miniaturized long-term stable and Chapter 2. LITERATURE REVIEW 15 selective sensor for continuous monitoring of the glucose level in blood or extracellular fluids, or subcutaneous tissues [13-20]. The aim of this research was to develope a needle type glucose sensor which preserves, in vitro and in vivo, characteristics suitable for tissue glucose monitoring [21-23]. The structure and membrane design of the intracorporeal glucose sensor, as well as its behavior in vitro will be briefly explained. In the next section the presently "developed electrodes" will be described and more detail will be given about the coated wire glucose sensor which is the major concern of the present study. 2.2 G l u c o s e S e n s o r s 2.2.1 E n z y m e E l e c t r o d e The enzyme electrode is one of the earliest type of glucose sensor developed. Enzymes are biological catalysts. They exhibit positive properties such as specificity and negative properties, denaturability or instability [24]. The enzyme electrode is also a sensitized electrode in that a basic electrode is overlayed with an enzyme layer to impart selectivity to a specific substrate. The substrates are various organic compounds which are acted on by the enzyme to either consume or produce an ion or gas which can be sensed by the electrode. The enzymatic reactions can be measured either amperometrically based on oxygen consumption, hydrogen peroxide production or potentiometrically. A m p e r o m e t r i c E n z y m e E l e c t r o d e s In 1962 Clark and Lyons [25] introduced an enzyme electrode device which was based on the enzymatic oxidation of glucose to gluconic acid in the presence of glucose oxidase (GOD) ac-cording to the following reaction: glucoteoxidase Glucose + 02 + H2O ^ Gluconic acid + H2O2 (2.1) Chapter 2. LITERATURE REVIEW 16 The oxygen uptake which was proportional to the amount of glucose present was measured by an Oi electrode. The search for an implantable glucose sensor was first suggested by Updike and Hicks [26]. They devised an electrode made by polymerizing a gelatinuous membrane over a polarographic oxygen electrode. Upon placing the electrode in contact with the biological solution or tissue; glucose and oxygen diffuse into the gel layer of immobilized enzyme. Figure 2.6 shows the operation of an enzyme electrode. The rate of diffusion of oxygen through the plastic membrane to the electrode was increased in the presence of glucose and glucose oxidase by the enzyme-catalyzed oxidation of glucose. From calibration curves of electrode response versus glucose concentration, the amount of glucose in whole blood was calculated. A special problem with these electrodes is the dependence on oxygen for operation and the variation in current output with changing Po2. This is particularly important in the context of monitoring glucose in vivo, because glucose is in molar excess over oxygen by several orders of magnitude, particularly in diabetic patients. Even when a second electrode is included without glucose oxidase and a difference current measured, the output is not entirely independent of Po 2 changes. In the literature it is stated that Updike [27] has employed another approach to the problem by enclosing the sensor in a thin hydrophobic membrane of cross-linked albumin or nylon which is more permeable to oxygen than glucose. Gough and workers [28] worked on the construction of a two dimensional oxygen electrode which measures both axially and radial (diffusion. Such a diffusion regime ideally allows the enzymatic and electrochemical reactions to occur uniformly in the plane of the membrane and across the electrode surface, with substrate concentration gradients developing perpendicular to this plane. Figure 2.6 shows the principle of operation of this type of glucose sensor. Clark designed an enzyme electrode based on the amperometric determination of hydrogen peroxide [29]. When the polarizing voltage is set at +600 mv, the following reaction occurred at the anode: H202 ^ 2H+ + 02 + 2e" (2.2) Chapter 2. LITERATURE REVIEW 17 Cathode of oxygen electrode I AXIAL DIFFUSION OF GLUCOSE AND OXYGEN I I Figure 2.6: I: Principle of Glucose Enzyme Electrodes Operating by Detection of Glucose Consumption, II: Schematic Drawing of Diffusion of Glucose and Oxygen Chapter 2. LITERATURE REVIEW 18 Other researchers such as Lobel and Rishpon [30] also designed an enzyme electrode based on the amperometric determination of hydrogen peroxide. The glucose measuring system consists of three electrodes: a glucose electrode, a counter electrode and a reference electrode connected to a potentiostat. The glucose working electrode consists of a fine platinum net anode pressed lightly against the glucose oxidase membrane. Between the enzyme membrane and the anode, a thin layer of gold was placed to improve the direct coating technique. Rehwald and Geibel [31] designed a micro-enzyme electrode for the continuous monitoring of glucose concentration in isolated tubule segments. This micro-electrode detects hydrogen peroxide, which is a product of oxidation of D — glucose by glucose oxidase immobilized at the tip of this electrode. The principle of such an enzyme electrode was first described by Clark and Lyons and Updike and Hicks [25-26]. A glass tube is melted, fused and pulled to a tip of 10-30 fim. The tip of the electrode is then dipped into a solution of a glucose oxidase, and immediately coated with a thin layer of cellulose acetate in acetone and cyclohexane to protect the electrode from interferences by high molecular weight substances and to stop the enzyme from diffusing away. Potentiometric Biosensors Potentiometric biosensors are based on the principles used in ion selective electrodes. The production of gluconic acid and hydrogen peroxide by the catalytic oxidation of glucose can be measured potentiometrically. The potentiometric changes are governed by the Nernst equation E' = E0 + (R!T/ZF)ln[analyte} (2.3) Where: E' = Measured Potential(V) E0= Constant Potential(V) R' = Gas constant (8.314 J/mole°K) T = Temperature °K Chapter 2. LITERATURE REVIEW 19 Z = Charge F = Faraday (96487 coulombs) Nagy et.al [32] described a potentiometric glucose sensor in which coimmobilized glucose oxidase and peroxidase were used to produce hydrogen peroxide, and then to catalyze the conversion of iodide to iodine: , Peroxidase _ . H202 + 2I~ +2H+ ^ 2H20 + I2 (2.4) The decrease in potential of the iodide-selective electrode was therefore proportional to the log of the glucose concentration. Wingard's group [33,34] used an enzyme that catalyzes the reaction of /3-D-glucose plus 02 to produce gluconolactone plus hydrogen peroxide. The redox potential of these electrodes is thought to be dependent on the glucose, oxygen and hydrogen peroxide concentrations in the solution and on functional groups on the platinum or carbon surface. It is suggested that at the platinum electrode a reduction is occurring, although other one-electron transformations very likely are present: H202 —>02 + 2H+ + 2e" (2.5) Pt{OH)2 + 2 H + + 2e~ —+ Pt + H20 (2.6) It was also stated that platinum oxide formation is thought to begin with O H groups attaching irreversibly to the surface layer of the platinum. The exact mechanism was unclear and it was suggested that a better understanding of platinum-oxygen-water interaction needs to be investigated. Unfortunately, implantable electrodes are still of limited use because of problems of biocom-patibility and electrode lifetime [29-41]. Other studies include enzyme electrodes which will not be discussed here and are not limited to the publications listed below [35-46]. Also the issues to Chapter 2. LITERATURE REVIEW 20 be resolved before implantation is considered are glucose oxidase membrane stability for a pro-longed time, electrode response variations, power requirements, calibration, and implantation site. Long term physiological side effects and electrode response, and long term problems due to changes in the immune system and repair mechanism of the body must also be considered. 2.3 Optical Electrode The fiber optic bioafnnity glucose sensor [47-49] is another type of sensor which detects glucose without using an oxidation-reduction reaction. The advantage of these sensors based on optics, are electrical isolation from the patient, absence of a reference electrode, and easy miniaturiza-tion. A short length of hollow dialysis fiber is connected to a fluorometer by a single optic fiber. The glucose-binding lectin concalvalin A, is immobilized on the inner surface of the dialysis membrane. The level of free fluorescein is measured by the fiber optic and is proportional to the glucose level over a range from 50-400 mg/dl. It is an optical approach which is based on the competitive binding of glucose with fluorescein labeled dextran(FITC-dextran) and concavalin A. Figure 2.7 shows the mechanism of how the competetive reaction takes place. Chapter 2. LITERATURE REVIEW 21 r )Con A Con A I I iF-Dex G Dialysis membrane , Rbre optic F-Dex Seal" G Excitation - Emte won Figure 2.7: The Fiber Optic Bioaffinity Sensor Chapter 2. LITERATURE REVIEW 22 ConA + Glucose ^ ConA — Glucose (2-7) ConA + FITC - dextran ^ ConA - FITC - dextran (2.8) Increasing the concentration of glucose will displace the first reaction to the right, forming more ConA-Glucose, and displace the second reaction to the left,releasing the FITC-dextran from ConA. By measuring the fluoresence of the bulk solution, the glucose concentration can be determined. Sensitivity to glucose in the physiological range was found, but further work was necessary to improve the sensitivity, the response time , and to reduce the sensitivity to amino acids and urea in the physiological range. 2.3.1 Sensors Based on Direct Electron Transfer Amperometric glucose sensors which are relatively oxygen insensitive have been constructed using 1,1'-dimethyl ferrocene to mediate glucose oxidase catalysed electron transfer between glucose and carbon based electrodes [50-52], The major potential advantage of the ferrocene-mediated glucose sensor is that transfer of electrons from glucose to the base electrode occurs without the participation of oxygen. Glucose + Godox —> Gluconic acid + GodTed (2-9) Godred + Medox —» Godox + Medred (2.10) Godox and Goa\.ed are the oxidized and reduced forms of glucose oxidase, and Medox and Medreci 3 X 6 the redox forms of the artificial coenzyme or mediator. The reduced mediator can be reoxidized at a base electrode and the current flowing will be proportional to the glucose concentration. Cass et al [53]. described a relatively large prototype amperometric glucose oxidase enzyme electrode in which the substituted ferrecinium ion acts as an alternative electron accepter to Chapter 2. LITERATURE REVIEW 23 oxygen according to the reaction sequence: Glucose + Godox —» Gluconolactone + GodTea- (211) Godred + FeR+ —> Godox + 2FeR + 2H+ (2.12) 2FeR —• 2FeR+ + 2e~ (2.13) Mor and Guarnacuia [54] used hexacyanoferrate(III) as an electron acceptor dissolved in the sample solution but with glucose oxidase trapped at the working electrode. There has been great interest in constructing biosensors with organic metal salts. Glucose oxidase adsorbed onto organic conducting salt will give catalytic currents at potentials higher than -0.08V. Another type of electrode is the conducting polymer which has a higher redox group density, better surface stability of the adsorbed polymer and reduced tendency for denaturating of the immobilized enzyme of the electrode, through reduction in surface tension. Nicolo and Lowe [55] described the entrapment of glucose oxidase in a polypyrrole matrix electrochemically deposited on a printed platinum electrode. The generation of an extremely reactive radical cation that reacts with neighbouring pyrrole species produces a polymer with a net positive charge. 2.4 I m p l a n t a t i o n a n d N e e d l e T y p e S e n s o r s Technology is now available to miniaturize and retain a conventional enzyme assay on the tip of a needle, the signal being integrated by pocket sized portable devices. Some preliminary studies will be cited to explain how these electrodes operate and were the measurments took place. Abel [56] and co-workers developed an implantable sensor measuring the interstitial glu-cose concentration. The prepared sensor consisted of self-manufactured polarographic H2O2 electrode. After implantation under local anaesthesia IG (Insuhn-Glucose) monitoring was immediately initiated. Chapter 2. LITERATURE REVIEW 24 Schichiri [57] and co-workers developed a needle type glucose sensor in which hydrogen-peroxide was generated by immobilized glucose oxidase, which generated a very weak direct current depending on the glucose concentration in surrounding fluid. Continuous glucose mon-itoring of five diabetic subjects for 99 hours showed a significant correlation existed between the subcutaneous tissue glucose concentration and plasma glucose concentration which was measured. Schichiri [58] improved the sensor's output by using a compatible membrane of alginate-polylysine-alginate(APA) around his original sensor. Characteristics of this membrane were examined during implantation in subcutaneous tissue. However, during monitoring, the sensor showed a gradual decline of the output responded to glucose. The electrode responded for approximately 10 days. Gough [59] and co-workers developed an enzyme electrode based on a layer of immobilized glucose oxidase and catalase coupled to a membrane-covered oxygen electrode. Sensors were implanted in anesthetized animals via the Formal vein and advanced into the Inferior Vena Cava. Experiments lasted approximately 10 hrs and it involved 2 to 4 glucose perturbations. A potentially implantable glucose sensor with direct electron transfer was developed by Pickup and Claremont [60]. The glucose sensor was relatively oxygen insensitive and was con-structed using l,l'-dimethyl ferrocene to mediate glucose oxidase catalyzed electron transfer between glucose and a carbon base electrode. Short term in vivo studies have been performed with the electrode implanted in subcutaneous or intermuscular tissue of rabbits and the sub-cutaneous tissue of anesthetized pigs. A correlation between blood glucose concentration and electrode response was found after an intravenous glucose load was applied. The problems associated with in vivo studies are surface fouling and as a result the research effort is shifting to the investigation of biocompatable membranes for these devices. The most urgent in vivo problem for solution is the drift of glucose sensor output with time. If this loss of sensitivity were predictable for a given electrode, it could be corrected, but, unfortunately, it usually is not. The main deposit found on needle-type electrodes after a few days in the Chapter 2. LITERATURE REVIEW 25 subcutaneous tissue was a coating of protein which acted to block the diffusion of glucose. 2.5 Ion Selective Electrode Ion selective electrodes are another type of biosensor which respond to a specific ion or protein in the body. They only respond to the free ionic species in solution. Such an electrode constituents a galvanic half cell, consisting of an ion selective membrane, an internal contacting solution or a solid contact(all solid state configuration), and an internal reference electrode. For convenience, these elements axe housed in a single body. The other half cell is an external reference electrode dipping into a reference electrolyte. If the ionic species is not distributed throughout the sample, the electrode will respond only to that ionic activity existing at the sensing tip of the electrode. The response of an ion-selective electrode may also be affected by the presence of proteins and other organic constituents which can coat the surface of the electrode membrane [61-68]. This is a complex problem which requires considerable research efforts to elucidate all of the possible interactions. Certain types of these electrodes work according to the liquid ion exchange theory which will be explained later. These devices may be classified into the following categories: 1. Solid state membrane electrodes. 2. Glass membrane electrodes. 3. Liquid ion-exchanger membrane electrodes. 4. Neutral carrier liquid membrane electrodes. 5. Special arrangments such as gas-sensitive electrodes, enzyme electrodes, etc. 6. Ion selective field effect transistor. The device which will be the focus of this thesis is the type which operates with the liquid ion exchanger membrane. Chapter 2. LITERATURE REVIEW 26 2.5.1 L i q u i d Ion Exchangers In addition to research on oxygen/enzyme based electrodes, several attempts have been made to design an ion selective electrode. Liquid ion exchangers are solutions of substances consisting of an ionogenic group which is attached to an organic molecule of proper size and configuration to make these compounds and their salts very sparingly soluble in aqueous electrolyte solutions [69]. Such systems were first studied systematically in 1933 by Beutner. These solutions are denoted as liquid ion exchangers because their overt behavior when in contact with electrolytic solutions is similar in many respect to that of the well known solid ion exchangers [61]. A difference between the solid and liquid exchangers is that in the liquid exchanger, the molecules carrying the ionogenic functional groups diffuse freely within the exchanger phase. Thus, the ionogenic functional compounds of the liquid exchangers act as "carriers" for the counter ions with which they are associated [70]. The theory behind the operation of liquid ion-exchange membranes has been developed based on the concept of zero-current potential. One of the first liquid membrane systems used diesters of phosphoric acid as a charged exchanger to form a selective Ca2+ sensor [71]. The ester functionalities were 8 to 16 carbon atoms long to render the ion exchanger hydrophobic. The successful fabrication of such membrane electrodes presents problems, many of which are mechanical. Thus the liquid ion exchanger, has certain requirements: 1. Any mixing of the phases must be minimal. 2. The ion exchanger must not be too soluble in sample solutions. 3. High viscosity. 4. Low cost [64]. The effective life times of such electrodes are effected by the gradual loss of liquid ion exchangers from the double reservoir through the cellulose acetate membrane. Sixteen organic and iono-genic salts of methyltricaprylyl ammonium ion(Aliquat 336S) in 1-decanol mediator have been Chapter 2. LITERATURE REVIEW 27 examined as prospective liquid ion exchangers using Orion model 92-20 electrode bodies. The body of a 92-20 model electrode can be adapted to test any commercial, liquid ion exchanger. Thus, the electrode after use is completely drained and refilled again with the new liquid ion exchanger. Selectivity of liquid ion exchange membrane electrodes depends on the following generalized equilibrium scheme lying far to the right [72]: RN + Mn+^RM + Nn+ (2.14) where: R = Large mobile organic exchanger in the membrane. M = Measuring ion. N = Ionogenic group. Problems arise when the membrane material reacts with an interferring material to give a water soluble complex. The liquid membrane electrodes contain the salts of methyl tricaprylyl ammonium ion (Ali-quat 336S) dissolved in 1-decanol which function effectively as organic phase components in liquid- liquid membrane electrodes for the determination of the respective anions in the con-centration range of 10 - 1 to 1 0 - 4 M : perchlorate, chloride, bromide, iodide, nitrate, sulfate, thiocyanate, acetate, formate, oxalate, and benzoate [73]. The electrode consists of an organic calcium electrode barrel in which the liquid phases prepared as described, were placed and an Orion 92-20 membrane was used to separate the organic phase from the test solution. The organic phase consisted of a 10v/v% solution of Aliquat 336S in 1-decanol which was converted to the proper form by repeated shaking with an aqueous solution of the sodium salt of the appropriate anion to effect the exchange [74]. Changes in initial potential were observed which require almost daily restandarization of all electrodes. It is noted that the life span of these electrodes is a month or longer if no mechanical problem such as leakage develops. Amino acid responsive electrodes based on the use of the quaternary ammonium salts in a liquid membrane phase have been prepared by Matsui and Freiser [75]. It was believed that the selectivity of these electrodes depends on the extractability of an ion pair involving Chapter 2. LITERATURE REVIEW 28 the interfering anion. For this reason, leucine, glutamic acid and alanine were used in the corresponding amino acid responsive electrodes. 2.5.2 Coated W i r e Ion Selective Electrodes The need to measure analytes in smaller sample volumes led to the development of the coated wire electrode (CWE) which uses components of the conventional ion selective electrode's (ISE) except that no internal aqueous filling solution is used. Instead, a conductor is directly coated with an ion responsive membrane (usually P V C based). This conductor can be metallic of any geometric shape (wire,disk,cylinder). These electrodes typically consist of a reference electrode (Ag/AgCl) immersed in an aque-ous reference solution in contact with a "membrane" which in turn contacts the solution under test. Among the simplest electrodes are fine metallic wires such as Ag and Cu that respond to changes in the activity of Ag+,Cu2+ respectively [76]. A wider variety of ions give poten-tiometric responses to electrodes in which a metal wire is coated with a poorly soluble salt of that metal, as with the Ag/AgCl electrode which responds to changes in activity of chloride ion. Although the salt layer is usually deposited by anodizing the metal electrode in a suitable electrolyte, recently it has been shown to be possible to incorporate the poorly soluble salt in a polymeric matrix, affixed to the metal wire, provided that the salt particles are in continuous contact from the outer surface of the polymer film to the metal surface underneath. If a solution of a salt or complex in a polymeric matrix such as polyvinyl chloride, epoxy resin, or methyl methacrylate painted on a platinum wire would function as an electrode responsive to one of the ions of the salt or complex, then a so-called liquid membrane electrode sensitive to a wide range of material could be produced in an inexpensive, compact form. Best results could be obtained after an initial conditioning, accomplished by soaking the electrodes for about an hour in either distilled water or in a dilute CaCl2 solution. At the early stages of this work, a simplified coated wire electrode was studied by Freiser and co-workers. Platinum wire coated with a P V C solution of calcium didodecylphosphate functions Chapter 2. LITERATURE REVIEW 29 as a calcium selective electrode [77]. Stworzewicz and Leszko prepared thiocyanate and chloride coated wire electrodes from the octadecyldimethyl benzyl ammonium chloride (0DBA-C1) or thiocyanate (ODBA-SCN) which was dissolved directly in the cyclohexanone solution of P V C [78]. A long chain aliphatic am-monium salt like Aliquat 336S was dissolved in decanol, and converted into an appropriate salt and mixed with a solution of P V C in cyclohexane. The platinum was dipped several times in the solution to obtain a 0.3 mm coating. The electrodes were left overnight in air and condi-tioned prior to use. The functioning of these electrodes requires that either the polymer-metal interface maintain a constant potential difference, or that this potential difference responds in a definite and reproducible manner to that of the solution-polymer interface. The potential of a platinum wire in an electrolyte solution which does not contain a redox couple will drift significantly in a relative short time period. Frequent restandarization is needed. Cattrall and Pui [79] described coated wire electrodes of the chloromercurate(II) and iodomer-curate(II) salts of the commercial quaternary ammonium compound Aliquat 336S, combined with P V C . A typical response is shown in the figure 2.8. The electrode showed severe drifting of the potential. Also, as the acidity was increased, the solution became yellow because of the formation of iodine, which no doubt, was responsible for the bad electrode drift. 2.6 Coated Wire Glucose Sensor The coated wire glucose sensor of Wilkins and Wilkins is claimed to be a potentiometric glucose biosensor [80-87]. Platinum wire is coated with a partially soluble barium-glucose chloride complex and is covered with a polyvinyl chloride matrix to give a maximum voltage and a minimum time to reach steady state. This typical ion selective electrode operates with the anion association complexes of a large quaternary ammonium ion such as Aliquat 336S. The electrode is reported to give a linearly decreasing potential with increasing glucose concentration without utilizing an oxidation reaction over the glucose range (40-200)mgd/_1(2.2-11.1 mmol/1). This type of sensor was first developed by Dr. Ebtisam Wilkins and Mr. Odeyle at the University of Chapter 2. LITERATURE REVIEW 30 MO, Figure 2. Response of the ctJorornercurate(II) electrode (A) 10 M free CT. (•) 3 M free CP. (O) 1 M total CT, (X) 0.3 M total CT Figure 2.8: The Results of a Mercury Electrode [79] Chapter 2. LITERATURE REVIEW 31 New Mexico [84]. Studies were made by a group of graduate students to examine the response of such glucose sensors and preliminary data were obtained on the following subjects: 1. The development of an implantable coated wire glucose sensor for an Artificial beta cell [84]. 2. Optimization study of the chemical component of the glucose sensor [85]. 3. Mechanism of the glucose sensor [86]. 4. Interferent tests of the glucose sensor with some body chemicals in physiological range [87]. 5. Testing of different membranes for the reduction of interference [87]. 6. Fabrication process [87]. The mechanism was said to involve the association of the Ba-glucose complex as more glucose enters the sensing membrane giving less free Ba2+ which thus produces a decrease in the potential. The work was started in 1978 in parallel with work on a polaragraphic enzyme electrode. The aim was to develop a micro scale electrode suitable for clinical and surgical use. Coated wire electrodes were initially prepared by dissolving a pentaacetate salt of glucose in sodium hydroxide [84]. A variety of compositions were used. The suspensions were formed using different concentrations of glucose pentaacetate/NaOH stirred with an Aliquat/decanol mixture. The tips of thin platinum wires (0.257 mm diameter) were dipped several times in a mixture of the aliquat and P V C and were air dried at room temperature. Another type of sensor was prepared using a barium salt of glucose synthesized by the addition of 100 ml of I M glucose to 50 ml of I M BaCl2- The crystal were dissolved in 20 ml 60 v/v% Aliquat:decanol and a mixture of P V C in cyclohexane was then added in a 1:3 volume ratio [84]. E l Degheidy [83] prepared the salt by adding 100 m l of 2M glucose solution to 100 ml of I M solution of barium chloride. He mentioned that the salt was believed to be an insoluble binary compound of glucose and barium. The sensor was called a novel, coated wire, liquid Chapter 2. LITERATURE REVIEW 32 membrane electrode, apparently the first successfully constructed for non-ionic species. He performed a continuous study on its response and reported reproducible results. Figure 2.9 shows the continuous glucose sensor potentiometric responses, with response times on the order of 2 to 4 min. Sensor performance was investigated with respect to parameters such as Aliquat to decanol ratio, amount of glucose complex, salt:polymer ratio, pH of solution and bias of applied potential to predict the membrane composition that provided optimum response. Al l sensors were made under the same ambient conditions. The stability of the sensor was assessed by conducting ageing tests. Experimental data obtained indicated that this was a very promising electrode for implanta-tion. Exactly what the mechanism was which causes the sensor response to glucose was unclear and further investigation was needed. In the earlier reports, it was suggested that the active sensor was a barium and glucose salt or complex [86]. The complex was said to undergo a reversible reaction between ionic and non-ionic states. A n N M R analysis of the glucose salt gave two peaks at the l a and 1/3 carbon positions that did not correspond to the usual values for glucose. It was found that the barium had changed the energy level of the carbon at that position by associating with the oxygens in an ionic configuration. Therefore, barium moves with glucose somehow bound to it. Ba * Glucose ^ Ba2+ + Glucose2" (2.15) Figure 2.10 shows a proposed molecular structure of the glucose salt [86]. It was mentioned previously that material leaching from the sensors was suspected as the cause of the sensor's responsiveness to glucose. In fact, an ionic combination of either barium and glucose, or glucose and chloride was believed to be the active ingredients in these devices. Radioactive glucose salt was used to make complexes by the same procedure as the original complexes were prepared in order to quantitatively detect any loss of leachable ions into the solution. Table 2.1 and 2.2 shows the results of barium , chloride and glucose leaching from sensors without a hydrogel membrane, with a positive pore membrane, and with a negative pore membrane respectively. It was found that since the negative pore membrane was much less effective than the positive Figure 2.9: Continuous Monitoring of Varying Glucose Concentration. Chapter 2. LITERATURE R E V I E W 34 SUGGESTED STRUCTURE: Figure 2.10: Proposed Molecular Structure of the Glucose Complex [86] Chapter 2. LITERATURE REVIEW 35 membrane, that indicates the glucose complex definetely contained a positively charged glucose component. Chapter 2. LITERATURE REVIEW 36 Table 2.1: Comparative Results of Barium and Chloride leaching test [86] Sensor Membrane Total % Leached Charge Cl Ba 1 None 11.7 2.55 17 None 22.1 4.20 10 None 23.7 5.60 21 Positive 2.77 0.02 15 Positive 12.9 0.06 19 Positive 10.88 0.04 18 Negative 1.15 0.10 5 Negative 9.10 0.33 8 Negative 6.50 0.20 Table 2.2: Comparative Results of Glucose Leaching [86] Sensor Membrane total % Leached Charge Glucose 4 None 7.70 10 None 4.42 8 None 4.96 2 Positive 2.30 3 Positive 1.98 20 Positive 2.64 17 Negative 3.75 16 Negative 3.28 13 Negative 4.12 Chapter 2. LITERATURE REVIEW 37 The mechanism was hypothesized, based on the glucose salt, that if glucose had become ionically bound or complexed with either barium or chloride, upon its dissolution it would exist in an ionic state. If excess glucose were added to the solution, the reaction would be expected to shift to the left, thereby causing the amount of free "ionic glucose" to decrease, and ultimately causing the potential created by the sensor to fall [86]. Finally, these sensors were covered with different combinations of charged pore hydrogel coatings before their selectivity and sensitivity to an addition of blood components such as ascorbic acid, uric acid, 1-cystine, glycine, billirubin, etc were deterrnined. The responses of such sensors were influenced by various factors such as temperature, environment, shape, different concentrations of salt and the experimental error in the fabrication process. Since the behavior of the electrode was still unclear, it was suggested that either positive or negative charges be applied to the pores of a hydrogel membrane which was coated as an outside membrane around the coated wire glucose sensor. From ref [87], it was found that membranes do have an important role in the reduction of interference and the generation of better response times. Using only one specific membrane around the glucose sensor seemed to reduced the amount of interference by up to 20%. Coated wire electrodes always responded but what was difficult to explain was the irreproducibility of the steady state potential for different coated wires in different concentrations of glucose. The term "Steady state" is used to indicate when the potential remained within i l m v for 1 minute within the testing solution. It was suggested that other techniques for fabricating these coated wire electrodes could improve the reproducibility of results. Changes in the methods of fabrication of these electrodes did seem to improve the reproducibility of the measured signal. Future studies were recommended to find the exact mechanism and behavior of the sensor: 1. Due to toxicity of barium, it would be reasonable to exchange the barium with another component such as magnesium, or lanthanides which could be less toxic. 2. The glucose sensor should be tested with real plasma or red blood cells. Chapter 2. LITERATURE REVIEW 38 3. Coating techniques must be improved by using different membranes. 4. Isolate the sensor mechanism and identify exactly the molecular structure of the glucose-metal complex. The work proposed for this research was to improve this type of electrode and to prepare the electrode for implantation in subcutaneous tissue. In the next chapter the objective of the present study and the preliminary investigation that was done will be described. Work has continued at the university of New Mexico where Dr. Wilkins and co-workers worked in parallel with others at John Hopkins University, Applied Physics Laboratory (Apl) [81]. The purpose was to develop a needle type "coated wire electrode" which would be used in conjunction with a lithium battery-powered telemetry system built by A P L which would transmit the glucose information to a hand-held receiving unit. 2.7 Preparation and Effect of Complexes Since in the literature it was stated that the mechanism of the coated wire glucose sensor was based on association and dissociation of the Ba-glucose complex, it was necessary to study the formation of the metal-sugar complex in greater detail. It was also desirable to determine if any other metal beside barium can make a metal- sugar complex for use in the preparation of the needle type sensors since Ba is known to be toxic. The preparation of the glucose complexes was studied using three different methods for finding: 1. A stronger metal-sugar formation by different fabrication techniques. 2. Replacement of Ba with another metal. In 1950 Lebedev and Liubin [88] showed that the binary compound of glucose and sodium chloride of the composition (C&H^O^N aCl + H2O, which forms well-defined crystals, de-composes rapidly when mixed with a small amount of H2O, forming large crystals of glucose and a mother liquor containing part of the glucose and almost all the sodium chloride. Chapter 2. LITERATURE REVIEW 39 Sharkov and Guzhavina [89] prepared binary compounds by slow dehydration, over sulfuric acid at 15°C , of concentrated aqueous solutions containing 1 mole of mineral salt, per 2 moles of glucose. The crystals formed were separated from the mother liquor on a porous plate, dried throughly first by means of filter paper and then in a drying oven at 60°C to constant weight and then analyzed quantitatively. The following salts formed binary compounds with glucose. NaCl, NaBr, Nal, KCl, KI, KBr, LiCl, LiBr, Lil, SrCl, SrCl2, CdCl2, CdBr2, MgCl2, BaCl2, ZnSOi, Na2SOA, NaNOz, KN03, Ca{N03)2, Mg(NOz)2. The ability of these binary compounds to decompose in the presence of H20, with the formation of glucose crystals, was tested and showed a yield of crystalline precipitate containing only 0.4% of the original glucose crystals after such treatment. Sugars occur in the solution which also contains salts. It is relevent to query whether any association exists between these two different types of compound. When this subject was reviewed in 1966, 172 references were cited. Neverthless, none of the important questions had been answered. Which sugars form complexes with cations? What are the structure of complexes? What are the conngurational or conformational features required for complex formation? What are the stability constants of these complexes? [90] N M R spectra give considerable information on the structure of sugar -cation complexes. Complex formation causes a downfield shift of some proton signals. The extend of the shift depends on the geometrical relation of the hydrogen-carbon bond to the cation. Crystalline addition compounds of sugars and various salts in stoichiometric proportions have been known for a long time. Their existence, however, does not constitute evidence that complex formation occurs to any significant extent in solution. It only indicates that the sugars, cations, and anions can fill the space in a regular packing usually held together by hydrogen bonds and by co-ordination of cations with oxygen atoms. When dissolved in water, the hydrogen bonds will be broken and water will usually displace the hydroxy-groups from their co-ordination with the cation. Whether there will be any complex formation in solution cannot be predicted from the crystal structure. It was also observed that, in the stable chair Chapter 2. LITERATURE REVIEW 40 form of sugars-in aqueous solutions, there are three oxygen atoms which are close to each other, forming an almost equilatral triangle (as do three synraxial hydroxy-groups), and it was postulated that these oxygen atoms, in an axial-equatorial-axial (ax-eq-ax) are the site of the formation of the complex. Figure 2.11 illustrates the above observation. Work continued with Franks and co-workers [91] to find if particular cations could effect the point of balance of the anomeric equihbrium. The technique used to monitor the changes was Raman spectroscopy. This technique has been applied to aqueous solutions of D-glucose by Barrett [92], Mathlouthi and co-workers [93,94], Sivchik and Zhbankov [95], She et al. [96] Vasko et al [97,98] and Spedding and Stamm [99]. Figure 2.12 shows the site formation of a complex on the addition of calcium ion compared to the original compound. The Anomeric regions of the Raman spectra of solutions of equilibrated D-glucose containing eight salts are present in figure 2.13. The differences between a and f3 anomers are described in figure 2.14. As indicated before, extensive studies were done by Angyal and co-workers [91] on complex-ing between sugars and cations. For strong complexing to occur, Angyal had proposed that an axial- equatorial-axial arrangement of hydroxyl groups on three consecutive carbon atoms must exist. They stated that a number of complexes of inorganic salts with such sugars have been isolated in solid state (the CaCh complex of D-mannose). D-glucose does not possess the required arrangment of hydroxyl groups, but in this study they suggest that Ca2+ does interact with D-glucose, causing a shift in the anomeric equUibrium, probably also through the mechanism of complex formation. The complexes were prepared according to the procedure given below: Aqueous solutions of D-glucose were prepared by dissolving a D-glucose in distilled water. Dissolution was aided by gentle heating for a very short time in a water bath at 80°C. Solutions required about 2 hours to reach anomeric equilibrium. Salts were added to those solutions and Raman spectra were recorded. Angyal observed no substantial change in the N M R spectra of dilute solutions of D-glucose on addition of CaC/2. He noted that these sugars do not possess the a-e-a sequence of three Chapter 2. LITERATURE REVIEW 41 0) Figure 2.11: Site Formation of a Complex on Addition of Calcium Ions [90] Chapter 2. LITERATURE REVIEW 42 Figure 2.12: Raman Spectra of Equilibrated 2.GM D-Glucose in the Range of (700 - 300)cm - 1 . A : Freshly Prepared aD Glucose, B: A n Equilibrated Solution of a and /9D Glucose, C: Freshly Prepared /3D Glucose [103] Chapter 2. LITERATURE REVIEW 43 IOOO 900 eoo cm"' Figure 2.13: Raman Spectra of Equilibrated 2.0M D Glucose Solution in the Range of (1000 - 800)cm - 1 Containing the Salt Indicated. Salt Concentrations Are 4.0M, With the Exception of LaCl3, Which is 3M.[104] Chapter 2. LITERATURE REVIEW 44 oxygen atoms favorable for complex formation [100]. Later, in studies of paper electrophoresis of polyols in a solution of calcium ions, he noted that practically all polyols and sugars show some mobility, even when other methods do not detect complex-formation [101]. Goulding [102] reported the separation of a and /3 anomers of D-glucose on a column of a Ca2+ or Sr2+ cation-exchange resin; /3D glucose is eluted before Q D glucose. He noted that the former has no a-e hydroxyl group neighbors, whereas the latter has one such pair and can form weak complexes. They concluded that Raman spectroscopy is sensitive to the presence of these weak complexes. Finch and co-workers [103] found that in aqueous solutions, the metal ions form the-octa-hedral aqua ion, or aqua complexCu(^ 2 0)^ + . The process of complex formation in aqueous solution is essentially a displacement of one set of ligands, water molecules of the aqua complex, by another, a diol: [Cu(H20)6}2+ + R(OH)2^[Cu(H20)4R(OH2}2+ + 2H20 (2.16) It was concluded again that polyols form complexes in aqueous solutions with tri-valent cations, of an ionic radius larger than 0.7° A. The interaction of D-glucose with a hydrated alkaline earth metal halide has been studied in solution and characterized by means of IR and N M R spectroscopy, x-ray powder diffraction and molar conductivity measurements by Heider Ah-Tajmir-Riahi [104,106]. The preparation of the salt included calcium halide(l mmol)in H20 (10ml) which was added to a hot solution of D-glucose(lmmol) in H20 (20ml) and the mixture was heated for 30 minutes at 8 0 ° C . Solutions were kept for 48 hrs at room temperature, and 7:3 acetone- ethanol solution was added to precipitate the complex. Figure 2.15 shows the spectra of D-glucose and its alkaline earth metal ions in the region of (1800-500) c m _ i . This was filtered off, washed several times with acetone and dried( C a C ^ ) - Ft.ir spectra were compared. He studied the spectra of the alkaline earth metal-D glucose adducts and free a-(3 D-glucose and their equilibrated solution and noted the effects.of binding of these metal ions on the sugar anomeric configurations. Similar observations were made earlier by Angyal and Davis for the D-glucose molecule in the presence of the Ca(II) ion in aqueous solution. In non-aqueous solution, a major down field shift upon Chapter 2. LITERATURE REVIEW Figure 2.14: F.t.ir Spectra of D Glucose and Its Alkaline-Earth Metal Ions in the Region (1800 - 500)cm_1.[104] Chapter 2. LITERATURE REVIEW 46 interaction of Mg(II), Ca(II) with D-glucose is observed which indicated a strong interaction between these metal ions and D-glucose. Similar observations showed complexes with soluble sugar acids (glucuronic acid) which appear to form stronger complexes with alkaline metals [105,106]. Rendleman [107] noted that published information on the alkali metals and alkaline earth metals with carbohydrates has often be vague and difficult to interperate, largely because ions of these metals do not readily form stable complexes in aqueous solution. Although the detection of complex-formation in solution is relatively easy, the determination of stability constants has not yet been accomplished. It has been noted that, in aqueous solution, the order of decreasing effectiveness in promoting electrophoretic migration of carbohydrates is (Ca2+, S r 2 + , B a 2 + ) > Mg2+ > Na+ > K+ > Li+ > NHA [107]. Chapter 3 P R E L I M I N A R Y E X P E R I M E N T A L A P P R O A C H 3.1 T h e Overall Objective of the Thesis Based on the literature available about the coated wire sensor, this research was designed to produce and test in vivo a miniaturized coated wire electrode. The objectives at that time were: 1. Better understanding of the mechanism proposed for the electrode operation and identi-fying the exact molecular structure of the glucose salt. 2. Developing a new complex since Ba is a toxic element. 3. Improvement of the fabrication process towards rniniaturization. • sensor shape. • sensor size. 4. Implantation in animals. • Tested in tissue (eq. subcutaneous tissue). 5. Biofunctionality of the glucose sensor to be tested for a long period of time. • Test membranes that were biocompatible. 6. Test the mathematical model which was proposed earlier for the glucose diffusion in and out of the membrane. As research progressed it was found that the transient behavior could not be due to the original proposed mechanism. The objective and the approach to this problem changed to: 47 Chapter 3. PRELIMINARY EXPERIMENTAL APPROACH 48 1. Reproducing the original data that was given in the literature. 2. Explaining the transient behavior of the glucose sensor. 3. To develop a mechanism to explain the observed behavior. The next section will explain the different experimental steps that were followed in order to arrive at the final conclusions for the preliminary experimental observations. 3.2 Preparation of the Complexes As indicated in section 3.1, one of the primary goals was to develop a new complex which could replace Ba since it is a toxic element. It was decided to prepare the complex by different techniques in order to find which technique can- produce stronger metal-glucose salts. IR and N M R spectra of the prepared salts were used to observe any metal-sugar interaction. Finally the prepared salts were used in the preparation of electrodes and their potentiometric behaviors were compared. Four methods were used to prepare the glucose-salt complexes used in the electrodes. 1. (3 D-glucose was mixed with an alkali or alkaline earth metal solution in the molar ratio of 1:1, 1:2, and 1:3. The complex was crystallized by slow dehydration over sulfuric acid at 15°C. Some of the compounds tested were BaCl2, calcium acetate, potassium nitrate, lanthanum acetate, cupric acetate, and ferric sulfate. Every complex has been given a code which has been used throughout this thesis. For instance: • Complex No.l contains BaCl2 • Complex No.2 contains calcium acetate. • Complex No.3 contains lanthanum acetate. • Complex No.4 contains cupric acetate. • Complex No.5 contains FeS04. Chapter 3. PRELWHNARY EXPERIMENTAL APPROACH 49 • Complex No.6 contains K N O 3 . 2. f3 D-glucose was mixed with an alkali or alkaline earth metal solution in the molar ratio of 1:1, 1:2, and 1:3. It was freeze dried for 30 hours. 3. The compound which was selected was added to a hot solution of D-glucose and the mixture was heated for a few hours at 60°C to form a more concentrated solution. The solution was kept for 48 hours at room temperature then a 7:3 v/v ratio of acetone-ethanol was added to recrystalhze the complex. Complexes of barium were prepared using this procedure which was described by Riahi [105]. 4. Copper acetate or copper chloride was added to a hot solution of D-glucose in a 1:1 molar ratio. The mixture was heated for a few minutes at 50-60°C. Since copper acetate oxidizes very quickly, the solution was then cooled quickly. It was kept for 24 hours at room temperature then a 7:3 methanol-ethanol mixture was added to recrystallize the complex. 3.3 Complex Analysis 3.3.1 Infrared Spectroscopy An analysis of the glucose salts in the liquid phase was done using an infrared interferometer equipped with a KBr beam-splitter. KBr pellets were prepared by mixing the prepared salts with KBr in a pehetizer. The infrared spectra of D-glucose and its alkaline earth metal halide adducts were recorded in the region of 500-4000 c m - 1 , and the results of the spectral analysis were recorded. The peaks shown in the observed spectra give an indication of the interaction of the metal with the sugar compounds. Even though this type of analysis is only qualitative, the spectra do give an indication of the metal-sugar binding. The spectra are described in the results section. Chapter 3. PRELIMINARY EXPERIMENTAL APPROACH 50 3.3.2 Nuclear Magnetic Resonance Spectroscopy The' N M R spectra were recorded on a Varian XL-300 multinuclear spectrometer operating on internal lock with a normal probe temperature of 31°C. Chemical shifts were measured from D20 as internal standard. The C 1 3 N M R were observed at 75 M H Z frequency. Changes in the chemical shift due to complex formation were measured. N M R were observed for selected com-plexes (No.l-No.6), and the observed spectra were compared to see if there was any indication of energy shift in the spectra. The chemical shifts of these signals varies slightly on addition of salts to the solution. 3.4 IR Results of Selected Complexes The IR spectra of these complexes at specific regions will reveal any indication of interaction between glucose and these metals. A mixture of calcium halide in water added to a hot solution of D-glucose and precipitated will form a complex. In our study we confirmed the formation of such a complex as shown in figure 3.1. In addition to the calcium complex we were able to prepare other complexes such as barium and copper metal with D-glucose in a similar manner to that of the calcium complex. Figure 3.2 shows the spectra of a Ba-glucose complex prepared by freeze drying and by recrystalization. These IR spectra are compared with the barium chloride spectrum in the regions of (450-1800 c m - 1 ) . The infrared spectrum of the recrystalized complex shows that there is sharp absorption bands at 920, 845, 772 c m - 1 , and a broad band at 654 c m - 1 . As can be seen, the freeze dried complex shows a weaker complex formation with the metal since the absorption bands can not be well identified. By looking at figure 3.3 which shows the whole spectra in the regions of(300-4000), a better explanation can be seen. The freeze dried complex shows a few strong absorption bands in the regions of 1019, 1599, 2926 c m - 1 and the recrystalized complex shows strong absorption bands in the region of 1013, 1613, 2900 c m - 1 . These absorption bands can not be seen in the spectrum of the original BaCl2. It can be concluded that there is a weak interaction between the barium metal and D-glucose. Similar comparisons were made for copper complexes. Figure 3.4 shows the spectrum of a D-glucose Chapter 3. PRELIMINARY EXPERIMENTAL APPROACH 51 Figure 3.1: IR Spectra of D-Glucose and Ca Complexes in the Regions of (500-1800 c m - 1 ) , a) Ca(D-Glucose)Cl 2 .4H 2 0, b) ^D-Glucose, c)aD-Glucose. Chapter 3. PRELIMINARY EXPERIMENTAL APPROACH 52 Figure 3.2: IR Spectra of Baxium Complexes in the Regions of (500-1800 c m - 1 ) , d) Ba-Glucose(Freeze Dried), e) B a C l 2 Compound, f) Ba-Glucose(Recrystallize). Chapter 3. PRELIMINARY EXPERIMENTAL APPROACH 53 tx»~gloco« V a n 3 *. < > 3 < 3 e IS «XM 36O0 3ZOO 2BOO 2<00 2000 16O0 12O0 BOO U^venurabers Figure 3.3: ER. Spectra of Barium Complexes in the Regions of(300-4000 cm *), g)Ba-Glucose(Freeze Dried), h) BaCl2 Compound, i) Ba-Glucose(RecrystalHze). Chapter 3. PRELIMINARY EXPERIMENTAL APPROACH 54 O 01 JO < Uavpnumbers Figure 3.4: IR Spectra of D-Glucose and Copper Complexes in the Regions of (500-1800 cm ) j) Cu(Ac)2, k) Glucose Compound, 1) Cu-Glucose(RecrystaUze), m) Cu-Glucose (Slow Dehy-dration) Chapter 3. PRELIMINARY EXPERIMENTAL APPROACH 55 TTZ—Tw, TWi r£Z TT.. •:•> Uavenumbers Figure 3.5: IR Spectra of D-Glucose and Copper Complexes in the Regions of (300-4000 c m - 1 ) , n) Cu-Glucose (dried freeze), o) Cu-Glucose (Recrystalize), p) a D-Glucose, q) /3 D-Glucose, r) Cu- Glucose (Slow Dehydration) Chapter 3. PRELIMINARY EXPERIMENTAL APPROACH 56 in comparison with copper acetate and copper complexes which were formed using techniques that were explained earlier. The IR spectrum shows strong absorption bands in the region of 802, 858, 923 c m - 1 for recrystalization of the copper complex and 767, 845,and 914 c m - 1 for the complex of copper made by slow evaporation. The a D-glucose has its stronger absorption in the regions of 776, 838, 915 respectively. In cupric acetate spectra, there is no indication of any bands in that region. A better explanation can be obtained from figure 3.5, where the whole spectra of these compounds are given. For the freeze dried copper complex, between the regions of 720-1130 c m - 1 , a strong absorption can be seen. The spectra appear similar between the 1900-4000 c m - 1 . It can also be observed that a D-glucose showed interactions with the copper metal for all the prepared complexes. Barium and copper do interact with D-glucose but the amount of complexation is small. However, complexes with soluble sugar acids(glucuronic acid) would appear to form stronger complexes. The complex formation can be important if the glucuronic acid would go through a competitive complexation with D-glucose. 3.4.1 Summary and Conclusions Sugars can form complexes in solutions which contain alkaline earth metals under various conditions as indicated from this work. Evidence has been accumulated to indicate that at least some sugars in particular in our study (D-glucose) form complexes with several cations. By using IR analyses of all the complexes, it was found that there was interaction between D-glucose and copper and barium metals. IR results showed strong absorption bands at 920, 845, and 772 for complex of barium formed by recrystalization in the region of (450-1800) and 1013, 1613, and 2900 c m - 1 in the regions of (1000-4000)cm_1. For the barium freeze dried complex, strong absorption bands were found in the regions of 1019, 1599, 2926 c m - 1 . IR results for copper complexes prepared by recrystalization showed strong absorption bands in the regions of 802, 858, 923 c m - 1 , 767, 845, 914 for copper complexes which were prepared by slow evaporation and 720-1130 c m - 1 for copper complexes which were prepared by freeze Chapter 3. PRELIMINARY EXPERIMENTAL APPROACH 57 drying. On the basis of these results , it can be concluded that: 1. There is an interaction between barium or copper and D-glucose. The amount of complex formation is very limited. 2. The binding of D-glucose to these alkaline earth metal ions appear to be via the a and /3 anomeric configuration. Three different preparation methods were analyzed (1) by slow evaporation, (2) By recrys-talization and (3) by freeze drying. Of these for copper, the freeze dried complex shows the lowest interaction with D-glucose and slow evaporation showed the highest. For barium, recrys-talization showed the highest interaction and the freeze drying showed the lowest interaction with D-glucose. 3.5 N M R Analysis The N M R analysis of different binary compounds with glucose were also observed. By looking at all these spectra, it is possible that few of these elements have changed the energy level of the carbon at the l a and 1/3 positions of the glucose molecule. Appendix C includes the N M R spectra. Table 3.1 includes the spectra at the positions of the l a and 1(3 carbon of the glucose complexes compared to the glucose molecule. By comparing the energy level at the l a and 1/3 carbon for pure glucose to metal-glucose salts, one can see that the metal- glucose interaction has possibly caused a small change in energy level of the carbon at l a and 1/3 position with the glucose molecule. Chapter 3. PRELIMINARY EXPERIMENTAL APPROACH 58 Table 3.1: N M R Analysis of Glucose and Glucose Salts Sample l/3(ppm) la(ppm) Glucose 93.405 97.220 Ca-glucose 93.394 97.213 K-glucose 93.492 97.279 Cu-Acetate-glucose 93.517 97.318 B a-glucose 93.484 97.258 Cu-sulfate-glucose 93.415 97.221 La-glucose 93.394 97.218 Fe-glucose 92.390 96.177 3.6 Experimental Apparatus and Materials Used for Constructing Electrodes 3.6.1 Lists of Material Used Table 3.2 lists the materials used , their shortened name and the supplier. Chapter 3. PRELIMINARY EXPERIMENTAL APPROACH 59 Table 3.2: List of Materials Used Substance M.W. g/gmole Manufacturer Sodium Dihydrogen Orthophosphate 156.01 Analar Disodium-Hydrogen Orthophosphate Anhydrous 141.96 Analar Embedding Capsules Beem, Size 3 Polysciences Methyl Tricapryl Ammonium Chloride (Aliquat 336S) 404.17 Aldrich Chemicals n-Decyl Alcohol 158.29 T.Baker Chemical Co. D glucose Anhydrous 180.16 B D H Beta-D glucose 180.20 Sigma Chemicals Lanthanum Acetate 343.07 A E S A R Calcium Chloride 110.99 Fischer Scientific Cupric Nitrate 241.60 Allied Chemicals Potassium Nitrate 101.10 B D H Chemicals Lanthanum Nitrate Pentaacetate 415.01 Aldrich Chemicals Ferrous Sulfate 278.02 American Sci.Chemicals Magnisum Nitrate 256.41 B D H Chemicals Ammonium Persulfate(APS) 228.20 Fischer Scientific Nitric Acid 63.01 Fischer Scientific Poly Vinyl Chloride(PVC) 140*31 BF-Goodrich Lanthanum Chloride 371.37 Aldrich Cupric Acetate 199.67 Aldrich Calcium Acetate 158.17 Aldrich Barium Chloride 208.25 Aldrich Tetrahydro Furan(THF) 72.11 B D H Chemicals 2-Hydroxyl Ethyl Methacrylate(HEMA) 130.14 Polyscience Tetra Ethylene Glycol Dimethyl Crylate(TEGDMA) Polyscience N,N-Dimethylamino Ethyl Methacrylate(DMAEM) Polyscience Methacrylic Acid(MA) 86.09 Polyscience Sodium Bisulflte(SB) Ethylene glycol 62.07 MaJlinckrodt(AR) Aluminum Oxide(99.99%) 101.96 U.B .C Chemistry dept. Acetone 58.08 Analar Methanol 32.04 Analar Chloroform-d MSD Isotopes Deuterium oxide MSD Isotopes cellulose Acetate Copper Selenide 142.50 Alfa Distilled Water 18.02 U.B .C . Chemical Engr. Saturated Calomel Electrode Platinum Wires GoodFellow Chapter 3. PRELIMINARY EXPERIMENTAL APPROACH 60 Table 3.3: Hydrogel Formations component Positive Pore (ml) Negative Pore (ml) H E M A 5.00 5.00 D M A E M 0.50 2.00 M A 0.00 0.50 T E G D M A 0.20 0.20 E T H Y L E N E 0.50 0.50 G L Y C O L W A T E R 4.50 4.50 APS 1.00 1.00 SB 1.00 1.00 3.6.2 Hydrogel Preparation Since there was an indication that the use of a hydrogel membrane can control the leaching of the material from the sensor according to the charges it carries, the hydrogels were formed around the coated wire glucose sensors after dipping the electrode into the coating solution. The original membrane was prepared according to reference [108,109]. The membrane was prepared as glucose sensitive membrane with a difference that in the original membrane glucose oxidase was used as a catalyst,instead for this, sodium bisulfite (SB) and ammonium persulfate (APS) were used as an inert catalyst and initiator. Preliminary results were for electrodes with a hydrogel coating in order to reproduce literature results. It is possible in the preparation of the hydrogel coating to produce either positive or negative pore charges. The proportions of the reagents used to create the different membranes may be found in Table 3.3. First, five components are mixed and then the initiator and the catalyst, APS & SB are added and stirred into the solution individually. The sensors were dipped into the solution to create a hydrogel coating. Chapter 3. PRELIMINARY EXPERIMENTAL APPROACH 61 3.6.3 Polyacrylamide Gel Preparation Polyacrylamide gel was prepared by mixing the monomer with the cross linking material [110]. When well mixed, the catalyst and finally the water were added to the mixture to form a thermally stable, water insoluble gel. The thickness of the gel was varied to see its effect on the electrode response. Table 3.4 shows the components and the amount needed for polyacrylamide gel formation. Chapter 3. PRELIMINARY EXPERIMENTAL APPROACH 62 Table 3.4: Polyacrylamite Formations component amounts Monomer 8.39 g N ' N D M A 0.066g D M A P N 0.0329ml W A T E R 23.7ml APS 0.31615g 3.6.4 Conditioning and Storage of the Sensors All electrodes were air dried in a dead air hood. The sensors were initially conditioned either in the 0.001M glucose buffer solution or phosphate buffer solution alone. 3.6.5 Preparation of Phosphate Buffer The buffer was made by dissolving 26.22g of NaH2P04H20 and 115.02g of Na2HP04 in 2 liters of distilled water. The pH of the solution was adjusted to pH 7.4. The total solution was brought up to a volume of 10L. 3.7 Preparation of the Electrodes Tests were made with a number of electrode types. The effect of all different types of elec-trodes were studied in order to understand the role of the Ba-glucose complex in these types of electrodes. 1. Coated wire sensors • Fabrication of identical coated wire sensors 2. Membrane electrodes 3. Glass membrane electrodes 4. Needle Type Sensor Chapter 3. PRELIMINARY EXPERIMENTAL APPROACH 63 3.7.1 Identification of the Compounds Used for Preparation of the Electrodes 1. n-Decyl Alcohol: C H3(CH2)8C H20H 2. Aliquat: A type of organic nitrogen compound in which the molecular structure includes a central nitrogen atom joined to four organic groups as well as to an acid radical . It is a cationic surface active compound and tends to be adsorbed on surfaces: its equation is CH3N{{CH2)7CH3)3Cl 3. Glucose complex: Ba-glucose complex, Ca-glucose complex, La-glucose complex and etc. 4. P V C : Polymer used {H2CCHCl)n 5. T H F : SolventC 4 ^8C 3.7.2 Coated Wire Electrode The coated wire sensors were prepared according to El-Deheidy's procedure [83]: The Ba-glucose complex was ground to a fine powder and seived to yield a product with 180 pm average diameter. 0.3 grams of the complex were weighed in a glass vial. 1.2ml of decanol and 0.8ml of the Aliquat were added to the salt, and mixed by a magnetic stirrer over night. 3 grams of polyvinyl chloride (PVC) were dissolved in 12ml of tetrahydro furan, then an equal volume of this was added to the salt mixture and stirred for approximately 4 hours. Platinum wires were dipped several times individually in the mixture until well coated, being careful to completely cover the metal tip. These were allowed to dry over night in a dead-air hood. The shape of these coated wire electrodes are shown in figure 3.6. 3.7.3 Fabrication of Identical Coated Wire Sensors The glucose sensors were again prepared according to El-Deheidy but in a mold. 1. The glucose salt was ground to powder approximately 180 pm. Chapter 3. PRELIMINARY EXPERIMENTAL APPROACH 64 2. 0.025 g of the salt were weighted and put into embedding capsule molds 6.35mm long X 9.53mm width. 3. 0.05 ml of decanol and 0.025 ml of Aliquat were added to the salt, which was mixed by a micro-magnetic stirrer. 4. 3g of P V C were dissolved in 12 ml of T H F . 0.08 ml of the above mixture was added into the embedding capsule beems until the mixture was well mixed. 5. A platinum wire 3.5 cm long(0.2 mm thick) was placed into the mixture and allowed to dry over night in a dead air hood. 6. The following day, the sensor was removed from the embedding capsule beem, shaped and dried with the use of tweezers. 7. Each sensor had approximately the same shape. 8. Approximately 14 sensors were made each time and tested. Figure 3.7 show the construction of these types of electrodes. Chapter 3. PRELIMINARY EXPERIMENTAL APPROACH Figure 3.6: The Shape of Coated Wire Glucose Sensors Chapter 3. PRELIMINARY EXPERIMENTAL APPROACH o f t e f l o n Figure" 3.7: Fabrication of Identical Coated Wire Glucose Sensors Chapter 3. PRELIMINARY EXPERIMENTAL APPROACH 67 3.8 Electrode Testing Equipment 3.8.1 Meters Used Three meters were used throughout the experimental observations. Because of the unavailabihty of a high impedance meter in Chemical Engineering Department, they had to be borrowed or rented. • meter no.l Keithly 616 electrometer.(input impedance=1012 ohms). • meter no.2 Keithly 617 electrometer.(input impedance=1014 ohms). • meter no.3 Keithly 640 vibrating capacitor electrometer in conjunction with a 640 input head with variable resistance 106 to 1012ohms. • Regulated Power Supply.(range (0-10v) ) 3.8.2 Testing Cell The dimension of the polyethylene testing cell is shown in figure 3.8. Water was circulated from a constant temperature bath through the jacket to maintain a constant electrode temperature. This cell was covered with a polyethylene piece which contained 6 electrical contacts with connectors to the meter. Sensors were connected to the meter for potentiometric measurements. 3.8.3 Potentiometric M e t h o d Using Needle Electrodes A schematic of the testing apparatus is shown in figure 3.9 During the test, the buffer entered the polyethylene cell at a constant rate until the volume of the fluid reached 200ml. The flow of the fluid was controlled from plastic feeder bags. The temperature of the cell was kept constant at 37°C. A miniaturized reference electrode was used in conjunction with the needle type glucose sensor. The potential created between the sensor and the reference electrode was measured using a potentiometer. The data were transferred to a chart recorder. Polymeric Chapter 3. PRELIMINARY EXPERIMENTAL APPROACH 4.50 " OD x 0.125 " O-Ring 0.25" Wall 4.25' 0.25 " NPT 5.50" OD x 0.25" 6 x 10-32 Tapped Holes on 5 "PCD 3.00' .3.00' 0.75" '0.375" NPT • 4.50" \ Figure 3.8: Schematic Drawing of the Electrochemical Cell Chapter 3. PRELIMINARY EXPERIMENTAL APPROACH 69 biomaterial membranes such as cellulose acetate, hydrogel, cellophane were selected and used in the electrodes to test for their responses. 3.8.4 Potentiometric Method Using Coated Wire Electrodes Potentiometric responses were also observed by immersing prepared electrodes in buffer solu-tions in conjunction with a saturated KCl reference electrode in an electrochemical cell which was either stagnant or stirred. Agitation of the solution was maintained by a magnetic stirrer. When the potential between the two electrodes reached steady state, it was defined as the base Une potential. /3D-glucose was added to bring the glucose concentration up to 100(mg/dl) . When the potential again reached a steady state reading, another incremental glucose addition was made to bring the concentration up to 600(mg/dl). The potentiometric results were ex-pressed as the change in potential with respect to the base line potential. Similar measurements were made with bare platinum pretreated wire. 3.9 Preliminary Experimental Results and Discussions This section includes some of the experimental data that were gathered in the early stages of this study. As shown in figure 3.10, electrodes were prepared with Ba — glucose complex and tested at various concentrations of glucose until an equilibrium potential was observed. As the potential reached a fixed time which was considered to be at a steady state potential, the coated wire glucose sensors were placed in another concentration of glucose. According to this procedure, voltage vs concentration of glucose curves were plotted. These electrodes were covered with hydrogel membrane to prevent leaching of the salt. The change of potential was first considered to be due to the effect of different glucose concentrations reported in the literature. In order to observe the response of the electrodes, 3 hydrophilic membranes have been used around the glucose sensor. In order to observe if the electrode is selective to glucose, different salt such as NaCl was added to the solution and changes in potential were observed. Figure 3.11 shows the response of the electrode to NaCl addition. As it is shown in this figure, Chapter 3. PRELIMINARY EXPERIMENTAL APPROACH 1. Temperature Controlled Polyethylene Cell 5. Deionized Water (500 ml) 2. Reference Electrode 6. Plastic Feed Controller 3. Glucose Sensor 7. Electrometer (Keithly 616, 617) 4. Buffer Solution (500 ml) 8. Chart Recorder Figure 3.9: Schematic Diagram of the Potentiometric Apparatus Chapter 3. PRELIMINARY EXPERIMENTAL APPROACH 71 NaCI did decrease the signal. The addition of the salt to the measuring solution was just to identify the amount of interference that could be caused by the salt since salt exists in tissues and for implantation purposes the electrode should not show a substantial response to the salt. The purpose of the needle type sensor was to miniaturize the electode for use in the monitor-ing devices. The data produced appeared to substantiate the model described in the literature. Work continued on miniaturizing and stabilizing the electrodes. Chapter 3. PRELIMINARY EXPERIMENTAL APPROACH 72 3 0 0 o 30 o GLU CON.=40(mg/ml) • GLU CON.=80(mg/ml) • GLU C0N.=120(mg/ml) o o O 2 4 0 • < . O > 10 TIME (min.) 3 0 0 \ro° H O < O 260 - i ' 1 ' 1 ' r I I i 1 1 1 1 1 r o GLU CON.=40(mg/ml) • GLU C0N.=8Q(mg/ml) « 5 6 TIME (min.) Figure 3.10: I: The Response of a Coated Wire Electrode, Complex N o . l in 40,80,120(mg/dl) Glucose Using Hydrogel Membrane as Outer Layer, II: The Response of a Coated Wire Elec-trode, Complex No.l in 40,80,120(mg/dl) Glucose Using Hydrogel Membrane as Outer Layer (Conditioned) Chapter 3. PRELIMINARY EXPERIMENTAL APPROACH 73 w o < o 60 120 160 CONCENTRATION (mg/ml) 200 80 120 160 CONCENTRATION (mg/ml) I I Figure 3.11: I: The Response of a Coated Wire Electrode, Complex No.6, Using Hydrogel Mem-brane as an Outside Layer, Measured in Different Concentrations of Glucose II: The Response of Coated Wire Electrode (Conditioned), Complex no.l, Using Tissue Membrane as an Outside Layer, Measured in Different Concentrations of Glucose Chapter 4 P R E P A R A T I O N O F T H E E L E C T R O D E S Since the prehminary measurements appeared to agree with the literature responses it was decided to continue the tests on miniaturized electrodes of the type proposed for subcutaneous implants. 4.1 Minia tur ized Electrodes This section includes the comparison of 3 types of electrodes (Glass electrodes,, needle electrodes, membrane electrodes). The preparation of these electrodes is described below. The prepared electrodes were conditioned prior to use in 180 (mg/L) glucose in H20. 4.1.1 Preparation of the Glass Electrode Since the coated wire electrodes did not have reproducible shapes, it was decided to prepare the electrodes in such a way that they were identical. Also different glucose complexes were tested and compared to observe which salt could give stable voltages most quickly. A schematic drawing of the glass electrode is shown in figure 4.1. The carrier used in this study was Ba-glucose complex mixed with the right amount of Aliquat/decanol as has been mentioned in earlier sections. The outer membrane used was initially cellulose acetate, but other membranes were tested later, to see the effect of coatings on the electrode response. Instead of mixing the polymer into the matrix, glass electrodes were prepared using a thin P V C layer as an outside membrane . The potential responses of these electrodes were measured in the same manner. 74 Chapter 4. PREPARATION OF THE ELECTRODES 75 4.1.2 Potentiometric Results Observed F r o m Glass Electrodes As stated earlier, different metal-sugar salts were prepared to see if it is possible to produce stronger glucose salt complexes. The term "stronger" means the strength and stability of the glucose salt according to its degree of dissociation. The electrodes were conditioned prior to use in 18(mg/dl) glucose in deionized H20. As shown in the figures below, glucose solutions up to 600(mg/dl) were added while changes of potential were still occurring. This was done to see if the transient behavior of the electrodes was effected by glucose addition. Figure 4.2 presents the potentiometric behavior of the electrodes prepared with different metal-sugar salts. Complex No.l represents the Ba-glucose complex whereas complex No.2 represents the Ca-glucose complex. Figure 4.3 shows the behavior of complex No.3 the La-glucose complex and No.4 the Cu-glucose complex. Again, glucose solutions were added to the measuring solution and the potentiometric behaviors were observed and compared. Figure 4.4 shows the behavior of complex No.5 a Fe- glucose complex in comparison to complex No.6 a K-glucose complex. It was observed that the addition of glucose solution does influences the electrode signal in some way. The cause of the change was considered to be the metal-sugar effect. About 50 electrodes were prepared with different metal-glucose complexes and their potentiometric behaviors were observed and compared with one another. The changes did not indicate any reproducibility of the electrode potential according to the observed initial and final voltages. It was observed that the potential would decrease more after the addition of glucose solution to the measuring cell. It was not well understood why the potentiometric behavior of different metal-glucose salt caused increases and sometimes decreases. The next section will describe the needle type sensor and its potentiometric behavior after the addition of glucose solution. Chapter 4. PREPARATION OF THE ELECTRODES 76 Platinum wire Carrier rC o r n P l e x & Liquid 1 I ion exchanger J < Glass tube < Membrane Figure 4.1: Schematic Drawing of the Liquid Membrane Glucose Sensor Chapter 4. PREPARATION OF THE ELECTRODES 77 Figure 4.2: Comparisons of Voltage-Time Responses of Conditioned Glass Electrodes, I: Com-plex N o . l , II: Complex No.2, Tested in Deionized Distilled Water, P V C Membrane, Final Concentration of glucose = (600mg/dl) ' Chapter 4. PREPARATION OF THE ELECTRODES 78 4.1.3 Preparation of Needle Type Sensors The glucose salt was ground to a fine powder and sieved to produce particles with a small size range (approximately 1 micron average). Micro'portions of the salt were mixed with Aliquat and decanol for at least one day. This small portion was injected into the glucose sensor which had been constructed for this research. The needle type glucose sensor consists of a fine platinum wire of 0.02 mm diameter. The tip of the platinum wire was melted in an oxygen natural gas flame to form a small ball. The wire was sealed to a micro glass capillary tube of about 1 mm diameter. The protected wire was then sealed with epoxy resin to the tip of a hypodermic needle (16-25 gauge). The electrode itself was insulated in a second soft glass tube for more protection. The sealed glass was then covered with different membranes, that were chosen to be tested during this experiment, with an O-ring. This electrode was examined in conjunction with a miniaturized reference electrode. A Schematic drawing of such a sensor is shown in Figure 4.5. These electrodes were first calibrated in phosphate buffer pH of 7.4. 4.1.4 Potentiometric Results Observed From Needle Type Sensor During the early stages of this research, the goal was to miniaturize the glucose sensor by inserting the Ba-glucose complex and Aliquat/decanol mixture in a needle gauge electrode in such a way that the coating would be supported. As mentioned previously in Experimental Section 3.9, the electrodes were prepared and tested in order to observe the potentiometric behavior of coated wire electrodes in comparison with needle type electrodes. Membranes were used such as P V C , cellulose acetate, hydrogel membrane and some times kidney tissues were used around the needle type electrode. The purpose was to see how the electrodes behave in a biological environment. Even though the tests did not sufficently describe the behavior of the electrodes at that time, the responses were compared. Figure 4.6 shows the.responses of a needle type electrode (16 gauge needle) which was initially tested in different concentrations of glucose using P V C and tissue as an outside membrane. Electrodes were conditioned prior to use as described in section 4.1. The purpose of this work was to prepare an electrode which can be Chapter 4. PREPARATION OF THE ELECTRODES 79 Figure 4.3: Comparisons of Voltage-Time Responses of Conditioned Glass Electrodes, I: Com-plex No.3, II: Complex No.4, Tested in Deionized Distilled Water, PVC Membrane, Final Concentration of glucose = (600mg/dl) Chapter 4. PREPARATION OF THE ELECTRODES 80 Figure 4.4: Comparisons of Voltage-Time Responses of Conditioned Glass Electrodes, I: Com-plex No.5, II: Complex No.6, Tested in Deionized Distilled Water, P V C Membrane j Final Concentration of glucose = (600mg/dl) 4. Glass Micro Pipette. 5. Platinum Wire. 6. Cap. 7. Epoxy Resin. 8. Soft Glass. 9. Filling Solution. 10. Selective Membrane. 11. O^Ring. Figure 4.5: Construction of a Needle Type Glucose Sensors Chapter 4. PREPARATION OF THE ELECTRODES 82 00 120 160 CONCENTRATION (mg/ml) 80 120 160 CONCENTRATION (mg/ml) I I Figure 4.6: I: The Response of 16 Gauge Needle Electrode, Complex No.6 Using P V C Membrane as an Outside Layer, Measured in Different Concentrations of Glucose, II: The Response of a Needle Type Electrode, Complex No.6, Using Tissue Membrane as an Outside Layer, Measured in Different Concentrations of Glucose Chapter 4. PREPARATION OF THE ELECTRODES 83 used for implantation. Figure 4.7 shows the potentiometric response of a needle type electrode (21 gauge needle). Even though the electrode potential continued to change, glucose solution was added to the testing cell to observe its potentiometric responses. There was no indication of when the equihbrium potential would be reached. It was observed that glucose addition decreased the signal by up to 20mv. As can be seen, the initial voltage and final voltage were initially much lower than observed with the coated wire glucose sensors and glass electrodes. Other metal-glucose complexes were also used to observe the response and behavior of these electrodes. Figure 4.8 represents the potentiometric behavior of needle type electrodes using Ca-glucose and K-glucose complexes. Glucose solutions were again added up to 300(mg/dl) when a stable signal was recognized. After the addition of a glucose solution to the measuring cell, it was seen that the potential of the Ca-glucose electrode and the K-glucose electrode increased while the Ba-glucose complex electrode potential had decreased. By comparing the initial potentials and the final potentials of the needle electrodes prepared with different complexes, no reproducibility of the same trend was evident. Finally the electrodes were prepared in a membrane form to see if stability of signal and reproducibility of results can be achieved. 4.1.5 Preparation of Membrane Electrodes Different combinations of membranes were used to test the sensitivity and the stability of these electrodes to glucose. Preparation of P V C membrane electrodes: Complexes were dissolved in a P V C matrix: 1. 0.6g of the glucose copper salt and 2.4 ml of decanol and 1.6 ml of Aliquat were mixed by a magnetic stirrer over night. 2. 8 ml P V C (8g/100 ml THF) was added to the solution and stirred over night. 3. Take 0.5 ml of above solution and pour into a glass ring(2.0cm) after 2 hours, add 0.3 ml of the solution again and leave to dry overnight. Chapter 4. PREPARATION OF THE ELECTRODES 84 no -j3D — glucose Time, (min) I Figure 4.7: Voltage-Time Response of Conditioned Needle Electrode, I: Complex No . l , Tested in Deionized Distilled Water, Final concentration of glucose = (300mg/dl) Figure 4.8: Comparisons of Voltage-Time Responses of Conditioned Needle Electrodes, I: Com-plex No.2, II: Complex No.6, Tested in Deionized Distilled Water, Final concentration of glucose = (300mg/dl) Chapter 4. PREPARATION OF THE ELECTRODES 86 The same kind of procedure was followed to make a Ba-glucose complex membrane and the results were compared with a BaCl2 compound and CuCl2 compound. 4.1.6 P o t e n t i o m e t r i c R e s u l t s O b s e r v e d F r o m M e m b r a n e E l e c t r o d e s Since constant drift was recognized, it was thought that using the coating material in a mem-brane form (solid state) could improve the stability of the electrode signal. P V C electrodes were prepared using the same optimal composition of membrane material proposed in the literature [83]. Membrane electrodes were prepared according to the procedure described in the exper-imental section described earlier. These electrodes were also conditioned in 0.001M glucose in buffer solution prior to use. Figure 4.9 shows the potentiometric behavior of a membrane electrode tested on 3 consecutive days. Complex No.4 (Cu-glucose complex) was used in this electrode. The electrode potential was measured for more than five hours and its continual drift was observed. Glucose solutions were added to the measuring cell and the electrode potentials were observed. Lower potentials were observed compared to glass and needle electrodes and a stable potential was not acheived in the phosphate buffer solution for a long period of time. The behavior of the electrode was very unusual and no conclusive results could be obtained. Questions that came to mind at that time were: Does glucose concentration increase have any effect on the electrode response or does the potential of these electrodes continuously change until it reaches an equihbrium value. Figure 4.10 shows voltage-time responses of BaCl2 salt, Ba-glucose complex and CuCl2 salt membrane electrodes. Again, the behavior of the electrodes was unusual. After leaving the electrodes for a period of time in buffer solution, glucose solution was added and the potentiometric change was observed. No consistency was observed in the data obtained. Figure 4.11 shows the voltage-time responses of FeCh salt and Cu(AC)2 salt in phosphate buffer solutions after addition of 18(mg/dl) KCl and glucose. The arrow indicates the addition points of such compounds. The inner solution of the electrodes contain either Cu(AC)2 or FeCh, oxidizing agents, which interact with any reducing sugars such as glucose. The purpose of this test was to determine if the oxidation can be measured when F e 3 + .ions Chapter 4. PREPARATION OF THE ELECTRODES 87 changes to Fe2+ ions. Even though glucose is a reducing sugar, a signal can not be measured in a reproducible manner after its addition. Membrane electrodes did not show any promising results. The behavior of these electrodes was not explainable. Potential drifts were observed and instability of the potential continued. There was no substantial change after the addition of glucose. The initial and final voltages were not reproducible. The constant drift of potential was recognized in all prepared electrodes and it was decided to attempt to find another type of membrane which could control the transient behavior. Another membrane that was looked at was a polyacrylamide gel. This gel was prepared in the laboratory arid according to ref [110] had been found to be an effective membrane because of its pore size. The pore sizes could be controlled by varying the chemical compositions. Figure 4.12 shows the voltage-time responses of Ba-glucose complex electrodes using polyacrylamite gel. The preparation of the membrane is explained in section 3.6.3. The electrodes were also conditioned prior to use in 0.001M glucose in buffer solution. The transient behavior of the electrodes were all observed until a stable signal was acheived. Glucose was then added to the measuring cell. The behavior of the electrodes indicated that addition of glucose only disturbs the signal. Could glucose have any effect on the electrode response or do these electrodes behave in the same manner without addition of glucose? At this point it was realized that what was being observed with these electrodes were the transient electrode responses and addition of different glucose concentrations only caused a new transient potential response. These electrodes do not behave as specific electrodes because they are not selective for any specific ion or molecule. The behavior of Ba-glucose or copper glucose or other complexes either in the form of glass electrodes, needle electrodes or membrane electrodes did not show any trend in potentiometric changes. The ^reproducibility of the electrode potentials did not bode any promising results. With the present evidence and the results obtained in previous months, it was possible to postulate that the mechanism that had been proposed did not explain the glucose concentration-potential changes. The potentials observed with these electrodes were caused by their transient response to some unknown cause which continued until a stable potential was Figure 4.9: Voltage-Time Curve of PVC Membrane Electrode A, Using Cu-glucose Complex, Tested in 3 Consecutive Days, 100ml pH=7.4, Final concentration of glucose = (300mg/dl) Chapter 4. PREPARATION OF THE ELECTRODES -i 1 1-c o S e _! , 1— -J l L. 89 SO 120 160 200 2*0 280 Time, (min) _ l , 1 . ISO 200 2S0 I I I ISO Tim«, (min) Figure 4.10: Comparisons of Voltage-Time Responses of PVC Membrane Electrodes, Using I: BaCh, II: Ba-Glucose, IQ: CuCl?., Final concentration of glucose = (300mg/dl) Chapter 4. PREPARATION OF THE ELECTRODES ) 90 Figure 4.11: Comparisons of Voltage-Time Responses of Hydrogel Membrane Electrodes, Using Inner Solution of I: Cu(AC)2, II: FeCl3 , Final concentration of glucose = (300mg/dl) Chapter 4. PREPARATION OF THE ELECTRODES rro j - , . r" 1 • 1 • ' — Time, (min) I V — glucose Time, (min) Figure 4.12: Comparisons of Voltage-Time Responses of Ba-Glucose complex, Using I: Poly-acrylamite with 3500mv Cellophane, LI: Polyacrylamite with lOOOmv Cellophane, IE: Polyacry-lamite with 0.02/im Hydrophilic membrane , ILH: 0.03/mi Hydrophobic Membrane, Tested in pH=7.4, Final concentration of glucose = (300mg/dl) "•• Chapter 4. PREPARATION OF THE ELECTRODES 92 achieved. It was next decided to determine if the literature results can be reproduced. In chapter 5 the procedure followed to test these coated wire glucose electrodes either amperometrically or potentiometrically according to the procedures reported in literature will be explained. C h a p t e r 5 R E P R O D U C I B I L I T Y O F T H E L I T E R A T U R E R E S U L T S Since, as was shown in the previous chapter, the response of the glucose complex electrodes was not reproducible it was decided to try to reproduce the results given in the literature. Their experimental setup was used and their procedures were followed to obtain potentiometric responses. 5.1 P o t e n t i o m e t r i c M e t h o d According to the literature [83-84], "The equipment and circuit used for the monitoring of the glucose sensor using the coated wire for each glucose concentration and for continuous monitoring of different glucose concentration solutions is shown in figure 5.11 and figure 5.Ill, respectively. The response time of the sensor is studied by the so-called immersing measur-ing technique for the continuous monitoring experiment. The glucose sensor is immersed in a thermostat (310K) buffer solution and the voltammetric current time curve is recorded contin-uously. After steady state is reached, glucose solution of a known concentration is added into the buffer instantaneously. Figure 5.2 shows the behavior of the sensor when tested continuously to monitor varying glucose concentrations ranging from 80-200 (mg/dl). The response time is between 2 to 4 min. and the voltage difference is reasonable for each concentration difference. It can be seen from figure 5.2 that the sensor attains the steady state values for various glucose concentrations in a reproducible manner." In order to reproduce the behavior of electrodes reported previously, an identical experi-mental setup was prepared and their procedure was followed to give potentiometric readings. 93 Chapter 5. REPRODUCIBILITY OF THE LITERATURE RESULTS 94 To Meter Thermometer Calomel Ref. Elec t rode Glucose'Sensor Glucose S o l u t i o n Water (38">C) Constant Temperature Water Bath J Klguro 6. Potentiometric T e s t i n g System. Seal -Paraf i lm paper-To dra in fi o Reference electrode Variable speed pump -Glucose so lut ion I I Figure 4. Continuous glucose monitoring system. Figure 5.1: I: Schematic Drawing of the Potentiometric Testing System, LT: Continuous Moni-toring of Varying Glucose Concentration [83] Chapter 5. REPRODUCIBILITY OF THE LITERATURE RESULTS Figure 5.2: Continuous Monitoring of Varying Glucose Concentration [83] Chapter 5. REPRODUCIBILITY OF THE LITERATURE RESULTS 96 1 2 1 ' 3 7 1 L r 8 J 1. Temperature Controlled Polyethylene Cell 2. Reference Electrode 3. Glucose Sensor 7. Electrometer (Keithly 616, 617) 8. Chart Recorder Figure 5.3: Schematic Diagram of the Potentiometric Apparatus Chapter 5. REPRODUCIBILITY OF THE LITERATURE RESULTS 97 A schematic of the testing apparatus is shown in figure 5.3. Potentiometric responses were also observed by immersing prepared electrodes in conjunction with a saturated KCl reference electrode in a buffer solution in an electrochemical cell, either stagnant or stirred as was used in reference [83,86]. Glucose solution of a known concentration ranging from 80-200(mg/dl) was added into the buffer solution at different times. The electrodes were conditioned prior to use in 18(mg/dl) glucose in H20 or 18(mg/dl) glucose in buffer of pH 7.4 as they were in the literature references. 5.2 Results of the Potentiometric Test Figure 5.4 represents the potentiometric response of an electrode with Ba-glucose complex as compared to a glucose "control" electrode. Glucose solutions were added to the measuring cell and the electrode potential began to decrease. After approximately one hour, the electrodes were then placed in fresh buffer solution. The electrode potential began to decrease further. It was interesting to see that the potential did not go back to the original value. Since glucose addition did not change the slope of the potentiometric curve; 1. What is the actual cause of the potentiometric drift? 2. How long would it take for the electrode to reach a stable potential? This type of behavior could be expected to result from concentration changes. According to the proposed mechanism [85] as the concentration of glucose increases, the potential decreases. By using a new starting solution, the electrode should show its original signal. Figure 5.5 represents the behavior of I: Ba-glucose complex electrode, II: glucose electrodes. Addition of glucose-buffer solution did not change the potentiometric response. No conclusive results were obtained which could explain the reason for the transient behavior of the electrodes. It was found that the electrodes prepared for this work did not show the reproducibility reported in the literature. Chapter 5. REPRODUCIBILITY OF THE LITERATURE RESULTS 98 The continuous trace of a Ba-glucose complex electrode taken from the literature [83] is shown in figure 5.2. It shows that the potential decreases as the glucose concentration increases. This curve in some respects is similar to the transient behavior observed in this work. The slope changes with glucose addition may be a temporary upset due to this addition. It was observed from the experiments that the tested electrodes with or without Ba-glucose complex responded transiently and a stable potential was not observed until 12-13 hours. After the addition of the glucose solution to transiently behaving electrodes, it was found that the slope of the potentiometric signal did not change in a reproducible manner. By exposing the electrode to a fresh buffer solution, the signal continued to decrease. The results did not correspond to the results published in literature. A different approach was undertaken in order to explain the obtained results. Chapter 5. REPRODUCIBILITY OF THE LITERATURE RESULTS 99 I I I /3D — glucose Figure 5.4: Voltage-Time Curve of I: Ba-Glucose Complex (0.2g) , II: Controlled Glucose (0.6g) Electrodes Conditioned in 18(mg/dl) Glucose in H20 ,Tested in Buffer pH=7.4 Chapter 5. REPRODUCIBILITY OF THE LITERATURE RESULTS 100 Figure 5.5: Voltage-Time Curve of I: Ba-Glucose Complex , II: Controlled Glucose Electrodes Conditioned in 18(mg/dl) Glucose in Buffer pH=7.4 ,Tested in Buffer pH=7.4 Chapter 6 N E W A P P R O A C H T O T H E P O T E N T I O M E T R I C O B S E R V A T I O N S Since the electrode responses to glucose concentration reported in the literature could not be reproduced, it became necessary to study all the factors which could effect the responses in order to determine why the present electrodes were different. 6.1 Potentiometric Observation Factors that could influence the responses of the electrodes are: 1. Platinum Wire Surface. 2. Aliquat Concentration. 3. n-Decyl Alcohol Concentration. 4. Glucose Complex Amount. 5. Type of Complexes Used. 6. The Amount of BaCl2 • 7. Conditioning Effect. 8. Polymer/Solvent Ratio. • PVC/THF. Studies in this part of the research were made to observe the behavior of the electrodes with respect to: 101 Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 102 1. Effect of each individual component on the electrode response. 2. Effect of coating on the electrode response. .3. Effect of membrane on the electrode response. 4. Effect of conditioning on the electrode response. 5. Effect of alcohol adsorption to the platinum surface. 6. Effect of the platinum surface on the electrode response. 6.1.1 Description of Electrodes No.l-No.18 In this section all the electrodes that were used air dried are considered. The electrodes are identified according to their electrode number, the surface cleaning solution, the concentration of the salt involved, (where c stands for Ba-glucose complex ), Aliquat concentration, decanol concentration and the solution in which the electrodes were tested and finally the meter used. Meter # 1 was the Keithly 616 electrometer, meter #2 was the Keithly 617 electrometer, and finally meter # 3 was the Keithly 640 vibrating capacitor electrometer. Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 103 Table 6.1: Electrode Description Electrode Pt surface complex Aliquat Decanol solution meter# # cleaning amount (g) volume (ml) volume (ml) used 1 Acetone 0.3c 0.8 1.2 H20 1 2 Acetone 0.3c 0.8 1.2 H20 1 3 Acetone 0.3c 0.8 1.2 H20 1 4 Acetone 0.05c 0.1 1.0 H20 1 5 Acetone 0.05c 0.1 1.0 H20 1 6 Acetone 0.05c 0.1 1.0 H20 1 7 Acetone 0.6c 0.8 1.2 H20 1 8 Acetone 0.6c 0.8 1.2 H20 1 9 Acetone 0.6c 0.8 1.2 H20 1 10 Acetone 0.3c 0.8 1.2 Acetate buffer 1 11 Acetone 0.3c 0.8 1.2 Acetate buffer 1 12 Acetone 0.3c 0.8 1.2 Acetate buffer 1 13 Acetone 0.6c 0.8 1.2 Acetate buffer 1 14 Acetone 0.6c 0.8 1.2 Acetate buffer 1 15 Acetone 0.6c 0.8 1.2 Acetate buffer 1 . 16 Acetone 1.06c 0.8 1.2 Citric buffer 1 17 Acetone 1.06c 0.8 1.2 Citric buffer 1 18 Acetone 1.06c 0.8 1.2 Citric buffer 1 Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 104 6.1.2 Testing Solutions The electrodes were immersed and tested in the following solutions. 1. Acetate Buffer Solution pH=5.0 2. Phosphate Buffer Solution pH=7.4 • Buffer solutions containing (40,100,120,200) mg/lOOml of glucose. 3. Distilled water 6.1.3 Treatment and Cleaning of the Platinum Wires • Heating in gas flame. • Cleaned in concentrated nitric acid and sulfuric acid. • Cleaned in acetone. • Polished in aluminum oxide and etched in Aquaregia. 1. anodized in 0.5M sulfuric acid at 1.9V. 6.1.4 Results Obtained For Platinum Wires Treated with Acetone At this stage of the study, when reproducing the literature results for coated wire glucose electrodes was not possible (as shown in earlier section), one factor that came to mind was the effect of surface cleaning. Initially, the platinum surfaces were cleaned in acetone for 6-7 hours. Table 6.1 indicates which electrodes were treated with acetone. A literature review on platinum use suggested other preparation and cleaning procedures for platinum surfaces. These were used in later stages of the study. Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 105 Effect of C o m p l e x Concentration on Electrode Response It was assumed that only a few important components could influence the potentiometric re-sponse of these electrodes. The initial studies were on : 1. The concentration of the Ba-glucose complex. 2. The Aliquat to decanol concentration ratio. 3. The ratio of the polymer to the ion exchanger. The Aliquat to decanol and the polymer to the liquid ion exchanger ratios were kept constant and the amounts were chosen according to the optimum literature values given before [82-86]. The constituents of the prepared electrodes are given in Table 6.1. It was thought that by increasing or decreasing the complex concentration, the potentiometric signal might be affected. Figures 6.11, and 6.Ill show the behavior of electrodes 1,2,3 compared to electrodes 7,8,9 which contain different concentrations of Ba-glucose complex. In both cases the signal decreased for about 10-12 hours until a stable potential was reached. The initial voltages were found to be between 700-800 mv. It has been reported in the literature, that a platinum wire immersed in an acidic or alkaline electrolyte, produces a potential of about 800mv [111]. Table 6.2 tabulates the initial voltage, the steady state voltage , the different complex concentrations, and finally the amount of glucose added to the measuring cell and the final potential observed. From table 6.2, it can be observed that, electrodes constructed identically did not give reproducible responses. The potential decrease from the initial to the steady state conditions for the identical electrodes 1,2,3 were 308, 123, and 335mv, and for electrodes 4,5,6 which contained one sixth the complex the decreases were 199, 358, and 395mv. The average potential for electrodes with 0.3g Ba-glucose complex is 255mv compared to 317.3mv for electrodes with 0.05g of complex and 165.6mv when 0.6g of complex were used. It was observed from the average values that the lower the concentration of the Ba-glucose complex initially dissolved in the ion exchanger, the higher the potential drift. When comparing the potential drift of identical electrodes, one can see that the deviation between their observed potentials is too high to be able to conclude Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 106 B - Q Electrode fj 1 o - o Electrode # 2 /s-A Electrode # 3 ^\\ ^ V , , 1 1_ t$5Z> 1 6 6 >0 ' 2 Time, (hrs) 550 -500 -Time, (hrs) I I 0 2 < I Figure 6.1: Comparisons Between the Potentiometric Responses of I: 0.3g Ba-Glucose Complex, II: 0.6g Ba-Glucose Complex" Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 107 that there can actually be difference using different amount of complex. This scatter could in fact be caused by an external effect which after the present tests was unknown. Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 108 Table 6.2: Stability of Coated Wire Sensors Before and After Addition of Glucose. (Pt Cleaned with Acetone) Electrode Starting Final Complex (Final mv After Addition of Glucose) # mv mv grams A B C D 1 738 430 0.3 2 703 580 0.3 3 773 438 0.3 4 730 531 0.05 533 537 536 539 5 815 457 0.05 456 458 457 464 6 823 428 0.05 430 430 430 430 7 820 695 0.6 691 693 698 700 8 820 614 0.6 607 608 606 603 9 775 609 0.6 608 593 590 590 10 695 419 0.3 425 429 433 11 735 500 0.3 504 507 508 12 732 401 0.3 405 407 408 13 710 502 0.6 506 508 508 14 685 433 0.6 431 427 426 15 700 382 0.6 382 382 382 16 689 630 1.06 630 630 630 17 765 492 1.06 493 492 491 18 749 694 1.06 694 694 693 Amount of glucose added: A=0.05g, B=0.1g, C=0.2g, D=0.3g S ta t i s t i ca l A n a l y s i s Only a statistical study can determine if the observations can lead to any conclusions. A brief explanation of the terms used in the statistical analyses will be given here. A detailed explanation of the results is given in appendix B. The analysis was done to find the degree of significance of the effect on the electrode potential of the amount of Ba-glucose complex used, amount of BaCl2, pH changes, the life time of the electrode, glucose addition to the measuring solution, and finally the changes to the measuring solution while testing the electrode. The statistical analysis used is based on a general linear model (GLM) which uses the method of least squares to fit the given parameters. " S A S " is a software system for data analysis [165]. Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 109 The input program indicates "Proc G L M " which analyzes data within the frame work of general linear models, hence the name G L M . G L M handles classification variables, which have discrete levels, as well as continuous variables, which measure quantities. Thus G L M can be used for many different analyses including: 1. Simple regression. 2. Multiple regression. 3. Analysis of Variance (ANOVA), especially for unbalanced data. 4 . Analysis of covariance. 5. Response surface models. 6. Multivariable analysis of variance ( M A N O V A ) . 7. Polynomial regression. Specification of Effects: Each term in an analysis-of-variance model is an effect that is some specified combination of classification variables. Effects are specified with a special notation using variable names and operators. There are two kinds of variables: classification (or class) variables and continuous variables. There are two kinds of operators: crossing and nesting. The values of class variables are called levels. Crossed effects are specified by joining class variables with asterisks, for ex-ample A * B . The crossed effects correspond to the order of the variables in the class statement in such a way that the right most variables in the cross index faster than the left most vari-ables. Nested effects are specified by placing a parenthetical field after a variable or interaction indicating the class variable within which the effect is nested, for example B(A). The order of the variables within nesting parentheses is made so that variables outside the parentheses index faster than those inside the parentheses. The model could have crossed variables or nested vari-ables which depend on how the data was initially presented. The F value is the ratio produced Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 110 by dividing the mean square of the model by the mean square error. It tests how well the model as a whole accounts for the dependent variable's behavior. A small significance prob-ability, P R > F indicates that some linear function of the parameters is significantly different from zero. A very small value for this probability leads to the conclusion that the independent variable contributes significantly to the model. The R square finds all possible combinations of independent variables for the model it considers. The significance level, PR> |T|, is the proba-bility of getting a larger value of T if the parameter is truly equal to zero. A very small value for this probability leads to the conclusion that the independent variable contributes significantly to the model. The statistical studies are shown in different sections throughout the thesis. T h e Effect of Glucose on the Potentiometric Response After the steady signal was reached, a known amount of glucose as a solid (to prevent dilution or addition of new buffer) was added to the measuring cell to observe the potentiometric changes. The sensitivity of these sensors to other salts was examined very superficially by adding a small amount of KCl to the testing solution. It is observed that electrodes are more sensitive to KCl than to glucose. By comparing the initial and steady state potentials from table 6.2, one can compare the potentiometric responses. Steady state responses of electrodes 1,2,3 are shown in comparison to electrode 7,8,9 in figure 6.2 1,11 after addition of glucose. According to the literature [86], it was thought that the change in potential involves the association of the barium glucose complex as more glucose enters the sensing membrane, thus causing a decrease in charge. At this stage, the progress of the research showed that the addition of glucose had no substantial effect on the response of these electrodes after steady state was reached. The mechanism which was hypothesized did not seem to explain the responses of the coated wire glucose sensor. A n experiment was designed to study the significance of the response observed after glucose addition. Four levels of glucose were used (50mg, 150mg, 350mg, 650mg/dl) respectively. Four levels of Ba-glucose complex amount were used in the electrodes representing 0.05g, 0.3g, 0.6g Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 111 900 950 1000 '050 Time, (hrs) I G-a Electrode #7 Time, (hrs) I I Figure 6.2: Comparisons Between the Steady State Potentials of I: 0.3g Ba-Glucose Compex, II: 0.6g Ba-Glucose Complex After Addition of Glucose and K C l Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 112 and l.Og dissolved in 60v/v% aliquat to decanol ratio. The design used different levels of time chosen from the curve of the continuous measurements. Initial and stable potentials were recorded after the addition of glucose. The effect of glucose addition was analyzed statistically using a nested model. For details of the observations refer to appendix B. The reason for using the nested model was that the platinum wires that were tested, were different each time tests were conducted. The results showed that addition of different glucose concentrations and the use of different complex concentrations in constructing the electrodes did not effect sensor responses significantely. From the t test, it was observed that the prob> \T\ for glucose 1.000 , Ba-glucose amount 0.2664 and sensor 0.0002 for the initial potential and 0.9717 for glucose, 0.0012 for Ba-glucose complex and 0.0129 for the sensor at steady state potential. The degree of significance can be detennined if the PR> \T\ value is less than 0.005. The results from the regression show that the data fit the GLM(general linear model) model well. The R2 value of the fit was nearly 1, which indicates that the data fits the general linear model. From the contour plot of the data, it can be concluded that the data are very reliable since no trends were shown. It can be concluded that the addition of different concentrations of glucose to prepared electrodes with different concentrations of the Ba-glucose complex had no significant effect. The Life Time of Electrodes The life time of the electrodes was also observed. The term "life time" corresponds to each electrode which was tested for several days to observe the change in potentiometric response. Drift of the potential was measured and compared. Figure 6.3 shows the transient behavior of electrodes 1,2,3 as tested on 3 consecutive days to observe if the initial and final potentials measured for identical electrodes are the same or if drift of potential can be recognized. Figure 6.31 represents day 1, 6.311 represents day 2, and 6.3111 represents day 3. One can see that the initial and steady state voltages decrease each time the electrodes were tested. This could indicate that materials are leaching from the electrodes. Figure 6.4 shows the steady state potential of these electrodes after the addition of glucose on 3 consecutive days. There are Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 113 differences in steady state voltages and potential drop with time. It can be seen from figure 6.4 that the steady state potentials are lower after each test. It is observed that continual drifting of the electrode is possible. How and why it occurs, was not known initially. The effect of time on electrode response was also statistically analyzed. The effect of time was analyzed using crossed effects as described in appendix B. The factorial design used six levels of time chosen from the curve of the continuous measurements, three levels of Ba-glucose complex in the electrodes, and 3 levels of repeat testing, where the same electrodes were im-mersed in buffer solutions on 3 consecutive days. The R square value was found to be 0.7367 which indicates that the model fits the experimental data sucessfully but with scatter. By comparing the P R > F values, it was found that testing the electrodes on different days will cause a change in initial and steady state potential. From the contour plots of the data, it can be concluded that the data is very reliable since no trends were shown. By comparing the F values generated from the program, it was found that the effect of time had the highest degree of significancy being 75.85 as compared to repeat tests which was 19.79 and sensor which was 9.45. Higher F value indicates the importance of that variable within the model which was observed to be time. That explains why the constant drifting of the electrode potential can not be reproduced. By testing the electrodes on three consecutive days, it was found that steady state voltage decreased as the number of repeat tests increased. The Effect of Testing Solution on Electrodes Since there was an indication of electrodes being tested in the phosphate buffer solution pH of 7.4 in the literature [82-86], all the electrodes were prepared in phosphate buffer in order to observe the electrode behavior similar to the literature. In order to investigate how these electrodes respond in different solutions, their behavior was tested in H20 and acetate buffer. The reason for this was to see if similar behavior would be observed as compared to electrodes that were tested in phosphate buffer solution. The effect of the measuring solution was observed by immersing the electrodes in acetate Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 114 I Dayl Day2 Time, (hrs) O-G Electrode ft 1 V o-o Electrode # 2 4-6 Electrode # 3 N NT——_^ ; ' • ' Timo, (hrs) o-c Electrode #1 o-o Electrode #2 Electrode #3 Time, (nrs) 1 Figure 6.3: Comparisons Between the Potentiometric Responses of Electrodes 1,2,3 Testing on 3 Different Days Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 115 I Day I I I Day2 > $00 Q - O Electrode * 1 o-o Electrode M 2 Electrode # 3 KCl I (3D Time, (hrs) '•• KCl glucose ! Y —A k /glucose \^^. Q - Q Electrode # 1 o-o Electrode # 2 Electrode § 3 0 f glucose] • Time, (hrs) I I I Day 3 E (3D — glucose \ Electrode *1 0 - 0 Electrode #2 b—ti Electrode #3 (3D — glucose X /3D - glucose Time, (hrs) Figure 6.4: Comparisons Between the Steady State Potentials of Electrodes 1,2,3 Testing on 3 Different Days After Glucose Addition Figure 6.5: Comparisons Between the Potentiometric Responses of I: 0.3g Ba-Glucose Complex in Deionized E20 , II: 0.3g Ba-Glucose Complex in Acetate Buffer Solution Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 1 1 7 Figure 6.6: Comparisons Between the Steady State Potentials of I: 0.3g Ba-Glucose Complex in Deionized H20 , II: 0.3g Ba-Glucose Complex in Acetate Buffer Solution After the Addition of Glucose and KCl Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 118 buffer and deionized water. Figure 6.5 1,11 shows the potentiometric behavior of electrodes 1,2,3 tested in H20 in comparison to electrodes 10,11,12 which were tested in acetate buffer solution. Comparison of the initial and final potentials indicates that different measuring solutions have no definite effect. Figure 6.6 1,11 shows the steady state potential before and after addition of an interfering component (KCl). Again, by addition of the interfering components, it was observed that the change of potential is similar in the measuring solution. Table 6.3 shows the effect of testing solution on electrodes with the same concentration of complex. Delta V represents the difference between starting and steady state potentials. Electrodes 1,2,3 and 10,11,12 which contained 0.3g of Ba-glucose complex when tested in different testing solutions did not show any significant difference in response. The effect of different testing solutions on electrode response were also statistically analyzed. The effect of testing solution was analyzed using a nested effect as described in appendix B. The model indicated 2 levels of buffer solution and 6 levels of electrodes. Even though the electrodes were identically prepared, the same electrodes were not tested in the other buffer solution which resulted in nested effects rather than crossed effects. Comparing the P R > F values, the results showed 0.5569 for buffer and 0.6546 for sensor indicating that the parameters have no significant effect on the potentiometric response. Since the degree of significance shows values higher than 0.005, this indicates that the effect of buffer and sensor are not statistically significant. The contour plot showed that the data is very reliable since no trend was shown. This again confirmed the reproducibility of the electrodes. Output results are shown in appendix B. Cleaning the platinum surfaces in acetone, resulted in very unstable and irreproducible behavior as was observed in the potentiometric-time curves. In the next section different surface treatments will be used and the individual components will be isolated in order to see what is actually causing the drift. Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 119 Table 6.3: Effect of Testing Solution on Electrode Response Electrode Starting Steady state Delta V Concentration Testing # mv mv mv Grams Solution 1 738 430 308 0.3 Deionized H20 2 703 580 123 0.3 Deionized H20 3 773 438 335 0.3 Deionized H 2 0 10 695 419 276 0.3 Acetate Buffer 11 735 500 235 0.3 Acetate Buffer 12 732 401 331 0.3 Acetate Buffer 7 820 695 125 0.6 Deionized H 2 0 8 820 614 206 0.6 Deionized H 2 0 9 775 609 166 0.6 Deionized H20 13 710 502 208 0.6 Acetate Buffer 14 685 433 252 0.6 Acetate Buffer 15 700 382 318 0.6 Acetate Buffer Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 120 Conclusions D r a w n F r o m Results To Date (Acetone Cleaned Platinum) By observing all the above tables, figures and the statistical analysis, it can be concluded that: 1. The testing solution had no significant effect on the electrode response. 2. Increasing the complex concentration in the electrode had no significant effect on the electrode response. 3. The effect of adding an interfering component (KCl) was greater than for the addition of glucose which is the component to be selectively measured. 6.1.5 Isolation of Each Component of the Electrode In order to observe the cause of the transient responses, further tests were made to determine the role played by each component of the electrodes. To do this, the components of which the electrode was constructed were removed one by one and the potentiometric behavior was measured. For example, in order to observe if the Ba-glucose complex has any effect on the electrode response, it was replaced with pure BaCl2 and or glucose. Also the platinum surface was cleaned by immersing the total surface in HNO3 for 6 hours before thermal treatment in the reducing part of a natural gas flame for 2 to 3 minutes. Electrodes were first prepared, air dried and finally conditioned. The electrodes were again tested for the effect of glucose addition, the effect of Ba-glucose complex, the effect of buffer change and finally the effect of platinum surface on electrode response. The effect of conditioning on electrode behavior will be described in later sections. 6.1.6 A i r Dried Electrodes The term "Air Dried" means testing the prepared electrodes without exposing them to the con-ditioning solution prior to use. Table 6.4 indicates the levels of complex mass (0.0g,0.1g,0.3g,0.6g,0.8g), BaCl2 mass (0.0g,0.1g,0.3g,0.6g,0.8g) glucose mass (0.0g,0.1g,0.3g,0.6g,0.8g), Aliquat volume (0.0ml,0.8ml) and decanol volume (0.0ml,1.2ml) used in the experimental design. It is necessary Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 121 Table 6.4: Description of the Levels Used for Air Dried Electrodes Platinum complex BaCh Glucose Aliquat Decanol P v c / T H F wire amount (g) amount (g) amount (g) volume (ml) volume (ml) ratio 1 0.3 0.0 0.0 0.8 1.2 0.25 1 0.0 0.3 0.0 0.8 1.2 0.25 1 0.0 0.0 0.3 0.8 1.2 0.25 1 0.0 0.0 0.0 0.8 1.2 0.25 1 0.0 0.0 0.0 0.0 1.2 0.25 1 0.8 0.0 0.0 0.8 1.2 0.25 1 0.0 0.8 0.0 0.8 1.2 0.25 1 0.0 0.0 0.8 0.8 1.2 0.25 1 0.0 0.0 0.0 0.8 1.2 0.25 1 0.0 0.0 0.0 0.0 1.2 0.25 1 0.6 0.0 0.0 0.8 1.2 0.25 1 0.0 0.6 0.0 0.8 1.2 0.25 1 0.0 0.0 0.6 0.8 1.2 0.25 1 0.0 0.0 0.0 0.8 1.2 0.25 1 0.0 0.0 0.0 0.0 1.2 0.25 1 0.1 0.0 0.0 0.8 1.2 0.25 1 0.0 0.1 0.0 0.8 1.2 0.25 1 0.0 0.0 0.1 0.8 1.2 0.25 1 0.0 0.0 0.0 0.8 1.2 0.25 1 0.0 0.0 0.0 0.0 1.2 0.25 1 0.0 0.0 0.0 0.0 0.0 0.00 1 0.3 0.0 0.0 0.0 1.2 0.25 to note that different parameters and levels were statistically analyzed according to a specific test. This table only describes the levels used in the statistical observations. 6.1.7 Description of Electrodes No.19-No.71 The description of these electrodes is given in table 6.5. In this section all the electrodes that were air dried are identified. The electrodes are identified according to their electrode number, the platinum surface cleaning solution (each was flamed), the concentration of the complex involved, (where c stands for Ba-glucose complex, g stands for glucose and b stands for barium chloride), Aliquat concentration, decanol concentration and the solution in which the electrodes Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 122 were tested and finally the meter used. Only Meter # 1 the Keithly 616 electrometer was used in these tests. Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 123 Table 6.5: Description of Air Dried Electrodes Electrode Pt surface complex Aliquat Decanol solution meter# # cleaning amount (g) volume (ml) volume (ml) used 19 HNOz 0.3c 0.8 1.2 Phosphate buffer 1 ; 20 HN03 0.3c 0.8 1.2 Phosphate buffer 1 21 HNOz 0.3c 0.8 1.2 Phosphate buffer 1 22 HNOz 0.6c 0.8 1.2 Phosphate buffer 1 23 HN03 0.6c 0.8 1.2 Phosphate buffer 1 24 HN03 0.6c 0.8 1.2 Phosphate buffer 1 25 HNOz 0.1c 0.8 1.2 Phosphate buffer 1 26 HNOz 0.1c 0.8 1.2 Phosphate buffer 1 27 HNOz 0.1c 0.8 1.2 Phosphate buffer 1 28 HNOz 0.8c 0.8 1.2 Phosphate buffer 1 29 HNOz 0.8c 0.8 1.2 Phosphate buffer 1 30 HNOz 0.8c 0.8 1.2 Phosphate buffer 1 31 HNOz 0.8c 0.8 1.2 Phosphate buffer 1 32 HNOz 0.8c 0.8 1.2 Phosphate buffer 1 33 HNOz 0.8c 0.8 1.2 Phosphate buffer 1 34 HNOz 0.6c 0.8 1.2 Phosphate buffer 1 35 HNOz 0.6c 0.8 1.2 Phosphate buffer 1 36 HNOz 0.6c 0.8 1.2 Phosphate buffer 1 37 HNOz 0.6c 0.8 1.2 Phosphate buffer 1 38 HNOz 0.6c 0.8 1.2 Phosphate buffer 1 39 HNOz 0.6c 0.8 1.2 Phosphate buffer 1 40 HNOz 0.8g 0.8 1.2 Phosphate buffer 1 41 HNOz 0.8g 0.8 1.2 Phosphate buffer 1 42 HNOz 0.8g 0.8 1.2 Phosphate buffer 1 43 HNOz 2.0c 0.8 1.2 Phosphate buffer 1 44 HNOz 2.0c 0.8 1.2 Phosphate buffer 1 45 HNOz 2.0c 0.8 1.2 Phosphate buffer 1 46 HNOz 0.8g 0.8 1.2 Phosphate buffer 1 47 HNOz 0.8g 0.8 1.2 Phosphate buffer 1 48 HNOz 0.8g 0.8 1.2 Phosphate buffer 1 49 HNOz 0.6g 0.8 1.2 Phosphate buffer 1 Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 124 Table 6.5 (continued) Electrode Pt surface complex Aliquat Decanol solution meter# # cleaning amount (g) volume (ml) volume (ml) used 50 HN03 0.6g 0.8 1.2 Phosphate buffer 1 51 HN03 0.6g 0.8 1.2 Phosphate buffer 1 52 HN03 0.6b 0.8 1.2 Phosphate buffer 1 53 HN03 0.8b 0.8 1.2 Phosphate buffer 1 54 HN03 0.8b 0.8 1.2 Phosphate buffer 1 55 HNO3 0.8b 0.8 1.2 Phosphate buffer 1 56 HNO3 0.8b 0.8 1.2 Phosphate buffer 1 57 HNO3 0.0 0.8 1.2 Phosphate buffer 1 58 HN03 0.0 0.8 1.2 Phosphate buffer 1 59 HN03 0.0 0.8 1.2 Phosphate buffer 1 60 HN03 0.0 0.8 1.2 Phosphate buffer 1 61 HN03 0.0 0.8 1.2 Phosphate buffer 1 62 HN03 0.0 0.8 1.2 Phosphate buffer 1 63 HN03 0.3b 0.8 1.2 Phosphate buffer 1 64 HN03 0.3b 0.8 1.2 Phosphate buffer 1 65 HN03 0.3b 0.8 1.2 Phosphate buffer 1 66 HN03 0.3b 0.8 1.2 Phosphate buffer 1 67 HN03 0.3b 0.8 1.2 Phosphate buffer 1 68 HN03 0.3b 0.8 1.2 Phosphate buffer 1 69 HN03 0.1c 0.8 1.2 Phosphate buffer 1 70 HN03 0.0 0.8 1.2 Phosphate buffer 1 71 HN03 0.8c 0.8 1.2 Phosphate buffer 1 Effect of Individual Components on Electrode Response The experimental design used for this part of the study was made to show the effect of each individual component on the sensor response. The factors which were kept constant during the experiment were: 1. The surfaces of the platinum wires were cleaned and washed with 70% HN03 acid, then heated in the non oxidizing portion of a gas flame for 2 minutes. . 2. All electrodes were tested in a phosphate buffer solution of pH of 7.4. 3. All electrodes were tested at a temperature of 37°C. Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 125 4. AU electrodes were air dried in a dead air hood before testing. 5. Aliquat /decanol ratio used was the optimum ratio reported in previous publication(60 v/v%)[84]. 6. The electrodes were dipped in such a fashion that the tip of the platinum would not be outside the coating layer. All electrodes were dipped 5 times within the same time interval. 7. Each set of electrodes was tested for a fixed time until the potential reached steady state. 8. Glucose was added after the signal had become stable so that the effect of glucose could be separated from other effects (for example time). Each coated wire electrode consists of a platinum wire, either the complex, BaCl2 or glucose, and a mixture of aliquat and decanol mixed in the PVC/THF. To investigate the effect of BaCl2 or pure glucose in the coating the pure component and the response was recorded. The potential differences of the sensors versus a double junction reference electrode (Sat ArC/)were determined either in stagnant or stirred phosphate buffer solution of pH 7.4. For each run 6 electodes were tested. After a steady state signal was observed, glucose was added to the buffer to see the change in potential. These electrodes were tested for several days to observe their total decay time. Effect of Glucose Addition to the Buffer Solution on Electrode Response When the potential between the two electrodes in buffer solution had reached steady state, this value was defined as base-line potential. A sufficient amount of /3D-glucose was added to bring the glucose concentration up to 100(mg/dl). When the potential reached a stable reading where potentiometric response was constant, another incremental glucose addition was made to bring the concentration up to 600(mg/dl). The potentiometric results were expressed as the change in potential with respect to the base line potential. Table 6.6 shows the initial Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 126 voltage, steady state voltage and final voltage observed for this type of electrode before and after glucose addition. The change in potential is the difference between the starting and the final potentials. The effects of glucose addition were observed after there was no indication of any potentiometric change. Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 127 Table 6.6: Stability of Coated Wire Sensor After Addition of Glucose. (Platinum Treated with HNOz Acid and flame treated) Electrode Starting Final (Final mv After Addition of Glucose) Time Measured # mv mv B C D Hrs 25 776 380 378 380 379 13 26 767 347 346 346 346 13 27 700 437 438 438 436 13 28 734 687 687 689 690 13 29 776 558 558 555 553 13 30 764 542 543 543 542 13 31 747 718 715 716 717 7 32 760 733 732 730 729 7 33 701 743 742 741 738 7 34 730 767 766 765 763 3 35 740 702 702 701 694 3 36 761 704 702 701 697 3 37 732 671 670 669 665 13 38 736 666 666 668 662 13 39 715 732 732 732 732 13 Amount of glucose added: B=0.1g, C=0.2g, D=0.3g In order to observe if glucose addition could effect the potentiometric response while the transient response was still taking place, glucose were added at different times. As indicated in figure 6.7, electrodes 19,20,21 were prepared using 0.3g Ba-glucose complex in comparison to electrodes 22,23,24 which were prepared using 0.6g Ba-glucose complex. /3D-glucose was added while the potential was still changing. It was thought that the Ba-glucose complex might have weak association with the ion exchanger thus this experiment was made to determine if the addition of /3D-glucose at an earlier time could influence the signal . It was found that the addition of the sugar did not change the potentiometric response in any way. Figure 6.8 represents the potential drift of electrodes 25,26,27 which contained O.lg of the Ba-glucose complex, and in this case the steady state potential was observed. Again, the effect of the addition of /3D-glucose to the measuring solution was negligible. It was observed again that glucose addition did not change the potentiometric response either transiently or at steady Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 128 state. The effect of an increase in concentration of the Ba-glucose complex, decreased the slope of the potentiometric response as was observed in section 6.1.4. Again, the effect of glucose addition was analyzed statistically using a cross model in com-parison with a nested model having platinum wires treated with HNO3 and flame. Results include comparisons between three levels of complex concentration contain 0.3,0.6,0.8g of Ba-glucose complex and 3 levels of glucose 100,300,600(mg/dl). The initial and final potentials were measured after each addition of glucose. Again, F values of different levels of complex concentration with glucose addition were compared to identify the degree of significance. The crossed model showed an R square of 0.51666 for initial voltages and 0.9107 for steady state voltages. It indicates that all possible combinations of independent variables used fits the G L M model. The closer the observed value to 1, the better the fit. The PR>F values for sources such as glucose levels, type levels, and sensor levels and glucose type interaction levels were found to be nearly 1, 0.0127, 0.00064, 1.000 for initial concentration and 0.999, 0.0001, 0.0288, and 1.000 after addition of each glucose level until stable potential was achieved. The T test indicated the significance of glucose, electrode type, and sensor when initial voltages were mea-sured. It was found that glucose addition,did not significantly affect the potential change even when the type of electrode was changed. The contour plot showed that the data is very reliable since no trends was shown. This confirmed that addition of /3D-glucose did not change the potentiometric response. For more detail information, refer to appendix B. The Effect of BaCl2 Salt In the previous section, it was found that for the range of concentrations of the Ba-glucose complex tested, glucose addition did not change the potentiometric response. In order to observe if the Ba-glucose complex has any effect on the electrode response, it was replaced with pure BaCl2 and glucose. Since the complex itself was prepared by mixing /3D-glucose with BaCl2 , it was the purpose of this test to detennine how the pure compounds would respond, if at all. Would the Ba-glucose complex response be any different from that with the pure Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 129 40 80 120 Time, (min) I I Figure 6.7: Comparisons Between the Transient Potentiometric Responses of I: 0.3g Ba-Glucose Complex, II: 0.6g Ba-Glucose Complex After Addition of Glucose Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 130 o-oElectrode #25 Electrode #26 a-^Electrode #27 6 9 Time, (hrs) <00 i H &-o o > 300 T—•—r {3D — glucose i • i 1 i 1 — r A 6 _ /3Z> — glucose / /3Z? — glucose - ] — i — ] — , — | -a. o - Q Electrode #25 o-oElectrode #26 ^ E l e c t r o d e #27 -e B-—Q - 1 — ' — l — i — l — i — 1 — i I . i • i . i • i U.S0.S.SS 13.60 13.65 t3.70 I3.7S 13.80 13.8S " 13.10 13.SS t< W.0S u . i o J J Time, (hrs) Figure 6.8: Comparisons Between the Steady State Potentials of O.lg Ba-Glucose Compl. Transiently, II: After Glucose Addition Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 131 compounds? The potentiometric responses of electrodes 40 to 71 are shown in table 6.7. These responses are compared with each other according to their differences in: 1. Ba-glucose complex concentration. 2. Decanol concentration. 3. Aliquat concentration. 4. Concentration of • BaCl2 • • Glucose. Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 132 Table 6.7: Stability of Coated Wire Sensors (Platinum treated with HNO3 Acid and Thermally Flamed Electrode Starting Final change in Time Measured Amount # mv mv mv Hrs Used(g) 40 775 421 354 6 0.8g 41 734 458 276 6 0.8g 42 749 382 367 6 0-8g 43 736 518 218 13 2.0c 44 723 671 52 13 2.0c 45 715 577 138 13 2.0c 49 689 344 345 13 0.6g 50 712 370 342 13 0-6g 51 738 391 347 13 0.6g 52 824 612 212 4 0.8b 53 785 750 35 4 0.8b 54 772 719 53 4 0.8b 55 802 767 35 3.5 0.8b 56 796 793 3 3.5 0.8b 57 765 325 440 13 0.0 58 805 318 487 13 0.0 59 813 580 233 13 0.0 60 769 469 300 5.5 0.0 61 790 424 366 5.5 0.0 62 788 400 388 5.5 0.0 63 760 345 415 13 0.3b 64 737 335 402 13 0.3b 65 762 502 260 13 0.3b 66 794 642 152 6 0.3b 67 744 556 188 6 0.3b 68 712 582 130 6 0.3b 69 718 551 167 10 0.1c 70 801 438 363 10 0.0 71 740 643 97 10 0.8c where c=Ba-glucose complex, b=BaCl2, g=glucose. Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 133 Figure 6.9 shows the response of BaCl2 electrodes when 0.0, 0.3 and 0.8 grams of the BaCl2 was used instead of the Ba-glucose complex dissolved in the liquid ion exchanger. It was observed that by increasing the BaCl2 concentration, the drift of the potentiometric response decreased. In order to see how glucose would play a role in these electrodes, the Ba-glucose complex was replaced by glucose. Figure 6.10 shows the responses of electrodes containing 0.8g glucose, 0.8g Ba-glucose complex, and 0.8g of BaCl2 . The potential behavior of the electrodes with 0.8g of glucose is similar to that when no BaCl2 was added to the ion exchanger. This could in fact bring out another point, that, if the Ba-glucose complex dissociates rapidly (if at all), there will be no substantial change between the BaCl2 electrode and Ba-glucose electrodes except that the concentration of the inorganic salt in the Ba-glucose complex is lower. However, since glucose does not carry any charge transfer, the movement of glucose out of the liquid ion exchanger will be due only to self diffusion. From this it could be implied that there is no association of glucose and the ion exchange sites, thus the presence of glucose in the ion exchange mixture has no major effect on the potentiometric response except that it could cause a slower passage of other ionic substances. Thus, this could in fact confirm that the Ba-glucose complex is partially soluble since glucose can not dissolve in the ion exchanger. To test the effect of ionic loading of the ion exchange resin, electrodes made with 0.0, 0.1 and 0.8 grams of Ba-glucose complex were tested and their responses are shown in figure 6.11. Again, it is seen that as the concentration of the salt increases, more ion exchange capacity is used up and thus the slope of the potential drift decreases. It is possible that the ion exchanger has a strong association with the BaCl2, leaving the sugar free to move since glucose is nonionic and does not associate with the ion exchanger. It can also be observed that by increasing the concentration of the pure compound (BaCl2), the slope of potential drift decreased. The principle on which this type of electrode works requires the complete association of the salt with the ion exchanger. This leads to the formation of a neutral molecule in the presence of the ion exchanger. It is understood at this point that the presence of the complex has no effect on the electrode selectivity (or lack of it) towards Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 134 glucose. In order to confirm that the complex decomposes, N M R analysis was performed on pure decanol, aliquat in decanol, Ba-glucose complex in an aliquat/decanol mixture, pure glucose and finally the BaCl2 in aliquat/decanol mixture. It was found that glucose does not dissolve in the ion exchanger. This indicates that Ba-glucose complex partially dissolves in the liquid ion exchanger. The N M R peaks obtained for these results are shown in appendix C. Table 6.7 shows the behavior of the coated wire electrodes when Ba-glucose complex has been replaced with the pure compounds (BaCl2, glucose). The transient and steady state potentials were compared. The description of the electrodes are given in section 6.1.7. It was observed that air dried electrodes with the same treatment and the same coating did not give reproducible potentiometric results. One thing that was not clear is the relation between the transient behavior of the prepared electrodes and the irreproducibility of their steady state potentials. Again, the behavior of these electrodes was tested for several days to evaluate their responses in buffer solution and to see how the potentiometric decay of these electrodes can effect the signal with time. It was observed that the larger the amounts of Ba-glucose complex there was, the smaller was the drift in potential. The potentiometric responses are shown in appendix D. It can be concluded that the transient response could be due to the gradual loss of the ion exchanger capacity. Reproducibility of electrode response was still impossible and the effect was not well understood. The air dried electrodes had to be compared with conditioned electrodes to observe the effect of conditioning on the electrode response. Since the original electrodes described in the literature had always been conditioned prior to use, the effect of conditioning had to be studied in order to understand if initial and steady state potentials can be reproduced after conditioning. This will be explained in the section on "Conditioned electrodes". Again, the effect of BaCl2 concentration was analyzed using a nested model. Voltage was chosen as the dependent variable, and the independent variables included 12 sensors, 3 levels of BaCl2 amount; 0.0, 0.3, 0.8 grams. Different levels of time (6) were chosen, since the measurements were taken continuously. The estimated R square was found to be 0.7241; It Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 135 o - o Electrode £57 » - o Electrode #56 Electrode #59 no too no no l w p 4M 400 350 J0O ISO Tim«, (hrs) T * o i « , ( h r s ) ' I ' 1 ' 1 f1 aeuffi . n. - - - --• »-o Electrode #55 • A-« Electrode #56 -• 1 . i . i -I I I I I Figure 6.9: Comparisons Between Transient Potentiometric Responses of I: O.Og BaCl2, II: 0.3g BaCl2, IH: 0.8g BaCl2 Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 136 ° - ° Electrode #40 °~° Electrode #41 Electrode #42 o-o Electrode #28 o-« Electrode #29 4 - « Electrode #30 Tims; (hr*) I I •so •M no TOO •so (00 550 soo 4SO 400 ISO SOO ISO o-o Electrode #55 « Electrode #56 I I I Tim*, (hrt) Figure 6.10: Comparisons Between Transient Potentiometric Responses of I: 0.8g Glucose, II: 0.8g Ba-Glucose Complex , III: 0.8g BaCl2 Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 137 j. £ 400 o-o Electrode #60 »>« o-o Electrode #01 roa Electrode #62 450 -too -iiO -SOO OO .... 'OO 1 1 . t . 1 « Tim«. (tm) 1 1 < i — r - . i o-aElectrode #25 o-o Electrode #26 ^Electrode #27 -, t . i _ i — . — i — , — i — . — I I o-a Electrode #28 »-o Electrode #29 4 - 4 Electrode #30 in Tim*, (tv-x) Figure 6.11: Comparisons Between the Potentiometric Responses of I: O.Og Ba-Glucose Complex II: O.lg Ba-Glucose Complex, III: 0.8g Ba-Glucose Complex. Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 138 indicates that all possible combinations of independent variables used fits the G L M model. The closer the value to 1, the better the fit. Comparing the F values between time, component and sensor(component), gave 233.37, 73.61, 10.21. This is indication that time had the biggest effect compared to component concentration, which confirms that the electrodes are being stabilized with time. The P R > F observed were 0.001 for BaCl2 and 0.001 for time and 0.001 for sensor which indicates that all the above levels are statistically significant. T h e Effect of Buffer Change on Electrode Response It was then decided to test these electrodes in a fresh buffer solution after they had reached their steady state potential to determine whether the phosphate concentration had any effect. The electrodes were placed in a new measuring solution after reaching steady state. Figure 6.12 shows that the potential again showed a steeper decrease even though there had been no change in potential when the electrode was in the first buffer solution which was initially of the same concentration. Table 6.8 shows the change in potential between the initial and final potential, Delta V and the time needed for the electrodes to reach an equihbrium potential. It can be seen that the second time the electrodes were immersed in the phosphate buffer solution the Delta V (relative to initial steady state value) is almost the same as observed initially, but the time to steady state is much shorter. The degree of significance of the electrode behaviors was analyzed with respect to time and with placing the electrodes in new buffer solution using the crossed modeL It was found that the change of buffer significantly changed the electrode behavior and the effect of time was significant. The R square observed was 0.988 which indicated that the data fit the general linear model. The effect of buffer change showed that the electrode potential decreases with time significantly. The computer output of the results generated are presented in appendix B. Conclusions of the Results Obtained for A i r D r i e d Electrodes By observing all the above tables and figures , it can be concluded that: Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 139 Figure 6.12: Observations Between the Steady State Potentials of 0.8g Glucose Compound When Changing the Buffer Solution Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 140 Table 6.8: The Effect of Buffer Change on Electrode Response Electrode Starting Final Delta V Time Measured # mv mv mv Hrs 46 711 540 171 10 47 703 477 226 10 48 755 539 216 10 46* 548 318 230 3 47*. 498 298 200 3 48* 543 339 204 3 * indicates that the electrodes were placed in a new buffer solution. 1. The Ba-glucose complex is not a necessary component for the type of response observed for the coated wire electrodes. Addition of /3D-glucose did not change the signal, which indicates that the coated wire glucose sensor is not responsive to glucose. 2. The effect of an increase in concentration of BaCl2 as replacement for the Ba-glucose complex in the electrode showed a greater association of the ion exchange sites with the BaCl2 than with the Ba-glucose complex. This could indicate greater availability of the Ba2Jr ion to an ion exchange site. 3. Since glucose is nonionic and does not dissolve in the ion exchanger, it is possible that its effect is insignificant. 4. By increasing the Ba-glucose complex or BaCl2 concentrations in the electrode, the rate of potentiometric drift decreased. 5. Potentiometric responses observed for coated wire glucose sensors with the same treatment did not show the same steady state potential. 6. It is possible that there may be a surface reaction between the components present in the electrodes and the platinum wire. 7. If surface potential has any effect on the electrode response, immersing only part of the platinum could result in different potentiometric responses. Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 141 8. The transient response of the electrode could be the result of substantial loss of the ion exchanger capacity. 6.1.8 Comparison of Literature Results With the Data Obtained From the previous literature [86], the mechanism was assumed to be dependent on the Ba-glucose complex which is partially dissociated inside the liquid exchange membrane into Ba2+ and glucose2~. As the external glucose concentration changes, the equihbrium inside the mem-brane shifts, and the amount of free Ba2+ changes. Increased external glucose concentration leads to greater association, and a lesser availabiUty of free Ba2+. The free Ba2+ leads to the electrical potential as: Ba * Glucose ^ Ba2+ + Glucose2' (6.1) It has been shown that the Ba-glucose complex does not appear to have any effect on the electrode response. In fact, glucose concentrations can not be measured with such an electrode. The transient response is most likely due to the loss of capacity of the liquid ion exchanger. Figure 6.13 taken from the literature shows the transient behavior of a coated wire glucose sensor and a control electrode which doesn't contain any Ba-glucose complex, when both were exposed to different glucose concentrations. One reason for the observations may be the fact that the potentiometric responses didn't reach equihbrium before they were tested in a higher concentration of glucose. The behavior of electrodes shown in figure 6.131, can be cross plotted to give 6.13II which shows the behavior of different electrodes in different concentration of glucose (According to the reference [83]). This type of measurement can lead to false conclusions. It is necessary to obtain a stable potential before changing the glucose concentration. Figure 6.14 shows the potentiometric responses of electrodes 69,70,71 until an equihbrium potential is reached. As can be observed from this figure, it takes considerable time for air dried coated wire glucose sensors to reach an equihbrium potential. It should be kept in mind that the electrodes whose responses were reported in the literature [83,84] were conditioned in 18(mg/dl) glucose in ^ 0 or 18(mg/dl) Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 142 glucose i n buffer prior to use. The ini t ia l and final potentials observed and the f inal observation w i l l be used later to explain a possible mechanism. For a better understanding of the action of l iquid ion exchangers and their potentiometric responses, the following model may be used. Since an ion cannot exist by itself i n the l iq-u i d membrane phase without violating the requirements of electroneutrality, it is necessary to provide a charge "site" of opposite sign as an addition to the solvent phase. The ion (qua-ternary ammonium ion) and the salt (inorganic compound) are dissolved as neutral molecules, essentially undissociated i n the base solvent. The salt ion must have a relatively high molecular weight to insure that the resulting salt w i l l be substantially insoluble i n water. The salt must be soluble i n the low dielectric constant solvent. In a l iquid membrane system, the site and the ion (quaternary ammonium ion) move together through the membrane phase. A t the membrane interface, a process of ion exchange can take place between the ions of the ion-site salt (Cl~) i n the organic phase and the free ions in the aqueous phase. The selectivity of the electrode depends oh the selectivity of the ion exchange process. In order for the electrode to display Nernst ian response to an ion A++ i n the presence of an interfering ion B++, it is necessary that the position of the following equil ibrium he far to the left. B++(aqs) + AR2(org) ^ BR2(org) + A++(aqs) (6.2) where: R = The charge "site" group. P u t t i n g this in other terms, selectivity requires that the site R forms a more stable complex w i t h the sought- for ion than with any potentially interfering ion i n the sample. This type of electrode can be responsive i f a highly insoluble salt can be formed. Referring to the literature [112], the evaluation of the PO\~ electrode was complicated by instabil i ty of the cell emf even after extensive soaking of the electrode. Potential " d r i f t " as great as 1 mv per 10 min was observed. B y reading the cell emf at a constant time interval after electrode immersion, crude calibration curves were obtained over the phosphate concentration range of 10~ 2 to 1 0 - 4 M . Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 143 Figure 6.13: Comparisons Between Responses of Coated Wire Glucose Sensors from Literature [83] Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 144 Figure 6.14: Comparisons Between the Potentiometric Responses of Electrode 69 (O.lg Ba-Glucose Complex), Electrode 70 (O.Og Ba-Glucose Complex), Electrode 71 (0.8g Ba-Glucose Complex) Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 145 What can be the cause of the transient potentiometric response in the electrodes in this study? The loss of liquid ion exchanger capacity could be caused by the phosphate ion in the buffer solution being exchanged with Cl~ ion of the aliquat ion exchange material until it equilibrates (According to Eq.6.1). By observing the data which showed similar trends, one can hypothesize that the electrode which carries a high concentration of Cl~ ions, when exposed to the phosphate buffer which has a high concentration of PO\~, will interchange phosphate for chlorine ions. Since the association of aliquat with the phosphate is weak, the potential will change transiently until all the Cl~ ions have been exchanged. That is when a stable potential is achieved. 6.1.9 C o n d i t i o n e d E l e c t r o d e s The term conditioned means the exposure of the air dried electrodes to the measuring solution for a period of time. As indicated in the previous literature [83,84], the prepared electrodes were conditioned in 0.001M glucose/buffer solution prior to use. The testing method for this part of the experiment were the same as for the previous experiments except that the prepared electrodes were conditioned for a period of time either in solution of 0.001M glucose in the phosphate buffer or just the phosphate buffer at a pH of 7.4. Table 6.9 indicates the levels of Ba-glucose complex mass (0.0g,0.3g), Aliquat volume (0.0ml,0.8ml) and decanol volume (0.0ml,1.2ml), used in the experimental design. 6.1 .10 D e s c r i p t i o n o f E l e c t r o d e s N o . 7 2 - N o . 1 4 4 In this section all the electrodes that were conditioned are identified. According to, the platinum surface cleaning solution, the concentration of the complex involved, Aliquat concentration, decanol concentration and the solution in which the electrodes were tested and finally the meter used. Meter #2 was the Keithly 617 electrometer. Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 146 Table 6.9: Conditioned Electrode Description Platinum wire complex amount (g) Aliquat volume (ml) Decanol volume (ml) P v c / T H F ratio 1 0.3 0.8 1.2 0.25 1 0.0 0.8 1.2 0.25 1 0.0 0.0 1.2 0.25 1 0.3 0.0 1.2 0.25 1 0.0 0.0 0.0 0.25 1 0.0 0.0 0.0 0.0 Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 147 Table 6.10: Description of Conditioned Electrodes Electrode Pt surface complex Aliquat Decanol solution meter# # cleaning amount (g) volume (ml) volume (ml) used 72 HN03 0.3c 0.8 1.2 Phosphate buffer 2 73 HN03 0.3c 0.8 1.2 Phosphate buffer 2 74 HN03 0.3c 0.8 1.2 Phosphate buffer 2 75 HN03 0.3c 0.8 1.2 Phosphate buffer 2 76 HN03 0.3c 0.8 1-2 Phosphate buffer 2 77 HN03 0.3c 0.8 1.2 Phosphate buffer 2 78 HN03 0.3c 0.8 1.2 Phosphate buffer 2 79 HN03 0.3c 0.8 1.2 Phosphate buffer 2 80 HN03 0.3c 0.8 1.2 Phosphate buffer 2 81 HN03 0,3c 0.8 1.2 Phosphate buffer 2 82 HN03 0.3c 0.8 1.2 Phosphate buffer 2 83 HN03 0.3c 0.8 1.2 Phosphate buffer 2 84 HN03 0.0 0.8 1.2 Phosphate buffer 2 85 HN03 0.0 0.8 1.2 Phosphate buffer 2 86 HN03 0.0 0.8 1.2 Phosphate buffer 2 87 HN03 0.0 0.8 1.2 Phosphate buffer 2 88 HN03 0.0 0.8 1.2 Phosphate buffer 2 89 HN03 0.0 0.8 1.2 Phosphate buffer 2 90 HN03 0.0 0.8 1.2 Phosphate buffer 2 91 HN03 0.0 0.8 1.2 Phosphate buffer 2 92 HN03 0.0 0.8 1.2 Phosphate buffer 2 93 HN03 0.0 0.8 1.2 Phosphate buffer 2 94 HN03 0.0 0.8 1.2 Phosphate buffer 2 95 HN03 0.0 0.8 1.2 Phosphate buffer 2 96 HN03 0.0 0.0 1.2 Phosphate buffer 2 97 HN03 0.0 0.0 1.2 Phosphate buffer 2 98 HN03 0.0 0.0 1.2 Phosphate buffer 2 99 HN03 0.0 0.0 1.2 Phosphate buffer 2 100 HN03 0.0 0.0 1.2 Phosphate buffer 2 101 HN03 0.0 0.0 1.2 Phosphate buffer 2 102 HN03 0.0 0.0 1.2 Phosphate buffer 2 Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 148 Table 6.10 (continued) Electrode Pt surface complex Aliquat Decanol solution meter# # cleaning amount (g) volume (ml) volume (ml) used 103 HNOz 0.0 0.0 1.2 Phosphate buffer 2 104 HNOz 0.0 0.0 0.0 Phosphate buffer 2 105 HN03 0.0 0.0 0.0 Phosphate buffer 2 106 HNOz 0.0 0.0 0.0 Phosphate buffer 2 107 HNOz 0.3c 0.8 1.2 Phosphate buffer 2 108 HNOz 0.3c 0.8 1.2 Phosphate buffer 2 109 HNOz 0.3c 0.8 1.2 Phosphate buffer 2 110 HNOz 0.0 0.0 0.0 Phosphate buffer 2 111 HNOz 0.0 0.0 0.0 Phosphate buffer 2 112 HNOz 0.0 0.0 0.0 Phosphate buffer 2 113 HNOz 0.0 0.0 0.0 Phosphate buffer 2 114 HNOz 0.0 0.0 0.0 Phosphate buffer 2 115 HNOz 0.0 0.0 0.0 Phosphate buffer 2 116 HN03 0.0 0.8 1.2 Phosphate buffer 2 117 HNOz 0.0 0.8 1.2 Phosphate buffer 2 118 HNOz 0.0 0.8 1.2 Phosphate buffer 2 119 HNOz 0.0 0.0 0.0 Phosphate buffer 2 120 HNOz 0.0 0.0 0.0 Phosphate buffer 2 121 HNOz 0.0 0.0 0.0 Phosphate buffer 2 122 HNOz 0.0 0.0 0.0 Phosphate buffer 2 123 HNOz 0.0 0.0 0.0. Phosphate buffer 2 124 HNOz 0.0 0.0 0.0 Phosphate buffer 2 125 HNOz 0.0 0.0 0.0 Phosphate buffer 2 126 HNOz 0.0 0.0 0.0 Phosphate buffer 2 127 HNOz 0.3c 0.0 1.2 Phosphate buffer 2 128 HNOz 0.3c 0.0 1.2 Phosphate buffer 2 129 HNOz 0.3c 0.0 1.2 Phosphate buffer 2 130 HNOz 0.0 0.0 0.0 Phosphate buffer 2 131 HNOz 0.0 0.0 0.0 Phosphate buffer 2 132 HNOz 0.0 0.0 0.0 Phosphate buffer 2 133 HNOz 0.0 0.0 0.0 Phosphate buffer 2 Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 149 Table 6.10 (continued) Electrode Pt surface complex Aliquat Decanol solution meter# # cleaning amount (g) volume (ml) volume (ml) used 134 HNOz 0.0 0.0 0.0 Phosphate buffer 2 135 HNOz 0.3c 0.0 1.2 Phosphate buffer 2 136 HNOz 0.3c 0.0 1.2 Phosphate buffer 2 137 HNOz 0.3c 0.0 1.2 Phosphate buffer 2 138 HNOz 0.0 0.0 0.0 Phosphate buffer 2 139 HNOz 0.0 0.0 0.0 Phosphate buffer 2 140 HNOz 0.0 0.0 0.0 Phosphate buffer 2 141 HNOz 0.3c 0.0 1.2 Phosphate buffer 2 142 HNOz 0.0 0.0 0.0 Phosphate buffer 2 143 HNOz 0.0 0.0 0.0 Phosphate buffer 2 144 HNOz 0.0 0.0 0.0 Phosphate buffer 2 Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 150 6.1.11 Potentiometric Results Obtained For Conditioned Electrodes These electrodes were conditioned for 12 hours in buffer solution prior to use. In order to com-pare the results with the literature data , 0.3g of Ba-glucose complex was used in the electrodes since that was the optimum composition indicated by previous workers for the construction of the electrodes [83,84]. A factorial design was set up in order to study the effect of each individual component on the electrode response. All these electrodes were tested with a Keithly 617 electrometer (input impedance 1014ohms). Again, unconditioned electrodes were also prepared to observe their potentiometric behavior with this meter since every meter has its own characteristics. Their responses were compared with those for conditioned electrodes and the results will be explained in later sections. To see how conditioning effected the potentiometric signal, a factorial design was used using a crossed effect model. Two levels of Ba-glucose complex (0.0,0.3g), two levels of aliquat (0.0,0.8ml), and 2 levels of decanol (0.0,1.2ml) were used. Continuous voltage-time effects were measured and analyzed statistically to see how the potential changes were effected by the 3 factors. The estimated R square was 0.783466 which indicated that the data agrees with the G L M model. The estimated P R > F values were 0.001 for Ba-glucose complex amount, 0.001 for decanol volume, 0.2436 for aliquat , 0.9456 for sensor effect and finally 0.001 for interaction effects between all the components presented in the electrode. This indicates that the effect of Ba-glucose complex and decanol were statistically significant. The estimated standard error was predicted to be around 8mv which was not statistically significant. Statistical analysis confirmed the degree of significance to the responses of each component used and how their characteristics effected the potentiometric response. The Effect of Complex Concentration on Air Dried and-Conditioned Electrodes Figure 6.15 shows the responses of air dried and conditioned electrodes, made using O.Og Ba-glucose complex. It was observed that the initial and final voltages were substantially less Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 151 Q-°Electrode § 90 ^ E l e c t r o d e # 91 ^ E l e c t r o d e # 92 o 2 0 -<0 6 0 8 0 Time, (min) I 7 0 0 > 1 • 1 • 1 ' 1 ' 1 R _ 0 - 0 Electrode #8* 6 0 0 o-o Electrode #87 A - A Electrode #88 5 0 0 -4 0 0 - ^ B - & O 3 0 0 - -2 0 0 - -1 0 0 1 1 . 1 . 1 . 1 I . I . ) 2 0 < 0 6 0 8 0 1 0 0 120 W 0 Time, (min) I I Figure 6.15: Comparisons Between Potentiometric Responses of I: O.Og Ba-Glucose Complex (Conditioned), II: O.Og Ba-Glucose Complex (Conditioned), Electrode No.84, Air Dried Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 152 for the conditioned electrodes compared to air dried electrodes. It can also be observed that conditioning of the coated wire glucose sensors reduced the slope of the potentiometric "drift" and that the potential takes less time to come to steady state. Air dried and conditioned electrodes contain the same coating materials and only differ in their time of exposure to the measuring solution. Why should conditioning the electrodes have any effect on the electrode behavior? Table 6.11 and Table 6.12 show the behavior of air dried and conditioned electrodes each containing 0.3 and 0.0 grams of Ba- glucose complex respectively. As observed by their change in mv between the electrodes prepared with 0.3g Ba-glucose complex with electrodes prepared with O.Og Ba-glucose complex, one can see that the differences observed are not distinguishable. A similar observation can be made for conditioned electrodes. Figure 6.16 shows the response of conditioned electrodes using 0.3g of Ba-glucose complex compared to an air dried electrode with the same complex concentration. Again, it can be observed that the electrode potential decreases with time indicating that stable potential is not achieved until equihbrium between the ions and the ion exchange resin is reached. By comparing the steady state potentials observed for air dried electrodes using 0.3g Ba-glucose complex with electrodes containing no Ba-glucose complex, one can see that the change in potential observed for the same treatment of the same coating material is different. By comparing the steady state potentials observed for conditioned electrodes with 0.3g Ba-glucose complex and with those electrodes with no Ba-glucose complex (Table 6.11 and 6.12), one can see that the change in potential for every individual coated wire is different regardless of the complex mass. The change in starting potential and the steady state potential is also different for each condition. One can see, that there is no indication whether the change is due to the decrease in the amount of Ba-glucose complex initially dissolved in the ion exchanger or is due to some other factor. One obvious observation is that for conditioned electrodes, there is no significant change between the potentials observed for electrodes with different complex concentrations. Comparisons of figures 6.15 and 6.16 results in the same conclusions. Figure Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 153 6.17 shows the data observed from the previous literature [83]. Platinum wires were coated with different concentrations of aliquat to decanol and Ba-glucose complex. 10 electrodes were coated and it was shown that the change was due to differences in the ratio of aliquat to decanol. As indicated in the literature, these electrodes were conditioned prior to use for some time. Since the electrodes were conditioned prior to use, comparisons could be made between the literature results and the results obtained from this thesis. Figure 6.18 shows the behavior of 9 electrodes with the same ratio of aliquat to decanol concentration independent of salt concentration, which were conditioned prior to use. Platinum wires that were first treated with HNO3 and flame, were coated with the same concentration of aliquat to decanol ratio. Even though the preparation of these electrodes was identical, the initial and steady state potentials observed from these same treated coated wire sensors were different from each other. This shows the irreproducibility of the coated wire electrode potentials. The cause of this irreproducibility will be explained in a later section. With the observations gathered from figure 6.18 which indicates that the reproducibility of coated wire electrodes is very minimal, it would be difficult to state that the observed differences between potentials of tested electrodes are only due to different aliquat to decanol concentrations. This could lead to a misinterpretation of the data. From previous discussions, it is known that by placing the coated wire electrodes in different buffer solutions, the potential would again drop until an equilibrium potential is acheived. Since the electrodes in the tests reported in the literature [83-86] were moved from a low to a high concentration of glucose in a new buffer solution, the break in the curve was interpreted as being due to different concentration of glucose, whereas it has been shown here that the change of potential is independent of glucose concentration. This could also lead to a misinterpretation of the data. Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 154 Table 6.11: Stability of Air Dried Coated Wire Sensor treated with HN03 Acid and Thermally Flame Treated Electrode Complex used Starting Steady state change in # grams mv mv mv 72 0.3 675 485 190 73 0.3 685 593 92 74 0.3 665 358 307 81 0.3 663 352 311 84 0.0 622 405 " 217 85 0.0 633 383 250 86 0.0 660 353 307 116 0.0 608 439 169 117 0.0 518 378 140 118 0.0 580 372 208 Table 6.12: Stability of Conditioned Coated Wire Sensor Treated with HN03 Acid and Ther-mally Flame Treated Electrode Complex used Starting Steady state change in # grams mv mv mv 78 0.3 265 190 75 79 0.3 236 194 42 80 0.3 160 129 31 82 0.3 167 114 53 90 0.0 156 114 42 91 0.0 132 111 21 92 0.0 161 105 56 93 0.0 141 81 60 94 0.0 165 82 83 95 0.0 175 100 75 Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 155 T h e Effect of Decanol in A i r D r i e d and Condit ioned Electrodes The effect of aliquat was now investigated. It had been hypothesized that the drift in condi-tioned electrodes was caused by exchange of ions with the aliquat. Electrodes were prepared using the procedure outlined in section 3.7.2 except that the treated platinum wires were coated with decanol entrapped in a polymeric matrix. Comparisons were made between the potentio-metric responses of air dried and conditioned electrodes. Figure 6.19 shows the responses of conditioned electrodes No.99, 100, 101, 102, 103. One can see that the removal of aliquat has caused the potentiometric drift to stop in the case of conditioned electrodes. In figure 6.20 it is seen that air dried electrodes still show a decrease in potential with time. Table 6.13 shows the responses of "air dried" and "conditioned" decanol electrodes using 0.0 and 0.3g of Ba-glucose complex as starting , steady state potential and change in potential. By comparing Table 6.11 and Table 6.13 which is for coated wire glucose sensors with aliquat, one can see higher initial voltages were observed with a lower steady state potential if aliquat is present. This could be caused by either: 1. By removing the aliquat, the ion exchange was stopped between the buffer and the elec-trode. 2. The potential decrease with air dried electrodes may be due to some effect, such as oxidation on the platinum surface. In the next section the response of bare platinum wires in buffer solution will be described. It was recognized that producing reproducible platinum surfaces when covered with a coating material and exposed to the measuring solution is difficult. It was necessary to determine how bare platinum responds to such environments. Conclusions of the Results Obtained for Condi t ioned Electrodes From the comparison of the data reported in the literature[83,84] and found in this work, it can be concluded that the Ba-glucose electrode does not respond to glucose concentration. In an Figure 6.16: Comparisons Between Potentiometric Responses of I: 0.3g Ba-Glucose Complex (Conditioned), II: 0.3g Ba-Glucose Complex (Conditioned), Electrode No.81, Air Dried Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 157 COMBINATIONS USEO FOR THE PREPARATION W E COATED HIRE SENSORS FOR THE SECOND SET TABLE S (Concluded) STEADY STATE VOLTAGE CEKERATEO 6T THE SENSORS Glucose S a l t <9«) A1tquat :decano l i ( V/Y ) A l i q u a t 336S 61ucose:Poly«er Ra t i o M 20 1:3 1.0 40 1:3 1.0 60 1:3 1.0 SO 1:3 1.0 100 1:3 z.o 20 1:3 2.0 40 1:3 Z.O 60 1:3 2.0 80 1:3 2.0 100 1:3 E(«V) USEO FOR SEO0N0 SET WITH INCREASED GLUCOSE LOAD Sensor I S11120 S11140 SI 1160 . S11130 SI 11100 Cone (no t ) 80 25.2 .39 31 138 107 100 2 3 . 7 33.7 28 .5 134.3 108.2 120 22.2 28.4 26 131 103.4 140 20 .7 23.9 23.S 127.8 101.6 160 19.2 17.8 21 124 9 9 . 8 Sensor f S11220 S11240 SI 1260 SI 1280 S112100 Cone (ng 1) 83 176 102 138.2 99.6 175.4 100 157 99 132.8 99 .4 172.4 120 138 96 127.4 99.2 169.4 140 119 93 122.0 99 .0 166.4 160 100 90 116.6 9 8 . 8 163.4 • • A l i q u a t : dec.no I 801 (V/V) • • A l i q u a t i d e c a n o l 1001 (V/V) A.—i: A l t q u a t : d e c a n o l 401 (V/V) O—O A U q u a t : d e c a n o l 601 (V/V) © — O A l i q u a t : d e c a n o l 201 (V/V) II 120 140 » cone (ag I) F l o w * 1 0 . Sensors ( U l ) v o l t a j e-concent ra t ion c u r v e s . Il l 200 ISO E(«V) • 100 IV a A H q u a t i d e c a n o l 1001 (V/V) .« A l 1qua t : de c ano l 801 (V/V) .A A l i q u a t : d e c a n o l 601 (V/V) A l iqu t t- . aecano l 401 (V/V) A U q u a f . d e c a n o l 201 (V/V) I 80 120 160 -» cone («g 1) F igure 1 1 . Sensors (112) w t t a o s - c o n c e n t r a t i o n c u r v e s . 200 Figure 6.17: Effect of Voltage-Cone. Curves on Electrode Response [83] I ' I 1 . I . 1 . 1 . 1 0 20 i 0 6 0 SO 100 120 " 0 160 ISO T i m e , ( m i n ) °-° Electrode #93 0 - 0 Electrode #94 M Electrode #95 1 1 • 1 . 1 i 1 1 1 1 J , i • I I I 2 3 * 5 6 7 6 9 T i m e , (hrs) Figure 6.18: Comparisons Between Potentiometric Responses of Aliquat to Decanol Ratio With-out Ba-Glucose Complex. Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 159 > 100 6 0 . 80 Time, (min) o-aElectrode # 99 o-oElectrode #100 Electrode #101 120 CO a $ wo 1 i I i | i i 1 j 1 j > 1 ! p - S B B -J i 1_ I I - 6 B 3 Q e-a Electrode #101 o-o Electrode #102 tr-A Electrode #103 20 40 60 8 0 100 Time, (min) 120 W0 160 180 Figure 6.19: Comparisons Between Potentiometric Responses of 1,11: Decanol Coating (Condi-tioned) Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 160 SOO I L o > 5 6 Time, (hrs) 200 o o > 1S0 EH-oElectrode # 99 o-oElectrode #100 ^Elec t rode #101 I I 60 80 Time, (min) 120 Figure 6.20: Comparisons Between Potentiometric Responses of I: Decanol Coating (Air Dried), II: Decanol Coating (Conditioned) Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 161 Table 6.13: Stability of Air dried and Conditioned Coated Wire Sensor Treated with HNO3 Acid and Thermally Flame Treated (Decanol Electrodes) Electrode Amount Salt Starting Steady state change in # grams mv mv mv 96 0.0a 503 397 106 97 0.0a 500 426 74 98 0.0a 556 432 124 99 0.0c 138 124 14 100 0.0c 134 125 9 101 0.0c 204 192 12 102 0.0c 151 133 18 120 0.0c 178 172 6 127 0.3a 502 474 28 128 0.3a 510 504 6 129 0.3a 478 493 15 135 0.3c 163 176 13 136 0.3c 158 160 2 137 0.3c 170 200 30 Term a indicates "Air Dried", Term c indicates "Conditioned". attempt to explain the response reported in the literature and in this study, it was found that: 1. By conditioning the prepared electrodes, the potential drift decreased. 2. Even with a better cleaning of the platinum wires used in the electrodes, reproducibility of the electrode response was still a problem. 3. By increasing the Ba-glucose complex concentration, the initial and final steady state potential was not significantly changed. 4. The effect of decanol addition to the platinum wires was not reproducible. 5. When the aliquat (quaternary ammonium ion) was removed, the transient behavior of the electrode stopped. 6. Surface oxides on the platinum wire could be a possible cause of irreproducibility of the electrodes. Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 162 The above tradings are just highlights of what could be observed from the potentiometric measurments. The final mechanism of the electrode behavior will be explained in detailed later in the thesis. 6.1.12 Effect of the Platinum Surface on the Electrode Response In this part of the research platinum wires alone were tested in the same measuring solution. After the platinum and the reference electrode had reached a stable potential, decanol was added in different portions to observe the changes in potential. To observe the reproducibility of the platinum electrode response, 3 to 4 treated wires were immersed and tested at the same time. Before use, the platinum wires were placed in HNOz and were thermally treated in the reduction part of a gas flame for approximately 2 minutes. The potentiometric behavior of treated platinum wires was observed by immersing the treated platinum wires in phosphate buffer in conjunction with a reference electrode (Sat KCl). The responses of the bare platinum wires were observed for 10 to 15 rninutes since steady state potential was reached very quickly, compared to the platinum wires coated with organic materials as is shown in figure 6.21. Table 6.14 shows the potentiometric behavior of treated platinum wires. The differences between the responses shown by these platinum wires could be due to differences in their surface conditions. Surface studies are described in the next chapter. Could the coating material in contact with the surface of the bare platinum have caused the electrodes to respond so differently from each other? As was observed and reported in the literature [112,113] ,the potentiometric response can be influenced by the type of pretreatment that the platinum surface was given. If the platinum surface treatment had no effect on the surface potential, then the potential values should have been the same for all treatments. As in-dicated in ref [112,113], platinum electrodes which have undergone extended anodic or cathodic polarization exhibit potentiometric properties that distinguish them from untreated platinum surfaces. It is now necessary to determine if platinum electrodes which have undergone usage Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 163 with organic coating materials could exhibit potentiometric properties that could distinguish them from untreated or retreated platinum surfaces. From table 6.14, one can observe that the steady state potential of bare platinum wires cleaned by the same method show differences in their potentiometric behavior. By looking at the table, one can see the starting mv, final mv and Delta V are very different for every platinum surface. The effect of individual surfaces were also analyzed for intravariability of the sensor. 22 electrodes which were treated with HNO3 and flame treated were potentiometrically measured. A crossed model was used to compare the least mean squares of the potentiometric mea-surements of individual electrodes with each other. Six sensors were examined at six different time intervals since continuous voltage readings were recorded. The R square was 0.906 which indicated that the model agrees with the experimental data. The statistical observation showed that the behavior of individual surfaces are significantly different. Results showed F values of 43.74 for time and 4.47 for sensors . The PR>F values were observed to be less than 0.005 which indicates that the effect is significant. This confirms that the irreproducible response of bare platinum electrodes is statistically significant. More detail results are shown in appendix B. The next section will test the effect of an oxygen containing organic material, decanol, on the platinum surfaces. Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 164 « 350 cn • O '300 _ , , , 1 , — o-o Electrode #104 0 - 0 Electrode #105 -Electrode #106 -§ i 1 1 1 1 20 40 Time, (min) Figure 6.21: Comparisons Between Potentiometric Responses of Bare Platinum (Air dried) Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 165 Table 6.14: Potentiometric Behavior of Treated Platinum Wires Electrode Starting Steady state Delta V # mv mv mv 110 319 202 117 111 217 209 8 112 203 237 34 113 202 167 35 114 186 104 82 115 261 132 129 119 298 140 158 120 190 134 56 121 196 116 80 122 178 144 34 123 147 158 11 130 223 236 13 131 239 243 4 132 223 232 9 133 198 231 33 134 235 191 44 138 390 199 191 139 380 134 246 140 247 134 113 142 390 123 267 143 380 158 222 144 418 195 223 6.1.13 Effect of the Alcohol Adsorption to the Platinum Surface This part of the study was done to see if decanol can change the platinum electrode signal. Since the original electrode coating contained decanol, it was very important to see the effect that it had on the electrode surface. Different volumes {pi) of decanol were added to the measuring cell to see if the electrode potential of platinum wire drops at the same rate after exposure of the surface to alcohol. Decanol is insoluble in H20, or buffer solutions, however trace amounts will be dissolved and there will be surface effects. Treated bare platinum electrodes were potentiometrically measured until a steady state potential was acheived, then a trace of decanol was added. As shown in figure 6.22, decanol addition did change the potentiometric Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 166 response. Table 6.15 shows the effect of the addition of decanol to treated platinum electrodes. By comparing the initial and final potentials of these electrodes, one can see that the steady state surface potentials were not the same and that the potentiometric change after addition of decanol to the measuring solution were not consistent with different platinum surfaces. This could be due to some surface effect. Since the bare platinum wires showed different responses to the addition of decanol, it was proposed to find if the same platinum wires could give reproducible results. Electrode 138 as shown in figure 6.231 was first stabilized in phosphate buffer solution. Decanol was then added to the measuring solution. It was then placed in a new buffer solution and the signal still decreased after the addition of decanol. It was found that every time the surface is exposed to the measuring solution, the voltage will drop until equilibrium potential is again reached. Electrode 140 as shown in figure 6.23II was stabilized in phosphate solution too. Again, decanol was added and the potential decreased from 134 to 72mv. After steady state potential was reached, it was placed in a new buffer solution, decanol was added and signal began to decrease again. It was possible that the cause of the decrease in potential was due to an interface effect at the air buffer interface. The effect of decanol addition to different platinum wires were analyzed using a crossed model. The effects of 22 sensors were analyzed according to their steady state potential and potential observed after addition of decanol. The estimated R square were measured to be 0.6366 which indicates that the model agrees with experimental data. The P R > F values were measured to be 0.5875 for sensor and 0.0004 for time which indicated that the addition of decanol had effected the potentiometric response significantly. The results of the t test also confirmed the significancy. The standard error predicted between the 22 electrodes were averaged to be 15mv which was not statistically significant. Finally, the contour plot showed the consistency of the data. Statistical results confirmed the significancy of addition of n-decyl alcohol to the response of each individual electrode. It was also observed that the change of potential due to the addition of the alcohol was not significantly different for different electrodes. The detail results are shown in appendix B. Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 167 ~i 1 i 1 1 1 1 ' 1 1 r too I 1 —' 1 Q—Q Electrode #132 o-o Electrode #13C M Electrode #134 20 40 eo too 120 Time, (min) o > 150 Q-Q Electrode #132 o-o Electrode #133 Electrode #134 U 0 ISO Time, (min) I I Figure 6.22: Comparisons Between Potentiometric Responses of I: Bare Platinum(Conditioned), Decanol Addition Figure 6.23: Comparisons Between Potentiometric Responses of Addition of n-Decyl Alcohol to I: Bare Platinum (New Buffer), II: Bare Platinum (New Buffer) Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 169 Table 6.15: Effect of Addition of Decanol to Treated Bare Platinums Electrode Starting Steady state After addition Change in # mv mv mv mv 130 223 231 168 63 131 239 240 196 44 132 223 232 132 100 133 198 231 146 85 134 235 191 110 81 138 390 199 176 23 140 247 134 72 62 142 390 123 83 40 143 380 157 106 51 144 418 195 116 79 Glass Platinum Wire Electrodes 1. Glass Electrodes are identified as electrode No.145-156 Since the platinum electrode potentials were not reproducible, it was thought that this could be due to an interfacial effect. The platinum air-buffer interface was covered by sealing the plat-inum inside a glass electrode and exposing only the end of the platinum wire to the measuring solution. Figure 6.24 and figure 6.25 shows the responses of glass electrodes to decanol addition. A smaller potential drop and a shorter stabilizing time are observed. Since the potential still decreased, and since the smaller change could be ascribed to a much smaller area of platinum surface, it is proposed that the air-liquid interface was not the cause of potential drop with decanol addition. It was concluded that the reproducibility of the surfaces was not due to the interfacial effect. It is now necessary to examine the platinum surfaces to see if the surfaces that could cause the potentiometric responses to be different. Could there be a surface quality of platinum wires which would effect the potential when they are coated with a aliquat/decanol polymer? Could decanol have any effect on platinum wires which could influence the irreproducibihty of the electrodes to increase? How can the surfaces be characterized ? Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 170 These questions and hypotheses led to the study of platinum wires electrochemically and by XPS (X-ray PhotoElectron Spectroscopy). Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 171 120 160 Time, (min) 120 160 Time, (min) 2*0 260 Figure 6.24: Comparisons Between Potentiometric Responses of Addition of n-Decyl Alcohol to I: Glass Electrodes (New Buffer), II: Glass electrodes (Air dried) Chapter 6. NEW APPROACH TO THE POTENTIOMETRIC OBSERVATIONS 172 Figure 6.25: Comparisons Between Potentiometric Responses of Addition of n-Decyl Alcohol to I: Glass Electrode (New Buffer), II: Glass Electrode (New Buffer) Chapter 7 X - R A Y P H O T O E L E C T R O N S P E C T R O S C O P Y To substantiate the proposition that the quality of the platinum wire used in the electrodes was the cause of the irreproducibility of the recorded potentials , the platinum surfaces and its electrochemical response were studied in greater detail. 7.1 Electrochemical Studies of Pla t inum Surfaces The principle requirements for a good electrode material for continuous concentration detec-tion are long-term stability, low residual current, and good electrocatalytic properties. In general, the active electrode area is quite different from the geometric area, a difference caused by variability in the surface microstructure, which results from physical roughness, adsorbed impurities, and surface oxides, among other factors. Noble metal electrodes [115,116] of platinum or gold have been widely employed for fun-damental electrochemical studies but have found limited usage in practical analytical applica-tions. The limitations of these electrode materials arise from their susceptibility to passivation through surface oxide formation and the tendency toward electrode fouling and poisoning [117-119]. These metals are notorious for problems associated with surface conditioning. Although considerable work has been done in this regard, their use is still very much of an art and it is difficult to achieve any degree of reproducibility in the results [116]. The problem of preparing and mamtaining a clean surface is one that persists in the study of all surface chemical and surface electrochemical reactions. Fundamental studies have been made on the characterization of the surface of platinum electrodes which will be briefly explained. 173 Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 174 Scortichini and Reuly [120-122] have described the surface characterization of platinum electrodes using underpotential deposition of hydrogen and copper. The surface of a platinum electrode pretreated by flame aimealing and quenching in sulfuric acid is shown to contain a high concentration of structural defects such as vacancies and self-adsorbed platinum atoms [123]. Adsorbed hydrogen is more strongly bound at the defects than on a uniform platinum surface. Pre-purification of the electrolyte was found to have no significant influence on the metal surface area available for hydrogen adsorption determined immediately after an electrochemical cleaning step. The effect of organic adsorbates on platinum surfaces have been classified, by Wieckowski [124]; into three groups: 1. The surface complexing processes with the participation of the adsorbate 7r-electron and the hybrid d-orbitals of the metal. 2. The adsorption in the second adsorbed-layer of the platinum interfacial region. 3. The electrochemical reactions of the organic molecules with the water molecules chemisorbed onto the platinum electrode. The oxidation of aliphatic alcohols on platinum as noted by Kokkinidis and Jannakovdakis [125], is markedly catalysed by foreign metal adsorbed-atoms deposited in the under potential range. In the case of methanol, the catalytic effect was more pronounced in basic media. In acidic media the catalytic activity was dependent on the length of the carbon chain and the number of hydroxyl groups. On bare platinum the formation and strong adsorption of organic intermediates resulted in the blocking of the surface active sites. This enhancement of the oxidation process by underpotential submonolayers has been interpreted in terms of the prevention of electrode poisoning by products. The adsorption of acetone on a platinum electrode from aqueous acidic solutions was in-vestigated by a radio tracer technique and cyclic voltammetry by Wieckowski et.al. [126]. The Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 175 formation of a 7r-electron complex between platinum and acetone was followed by surface poly-merization, the length of the chain being dependent on the bulk concentration of the acetone. The adsorption of glucose on a platinum electrode in 0.5M H2S04 was observed by Nikolaeva [127] et.al., to exhibit a maximum at 0.2V. At more anodic potentials, adsorption decreased due to the oxidation of the chemisorbed particles. A voltammetric study of oxygen chemisorption on platinized platinum electrodes in acid so-lutions was carried out by Druz and Novikova [128]. They observed that both oxygen chemisorp-tion and electrode surface oxidation occurred simultaneously irrespective of the electrode pre-treatment. Specific adsorption of oxygen decreased with increasing temperature but the amount of charge required for the reduction of surface oxides was temperature independent [123]. A voltammetric study of hydrogen adsorption-desorption on platinum surfaces was made by Clavilier and co-workers [129] to find the crystalline surface structure of platinum at the atomic level. The author concluded that a change in crystalline surface structure was the consequence of activation by repeated oxidation and reduction of the oriented platinum surface [129,130]. Oxygen electrodes which were prepared by bubbling oxygen gas over a platinum wire im-mersed in acid or alkaline electrolyte, give a potential of about 0.8V (vs Pt /H 2 in the same solution)[131-132]. In the published literature [133,134-139], there is considerable evidence that oxygen may be dissolved in bulk platinum metal. The work function of a platinum electrode was found by Kalish and Burshtein [137] to increase when oxygen was adsorbed on the plat-inum surface due to the presence of the negative dipoles of the adsorbed oxygen. After about 11 hours, they noted that the work function decreased, from which they concluded that oxygen must have diffused into the platinum metal. In confirmatory tests, they permitted oxygen to be adsorbed anodically on platinum at various temperatures and then determined the amount of oxygen adsorbed on the metal surface from constant current cathodic stripping curves obtained in a stirred N2 solution [133]. It was mentioned that only a monolayer of oxygen exists on the platinum surface with the remainder, which can diffuse to uncovered surface sites, dissolved in the metal [140]. Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 176 7.2 X- ray PhotoElec t ron Spectroscopy There are numerous reports of oxidative and reductive reactions taking place at crystalline and polycrystalline platinum surfaces. A variety of analytical techniques have been utilized in these studies; Coulometry [141], Cyclic voltammetry [142-146], elhpsometry [146,147], and X-ray photoelectron spectroscopy [147,152][XPS]. The electrochemical and elhpsometry experiments gave direct measurements of the oxide layer thickness, however, identification of the chemical species on the platinum surface can only be inferred from the data. XPS is an analytical technique capable of directly evaluating the chemical components on an electrode surface [147-152]. XPS has also been used for monitoring the appearence or disappearence of foreign materials at electrode surfaces due to "cleaning" or pre-conditioning procedures [147,150-151]. Studies on the surface properties of platinum electrodes as a result of thermal pretreatment or electrochemical treatment have been made by Proctor and co-workers [152]. In their study they emphasized the usefulness of XPS to probe such surfaces qualitatively while at the same time suggesting caution where quantitative analyses axe required [152]. The purpose of the present work with XPS (X-ray photoelectron spectroscopy) was to iden-tify the type of oxidation taking place, the type of contamination, and the amount of oxidation which the platinum surface is exposed to. The potential of the same platinum electrodes was also measured to compare the behavior of individual platinums with each other and with their level of oxidation. 7.3 Introduction to X P S X-ray photoelectron spectroscopy (XPS) is used for surface analysis in a wide variety of areas such as electrochemistry, corrosion studies, microelectronic materials processing and the analysis of biological materials. The advantages of the technique are that it is nondestructive, it often requires no special Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 177 sample preparation, it allows for the direct probing of core level and valence electrons, and it can differentiate between different chemical states. It allows the detection of adsorbates to depths down to a few atomic layers and can detect all the elements except for H and He. The sensitivity of the instrument is on the order of 0.3% [153]. The photoelectric effect was first observed by Hertz in 1887. He found that the interaction of electromagnetic radiation of sufncent frequency with the surface of a metal produced a photoelectric current. The photoelectric effect is described by the following relationship. EK = hv- Eb (7.1) where: hv = Energy of the ionizing photon. h= Planks constant. v = Frequency of radiation. Eb= Binding energy. EK = Kinetic energy of the ejected electron. The sample is in some initial state prior to the XPS experiment. It is then bombarded by a beam of photons which perturb that state, causing photoelectron emission and leaving the atom in an unstable final state which must relax back to the ground state. For example, a Is electron vacancy results in a 2s or 2p electron falling into the Is shell, giving off energy in the process. The photoemission event is extremely rapid. In the range of atomic orbitals commonly examined by XPS, the dominant relaxation mode tends to be Auger emision, especially for polymers. The Auger electron has a discrete kinetic energy and is detected by the analyzer in exactly the same manner as a photoelectron. The photoelectric and Auger processes are depicted in figure 7.1 [153,154]. The p,d,f levels become split upon ionization, leading to vacancies in the Pi/2, P3/2, d 3 / 2 , d5/2, fs/2 and f7/2 levels in the ratio 1:2 for P levels, 2:3 for d levels, and 3:4 for f levels. Table 7.1 is useful for remembering the atomic orbital nomenclature. The spectral lines or bands observed in x-ray photoelectron spectra are identified in terms of the quantum number,n, with values of 1,2,3,4,5 or 6 and the angular momentum quantum number,!, with Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 178 Figure 7.1: Diagram of (I) Photoelectric Process, (II) Auger Process Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY values of 0.1,2,3, commonly called s,p,d or f respectively [155]. Table 7.1: Atomic Orbital Nomenclature [155] Atomic Orbital Nomenclature Principal quantum number: Angular momentum quantum number (/): Subscript: Roman numerals: Example: 1,2,3,4,5 K, L, M, N, O 0. 1.2.3 s.P.dJ /+ 1/2 or 7- 1/2 1, II, III, IV, V, VI, VII 5v\pi'\pV\d3'2,d"2,f*'\f'2 4d 5 / 2 •*-principal quantum number orbital type 0 = 2) spin-orbit splitting (/+ 1/2) Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 180 The field of X-ray photoelectron spectroscopy (XPS) was developed by Kai Siegbahn's group during the period from 1955-1970 [153,156]. In 1970 XPS became a common surface analysis technique for studying well defined surfaces when ultra-high vacuum systems (pressure = 1 0 - 9 torr) were used. The procedure is as follows: An atom is ionized by an X-ray photon of energy 1254 eV and an electron, with a binding energy of Eb is ejected. The electron then travels through a potential $, which arises from the spectrometer, and is detected with kinetic energy Ek, Ek = hv- E b - $ (7.2) Because each element exhibits a unique set of orbitals, a spectrum of electron intensity as a function of kinetic energy will identify the elements interacting with the X-rays. Figure 7.21 shows the spectrum of a clean platinum surface [157]. Figure 7.211 shows a typical x-ray photoelectron spectrum, obtained from platinum, which displays peaks corresponding to the discrete binding energies of the electrons of different elements on the surface. By comparing the figures, it can be observed how contarninents can effect the surface. A n atomic peak is identified according to the energy level and the total angular momentum, j , of the photoelectron. The intensities of a transition, which is the area underneath the corresponding peak, for an atomically clean sample is measured by XPS as [153,157]: finfty h = fcrkVeAD Pk(z)exv(-z/A)dZ (7.3) Jo Where: f= X-ray flux. tTfc = Photoelectric cross section. $ = Angular efficiency factor. e= Efficiency of photoelectric process. A = Area from which photoelectrons are detected. D= Detection efficiency for emitted electrons. Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 181 Mg K a I P M l J Figure 7.2: I: Spectrum of a Clean Platinum Surface [157], II: Typical X-ray PhotoElectron Spectrum of a Platinum Electrode With Surface Contamination Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 182 A= Effective escape depth for transition. pk(z) = Atomic concentration of element K. Thus, by measuring the intensities of the characteristic peaks originating from each element in an XPS study, it is possible to obtain relative quantitative analyses. Practically all XPS instruments collect data digitally and are fully computer interfaced. Analysis of the spectra is concerned with identifying the shape of signal peaks and methods of describing them. It was also concerned with the overlap between the peaks and how to separate the peaks, if not by improving instrumental resolution, then by appropriate mathematical processing of the data. This is important because the chemical shifts are not very large. A peak may consist of a number of subpeaks which can be separated and resolved mathematically. There are certain facts that can influence the results [158,159]. The overall XPS results are given in terms of binding energies and relative intensities. Normally, a wide scan is used to identify the elements and the high resolution scan (narrow scan) is used to obtain peak positions, chemical states, and peak areas for quantification. XPS provides a total elemental analysis, except for hydrogen and helium, of the top 10 — 250° A (depending on the sample and instrumental conditions) of any solid surface which is vacuum stable or can be made vacuum stable by cooling. Inorder to remove any degree of contamination from the top layer of a metal surface which is physically adsorbed, argon ions were sputtered into the chamber for 5 minutes. After removing the contaminants, it is easier to observe the platinum surface for any degree of oxidation. Electrode No.14 was argon sputtered for 5 min in the vacuum chamber. The results before and after sputtering are shown later. The main objective of this part of the experiment was to provide information about chemical state of the platinum surface and to determine (if possible) what type and the amount of oxidation the surface had been exposed to. Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 183 7.4 Instrumentation The analysis was performed in the new Leybold Max (micro Area XPS) 200 XPS Surface Analysis System in the U . B . C . Chemistry Department. The basic experimental apparatus used is shown in Figure 7.3 and Figure 7.4. The chambers are constructed of stainless steel and vacuum seals between flanges are made using copper gaskets compressed against the inner walls of the flanges. Pressures are reduced from 1 atmosphere to 1 0 - 3 mbar by a rotary pump. Turbomolecular pumps reduce the pressure to 10~8 mbar. A differential ion pump is used to evacuate the x-ray source. Typical basic pressures are 4*10 - 1 0 mbar in the analysis chamber and 2*10~8 mbar in the transfer chamber. The spectrometer is equipped with a dual anode x-ray source that allows excitation to be by Mg K a or A l K a radiation. The anode, Mg or A l , is at a high positive potential and is bombarded by electrons sputtered from a hot filament which excites inner electrons in the target atoms which then may be ejected from the atoms. Electrons from higher energy levels in the atoms may then fall into the inner orbital vacancies and X-rays are emitted from these transitions. For the dual anode source, Mg or A l K a radiation may be used by a simple switching of filaments. A thin aluminum window is placed between the source and the analysis chamber to act as a barrier to electrons and to prevent contamination. The central part of the spectrometer is an electron energy analyzer (EA-200), which, for this instrument, is of the concentric hemispherical type. Photoelectrons with various kinetic energies are focussed and retarded by a lens system to an energy referred to as the pass energy. A potential is applied between two concentric hemisphrical shells so that only electrons with a particular pass energy can travel on circular orbits. The electrons are then counted by a multichannel detector. The spectrometer is interfaced to a HP 100 based micro processor using data system DS 100 software. The computer controls the selection of instrumental parameters, data acquisition and data processing. The x-ray photoelectron spectrometer was operated using the Mg K a -Figure 7.3: The Experimental Apparatus of XPS Surface Analysis System Figure 7.4: Data Acquisition and Data Processing Equipment Which Is Connected With XPS Spectrometer Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 186 radiation source (1253.6 eV), at a power of 15.0 kV. Emission current was held at 20mA, the pressure was reduced from atmospheric to below 2*10 _ 8 mbar going to the transfer chamber then to 1.22*10_9 mbar in the analytical chamber and was set at a basic pressure when the experiment was running at 1*10 - 8 mbar. Data were collected and stored in the data aquisition unit model 865986, then digitized and finally stored on a computer model Sony 2436EA600. 7.5 Experimental Section 7.5.1 Reagents Used • Commercial grade platinum wires 0.25 mm in diameter purchased from Good Fellow, Cambridge, England , composition (ppm) Ag 5, A u 10, Ca 1, Cu 10, Fe 10, B 4, Mg 1, Na 1, Pb 1, Pd 50, Rh 10, Si 2,. • Commercial grade platinum wires 0.25 mm in diameter purchased from Sargent-Welch Scientific Co., Composition: Not available. • Nitric acid(70%) from Fischer Scientific. • Poly Vinyl Chloride(140*31) from B.F.Goodrich. • Tetrahydro furan from B D H chemicals. • n-Decyl Alcohol from B D H chemicals. • Aliquat(Quaternary ammonium salt) from Aldrich Co. • Distilled water available at the Chemical Engineering Dept U B C . 7.5.2 Apparatus The electrochemical experiments were conducted in a two electrode cell. The cell contained a saturated KCl reference electrode and a platinum wire as the working electrode. These platinum wires were either cleaned or coated with a ion exchanger/polymer film. Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 187 A high impedance Keithly 616 electrometer was used for the potentiometric measurement and the data were recorded continuously. The cell operated at room temperature, approximately 27°C. The electrodes were immersed in phosphate buffer of pH 7.4 and data were collected continuously. Electrodes were stored in clean sampling jars for XPS spectra analysis. Each sample was mounted onto a sample holder made of aluminum and then introduced into the XPS transfer chamber, which was immediately evacuated to 10~8mbar. Each sample holder, in turn, could then be transferred on to the manipulator, in the analysis chamber. The manipulator facilitates accurate positioning of samples and allows translation in up to 100 different positions. For coated wire electrodes, the coating was removed gently by pulling the coating away to expose the surface. 7.6 Preparation A n d Modification Of Platinum Surfaces Cleaning Procedure (Coated Electrodes): After the coating had been removed, the platinum wires were washed in concentrated HNOz for a few hours. After conditioning in acid, the wires were heated in the nonoxidizing portion of a natural gas-air flame for approximately 2 minutes. The timing of the flame was chosen according to reference [113]. The platinum wires were then dipped individually in a mixture of liquid ion exchanger/polymer until well coated, being careful to completely cover the metal tips. These were allowed to dry over night in a dead air hood. The shape of these coated wire electrodes are shown in Figure 7.5. The liquid ion exchanger/polymer coating contained 0.3 grams of the glucose complex added to 1.2ml of n-decyl alcohol and 0.8 ml of the quaternary ammonium salt. The coating liquids were mixed by a magnetic stirrer over night. 3 grams of polyvinyl chloride were dissolved in 12 ml of tetrahydro furan, then an equal volume of this was added to the salt mixture and stirred for approximately four hours, prior to dipping. Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 188 Figure 7.5: The Shape of Coated Wire Glucose Sensor Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 189 7.7 History of Platinum Wires The effect of surface treatment was still unknown, therefore, a group of new and reused plat-inum electrodes (previous coated electrodes) were given different pretreatments prior to their measurements. Table 7.2 gives a brief description of the electrodes that were used for the XPS spectral studies. The number indicated for file number is the spectra number in wide and nar-row scan which might appear in some of the comparisions of the spectra shown in later sections. A more detailed explanation is given below. • Electrode No. l , reused platinum wire, treated with HNOz for 12 hours and flamed for 2 rninutes, decanol was added, potentiometric response was observed, XPS was analyzed. • Electrode No.2, reused platinum wire, treated with HNOz for 12 hours and flamed for 5 seconds, decanol was added, potentiometric response was observed, XPS was analyzed. • Electrode No.3, reused platinum wire, treated with HNOz for 12 hours and flamed for 2 minutes, potentiometric response was observed, XPS was analyzed. • Electrode No.4, reused platinum wire, treated with HNOz for 12 hours and flamed for 5 seconds, potentiometric response was observed, XPS was analyzed. • Electrode No.5, As received platinum wire(from Good Fellow), XPS was analyzed. • Electrode No.6, As received platinum wire, treated with HNOz for 12 hours and flamed for 2 minutes, XPS was analyzed. • Electrode No.7, As received platinum wire, treated with HNOz for 12 hours and flamed for 2 minutes, potentiometric response was observed, XPS was analyzed. • Electrode No.14, reused platinum wire, treated with HNOz for 12 hours and flamed for 2 minutes, coated with coating liquids, removed the coating, XPS was analyzed, potentiometric response was observed. • Electrode No.19, reused platinum wire, XPS was analyzed. Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 190 • Electrode No.25, reused platinum wire, treated with HNO3 for 12 hours and flamed for 2 minutes, coated with coating liquids, removed the coating, XPS was analyzed, potentiometric response was observed. The numbering of these electrodes are separate from the previous sections. Flame or thermally treated platinum surfaces were prepared by heating the metal surface in the non-oxidizing portion of the flame. Table 7.2: History of the Platinum Electrodes Electrode Description File (Narrow) File (Wide) 1 2min flame, HNOz, Decanol exposure(reused) 215 214 2 5sec flame, HNOz, Decanol exposure(reused) 226 225 3 2min flame(reused), HNOz, 210 209 4 5sec flame(reused), HNOz, 228 227 5 As recieved(no treatment), 263 262 6 As recieved, 2min flame,HNOz, 268 267 7 Treated,lst Potential measur., 270 269 14 Coated wire electrode(reused) 224 223 19 Untreated platinum wire(reused) 230 229 25 Coated wire electrode(reused) 234 233 4(rec) 2min flame(reused), HNOz, 236 235 Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 191 7.8 Mathematical Background and Procedure As described earlier, an atomic peak is identified according to the energy level and total angular momentum, J, of the photoelectron. The intensities of a transition, which is the area underneath the corresponding peak, for an atomically clean sample is measured by XPS. The corresponding plots are given in terms of intensity vs. binding or kinetic energies. A peak may consists of a number of subpeaks which can be separated and resolved mathematically. This needs to be done in order to identify the type of contaminants which covers the surface. Contaminants could involve any type of oxidation levels which could be chemisorbed into the platinum lattice. The procedure followed to obtain results included a curve fitting of the experimental data to give the best fit to theory. In order to do curve fitting for this study, an original peak distribution of pure platinum surface was taken as a reference peak. Figure 7.6 indicates the reference data obtained from literature indicating the binding energies of 4/7/2 for the original platinum as compared to contaminated platinum surfaces. As indicated in figure 7.71, the corresponding plot shows a clean platinum surface. It indicates the original binding energies of 4 / 7 / 2 at 70.9ev and 4 / 5 / 2 at 75ev. The difference between the two peaks is identified to be 3.35ev. The ratio between the heights of the first peak compared to the second peak is observed to be 4 to 3. By setting a window around the spectra and guessing a position where the original peak can start, one determines a starting position. If a platinum surface is exposed to contaminants, the observed peak will cover a certain area of the platinum which has to be identified mathematically. The difference between the original platinum and PtO, and Pt{0H)2 were given, indicating a maximum chemical shift of about 2.7 ev from the metal peak which corresponds to the oxide I species(found by Hammond and Winograd and Dickinson et.al); which is thought to be an intermediate oxide of +2 oxidation state and another shift of about 1.2 ev from the metal peak which corresponds to another Pt+2 species such as Pt(0H)2 [155]. Figure 7.71 shows a typical spectrum of a clean platinum surface. Figure 7.711 shows a typical spectrum of the experimental and calculated peaks for electrode No.3. The sum of the analytical peaks (individual curves) were added to the original background to yield the fitted spectrum, Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 192 Platinum, Pt Number 78 COMPOUND 4(, A BINDING ENERGY, eV Pt 1 0 PttPPhJ, R Pt(PPh,), 1 R PtfPBuACI, . i . CAB Pt(PPh,),CI, CAB PKPPhJ jCI , i R P t f PPh^Me , R PtfPPh.l.Ph, R Pt(PPhJ,I, R Pt(PPh,),HCI R Pt(OH), H W Pt(PPh,),C,H< CAB P t fPPh j l jCH . M M R PtfPPh.l .C.F, M M R PUPPh J .C tCN) , M M R PtfPPhJ.O, . $ '• si * i R PttPEWjCI, & R K ,RC I 4 1v® R K,PtCI. * " E P C PtCI, 1 E P C K,Pt(CN). I R PtO 1 K W D PtO « ••:-v- E P C RfSOJ.-H.O 'is 1 H W PtO, 1 K W D PtO, 1 E P C PtCI, I E P C K.PtCI, I LB K ,RC I , v E P C K .RC I , 1 R Pt(PMe,Ph),CI. r LB PtlPEt.hCI. 1 LB Pt(PEt3),CI. i R KJRICNJ.CI.-SHJO i C L Pt(EtNHd,CI,-4H,0 r C L Figure 7.6: Table of a Clean Platinum Surface With Comparison to Contaminated Surfaces of Platinum [157] Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 193 Figure 7.7: Comparison of Clean Surface with Experimental and Calculated Peak for Electrode No.3 Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 194 which could be compared to the original spectrum. This analysis was carried out on the most sensitive layer of platinum, Pt4f, the narrow scan surface of the platinum metal. The procedure followed is described in BIGFIT [160]. The BIGFIT function in the XPS computer is used to perform the background subtraction on up to sixteen lines or groups of lines and to integrate or fit the resulting peaks [160]. If only background subtraction or background subtraction with integration is performed, the result and spectrum will consist of background subtracted lines or group of lines. It can be processed in the same way as the original. Should background subtraction be performed with fitting, the result and spectrum will consist of the adapted analytical function added to the original background so that the result can be directly compared with the original spectrum. Two types of background functions are available: 1. First degree polynomial. 2. Non-linear model function, proportional to the integral of the unscattered or elastically scattered particles. For background subtraction the user sets one or two markers (energy values) on the low energy as well as the high energy side of the line or group of lines (see fig 7.6) [161] . These markers define the background limits below and above the peak. When single markers are set, the corresponding channels are taken as reference points for the background function. At the same time the markers thus set are the starting or the finishing points of the background subtraction. Setting pairs of markers (on either side of the line) defines limiting windows. The marker nearest to the peak determines the finishing point of the background subtraction and integration (or fitting). If the background subtraction is non-linear, the values in the windows are averaged. The results define the background level above and below the peak [160]. The determination of the background values between the limits on either side of the peak is based on the following model: Some of the electrons have suffered energy loss because of scattering in the sample and these are detected at lower energy values. The number of such Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 195 particles increases towards lower kinetic energies proportional to the number of unscattered or elasticaJly scattered particles. —j- = nelast (7.4) where: neiast = Count rate in the peak without a background of higher energy peaks; nelast = nl - nb n l = Measured count rate This yields the following iterative algorithm for the calculation of neiast. {neiast)l+1{E) = n{E) - n ^ ) - Constant nlelaste (7.5) constant - E ^ (7.6) where: [n(E1) - n(E2] /'E2 ' £j nelast£ For non-linear background subtraction the program (BIGFIT) will calculate the values E\, E2, and n(E2) from the reference points. The background subtraction is then performed according to the above model. The integration option is used for determining the peak area by simple intensity summation. The fit option is used for fitting a Gaussian/Lorentzian or an intermediate function to peaks or doublets of a spectrum using the least squares method. Fitting is carried out simultaneously for all peaks of a region. This makes it possible to fit superimposed peaks or doublets. The parameters of a peak, which can be optimised (fitting parameters), are the mixing ratio for a Gaussian/Lorentzian intermediate function, the peak position, height, and the peak width which is defined as the full width at half-maximum(FWHM values). When processing doublets, hot only the parameters mentioned above are optimised, but also the distance between the doublet lines, their intensity ratio and the F W H M value of the second peak. For fitting, the user specifies estimated values for all parameters which are to be optimised. Then, the system will optimise all parameters or a selection of these parameters by fitting an analytical spectrum [160]. Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 196 \ / • -V-- - A / \ single marker pair of markers (window) - A A- / / . . . / . - • / \ \ energy E Literature: D.A. Shirley: Phys. Rev. B5, 4709 (1972) I I Ei E 2 Figure 7.8: I: Peak With Markers for Background Subtraction (a single marker on the low energy side of the peak, a pair of markers on the high energy side), II: Literature: D . A . Shirley: Phys. Rev. B5, 4709 (1972) Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 197 The amount of Gaussian or Lorentzian character is indicated by a parameter known as the mixing ratio, which takes the value of 0 for a pure Lorentzian peak and the value 99 for a pure Gaussian peak. Each parameter was first estimated and then optimized separately using a combination of the gradient method and the polynomial approximation [162]. The criterion for "best fit" was the minimization of the CHI square value, X2, where X is defined as X I A {Fm - Ff)2 \n- nfit fr[ Fm where: Fm = Measured count rate. Ff = Intensity value of fitting function, n = Number of points in fitted region. rijit — Number of fitting parameters. An initial optimization of the mixing ratio yielded a value of 1; this value was used in all subsequent fits. Non linear background subtraction was performed on the smoothed high reso-lution spectra. The resulting spectrum will consist of the adapted analytical function added to the original background so that the result can be directly compared with the original spectrum. The mathematical method on which fitting is based is a combination of the gradient method and polynomial approximation. The algorithm used was developed by D.W.Marquardt [160]. On the basis of the estimated values, the method determines those values of the fitting param-eters for which the corresponding reduced CHI square has a local rriinimum. The Gaussian/Lorentzian intermediate function can be deteranned using the following equa-tion [163]: )2)n where: fn{E) = Intensity value of fit function. I0 = Peak height. E = Kinetic energy (binding energy =1253.6(ev) -K.E.) Ea = Peak centre. = 7 7 - ^ 7 l b - T ^ (7-8) Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 198 F W H M = Full width at half-maximum. C = 4(2 1 / , n) — 1 = Normalization constant. N = 1 for Lorentzian and n —> oo for Gaussian. The integration option was used to determine the peak areas by simple intensity summation and to identify the total amount of contaminated elements observed on the surface. Table 7.3 shows the computer output of the curve fitting routine called (BIGFIT). The computer output for different electrodes indicates two relative intensities for each individual peak which results from the addition of the peaks compared to the different oxidation type. The rest of the outputs are presented in Appendix D. Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 199 spet'irun 224 .BF2 ! I Pt e l e c t r o d e t r e a t e d w i t h decanol< tt4>;Ar i o n s p u t t e r i n g 5 win. ! l a o e l I n a r r o u scan f o r NaIs.013+Pt4p3/2,CIs ,Si2p+ptif . i I o r i g i n a t o r : Chen E . S h i v a Sharaeh. o p e r a t o r : PU + + + + + ^ H 4 + -+ + + I s p e c i f i c a t . I r e g i o n l r . c h i I r a t l p o a i t I i n t e n s . I f u h m INI s e n . f I r e l . i n t I IPt if I IPSO I IPt :OH )2 41.10113 11 I 83.I4I5.669E2II.91 IXI 1.001 36.1781 I-t 1-1-3.351-1•.75 1-1 I - l I-l-2.7011.483E2I-1 1-1 1=3-3.35I-3-.75 I - l I - l I—I — I.2211.602E2I-I I - l I-S-3.35I-5-.75 I - l IXI 1.001 27.1331 IXI 1.001 9.46GI IXI 1.001 7.0991 IXI I.001 11.4991 I x l 1.001 8.6251 I I Pt e l e c t r o d e t r e a t e d w i t h decano1<«14)ifir i o n s p u t t e r i n g 5 min. I l a b e l I n a r r o u s c a n f o r Nals,015+Pt4p3/2,C1s,S12p+pt4f. I I o r i g i n a t o r : Chen E . S h i v a Sharaeh. o p e r a t o r : PU + + + + + + + + +_+ + + : I s p e c i f i c a t . I r e g i o n l r . c h i I r a t l p o a i t I i n t e n s . I f u h n I N I s e n . f I r e l . i n t I 10 Is • IC Is IPt 4f7/2 17. S70E2I I0.000E0I I3.946E3I I x l .781 50.2421 IXI .341 0.0001 IXI 3.901 49.7581 Table 7.3: Values) Computer Output Results of a Typical Run, a) (Curve Fitting), b) (Integrated Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 200 The overal XPS results (Integrated values) are given in terms of relative peak areas which are presented in table 7.4. This table shows the degree of oxygen and carbon contamination. Table 7.5 shows the results obtained from curve fitting. This table shows the type and amount of oxidation platinum wires were exposed to. It was observed that coated wire electrodes No.14 and No.25 contained higher ratio's of platinum oxidation of different types than did treated platinum electrodes. As was shown in table 7.5, electrodes No.14 and No.25 which were coated with a polymer containing decanol(long chain alcohol) had the highest oxidation of the platinum surfaces. The high contamination by O(ls) and C(ls) can be explained by the fact that the wires still maintained some outside coating compared to clean platinums. The platinum electrodes (No.14 and No.25) that were exposed to decanol showed higher oxygen adsorption compared to the treated platinums. The platinum wire that was not treated also contained a high oxygen to platinum ratio. It can therefore, be concluded that different surface treatments can cause different platinum oxidation levels. In order to keep a reference point , these platinum wires have to be first cleaned since the contaminants can block all the active sites present at the surface of the metal. By treating the new electrodes thermally and by exposing the surfaces to acid, one can observe a base initial surface inorder to have meaningful responses. 7.9 R e s u l t s a n d D i s c u s s i o n The platinum electrodes were tested potentiometrically and their surfaces were analyzed by XPS to determine if there is any correlation between oxidation level and potentiometric decay. 7.9.1 E l e c t r o c h e m i c a l studies As indicated" in the literature, platinum electrodes which have undergone extended anodic or cathodic polarization exhibit potentiometric properties that distinguish them from untreated platinum surfaces. If the platinum surface treatment had no effect on the surface potential, then the values should have been the same for all treatments within experimental error. Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 201 Table 7.4: Area Ratio's For Platinum Electrodes (Integrated value) Electrode Ols/Pt4f Cls/Pt4f Nals/Pt4f O l s / C l s 1 0.3602 0.2624 1.3724 2 1.1080 2.396 0.4623 3 2.571 4.343 0.650 0.5919 4 3.897 .4.280 1.253 0.9106 5 c 0.372 . 1.34 0.2776 0 7 1.19 0.446 2.668 14 1.0097 19 0.329 1.6062 0.205 25 0.1042 0.8343 0.1042 4(rec) 1.073 0.92145 1.1651 Table 7.5: Oxidation Ratio's For Platinum Electrodes (Curve fit value) Electrode PtO/Pt4f Pt (0#) 2 Pt4f PtTot/Pt4f 1 0.0928 0.2473 0.340 2 0.1022 0.2839 0.386 3 0.1235 0.1868 0.310 4 0.0363 0.2957 0.332 5 0.1380 0.00 0.138 6 0.00 0.00 0.00 7 0.24 0.00 0.24 14 0.2616 0.3178 0.579 19 0.1242 0.2986 0.423 25 0.1919 0.2803 0.472 4(rec) 0.1589 0.2169 0.376 Electrode 14 has been argon sputtered for 5 minutes. Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 202 Potentiometric measurements were observed after the treatments of the platinum wires as summarized in table 7.2 section 7.7. Figure 7.91 shows the potentiometric responses of the bare platinum electrodes mentioned in table 7.2 recorded until stable potentials were achieved in phosphate buffer solution. There was a large range of initial potentials from 118 to 265v. The steady state potentials ranged from 280 ± 25mv. Figure 7.911 shows the effect of the addition of alcohol to the measuring solution. Electrode No. 19 which was not flame treated did not change potentiometrically after addition of n-decyl alcohol. Electrodes No.l and No.2 which were treated for.2min, and 5sec in a flame responded the most and electrodes No.14 and No.25 which were coated wire sensors responded minimally after the addition of n-decyl alcohol. Figure 7.10 shows the potentiometric responses of the bare platinum electrodes retreated with HN03 and flame treated for 2 minutes. The potential of the electrodes after stabilization ranged from 280 ± 50mv. After the addition of n-decyl alcohol to the measuring solution, it can be seen that all the electrode voltages decreased differently. This could be due to the platinum surface structure itself. If the platinums had the same uniform structure on their surfaces, then the electrodes should have behaved in the same way. But it has already been shown that electrodes No.14 and No.19 and No.25 potentiometrically decrease much less than when the electrodes were retreated and were not coated. Table 7.6 [114] shows the potentiometric response of bare platinum electrodes treated as indicated and measured in air-saturated pH 7.4 phosphate buffer. By comparing the potentio-metric values from the literature as given in table 7.6 with figure 7.91, 7.101, one can see that the potentiometric responses of the bare platinum surfaces are within the same range. It can be proposed that by coating and by leaving the platinum surfaces untreated, the active sites of platinum wires may be blocked which can cause a reduction in their response to alcohol. That could also mean that the surface of the platinum saturates with oxygen. Whether this happens to the surface can be determined from the XPS analyses. In order to observe how bare platinum responds potentiometrically to weak and strong acids, /3D-glucose and n-decyl alcohol additions were made at steady state. These were chosen because Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 203 both compounds were used in preparing the original electrodes. The steady state potentials were found to be within the same range reported in the literature [114]. Figure 7.111 shows the behavior of two treated bare platinum electrodes. The word treated means that both platinum wires were cleaned with HN03 and were thermally flamed. Table 7.7 is the potentiometric data with respect to time before and after the addition of /3D-glucose. After steady state potential was achieved, different amount of glucose were added to bring the final glucose concentration up to 600(mg/dl). Even though the observed potential change is very small, it can be seen that the effect of glucose addition is not similar for the two platinums. This could be due to the difficulty encountered in making reproducible electrode surfaces. One explanation could be the effect of different contaminants on the surface of the platinum. The difference could also be the effect of the oxides at the surface of the platinums. Figure 7.11II shows the potentiometric effect of decanol on three treated bare platinum electrodes. Table 7.8 shows the potentiometric data with respect to time before and after the addition of n-decyl alcohol. Again, addition of n-decyl alcohol did not show the same change of potential with each of the prepared electrodes. It can be observed that either surface contamination or oxides formation at the surface of the electrodes, can effect the potentiometric responses. Another possibility could be that a previous organic coating, even if removed, or if not cleaned, the platinum surface can effect the potentiometric response. To determine whether the surface effects could be quantified, the same electrodes were examined using XPS. Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 204 Figure 7.9: Potentiometric Responses of Bare Platinum Electrodes Treated as Indicated in Table 7.2 in Phosphate Buffer pH of 7.4, I: Bare Platinum Surfaces, IT: Bare Platinum Surfaces After Addition of n-Decyl Alcohol Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 205 Figure 7.10: Potentiometric Responses of Bare Platinum Electrodes Retreated and Tested in Phosphate Buffer pH of 7.4, I: Bare Platinum Surfaces, II: Bare Platinum Surfaces After Addition of n-Decyl Alcohol Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 206 measd potential* (mV platinum treatment vs. Ag/AgCl) cathodic 384 ± 52 double layer 362 ± 35 anodic 307 ± 40 flamed 305 ± 34 ° Mean ± standard deviation for six to seven electrodes. See text for statistical comparisons of means. Table 7.6: Potentiometric Response of Bare Platinum Electrodes Treated as Indicated and Measured in Air-Saturated pH 7.4 Phosphate Buffer [114] Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 207 7.9.2 X P S S t u d i e s I n t e r p e t a t i o n o f T h e D a t a The platinum wires listed in table 7.2 were analyzed to determine the contamination and the type of oxidation on their surfaces. The primary objective was to determine if the surfaces were clean and whether they showed the same type of surface contamination and amount of oxidation. The study concentrated on the outer electron shell of the platinum which is most sensitive to oxidation. The XPS measurment can indicate the type of oxidation and the ratio of the individual types at the surface of the platinum. In order to have a better explanation of the experiental data, a few important parameter have to be mentioned. 1. The thickness of the platinum wire was 0.25mm. 2. The small surface area can causes roughness at the surface. 3. The surface is not smooth. 4. Small sample preparation uses low resolution, and low resolution leads to low intensity which can cause high noise in the spectra . This can be eliminated by using smooth background subtraction. Figure 7.12 shows qualitatively the contamination levels of platinum No.5 "As received", as compared to platinum No.6 and platinum No.7. By looking at these figures, one can see that electrode No.5 shows a lower oxygen peak in figure 7.121 as compared to electrode No.7 (7.12III) but a higher platinum (4f) peak. Because of the treatment given electrode No.6 , one can see the surface of the platinum in 4p 3 / 2 or pt(4d) clearly since no oxygen or carbon contamination is at the surface. The intensities observed are higher for cleaned surfaces compared to an as received surface and to an as received surface after potential testing. The degree of contamination and chemisorption on platinum surfaces and how it corresponds to different levels of oxidation will be explained later. A qualitative interpretation of the surface contaminants can also be Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 208 Figure 7.11: Potentiometric Responses of Bare Platinum Electrodes Treated and Measured in 200ml p H 7.4 Phosphate Buffer After Addition of Different Amount of I: Glucose, II: n-Decyl Alcohol Intana Icp»; l l n Figure 7.12: Comparisons Between X-ray PhotoElectron Spectra of Platinum Wire I: No.5 IT No.6, HI: No.7 Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 210 obtained from figure 7.13 which shows the contamination levels at each particular orbital. Figure 7.131 shows the oxygen contamination of the surface for platinum No.5(263), No.6(268), No.7(270). It is highest for platinum No.7 which had been tested potentiometrically and lowest for platinum No.6 which had not. Figure 7.13II shows carbon contamination of the surface for platinum No.5, No.6, and No.7. The carbon contamination is highest for platinum No.5 (As received) and lowest for platinum No.6 which had been cleaned. Figure 7.13111 shows the total platinum surface contamination which is highest for platinum No.7 and lowest for platinum No.6. This can lead us to the conclusion that by treating platinum surfaces, one can remove the contaminants from the surface. As mentioned in the literature [114], the "As received" platinum electrodes show little evi-dence of any surface oxide film, as observed by the pt (4f) spectrum. It is common to find large amounts of oxygen and carbon in association with an uncleaned surface. This contamination which is mainly organic in nature, is indicated by the 0(ls)/Pt(4f) and C(ls)/Pt(4f) ratios. It can also be expected to come from hydrocarbon contamination when the platinum has been coated by a polymer. The type of functional group of the carbon attached to the surface was not observed and is beyond the scope of this study. The literature results agree with our ex-perimental observations [114]. Figure 7.14 shows the wide scan spectra of electrodes No.3 , No.14 and No.25. These electrodes differ markedly in their history, but all were reused wires. Electrodes No.14 and No.25 had an organic coating scraped from them just prior to analysis while No.3 was acid and flame cleaned before analysis. Concerning contamination, the surface of electrode No.3 contains Na(ls), O(ls), C(ls), Na(a), No.14 contains 0(a), O(ls), and No.25 contains O(ls), C(ls). Electrode No.25 (spectra 233) shows a low oxygen peak and carbon contamination in the Is shell and a high platinum peak. Electrode No.14 (spectra 223) shows a higher oxygen peak and lower platinum peak and finally electrode No.3 shows a lower oxygen peak compared to electrode No.14 but higher oxygen peak compared to electrode No.25. One can see that the degree of contamination is different in those electrodes. When the surface is exposed to either air or H20, oxygen molecules cover the surface layer, making it impossible to Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 211 Figure 7.13: XPS Results of Platinum Wires No.5, No.6, No.7 (Narrow Scan) I: Oxygen level, fl": Carbon level, UL Platinum level Figure 7.14: Compansons Between X-ray Efectrophoton Spectra of Platinum Wire No.3, No. 14 and No.25 Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 213 detect the platinum surface. If oxygen is shown to be at the surface it does not mean that the surface of the platinum is oxidized. The oxygen can be adsorbed to the outer surface layers. The overall XPS results (integrated values) are given in terms of relative peak areas which are presented in table 7.3. Table 7.4 shows the integrated results which indicate the degree of oxygen and carbon contamination. By comparing the results given in the table 7.4 to figure 7.14 , the following interpretation can be made: 1. The total level of oxidation chemisorbed to the platinum surface is the highest for elec-trodes No.14 and 25 and lowest for No.3. Narrow scans usually show much better results because data are accumulated over only a narrow energy region. This scan presents a high resolution spectrum which shows peak position, chemical states, and peak areas for quantification. A typical narrow scan spectra of a clean platinum surface is shown in figure 7.7. The difference between the clean surface spectrum and those discussed above is not only due to the presence of contaminants at the surface but also due to oxidation of the platinum. When platinum wires are exposed to air, carbon and oxygen and even sodium can physically be adsorbed to its surface. For that reason, once platinum wires have been treated or exposed to air, certain elements can be physically adsorbed obscuring the platinum surface. Table 7.4 shows that the highest C(ls) and 0(ls) contamination was on electrodes No.2, No.3, No.4. As mentioned before, since the carbon and oxygen contamination had covered the platinum surface, argon sputtering was used on the surface of the platinum No.4 to expose the surface. Argon sputtering can roughen the surface and can get rid of the adsorbed contaminants. As can be seen in figure 7.15 , spectrum 223 which was observed after sputtering shows that all the carbon contamination was removed and the platinum surface is shown in the spectrum. By comparing the spectra one can see the differences in the elements present before and after sputtering. Argon sputtering was only done on electrode No.14 which had been recoated with polymer and which showed a large contamination from carbon and oxygen. Figure 7.16 shows the spectra produced by using a wide scan on electrode No.4 and electrode No.4(rec). No.4(rec) Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 214 XPS SP 217 XPS SP 223 wide scan wide scan . 4000 s 2000 s !rter>a; *_v 550 400 SCO 60C ;?30 • 018 + Cls ;ooo •1300 BOO POO *Pt4f Figure 7.15: XPS Results for Electrode No.14 Before and After Sputtering. Spectra 217 Rep-resents Before Sputtering and Spectra 223 Represents After Sputtering. Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 215 was recleaned with HNOz and flame after XPS testing, to see if any change in oxidation level or degree of contamination would be observed. After treatment of the electrode one can see that the amount of contamination had decreased(refer also to table 7.4) . The surface of electrode No.4 contains Na(ls), 0(ls), C(ls), Na(a), and Pt(4f) whereas electrode No.4(rec) contains a higher platinum surface(4f)peak and 0(ls)peaks. Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 216 By using a high ultra vaccum for a long period of time, one can also get rid of the physically adsorbed materials from the surfaces. This could be a reasonable approach for decontaminating surfaces but it is not very feasible nor economical. The contaminants could include absorbed water , sodium contamination caused by exposure to the skin , and oxygen contamination caused by air or water . By comparing the results shown in figure 7.17, one can see that U H V reduced the Na(ls) contamination and increased the 0(ls) to pt(4p 3/ 2) ratio. 7.9.3 Comparisons of Electrochemical Observations with XPS Results Oxidation ratio's of different types of oxidation which were observed from XPS studies were plotted against potentiometric responses of initial, and steady state potentials before and after the addition of decanol. No significant observation were seen from these plots. The plotted observations can be found in appendix D . Finally comparison was made between the "AVoltage" after addition of decanol. " AVoltage" is calculated as the steady state potentials before and after addition of decanol to the buffer solution. It was the purpose of this test to determine if there was any correlation between the platinum wires that were treated and retreated. The term "treated" indicates electrodes presented in table 7.2 according to the history of platinum wires given. The term "retreated" means that the bare platinum electrodes were retreated with HNOz and flame treated for 2 minutes before being reused. Table 7.7 describes the different oxidation and contamination levels of the treated and retreated platinum surfaces and the potentiometric change for each after the addition of decanol. The total contamination is calculated by adding the carbon and oxygen contamination measured by XPS. This table shows the "AVoltage" changes of treated bare platinum electrodes with respect to their total surface oxidation and total contamination. It can be seen that the "AVoltages" observed for treated platinum wires exposed to organic coatings (No.14, No.25) or were not cleaned (No.19), did not-show a substantial decrease in potential after the addition of decanol. This could be due to saturation of the platinum surface with oxygen. Another possibility could be that a previous organic coating, even if removed can effect the potentiometric response. - As Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 217 Figure 7.16: Comparison of XPS Results of Electrode No.4 and Electrode No.4 Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 218 XPS XPS 212 210 Ns la Na la 1. a 1. a Sap 10. 90 Aug 23. 90 i n t e n s i t y ~ l i n XPS SP 212 XPS SP 210 0 Is 0 IS Sap 10. 90 Aug 23. 90 II H - i-Slfi 510 50b Oind. «r>«rgv [s)VJ l i n Figure 7.17: XPS results for comparing the spectra of electrode No. after exposure to a UHV for 3 weeks, Spectra 212 represents the data Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 219 observed for retreated electrodes (Table 7.7), even though the same treatment was given to all the platinum electrodes, the same surface behavior was not produced. Platinum wires that contained a high contamination level also responded less to the addition of decanol (see Table 7.7). By comparing the potentiometric responses with the XPS results, one can see that the correlation is not only dependent on the oxidation level but also on the contaminants adsorbed to the surface. Also, all platinum wires mentioned in Table 7.2 are reused platinum wires. This could also indicate that by reusing platinum wires even if cleaned each time, could result in a different morphology of the platinum surfaces. 7.9.4 Conclusion The chemistry of the interaction of a platinum metal surface with oxygen species in aqueous solution is a complex process that is not well understood. The major species that are thought to be involved in the process are: PW2, PtO, and Pt(0H2)- The process of oxide formation has been described as initially an oxidation in which OH groups are reversibly attached to the surface layer of platinum atoms. At or near monolayer coverage, the surface atoms of platinum and OH undergo an irreversible rearrangement in which the OH groups migrate into the platinum atomic lattice. The active electrode area is quite different from the geometric area. The difference is caused by variability in the surface microstructure which results in physical roughness, adsorbed impurities and surface oxides. This can create fouling and poisoning of the electrodes. In an attempt to make reproducible electrodes, the platinum wires were placed in a con-centrated nitric acid environment and were then thermally treated in a natural gas flame for approximately 2 minutes. As indicated in literature [116], the amount of oxygen increases with the time of flame treatment. This oxygen increase can not be accounted for by platinum oxidation. The oxidation type, the amount of oxidation and the amount of contaminants on the surface Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 220 were measured and it was found: 1. Irreproducibility of platinum electrodes depends on • Type of treatment • The exposure of the surface to contaminants. • Surface oxidation and the amount of oxygen. Quoted from literature [164], "The miniaturization of CWEs (Coated wire electrodes) the limit being the available micro electrode systems, allows measurements of such volumes (micro liters) at any angle or position. This is particularly useful for the in vitro and in vivo biomedical and clinical monitoring of anatytes. The disadvantage with CWEs is the initial life time and durability of the membrane coating. Unlike the highly stable ISEs (Ion selective electrodes), CWEs usually last only for a maximum of six months, with many enduring just one or two experiments. Also, there are problems with reproducibility caused by the fluctuation of the membrane defined internal reference potential". As observed from this part of the thesis, this irreproducibihty could be due to the platinum wire itself which could cause the fluctuations in the membrane defined internal reference potential. In order to avoid this situation, it might be possible to saturate the platinum wires with oxygen causing O H groups not to reversibly attach to the surface layer of platinum atoms. This could result in a base potential for a platinum wire electrode. The behavior of individual platinum wires was also statistically analyzed to see if the stan-dard error between the responses of these electrodes are statistically significant according to their potentials and oxidation levels observed. The platinum electrodes were grouped as 2 types: treating electrode 1,2,3 as type 1 and electrode 14,19,25 as type 2. The model was based on G L M having Y l as the initial voltage, Y2 as steady state voltage, Y3 as the final change after addition of decanol, and finally the oxidation found at the surface. The standard error was predicted to be 13.5 for initial, 7.48 for the change of mv and 0.09 for oxidation which were found to be statistically insignificant. The contour plots of the predicted values versus residuals Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 221 for both voltages and the oxides showed that the data is reliable. The results obtained indicated that reused platinum wires cannot potentiometrically produce the same initial and steady state potentials even though their surfaces were treated for the removal of contaminants. The next chapter will describe the overall conclusions from this study. Further recommenda-tions are given on how to improve the overall behavior of coated wire sensors without platinum oxidation effects. Chapter 7. X-RAY PHOTOELECTRON SPECTROSCOPY 222 Table 7.7: Comparisons Between Treated and Retreated Platinum Surfaces According to Levels of Contamination and Oxidation Electrode AVoltage AVoltage Total Total # Treated Retreated Oxidation Contamination Platinum Platinum PtTot/pt4f 01s+Cls/pt4f 14 15 41 0.579 1.0097 25 16 44 0.472 0.9385 19 17 45 0.423 1.9352 4 23 24 0.332 8.177 3 19 41 0.310 6.914 2 39 51 0.386 3.504 5 42 - 0.138 1.712 1 43 60 0.340 0.6226 7 46 - 0.240 1.0097 Chapter 8 DISCUSSION OF RESULTS In previous chapters, important observations were made which are the major key findings in tliis thesis. Most importantly, this thesis showed why the coated wire glucose sensor would not work in the manner reported by other workers (81,87). The coated wire glucose sensors were reported to measure different glucose concentrations. The proposed mechanism was based on the association and dissociation of a Ba-glucose complex as glucose concentration changed. From the data gathered in the present research, it was observed that glucose addition does not change the potentiometric response. It was also possible to explained how this misinterpretation of the reported results could have occurred. The principle of operation of a coated wire electrode is very similar to that for a liquid ion-selective electrode except that no internal aqueous solution is used, instead a conductor is directly coated with an ion responsive membrane. When the sensor is immersed in a buffer solution, there is a concentration difference between the two phases; the membrane and the solution. A proposed qualitative mechanism for the generation of the observed potentiometric re-sponse is given below. It was observed that by placing the coated wire electrodes in different buffer solutions independent of glucose addition, that the potential would drop until an equi-hbrium potential was achieved. The potentiometric response is influenced by: 1. Platinum wire surface effect. 2. Mass of BaCl2 used. 3. Volume of liquid ion exchanger used. 223 Two types of coated wire electrodes classified as "Air dried" and "Conditioned" were studied. It was observed, that it took considerable time for air dried coated wire glucose sensors to reach a stable potential, whereas by exposing the prepared electrodes to buffer solution prior to use (conditioned), the potentiometric drift and the time it took to reach a stable potential decreased. Two possible sources are suggested for the observed potentials; platinum surface potential and exchange equilibrium potential. In order to investigate these phenomena, experiments were conducted to determine the role played by each component of the electrode. The components of which the electrode was constructed were removed one by one to see the effect of the missing component on the potentiometric response. It was observed that: 1. By increasing the Ba-glucose complex mass, the slope of the potentiometric drift decreased (see fig 6.11). 2. By increasing the BaCl2 mass, the slope of the potentiometric drift decreased (see fig 6.9). 3. By using the same mass of glucose, BaCl2, and Ba-glucose complex, it was found that using BaCl2 gave less potential drift compared to the Ba-glucose complex and Ba-glucose complex showed less potential drift compared to glucose (see fig 6.10). The equilibrium affecting the potential within this system is: {POl-){aqs) + 3{CH3NR3)Cl{org) ^ (C H3N R3)(POl~)(org) + 3Cr{aqs)) (8.1) Where: (aqs) = Aqueous phase. (org) = Organic phase. If the membrane has no fixed mediators, the sensor potential would be equivalent to that for the exchange equilibrium . 224 Iii the case of having BaCU within the liquid exchanger; Increasing the concentration of BaCl2 in the organic phase, wouldresult in a longer time for Cl~ ions to be extracted from the organic phase to the aqueous phase. The final potential would be observed given by: a{Cl-)\org)a{POl-){aqs) a(P04i-){org)a{Cl-)3(aqs) Vproposed - Vo + Kl jnt ^" „ _ 3 _ w N ^ , ^ (°-2) It was hypothesized that the equihbrium potential was the result of two potentials, one due to surface oxides on the platinum and the other to an ionic exchange of chloride ion with phosphate ions on the liquid ion exchange material. 'Measured * proposed ~r »' platinum (8.3) The electrodes were also conditioned prior to use. The experimental observation indicated that: 1. By conditioning the prepared electrodes, the potential drift decreased. The starting and steady state potentials decreased from 800-350mv to 200-150mv. 2. By increasing the Ba-glucose complex mass, the initial and final potentials were not changed significantly to indicate any conclusive results, (see fig 6.15, 6.16) 3. Studies on air dried and conditioned electrodes gave a better understanding of why the potential decreases with time. By conditioning of these electrodes prior to use, the slope of the potentiometric response decreased. This indicated that the aliquat, contained in the electrode, caused the voltage drift shown by the electrodes. 4. By removing the aliquat (quaternary ammonium ion), the transient behavior of the elec-trode stopped. For platinum electrodes coated only with P V C and decanol, since there is no indication of any charge transfer, the potential observed could only caused by the surface potential. Then: ^Measured = Vplatinum (8-4) 225 The XPS results indicated that two types of oxidation occurred at the platinum surface. The process of platinum oxide formation has been described initially as an oxidation in which O H groups reversibly attach to the surface layer of the platinum atomic lattice. Thus possible sources of oxidation which occurred during the experiments could be explained by the following reactions at the platinum surface: Pt + H20—>PtOH + H+ + e- (8.5) PtOH —> PtO + H+ + e~ (8.6) Pt + 2H20—• Pt{OH)2 + 2H+ + 2e~ (8.7) It was recognized that producing reproducible platinum surfaces when covered with a coating material exposed to the measuring solution is difficult. As explained in chapter 7, the irrepro-ducibility of platinum surfaces was found to be affected by. the degree of contamination and the type and amount of oxidation to which the platinum surfaces are exposed. Coated wire electrodes are covered with a decanol layer which has been shown to have a large effect on the potential of platinum electrodes. As evidenced in literature, rearrangment of the platinum atomic lattice also has been shown to occur when oxygen dissolves in the metal (145). Accord-ing to the experimental observations in the present research, every platinum surface could have a different level of oxidation and contamination. Since the morphology of platinum changes according to the surface conditions, the electrodes could not possibly give identical responses. This could be the cause of the variability in potential response of the electrodes. In summary, the postulated mechanism, since decanol can influence or react with platinum surfaces to give one or more oxide species, appears to be consistent with our data. The type and amount of platinum surface oxides may vary with the method of platinum pretreatment. However, our present knowledge of the platinum oxide surface chemistry, and specifically, which oxide can cause a greater effect on the potentiometric response, is very limited. 226 Chapter 9 C O N C L U S I O N S A N D R E C O M M E N D A T I O N S 9.1 Conclusions The behavior of the coated wire glucose sensors has been studied in detail. This study has shown that the Ba-glucose complex in the coated wire sensor does not have the effect of causing a response to glucose concentration. From the experimental and analytical results presented in chapters 4, 5, 6 and 7, the fol-lowing specific conclusions may be drawn: 1. Ba-glucose complex in glass, needle, coated wire and membrane glucose electrodes did not cause a change in potential as the glucose concentration changed. 2. The effect of surface reactions or contamination plays an important role on the repro-ducibility of the potential of platinum surfaces. 3. Once having been coated, the structure of the platinum surfaces has changed. This would suggest that platinum electrodes should not be used for potentiometric measurements unless the surface oxidation effects can be controlled potentiometrically. 9.2 Recommendation Coated wire glucose sensors were found not to be suitable for sensing glucose. A few recom-mendations which can give further insight into the development of glucose sensors 1. Glucose sensor oxidation should be measured amperometrically. 227 Chapter 9. CONCLUSIONS AND RECOMMENDATIONS 228 2. Using a metal-glucose salt which makes a weak association between the metal and the salt for detection of glucose molecule should not be used in constructing electrochemical sensors. 3. The platinum wire should be replaced with another type of metal on which surface con-tamination and surface oxidation can not occur. 4. Totally oxidized platinum surface could be a constant basic surface which may be used for coated electodes. Nomenclat ure tableO A Area from which photoelectrons are detected. C Normalization constant. D Detection efficiency for emitted electrons. E'- Measured Potential(V). E Kinetic energy (binding energy =1253.6(ev) -K.E.) EO Constant Potential(V). EB Binding energy. EK Kinetic energy of the ejected electron. E0 Peak centre. e Efficiency of photoelectric process. F Faraday (96487 coulombs). Fm Measured count rate. Ff Intensity value of fit function. FWHM Full width at half-maximum. f X-ray flux. Intensity value of fit function. h Planks constant. hv Energy of the ionizing photon. h Peak height. N 1 for Lorentzian and n —* oo for Gaussian. n Number of points in fit region. KeLatt Count rate in peak without background of higher energy peaks; 229 Nomenclature 230 nelast "1 - "fe n l Measured count rate. Ufa Number of fit parameters. R Large mobile organic exchange(sites) in the membrane. R' Gas constant (8.314 J/mo\e°K). T Temperature °K. Z Charge. Greek Letters A Effective escape depth for transition. Pk(z) Atomic concentration of element K. v Frequency of radiation. Photoelectric cross section. \£ Angular efficiency factor. References 1 Vander, A . and Sherman, J., Human Physiology, McGraw Hill, New York, 1975. 2 Gough, D . A . , Aisenberg, S., Colton, C.K. , Giner, J., Soeldner, J.S., "The status of electro chemical sensors for in vivo glucose monitoring", Diabetes Care, 10-22, 1982. 3 Wilkins, E. , and Wilkins, M . , " The coated wire electrode glucose sensor", Hormone and Metabolic Research, 20, 50-55, 1988. 4 Pickup, J .C. , and Rothwell, D. , "Technology and the diabetic patient", Med & Biol Eng & Compt, 22, 385-400, 1984. 5 Albisser, M . A . , Botz, C.K., and Marliss, E.B. , "Studies with an artificial endocrine pan-creas", Arch Intern Medicine, Vol 137, 639-649, 1977. 6 Pfeiffer, D.F. and Clemens, A . H . , "The artificial beta cell , A continuous control of blood sugar by external regulation of insulin infusion( Glucose controlled insulin infusion sys-tem)", Horm Metab Res, 487, 339-342, 1974. 7 Schubert, W., Barschmidt, P., Nagel, J., Thull, R., Schaldach, M . , " A n implantable arti-ficial pancreas", Med&Biol Eng&Comput, 18, 527-537, 1980. 8 Schade, S.David., Eaton, R.Philip., Friedman, M.Neal., "Five-day programmed intraperi-toneal insulin delivery in insulin-dependent Diabetic man", J of Clinical Endocrinology & Metabolism, Vol 52, No.6, 161-166, 1981. 9 Higgins, I.J., Potter, W.G. , Turner, A.P.F. , Biosensors,Vol 1, No.3, 213-320, 1985. 10 Bruno, W., Volk, Arquilla., Edward, R., "The Diabetic Pancreas", Plenum Pubhshing Corporation, New York, N.Y, 1985. 231 References 232 11 Reach, G. , " A method for evaluating in vivo the functional characteristics of glucose sen-sor", Biosensors, Vol 2, 211-220, 1986. 12 Mcic, Krull.U.J., "The Biosensor revolution:Fad or reality?", Canadian Chemical News, 13-16, March 1988. 13 Lucisano, Y.Joseph., Armour, CJon. , Gough, A.David., "In vitro stability of an oxygen sensor", Anal Chem, Vol 59, No.5, 736-739, 1987. 14 Peirce II, E.Converse., Dibelius, R.Norman., "The membrane lung: studies with a new high permeability co-polymer membrane", Vol 10,Trans Amer Soc Artif Int Organ, 220-226, 1968. 15 Biermann, E . , Armour, J . C , Mckean, B.D., Lucisano, J.Y., Gough, D.A. , " Chronic im-plantation of glucose sensor in dogs", Abstracts, Diabetologia, Vol 30, No.7, 499A, July 1987. 16 Bessman, S.P., Schultz, R.D. , "Sugar electrode sensor for the artificial pancreas", J.Horm Metab Res , Vol 4, 413-417, 1972. 17 Yao, S.J., Chan, Lap-Tak., Wolfson, K.Sidney., "The low potential approach of glucose sensing", I E E E Transactions on biomedical Eng, Vol.Bme-33, No.2, 139-146, Feb 1986. 18 Mascini, Marco., Guibault, G . , "Clinical uses of enzyme electrode probes", Biosensors, 2, 147-172, 1986. 19 Rehwald, Wolfgang., Geibel, John., & Gstrein, Edith., " A microelectrode for continuous monitoring of glucose concentration in isolated perfused tubule segments", Pflugers Arch, 400, 398-402, 1984. 20 Fischer, U . , Ertle, R., Abel, P., "Assessment of subcutaneous glucose concentra-tion,Validation of the wick technique as a reference for implanted electrochemical sensors in normal and diabetic dogs", Diabetologia, 30, 940-945, 1987. .References 233 21 Fischer, U . , Abel, P., Brunstein, E . , Rebrin, K . , and Salzsieder, E. , " Interstitial glucose concentration as measured by implanted sensors", Biosensors, 255,259, 1987. 22 Turner, A.P.F. , "Miniaturised Sensors", Biosensors Fundamental and Applications, Oxford University Press, 229-233, 1987. 23 Kerner, W., Zier, H . , Steinback, G., et al., "Measurment of subcutaneous glucose concen-tration in sheep with an amperometric enzyme electrode", Diabetologia, 30, 539A, 1987. 24 Lobel, Ester., Rishpon,Judith., "Enzyme electrode for the determination of glucose", Anal Chem , Vol 53, 51-53, 1981. 25 Clark, L . C . and Lyons, C , "Electrode system for continuous monitoring in cardiovascular surgery", Ann N Y Acad Sci , 102, 29, 1962. 26 Updike, S.J., and Hicks, G.P., "The Enzyme electrode", Nature, 214, 986, 1967. 27 Updike, S.J., Shults, M . , and Ekman, B., "Implanting the glucose enzyme electrode: prob-lems, progress and alternative solutions", Diabetes Care 5, 207, 1982. 28 Gough, D . A . , Lucisiano, J.Y., and Tse, P.H.S., "Two- dimentional enzyme electrode sensor for glucose", Anal Chem, 57m, 2351-2357, 1985. 29 Clark, L . C , "Membrane polarographic electrode system and method with electrochemical compensation", U.S.patent, No.3, 539, 455, 1970. 30 Lobel, E . , and Rishpon, J., "Enzyme electrode for the determination of glucose" , Anal Chem, 53, 51-53, 1981. 31 Volkl, Harold., Geibel, John., Rehwald, Wolfgang., lang, Florian., " A microelectrode for continuous monitoring of redox activity in isolated perfused tubule segments", Plugers Arch, 400, 393-397, 1984. References 234 32 Nagy, G. , Vonstrop, L . H . , and Guilbault, G . G . , "Enzyme electrode for glucose based on an iodide membrane sensor", Anal Chim Acta, 66, 443, 1973. 33 Wingard, L.B. , Schiller, J .G. , Wolfson, S.K., et al."Immobilized enzyme electrodes for the potentiometric measurement of glucose concentratiomimmobilization techniques and materials", J . Biomed Mat Res, 13, 921, 1979. 34 Wingard, L.B. and Castner, J., "Potentiometric biosensors based on redox electrodes, in biosensors: Fundamentals and applications (Turner, A.P.F. , Karube, I., and Wilson, G.s. eds.), Oxford University Press, Oxford, 153, 1987. 35 Caras, D.Steve., Petelenz, Danuta., Janata, Jiri., "PH-based enzyme potentiometric sen-sors,glucose sensitive field effect transistor", Analytical Chem, 57, 1920-1923, 1985. 36 Fischer, U . , Abel, P., " A membrane combination for implantable glucose sensors measure-ments in undiluted biological fluids " , Trans A m Soc Artif Inter Organs, Vol 28, 245-248, 1982. 37 Foulds, C.Nicola., Lowe, R.Christopher., "Enzyme entrapment in electrically conducting polymers " , J.Chem Soc Faraday Trans, 82, 1259-1264, 1986. 38 Cobell, Claudio., and Mari, Andrew., "Control of diabetes with artificial systems for in-sulin delivery-Alogorithm independent limitations revealed by a modeling study", I E E E Transactions on Biomedical Eng, Vol Bme-32, No.10, 840-845, October 1985. 39 Gough, David.A., Lucisano, YJoseph., Tse, Pius.H.S., "Two dimensional enzyme electrode sensor for glucose", Anal Chem, Vol 57, No.12, 2351-2357, Oct 1985. 40 Niijima, Akira., and Mei, Noel., "Glucose sensors in viscera and control of blood glucose level", Nips, Vol 2, 164-167, Oct 1987. 41 Leypoldt, KJohn. , Gough, David.A., "Model of a two-substrate enzyme electrode for glucose", Anal Chem, Vol 56, No.14, 2896-2903, Dec 1984. References 235 42 Gough, DavicLA., Armour, CJon. , Lucisano, YJoseph., Mckean, BrianJ)., "Short term in vivo operation of a glucose sensor", Vol 17, Trans A m Soc Artif Intern Organs, 148-150, 1986. 43 Schick, G.K. , Mageaura, V . G . , and Ruber, C O . , "Amperometric non enzymatic determi-nation of serum glucose by means of a nickle- catalyst electrode", Clin Chem, Vol 24, No.3, 448-450, 1978. 44 Luksha, Eugene., "Glucose sensor", United State Patent, Pat. No. 4,679,562, July 14, 1987. 45 Gough et al., "Complete glucose monitoring system with an implantable telemetered sensor module", United State Patent, Pat. No. 4,703,757, Nov 3, 1987. 46 Gebhardt, U . , Luft, G. , Richter, G.J., "Development of an implantable electrocatalytic glucose sensor", Bioelectrochem and Bioenergetics , Vol 5, 607-624, 1978. 47 Lowe, C.R., "The affinity electrode" ,FEBS Letters, Vol 106, 405-408, Oct 1979. 48 Schultz S.Jerome., Mansouri, Sohrab., Goldstein, J.Irwin., "Affinity sensor: A new tech-nique for developing implantable sensors for glucose and other metabolites", Diabetes Care, Vol 5, No.3, 245-253, May-June 1982. 49 Mansouri, S., and Schulz, J.S., " A miniaturize optical glucose sensor based on affinity binding", Biotechnology, 2, 885, 1984. 50 Claremont, D.J. , Penton, Claire., and Pickup, J . C , "Potentially implantable ferrocene mediated glucose sensor", J.Biomed Eng, 8, 272-275, 1986. 51 Pickup, J . C , Claremont, D.J. , " A potentially implantable glucose sensor with direct elec-tron transfer", Diabetic Research and Clinical Practice, suppl.l, S447, 1985. 52 Claremond, D.J . , Sambrook, I.E., Penton, C , Pickup, J . C , " Subcutaneous implantation of a ferrocene mediated glucose sensor in pigs", Diabetologia, 29, 817-821, 1986. References 236 53 Cass, A . E . G . , Davis, G . , Francis, G.D. , et.al," Ferrocene-mediated enzyme electrode for amperometric determination of glucose', Anal Chem, 56, 667, 1984. 54 Mor, J.R., and Guarmaccia, R., "Assay of glucose using an electrochemical enzymatic sensor", Anal Biochem, 79, 319, 1977. 55 Nicola, C.Foulds and Lowe, Christopher.R., "Enzyme entrapment in electrically conducting polymers", J.Chem Soc Faraday Trans I, 82, 1259-1264, 1986. 56 Abel, P., Muller, A. , and Fischer, U . , "Experience with an implantable glucose sensor as a prerequisite of an artificial beta cell", Biomed Biophys Acta, 5, 577, 1984. 57 Schichiri, Motoaki., Asakawa, Nobuyuki., Yamasaki, Yoshimitsu., Kawamori, Ryuzo., and Hiroshi, abe., "Telemetry glucose monitoring device with needle-type glucose sensor: A useful tool for blood glucose monitoring in diabetic individuals", Diabetes Care, Vol9, No.3, 298-301, May-June 1986. 58 Sc^chiri, Motoaki., Yamasaki, yoslumitsu., Nobuyuki, Ueda., Sekiya, Masashi., "Im-plantable glucose sensor, problems awaiting solutions for long-term clinical application", Biosensors International Workshop , G B F , Volume 10, 95-100, 1987. 59 Gough, David.A., Armour, Jan.C, Lucisano, Joseph.Y., Mckean, BrianJD., "Short-term in vivo operation of a glucose sensor", Trans A m Soc Artif Intern Organs, Vol 11, 148-150, 1986. 60 Claremont, D.J. , Penton, Claire., and Pick up, J . C , "Potentially implantable ferrocene-mediated glucose sensor", J.Biomed Eng, 8, 272-275 , 1986. 61 Shean, M.Gerald., Sollner, K. , "Carrier mechanism in the movement of ions across porous and liquid ion exchanger membranes", Annals New York Academy of Science, 759-791, 1964. References 237 62 Laitinen, A.Herbert., Harris, E.Walter., "Membrane potentials and ion selective elec-trodes", Chemical Analysis, Chapter 13, 239-249, McGraw Hill Company. 63 Berman, Herbert., "Ion selective microelectrodes", Plenum Press", 1973. 64 Moody, J .G. , Thomas, R.D.J. , "Selective Ion Sensitive Electrodes", Merrow Pubhshing Co, 1-135, 1971. 65 Arnold, Mark.A., Solsky, Robert.L., "Ion selective electrodes", Anal Chem, 58, 84R-101R, 1986. 66 Morf, W.E . , "The principles of ion-selective electrodes and of membrane transport", Else-vior Scientific Pubhshing Company, Oxford, New York, 211-263, 1981. 67 Baiulescu, George.E., Cosofret, Vasile.V., "Applications of ion- selective membrane elec-trodes in organic analysis", John Wiley and Sons inc, 94-107, 1977. 68 Sollner, K. , "Membrane Electrodes", Ann N . Y Acad Sci, 148, 154, 1968. 69 Sollner, Karl., Shean, Gerald.M., "Liquid ion exchange membranes of extreme selectivity and high penniability for anions", Anal Chem, Vol 19, 1901-1902, 1964. 70 Eisenman, George., "Similarities and Differences between liquid and solid ion exchangers and their usefulness as ion specific electrodes", Anal Chem, Vol 40, No.2, 310-320, February 1968. 71 Ross, J.W., "Calcium-selective electrode with liquid ion exchanger", Science, 156, 1378-1379, 1967. 72 Sollner, Karl., "Ion exchange membranes", Annals New York Academy of Science, Vol 139, 177-203, 1966. 73 Coetzee, C.J., and Preiser, H . , "Anion-responsive electrodes based on ion association ex-traction system", Anal Chem, 40, 2071, 1968. References 238 74 Coetzee, C.J. , Preiser, Henry., "Liquid-liquid membrane electrode based on ion association extraction systems", Anal Chem, Vol 41, No.8, 1128-1130, July 1969. 75 Matsui, M . , Freiser, H . , "Amino acid-responsive liquid membrane electrodes", Anal Lett, 3(4), 161-167, 1970. 76 Cattrell, R.W., "Coated wire ion selective electrodes", Anal Chem, Vol 43, No.13, 1905-1906, Nov 1971. 77 James, Helen., Carmack, Gary., Freiser, Henry., "Coated wire ion selective electrodes", Anal Chem, Vol 44, No.4, 856, April 1972. 78 Stworzewicz, T . , Czapkiewicz, J., Leszko, M . , "Selectivity of coated wire and liquid ion-sensitive electrodes", Jagellonia univ., Chem. Dept, Craw, Poland, Akad.Kiado, Budapest, Hungary, 259-267, 1972. 79 Cattrell, R.W., Chin, Pui.Poh., "Coated wire ion selective electrodes for the determination of mercury II", Anal Chem, Vol 48, No.3, 552-556, March 1976. 80 Wise, Donald.L., "Electrochemical glucose sensor-technology, Applied Biosensors", But-lerworth Publisher, Chapter 8, 227-245, 1989. 81 News, "Implantable glucose monitor", Anal Chem, Vol 61, No.19, October 1989, 1077A. 82 Wilkins, E. , Wilkins, M . G . , "Implantable glucose sensor",J. Biomed Eng, Vol 5, No.4, 304-315, Oct 1983. 83 E l Degheidy, M . , Wilkins, E.S., and Soudi, omnia., "Optirnization of an implantable coated wire glucose sensor", J.Biomed Eng, Vol 8, 121-129, 1986. 84 Odeylale, O.Charles., " The development of an implantable coated wire glucose sensor for an artificial beta cell", Master thesis, U N M , December 1980. References 239 85 El-degheidy, Mourad.Maamoun., "The optimization study of an implantable coated wire glucose sensor", Master Thesis, U . N . M . , August 1982. 86 Oguin, P., "The mechanism of action of an implantable coated wire glucose sensor", Master Thesis, U . N . M . , 1985. 87 Sharareh, S., "Interference effect of common components of blood with glucose sensor performance", Master Thesis, U . N . M . , 1986. 88 Lebedev, N . V . , and Liubin, B.O. , J .Appl Chem, 23, 4, 409, 1950. 89 Sharkov, V.I. , Guzhavina, V . , "Binary compounds of glucose with mineral salts", J.Appl Chem, USSR, Vol 31, 1749-1751, 1958. 90 Angyal, S.J., "Sugar-cation complexes- structure and applications", Chem Soc Rev , 415-429, 1981. 91 Franks, Felix., Hall, John.R., Irish, Donald.E., and Norris, Kathleen., "The effect of cations on the anomeric equilibrium of D-glucose in aqueous solutions- A Raman-spectral Study " , Carbohydr Res, 157, 53-64, 1986. 92 Barrett, T . W . , Spectrochim Acta, Part A , 37, 233-239, 1981. 93 Mathlouthi, M . , Luu, D . V . , "Laser-raman spectra of D-glucose and sucrose in aqueous solution"., Carbohydr Res, 81, 203-212, 1980. 94 Mathlouthi, M . , Luu, C., Meffroy-Biget, A . M . , and Luu, D.V. , " Laser-raman study of solute-solvent interactions in aqueous solutions of D-fructose, D-glucose and sucrose", Car-bohydr Res, 213-223, 81, 1980. 95 Sivchik, V . V . , Zhbankov, R .G. , Zh.Prikl Spektrosk, 28, 1038-1045, 1978. 96 She, C.Y. , Dinh, N.D., Tu, A . T . , "Laser-raman scattering of glucosamine, N-acetyl glu-cosamine, and glucuronic acid", Biodiim Biophys Acta, 372, 345-357, 1974. References 240 97 Vasko, P.D., Blackwell, J., and Koeing, J.L., "lnfared and raman spectroscopy of carbo-hydrates, Part I: Identification of 0 -H and C - H related vibrational modes for D-glucose, mallose, cellobiose, and dextran by deuterium-substitution method", Carbohydr Res, 19, 297-310, 1971. 98 Vasko, P.D., Blackwell, J., and Koenig, J.L., "lnfared and raman spectroscopy of carbohy-drates, Part II: Normal coordiate analysis of a D-glucose " . , Carbohydr Res, 23, 407-416, 1972. 99 Spedding, F . H . , Stamm, R.F. , "The raman spectra of the sugars in the solid state and in solution", J.Chem Phys, 10, 176-183, 1942. 100 Angyal, S.J., " Complexes of carbohydrates with metal cations", Aust.J Chem, 25, 1957-1966, 1972. 101 Angyal, S.J. and Mills, J .A., AustJ Chem, 32, 1993-2001, 1979. 102 Goulding, R.W., "Liquid chromatography of sugars and related polyhydric alcohols on cation exchangers, The effect of cation variation", J.Chromatogr, 103, 229-239, 1975. 103 Finch, Paul., Matulewicz, Maria.Cristina., and Helmut, Weigel., " Complexes of copper (n), Calcium, and other metal ions with carbohydrates: Thin-layer ligand-exchange chro-matography and determination of relative stabilities of complexes", Carbohyd Res, 97, 181-188, 1981. 104 Riahi, Heidar-AJi.Tajrnir., "Interaction of D-glucose with alkaline- earth metal ions. Syn-thesis, Spectroscopic, and structural characterization of Mg(LT)- and Ca(II)- D-glucose adducts and the effect of metal-ion binding on anomeric configuration of the sugar", Car-bohyd Res, 183, 35-46, 1988. 105 Riahi, Heidar-AU.Tajrnir., "Sugar interaction with metal ions: Synthesis, spectroscopic and structural analysis of Zn(II), Cd(LT), and Hg(H) complexes, containing D-glucuronate anion", J . Inorg Bio Chem, 26, 23-33, 1986. References 241 106 Riahi, Heidar-AIi.Tajmir., "Infrared spectra of crystalline B-D glucuronic acid and its Na+, K+, and Rb+ salts", Carbohyd Res, 125, 13-20, 1984. 107 Rendleman, J .A. , "Complexes of alkali metals and alkaline-earth metals with carbohy-drate", Agriculture Research Service, Illinois, 209-271, 1972. 108 Rathner, B.D. , Horbett, T . A . , "Enzymatically controlled drug release system", Drug and Enzyme Targeting (Methods in Enzymology), Academic Press, N.Y, 1-20, 1984. 109 Kost, Joseph., Horbett, Thomas.A., Rathner, Buddy., Singh, Manindor., "Glucose-sensitive membranes containing glucose oxidase: activity, swelling, and permeability stud-ies", J . Biomed Mat Res, Vol 19, 1117-1133, 1985. 110 White, MalcolmX., "The permeability of an acrylamide polymer gel", J.Amer Chem Soc, May 27, 1563-1564, 1960. 111 Bard, J.Allen., "Encyclopedia of electrochemistry of the elements", Electrochemical Stud-ies, Vol 11, 210-238, 1984. 112 Nagelberg, Irving., Braddock, Leroy.1., Barbero, Giulio.J., " Divalent phosphate electrode", Science, Vol 166, 1403-1404, Dec 1969. 113 Proctor, Andrew., Castner, James.F., Wingard, Jr., Lemvel, B., and Hercules, David.M., "Electron Spectroscopic(ESCA) studies of platinum surfaces used for enzyme electrodes", Anal Chem, 57, 1644-1650, 1985. 114 Castner, James.F., Wingard, Jr.Lemuel.B., "Alterations in potentiometric response of glu-cose oxidase platinum electrodes resulting from electrochemical or thermal pretreatments of a metal surface", Anal Chem, Vol 56, No.14, 2891-2896, Dec 1984. 115 Bard, AUen.J., "Hydrodynamic voltarnmetry in continuous- flow analysis", Electroanalyt-ical Chem, Vol 16, Marce., Dekker, Inc, New York, 115-117, 1989. References 242 116 Bard, J.Allen., "Hydrodynarnic voltarnmetry in continuous-flow analysis, metal elec-trodes", J. Elect Anal Chem, Vol 16, 115-117, 1989. 117 Woods, R., Bard, A.J . , Electroanalytical Chemitry, Vol9, Marcel Dekker, Inc, New York, P.20, 1976. 118 Giner, J., Electrochim Acta, 9, 63, 1964. 119 Adams, R .N. , "Electrochemistry at Solid Electrodes", Electrochemistry, Marcel Dekker, New York, 60-70, 1969. 120 Scortichini, C.L. , and Reilly, C .N. , " Surface characterization of platinum electrodes using underpbtential deposition of H and Cu., Part Lpt(lOO)", J.Electroanal Chem, 139, 233-246, 1982. 121 Scortichini, C.L. , and Reilly, C .N. , " Surface characterization of platinum electrodes using underpotential deposition of H and Cu., Part II:(pt 110, pt 111)", J.Electroanal Chem , 139, 247-264, 1982. 122 Scortichini, C.L. , and Reilly, C .N. , " Surface characterization of platinum electrodes us-ing underpotential deposition of H and Cu., Part HI: Surface improvement of the flame-annealed pt(100) and pt ( l l l ) electrodes via potential cycling", J.Electroanal Chem, 139, 265-280, 1982. 123 Pletcher, D. , "Adsorption at solid electrodes", Electrochemistry, Vol 10, 51-65, 1985. 124 Wieckowski, A . , Electrochim Acta, 26, 1121, 1981. 125 Koklcinidis, G . , and Jannakoudakis, D. , J.Electroanal Chem, 153, 185, 1983. 126 Weickowski, P.Zelenay., Szlarczyk, M . , and Sobkowski, J., J . Electroanal Chem, 135, 285, 1982. 127 Nikolaeva, N . N . , Khazova, O.A. , and Vasilev, Yu.B. , Elecktrolriiniya, 16, 1227, 1980. References 243 128 Druzand, V . A . , Novikova, Z .N. , Zh.Fiz Khim, 54, 1818, 1980. 129 Clavilier, J., "Electrochemical surface characterization of platinum electrodes using ele-mentry electrosorption processes at basal and stepped surfaces", Electrochemical Surface Science, chapter 14, 202-216, 1988. 130 Clavilier, J., "The role of anion on the electrochemical behavior of a (111) platinum surface; A n unusual splitting of the voltammogram in the hydrogen region", Electroanal Chem, 107, 211-216, 1980. 131 Hoare, J.P., J.Electrochem Soc, 109, 858, 1962. 132 Bockris, J . O M . , Huq, A . K . M . S . , Proc Roy Soc, Ser.A, 237, 277, 1956. 133 Obrucheva, A . D . , Zh.Fiz Khim, 26, 1448, 1952. 134 ffickling, A . , Trans Faraday Soc, 41, 333, 1945. 135 Hoare, J.P., J . Electrochem Soc, 116, 612, 1969. 136 Hoare, J.P., ibid., 116, 1390, 1969. 137 Kalish T . V . , and Burshtein, R.Kh. , Dokl, Akad., Nauk USSR, 81, 1093, 1951; 88, 863, 1953. 138 Luk'yanycheva, V.I., and Bagotskii, V.S. , ibid., 155, 160, 1964. 139 Shirnizu, H . , Electrochim Acta, 14, 55, 1969. 140 Will, V . F . G . , Knorr, C .A.Z. , Elektrochim, 64, 258-269, 1960. 141 Sasaki, K. , and Nishigakierchi, " Cathodic reduction of adsorbed oxygen on platinum", Y. , Electrochim Acta ,16, 1099-1106, 1971. 142 Burke, L .D. , Roche, M.B .C . , " Enhanced oxide growth of a platinum electrode in acid under potential cycling condition", J. Electroanal Chem, 137, 175-181, 1972. References 244 143 Clavillier, J., Armond, D. , Wu, B.L., " Electrochemical study of the initial surface condition of platinum surfaces with (100) and (111) orientation", J. Electroanal Chem , 135, 159-166, 1982. 144 Koklowska, Angerstein.fi., Conway, B.E. , Sharp, W.B.A. , " The real condition of elec-trochemically oxidized platinum surfaces. Part I: Resolution of component processes", J . Electroanal Chem, 43, 9-36, 1973. 145 Hoare, J.P., "Encyclopedia of Electrochemistry of the elements", Bard, A . J . , Ed. , Marcell, Dekker., New York, Vol 11, chapter 5, 1974. 146 Reddy, A . K . N . , Genshaw, M . A . , Bockris, J . O . M . , J . Chem Phys, 148, 671-676, 1963. 147 Dickinson, T . , Povey, A . F . , Sherwood, P .M.A. , " X-ray photoelectron spectroscopic studies of oxide films on platinum and gold electrodes", J . Chem Soc Faraday Trans, 71, 298-311, 1975. 148 Allen, G.C. , Tucker, P .M. , Capon, A. , Parsons, R., " X-ray photoelectron spectroscopy of adsorbed oxygen and carbonaceous species on platinum electrodes'-, J . Electroanal Chem, 50, 335-343, 1974. 149 Kim, K.S., Winograd, N. , Davis, R.E. , " Electron spectroscopy of platinum -oxygen sur-faces and application to electrochemical studies", J . A m Chem Soc, 93, 6296-9297, 1971. 150 Hammond, J.S., Winogard, N.J., " XPS- spectroscopic study of potentiostatic and gal-vanostatic oxidation of platinum electrodes in H 2 SO 4 and HCIO4", Electroanal Chem, 78,55-69,1977. 151 Kover, L. , Ughelyl, C.S., Berenyl, D. , Varge, D . , Kadar, I., Kover, A . , Miller, J., " X-ray photoelectron spectroscopic investigation of electrochemically oxidized and reduced platinum surfaces", J . Electron Spectrosc Relat Phenom, 14, 201-214, 1978. References 245 152 Bancroft, G . M . , Adams, I., Coatsworth, K . L . , Bennewitz, C D . , Brown, J .D. , Westwood, W.D. , Anal Chem, 47, 586-588,. 1975. 153 Briggs, D. , Seah, M.P., "Practical surface analysis by Auger and X-ray photoelectron spectroscopy, John and Wiley, 1983. 154 Kasza, Rita.V., " A study of oxidized GaAs(lOO) surfaces by angle- dependent X-ray pho-toelectron spectroscopy", B.S., U.B.C. , 1-32, April 1990. 155 Briggs, D. , "Handbook of X-ray and ultraviolet photoelectron spectroscopy, Basic princi-ples of photoelectron spectroscopy", Heyden and Son Ltd., 1977. 156 Hochella, M.F . , In: Reviews in mineralogy, Ed. F .C . , Hawthorne, 18, 263, 1988. 157 Wagner, C D . , Riggs, W . M . , Davilenberg, L .E . , "Handbook of X-ray photoelectron spec-troscopy, Physical electronics devision", Perkin-Elmer. Corp., 1979. 158 Andrade, Joseph.D., "Biomedical polymer, Surface chemistry and physics, "X-ray photo-electron spectroscopy", Plenum Press, New York, Vol 1, 1985. 159 Andrade, Joseph.D., "Biomedical Polymers, Protein adsorption. The study of interfacial proteins and biomolecules by X-ray photoelectron spectroscopy, Vol 2, Plenum Press, New York, 1985. 160 DS 100 Users Guide, " The bigfit spectrum function", Leybold A G , 85-97. 161 Shirly, D . A . , Phys Rev, B5, 4709, 1972. 162 Hawthorne, F .C . , and Waychunas, G.A. , In: Reviews in Mineralogy, 18, 63, 1988. 163 Maddarns, W.F. , Appl Spec, 34, 245, 1980. 164 Gennett, Thomas., and Purdy, William.C, "Electrochemical sensors, Part 1: A review of their theory", American Laboratory, 60-64, February 1991. References 246 165 Sas User's Guide: Statistics, version 5 addition, Sas Institute Inc., Cary, N C , Ch(9-10), 1985. 166 Box, George. E.P., Hunter, William.G., Hunter, J.Stuart., " Statistics for experiments", John Wiley and Sons, 306-351, 1978. Appendix A Description of Electrodes This section includes an abbreviated description of the electrodes that were tested during this part of the research. The electrodes are identified by number, the surface cleaning solution, the concentration of the salt involved, (where c stands for Ba-glucose complex, g stands for glucose and b stands for barium chloride ), aliquat concentration, decanol concentration and the solution in which the electrodes were tested and finally the meter used. Meter # 1 was the Keithly 616 electrometer, meter #2 was the Keithly 617 electrometer, and finally meter # 3 was the Keithly 640 vibrating capacitor electrometer. 247 Appendix A. Description of Electrodes 248 Table A . l : Electrode Description Electrode Pt surface complex Aliquat Decanol solution meter# # cleaning amount (g) volume (ml) volume (ml) used 1 Acetone 0.3c 0.8 1.2 H20 1 2 Acetone 0.3c 0.8 1.2 H20 1 3 Acetone 0.3c 0.8 1.2 H20 1 4 Acetone 0.05c 0.1 1.0 H20 1 5 Acetone 0.05c 0.1 1.0 H20 1 6 Acetone 0.05c 0.1 1.0 H20 1 7 Acetone 0.6c 0.8 1.2 H20 1 8 Acetone 0.6c 0.8 1.2 H20 1 9 Acetone 0.6c 0.8 1.2 H20 1 10 Acetone 0.3c 0.8 1.2 Acetate buffer 1 11 Acetone 0.3c 0.8 1.2 Acetate buffer 1 12 Acetone 0.3c 0.8 1.2 Acetate buffer 1 13 Acetone 0.6c 0.8 1.2 Acetate buffer , 1 14 Acetone 0.6c 0.8 1.2 Acetate buffer 1 15 Acetone 0.6c 0.8 1.2 Acetate buffer 1 16 Acetone 1.06c 0.8 1.2 Citric buffer 1 17 Acetone 1.06c 0.8 1.2 Citric buffer 1 18 Acetone 1.06c 0.8 1.2 Citric buffer 1 19 HNO3 0.3c 0.8 1.2 Phosphate buffer 1 20 HN03 0.3c 0.8 1.2 Phosphate buffer 1 21 HN03 0.3c 0.8 1.2 Phosphate buffer 1 22 HNO3 0.6c 0.8 1.2 Phosphate buffer 1 23 HN03 0.6c 0.8 1.2 Phosphate buffer 1 24 HN03 0.6c 0.8 1.2 Phosphate buffer 1 25 ' HN03 0.1c 0.8 1.2 Phosphate buffer 1 26 HNO3 0.1c 0.8 1.2 Phosphate buffer 1 27 HN03 0.1c 0.8 1.2 Phosphate buffer 1 28 HNO3 0.8c 0.8 1.2 Phosphate buffer 1 29 HNO3 0.8c 0.8 1.2 Phosphate buffer 1 30 HNO3 0.8c 0.8 1.2 Phosphate buffer 1 Appendix A. Description of Electrodes 249 Table A . l (continued) Electrode Pt surface complex Aliquat Decanol solution meter# # cleaning amount (g) volume (ml) volume (ml) used 31 HN03 0.8c 0.8 1.2 Phosphate buffer 1 32 HNOz 0.8c 0.8 1.2 Phosphate buffer 1 33 HNOz 0.8c 0.8 1.2 Phosphate buffer 1 34 HNOz 0.6c 0.8 1.2 Phosphate buffer 1 35 HNOz 0.6c 0.8 1.2 Phosphate buffer 1 36 HNOz 0.6c 0.8 1.2 Phosphate buffer 1 37 HNOz 0.6c 0.8 1.2 Phosphate buffer 1 38 HNOz 0.6c 0.8 1.2 Phosphate buffer 1 39 HNOz 0.6c 0.8 1.2 Phosphate buffer 1 40 HNOz 0.8g 0.8 1.2 Phosphate buffer 1 41 HNOz 0.8g 0.8 1.2 Phosphate buffer 1 42 HNOz 0.8g 0.8 1.2 Phosphate buffer 1 43 HNOz 2.0c 0.8 1.2 Phosphate buffer 1 44 HNOz 2.0c 0.8 1.2 Phosphate buffer 1 45 HNOz 2.0c 0.8 1.2 Phosphate buffer 1 46 HNOz 0.8g 0.8 1.2 Phosphate buffer 1 47 HNOz 0.8g 0.8 1.2 Phosphate buffer 1 48 HNOz 0.8g 0.8 1.2 Phosphate buffer 1 49 HNOz 0-6g 0.8 1.2 Phosphate buffer 1 50 HNOz 0.6g 0.8 1.2 Phosphate buffer 1 51 HNOz 0.6g 0.8 1.2 Phosphate buffer 1 52 HNOz 0.8b 0.8 1.2 Phosphate buffer 1 53 HN03 0.8b 0.8 1.2 Phosphate buffer 1 54 HNOz 0.8b 0.8 1.2 Phosphate buffer 1 55 HNOz 0.8b 0.8 1.2 Phosphate buffer 1 56 HNOz 0.8b 0.8 1.2 Phosphate buffer 1 57 HNOz 0.0 0.8 1.2 Phosphate buffer 1 58 HNOz 0.0 0.8 1.2 Phosphate buffer 1 59 HNOz 0.0 0.8 1.2 Phosphate buffer 1 60 HNOz 0.0 0.8 1.2 Phosphate buffer 1 Appendix A. Description of Electrodes 250 Table A . l (continued) Electrode Pt surface complex Aliquat Decanol solution meter# # cleaning amount (g) volume (ml) volume (ml) used 61 HNOz 0.0 0.8 1.2 Phosphate buffer 1 62 HN03 0.0 0.8 1.2 Phosphate buffer 1 63 HNOz 0.3b 0.8 1.2 Phosphate buffer 1 64 HNOz 0.3b 0.8 1.2 Phosphate buffer 1 65 HNOz 0.3b 0.8 1.2 Phosphate buffer 1 66 HNOz 0.3b 0.8 1.2 Phosphate buffer 1 67 HNOz 0.3b 0.8 1.2 Phosphate buffer 1 68 HNOz 0.3b 0.8 1.2 Phosphate buffer 1 69 HNOz 0.1c 0.8 1.2 Phosphate buffer 1 70 HNOz 0.0 0.8 1.2 Phosphate buffer 1 71 HNOz 0.8c 0.8 1.2 Phosphate buffer 1 72 HNOz 0.3c 0.8 1.2 Phosphate buffer 2 73 HNOz 0.3c 0.8 1.2 Phosphate buffer 2 74 HNOz 0.3c 0.8 1.2 Phosphate buffer 2 75 HNOz 0.3c 0.8 1.2 Phosphate buffer 2 76 HNOz 0.3c 0.8 1.2 Phosphate buffer 2 77 HNOz 0.3c 0.8 1.2 Phosphate buffer 2 78 HNOz 0.3c 0.8 1.2 Phosphate buffer 2 79 HNOz 0.3c 0.8 1.2 Phosphate buffer 2 80 HNOz 0.3c 0.8 1.2 Phosphate buffer 2 81 HNOz 0.3c 0.8 1.2 Phosphate buffer 2 82 HNOz 0.3c 0.8 1.2 Phosphate buffer 2 83 HNOz 0.3c 0.8 1.2 Phosphate buffer 2 84 HNOz 0.0 0.8 1.2 Phosphate buffer 2 85 HNOz 0.0 0.8 1.2 Phosphate buffer 2 86 HNOz 0.0 0.8 1.2 Phosphate buffer 2 87 HNOz 0.0 0.8 1.2 Phosphate buffer 2 88 HNOz 0.0 0.8 1.2 Phosphate buffer 2 89 HNOz 0.0 0.8 1.2 Phosphate buffer 2 90 HNOz 0.0 0.8 1.2 Phosphate buffer 2 Appendix A. Description of Electrodes 251 Table A . l (continued) Electrode Pt surface complex Aliquat Decanol solution meter# # cleaning amount (g) volume (ml) volume (ml) used 91 HNOz 0.0 0.8 1.2 Phosphate buffer 2 92 HNOz 0.0 0.8 1.2 Phosphate buffer 2 93 HNOz 0.0 0.8 1.2 Phosphate buffer 2 94 HNOz 0.0 0.8 1.2 Phosphate buffer 2 95 HNOz 0.0 0.8 1.2 Phosphate buffer 2 96 HNOz 0.0 0.0 1.2 Phosphate buffer 2 97 HNOz 0.0 0.0 1.2 Phosphate buffer 2 98 HNOz 0.0 0.0 1.2 Phosphate buffer 2 99 HNOz 0.0 0.0 1.2 Phosphate buffer 2 100 HNOz 0.0 0.0 1.2 Phosphate buffer 2 101 HNOz 0.0 0.0 1.2 Phosphate buffer 2 102 HNOz 0.0 0.0 1.2 Phosphate buffer 2 103 HNOz 0.0 0.0 1.2 Phosphate buffer 2 104 HNOz 0.0 0.0 0.0 Phosphate buffer 2 105 HNOz 0.0 0.0 0.0 Phosphate buffer 2 106 HNOz 0.0 0.0 0.0 Phosphate buffer 2 107 HNOz 0.3c 0.8 1.2 Phosphate buffer 2 108 HNOz 0.3c 0.8 1.2 Phosphate buffer 2 109 HNOz 0.3c 0.8 1.2 Phosphate buffer 2 110 HNOz 0.0 0.0 0.0 Phosphate buffer 2 111 HNOz 0.0 0.0 0.0 Phosphate buffer 2 112 HNOz 0.0 0.0 0.0 Phosphate buffer 2 113 HNOz 0.0 0.0 0.0 Phosphate buffer 2 114 HNOz 0.0 0.0 0.0 Phosphate buffer 2 115 HNOz 0.0 0.0 0.0 Phosphate buffer 2 116 HNOz 0.0 0.8 1.2 Phosphate buffer 2 117 HNOz 0.0 0.8 1.2 Phosphate buffer 2 118 HNOz 0.0 0.8 1.2 Phosphate buffer 2 119 HNOz 0.0 0.0 0.0 Phosphate buffer 2 120 HNOz 0.0 0.0 0.0 Phosphate buffer 2 Appendix A. Description of Electrodes 252 Table A . l (continued) Electrode Pt surface complex Aliquat Decanol solution meter# # cleaning amount (g) volume (ml) volume (ml) used 121 HNOz 0.0 0.0 . 0.0 Phosphate buffer 2 122 HN03 0.0 0.0 0.0 Phosphate buffer 2 123 HN03 0.0 0.0 0.0 Phosphate buffer 2 124 HNOz 0.0 0.0 0.0 Phosphate buffer 2 125 HNOz 0.0 0.0 0.0 Phosphate buffer 2 126 HNOz 0.0 0.0 0.0 Phosphate buffer 2 127 HNOz 0.3c 0.0 1.2 Phosphate buffer 2 128 HNOz 0.3c 0.0 1.2 Phosphate buffer 2 129 HNOz 0.3c 0.0 1.2 Phosphate buffer 2 130 HNOz 0.0 0.0 0.0 Phosphate buffer 2 131 HNOz 0.0 0.0 0.0 Phosphate buffer 2 132 HNOz 0.0 0.0 0.0 Phosphate buffer 2 133 HNOz 0.0 0.0 0.0 Phosphate buffer 2 134 HNOz 0.0 0.0 0.0 Phosphate buffer 2 135 HNOz 0.3c 0.0 1.2 Phosphate buffer 2 136 HNOz 0.3c 0.0 1.2 Phosphate buffer 2 137 HNOz 0.3c 0.0 1.2 Phosphate buffer 2 138 HNOz 0.0 0.0 0.0 Phosphate buffer 2 139 HNOz 0.0 0.0 0.0 Phosphate buffer 2 140 HNOz 0.0 0.0 0.0 Phosphate buffer 2 141 HNOz 0.3c 0.0 1.2 Phosphate buffer 2 142 HNOz 0.0 0.0 0.0 Phosphate buffer 2 143 HNOz 0.0 0.0 0.0 Phosphate buffer 2 144 HNOz 0.0 0.0 0.0 Phosphate buffer 2 1. Glass Electrodes were prepared and identified as electrode No.145-156 A p p e n d i x B S T A T I S T I C A L A N A L Y S I S " S A S " is a software system for data analysis [165]. The regression equation predicts a response variable as a function of regressor, variables and parameters, adjusting the parameters such that a measure of fit is optimized. For example, the equation for the ith observation might be: Yi = Po + PiXi + Et (B.l) where: Yi= response variable Xi= regressor variable Po and Pi = unknown parameters to be estimated Ei= error term The statistical analysis used is based on a general linear model(GLM) which uses the method of least squares to fit the given parameters. G L M handles classification variables, which have discrete levels, as well as continuous variables, which measure quantities. Analysis of variance models require independent variables that identify classification levels. In the SAS system these are called class variables. The values of a class variable are called levels. G L M is used in most unbalanced situations, that is, models where there are unequal numbers of observations for the different combinations of class variables specified in the model statement. B.0.1 Input D a t a The input, program indicates: Proc G L M ; 253 Appendix B. STATISTICAL ANALYSIS 254 Class (Independent variables); Model(Dependent variables) = (Independent variables) * Interactions; Lsmeans effects /diff solution; proc reg; Model (Dependent variables) = (Independent variables) /p r cli elm; output out=C ; p=pred r=resid; run; proc plot data=C; plot pred*resid; B.0.2 T y p e I, II, III, I V S A S Tests Type I sums of squares, also called sequential sums of squares, are the incremental improvement in error sums of squares as each effect is added to the model. The type II tests can also be calculated by comparing the error sums of squares for subset models. The type II sums of squares are the reduction in error sums of squares due to adding the term after all other terms have been added to the model except terms that contain the effect being tested. Type III and type IV sums of square, sometimes referred to as partial sums of squares, are considered by many to be the most desirable tests. These sums of squares can not in general be computed by comparing model sums of square from several models using GLM's parameterization. They are computed by constructing an estimated hypothesis matrix land then computing the steady state associated with the hypothesis L/3 = 0.0 B.0.3 M a n o v a Statement If the model statement includes more than one dependent variable, additional multivariate statistics can be requested with the Manova statement. When a Manova statement appears, Appendix B. STATISTICAL ANALYSIS 255 G L M enters a multivariate mode with respect to the handling of missing values: observations with missing independent or dependent variables are excluded from the analysis. Crossed Effect G L M reorders the terms to correspond to the order of the variables in the class statement; the order of the column is such that the right most variables in the cross index faster than the left most variables. Nested Effect The nesting operator in G L M is more a notational convenience than an operation distinct from crossing. Nested effects are characterized by the property that the nested variables never appear as main effects. The order of the variables within nesting parentheses is made to correspond to the order of these variables in the class statement. The order of the column is such that variables outside the parentheses index faster than those inside the parentheses. The next section will explain the output "Description of the variables", generated from SAS. B.0.4 Output Results The program asks for: P- Calculates predicted values from the input data and the estimated model. The printout includes the observation number, the first D variables if specified, the actual and predicted values, and the residuals. R- Request that the residual be analyzed. The printed output includes every thing requested by the p option plus the standard errors of the predicted and residual values, the studentized residual, and Cook's D statistic to measure the influence of all observations on the parameter estimates. C L M - Prints the 95% upper- and lower- confidence limits for the expected value of the dependent variable(mean) for each observation. It takes into account only the variation in the Appendix B. STATISTICAL ANALYSIS 256 parameter estimates, not the variation in the error term. CLI- Requests the 95% upper- and lower- confidence limits for an individual predicted value. The confidence limits reflect variations in the error, as well as variation in the parameter estimates. The G L M produces the following printouts: The overall analysis of variance table breaks down the corrected total sum of squares for the dependent variable into the portion attributed to the model and the portion attributed to the error. The mean square error (MS error), is an estimate of sigma*2, the variance of the true errors. The F value is the ratio produced by dividing MS model by MS error. It tests how well the model as a whole accounts for the dependent variable's behavior. An F test is a joint test in that all parameters except the intercept are zero. A small significance probability, PR>F , indicates that some linear function of the parameters is significantly different from zero. R-square, i t 2 , measures how much variation in the dependent variable can be accounted for by the model. R-square, which can range from 0 to 1, is the ratio of the sums of squares for the model divided by the sum of squares for the corrected total. In general, the larger the value of R-square, the better the model's fit. C .V. , The coefficient of variation, which describes the amount of variation in the population, is 100 times the standard deviation estimate of the dependent variable, root M S E , divided by the mean. Root-MSE, estimates the standard deviation of the dependent variable and equals the square root of MS (error). Mean, Sample mean of the dependent variable. The type I ss measures incremental sums of squares for the model as each variable is added. The type III ss is the sum of squares that results when that variable is added last to the model. Appendix B. STATISTICAL ANALYSIS 257 This section of the output gives the estimates for the model parameters, the intercept and the coefficients. 1. T for Ho'- Parameter=0 is the student's t value for testing the null hypothesis that the parameter equals zero. 2. The significance level, PR> |T|, is the probability of getting a larger value of t if the parameter is truely equal to zero. A very small value for this probability leads to the conclusion that the independent variable contributes significantly to the model. 3. The STD Error of estimate is the standard errors of the estimate of the true value of the parameter. 4. Adj R-square, is a version of R-square that has been adjusted for degrees of freedom for each model selected. It is calculated as: l-(n-i)(l- R2)/(n- P) (B.2) • n=number of observations. • p=number of parameters including intercept. • R-square= 1-SSE/TSS • TSS=total sum of squares corrected for the mean for the dependent variable. • SSE=error sum of squares. • i=l if there is an intercept otherwise is zero. The block design was completely randomized [166]. All the sensors were made at the same ambient conditions, although they were not tested simultaneously. This section will show the statistical analyses that were'gathered and were.explained briefly in the thesis. The analyses were done to find the degree of significancy on amount of Ba-glucose complex used, the amount of BaCl2 , the effect of pH changes on the electrode response, the effect of life time, and the effect of the intervariability of the sensor, and the glucose addition to the measuring solution, and finally the effect of changing the measuring solution while testing the electrode. Appendix B. STATISTICAL ANALYSIS 258 SAS GENERAL LINEAR MODELS PROCEDURE CLASS LEVEL INFORMATION CLASS LEVELS VALUES COMP 3 0 1 2 SEN 12 1 2 3 4 5 6 7 8 0 10 11 12 NUMBER OF OBSERVATIONS IN DATA SET = 205 i r u n sas:sas sercom=-log s p r i DATA COATED W; i n f i l e cards; INPUT VOLTAGE TIME SEN COMP PROC GLM; CLASS 90MP SEN; MODEL V0LTAGE-=C0MP SEN (COMPK TIME; PROC GLM; CLASS SEN COMP; MODEL VOLTAGE = COMP SEN(COMP) TIME/ SOLUTION; LSMEANS COMP SEN(COMP) / STDERR PDIFF; proc reg; model voltage=sensor comptfXJp r, c l i elm; output out=c p=pred r^t'esla. CARDS; 765.0 785.0 777.0 765.0 694.0 605.0 525.0 482.0 464.0 452. 445 445 445. 437 325 805.0 815.0 808.0 807.0 802.0 703.0 655.0 614.0 614.0 575.0 548.0 489.0 471.0 467.0 457.0 456 454 451 437. 318. 813, 829 836 830 818.0 787.0 778.0 777.0 764.0 752.0 733.0 723.0 709.0 702.0 701.0 701.0 695.0 693.0 688.0 686.0 0.00 1.40 4.90 6.37 17.40 34.30 51.03 68.40 64.55 110.38 129.50 151.22 158.85 208.48 651.12 0.00 1.18 5.42 6.00 17.72 34.00 51.32 52.43 68.77 84.00 110.58 130.00 150.85 159.15 160.52 162.20 165.60 166.25 208.87 651.70 0.00 1.62 4.52 6.73 17.22 34.53 50.73 52.85 67.83 84.27 111.00 129.17 151.67 157.77 159.00 167.00 209.28 210.00 213.05 215.42 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 . 0 t) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 = - o u t par =;si ze =6000k run; plot data"=c; proc p l o t pred*resid; 580.0 651.83 3 0 380.0 835.50 7 769.0 0.00 4 0 380.0 838.00 7 772.0 1.37 4 0 379.0 840.00 7 740.0 3.62 4 0 767.0 0.00 8 687.0 15.43 4 0 772.0 1.18 8 675.0 24.33 • 4 0 775.0 6.28 8 673.0 32.88 4 0 772.0 27.00 8 632.0 49.00 4 0 764.0 45.02 8 597.0 68.00 4 0 691.0 101.43 8 559.0 100.32 4 0 571.0 156.83 8 549.0 121.43 4 0 463.0 242.12 8 539.0 136.82 4 0 417.0 368.68 8 532.0 149.83 4 0 381.0 580.55 8 531.0 167.97 4 0 378.0 588.55 8 469.0 324.70 4 0 347.0 812.70 8 790.0 0.00 5 0 346.0 826.97 8 785.0 1.58 5 0 346.0 832.05 8 754.0 2.65 5 0 346.0 834.28 8 629.0 15.75 5 0 346.0 840.32 8 612.0 24.03 5 0 700.0 0.00 9 597.0 33.23 5 0 670.0 0.77 9 570.0 48.73 5 0 606.0 6.53 9 545.0 68.50 5 0 530.0 26.40 9 516.0 99.53 5 0 499.0 45.57 9 499.0 121.58 5 0 457.0 100.47 9 499.0 136.47 5 0 448.0 157.47 9 478.0 150.28 5 0 447.0 241.75 9 467.0 .167.55 5 0 440.0 369.08 9 424.0 324.90 5 0 445.0 580.30 9 788.0 0.00 6 0 442.0 588.87 9 727.0 1.00 6 0 439.0 811.80 9 699.0 4.00 6 0 439.0 827.33 9 701.0 5.00 6 0 438.0 830.08 9 689.0 15.13 6 0 437.0 835.88 9 661.0 24.70 6 0 438.0 840.58 9 630.0 32.60 6 0 734.0 0.00 10 2 567.0 49.17 ' 6 0 746.0 2.50 10 2 519.0 67.60 6 0 748.0 3.00 10 2 470.0 99.83 6 0 766.0 18.97 10 2 454.0 122.57 6 0 778.0 32.52 10 2 443.0 136.00 6 0 784.0 44.63 10 2 446.0 150.77 6 0 782.0 86.92 10 2 439.0 167.18 6 0 775.0 122.75 10 2 404.0 325.18 6 0 771.0 146.90 10 2 400.0 335.22 6 0 769.0 161.00 10 2 776.0 0.00 7 1 766.0 197.32 10 2 780.0 1.58 7 1 761.0 270.75 10 2 767.0 6.00 7 1 761.0 296.68 10 2 756.0 26.83 7 1 760.0 336.02 10 2 755.0 45.33 7 1 760.0 342.03 10 2 690.0 100.87 7 1 760.0 369.67 10 2 594.0 157.17 7 1 758.0 394.77 10 2 510.0 242.42 7 1 688.0 725.17 10 2 496.0 269.45 7 1 776.0 0.00 11 2 469.0 368.47 7 1 765.0 2.00 11 2 425.0 581.20 7 1 766.0 3.38 11 2 424.0 588.28 7 1 744.0 19.73 11 2 380.0 813.00 7 1 738.0 32.13 11 2 379.0 826.53 7 1 734.0 45.00 11 2 380.0 831.15 7 1 729.0 68.55 11 2 725.0 711.0 698.0 693.0 674.0 643.0 639.0 642.0 641.0 642.0 641.0 558.0 764.0 750.0 758.0 776.0 780.0 782.0 785.0 782.0 779.0 778.0 774.0 761.0 757.0 752.0 751.0 751.0 743.0 543.0 86.38 123.08 146.35 161.43 197.03 271.28 296.40 331.32 341.68 370.10 394.40 725.62 0.00 2.00 3.72 18.50 32.85 44.32 87.28 122.42 147.27 160.65 197.77 270.35 297.27 330.43 342.42 351.05 395.15 724.72 1 i 11 11 11 11 11 11 11 11 11 11 11 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 Table B . l : The Effect of BaCl2 Concentration on Electrode Response Appendix B. STATISTICAL ANALYSIS 259 Table B.2 : The Effect of BaCl2 Concentration on Electrode Response (continues). SAS GENERAL LINEAR MODELS PROCEDURE DEPENOENT VARIABLE: VOLTAGE SOURCE DF MODEL 13 ERROR 192 CORRECTED TOTAL 204 SUM OF SQUARES 3 2 2 0 1 9 9 - 6 1 9 3 3 5 6 0 1 2 2 6 8 8 8 . 0 1 9 6 8 8 8 0 4 4 4 7 0 8 7 . 6 3 9 0 2 4 4 0 MEAN SOUARE 2 6 8 3 4 9 . 9 6 8 2 7 7 9 7 6 3 9 0 . 0 4 1 7 6 9 2 1 F VALUE 4 2 . 0 0 PR > F 0 . 0 0 0 1 ROOT MSE 7 9 . 9 3 7 7 3 6 8 3 R-SOUARE ; v . 0 . 7 2 4 1 U . ?F_?e VOL 1 .•(...; 625 10731707 SOURCE COMP SEN(COMP) TIME DF 1 TYPE I SS 12O0212.8S9O2439 528745.75516708 1491241.00514413 0 . 0 0 0 1 0 . 0 0 0 1 0 . 0 0 0 1 TYPE III SS 9 4 0 7 5 0 . 4 7 8 3 2 8 4 5 5 9 1 9 4 4 . 2 9 9 4 7 2 7 5 1 4 9 1 2 4 1 . 0 0 5 1 4 4 1 3 F VALUE 7 3 . 6 1 1 0 . 2 9 2 3 3 . 3 7 PR > F 0 . 0 0 0 1 0 . 0 0 0 1 0 . 0 0 0 1 GENERAL LINEAR MOOELS PROCEDURE OEPENDENT VARIABLE: VOLTAGE SOURCE DF MODEL 12 ERROR 192 CORRECTED TOTAL 204 SUM OF SOUARES 3220199.61933560 1226888.01968860 4447087.63902440 MEAN SOUARE 268349.96827797 6390.04176921 F VALUE 4 2 . 0 0 PR > F 0 . 0 0 0 1 ROOT MSE 7 9 . 9 3 7 7 3 5 8 3 R* SQUARE C V . 0 . 7 2 4 1 1 4 1 2 . 7 8 7 8 V O L T A G E MEAN 6 2 5 . 1 0 7 3 1 7 0 7 SOURCE COMP SENICOMP) TIME DF T Y P E I SS 1200212.85902439 528745.75516708 1491241.00514413 93.91 9.19 233.37 0 . 0 0 0 1 0 . 0 0 0 1 O .O0D1 TYPE III SS 9 4 0 7 5 0 . 4 7 8 3 2 8 4 5 5 9 1 9 4 4 . 2 9 9 4 7 2 7 5 1 4 9 1 2 4 1 . 0 0 5 1 4 4 1 3 F V A L U E 7 3 . 6 1 1 0 . 2 9 2 3 3 . 3 7 PR > F 0 . 0 0 0 1 0 . 0 0 0 1 0 . 0 0 0 1 PR > | T | INTERCEPT COMP SEN(COMP) ESTIMATE PARAMETERS 831.98130037 B 42.61 0. 0001 -228.81881169 B -8.29 0 . 0001 -183.38985074 B -6 . 56 0 . 0001 0.00000000 B 0 2.25967817 B O.OB 0 . !9374 0 31.81052943 B 1.19 0. 2370 0 190.73739815 B 7. 19 0 .0001 0 54.09874017 6 1.85 0. 0660 0 14.14390957 B 0 . 4 6 0 .6293 0 0 . 0 0 0 0 0 0 0 0 B 1 55.05051959 B 2.00 0 !04S5 1 37.96447275 B 1.34 0 . 1808 1 0.00000000 B 2 6.13125219 B 0.23 0 ! 8183 2 -63.36582671 B -2.41 0 .0169 2 0.00000000 8 -0.39959622 -1s!2B 0 !oooi STO ERROR OF ESTIMATE 19.52664901 27.59028626 27.93780749 28!73243068 26.81887676 26.54019223 29.25552762 29.25554394 27^47006407 28.26225795 26!e4593403 26.29339360 GENERAL LINEAR MOOELS PROCEOURE LEAST SQUARES MEANS VOLTAGE LSMEAN 571.547935 599.140164 732.448846 STD ERR LSMEAN PROS > |T| HO:LSMEAN=0 PROB > |T| HO: LSMEAN(I)=LSMEAN(J) I/J 1 2 3 11:38 MONOAY. DECEMBER 2 4 , 1990 8.486122 12.430326 10.783782 0.0 0.0 0.0 0.0841 0 . 0 0 0 1 0.0001 0.0001 0.0841 0.0001 TO ENSURE OVERALL PROTECTION LEVEL. ONLY PROBABILITIES ASSOCIATED WITH PRE-PLANNED COMPARISONS SHOULD BE USED. SEN COMP VOLTAGE STD ERR PROB > |T| LSMEAN LSMEAN LSMEAN H0:LSMEAN-0 NUMBER 1 0 524.965904 20.772209 0.0001 1 2 0 554.616755 18.002983 0.0001 2 3 0 713.443624 17.548081 0.0 3 4 0 576.804966 21.578517 0.0001 4 5 0 536.850136 21.578759 0.0001 5 6 0 522.706226 20.173615 0.0001 6 7 1 623.185707 19.612605 0.0001 7 6 1 608.099660 20.593969 0.0001 8 9 1 568.135187 20.593543 0.0001 9 10 2 757.658290 18.84(604 0.0 10 11 2 688.159211 18.341257 0.0 11 12 2 751.525038 18.842027 0.0 12 PROB > | T | HO: LSMEAN1I)=LSMEAN(J) I/J 1 2 3 4 5 6 7 8 9 10 11 12 1 0 .2805 0. .0001 0.0827 0 .6896 0 .9374 0 .0009 0 .0088 0 . 1470 0, 0001 0 .0001 0.0001 2 0.2805 0. .0001 0 . 4 2 4 9 0 .5269 0 .2370 0 .01 19 0, .0645 0 .6240 0. .0001 0. .0001 0.0001 3 0.0001 0 !oooi 0.0001 0, .0001 0 .0001 0 .0009 0 ,0001 0 .0001 0. 0675 0. ,3200 0.1406 4 0.0827 0 .4249 0 0001 0 . 1878 0 .0660 0 . 1203 0 ,3353 0 ,7753 0. 0001 0. .0001 0.0001 5 0.6896 0 .5269 0 .0001 0.1878 0 .6293 0 .0041 0, 0235 0 .3036 0 ,0001 0. .0001 0.0001 6 0.9374 0 .2370 0. .0001 0.0660 0 !6293 0 .0006 0 .0049 0 .1227 0. .0001 0 .0001 0.0001 7 0.0009 0. .0119 0. .0009 0.1203 0 .0041 0. !o006 0, .534 7 0 .0465 0. 0001 0. .0167 0.0001 8 0.0088 0 .0645 0. ,0001 0.3353 0 .0235 0 .0049 0 !5347 0 .1808 0. 0001 0 .0034 0.0001 9 0. 1470 0 .8240 0 .0001 0.7753 0 .3036 0 . 1227 0 .0485 0 . 1808 0 0001 0 .0001 0.0001 10 0.0001 0 .0001 0. 0875 0.0001 0 ,0001 0 .0001 0 ,0001 0, ,0001 0 .0001 0 . 0089 0.8183 11 0.0001 0. .0001 0. 3200 0.0001 0 .0001 0 ,0001 0 ,0167 0 .0034 0 0001 0. 0089 0.0169 12 0.0001 0. 0001 0. 1406 0.0001 0. 0001 0. ,0001 0. 0001 0 0001 0 0001 a. 8183 0, 0169 N 0 1 E : TO ENSURE OVERALL PROTECTION LEVEL. ONLY PROBABILITIES ASSOCIATED WITH PRE-PLANNED COMPARISONS Appendix B. STATISTICAL ANALYSIS 260 PROC GLM; CLASS SEN COMP; MODEL VOLTAGE = COMP SEN(COMP) TIME TIME«SEN TIME * SEN(COMP) /SOLUTION; LSMEANS COMP SEN(COMP) / STDERR PDIFF; p r o c r e g ; model v o l t a g e = s e n comp/p r c l i elm; out p u t out=c p=pred r = r e s i d ; run; p r o c p l o t data=c; p l o t pred«resid; Stun s a s : s a s sercom=-log s p i i n t = - o u t par =; s i z e = 6000lc DATA COATED W; i n f i l e c a r d s ; INPUT VOLTAGE TIME SEN COMP; CARDS; 642.0 356. 20 4 1 765.0 0 00 1 0 642.0 361. 72 4 1 673.0 32 88 1 0 740.0 0. 00 5 1 632.0 49 00 1 0 748.0 1. 00 5 1 597.0 68 00 1 0 739.0 5. 92 5 1 559.0 100 32 1 0 730.0 22. 72 5 1 549.0 121 43 1 0 729.0 31. 72 5 1 539.0 136 82 1 0 730.0 50. 42 5 1 532.0 149 83 1 0 723.0 80. 23 5 1 531 .0 167 97 1 0 713.0 98. 52 5 1 469.0 324 70 1 0 684.0 134. 82 s 1 790.0 0 00 2 0 677.0 144. 35 5 1 785.0 1 58 2 0 663.0 170. 40 5 1 754.0 2 65 2 0 648.0 191. 83 5 1 629.0 15 75 2 0 601.0 261. 17 5 1 612.0 24 03 2 0 597.0 281. 00 5 1 597.0 33 23 2 0 580.0 310 02 5 1 570.0 48 73 2 0 557.0 355 63 5 1 545.0 68 50 2 0 556.0 362 00 5 1 516.0 99 53 2 0 712.0 0 00 6 1 499.0 121 58 2 0 682.0 1 20 6 1 499.0 136 47 2 0 681.0 5 67 6 1 478.0 150 28 2 0 685.0 32 10 6 1 467.0 167 55 2 0 696.0 50 00 6 1 424.0 324 90 2 0 703.0 80 45 6 1 788.0 0 00 3 0 703.0 98 17 6 1 727.0 1 00 3 0 701.0 135 35 6 1 699.0 4 00 3 0 696.0 170 97 6 1 701.0 5 00 3 0 679.0 191 52 6 1 689.0 15 13 3 0 608.0 261 58 6 1 661.0 24 70 3 0 618.0 282 00 6 1 630.0 32 60 3 0 606.0 309 40 6 1 567.0 49 17 3 0 582.0 356 00 6 1 519.0 67 60 3 0 582.0 362 28 6 1 470.0 99 83 3 0 802.0 0 00 7 2 454.0 122 57 3 0 810.0 1 28 7 2 443.0 136 00 3 0 807.0 4 25 7 2 446.0 150 77 3 0 799.0 35 .78 7 2 439.0 167 18 3 0 791.0 51 .75 7 2 404.0 325 18 3 0 793.0 63.72 7 2 400.0 335 22 3 0 781.0 93.75 7 2 794.0 0 00 4 1 781.0 95.53 7 2 805.0 1 43 4 1 773.0 140 .82 7 2 798.0 4 75 4 1 778.0 166 .02 7 2 795.0 5 33 4 1 770.0 195 .77 7 2 757.0 32 45 4 1 768.0 197 .82 7 2 742.0 49 07 4 1 767.0 198 .77 7 2 730.0 80 75 4 1 767.0 200 .42 7 2 730.0 98 05 4 1 796.0 0 .00 8 2 729.0 135 08 4 1 787.0 1 .02 8 2 722.0 144 00 4 1 793.0 36 .12 8 2 726.0 170 62 4 1 791.0 52 .08 8 2 718.0 192 07 4 1 789.0 53 .57 8 2 688.0 260 68 4 1 790.0 55 .10 8 2 671.0 282 10 4 1 788.0 58.92 8 2 661.0 309 12 4 1 786.0 63 .20 8 2 792.0 94 .10 8 2 792.0 180 .00 8 2 792.0 189 .58 8 2 792.0 195 .43 8 2 792.0 200 70 8 •2: The Effect of Ba-Glucose Complex on Electrode Response Appendix B. STATISTICAL ANALYSIS 261 1 1 : 4 2 F R I D A Y , J A N U A R Y 1 1 . 1 9 9 1 D£P V A R I A B L E : V O L T A G E A N A L Y S I S O F V A R I A N C E M O D E L E R R O R C T O T A L 2 1 1 3 1 1 5 R O O T M S E O E P M E A N C V . S U M O F . S Q U A R E S 7 6 3 5 1 9 . 5 2 6 9 0 6 7 7 . 3 4 1 4 5 4 3 9 6 . 8 6 7 6 . 1 9 1 8 1 6 7 2 . 5 3 4 5 1 1 . 6 2 6 4 4 P A R A M E T E R E S T I M A T E S M E A N S Q U A R E 3 8 1 7 5 9 . 7 6 6 1 1 3 . 9 5 8 7 7 R - S Q U A R E A O J R - S Q V A R I A B L E D F I N T E R C E P S E N C O M P F V A L U E 8 2 . 4 4 1 0 . 5 2 5 0 0 . 5 1 6 6 P A R A M E T E R E S T I M A T E 1 6 2 0 . 7 3 2 9 1 1 - 1 9 . 3 2 5 2 0 3 0 2 1 1 5 7 . 9 6 7 6 2 P R O B > F 0 . 0 0 0 1 S T A N D A R D E R R O R 2 4 . 2 1 1 5 8 4 8 5 S . 7 4 8 7 2 4 8 6 2 7 . 8 3 4 1 2 6 7 4 T F O R H O : P A R A M E T E R S ) 2 5 . 8 3 8 - 1 . 9 8 2 5 . 6 7 5 P R O B > | T | 0 . 0 0 0 1 0 . 0 4 9 9 0 . 0 0 0 1 1 2 3 4 5 6 7 8 6 1 0 11 1 2 1 3 1 4 1 5 1 6 17 1 8 1 9 2 0 2 1 2 2 2 3 P R E D I C T S T D E R R L O W E R 9 5 X U P P E R 9 5 X L O W E R 9 5 X U P P E R 9 5 X S T D E R R S T U D E N T A C T U A L V A L U E P R E D I C T M E A N M E A N P R E D I C T P R E D I C T R E S I D U A L R E S I D U A L R E S I D U A L 7 6 5 0 6 0 1 . 4 1 8 2 2 0 0 5 6 9 3 6 3 3 5 4 4 3 2 7 5 9 6 1 6 3 . 8 7 6 4 9 1 0 2 1 3 8 7 6 7 3 0 6 0 1 . 4 1 8 2 2 0 0 5 6 9 3 6 3 3 5 4 4 3 2 7 5 9 8 7 1 . 5 9 2 3 7 8 4 9 1 0 0 9 3 6 0 6 3 2 0 8 0 1 . 4 16 2 2 0 0 5 6 9 3 6 3 3 5 4 4 3 2 7 5 9 6 3 0 . 5 9 2 3 7 6 4 9 1 0 0 3 9 9 9 5 9 7 0 6 0 1 . 4 16 2 2 0 0 5 6 9 3 6 3 3 5 4 4 3 2 7 5 9 6 - 4 . 4 0 7 7 7 6 4 9 1 0 - 0 0 5 7 6 5 5 9 0 6 0 1 . 4 16 2 2 0 0 5 6 9 3 6 3 3 5 4 4 3 2 7 5 9 6 - 4 2 . 4 0 7 7 7 6 4 9 1 0 - 0 5 5 4 4 5 4 9 0 6 0 1 . 4 1 8 2 2 0 0 5 6 9 3 6 3 3 5 4 4 3 2 7 5 9 6 - 5 2 . 4 0 7 7 7 6 4 9 1 0 - 0 6 8 5 1 5 3 9 0 6 0 1 . 4 1 8 2 2 0 0 5 6 9 3 6 3 3 5 4 4 3 2 7 5 9 8 - 6 2 . 4 0 7 7 7 6 4 9 1 0 - 0 8 1 5 9 5 3 2 0 6 0 1 . 4 1 6 2 2 0 0 5 6 9 3 8 3 3 5 4 4 3 2 7 5 9 6 - 6 9 . 4 0 7 7 7 6 4 9 1 0 - 0 9 0 7 4 5 3 1 0 6 0 1 . 4 16 2 2 0 0 5 6 9 3 6 3 3 5 4 4 3 2 7 5 9 6 - 7 0 . 4 0 7 7 7 6 4 9 1 0 - 0 9 2 0 5 4 6 9 0 6 0 1 . 4 16 2 2 0 0 5 8 9 3 6 3 3 5 4 4 3 2 7 5 9 6 - 1 3 2 . 4 7 6 4 9 1 0 - 1 7 3 1 0 7 9 0 0 5 6 2 . 1 11 4 0 4 0 5 5 9 5 6 0 4 7 4 2 5 5 7 3 8 6 2 0 7 . 9 7 7 3 5 5 7 2 6 8 7 8 7 8 5 0 5 6 2 . 1 1 1 4 0 4 0 5 5 9 5 6 0 4 7 4 2 5 5 7 3 8 6 2 0 2 . 9 7 7 3 5 5 7 2 6 2 3 2 7 5 4 0 5 8 2 . 1 11 4 0 4 0 5 5 9 5 6 0 4 7 4 2 5 5 7 3 8 8 1 7 1 . 9 7 7 3 5 5 7 2 2 2 2 4 6 2 9 0 5 8 2 . 1 11 4 0 4 0 5 5 9 5 8 0 4 7 4 2 5 5 7 3 8 6 4 6 . 9 1 7 5 77 3 5 5 7 0 6 0 6 5 6 1 2 0 5 8 2 . 1 11 4 0 4 0 5 5 9 5 6 0 4 7 4 2 5 5 7 3 8 6 2 9 . 9 1 7 5 7 7 3 5 5 7 0 3 8 6 8 5 9 7 0 5 8 2 . 1 1 1 . 4 0 4 0 5 5 9 5 6 0 4 7 4 2 5 5 7 3 8 6 1 4 . 9 1 7 5 7 7 3 5 5 7 0 1 9 2 8 5 7 0 0 5 8 2 . 1 11 4 0 4 0 5 5 9 5 6 0 4 7 4 2 5 5 7 3 8 6 - 1 2 . 0 8 2 5 7 7 3 5 5 7 - 0 1 5 6 2 5 4 5 0 5 8 2 . 1 11 4 0 4 0 5 5 9 5 6 0 4 7 4 2 5 5 7 3 8 6 - 3 7 . 0 8 2 5 7 7 3 5 5 7 - 0 4 7 9 4 S 1 6 0 5 8 2 . 1 11 4 0 4 0 5 5 9 5 6 0 4 7 4 2 5 5 7 3 8 6 - 5 6 . 0 8 2 5 7 7 3 5 5 7 - 0 8 5 4 3 4 9 9 0 5 8 2 . 1 11 4 0 4 0 5 5 9 5 6 0 4 7 4 2 5 5 7 3 8 6 - 8 3 . 0 8 2 5 7 7 3 5 5 7 - 1 0 7 4 0 4 9 9 0 5 6 2 . 1 11 4 0 4 0 5 5 9 5 6 0 4 7 4 2 5 5 7 3 6 6 - 6 3 . 0 8 2 5 7 7 3 5 5 7 - 1 0 7 4 0 4 7 8 0 5 8 2 . 1 1 1 4 0 4 0 5 5 8 S 6 0 4 7 4 2 5 5 7 3 8 6 - 1 0 4 . 1 7 7 3 5 5 7 - 1 3 4 5 5 4 6 7 0 5 8 2 . 1 11 4 0 4 0 5 5 9 5 6 0 4 7 4 2 5 5 7 3 8 6 - 1 1 5 . 1 7 7 3 5 5 7 - 1 4 8 7 7 1 1 : 4 2 F R I D A Y , J A N U A R Y 1 1 . 1 9 9 1 P R E D I C T S T D E R R L 0 W E R 9 5 1 U P P E R 9 5 X L 0 W E R 9 5 X U P P E R 9 5 X S T D E R R S T U D E N T A C T U A L V A L U E P R E D I C T M E A N M E A N P R E D I C T P R E D I C T R E S I D U A L R E S I D U A L R E S I D U A L 2 4 4 2 4 0 5 8 2 . 1 1 1 4 0 4 0 5 5 9 5 6 0 4 . 7 4 2 5 . 5 7 3 8 . 6 - 1 5 8 . 1 7 7 3 5 5 7 - 2 0 4 3 8 2 5 7 6 8 0 5 8 2 . 8 1 3 6 7 8 0 5 3 5 7 5 8 9 . 9 4 0 5 . 5 7 2 0 . 0 2 2 5 . 2 7 6 9 8 6 2 2 9 2 S 8 2 6 7 2 7 0 5 6 2 . 8 1 3 6 7 8 0 5 3 5 7 5 6 9 . 9 4 0 5 . 5 7 2 0 . 0 1 6 4 . 2 7 6 9 8 6 2 2 1 3 3 4 2 7 6 9 9 0 5 6 2 . 8 1 3 6 7 8 0 5 3 5 7 5 8 9 . 9 4 0 5 . 5 7 2 0 . 0 1 3 6 . 2 7 6 9 6 6 2 1 7 6 9 7 2 8 7 0 1 0 5 6 2 . 8 1 3 6 7 8 0 5 3 5 7 5 8 9 . 9 4 0 5 . 5 7 2 0 . 0 1 3 8 . 2 7 6 9 8 6 2 1 7 9 5 7 2 9 6 8 9 0 5 6 2 . 8 1 3 6 7 8 0 5 3 5 7 5 8 9 . 9 4 0 5 . 5 7 2 0 . 0 1 2 6 . 2 7 6 9 8 6 2 1 6 3 9 8 3 0 6 6 1 0 5 6 2 . 8 1 3 6 7 8 0 5 3 S 7 5 8 9 . 8 4 0 5 . 5 7 2 0 . 0 9 8 . 2 4 2 7 7 6 9 8 6 2 1 2 7 6 1 3 1 6 3 0 0 5 6 2 . 8 1 3 6 7 8 0 5 3 5 7 5 8 9 . 9 4 0 5 . 5 7 2 0 . 0 6 7 . 2 4 2 7 7 6 9 8 6 2 0 8 7 3 4 3 2 5 6 7 0 5 6 2 . 8 1 3 6 7 8 0 5 3 5 7 5 8 9 . 9 4 0 5 . 5 7 2 0 . 0 4 . 2 4 2 7 7 6 9 8 6 2 0 0 5 5 1 3 3 5 1 9 0 5 6 2 . 8 1 3 6 7 8 0 5 3 5 7 5 8 9 . 9 4 0 5 . 5 7 2 0 . 0 - 4 3 . 7 5 7 3 7 6 9 8 6 2 - 0 5 6 8 4 3 4 4 7 0 0 5 8 2 . 8 1 3 8 7 8 0 5 3 5 7 5 8 9 . 9 4 0 5 . 5 7 2 0 . 0 - 9 2 . 7 5 7 3 7 6 9 8 6 2 - 1 2 0 4 9 3 5 4 5 4 0 5 6 2 . 8 1 3 6 7 8 0 5 3 5 7 5 8 9 . 9 4 0 5 . 5 7 2 0 . 0 - 1 0 8 . 8 7 6 9 8 6 2 - 1 4 1 2 7 3 6 4 4 3 0 5 6 2 . 8 1 3 6 7 8 0 5 3 5 7 5 8 9 . 9 4 0 5 . 5 7 2 0 . 0 - 1 1 9 . 8 7 6 9 8 6 2 - 1 5 5 5 6 3 7 4 4 6 0 5 6 2 . 8 1 3 6 7 8 0 5 3 5 7 5 8 9 . 9 4 0 S . S 7 2 0 . 0 - 1 1 6 . 8 7 6 9 8 6 2 -1 5 1 6 6 3 8 4 3 9 0 5 6 2 . 8 1 3 6 7 8 0 5 3 5 7 5 8 9 . 9 4 0 5 . 5 7 2 0 . 0 - 1 2 3 . 8 7 6 9 8 6 2 -1 6 0 7 5 3 8 4 0 4 0 5 6 2 . 6 1 3 6 7 8 0 5 3 5 7 5 8 9 . 9 4 0 5 . 5 7 2 0 . 0 - 1 5 8 . 6 7 6 9 6 6 2 - 2 0 6 2 2 4 0 4 0 0 0 5 6 2 . 8 1 3 8 7 8 0 5 3 5 7 5 8 9 . 9 4 0 5 . 5 7 2 0 . 0 - 1 6 2 . 8 7 6 9 6 6 2 - 2 1 1 4 1 4 1 7 9 4 0 7 0 1 . 4 11 2 7 3 3 6 7 9 1 7 2 3 . 7 5 4 4 . 9 8 5 7 . 9 9 2 . 6 0 0 3 7 7 3 7 4 9 1 1 9 6 8 4 2 8 0 5 0 7 0 1 . 4 11 2 7 3 3 6 7 9 1 7 2 3 . 7 5 4 4 . 9 8 5 7 . 9 1 0 3 . 6 7 7 3 7 4 9 1 3 3 8 9 4 3 7 9 8 0 7 0 1 . 4 11 2 7 3 3 6 7 9 1 7 2 3 . 7 5 4 4 . 9 8 5 7 . 9 9 6 . 6 0 0 3 7 7 3 7 4 9 1 2 4 9 5 4 4 7 9 5 0 7 0 1 . 4 11 2 7 3 3 6 7 9 1 7 2 3 . 7 5 4 4 . 9 8 5 7 . 9 9 3 . 6 0 0 3 7 7 3 7 4 9 1 2 0 9 7 4 5 7 5 7 0 7 0 1 . 4 11 2 7 3 3 6 7 9 1 7 2 3 . 7 5 4 4 . 9 8 5 7 . 9 5 5 . 6 0 0 3 7 7 3 7 4 9 0 7 1 8 6 4 6 7 4 2 0 7 0 1 . 4 11 2 7 3 3 8 7 9 1 7 2 3 . 7 5 4 4 . 9 8 5 7 . 9 4 Q . 6 0 0 3 7 7 3 7 4 9 0 5 2 4 7 4 7 7 3 0 0 7 0 1 . 4 11 2 7 3 3 6 7 9 1 7 2 3 . 7 5 4 4 . 9 8 5 7 . 9 2 8 . 6 0 0 3 7 7 3 7 4 9 0 3 6 9 6 4 8 7 3 0 0 7 0 1 . 4 11 2 7 3 3 6 7 9 1 7 2 3 . 7 5 4 4 . 9 8 5 7 . 9 2 8 . 6 0 0 3 7 7 3 7 4 9 0 3 6 9 6 4 9 7 2 9 0 7 0 1 . 4 11 2 7 3 3 8 7 9 1 7 2 3 . 7 5 4 4 . 9 . 8 5 7 . 9 2 7 . 6 0 0 3 7 7 3 7 4 9 0 3 5 6 7 5 0 7 2 2 0 7 0 1 . 4 11 2 7 3 3 6 7 9 1 7 2 3 . 7 5 4 4 . 9 8 5 7 . 9 2 0 . 6 0 0 3 7 7 3 7 4 9 0 2 6 6 2 5 1 7 2 6 0 7 0 1 . 4 11 2 7 3 3 6 7 9 1 7 2 3 . 7 5 4 4 . 9 8 5 7 . 9 2 4 . 6 0 0 3 7 7 3 7 4 9 0 3 1 7 9 5 2 7 1 8 0 7 0 1 . 4 11 2 7 3 3 6 7 9 1 7 2 3 . 7 5 4 4 . 9 8 5 7 . 9 1 6 . 6 0 0 3 7 7 3 7 4 9 0 2 1 4 5 5 3 6 8 8 0 7 0 1 . 4 11 2 7 3 3 6 7 9 1 7 2 3 . 7 5 4 4 . 9 8 5 7 . 9 - 1 3 . 3 9 9 7 7 7 3 7 4 9 - 0 1 7 3 2 5 4 6 7 1 0 7 0 1 . 4 11 2 7 3 3 8 7 9 1 7 2 3 . 7 5 4 4 . 9 8 5 7 . 9 - 3 0 . 3 9 9 7 7 7 3 7 4 9 - 0 3 9 2 9 5 5 6 6 1 0 7 0 1 . 4 11 2 7 3 3 6 7 9 1 7 2 3 . 7 5 4 4 . 9 8 5 7 . 9 - 4 0 . 3 9 9 7 7 7 3 7 4 9 - 0 5 2 2 1 5 6 6 4 2 0 7 0 1 . 4 11 2 7 3 3 6 7 9 1 7 2 3 . 7 5 4 4 . 9 6 5 7 . 9 - 5 9 . 3 9 9 7 7 7 3 7 4 9 - 0 7 6 7 7 5 7 6 4 2 0 7 0 1 . 4 11 2 7 3 3 6 7 9 1 7 2 3 . 7 5 4 4 . 9 6 5 7 . 9 - 5 9 . 3 9 9 7 7 7 3 7 4 9 - 0 7 6 7 7 5 8 7 4 0 0 8 8 2 . 1 7 4 3 6 4 6 6 7 3 6 9 6 . 8 5 2 6 . 5 8 3 7 . 7 5 7 . 9 2 5 5 7 7 8 3 7 4 0 7 4 4 2 5 9 7 4 8 0 6 8 2 . 1 7 4 3 6 4 6 6 7 3 6 9 8 . 8 5 2 6 . 5 8 3 7 . 7 6 5 . 9 2 5 5 77 8 3 7 4 0 8 4 7 0 6 0 7 3 9 0 6 8 2 . 1 7 4 3 6 4 6 6 7 3 6 9 6 . 8 5 2 6 . 5 8 3 7 . 7 5 6 . 9 2 5 5 77 6 3 7 4 0 7 3 1 3 8 1 7 3 0 0 6 8 2 . 1 7 4 3 6 4 6 6 7 3 6 9 6 . 8 5 2 6 . 5 8 3 7 . 7 4 7 . 9 2 5 5 7 7 8 3 7 4 0 6 1 5 7 8 2 7 2 9 0 6 8 2 . 1 7 4 3 6 4 6 6 7 3 6 9 6 . 8 5 2 6 . 5 8 3 7 . 7 4 6 . 9 2 5 5 7 7 8 3 7 4 0 6 0 2 9 6 3 7 3 0 0 6 8 2 . 1 7 4 3 6 4 6 6 7 3 6 9 8 . 8 5 2 6 . 5 8 3 7 . 7 4 7 . 9 2 5 5 7 7 8 3 7 4 0 6 1 5 7 6 4 7 2 3 0 6 8 2 . 1 7 4 3 6 4 6 8 7 3 6 9 6 . 8 5 2 6 . 5 8 3 7 . 7 4 0 . 9 2 S S 7 7 8 3 7 4 0 5 2 5 6 6 5 7 1 3 0 6 8 2 . 1 7 4 3 8 4 6 6 7 3 6 9 6 . 8 5 2 6 . 5 8 3 7 . 7 3 0 . 9 2 5 5 7 7 8 3 7 4 0 3 9 7 3 6 6 6 8 4 0 6 8 2 . 1 7 4 3 6 4 6 6 7 3 6 9 8 . 8 5 2 6 . 5 8 3 7 . 7 1 . 9 2 5 5 77 8 3 7 4 0 0 2 4 7 6 7 6 7 7 0 6 8 2 . 1 7 4 3 8 4 8 6 7 3 6 9 6 . 8 5 2 6 . 5 8 3 7 . 7 - 5 . 0 7 4 5 77 8 3 7 4 - 0 0 6 5 2 6 8 6 8 3 0 6 8 2 . 1 7 4 3 6 4 6 6 7 3 6 9 8 . 8 5 2 6 . 5 8 3 7 . 7 - 1 9 . 0 7 4 5 7 7 8 3 7 4 - 0 2 4 5 1 6 9 6 4 8 0 6 8 2 . 1 7 4 3 8 4 8 8 7 3 8 9 6 . 6 5 2 6 . 5 8 3 7 . 7 - 3 4 . 0 7 4 5 77 6 3 7 4 - 0 4 3 7 8 7 0 6 0 1 0 6 8 2 . 1 7 4 3 8 4 6 6 7 3 6 9 6 . 8 5 2 6 . 5 8 3 7 . 7 - 8 1 . 0 7 4 5 7 7 8 3 7 4 -1 0 4 1 6 7 1 5 9 7 0 6 8 2 . 1 7 4 3 6 4 6 6 7 3 6 9 6 . 8 5 2 6 . 5 8 3 7 . 7 - 8 5 . 0 7 4 5 77 6 3 7 4 -1 0 9 3 0 Appendix B. STATISTICAL ANALYSIS 262 Table B.3 : The Effect of Ba-Glucose Complex on Electrode Response (continues). SAS GENERAL LINEAR MODELS PROCEDURE CLASS LEVEL INFORMATION CLASS LEVELS VALUES SEN S 1 2 3 4 5 8 7 B COMP 3 0 12 NUMBER OF OBSERVATIONS IN DATA SET = 118 SAS GENERAL LINEAR MODELS PROCEDURE DEPENOENT VARIABLE: VOLTAGE SOURCE DF MODEL 15 ERROR too CORRECTED TOTAL 115 1 1 : 4 2 FRIOAY. J » N U A R V 11. 1 9 9 1 SOURCE COMP SEN(COMP) TIME TIME-SEN TIME'SEWCOMPI OF 2 5 1 7 0 SUM OF SOUARES 1294869.40297211 159727.45909886 1454396.86208697 TYPE I SS 740430.91104856 43444.68762062 382553.93112861 128239.87317612 0.00000000 MEAN SQUARE 86311.29353147 1597.27459097 F VALUE 231.78 5.44 239.50 11.47 PR > F 0.0001 0.0002 0.0001 0.0001 .04 0.0001 0.890176 5.9426 ROOT MSE VOLTAGE MEAN 39.96591787 672.53448276 OF TYPE III SS F VALUE PR > F 2 106496.95428962 33.34 0.0001 5 18416.73046843 2.31 0.0499 1 284775.48233280 178.29 0.0001 7 128239.87317612 11.47 0.0001 0 0.00000000 PARAMETER INTERCEPT COMP SEN(COMP) TIME TIME'SEN TIME'SEN(COMP) 0 1 2 1 2 3 4 5 8 7 8 1 2 3 4 5 6 7 8 1 2 3 4 5 e 7 ESTIMATE 790.01929711 B -125.02997785 B -77.64096512 B 0.00000000 B 13.62826649 B 11.15926294 B 0.00000000 8 72.69483197 B 38.37397256 B 0.00000000 B 14.27193462 B 0.00000000 B 0.00826324 B •0.82513268 B -1.09723372 B -1.05160574 B -0.40942124 B -0.54700180 B -0.32994890 B -0.19715514 B 0.00000000 B 0.00000000 B 0.00000000 8 0.00000000 B 0.00000000 B 0.00000000 B 0.00000000 B 0.00000000 B 0.00000000 B T FOR HO: PR > |T| STD ERROR OF PARAMETERS ESTIMATE 44.01 0.0001 17 .95135707 -5.55 0.0001 22 .51199734 -3.20 0.0019 24 .27348610 o!54 0.5870 25 .00822657 0.55 0.5818 20, .19378854 3!26 0.0014 22! 15140262 1.72 O.OB77 22. ,25139372 0^56 o!5757 2S] 41856806 o!o5 o!9S78 0] 15558878 -3.87 0.0002 0. 21314497 -5.54 0.0001 0 . 19797763 -5.76 0.0001 0 . 16275496 -2.35 0.0206 0. 17403591 -3.14 0.0022 0 . 17436950 -1.88 0.0631 0. 17555544 •0.94 0.3469 0. 20947251 SAS GENERAL LINEAR MODELS PROCEDURE LEAST SQUARES MEANS COMP VOLTAGE LSMEAN 0 NON-EST 1 NON-EST 2 NON-EST SEN COMP VOLTAGE STO ERR PROB > |T| PROB LSMEAN LSMEAN HO:LSMEAN=0 I/J 1 0 581.233794 12.652583 0.0001 1 2 0 546.326028 11.458155 0.0001 2 0 3 0 540.582500 10.236471 0.0001 3 0 4 1 737.248768 9.817016 0.0 4 0 5 1 888.526123 9.939888 0.0 5 0 6 1 874.028538 10.738952 0.0 6 0 7 2 781.772321 10.913176 0.0 7 0 8 2 791.004407 11.936294 0.0 8 0 > |T| HO: LSMEANlI>=LSMEAN(J) 1 2 3 4 5 6 7 8 0, .0435 0. 0141 0. 0001 0. ,0001 0 .0001 0. 0001 0 , . 0 0 0 1 10435 0, ,7093 0 ,0001 0.0001 0. .0001 0 . ,0001 0. ,0001 .0141 0 ! 7093 0 .0001 0 .0001 0 .0001 0 ,0001 0 . 0 0 0 1 .0001 0.0001 0 !oooi 0 .0005 0.0001 0 .0032 0 .0006 .0001 0 .0001 0 .0001 0 .0005 0 .3951 0 .0001 0 . 0 0 0 1 .0001 0 .0001 0 .0001 0 .0001 0 !s951 0 .0001 0 . 0 0 0 1 .0001 0 .0001 0 .0001 0 .0032 0 .0001 0 !oooi 0 .5694 .0001 0 .0001 0 .0001 0 .0008 0 .0001 0 .0001 0 .5694 Appendix B. STATISTICAL ANALYSIS 263 11:42 FRIOAY, JANUARY 11. 1991 PREDICT STD ERR LOWER95X UPPER95X LOWER95% UPPER95X STD ERR STUDENT ACTUAL VALUE PREDICT MEAN MEAN PREDICT PREDICT RESIOUAL RESIDUAL RESIDUAL 72 580 0 682 1 7.4364 667 3 696 8 526 5 837.7 102. 1 77 6374 -1 3114 73 557 0 682 1 7.4364 667 3 696 8 526 5 837.7 - 125. 1 77 8374 -1 6069 74 556 0 682 1 7.4364 667 3 696 8 526 5 837.7 - 126. 1 77 8374 -1 6197 75 712 0 662 7 13.1753 636 6 668 9 505 7 819.8 49 2507 77 0738 0 6390 76 682 0 662 7 13.1753 636 6 688 9 505 7 819.6 19 2507 77 0738 0 2498 77 681 0 662 7 13.1753 636 6 688 9 505 7 819.8 18 2507 77 0738 0 2368 76 685 0 662 7 13.1753 638 6 688 9 505 7 819.8 22 2507 77 0738 0 2887 79 696 0 662 7 13.1753 636 6 688 9 505 7 819.8 33 2507 77 07 38 0 4314 BO 703 0 662 7 13.1753 636 6 688 9 505 7 819.8 40 .2507 77 07 38 0 5222 81 703 0 662 7 13.1753 636 8 688 9 505 7 819.8 40 .2507 77 0738 0 5222 82 701 0 662 7 13.1753 636 6 688 9 505 7 819.8 38 2507 77 0738 0 4963 83 698 0 662 7 13.1753 636 6 688 9 505 7 819.6 33 2507 77 07 38 0 4314 84 679 0 662 7 13.1753 636 8 688 9 505 7 619.8 16 2507 77 0738 0 2108 85 608 0 662 7 13.1753 636 6 666 9 505 7 819.6 -54 7493 77 0738 -0 7103 86 618 0 662 7 13.1753 636 6 688 9 505 7 819.8 -44 7493 77 0738 -0 5806 87 606 0 662 7 13.1753 636 6 688 9 505 7 819.8 -56 7493 77 0738 -0 7363 88 582 0 662 7 13.1753 636 6 666 9 505 7 819.8 -80 7493 77 0738 -1 0477 89 582 0 662 7 13.1753 636 6 666 9 505 7 819.8 -80 7493 77 0738 -1 0477 90 802 0 801 4 14.0495 773 6 629 2 644 0 958.6 0 6083 76 9192 .0079079 91 810 0 801 4 14.0495 773 6 629 2 644 0 958.8 8 6083 76 9192 0 1119 92 807 0 801 4 14.0495 773 6 629 2 644 0 958.8 5 6083 76 9192 0 0729 93 799 0 801 4 14.0495 773 6 829 2 644 0 958.8 -2 .3917 76 9192 -0 0311 94 791 0 801 4 14.0495 773 6 829 2 644 0 958.8 -10 .3917 76.9192 -0 1351 95 793 0 801 4 14.0495 773 6 629 2 644 0 958.8 -8 .3917 76 9192 -0 1091 96 781 0 601 4 14.0495 773 6 829 2 644 0 958.8 -20 .3917 76 9192 -0 2651 97 781 0 801 4 14.0495 773 6 829 2 844 0 958.8 -20 .3917 76 9192 -0 2651 98 773 0 601 4 14.0495 773 6 829 2 644 0 958.8 -28 .3917 76 9192 -0 3691 99 778 0 801 4 14.0495 773 6 829 2 644 0 958.8 -23 .3917 76 9192 -0 3041 100 770 0 801 4 14.0495 773 6 829 2 644.0 958.8 -31 .3917 76 9192 -0 4081 101 788 0 801 4 14.0495 773 6 629 2 644 0 958.8 -33 .3917 76 9192 -0 4341 102 767 0 801 4 14.0495 773 6 829 2 644 0 958.8 -34 .3917 76 9192 -0 4471 103 787 0 801 4 14.0495 773 6 629 2 644 0 958.6 -34 .3917 76 9192 -0 4471 104 796 0 782 1 13.6693 755 0 809 1 624 8 939.3 13 9335 76 9B77 0 1810 105 787 0 782 1 13.6693 755 0 809 1 624 8 939.3 4 9335 76 9877 0 0641 108 793 0 782 1 13.6693 755 0 809 1 624 8 939.3 10 .9335 76 9877 0 1420 107 791 0 782 1 13.6693 755 0 809 1 624 8 939.3 8 9335 76 9877 0 1160 108 789 0 782 1 13.6693 755 0 609 1 624 8 939.3 6 9335 76 9877 0 0901 109 790 0 782 1 13.6693 755 0 809 1 624 8 939.3 7 9335 76 9877 0 1030 110 788 0 782 1 13.8693 755 0 809.1 624 8 939.3 5 9335 76 9877 0 0771 111 786 0 782 1 13.6693 755 0 809 1 624 8 939.3 3 9335 76 9877 0 0511 112 792 0 782 1 13.6693 755 0 809 1 624 8 939.3 9 9335 76 9877 0 1290 113 792 0 782 1 13.6693 755 0 809 1 624 8 939.3 9 9335 76 9877 0 1290 114 792 0 782 1 13.6693 755 0 809 1 624 8 939.3 9 933S 76 9877 0 1290 115 792 0 782 1 13.8693 755 0 809 1 624 8 939.3 9 9335 76 9877 0