x '3 ~L~IH “a ' ‘*" 3 :. "y I 'I‘flb . . . . ‘ 9‘. oft" ' Iw} ! Q 5 WM: -V'. 2 -;""~‘ 3 gum-5" " v- . ffé‘xE-firw 'w.’ ' a 5' 1r?“ . ’7- -‘-{4.‘ . ‘3‘). .‘ ‘8 s';_-.‘_ 7‘..." }‘."I. v I u . . ‘ ‘ .... L3 “Xx-3‘1 ‘ . . VII ' II n ‘ 'L‘I“":" ‘-:.. ‘ w --.w--.—. . .. . 1? 'g. - u» L' I ‘~‘ .‘ . 'I 4“'~ . “H's-u... A ' 'ct- “Nu - ‘.'.-'-y'"...'a‘ '\4 .' .. ‘ .‘ L ' ,flc) flog“. ‘ . 7 . . '1 . , .., ‘ A a ‘t :3 d H ‘ I} “ M ‘u' I\ ‘1 '1 :45. . : ‘ ‘ ‘ k . ‘ ' ' .- .i,‘ if, , . . "a Ié‘ ‘I‘ . I” \itli‘ ".u .‘ ”1"" 'é't“ *3. ' ' h .w 4""- 011651” ':k‘):"'o V: I -\ ’ ' I R; Jr? . ;“.;£..'L.m 3+. " ' " w“ '. ‘ 3:3,: ,_ [I] m . L‘ l ‘ . .. h" v ‘ V s a ’ , ' . ‘. ‘ ‘- ‘1'” ~' ‘V ‘y" w 1“. A " ”f ( : - '(‘- . r | Q. '“\\‘N . ‘ ‘ ‘I ‘1 ‘ .1 .I I ‘ '_ I. ' "'m V'IM'I‘ 'r I I ‘ . ‘I V II I'EI’”R’$I"J/qb‘lfl .. H a J ' .I I ‘ " I . L I I“. |‘ «2‘7“ n _. 2-? L" 3 n u I u ‘Ir ‘ ‘ a I‘ I H" l.‘ IJ. n ‘fi. vI 2%, ‘ ' . ., . “ .mlu 3+. . . ‘. , ‘ ' .ul'”? mu" 1.4- “H ‘ ' ' yawlfu "“-' ‘, H7?“ #5.? _ \ 5““ ,J‘“ > .Al‘),I'-or 7 xi“? W l. LIBRAR y Michigan State University This is to certify that the thesis entitled Computer Controlled Coulostatic Instrumentation and Its Application to the Study of Mercury Film on the Platinum Surface presented by Minchen Wang has been accepted towards fulfillment of the requirements for Ph.D. degree in Analytical Chemistry f2? 24 I Major professor \ Date 52% /§-,. ’97/ 0-7639 cxrnnnunzcxrnxrnrrrlcourosmnmmc INSTRUMENTHAND ITS.APPLICHTIGN $O‘THE STUDY’OFNMERCUR!’FIINICRIPIAEINUM.SUREACE. mm A.DISSERINEE3§ Eiflnfixied.to Michigan State [hiversity Inpartialfiilfillmtoftherequjxanerrt forthedegreeof DOCTOR.OF'PHILOSOPHY Deparmentofdanistry 1978 const daliv elect curre Pulse immed by th Per d seque PPOgr QUite film 1 three ABSTRACT COMPUTER CONTROLLED COULOSTATIC INSTRUMENT AND ITS APPLICATION TO THE STUDY OF MERCURY FILM ON PLATINUM SURFACE By Minchen Wang A computer controlled coulostatic instrument was constructed. The basic function of this instrument is to deliver a predetermined amount of charge to the working electrode of an electrochemical cell by supplying constant current pulses. It can deliver currents up to‘: 2.5 mA with pulses as short as 0.56‘ns. The potential response of the cell immediately after the coulostatic charging pulses is followed by the computer. The fastest sampling rate is about 13‘ns per datum point. Under computer control, other polarization sequences can be performed simply by using the appropriate programs. This capability makes the coulostatic instrument quite versatile. This instrument was used in the study of the mercury film formation on a platinum surface. On the platinum surface, three structures of mercury deposits were identified, namely; re ho pl. int on 1 con; to t be d. reac1 thick trans deCre surfs mercury droplets, mercury patches and smooth mercury films. The structure of the mercury film depends on the deposition overpotential and the pretreatment of the platinum prior to the mercury deposition. 0n the freshly polished platinum surface, mercury drOplets are formed at small deposition overpotentials while at higher deposition overpotential, mercury patches are formed. By repetitive oxidation and reduction of the platinum surface, an activated, more homogeneous platinum surface is produced which can be easily plated with smooth mercury film. The plated mercury reacts with the platinum to form an intermetallic compound. The rate of this reaction is dependent on the thickness of the intermetallic compound. For a given compound thickness, the transformation of the metallic mercury to the compound form is proportional to t%. This relation can be described in terms of the diffusion process of the reactants through the intermetallic compound layer. As the thickness of the compound increases, the rate of the transformation of the metallic mercury into the compound decreases. A new approach to plate a mercury film on a platinum surface was also devised. In this approach, the chloride ion pr on of ef pl: is utilized as a masking agent during the mercury plating process. The chloride ion forms an insoluble calomel film on the deposited mercury surface which prevents the formation of mercury droplets. This approach is found to be quite effective for creating a thin smooth mercury film on the platinum surface. gr; H103 ser' pre; Enid aPPP Help are 1 While thank throu- ACKNOWLEDGEMENT I would like to express my thanks to Professor Christie Enke for his guidance and patience during the course of my graduate study. His advice and friendship in many critical moments will always be remembered. Thanks to Professor Weaver for taking the hardship and serving as my second reader at the last moment of thesis preparation. Thanks go also to the other members of my guidance committee. I also like to express my special appreciation to Dr. Atkinson for his help and friendship. Helps from my fellow members of the Enke research group are highly appreciated. Thanks to my wife, Hwai-Lin, who typed the whole thesis while taking a full time Job to support our family. Thanks to my parents-in-law for their selfless help. And above all, thanks to my parents who have always shown their support throughout my life. Chapt LIST LIST I. II. III Chapter TABLE OF CONTENTS Page LIST OF TABLESCOOOOOOOOOOOOOOOOOOOOOOOOOCOOOOOOOOOOOOOVii LIST OF FIGURESOOOOOO0.0...OOOOOOOOOOOOOOOOOOOIOO0.000Cix I. A COMPUTER CONTROLLED GENERAL PURPOSE COULOSTATIC SYSTEM...OOOOOOOOOOOOOOOOOOO.0...00.0.1 A. B. C. D. Introduction to the Coulostatic Instrumentation...............................1 Instrument....................................7 Performance of the Instrumental syStemoeeeeeeeeeeeeoeeeeeeeeeeeeeoeoeee0.0.0023 Applications of the General Purpose Coulostatic Instrument.......................51 II. INTRODUCTION TO THE STUDY OF THE MERCURY FILM 0N PLATINUM.........................6# A. B. C. D. E. The Search for a Mercury Film Electrode....................................64 Mercury Film Preparation.....................68 Interaction between Mercury and Platinum eoeeeoeeooeoeeeeeoeeeeeeone00000000073 Additional Problems with Platinum- Based Mercury Film BleCtrOdesoooooooeooo0.0.074 Objective Of StUdyoeone.eoooeeeeeeeeeoeooeeee75 III. EVALUATION OF THE FACTORS AFFECTING THE STRUCTURE OF THE MERCURY ELECTROPLATED ON [PI-IE PLATINUM SURFACE..OOCOCOCOOOOOOOOOO0.0.0.077 111 Chapter A. B. C. D. E. F. Page Platinum Surface State Characterization.......77 Methods of Examining the Smoothness of the Mercury Film on the Platinum..............83 Experimental Set-up...........................8h Experiments and ResultSOOOOOOOOOOOOOOIO00.0.0.9h 1. Creating Platinum Surface with Varying Roughness FactorS.................9# 2. Influence of the Potential and the Roughness Factor on the Mercury ' DeP081tionooeoeoooeeeeoe00000000000000.0010} DichSSiOneooeeeeeeeeeeeooeoeeeeeeeeeeeoeeoeeo‘11 1. Structure of the Mercury Deposit VGIBus the Plating Potential..............111 2. Effect of Surface Roughness Factor on the Structure of the Mercury Deposit......ll7 3. Other Factors that Influence the Formation of the Mercury Deposit..........120 COROlHSiODOOeeeeeeeeeeoeeeeeoecoo-00000000000123 A NElV METHOD OF MAKING A SMOOTH MERCURY FILM ON A PLATIMIM SURE‘ACE...COOOCOOOOOOOOOOOOCCC125 A. B. C. Introduction.................................125 Procedures of Making a MFE Using Chloride Ion as a Masking Agent..............127 Results and DiscuSSioneeeeeoeeeoeeoeoeee00000127 1. PrOdUCt Of Primary Plating...............127 2. Continual Mercury Plating................131 3. Factors that Influence the MFE Obtained..131 iv summit» V. mum W THE mm mm m m m Moooooooooooooooooeoo090000000134 A. Inmtim...O...OOOOOOOOOOOOOOOOOOOOOO0.0.0.00134 1. Scope of the Pt-Hg Reaction Rate Sm...OOIOOOOOOOOOOOOIOOOOO.00.00.000.134 2. Relevant Results Obtained by Other Vbrkers...l35 B. mimalOOOOOOOOOOOOOOOOOOOOOOOOOOOO000......139 C. Wilma]. mall-QC...OOOOOOOOOOOOOOOOOOOO0.0.0140 1. Methods Used for: tie Pt-Hg mm mm WYOOOOOOOOOOOOOOOOOOOOO0.00.140 2. Anodic Stripping of MP!) in Nlerams Solutim...........................143 3. Disappeararm of Metallic Hg on tm WOOD...0.0.0.0...0.0.0.0000000000000145 D. Dimsj-mOOOOO00......OOOOOOOOIOOOOOOOOOOO0.0.0.155 1. Reaction of Platinuu and Mercury on WOOOOOOOOOO...OOOOOOOOOOOOOOOOOO00......155 2. Accmmting for the Pt-Bg Reactim Rate.......lS7 VI. Putin's Study........................................164 APPENJDK l Carpater Interface...........................165 APPENDIX 2 Progran Listingsnuu.......................168 Macros and Registers Assigment.....................169 Program am.m..................................172 Suhroutim 01mm.mc...............................175 Program SIDPE.E'IN...................................178 Progran CAPF3.E'1N...................................180 Suhrmtine CAPMB.MC................................184 Subrmtim m.m...............................186 Sihmutine NIP.F'IN..................................187 Progran W.E'1N..................................189 W VDEPm.MC...............................191 Program QWJ'IN..................................l93 Subroutim Gm.MAC...............................195 Program DEPOSE‘.P‘IN..................................198 Submutine DWMACNHHH.......................200 Program EI'REAPJ'IN..................................202 Submutim EI'RFAMMPC..."..........................204 Progran PIA'INZ.MAC..................................20‘7 Progran SIRIPF.F'IN..................................210 Smroutine STRICN.MC...............................212 WOOOOOOOOOOOOOOOOOOOOOOO00......0.......0.0.0.0214 1-. 1.10 Table 1-1 Slopes of Channel 1 Slopes of Channel 1 Slopes of Channel 2 Slopes of Channel 2 Slapes of Channel 3 Slopes of Channel 3 Variation LIST OF TABLES Page the Pulse Charge Content, Current with Negative Current.....................34 the Pulse Charge Content, Current with Positive Current.....................34 the Pulse Charge Content, Current with Negative Current.....................38 the Pulse Charge Content, Current with Positive Current.....................38 the Pulse Charge Content, Current with Negative Current.....................h2 the Pulse Charge Content, Current with Positive Current.....................#3 of the SCC with the Pulse Width for Current Channel 1eeeeeeeoeeeeeoeeeeeeeeoeeeeeoeeH? Variation of the SCC with Pulse for Current Channels 2 and 300eoeeeeeeeeeeeeeeoeeeeeeeeeeooeeeeoh? Charge Produced by a 1.56‘ns, 2.5 mA Current U PUIBeeoeoeeeeooeeeoeeeeeeeeeoeeeeeeeeeeeeeeeeeeeeeee#8 Determination of Cell Capacitance and Faradaic Resistance with High Amplitude Current Pulses.......55 vii \N 5-. Table 1-11 3-1 5-2 Page Determination of Cell Capacitance and Faradaic Resistance with Low Amplitude Current Pulses........60 Platinization of Platinum with Alternating Current Steps......................................102 Potential Dependence of the Mercury Deposit on the Bright Platinum Electrode...................106 Dependence of the RF of the Mercury Deposited at 200 mV..........................................106 Disappearance of Metallic Mercury in the Current Reversal Experiment........................1Q? Disappearance of Metallic Mercury in the Aging Experiment.........................................152 viii Figure 1-1 1-12 1-13 1-H. 1-15 1-16 LIST OF FIGURES Page Equivalent Circuit and Current Impulse Response of Electrochemical System...................4 Computer Controlled Coulostatic System...............8 Pulse Generator.....................................12 Pulse Generator Circuit.............................14 Current Generator...................................17 Digital to Analog Converter Card....................19 Data Acquisition System.............................21 Shortest Current Pulses.............................27 Cathodic Charging Curve with Current Channel 1......31 Anodic Charging Curve with Current Channel 1........33 Cathodic Charging Curve with Current Channel 2......36 Anodic Charging Curve with Current Channel 2........37 Cathodic Charging Curve with Current Channel 3......h0 Anodic Charging Curve with Current Channel 3........41 Cathodic Charging Curve with Current Channel h......#5 Potential Decay of a Dummy Cell with High Amplitude Current PulBGSOO000.000.000.00000000053 ix Fig: 1-20 3-2 3-3 3-4 3-5 5~6 5~7 5-8 3-9 3~lo iFigure 1-17 1-18 1-19 1-20 3-6 3-7 3-8 3-10 Page Curve Fitted Potential Decay of a Dummy Call with High Amplitude Current Pulses.................5# Potential Decay of a Dummy Cell with Small Amplitude Current Pulses...........................57 Improvement of Data through the Curve Fitting and Signal Averaging Process.......................59 Simulated Potentiostatic Polarization by the Coulostatic Instrument.............................62 Flow Cell..........................................88 Potential Response to the Solution Oxygen..........91 Flow Diagram of Hg Deposition Experiment...........93 Cyclic Chronopotentiograms of Pt Electrodes with and without Aqua Regia Treatment..............95 Cyclic Chronopotentiograms of Pt Electrodes in 0.1 M HCl and 0.1 M H0104.......................99 Creating Platinized Pt Electrodes by Alternating Current Steps.........................101 Two Types of Mercury Deposit on Pt Surfaces.......105 Plating Mercury on Pt Surface with Small Current, -2h‘pA............................................109 Plating Mercury on Pt Surfaces with Large Current, ~1000‘pA..........................................11O Plating Mercury on Thick Hg Coated Pt Electrodes With Large Current, -1000 Moose-eeeeeeeeeoeeeeeQOI12 \n 5-1 5~7 5-8 5‘9 Figure 5-6 Page Mercury Plating at a Constant Potential in 1.2 N HCl and 0.1 M HgClZ Solution................128 Potential-Charge Response of MFE under Anodic Current in 0.1 M H0104.....................137 Current Programs used for Pt-Hg Reaction Rate Study........................................142 Anodic Potential-Charge Characteristic of the MFE in 0.1 M HgClO4 Solution..................1AA Potential-Charge Response with Current Reversal Polarization.............................1A6 Potential-Charge Variation during Reduction of Mercurous Ions from 0.1 M HgClO Solution 4 Using High Reduction Current......................1A9 Potential-Charge Variation during Repetitive Deposition-Stripping of Hg with Varying Aging Intervals.........................................151 Disappearance of Metallic Mercury with Aging......154 Effect of Pt-Hg Compound Thickness on the Reaction Rate Of Pt and ngeeeeeeeeeeeeoeeeeeeeeoe156 Hg Concentration Profile in the Pt-Hg Compound....158 A. stu occ of ' trai syst CHAPTER I A COMPUTER CONTROLLED GENERAL PURPOSE COULOSTATIC SYSTEM A. Introduction to Coulostatic Instrumentation Several methods have been developed over the years to study the charge tranSfer and mass transport processes occuring at electrode surfaces. To avoid the interferences of the mass transport phenomenon, most methods of charge transfer study rely on the response of the electrochemical system immediately after an excitation signal has been applied. Among these methods is the coulostatic technique which was independently and simultaneously proposed by Reinmuth (A2) and Delahay (28). In this method, a known amount of charge is delivered to the electrode double layer in a very short time. The application of charge changes the potential of the electrode to a new value. Thus the electrode is disturbed from the reaction conditions (equilibrium, steady state, no reaction) which existed prior to the charge injection. The potential change may induce a change in the electrochemical reaction conditions which consumes the charge stored on the electrode double layer. As a result, the potential relaxes back toward the prior value. The rate of the potential decay is a function of the rate of the electrochemical reaction and the value of the double-layer cc De st eli (1+5 of cha the at - This as t time ovepl Hausa depen tErmi equiv Cich Paral capacitance. Thus the rate of the electrochemical reaction induced by the charge pulse can be deduced from the potential- time curve. Two techniques have been used to accomplish the coulostatic charging process. The earlier technique of Delahay and Reinmuth used a small, high quality capacitor to store the charge which was then discharged across the electrochemical cell. The second technique, by Wier and Enke (45), used a current impulse to deliver the desired amount of charge. This method has two advantages over the capacitor charging method: the potential during pulsing can be lower than with the capacitor method in which the peak potential at the first instant of charge dumping must be quite high. This high voltage drives the input of measuring devices such. as the oscilloscope to their limit. Since a certain period of time is required for the device to recover from being overdriven, the information within this period of time is unusable (#3). In addition, since the pulse duration is not dependent on cell resistance, it is better defined and it is terminated more sharply than with an exponential decay. An electrochemical cell can be analyzed in terms of an equivalent circuit as shown in Figure 1-1. A. This equivalent circuit is composed of a double-layer capacitor Cd, in parallel with resistors Rf and Zn, and in series with another $02 as 1‘68 At unt: resistor Rs' The double-layer capacitance is determined by several factors: the material and the surface condition of the electrode; the composition of the electrolyte and the potential of the electrode. The faradaic resistance, Rf, is a function of the rate of the electrochemical reaction occuring at the electrode surface. The mass transport impedance is Zm and R8 is the solution resistance which is a function of the electrolyte concentration in the cell solution. When a current impulse, I, as in Figure 1-13, is applied as an excitation function to this equivalent circuit, the response function, E, is observed as shown in Figure 1-10. At the begining of the impulse, the potential rises immediately until it is equal to 1tRs' Then the potential increases gradually as Cd charges until the current pulse is ended. The potential now drops by itRS. After this point, the discharge of the double-layer capacitor dominates the potential variation in the circuit. In the absence of mass transport limitation, the potential follows a simple exponential decay law with 7k: qt=0 exp (-t/Rde). The exchange current Io of the reaction can be calculated directly from the slope of the logi[ 1g t curve from the relationship: Slope = (nF/2.303RT) (Io/Cd) Ei Fi [—me—ZWCD- s f A, EQUIVALENCE CIRCUIT FOR ELECTROCHEMICAL CELL I lit _L, ark— t 4~> Ba EXCITATION FUNCTION: CURRENT IMPULSE 71 3 I1 OexP (“t/Rfcd). t 3> c . POTENTIAL RESPONSE Figure 1-1. Equivalent Circuit and Current Impulse Response of Electrochemical System at th EU! OCC the the curv cell the enha Coul the In the case of a reaction that is mass-transport limited throughout the pulse response, the potential decay follows the equation : _ + 2nFc°D% % 1L " Hzcd t In either case, the double-layer capacitance Cd can be obtained experimentally by the ratio of the total charge in the current impulse to the initial potential‘qlt=0. For an accurate determination of 7Lt=0’ the pulse duration must be short enough so that a negligible amount of reaction occurs during the pulse time. There are two significant advantages associated with the coulostatic technique : the double-layer capacitance at the electrode can obtained from the same potential-time curve and the measured potential need not be corrected for cell resistance because there is no current through‘R8 when the potential is being measured. The capability of a coulostatic system is greatly enhanced with an on-line computer. The computer controlled coulostatic system improves this polarization technique in the following aspects : is by l the Pole CUI‘I‘ of a capab Plati Plati adsor COVGr be Ca a Cle aCch 1. Capability of performing other types of polarization techniquesa(50). 2. Versatility of experiment control. 3. Ease of data processing including signal averaging. These advantages are illustrated in the following examples : Controlled potential polarization has been achieved with the computer controlled coulostatic system. The potential is maintained within a few millivolts of the desired value by constantly adding short charging pulses as necessary to the electrical double layer (30). Controlled current polarization is realized by adjusting the pulse width, the current amplitude, and the pulse repetition rate. The experiment can be programmed according to the need of a particular electrochemical system under study. This capability has been critical in the study of the mercury- platinum interaction reported here. The behavior of a platinum electrode is greatly affected by trace impurities adsorbed on the platinum surface, and by the degree of coverage of the oxide film on the surface. The electrode must be carefully pretreated Just before the experiment to generate a clean and reproducible electrode surface. This is accomplished by programming the potential of the electrode. - 7 - The electrode is first held to a positive potential to oxidize all impurities and then is changed to a negative potential to remove the oxide film on the surface. Improvement of the data collected is realized by the averaging of the data obtained from repetitive runs of an experiment. The signal-to-noise ratio increases at a rate proportional to the square root of the number of averages for random noise. The digitization error which arises from the resolution of analog-to-digital conversion is improved as well. B. Instrument The computer controlled coulostatic system is shown in Figure 1-2. It consists of a minicomputer and its peripherals, a pulse generator, a current generator, a sample cell and an analog data conversion unit. All controls and data collection are performed by the minicomputer under program control. When a current pulse is needed, the computer sets up the pulse width, pulse amplitude and the current channel number. Then it issues a pulse command upon the request of the Operator or of the program. The data collection sequence is started at a predetermined rate. These data are either transferred to a mass storage device for future analysis or sopmhm oaemumoaaoo coHHonpmoo nopsmsoo OHF Ammo m4mzma+ 1\\4 m a" 111”; a m “u >ma. 1\\A m ..)/fi_. a a >na. - 13 - National Semiconductor CO.. These switches have a switching speed of 150 nsec and an "0N" resistance Of 3051, which allows extremely short, reproducible current pulses to be produced. The high speed Operational amplifier is a AM405-2 manufactured by DATEL CO.. It has a slew rate of 150 Vflhsec with a settling time of 400 nsec. A small capacitor, 20pf, is used to stablize the Operational amplifier. The voltage control is shown in Figure 1-6. The device select codes from the interface buffer are decoded by two BCD decoders. The decoders were wired to decode device 14. The device select signal is gated by the timing pulse, IOP2. When both are issued from the CPU, a command pulse, NEW AMP. CMD. is produced which latches the pulse amplitude data from the interface buffer into three A-bit data latches (5,8, and 12). These latched data control the voltage output of the DAC. The DAC is a 12-bit digital to analog converter, DAC 850 from Burr-Brown Co. The DAC was wired to produce voltages between +5.000 V and -5.000 V. Since four channels are available, the current generator provides four decades of current range with the same resolution. The maximum current is 3 2.5 mA. - 19 - IOPZ NEH AMP. CMD. >—1>D 2H V OUT le-o-SV Figure 1-6. Digital to Analog Conv. Card 88 V0 pri dec wit and Elec for fast manue an at Volta -20.. c. Data Acquisition System This section of the instrument amplifies the voltage sensed by the reference electrode, digitizes the resulting voltage and transfers the digitized data to the CPU under program control. The circuit is shown in Figure 1-7. Three control signals are produced by the device select decoders and the timing pulses. Device code 56 is combined with IOP1, IOP2, IOPA to generate commands CONV. CMD., READY and DATA IN CMD. respectively. As shown in Figure 1-5, the voltage at the reference electrode is sensed by a follower amplifier. A CA31AO is used for its extremely high input impedance of 10129 and reasonably fast response (slew rate 8 Vfins). The voltage is amplified by a second CA3140. The amplification factor is selected manually with a rotary switch. The output voltage is fed to an absolute value amplifier which produces a positive output voltage for an input voltage of either polarity. After the conversion, the voltage Vin is fed to the sample and hold circuit in Figure 1-7. An HA2425 gated op amp was chosen to perform the sampling and holding of the voltage input, Vin. (An option is also available which uses a follower Op amp (CASIHO) instead of the absolute value amplifier.) In this case, two jumper wires on the ADC have to be adjusted .. a] - DATA 1N CMD. II!!! 3 COM V. (MD SKF> T 0 1,e,ie,13 C 11 U FIGURE'1-7. DATA ACQUISITION SYSTEM -22.. accordingly to accommodate i 5 V voltage range.) When the data conversion command, CONV. CMD. is issued, two short controlling pulses are simultaneously generated by monostables, M1 and M2. M1 generates a 11‘nsec long pulse. This pulse is connected to the gate control of the sample and hold circuit. The output amplifier of this gated op amp is isolated from the input amplifier and the voltage is held at the same level by the holding capacitor during this pulse period. The other monostable M2 generates a l‘usec pulse. The falling edge of this pulse triggers the analog to digital conversion sequence of a A/D converter. The converter used is a ADC-12Z from DATEL, block 10. It takes 8Jpsec to do one conversion. The control signal READY is issued by the CPU during this period to check the status of M1. When the conversion is finished, Q of M1 goes to HI. This makes the SHE line go L0 and clears the flag bit of the control and status register of the DR11-C. The CPU acknowledges that and issues the DATA IN CMD. signal which latches the data from the ADC to three 4-bit data drivers (blocks 8, 11 and 12). These data are passed to the input register Of the DR11-C and processed by the computer. - 23 - C. Performance of the Instrumental System The most important performance factor for a coulostatic instrument is the ability to deliver a precisely known amount of charge in a short time. Ideally, a current pulse of rectangular shape is generated by the pulse generator and the charge content Of the current pulse is equal to the product of the pulse width and the current amplitude. However, several factors undermine the performance Of the coulostatic system, causing the system performance to deviate from the ideality. It is critical to determine these factors and evaluate their corresponding effects for the usage of such system. 1. Scope of the Instrument Performance a. Pulse Generator Although the timing pulses at the output of the timebase gate are the complete pulse cycles, the pulse width produced is not an exact integer multiple 0f,fl8. This is due to the delays caused by the output Flip-Flop and the logic gates. These delays increase the pulse width by 0.56,ps. The shortest pulse produced is then 0.56jus. b. by and to err set How rel cha: app; in i the Thi PGSJ imp; Perc iUs: Chen The I the the} 81 8 dis. - an - b. Current Generator The performance of the current generator is controlled by the digital to analog converter (DAC), the current switch and the Operational amplifiers. Since the 12-bit DAC is wired to produce 3,5 V, the resolution is 2.4A14 mV. The relative error caused by the resolution depends on the voltage setting. It increases as the voltage setting becomes smaller. However, since four current channels were implemented, the relative error can be reduced by switching from a higher amplitude current channel to a lower amplitude current channel. The current switch has an on-resistance of approximately 3011.. The current limiting resistor is 2 KII in current channel 1. The on-resistance of the switch causes the current to drop by 1.5 % from the theoretical value. This error was offset using the matched current setting resistor. A current limiting resistor of 1961flis actually implemented which reduces the error to within.one half percent. The on-resistance of the current switch causes only insignificant error for other current channels. The other characteristic Of the current switch is the switching speed. The typical value is 150 as. The longer the current pulse, the less error is caused by the rise time limitation. When the current pulse is very short, this factor become more significant because the shape of the current pulse is distorted. Ce Th is The de} I‘GC and 35’ To 01’ 1’01 the tes - 25 - c. Operational Amplifiers The slew rate and the settling time of the Operational amplifier also affect the shape of the current pulse produced. These two effects are more significant when the current pulse is short. d. Data Acquisition System Since the analog to digital converter is wired to give a 10 V range, the digitization error is thus 2.4A1h mV. The relative error contributed by the digitization error depends on the voltage to be measured. Its magnitude is reduced by the prOper selection of the amplification factor and by the data averaging technique employed. In summary, the current pulse produced by the coulostatic system is more ideal when the pulse width is not too short. TO use a very short current pulse, the actual charge content Of the pulse needs to be determined experimentally. In the following sections, the shapes of the extremely short, high amplitude current pulses are examined, and the results of the charge content determinations and the reproducibility tests are discussed. 2.. anq in] dun A 1: ins the Wit f r0 an; the Thf is -26- 2. Shapes of the Extremely Short Current Pulse The extremely short current pulses were observed with an oscilloscope. The results of two such pulses are shown in Figure 1-8. a. 0.56 ps Pulse A 2.5 mA, 0.56,ps current pulse was dumped into a dummy cell which consists of a 5 K11 precision resistor. A triangular wave was Observed as is shown in Figure 1-8.A instead of the expected square wave. The half-width of the pulse is 0.5‘ps. The peak height is only 6 V in contrast with the theoretical value of 10 V. This discrepancy comes from the relatively slow response of the buffer Operational amplifier which is unable to drive the current pulse through the cell from 0 V to 10 V during the pulsing period. This pulse is not practical because the charge it carries is too small to be useful and the charge content of this pulse varies depending on the resistance of the cell solution. be 1056 )18 Pulse A 1.56,us long, 2.5 mA high current pulse was dumped into a dummy cell which consists Of a 1 Kit, 1 % resistor. The appearance of the wave is shown in Figure 1-8 B. The half-width - 27 - Figure 1-8. Shortest Pulses - 23 - of this wave is 1.56 ms and the amplitude is 2.5 V. Voltage ringing is present as a result of the sudden voltage change. The same current pulse was dumped into a resistor of 5’K2 The result wave is shown in Figure 1-8 C. In this figure, the horizontal scale is 1‘ps per division while the vertical scale is A V per division. A trapezoid is observed. The voltage ringing is still present but is reduced as compared with the previous wave. As the pulse width increases, the current pulse approaches that Of an ideal square wave. Figure 1-8 D shows the voltage pattern Observed at the input of sample and hold circuit after the voltage of Figure 1-8 B is amplified eight times. The horizontal scale is 0.5,ps/div as shown in Figure 1-8 B, while the vertical axis is 5 V/div. The distortion is quite Obvious. This is because the operational amplifiers used for the amplification and voltage follower are the type CA3140 which have a slower response than the buffer Operation amplifier. However, this distortion problem does not affect the data acquired ' because the fastest data rate is only 13,»s per datum point which allows sufficient time for these amplifiers to follow the voltage input. 3. C' a. D gener conte calcu anpli of th of th Pulse Pelat Pulse short dummy The v 'T‘ «he p FTN a Pregr or th QUART vaUi - 29 - 3. Charge Content of the Current Pulse a. Dependence of the Pulse Charge Content on the Current Intensity Since the current pulses produced by the current generator are not exactly rectangular in shape, the charge content of the current pulses are different from the values calculated using the equation Q = ’1, x 4‘, , where 4, is the amplitude of the current pulse and Q is the charge content of the current pulse of "(.118 long. Thus the charge content of the current pulses was determined experimentally. The pulse durations chosen for the following examples are relatively short, because the distortion of the current pulse becomes significant when the pulse width is very short. A 0.01 pF, 0.5 % precision capacitor is used in the dummy cell to store the charge carried by each current pulse. The voltage developed is measured under program control. The program used is called QUANT which consists of a FORTRAN. FTN and two subroutines QUANTM.MAC and AMP.FTN. The main program reads in the current channel to be tested, the name of the data file to be created and then calls the subroutine QUANTM to perform the discharging and the subsequent data acquisition. The amplification factors are manually selected. Cl. Cu: 3-5 res 31. thre expE eQUe data Stan AGV1. and I this - 3o - These factors are stored in the subroutine AMP.FTN and are retrieved by the main program to process the acquired data. The pulse command is issued by the Operator. After each current pulse, an indicator light turns on which indicates that the processor is ready for the next pulse command. Five pulses of different width were chosen for each current setting. Five current settings were examined each case. Each of the four current channels was examined. 1). Channel 1 Figure 1-9 shows the charge content vs. pulse width with current channel 1, using negative current pulses. Curves 5, 4, 3, 2 and 1 currespond to the amplitudes of 2.500 mA, 2.000 mA, 1.500 mA, 1.000 mA and 0.500 mA respectively. The pulse widths are 1.56 ms, 7.56 ms, 15.56 )18, 31.56,us and 39.56,us. Each plotted point is an average Of three data points from three separate runs of the identical experiments. The length of the data point in the figure is equal to twice the standard deviation associated with that data point. It seems that curve 2 carries the greatest standard deviation. With a 31.56lps pulse, the standard deviation is 3.3 %. With a 39.56,ns pulse, it is 1.2 %, and within 1 % with others. Other experiments showed that this is not always the case. 1.2 n. . JDDUOKQHZ CF) MOIaummoz .m Hangman peoasoo mmumno omasm on» mo momoam .mu. canoe - g3 - 38.0 H 33.0 a 8m 893% «3103\Hsooonm: wmwm.o omwm.o comm.o omwm.o ommm.o 00m maVasooosma :N:No.o mum—0.0 Ob¢—o.o mmmoooo wmmoo.o omoam «a o.mm 0.0N o.mp 0.0— o.m ovzuddmad pumhhfio paonaso 0>wpfimom .m awesome unopaoo owumno ended on» we mmmoam .mlp magma 999 A1 513‘ thrc At 1 bale the are - 44 - 0). Channel 4 The pulses chosen are 119.56 1.18, 99.561118, 1199.56 ps, 999.56/ps and 3999.56‘us long. The current amplitudes are -o.5oo M, -1.000 ml, -1.500 pA, -2.ooo )1A and -2.500 pm. A log-log plot is used to display the data as in Figure 1-15. The first two data points of each curve deviate significantly from the straight line drawn through the other three points. This is caused by the offset of the amplifier. At such level of charge quantity, the amplifier needs to be balanced very carefully. In addition, any stray charge from the environment affects the result significantly. The charges 11 are in the range between 10' coul. and 10-9 coul.. .DOUOEQHE «mwc<10 004 _ 45 - POTENTIAL DECAY O. 900—- _'u 'h :3 O _- I L) o / DC 0 “ . H 5/ E /N . W -_ 3/ Nu CD c: ‘ 3 ' < I / U —— 1. (D C) _ .J 1‘ '6.@@@ t g 1 1 1 IL 1 1 1 TL I 2.000 9.898 Figure 1-15. LOG PULSE WIDTH, ,usgc Cathodic Charging Curve with Current Channel A - 45 - b. Dependence of the SCC on the pulse width The data obtained in the previous sections are used to calculate the specific charge content for each pulse width. The results are listed in Table 1-7 and Table 1-8. Table 1-7 includes the results obtained using current channel 1. Table 1-8 includes results obtained using current channel 2“ and channel 5. Table 1-7. Variation of the SCC with the Pulse Width for Current Channel 1 P' W. ’ h from I=ml 1.56 7.56 15.56 31.56 39.56 slope** 2.5 1.012 0.996 0.992 0.992 0.980 0.993 2.0 1.011 0.992 0.993 0.991 0.993 0.989 1.5 1.015 0.990 0.990 0.992 0.991 0.988 1.0 1.001 0.991 0.995 0.979 0.984 0.986 0.5 1.080 1.000 0.999 0.993 0.990 0.993 -2.5 0.988 0.989 0.993 0.993 0.993 0.983 -2.0 1.001 0.989 0.995 0.999 0.993 0.987 -1.5 1.001 0.976 0.990 0.993 0.993 0.988 -1.0 0.978 0.989 0.990 0.992 0.990 0.986 ~0.5 0.954 0.989 0.983 0.999 0.996 0.977 ~ - ~31\ I H Table 1-80 - 47 - a) Current Channel 2 Variation of the SCC with the Pulse Width for ‘.w I '1'“; 79.56 1119.56 299.56 mm" ’mA slope 0.250 1.000 1.000 1.000 0.991 -0.250 0.992 0.996 0.992 0.989 * This value is exceptionally high. b) Current Channel 3 tOw,p8 I 999.56 999.56 1999.56 fromu ’ slepe 25.0 0.996 0.996 0.992 0.984 -25.0 0.984 0.986 0.992 0.988 ** obtained from section a th 10 - 48 - c. Reproducibility The reproducibility of the system is generally quite good. A typical case is shown which utilized a 1.56/us wide, 2.5 mA current pulse. The set-up is the same as the charge content determination experiments. The results are listed in Table 1-9. Table 1-9. Charge Produced by a 1.56Jus, 2.5 mA Current Pulse in nanocoul. 3.891 3.899 3.899 3.925 ‘ 3.861 3.876 3.896 3.913 3.831 3.325 average a 3.8841: 0.03118 nanocoul. The reproducibility of the pulse generator is better A than 1 % for consecutive tests. Yet the long term drift is less favorable, approximately‘i 2 %. 3. the is : of 1 CaUSe disc< m). redU1 stat.J - 49 - 3. Discussion No regular pattern of SCC variation was observed when the current amplitude was changed. This indicates that SCC is independent of the current amplitude and that the variation of the SCC is caused by other parameters. Except for the very short current pulse, the SCC is 1 % less than unity, the average value is 99.2. With the shorter current pulse, the SCC value appears higher than the averaged value for the anodic pulse and lower than the . averaged value for the cathodic pulse. Thus it is imperative to run calibrations along with the experiment when these short current pulses are used in order to obtain a more precise results. The SCC is used for the longer current pulse without significant error. Although the reproducibility of the current pulse in the consecutive run is better than 1 %, the instrument does show a long term drift. This drift causes the result to vary by approximately': 2 %. It was discovered that most of the errors came from a defective A/D converter. With a new A/D converter, the error is reduced to below 1 %. The error has two types, the systematic error and the statistical deviation. The systematic errors originated from the the from erro the A de be 1. tech: tech: 0388‘ capa' can ' with. as r. the For the - 50 - the calibration error and the non-ideal characteristics of the system. This type of error can be compensated for by the careful calibration work. The statistical error comes from the noise pick-up from the environment, the digitization error as well as other noise sources such as shot noise in the system. This type of error can be quite large. A deviation of several percent is possible. Such error can be largely corrected by employing various data manipulation techniques, such as data averaging and data smoothing techniques. These techniques are quite effective, in many cases, in reducing the statistical errors. In summary, the coulostatic system constructed is capable of delivering current pulses as short as 1.56,ps at i 2.5 mA. The charge content of current pulses delivered can be calculated for the current pulses over 104us long to within 1 % error. The potential variations can be followed as fast as 134ns per data point. This data rate approaches the conversion time, 8 us, of the A/D converter used. For any faster application, both the pulse generator and the data aquisition system would have to be modified. - 51 - D. Application of the General Purpose Coulostatic Instrument 1. Determination of the Double Layer Capacitance and the Faradaic Resistance The capacitance and the resistance of a dummy cell were determined with the coulostatic technique. The dummy cell as shown in Figure 1-1 A, consists of a 0.01 pF capacitor in parallel with a precision resistor which is either 100 K51 or 200 K11 and a resistor RE which is 250 Kit. A program called CAP} was written to perform the experiments. This program is composed of a FORTRAN main program CAPFS. FTN , a MACRO subroutine CAPM3.MAC and FORTRAN subroutine LINFIA.FTN. The main program reads in the pulse width, pulse height data, data points to be fitted to an exponential equation and the name of the data file for storing the collected data, then it calls the subroutine CAPM3.MAC to deliver the current pulse and collect data. The first data is collected 10‘ps after the pulse is initiated. One hundred data readings are collected at intervals of 39.58,ns. The logrithm of the collected data is calculated and fitted to a linear equation : 1‘3. 10g 111 = log 119 - R C VJ 3. 881 3610 the - 52 - where 111 is the potential at time ti’ no is the extrapolated value of 111 at t=0. The capacitance C is calculated by dividing the charge of the pulse by 710. Current channel 1 was used exclusively in the following experiments. Large amplitude current pulses as well as small amplitude current pulses are included for comparison. a. Large Amplitude Current Pulses Figure 1-16 shows the potential vs. time curve of several large amplitude current pulses. Curve 1 corresponds to the decay from a 3.56Ips long, -2.000 mA current pulse. Curve 2 corresponds to that of a 1.56,us long, 2.000 mA current pulse. Curve 3 corresponds to that of a 1.56jus long, 1.000 mA current pulse and curve A corresponds to that of a 1.56Jps long, 0.500 mA current pulse respectively. Each point represents a single data reading. The potential decay appears exceptionally smooth. These data were fitted to a linearized exponential equation and replotted in Figure 1-17. The capacitance and resistance are calculated accordingly. Table 1-10 lists the results obtained from these experiments plus some additional experiments. 8. h 40>! - 53 - POTENTIAL DECAY 8. 08¢- MVOLT X10 1 I TIME MICROSEC. x10 '3 Figure 1-16. Potential Decay of a Dummy Cell with High Amplitude Current Pulses. 0:0.01 M“, R=1OO Kn IF uCNl-‘d F187; - 54 - POTENTIAL DECAY 8. ODD" MVOLT 1 l - X10 0... ': TIME MICROSEC. x10 " Figure 1-17. Curved Fitted Potential Decay of a Dummy cell " . with High Amplitude Current Pulse. C = 0.01 m“, R: 100 Km - 55 - unosaaoaxo 000:» ad eosnomaoa mes sodnooaaoo com oz ” 0002 a _o.o .flum co. u m amé .0156uo mm._o_ mm.aa mm..e_ me.ne_ me.ne_ e.aa Ame . ex.m 880.6 868.6 $86.6 $86.6 886.6 226.6 .3 . 036 <3 0000—1 OOOoNl 0000—1 oomoo 0000— OOOoN adfivfiflmad vdohhfio m1. mm.n mm._ need; eeaem emaem eeeeaaeee eeeeuee amen eves eeeeeeaeem ofidodhdh and mondpdodmdo HHoo «0 defiymnflahoaon .opl— sands V8 th ca :11 da da pul 0b1 P11] Cu: f0 (11 f0 In 1 Th -55.. The normal value of RC is 1.0.nF-K1L. The experimental values with a 5.56.ns current pulse shows a better agreement than values obtained with a 1.56,ns current pulse. In both cases, the RC product agrees with the predicted value to within 3 %. With these high current amplitudes, the fitted data shown in Figure 1-17 completely overlaps the original data shown in Figure 1-16. b. With Small Current Amplitude Pulses The potential decay vs. time with small current amplitude pulses are shown in Figure 1-18. Curve 1 was obtained with a 1.56 ms, 200 pA current pulse. Curve 2 was obtained with a 1.56.us, 100 nA current pulse, curve 3 and curve A were obtained with a 5.56‘ns, -50‘uA and a 1.56.ns, +50.uA current pulses respectively. These curves appear noiser than the curves obtained with higher amplitude current pulses. Except for curve 4, all the noise could be attributed to the digitization error. Since the amplification factor is 16 for all these data, the digitization error is i 0.15 mV. In curve A, the small data values are buried under the noise. The maximum voltage in curve A is 2.0 mV. Two techniques were adopted to improve the signal to noise ratio of the collected data. One of them is signal averaging. Curve 5 in Figure 1-18 shows the result using 2. hno>z F11 -57... POTENTIAL DECAY 2.600” ~1- I— on .J 1 C3 :> CD .:1 IE '4 ‘3 )< fir . 3 1 . '9 002... 0.000 4.000 TIME MICROSEC. X10 Figure 1-18. The Potential Decay of the Dummy Cell Using Some Small Current Pulses W1 re he Ch. d0 - 53 - data averaging..A Fortran.routine called AVG was written to do this Job. Corresponding data from eight separate experiments were averaged together. The current pulse used was 2.56 [.18 long and +50 pA high. The improvement is significant. The digitization error is greatly reduced. The second technique is to fit the equation to the linearized exponential equation using the first 20 points. The results are shown in Figure 1-19 with curves I-A. The capacitance and the resistance of the dummy cell were calculated from these fitted data. Table 1-11 lists the results. As the charge of the pulse become smaller, the capacitance and the resistance obtained experimentally deviate more from the correct values. This deviation comes from the inaccurate estimate of the real charge content in the current pulse used. Although in each successive run the charge content is within 1 %, it does drift over time. The accuracy of the resistance measurement relies on the accuracy of the capacitance measurement which in turn depends on the accuracy of the charge determination. The accuracy of the time constant, RC, does not depend on the charge determination. - 59 - POTENTIAL DECAY 2 . 8 0 0" -+. b-'~ . .J 1 C3 :> CD ' 1!: v4 1 >< . . ........... "'e ....... 1 ....... ' " "Zia-.0: ...................... - M "5.333122::::::-~:::: """""" a ......... ' ----- . ...:°;;;:,:2:::::::m;;;;;;:: 9 - 9 9 9 WWW-H 0.000 4.000 TIME MICROSEC. x10 ’ Figure 1-19. Improvement of the Data through the Curve Fitting and Signal Averaging Technique -60- .0000 omen» no cosuomnea use sedueeunoo com 02 ” ouoz a .o.o .nnm .00. u m Amesam> Hmauonv am.o.el_o.ouo wm.~m mm.¢m— w.mo— pm.Nm N.—o— m¢—oooo nanocoo mmmoooowmopo.o :mm0000 mmm.m no.0. no.0— Nw:.w mmm.m 0.0m! 0.0m 0.00—1 0.00— 0.00m1 ee.em nee . exem eopo.o hex . exec mmwm.m um 40. o.oom+ eeeeaaeae eeeeeeo an.“ «new mm“ . «36H; 00.25 mondsm ovzufidnad unohhfio Hanan Ava: oondumdmom Oddvundm dad ooddufiodmdo Hdoo «0 Odaddfiahovon .p—l— wands 2. P< perfc This the 1 requ: the ‘ inje: retUJ inst; the a Call: frOn Pote: Simpj ampl' The . 1~20 at t‘ the the 1 caUs. 2. Potentiostatic Polarization with the Coulostatic Instrument The potentiostatic polarization technique can be performed by the computer controlled coulostatic instrument. This is possible because the computer continuously monitors the potential of the electrochemical cell and makes the required adjustment in a very short time. A small shift of the potential value causes the computer to issue command to inject additional charge which forces the potential to return to the predetermined value. The precision of the potential maintainence by this instrument relies on the electrochemical system as well as the algorithm of the program used for the purpose. A program called VDEPO was written for the deposition of the mercury from the mercurous perchlorate solution at a constant potential. The algorithm for the process control is quite _ simple. It uses a continuous cathodic current with varying amplitude. The potential to be maintained is 200 mV vs. R.H.E.. The variation of the potential vs. time is shown in Figure 1-20 a. A current pulse of very high intensity is applied at the begining until the potential drops to 200 mV and then the current amplitude is reduced which eventually brings the potential back to about #00 mV. The anodic potential causes the computer to increase the current amplitude again. ../IN..®®® .0 nopmnosow oHumpmoHsoo on» an coaumsuumaom oaumumowpsmuom woumHsEHm .omnp meandm -62- 01X 'SA 'AN 01X 'SA 'AH ‘0‘ of o (a 00w 0 Se. 0 ‘0 00' o o o o C o no a o o o o O o 000‘ e o 9 so». 0 to. O. o co .0 o o .00 o e. or Ono-9'0 ‘e‘ ooe’ofooc . o 0‘00 no" o‘looooa. 00‘, 0.000... I o‘oi‘oe‘oxogl‘fi 3'H'H 3'H'H 3;. ./,oo.~. -63- The resulting potential shows a damped oscillation which converges to a constant value within approximately 0.1 sec. Figure 1-20 b, shows the same potential-time curve at an expanded scale during the 200 sec deposition period. The computer is able to maintain a fairly constant potential after it converges. The oscillation of the potential observed using this technique can be substantially reduced with a more sophisticated program. However, the program deveIOpment is time consuming. Better results can easily be obtained with a regular inexpensive potentiostat if the special features of capacitance measurement, no IR error, quantized charge addition, and change in polarization technique in mid-experiment are not needed. 001 ca] Dur I‘ed; - 64 - CHAPTER II INTRODUCTION TO THE STUDY OF THE MERCURY FILM ON PLATINUM A. The Search for Mercury Film Electrodes Despite the well-known problems associated with the environmental hazard, mercury has played and will still continue to play an important role in electroanalytical chemistry. One of the reasons is that mercury electrodes provide high sensitivity and reproducibility unmatched by other types of electrodes used for trace metal analysis. In many cases, the sensitivity is even higher than with spectroscopic techniques. This is demonstrated by the technique called anodic stripping voltammetry (48). In this technique, the mercury electrode is first kept at a selected cathodic potential for a predetermined time. This is commonly called the preconcentration step in the stripping analysis. During this step, the trace metal ion in the solution is reduced on or into the mercury. After preconcentration, the electrode is allowed to sit idle for some time, typically several minutes, without any current passing. After this period of time, the electrode is made anodic to dissolve the metals that had accumulated during the preconcentration step. By analyzing the resulting stripping response, the concentration of the metal in the original solution can be 50 Dr E18 to F1 -55.. determined. Successful determinations down to several picogram per m1 concentration have been achieved (#8). The sensitivity achieved by this method depends on several factors : the experimental conditions during the preconcentration, the eleCtroanalyticalmethod adopted for the stripping step (49), and the electrode used. One of the most commonly used electrodes for the stripping analysis is the hanging mercury dr0p electrode. However, detailed analysis of the stripping technique indicates that increasing the area to volume ratio of the mercury results in a significant improvement of this technique. This causes an increasing attractiveness of the mercury film electrodefbr application in stripping analysis. The advantages of using mercury film electrode can be seen clearly by the following comparisons : During the preconcentration step the concentration of the interested metal in the mercury is given by equations (1) and (2) for HMDE (Hanging Mercury Drop Electrode) and MFE (Mercury Film Electrode) respectively (50). whe inc in -55.. 5 I (1) C = Li: mole/cm3 41rrg nF I (2) C = L z mole/cm3 A 1L nF where IL : limiting current in Amperes : time of preconcentration in Seconds d ’1 . radius of the mercury drop, typical value 0.5 mm L : thickness of the mercury film, normally 1-100 )111 n oxidation state of the metal ion “*1 : Faraday constant Under the same preconcentration process, this means an increase in concentration by two to three orders of magnitude in the thin film of one micron thickness as compared to HHDE with the same surface area. There is also an advantage of the MFE in the duration of the rest period. Equations (1) and (2) are derived under the assumption that the metal distribution inside the mercury is uniform. This condition is met some time after the preconcentration is stopped. For an HMDE of usual dimensions, this uniformity is achieved approximately 200 sec after the electrolysis. With an urn of 10'”3 cm thickness, uniformity is reach perfo. for t1 mercu might metal the r. the re and c: is it: Quite the t other with Chara This allow and Surf trad~ advaz Stabi -57.. reached to within 0.5 % in 5-4 sec. During stripping, the performance of these two electrodes is compatible except for the extremely thin mercury film. When an extremely thin mercury film is used, the dissolution of one metal element might be completely finished before the ionization of another metal with a close oxidation potential. This effect improves the resolution of this technique. Since the film is built on some rigid, inert support, the resulting electrode provides good mechanical stability and could be used under severe conditions. One such example is its use under high speed rotation. This property is quite important during the preconcentration step to decrease the time needed to reduce a desired amount of metal. The mercury film electrode could also be used in many other electrochemical applications replacing other electrodes with improved performance. Such electrodes exhibit characteristics similar to that of the mercury dr0p electrode. This includes the high hydrogen reduction overpotential which allows the observation of a wide range of cathodic processes and a good surface reproducibility of the reformable liquid surface. The MFE could be used in many applications traditionally performed by the dropping mercury with the advantage of easy electrode handling and good mechanical stability. Shm r88 -53.. B. Mercury Film Preparation To be a candidate for the supporting material for the mercury film, a material should possess several qualities : good electrical conductivity, wettability with mercury while maintaining chemical inertness toward mercury, and chemical inertness toward metals dissolved in the mercury. No material that has been used meets all these criteria. These materials include various constructions of graphite, silver, platinum and nickel. The graphite based mercury film has drawn the attention of many researchers lately. The materials tested include glassy carbon (58), graphite-epoxy (55), wax impregnated graphite (W10) (56, 57), and radiation cured polymer-graphite (39). The mercury film.could be made extremely thin on these electrodes because the mercury forms no compound with carbon. The sensitivity is quite high with these electrodes. However, an examination of the films made on these electrodes showed that the mercury forms draplets instead of a smooth film (58). These electrodes are quite stable in basic or neutral solution but deteriorate quickly in acidic media. The deterioration rate depends on the construction of the electrode. The glassy carbon which is the most expensive among these electrodes shows the best stability (60). The reproducibility of the results is not good. This is not surprising knowing that the drc fil dis com as . 8ta1 the: and 6-8 that With resu film Oste 18 c. The amal thi0] £11D SEVe is e -59.. dr0plets change sizes with the potential employed for film making (58). Gold (51, 52) is readily wet by mercury. However, gold dissolves in mercury and forms a series of intermetallic compounds. The resulting film is thus not pure mercury and as such as is useless for reducing other metals for stripping analysis. Using silver as a substrate, Zakharov and Trushina stated that the results obtained with mercury films thinner than 2.um are irreproducible (53). According to Igolinskii and Stromberg (54), Silver electrodes with mercury films of 1-5 mm thickness gave reproducible results only during 6-8 sec after plating, while Gardiner and Rogers (6) reported that a few days of reproducible performance were obtained with 2-4 pm thickness film. These seemingly contradictory results reflect the fact that the property of the mercury film depends on the film preparation procedure. StoJek, Osterpczuk and Kublik (6h) found that this irreproducibility is caused by the amalgamation of silver at the contact. The amalgamation rate of silver is greatly reduced as the amalgam thickness is increased. By controlling the thickness of this amalgam layer, they found that mercury films as thin as 0.1,nm gives reproducible results within several hours of preparation (62, 65, 60). This improvement is encouraging for the silver based mercury film electrode. most 5000 were were Unde wire DGPC‘ with surf. 1‘9 311. Vand 111:3 - 7o - Platinum, due to its chemical inertness, has been the most widely used solid electrode material. It has been successfully used as the supporting material for stationary mercury electrodes (12, 15, 59, A0). Various approaches were utilized to make the mercury platinum contact. Underkofler and Shain (59) etched the tip of a platinum wire sealed in a piece of soft glass and then deposited mercury drop on it. Ramaley, Brubaker and Enke (15) plunged the electrochemically reduced platinum into a mercury pool without interrupting the current to wet the platinum surface with mercury and then deposit more mercury on the resulting platinum surface from mercuric nitrate solution. ‘Vandeer’Leest (40) merely immersed the platinized platinum in mercury pool. Moros (12) made the mercury platinum contact by immersing the platinum in sodium amalgam. Provided that the contact is perfectly covered with mercury, the polarization curves obtained with such electrodes differ very little from those obtained with the HMDE. This type of electrode has been used for as long as two months with no detectable changes in its residual current or sensitivity (6). While making a platinum supported hanging mercury drop is easy and reliable, the construction of a smooth mercury film on platinum is difficult. When a small amount of mercury is deg fou Thi mer big is a 1 pro Cons Cont the obta OCC -71- deposited on the platinum, mercury droplets are frequently found instead of the desired smooth mercury film (7. 10). This phenomenon causes no problem if a large amount of mercury is deposited because small droplets grow into bigger drops and eventually a coalesced mercury coating is formed. As long as the mercury coverage is complete, a large drop can form that should have electrochemical properties similar to the HMDE. The success of making the smooth mercury film depends on some empirical step in the process. Marple and Rogers (7) made the mercury film by plating mercury from saturated mercuric nitrate followed by cathodizing the electrode in saturated potassium chloride solution. The last step is essential and must be repeated frequently. Neeb (10) prepared mercury plated platinum electrodes by electrolysis at -1.5 V vs. S.C.E. in 0.1 % Hg(N03)2. The electrode obtained consists fine mercury draplets. By using the solution containing 0.005 % Hg(N03)2, 0.5 M NH 0H and 0.05 M EDTA at 11 the same cathodic potential, a smooth mercury coating was obtained. At this potential vigorous hydrogen reduction occurs. Decreasing the cathodic potential to obtain a 100 % current efficiency for the mercury reduction reaction in the same solution produced mercury droplets. Hartley, Hiebert and Cox (14) employed a current of 10 mA/cmz to deposit mercury from a solution containing 0.05 M Hg(NO and 0.1 M HClO 3’2 solution to form mercury drOplets. The electrode 1. WE 216 of de and - 72 - was then subjected to a high cathodic current in 0.1 M HClO,+ with hydrogen gas generated by the electrode reaction. Smooth mercury films were observed. Except for the Marple and Rogers' method, the success of these methods to produce smooth mercury film seems to depend on the large cathodic potential employed which greatly reduces the surface tension of the mercury droplets. In Marple and Rogers' method, two 1.5 volt batteries were used with saturated Hg(N03)2 solution as electrolyte. The potential would have been quite cathodic at the working electrode. One other important but less apparent factor in the successful method is the conditioning of the platinum electrode. The role played by the K01 which was used before and after the deposition of mercury in Marple and Rogers' method is not clear. Hartley, Hiebert and Cox's method employed an electrode cleaned by dissolving the mercury deposited on the platinum electrode with hot nitric acid. The platinum surface obtained this way is quite different from the fresh platinum surface obtained by mechanical polishing. The surface after removal of the mercury appears spongy showing brown to black color depending on the degree of the reaction between the platinum and the mercury at the contact (17, 29). mer01 that nerc1 in t1 film Co II Pole: (13, POte Vs. the : inte: reSn. - 73 - One problem associated with the creation of smooth mercury films by the induced merging of mercury droplets is that little control can be exercised at the interface between mercury and platinum. Ion or solvent molecules may be entrapped in the interface and thus affect the behavior of the mercury film electrode especially when thin films are employed. C. Interaction Between Mercury and Platinum The most undesirable property of the mercury-platinum system for making thin mercury films is that platinum reacts with mercury. This process proceeds at a significant rate (14, 17), and transforms deposited mercury into an intermetallic compound which cannot be used for stripping analysis purpose. When the mercury coated platinum is subject to anodic polarization, three distinct processes can be identified (15, 15, 29). With constant current, Brubaker found three potential plateaus at 0.65 volt, 1.25 volt and 1.80 volt vs. R.H.E. which correspond to the ionization of the mercury, the ionization of the mercury from the mercury-platinum intermetallic compound and the oxidation of the solvent respectively. the exis Sur com PtH bod The Cad: than - 74 - Studies on the mercury-platinum system indicate that the intermetallic compound between platinum and mercury exists in many forms. An X-ray study by Plaskin and Survorovoskaya (19) found three phases of the intermetallic compound with structures corresponding to PtHg, PtHga, and PtHg#. The crystal structure of Ptth was identified as body-centered cubic with ten atoms in one unit cell (51). The structure of PtHga is tetragonal with three atoms in one unit cell. Only one stoichiometry has been found at the inter face between the platinum and mercury metal (15, 55, 41); it has the structure of Ptng. Judging from the long life of the platinum-based mercury drop electrode, the mercury-platinum reaction rate must slow down as the intermetallic compound layer thickens. This view has been confirmed directly or indirectly by several studies (15, 14). D. Additional Problems with Platinum-Based Mercury Thin Film Electrode Gardiner and Rogers (6) used the thin mercury film for their anodic stripping analysis to determine the concentration of cadmium and discovered that the recovery of the reduced cadmium determined by the stripping current peak reached a maximum value until the film thickness reached 1‘pm. No In b. .5. 811 Of Dr 8121 Whj of wit - 75 - explanation was given for this phenomenon. Kemula, Kublik and Galus (9) discovered that when the stripping analysis of Cd and Zn was performed with a hanging mercury drop and a thin mercury film electrodes, Cd was stripped off at the same potential from both electrodes while the Zn peak was missing with the film electrode. The behaviors of antimony and tin were found analogous to zinc. These authors suggested that intermetallic compounds formed between platinum (or the Pt-Hg compound) and these metal might be responsible for such behavior. E. Objective of the Study To summarize, mercury film formation on the platinum surface is dependent on the surface state of the platinum prior to the mercury deposition. Understanding of the nature of this relation paves the way for devising effective procedures for film formation. This is essential for the study of the formation for the platinum-mercury system in which a growing layer of intermetallic compound exists. The objective of this study is to correlate the nature of the film formed and the surface state of the platinum with the aim of divising an effective process for constructing -75.. reproducible and smooth mercury films, and to determine the rate of the intermetallic compound formation so that a sound prediction may be made on the conditions under which the mercury plated platinum electrode can be effectively used. A... use emp1 in e Eru‘ new 3111‘. The the - 77 - CHAPTER III EVALUATION OF THE FACTORS AFFECTING THE STRUCTURE OF THE MERCURY ELECTROPLATED ON THE PLATINUM SURFACE A. Platinum Surface State Characterization As discussed in the preceding chapters, the processes used for making mercury coated platinum have been largely empirical. The success of obtaining a good mercury coating in each attempt depended on chance as was observed by Brubaker. If any kinetic data is to be drawn on the mercury-platinum interaction, this situation must be avoided. A method of making a smooth reproducible mercury film on the platinum surface has to be found. To be able to accomplish this, a good understanding of the platinum surface under the experimental conditions is necessary. The dependence of the structure of the mercury deposit on the surface states of the platinum and the other conditions of the experiment has to be determined. The effect of the surface state of the platinum electrodes will be considered first. 1. pl. pl. 811] fox the As ree 11110 has thi the and be Pla- and is 1 SUI‘ hat: it : pla~ - 7g - 1. The Characterization of the Surface State of Platinum The electrochemical behavior of a platinum electrode is greatly influenced by the treatment performed on the platinum. For instance, the catalytic activity of the platinum is found to be greatly increased after it has been subjected to anodic-cathodic cycling. This phenomenon was found decades ago (3) and was attributed to the removal of the impurities irreversibly adsorbed on_the surface. As a consequence of the great interest shown by numerous researchers on platinum in the past several decades, our understanding of the electrochemical behavior of this metal has been improved (8). The electrochemical properties of this metal are dependent on its surface structure (56), the preparation (1), the electrolyte solution used (2), and the process being carried out (8). A description of the state of the platinum surface can be divided into the chemical and physical aspects. The platinum surface exposed to the air always has an oxide film and this is one of the reasons that this type of platinum is not in the active state. Mercury can't wet this type of surface even after a long time of contact (49). Although the nature of the surface oxide is still controversial (8). it is generally agreed that the adsorption of oxygen on the platinum surface in the anodic process is irreversible, th be th ma 60' ea: on is Pla 0n to hec SUI‘ -79... that the irreversibility increases as the potential applied becomes more anodic, and that oxygen is capable of entering the platinum crystal lattice (18). When the potential is made cathodic enough, the platinum electrode surface is covered with an adsorbed layer of hydrogen. In contrast to the oxide film, the adsorbed hydrogen film can be removed easily. The adsorption and the ionization of the hydrogen on and from the platinum surface is a reversible process. The physical state of the platinum surface is associated with its physical structure on the surface. Although the nature of the dependence of the electrochemical properties on the physical state is not fully understood, the importance is well recognized. For instance, the activity of the platinum surface for hydrogen adsorption was found to vary among different crystal planes. The control of the physical structure of the platinum is difficult. Platinum forms a face-centered cubic crystal. Both single crystal and polycrystalline platinum electrodes have been used in electrochemical studies. The polycrystalline platinum which is most commonly used is composed of numerous grains of platinum crystallites. Several grain boundaries are exposed on the surface. As the initial step, the platinum electrode to be used in the electrochemical study is commonly polished mechanically or electrochemically in order to create a fresh surface. This process creates a microscopically rough (1') SU be nec of re inc - go - surface. The surface actually exposed is always larger than the geometric area. The ratio of the actual area to the geometric area is called the roughness factor. This value varies from almost unity for a very smooth surface to several thousand for platinized platinum (2h). The most common method of measuring the actual surface area is to measure the quantity of charge required to completely cover the platinum surface with hydrogenatoms. One hydrogen atom is assumed to be adsorbed by one surface platinum atom (h, 21). 2. Effects of Pretreatment on the Surface State of Platinum After the initial mechanical polishing, it is always necessary to pretreat the platinum surface by a combination of chemical and electrochemical methods in order to obtain reproducible results from the platinum electrodes. The major concern is to remove any trace of impurities adsorbed on the platinum surface. Various processes have been proposed and according to their promoters, reproducible results have been obtained (1, 2, 17). These processes in general consist of a cleaning step followed by a treatment step which includes electrochemical oxidation and reduction. Inspite of our incomplete knowledge of the platinum oxidation process, most of these methods prescribe a strong oxidation. The platinum electrodes form different treatment methods show different behavior. If the platinum is cycled between the anodic and cat inc sur Man Dur sur and aPP cat red SLII‘ is thr. -8]... cathodic potentials, the resulting electrodes show an increase in catalytic activity (3). In addition, the platinum surface is altered by the anodic-cathodic cycling process. Many observers obtained a smoother platinum surface (25, 24). During the anodic cycle, the platinum at the electrode surface is oxidized. A certain amount of platinum is ionized and dissolved in the solution which depends on the potential applied and the electrolytes used. During the following cathodic cycle, part of the dissolved platinum ions are reduced. Thus many observers reported that the platinum surface was platinized by the potential cycling process (5, 25). The reduction of the oxide film on the platinum surface would also be obtained without the occurrence 0f surface platinization (1: 5). F. Anson (5) reported that the reduction of the oxidized platinum with hot 12 M HCl produced platinum without forming platinization. This type of platinum surface was named "stripped platinum" to differentiate it from the electrochemically reduced surface, "red platinum". 5. Classification of Platinum Surface According to the Surface Roughness Factor The electrochemical behavior of the platinum electrode is greatly affected by the platinum surface state. At least three factors concerning the surface roughness influence the st] ac‘ dis prc 51. -32.. structure of the mercury deposit formation, namely the activity of the surface platinum, the local electrical field distribution on the platinum surface during the deposition process and the crystal structures of the platinum surface. It is logical to classify the platinum surface according to roughness factor in the study of the effects of the surface state on the mercury film formation. Three types of surface can be differentiated : a. Bright platinum. This type of surface is characterized by a small roughness factor, typicall between 1 and 2. b. Lightly platinuzed platinum. The platinum is this category has roughness factors ranging from 5 to well over 10. c. Heavily platinized platinum. The platinum surface in this class has a high value of the roughness factor, typically between 100 and several thousands. From the discussion that follows in this chapter, it is apparent that the roughness factor is an important factor in characterizing the state of the platinum surface. Its effects on the mercury deposition are examined. The effects of the surface tension on the mercury deposition as suggested by Hartley, Hiebert and Cox (14) and real role of potassium ch: B. twc pot phi de; mer me: - 85 - chloride during the mercury plating are also discussed. B. Methods of Examining the Smoothness of the Mercury Film on the Platinum The smoothness of the mercury deposit is examined by two methods : microscopic examination and analysis of the potential-time behavior of the resulting electrode. The physical shape of a thin coating of electrochemically deposited mercury is examined under a microscope. If the mercury coverage of the surface is complete, the resulting surface looks like a mirror. Otherwise, the surface appears rough. With a 10X object lens, mercury dr0ps with a size as small as two microns can be resolved. The platinum surface which is completely covered with mercury shows a very large hydrogen overpotential even with a very thin mercury film. With a platinum electrode in 0.1 M HCth electrolyte, hydrogen evolves vigorously at potentials more anodic than -0.1 V vs. R.H.E. With smooth mercury film electrodes, no hydrogen evolution is detected above -0.6 V vs. R.H.E. The cathodic hydrogen overpotential of a mercury plated platinum electrode varies with the states of the mercury coverage. C. ac sol use aci the whi des has “8e: - 54 - C. Experimental Set-Up 1. Chemicals a. Mercury Solution Mercury forms ions with two common oxidation states in solution. Both mercuric and mercurous solution have been used as the plating solution for mercury. Since in perchloric acid solution, no significant difference is expected in the choice of the oxidation state of the mercury ion from which to plate. In the study of the film making process described in this chapter, mercurous ion was used exclusively. Mercurous perchlorate was prepared following Pughe's method (22). Reagent grade mercuric oxide was dissolved in conc. perchloric acid. The solution was shaken with triply distilled mercury which converted all mercuric ion into mercurous ion. The solution was filtered and diluted to have 0.1 M in Hg;+ and 0.1 M in perchloric acid. The solution was stored in brown bottles. No disproportionation of the Hg(I) ion was detected after several months of storage. The exact concentration of the mercurous perchlorate was determined by a gravimetric method. Sodium chloride was used to precipitate the mercurous ion. The precipitate was 2L 1:}. acj for 1‘11) ele Wit Eur Pea Of. dESC - 35 - filtered and dried at 110°C for an hour before weighing. The concentration was found to be 0.107 M. b. Platinum The platinum electrode is made by sealing a piece of 24 guage platinum wire in soft glass tubing so that only the cross-section of the wire is exposed. The platinum is reported to be 99.9 % pure. The tip of this platinum wire is polished with 600 grit silicon carbide polishing paper each time before a new mercury deposition in order to have a fresh platinum surface unless noted otherwise. After polishing, the electrode is cleaned by dipping it in boiling concentrated perchloric acid briefly. The electrode is then etched in hot aqua regia for one minute or in room temperature aqua regia for over five minutes to amooth the surface. A bright platinum electrode is obtained by reducing the resulting electrode with hot 12 M HCl or with reducing current. The platinum surface obtained after this step is free from oxide and is ready for the electroplating of mercury. The geometric area of this electrode is 2.5 x 10"3 cm2. The preparation of the platinized platinum will be described in the eXperiment section. A series of oxidation- -86.. reduction cycles achieved by means of alternating current steps are employed. c. Cases The working solutions are deaerated by bubbling purified H2 gas or N2 gas through the solution. Purification of the H2 gas is done with a catalytic column produced by Englehard Industries. The residual oxygen content is less than 1 ppm. The N2 gas from the gas cylinder is deoxygenated with a hot copper catalytic column. Before going into the working solution, gases are saturated with water. d. Perchloric Acid The 0.1 M perchloric acid is made by diluting 7O % reagent grade perchloric acid. e. Sulfuric Acid The platinization of the platinum is carried out in 1 F H2504. Reagent grade H2504 is used for dilution. f. H01 Reagent grade 12 M HCl was used to reduce the surface oxide. 2. 0f dro thi 3. 3. Sur for EXP mer Tef WOr fOr Per ele - 37 - 2. Microscope A microscope from Bausch and Lomb is used for inspection of the mercury deposits. A 10X object lens is used. Mercury dr0plets with diameter of 2 microns can be resolved under this condition. 5. Flow Cell a. Construction It was suggested by some workers that the platinum surface could not be considered clean after exposure to air for as brief a time as two seconds (8). To prevent the exposure of the treated platinum to the air before the mercury deposition, a flow cell made of pyrex glass and Teflon stopcocks has been used. With this flow cell, the working solution is switched from the perchloric acid used for electro-cleaning of the electrode to the mercurous perchlorate used for plating mercury without exposing the electrode to the atmosphere. The flow cell is composed of two solution reservoirs and a reaction chamber as shown in Figure 5-1. The reservoirs and the chamber are connected by a two way Teflon st0pcock. - 88 - purified N or H 2 2 r 2 gas _ S-B ’ outlet ... we R-1 R-Z o h“ "' ‘ O ‘ (working \\4 111.3. g? electrode a 1 male ground glass adaptor to waste Figure 5-1. Flow Cell One At gla wit be cou' chal frox of 1 t0 1 the V010 IPOm elec dead reSe the air. Cham , ..‘fl - 39 - One solution is allowed to enter the chamber at a time. At the top of the reaction chamber, an adapter made of ground glass is constructed so that the working electrode (sealed with silicon rubber in a male ground glass adapter) could be inserted. A platinum wire mesh, which serves as the counter electrode, is sealed at the lower part of the chamber. The reference electrode compartment is isolated from the chamber by a medium pore glass frit. The diameter of the chamber at the working electrode has been minimized to reduce the solution dead volume. b. Performance The most important parameters of the flow cell are the dead volume and the dead time. These are the solution volume and the time needed for the solution to be transported from the bottom of the solution reservoir to the working electrode. The dead volume is 0.65 cm3 for this cell. The dead time varies slightly with the solution volume in the reservoir. It is determined primarily by the orifice at the exit stopcock. The typical value is 0.25 second. It is critical that the reaction chamber must be air-tight, allowing no air oxygen to diffuse into the chamber during the experiment. The performance is examined by measuring the potential variation with time as is shown The bubt ope: is a the B b‘ dec: of 3 Sud: at- - 90 - in the experiment described below and in Figure 5-2. The solution was exposed to the air prior to point-A. The potential was quite anodic. Then hydrogen gas was bubbled through the reservoir 3:1 solution with st0pcock.§:§ Open to this reservoir. The gradual decrease in potential is a noticible indication that H2 molecules diffused into the reaction chamber. The solution flow was started at time B by Opening the exit stopcock.§:2. This produced a sharp decrease of potential which indicated that the concentration of H2 molecules at the working electrode surface was suddenly elevated. The potential stablized quickly as shown at time D. The value of the potential agrees with the value calculated by the Nernst equation for the H2 reduction reaction. The chamber was isolated by closing both entrance stopcock §2A and exit stopcock.§:2. The potential remained unchanged over tens of minutes. The stopcock.§=A was opened to reservoir 3:; which was bubbled with N2 and the stopcock §:2 was opened allowing the N2 bubbled solution to enter the chamber and replacing the H2 saturated solution. The gradual increase of potential was observed indicating that the H2 concentration at the working electrode surface was diminishing. Finally, the reservoir was exposed to the air and the solution allowed to pass into the chamber at time F by opening stopcock S-D. The potential returned to the anodic potential. - 91 - :mwhxo soapsaom on» on Hamo Roam 0:» mo ommoammm Heapsouom 2H: .mzHe 0.0m show one. .mum mesmaa :.o 1 .m.o.m .m> .m :Moo PI by Cu DO 011. the ele the ar - 92 - This experiment clearly shows that the flow cell satisfactorily keeps an air-tight environment. All the potential values reported in the whole thesis refer to the reversible hydrogen potential (R.H.E.) unless specified otherwise. 4. Typical Experimental Sequence The computer controlled coulostatic system offers the user great flexibility in varying experimental sequences according to the requirements of the experiment. A typical experimental sequence is shown in Figure 5-5 which includes the electrochemical pretreatment of the platinum electrode and the electrodeposition of mercury. The experimental sequence is initiated with the pretreatment of the electrode. The process is controlled by a program called ETREA. Repeated cycles of anodic-cathodic current pulses are applied to the cell with the resulting potential-time curves shown on the computer terminal for on-line inspection. The working electrode is discarded if the results show that any impurity has remained on the electrode. If the result shows a clean platinum surface, the mercurous solution is introduced. Within a short time, a reduction current pulse is issued which deposits the amount of mercury selected by the operator. The computer ..93 - OPERATOR COMPUTER START I V EXAMINE t: PRETBEAT v-0 cunvg J > ELECTRODE 1 INTRODUCE ' MERCUROUS SOL’N , spun ngcTno-A LYSIS PULSES EXAMINE J<%—RINB BELL-—___J HG DEPOSIT E110 FIGUREITG. HG DEPOSITION EXPERIMENT - 94 - rings a bell as soon as the reducing pulse is ended. The electrode is removed and examined with a microsc0pe to inspect the structure of the mercury deposit. D. Experiments and Results 1. Creation of Platinum Surfaces with Varying Roughness Factors a. Anodic-Cathodic Cycling of Platinum and Roughness Factor Measurement A clean platinum electrode exhibits a characteristic potential-charge relationship under constant current anodic-cathodic cycling (reverse current chrono- potentiometry). Typical potential-charge responses are shown in Figure 5-4. A solution of 0.1 M HClO was used 4 as the electrolyte for both curves. Curve 2 resulted from a freshly polished, hot concentrated perchloric acid treated platinum electrode. Curve 1 was obtained with the same electrode after it was boiled in hot aqua regia for one minute following the concentrated perchloric acid treatment. These curves are characterized by several distinctive regions. 0n the anodic current side, the curve starts with the hydrogen desorption region. In this region, the -95- 1. 8550f"- o: 3: 5°." 6 ‘3. :’ >< >' z: ”9'1599 : 2+ 1 t T i e : 5.000 )JCOUL. X10 ' Figure 3-4..0yclic Chronopotentiogram of Pt Electrode in 0.1 M HClOA -96- potential rises slowly until all hydrogen that was adsorbed on the platinum surface is depleted (at a potential of 0.2 V). The potential rises sharply thereafter until a new electrochemical reaction begins. The fast rising potential region is associated with the electrical double layer charging. The new reaction, the oxidation of the platinum surface, begins at about 0.8 V. At 1.8 V, the oxidation reaction becomes the formation of oxygen gas. Shortly after this point, the current polarity is reversed which results in a very sharp plunge of the potential down to 0.6 V. The oxide just formed begins to be reduced. The potential continues to drop, passing the brief double layer charging region, into the hydrogen adsorption region and eventually to the begining of the hydrogen evolution reaction. The potential remains fairly constant after that point. These curves are used for two purposes : (1). To ensure the surface.cleanliness, These curves serve as the fingerprints for the clean platinum surface. Surface contamination causes a deviation from these curves and calls for further cleaning of the surface. Furthermore, this cycling process drives the potential above 1.8 V which is enough to oxidize some contaminants. In 0.1 M H0104, the anodic-cathodic cycling causes no apparent surface - 97 - roughening up to several thousand cycles. The surface area remains unchanged. (2). To determine the roughness factor, The roughness factor (RF) is determined by multiplying the time required during the hydrogen ionization or adsorption by the current used to get the charge equivalent of real surface area and dividing this value by the charge equivalent to the geometric area. One cma polycrystalline platinum surface adsorbs about 210 microcoul. charge equivalent of hydrogen. b. Bright Platinum The platinum surface obtained by mechanical polishing shows a rather large roughness factor. Curve 2 in Figure 5-4 was obtained with this type of electrode. A 2.5,ncoul. charge was used for ionization of the adsorbed hydrogen. This corresponds to a roughness factor of 4.8. After aqua regia treatment, the surface area was reduced because the fine platinum grains on the surface were dissolved in the solution. Curve 1 in the same figure was obtained with an aqua regia treated electrode. It shows a roughness factor of 1.6. The platinum surface obtained in this way is used as the "bright platinum" in the studies that follow. - 98 - c. Lightly Platinized Platinum Platinum is reported to be platinized following the anodic-cathodic cycling process (5. 25). This phenomenon was utilized to make the lightly platinized platinum in this study. The method used is a modification of the method used by F. Anson (5). Instead of using 60 Hz ac as in Anson's study, alternating current steps were used in this study. Each step lasts 15 msec, and has an amplitude of either +1.0 mA or -1.0 mA. The whole process is under the control of the program called PLATNZ.MAC. A multiple of 20,480 current steps is chosen by the operator. Since the platinum dissolution in halogen acid is one to two orders of magnitude higher than in other acid (25), HCl solution seems to be the choice for the platinization. However, in HCl, the evolution of chlorine gas occurs at very low potential, and electrically insulating bubbles are formed on the platinum surface. Thus HCl is not a good choice for the purpose of platinization. In Figure 5-5, curve 2 shows a characteristic potential- charge curve in HCl. Curve 1 was measured in 0.1 M H010,+ solution. Solutions of 2.0 M and 0.1 M HClOI+ were tested as platinization electrolytes. No apparent surface roughness - 99 - 'T .. v m J- I .. um .. 0:. .. a; o -- iii ;: -r . ~11- > z .. qt ‘9-199 ::::::44:4~r::::‘1=1-1-;| 0.000 ' 2.000 ANODIC fiCOUL. CATHODIC x10 '1 Figure 5-5. Cyclic Chrono potentiogram of Pt Electrode in (1) 0.1 M Perchloric Acid (2) 1.0 M HCl ~100- increase was detected after 20,480 current steps and these electrolytes were therefore abandoned. The 1.0 M HZSO,+ solution used by F. Anson (5) gave satisfactory results. Variations of the potential-charge curves measured in 0.1 M HCth with an increasing number of current steps are seen in Figure 5-6. The y-coordinates of the curves are moved successively downward by 0.1 V for clarity. The expansions of the hydrogen ionization region and hydrogen adsorption region are on the graph. Table 5-1 lists the roughness factors corresponding to these curves. Apparently the roughness factor does not vary linearly with the number of the current steps. This might by caused by the interference of the gas bubbles generated during the process. The roughness factor also varies from one electrode to another with the same number of current cycles. An oscillosc0pe is quite helpful in estimating the roughness factor attained during the process. This is done by connecting the sc0pe probe to the output of the absolute value amplifier (Fig. 1-5) and observing the time required for hydrogen adsorption in each current cycle. - 101 - 1.900" :1: __ 11‘? v a; 3 > x >' z: 1 -0.100 3 —e:e::i—Tae{:::+lr:e:: 0.000 2.00 ANODICpCOUL. CATHODIC X10 Figure 5-6. Platinized Electrode. Y-coordinates is drawn with respect to curve 1, other curves are shifted downward successively by 0.1 volt. - 1023- m.m.A o.m_A s..— m.¢ m.m m e m N P m.¢ modemufloapmaa oaomom houomm mmosswsom A mampm vacuums mafipmsuopamv owe.o~ so caeasasa cease mmoam pmoaaso wsHpmsAopH< npfia SfidHHMHm mo QOHuMNHzHHGHm .—1m manna -103- 2. Influence of the Potential and the Roughness Factor on the.Mercury Deposition The effects of the potential and the roughness factor on the structure of the mercury deposit are determined by examining the structure of mercury deposit formed on bright platinum as well as on platinized platinum electrodes under the following conditions : a. potentiostatic, b. galvanostatic. With the potentiostatic method, the potential is suddenly stepped to a predetermined value and is then held at the same potential until the process is over. With galvanostatic method, a current step is applied. The potential varies with time during the process and is used for the analysis of the plating process. a. Potentiostatic Experiments (1). Bright platinum electrode Two schemes were used to examine the potential effects of the mercury deposit : (a) the deposition was carried out at several selected potentials and the resulting deposit was examined under microscope, (b) the mercury deposit obtained at a moderate overpotential, ie. 400 mV vs. R.H.E., was held at several more cathodic potential values for several - 104 - minutes in 0.1 M HClO,+ solution. The electrode surface was examined with a microscope prior to the application of a new potential to detect any difference in the appearance of the mercury deposit. (a). The mercury plating was conducted at some selected potentials. The results obtained are listed in Table 5-2. Two distinct appearances were detected, the mercury droplets and the mercury patches. The mercury droplet appears spherical or oval in shape. Each droplet has a small contact area with the platinum. The shape of the mercury patches is quite irregular. Each patch has a large platinum-mercury contact area. A drawing of these two configurations is in Figure 5-7. Most frequently, both configurations of mercury deposit appear to distribute randomly over the platinum surface with varied sizes. (b). A mercury coated electrode with mercury patches produced at 400 mV was held at more cathodic potentials in 0.1 M HClO until the potential was as low as -0.8 V. 4 Hydrogen gas was evolving vigorously at such a cathodic potential. Several mercury patches merged together to form bigger mercury patches, but some area of the platinum surface appeared to remain free from mercury coverage under microscopic inepection. - 105 - 000 Pt 0.24mm (a) Mercury Droplets Q <33 0} Pt (b) Mercury Patches Figure 5-7. Two Types of Mercury Deposit on Platinum Surface - 106 - Table 5-2 Potential dependence of the mercury deposit on the bright platinum electrode Potential mV vs. R.H.E. Appearance of Hg deposit 700 mercury droplets and patches 650 mercury patches 400 mercury patches 200 mercury patches -50 mercury patches with Habubble Note : Rest potential = 750 mV The rest potential in 0.1 M HgClO & 0.1 M HClO 1+ 4 is 0.750 V. Table 5-5 Dependence of the roughness factor of the mercury deposited at 200 mV RF ' Appearance of plated mercury 3.9 mercury patches and scattered smooth mercury films 5.8 frequently smooth film 5.7 smooth film 9.5 smooth film Note : Smooth mercury film is obtained with electrode having RF 2 5.8. - 107 - (2). Lightly_platinized platinum electrode Platinum electrodes with various roughness factors were electroplated with mercury at 200 mV. The results are listed in Table 5-5. b. Galvanostatic Experiments In galvanostatic cases, a current step is applied to produce mercury plating. The measured potential always shifts in the cathodic direction with time. The rate of the potential change depends on the current density applied. If the current density is quite small, the potential remains quite close to the rest potential. If a large current density is used, the concentration of the mercurous ion near the electrode surface drops to zero in a short time and this causes the potential at the platinum working electrode to become sufficiently cathodic for H2 evolution. The platinized electrode used in the following experiments are chosen to have high RF which, according to results shown on Table 5-5, produces smooth mercury films at a suitable potential. (1). Small current density The potential-charge response to a current of -24,nA 3.’ - 108 - for both the bright and lightly platinized platinum electrodes are shown in Figure 5-8. Curve 1 was obtained with a bright platinum which had a roughness factor of 1.0. Curve 2 was from a platinized platinum with RF of 9.5. The potential moves slowly cathodic. The charge passed during the electrolysis is enough to deposit over 600 monolayers of mercury. In both cases, mercury droplets were found. Much finer, evenly distributed mercury droplets were formed on the surface of the platinized electrode which might cause its potential response to be slightly anodic compared to the other electrode since the area exposed by the finer mercury droplets with same amount of mercury must be larger. (2). High current density A current of ~1.0 mA was used to plate mercury. An unplatinized platinum with a RF of 2.? gave curve 1 in Figure 3-9. while a lightly platinized platinum with RF of 9.5 gave curve 2. The overall platings took 10 sec.. The exhausting of the mercurous ion at the electrode surface caused the potential to plunge from 0.75 V to less than -0.5 V within one tenth of a second. The potential measured with the platinized electrode started at a more anodic value than that of the other electrode but soon became more cathodic. Hydrogen bubbles were formed in both cases. However, smoothly plated mercury was found on the platinized platinum electrode surface while mercury patches - 109 - 1.00U"’ 0.500. 0.00 2. 00 Figure 5-8. Plating of Mercury with -24,uA Current on 1. Bright Platinum 2. Lightly Platinized Platinum - 110 - 8.000'. B.H.E. VS. X10 -2 1.1.1.1.1.1.1.1.LII.IL1.1.|.1 MV. lfl.’l‘r'l'l1l'lri'l'l'ljl'l '8-999W 0. 00 1.000 )JCOULL. x10 " Figure 5-9. Plating of Mercury with -1000.0.nA Current on 1. Bright Platinum 2. Lightly Platinized Platinum - 111 - were formed on the unplatinized electrode. To compare the potential response, platinum covered with thick film of mercury, 57.9 mcoul, was tested under identical experimental conditions. Figure 5-10 shows the resulting potential-charge response. The X-coordinate of this figure has smaller scale so that the depletion of the surface mercurous ions is shown. The hydrogen overpotential of this electrode is at least 0.5 V more cathodic than the two freshly plated electrodes in Figure 5-9. B. Discussion 1. Structure of the Mercury Deposit versus the Plating Potential The mercury electrodeposited on the platinum surface is found to have three types of structure : mercury droplets, mercury patches and smooth film. At small potential polarization, both potentiostatic and galvanostatic experiments produced mercury droplets. This is true with bright platinum as well as with the lightly platinized platinum. At large overpotentials, mercury patches and/or mercury films were produced. The transition from one type of behavior to the other occurs at about 700 mV. Both mercury patches and droplets were found on the same surface at this potential. Formation of the droplets at low overpotential indicates - 112 - 1.00 P u] :E 5‘? :52 > x >' z '1-509 *‘e:;:::::::::::::‘:: 0.000 2.000 100001.. x10 '3 Figure 5-10. Plating of Mercury with -1000.0 11A Current on a Thick Mercury-coated Platinum Electrode - 113 - that the deposition of the mercury on the platinum crystal lattice is not favorable at such potentials. This situation cannot be caused by the oxide film on the platinum surface because it had been deliberately reduced. The great majority of the platinum surface sites were free from oxide before mercury was plated. Furthermore, the reduction of oxide film on the platinum surface is an irreversible process , and a rather large overpotential is needed. The anodic-cathodic cycling in Figure 5-4 indicates that the reduction of the oxide film on platinum starts below 0.55 V. Thus if it had been caused by the residual surface oxide, the same structure of mercury deposit would have been produced for mercury deposited at the potential above 0.55 V. However, this was not the result observed at 0.65 V. The possibility of the interference from the hydrogen adsorption is remotely slim since its adsorption requires a potential as low as 0.5 V. In order to bring the mercury into direct contact with the platinum surface as in the patch structure, several kinetic steps are involved. These are the transportation of the mercurous ion to the platinum surface, the adsorption of the mercurous ion by the platinum surface in order to be reduced, the dehydration of the ion, and the incorporation of the mercury atom onto the surface crystal lattice of the platinum. All these steps, including the last two, require a certain amount of activation energy. The activation - 114 - energy required for the last step is much less when a mercury atom is deposited on a mercury surface. The droplet configuration of the mercury deposit indicates that the deposition of the reduced mercury on the existing mercury surface is prefered to deposition on the platinum surface. As the overpotential is increased, the energy transfered per mole of mercury during the reduction is increased according to AG = -n.FAE, where AB is the overpotential, F is the Faraday constant, and n is the number of electrons associated with the reaction. The high potential of the transition of the structure from the mercury dr0plets to the mercury patches seems to indicate that a significant energy barrier exists for the deposition of the mercury on the average platinum surface sites. The magnitude of this energy barrier depends on the structure of the platinum surface and is expected to vary among sites with different crystal environments (26). The bright platinum surface structure is complicated by many factors. Since the electrode is made with polycrystalline platinum, numerous platinum crystallite boundaries are present. Various crystal planes are exposed and numerous crystal defects are present on the surface. The defects include the edge vacancies, holes, steps and kinks on the - 115 - platinum surface. These surface features create various crystal environmentscnithe surface and thus present various crystal sites, each carrying a different energy requirement for the mercury deposition. The bright platinum surface created by mechanical polishing and aqua regia etching most frequently posesses a surface which is random in the constitution and distribution of these surface features. As a consequence, the distribution of the mercury deposit on the platinum surface under potentiostatic conditions will be random. This was confirmed by the experiments using bright platinum electrodes which produced either droplets or patches. Furthermore, at the potential of 700 mV, on part of the surface mercury patches were formed while on an other part of the surface mercury droplets were found. This is a strong indication of the non-uniformity of the platinum surface structure. However, in addition to the non-uniformity of the energy requirement associated with the crystal structure, a second type of non-uniformity might be operative. This is the non-uniformity of the local electrical field caused by the rough surface condition. Which of these two factors, the energy barrier or the electric field strength, is mainly responsible for the simultaneous formation of these two mercury deposit is not certain. - 116 - No noticible difference was observed using the various overpotentials which result in the formation of mercury patches. This indicates that some sites of the platinum surface are always prefered for the mercury deposition in spite of the increase in the overpotential. A related study was done by Hassen, Unterecker and Bruckenstein who used the HgClOu solutions with concentration between 10'“ m to 10"6 M for the deposition of mercury. The potential during the deposition was held at 0.0 V vs. S.C.E. They found that on the mercury plated platinum surface, the capability to adsorb hydrogen is reduced. One deposited mercury atom displaces two hydrogen atoms when the mercury coverage of the platinum surface is low. When the coverage exceeds :~v55 %, fewer hydrogen atoms are displaced for the same amount of mercury deposited. This suggests that there is a competition for the reduced mercury atom between the platinum atom and the mercury nuclei already deposited. As the overpotential is increased, more sites on the platinum surface become available for the mercury deposition to occur so that a wider area of mercury-platinum contact becomes possible. Since deposition of mercury on these nuclei is energetically more favorable than on the platinum atom, faster mercury deposition is expected on these sites. Therefore, the overall mechanism leads to the formation of mercury patches. -117- Assuming that the platinum surface exhibits no specific adsorption for the mercurous ion, the mercurous ion at the surface of the platinum in 0.1 M Hg2(ClO#)2 solution constitutes less than one in 500 of the molecules and ions in contact with the platinum surface. Thus the usage of high overpotential seems to have no advantage. Hydrogen evolution occurs at much higher overpotential and creats a new situation. The evolving Ha gas disturbs the solution with gas bubbles. These bubbles tend to exist preferentially at the platinum sites that are poorly covered with mercury. Since the H2 bubbles block these sites from the solution ions, no real gain could be expected. 2. Effect of Surface Roughness Factor on the Structure of the Mercury Deposit The results shown on Table 5—5 clearly demonstrate that at suitable potentials, a smooth mercury film is produced on the platinized platinum surface with higher RF value. Thus the platinization must assist the mechanism which leads to a smooth mercury film formation. What is happening during the platinization process which is comprised of repetitive oxidation and reduction cycles ? - 118 - During the platinization process in H280“, a topological change is induced. Surface roughening was clearly observed in these eXperiments since the platinum surface changes appearance during the process. It loses the shiny metallic appearance, turns gray, and eventually becomes dark. The surface roughness factor increases as the process continues. Roughening has been explained in terms of redistribution of surface metal atoms brought about by forming and breaking platinum-oxygen bonds (2, 25). Biegler (25) showed that the roughening only took place when more than one oxygen atom was associated with each surface platinum atom on the anodic step. Unterecker and Bruckenstein (27) demonstrated with the potential cycling method that the roughening process involved dissolution and redeposition of surface platinum. Thus a new platinum surface is generated by anodic-cathodic treatment. Since the mercury droplets formed on the platinized platinum electrode showed a uniform distribution and a fine uniform diameter under the galvanostatic condition with a small current, a platinum surface with a more uniform surface condition must be created by the platinized treatment. Since an increase of RF value implies an increased amount of surface defects, the finer diameter droplets configuration with this surface condition would mean that these defect -119- sites are energetically more favorable for the reduced mercury atom. This is consistant with the theoretical calculation which predicts that the activation energy for the incorporation of the an atom during the electroplating process is less on the defect sites of crystal surface (44). What is the threshold RF value for the smooth mercury film formation ? Table 5-5 shows that at a RF of 5.8, a smooth mercury film was obtained. It seems that the threshold RF for smooth film formation should be below this value. However, the concept of the threshold should be taken with caution since the RF value obtained at low platinization is easily affected by the surface condition prior to the platinization. The RF value alone cannot be used to specify the surface condition unless a treatment had been applied which could exactly reproduce the surface condition. This fact is demonstrated in the example shown in Figure 5-5. Two bright platinum electrodes were used. One of them, curve 1 had a RF value of 5.7. Despite of the high RF, the creation of a smooth mercury film on such a surface has been found to be very improbable. It is interesting to compare the potential responses of a lightly platinized platinum electrode with a bright platinum electrode using a high plating current as shown in Figure 5-9. - 120 - The response potential of the lightly platinized platinum electrode (curve 2), which ended with a smooth mercury film, started at 0.1 V more anodic than the other electrodes, but became more cathodic as more mercury was plated. This means that the hydrogen overpotential on this electrode was increasing. This clearly indicates the mercury coverage of the platinized platinum electrode was improving. The potential response using the bright platinum is also interesting. During the plating process, the potential with hydrogen evolving stayed fairly constant. This value, -O.6 V, is quite cathodic compared to that of the platinum electrode which is about -0.05 V in 0.1 M HClO# solution. Except for the beginning, the potential did not vary with the charge consumed. This evidence strongly suggests that the reduction of mercury ions occurred on the existing mercury surface instead of the platinum surface which resulted in mercury patches. 5. Other Factors that Influence the Formation of the Mercury Deposit 1. Hydrogen Adsorption Hassan, Untereker and Bruckenstein (17) studied the platinum surface covered with a submonolayer of mercury - 121 - and found that the hydrogen adsorbed decreased in direct proportion to the amount of mercury deposited up to 55 % mercury coverage of the electrode surface, then a smaller amount of hydrogen is displaced by an equal amount of mercury deposit. This indicates that mercury atoms must occupy the platinum sites at the expense of the hydrogen. However, the occurrence of hydrogen evolution did not seem to interfere with the smooth mercury film formation, except that H2 bubbles that stick to the surface block current passage at those locations. b. Surface Tension Hartley, Hiebert and Cox (14) showed that the mercury droplets on the platinum surface merged to form a smooth film at a high cathodic potential. My results indicate that the platinum electrode covered with mercury patches produced at 400 mV showed an independence of the potential down to -O.8 V (section D-2, 1.b.). Although a 6 V battery was used by those cited workers, the overpotential could not have been more cathodic than -0.8 V unless a good coverage of mercury had already existed which is not the case for mercury droplets. The difference might come from the structure difference between these two types of deposits. The droplets, due to their lack of interaction with the platinum surface, - 122 - appear to be more susceptible to the influence of the surface tension variation as well as the physical agitation created by hydrogen evolution. The mercury patches, due to their strong interaction with the platinum surface as indicated by the contact area between the mercury and the platinum, are less susceptible to these influences. In addition, the interface between the mercury patches and the platinum becomes amalgamized quickly and loses the mobility of the liquid mercury. This bending adds more inertness to the mercury patches. c. Chloride ion It was pointed out by Gardiner and Rogers (6) that treatment of the electrode with potassium chloride before and after the plating of mercury on the platinum surface is both essential and critical to have good mercury plating. This does not corespond to our experience as is shown by the results obtained above. The plating of a smooth mercury film on platinized platinum needs no KCl treatment. However, chloride ion was found to play an interesting role in the mercury coating process. It was discovered that the platinum surface reduced hot concentrated HCl could easily adsorb traces of chloride ion on the surface unless great precautions were taken. Since chloride ion reacts with mercurous ion easily, when such a surface was plated with - 123 - mercury, precipitates of mercurous chloride could easily be produced on the platinum surface. Since calomel precipitate shows a high affinity for the mercury surface (65), it is easily adsorbed on the mercury surface. The calomel film interferes with the growth of the mercury drop due to its poor electrical conductivity. This forces the further deposition of mercury to proceed at other locations on the platinum surface. As a result, the creation of a more evenly distributed mercury coating becomes easier. This process is explored further in the next chapter. F. Conclusion The structure of mercury deposited on the chemically or electrically reduced platinum depends on the plating- potential and the surface structure of the platinum. When a small polarization potential is used for mercury plating, mercury droplets are formed. At higher polarization potential, mercury patches or mercury film are produced. To create a smooth mercury film on bright platinum directly by an electroplating process is difficult because the platinum surface is quite inhomogeneous. The situation is changed when the surface is platinized by current cycling -121.- which results in an increase in the surface area and probably a more homogeneous surface. The plating of a smooth mercury film is thus made easier. -125- CHAPTER IV A NEW METHOD OF MAKING A SMOOTH MERCURY FILM ON A PLATINUM SURFACE A. Introduction One major factor seems to be responsible for the difficulty of making a smooth mercury film on a platinum surface. The platinum surface is normally quite inhomogeneous. The surface atoms display various affinities toward mercury atoms. This causes a preferential deposition of mercury atoms on certain sites when mercury is reduced electro- chemically. Once some mercury nuclei are formed, these nuclei immediately become the center for newly reduced mercury because they compete much more favorably for mercury than the bare platinum atom. To solve this problem, several processes have been devised which include the usage of high negative reduction potentials (7), the immersion of platinum in a mercury pool (15) and the complexation of mercuric ions with EDTA Species. The fundamental principle of these attempts is that by varying the experimental condition, the tendency of preferential deposition of mercury might be made less to allow a more homogeneous coverage of the Pt surface with Hg. ~126- It was found during this study that there is a direct way of solving the mercury preferential deposition problem. In this method, a masking agent is utilized during the initial mercury plating to prohibit the deposition of mercury on the mercury nuclei. By masking, it forces the mercury plating to proceed on the bare platinum. Once the whole platinum surface is covered with this primary mercury layer, the masking agent can be removed and the mercury plating is resumed until a desired thickness is obtained. This process would be quite advantageous over the methods now being used because complete coverage of mercury on the platinum is quaranteed and the effect of the platinum surface status does not seem to have any direct influence on the success of the method. The masking agent chosen must be able to react with mercury atom on the platinum surface preventing further deposition of mercury and yet leave the platinum surface uncovered for mercury deposition. Chloride ion which forms a very insoluble compound in the presence of mercurous ion, is chosen as the masking agent in this study. -127— B. Procedures of Making An MFE Using Chloride Ion As A Masking Agent The platinum electrode to be coated with mercury is cleaned in hot perchloric acid and treated with aqua regia for several minutes. Two plating baths are used. One contains 1.2 N HCl and 0.1 N HgClZ. The other contains 0.1 N HgCth and 0.1 N H0104. The plating is first carried out in a HgCl2 bath. A potential of approximately 1 V is used for this primary plating process for several seconds. Figure u-l shows a typical current response. Then the second solutionis introduced after a brief dipping in hot conc H01 and rinsing with distilled water. The mercury plating is then resumed. A typical reduction current used is 5 mA/cma. The reduction is continued until a desired thickness of mercury plate is obtained. C. Results and Discussion 1. Product of the Primary Plating The electrode surface appears gray after the primary plating. Under microscOpic examination, the platinum surface is covered with a calomel precipitate which has a rugged spongy appearance. - 128 - 80.0 . 60.0 . 40.0 :- uA 20.0. 0.0 1 1 1 1 1 T START PLATING Figure 4-1. Mercury Plating at a Constant Potential, Electrolyte Contains 1.2 N HCl and 0.1N HgCla - 129 - Several reactions are involved during the film making process. At the primary mercury plating stage, these are : a. Electrochemical Reduction of Hg(ll) Ion (a-l) Hg++ + Ze' + (Pt) --9’ Hg(Pt) slow (a-z) Hg++ + 2e” + (Hg) -—-> Hg(Hg) fast b. Chemical Reduction of Hg(II) Ion Hg(Hg) + Hg” ——> Hg;+ + (H5) + c. Precipitation of Hg; by Chloride Ion Hg2+ + 2 01‘ ——> Hg2012 l When the reduction current is applied, mercuric ions are plated on the platinum surface as indicated by reaction (a-l). The initial mercury is deposited on the most active sites of the platinum surface. Once these nuclei of mercury deposition are formed, a competition starts between depOsition on the unplated Pt surface and deposition on these nuclei as indicated by reaction (a-2). The former sites are more favorable. In the meantime , a second reduction takes -130- place on these favored sites; that is,the chemical reduction of mercuric ion by mercury shown as reaction b. This reaction is known to be quite fast (37). The product of this reaction is mercurous ion which forms precipitate with the chloride ion in the solution immediately by reaction 0. Because the calomel formed is quite insoluble in the aqueous solution, it immediately precipitates and is adsorbed on the mercury surface. The success of this plating process relies on the peculiar combination of reactions b and c. From reaction c, it is seen that the chloride forms a precipitate only with mercurous ion. But as reaction b indicates, the only place where these ions are generated is the place where mercury already exists. Thus, the precipitate is only formed on the place where mercury has been deposited. Once the precipitate forms, it creates an insoluble film over the mercury deposit and retards the electroreduction on these places. This forces the reduction to continue on places where mercury has not yet been formed. In this way, the plating of the mercury goes on evenly over the whole surface of the platinum electrode. ..13]- 2. Continual Mercury Plating By this time, the platinum electrode surface is covered with a thin film of calomel. This film is removed by the reduction reaction in hot conc 361. The mercury plating proceeds smoothly after this step. After an appreciable amount of mercury is plated, the electrode surface appears mirror-like under the microscope. It is interesting to note the difference in mercury obtained with and without the primary plating step. with the same reducing current density without the primary plating step, the platinum electrode surface is covered with mercury droplets. 3. Factors that Influence the MFE Obtained a. Geometry of the Electrode Electrodes with other geometry have been prepared including a spherical Pt electrode and a wire electrode. The spherical electrode is made by heating the tip of a Pt wire electrode. At the melting point, the Pt melts and forms a small Pt ball at the wire tip of about 1 mm in diameter. The wire electrode is made by sealing one end of - 132 - a piece of platinum wire in the soft glass leaving an approximately 1 cm length of wire exposed. The diameter of the electrode is approximately 1 mm. The plating procedure described earlier is performed. A smooth mercury film was obtained in both cases. b. Influence of Aqua Regia The pretreatment of aqua regia on the platinum surface seems essential for the success for making smooth mercury film after the electrode is polished with fine grit silicon carbide sandpaper (650 grit). Several attempts to plate mercury on the planar platinum electrode following all the procedures described above except the aqua regia treatment resulted in mercury droplets on the platinum surface. Since the Pt surface after the sandpaper polishing is quite non-uniform physically as evidenced by the random scatches and marks seen under microscopic examination. This causes the platinum to have a non-uniform surface activity and thus causes the failure to obtain a smooth mercury coating. - 133 - One other reason could be due to the surface changes associated with the aqua regia treatment. The aqua regia attacks the surface Pt atoms. Various compounds between chlorine and platinum have been found including PtCl, 3 and PtClQ. It is quite possible that after the aqua regia treatment, chlorine atoms are chemisorbed PtCla, PtCl on the Pt surface. The chloride covered surface might be more hospitable for the plating of mercury than aplatinum surface covered with either hydrogen or oxygen. This might be the reason that the Pt surface without treatment of aqua regia is difficult to plate with a smooth mercury film. - 154 - CHAPTER V DETERMINATION OF THE Pt-Hg REACTION RATE ON THE MERCURY COATED PLATINUM ELECTRODE A. Introduction 1. SCOpe of the Pt-Hg Reaction Rate Study The application of the mercury film electrode (MFR) in electrochemical studies has not gained its full potential for some apparent reasons. Making a smooth mercury film is difficult. The supporting materials used for the mercury film show either high affinity with mercury which results in a fast Pt-mercury'. reaction or no affinity which results in the formation of mercury droplets loosely hanging from the surface of the supporting material. Platinum shows a certain affinity for mercury. The hanging mercury drop electrode supported by platinum has been shown to possess good stability with time. But the application of platinum as the supporting material for a mercury film electrode is handicapped by the difficulty involved in creating a thin smooth mercury film. As a consequence, the kinetic data for the Pt-Hg reaction is lacking from the literature partly due to the difficulty of achieving reproducible -155... results (29). This knowledge is important for the application of MFE in the electrochemical studies. The study of the effect of the surface state on the mercury deposit described in the previous chapters revealed that two types of mercury deposit exist on the platinum surface, namely mercury droplets and mercury patches. By lightly platinizing the platinum surface, the platinum surface can be coated with a smooth mercury film by an electroplating process. This finding is crucial for the study of the reaction rate of platinum and mercury because unless the mercury forms a smooth film on the platinum surface, the rate study can't be expected to generate reproducible data. 2. Relevant Results Obtained by Other Workers a. Existance of the Pt-Hg Intermetallic Compound The reaction of the platinum with the mercury coat leads to the formation of a Pt-Hg intermetallic compound. This compound has a definite composition. The structure has been determined as Ptng (15, 33, 41). The intermetallic compound exhibits several electrochemical properties different from either the platinum electrode or the pure mercury electrode. The oxidation of the mercury in the intermetallic compound occurs at a more anodic potential than that of metallic mercury. This property has been used in the scanning voltammetr1c(17) and the Chronopotentiometric (29) studies of the Hg-Pt interaction. With a constant anodic current, the chronopotentiogram of the mercury stripping reaction using a mercury coated platinum electrode shows three potential plateaus. A typical plot is shown in Figure 5-1. The stripping is carried out in 0.1 MTHClog solution. The potentials of these plateaus are 0.54 V, 0.12 V and 1.7 V, corresponding to the ionization of the metallic mercury, the ionization of Hg in the Pt-Hg intermetallic compound and the oxidation of the supporting electrolyte respectively (29). Similar results were obtained with the scanning voltammetric method(17)o Since mercury has three oxidation states, 0, +1 and +2, the ionized product of the mercury oxidation has the possibility of taking either one of the two oxidized states. Theoretical calculations indicate that the oxidation state of the mercury ion generated by electrochemical oxidation depends on the applied potential. According to Hassen, Untereker and Bruckenstein (17), at E ( 0.85 V, only Hg:+ is produced; at E ) 1.10 V, only Hg++ ion is produced; while at 1.10 V >'E > 0.85 V, a mixture of Hg;+ and Hg++ is produced. Thus, in Fig. 5-1, the ionization product -137... H6 STRIPPINB 1. 800" X10 00498 i I 1 3.000 6.000 )JCOUL. x10 '4 , Figure 5-1. Potential-Charge Response of MFB under Anodic Current in 0.1 M HClOl+ -138- during the second potential plateau is Hg++ ion while the ionization product during the first plateau is Hg;+. This was confirmed by UV spectrophotometric methods on the resulting solutions (29) and also by the electrochemical method (17). b. Pt-Hg Reaction Rate Study The rate of Pt-Hg reaction on the MFE is determined by measuring the quantity of the compound formed with time. Both a gravimetric method (53) and electrochemical methods have been used (29, 17). Barlow and Planting used a gravimetric method to determined the Pt-Hg reaction rate at elevated temperature. This technique is quite insensitive and is only applicable with a large quantity of Pt-Hg compound. The cyclic voltammstric technique was used by Hassen, Untereker and Bruckenstein (17) in their study of the Pt-Hg interaction involving a few atomic layers of Hg. Brubaker (29) used the chronopotentiometric method in his mercury-platinum system study. The second method enjoys the advantage of simple calculation for the number of coulombs involved in the process. With the computer controlled coulostatic system, the Pt-Hg reaction can be studied within a much shorter period -159- of time than possible by other previous methods. Since a known amount of charge is delivered in each case, there is little difficulty in calculating the charge involved. In addition, the sequence of the polarization process is easily modified to meet the need of the study. B. Experimental 1. Preparation of the Platinum Surface for the Rate Study The reaction rate of Pt with Hg is apparently affected by the amount of Pt-Hg compound between the metallic mercury and the platinum. An analysis of the rate data has to include a consideration of the amount of the Pt-Hg compound already formed. For simplification of the mathematical treatment, the Pt-Hg compound covered Pt electrode is used. Such an electrode only has a layer of Pt-Hg compound on the surface, and is prepared by plating a certain amount of mercury and then stripping the metallic mercury off the surface after various periods of aging time allowed for the reaction to preceed. The platinum electrodes in this study were lightly platinized using the procedure described previously. The RF of the lightly platinized electrodes ranged between 4.0 and 10.0. - 140 - 2. Experimental Set-Up The set-up is essentially identical to that used in the previous chapters. The Pt-Hg compound covered Pt electrode, PMCCPE, is obtained from the smooth MFE under the control of a program called STRIP. Under program control, the coulostatic generator issues an anodic current to ionize the the metallic mercury on the NFE. The potential during the stripping is measured every 30 microseconds. The stripping current is terminated when the metallic mercury on the surface is exhausted which is indicated by a sharp rise of potential. The potential value for the termination of current is 1.0 V vs. R.H.E. C. Experimental Results 1. Methods Used for the Pt-Hg Reaction Rate Study Two polarization techniques have been used for the Pt- Hg reaction .rate study : the current reversal technique and the aging technique. In both cases two steps with mercury plating and stripping are included. The mercury to be plated is only a very small fraction compared to the amount of existing Pt-Hg compound on the electrode surface so that the amount of Pt-Hg compound can be treated as constant during study. - 1g] - a. Current Reversal In this technique, a short reduction current pulse is applied,immediately followed by an anodic current pulse. The current program is shown in Figure 5-2a. If the reducing current is made so that only reduction of mercurous ion is allowed, the charge consumed is equal to the amount of plated mercury. During the stripping step, part of the deposited mercury is oxidized and removed at a low potential value. The remaining plated mercury reacts with platinum and become indistinguishable from the existing compound. The mercury in the compound is oxidized at a higher potential. Analyzing the resulting potential-time curve during the stripping step gives the quantity of mercury which remained in metallic state. The difference in charge quantities used for plating of mercury and for removing metallic mercury gives the amount of mercury transformed into Pt-Hg compound. b. Aging Method In this technique, an additional period is incorported in the polarization program. The scheme is shown in Figure 5-2b. The waiting period is'21ncluded to vary the time available for the Pt-Hg reaction. The reaction rate can be determined in the same way as the previous technique. - 142 - Plating Stripping anodic, Ia ' i’uA ' L— cathodic, -Ia Figure 5-2 a . Current Program Used for Current Reversal Technique Plating Aging Stripping J L. wait "A """ L l cathodic + , t t=0 Figure 5-2 b . Current Program Used for Aging Technique “. - 143 - 2. Anodic Stripping of MFE in Mercurous Solution Since the transformation of the metallic mercury into Pt-Hg compound is known to be fast,particularly when the quantity of the Pt-Hg compound is small, it is impractical to change the supporting electrolytes during the deposition and the stripping of mercury. Thus the whole plating- stripping process has been performed in one supporting electrolyte, 0.1 M HgClOl+ in 0.1 M HCIOH. The potential response of the mercury stripping process from the MFE in 0.1 M HClO# solution containing 0.1 M HgClO z, is similar to the response in the HClO solution without 1+ HgClO#. But the corresponding numerical values of the potential plateaus are different in these two solutions. Figure 5-3 shows a typical potential-charge variation during stripping in 0.1 M HCIOA electrolyte solution containing 0.1 M HgCth. This figure shows the stripping of the mercury on the MFE. The horizontal curve comes from the oxidation of the metallic mercury, a 11 mcouls of charge was consumed before the potential started to rise. The curve leveled off for the second time at 1.h0 V. The potential hump at the transition appears only when the stripping is started with metallic mercury on the electrode surface. -u.z+- coausaom msohconmz z —.0 ca max as» we oHpmfinmpomumco omnmnolamaucmuom canons .mlm mA:Mdm m- s; ...Dooozo .<5 00— H u ouspaamsa psonnzo .coapdaom msousoaoz 2 p.0 ca mm: commanpm assume: oaaampoz «o pcosaaomxm Hmmne>mm pmonnso .ssm ouswam e- six .m:ooozom usenaso cu museum: ofiaampez Ho oocmnmonmmmfim .le OHQMB - 148 - charge is required to plate one monolayer of mercury from mercurous ion. It takes a different period of time among the different current amplitudes used to supply this quantity of charge. Thus the actual contact time between the newly plated mercury and the Pt-Hg compound surface differs among these three cases. This should lead to‘a variation of the quantity of the mercury transformed into the compound. b. Reduction of Mercurous Ion with High Amplitude Current A reduction current of high amplitude, -2.400 mA, was chosen for the aging technique experiments. With this current amplitude, the mercurous ion at the electrode surface is depleted within a short time. Soon after the current is started, the charge supplied becomes predominantly used for the reduction of the supporting electrolyte. A potential- time curve is shown in Figure 5-5. At the beginning, the potential drops gradually which indicates that the concentraa tion of the mercurous ion at the electrode surface is decreasing. Then, the potential plunges sharply - 149 - 1.BGU"' m n I m'm . u d 9 l >u-c x V =: ..- 2 H -0.580 ;:;I;:;:;:g::::4:l:l:|i‘| 0.000 'rEJQ' CHARGE,NANOCOUL. Figure 5.5. Potential-charge variation during reduction of mercurous ions from 0.1 M HgClO# using high reduction current -150- to a rather cathodic potential. After that, the potential varies only slowly with time. The reduction of hydrogen ion becomes predominant. To have a 100 % current efficiency for the reduction of mercurous ion, the current pulse should be restricted to within 20 pcoul. In the following experiments described, the current pulses used are 1 msec. long which contains 2.A‘pcoul. of charge. According to calculation, this amount of charge is equivalent to 6 mercury monolayers on the electrode used. An aging period follows the mercury deposition, and then an anodic current pulse of 1.0 mA is issued. The potential-charge variation is recorded during this step at h0,us intervals. This deposit-strip cycle is repeated with a longer aging period after several seconds with the same electrode. The whole process is repeated several times after the last cycle. Then, a longer anodic current is issued which ionizes the remaining mercury in the Pt-Hg compound. The quantity of mercury in the compound is determined from the resulting potential-charge curve. Figure 5-6 was obtained from one of the experiments. This electrode required 0.60 mcoul. charge to remove the mercury in compound. Table 5-2 lists the results obtained in this particular experiment. .60, is the averaged value from four repetitions of one single experiment. - 151 - . UH . I 00 0C ' 3 a; ‘9 I 3 ease )- '“ -.. .flfifi 3 2 2 >< q 2 1 ' M . .1 p::“‘ I] 111 .P - . - - :cc‘;$$$c“ [lip/[If/A 111 ; 'Icfrcfchff;:IIc/CIF;IC;;;IIIf;I’I‘I“;;:;;;;fifff;ffffffffff/I/// -_ 1 CHARGE, )JCOUL. Figure 5-6. Potential-charge variations during repetitive deposition-stripping of mercury with varying aSing intervals - 152 - 9.0 H 00.. 091m 80 u :00 000.. No.0 H 3.0 000.0 8.0 H -.0 m0m.0 6.0 H 9.0 i$0.0 6.0 H no.0 000.0 2300* .34 .000 .039 .gooa 00¢ a $79.5 .psosanmmxm msfiw< 0:» ca humane: ofiaamuoz «o eccenmommmmwa .mlm mHnt -153- The aging period, T e’ is the sum of the time interval as during the deposition of mercury, the waiting period, and the time required to strip all the metallic mercury remaining on the surface compound. AQ, the charge equivalent of mercury converted to the Pt-Hg compound increases with the aging period. A linear relationship is found when AQ is )1/2 8 . Figure 5-7 shows the results from plotted vs. (Tag electrodes with varying amountsi.2L (X,p) Apply the Laplace transform to equation (4) with respect to X to obtain, 30 x _ 2 g .22 (5) P C(q.p) ’ D” C(cup) " q Camp) " ( ax )x=0) Move D q2 C to the right side of equation and (gap) 2 divide equation (5) by ( q - ‘%E' which gives x=0 ) D 3C (6’ Cum) " (1.2 P ‘ ‘1 C(¢,p) " ‘ a x ’ .— By solving equation (6) for x using reverse Laplace transformation, we have .. it} (7) c(x,p) — D C(¢’P) Cosh ( D ) x ac i + D ( (X’P) Sinh (-%—)‘ x E13 3 x )x=0 - 161 - By rearranging equation (7), we have 1 (ii-)3" C 2' 3C X D (8) C(x,p) = ( Jgafl + ;I-D? max )x=0) e + ( Jgal’L .. :5;- ( aiP ) =O)e The first term on the right side of equation (8) is 0, because as X ——§oo , C(x p) = 0. So that 9 1 D7 achhp) (9) C(flep) = " Tap ( ax )x=o Since the concentration of the metallic mercury at x=0 is equal to 1, l d U C(¢.t) ’ Cum) = L ( C(M)’ = T Thus, - 162 - or X <1” (“drake ‘ “W Reverse the transformation of equation (11) with reapect to 1 which gives dN BC. 1 (12) """"""" = ( —"""', ) _, = " ‘E dt a}; X-O D2 ti. “E .. 2 0 1T2 D, The decrease of metallic mercury dN is shown proportional to the square root of time. - 163 - A.linear relationship between N, the amount of metallic i mercury, and ta is expected from this derivation. For Pt-Hg reaction on the PMCCPE, this relationship is observed in the experiment as is shown in Figure 5-7. In this derivation, Fick's Laws of Diffusion are used. The actual transportation mechanism of the metallic mercury through the reaction is not known. Since only one crystalline structure between mercury and platinum has been found on the mercury coated platinum electrode, the mercury atom transport in the Pt-Hg compound must be different from the diffusion phenomena occurring in the liquid state. Equation (2) shows that the decrease of metallic mercury is proportional to t% with the proportional constant -§§;; . Unlike the diffusion process in liquid state, D,"the diffusion constant, shows a large variation when the amount of compound is varied. -164- CHAPTER VI FUTURE STUDY A smooth mercury film electrode obtained by the chloride ion masking procedure has been applied in the anodic stripping analysis of cadmium ion. A preliminary study shows about a 10-fold increase in sensitivity, down to 10'"7 M using the film electrode as compared to the hanging drop mercury electrode. Higher sensitivity is possible through the optimization of the procedures. This can be achieved through the usage of rotational MFE electrode, thinner mercury film, and the coulostatic anodic stripping technique. APPENDICES - 165 - APPENDIX 1: Computer Interface A special computer interfacing circuit called the 11-8 Emulator was designed by Dr. Enke's research group for the PDP 11/40 computer. This interface enables the PDP 11/40 to generate several control signals similiar to the PDP 8 computer. The complete interface consists of a general device interface, the 11-8 Emulator and an interface buffer. The block diagram is shown in Figure A. 1. The general device interface, DR11-C, provides the logic and buffer registers necessary for program - controlled parallel transfer of 16-bit data between a PDP 11 system and the external device. It has three registers : control and status, input buffer and output buffer. Each register has an address. Data is written into or read from the corresponding register when a particular register is addressed. The 11-8 Emulator provides the logic necessary to generate control signal lines and data lines similiar to these of the PDP 8 system. These control lines are three IOP lines, six DS line and one SKP line. - 166 - .mpzn .v mxm z" DECHARGE. OTHERWISE ‘0 - ho .- GUANM CONT: CKA4: COMP: SUM: DIV: RETURN: FINISH: I INPUT: READER: PUTRI: -176- WRITE MSI MOV *16011.0UT NOP TSTB CSR BPL G2 RSTCK A1 MOV #DATABF.R2 BIS #200.CKINIT SENT PULSE BIT #100.CKCSR BEG CKA4 BIS #200.CKINIT SENT CONVRT FLAG DONE.BI DATAAG ADREAD.(R2) COM (R2) BIC DEC BNE MOV MOV ADD SOB MOV BIC ROR SOB MOV #17000O.(R2)+ NPOINT CKA4 #20.R4 #DATABF.R2 (R2)+.R3 R4.SUM #4.R4 #1.R3 R3 R4.DIV RET.R5 MOV R3.@2 DATA BUFFER HEAD CLEAR LEAST SIGNIFICANT BIT DEVIDE BY 16 SAVE R0 RO->BUFFERHEAD SET ADDRESS FOR LINE BUFFER SET MAX WORD COUNT SAVE R5 POINT TO BUFFER HEADER STORE THE CONTENT OF R1 BYTE COUNT GUANM COUNT1: LINBK: BUFHD: M51: DATABF: NPOINT: NPLSE: PVOL: PH: COUNT: RET: DATAHD: TMP: N1: TEASE: CHANNO: PL: ICHAN: CKBASE: INC R0 TSTB (R1)+ BNE COUNTI MOV R0.4(R5) MOV #LINBK.RO .INIT R0 .WRITE R0.R5 .WAIT R0 .RLSE R0 MOV (SP)+.R5 RTS R5 .EVEN .WORD FINISH .WORD 0000 .RAD5O IKB/ .BYTE 1.0 .RADSO /KB/ .WORD 6 .BYTE 6.1 .WORD 6 .WORD 0 .BYTE 15.12 - 177.- o I c I .ASCIZ /SHORT CAP./ .EVEN .BLKW .WORD .WORD .WORD .WORD .WORD .WORD .WORD .WORD .WORD .WORD .WORD .WORD .WORD .WORD .END GUANM O MOOOOOOOOOOOOOM COUNTING THE NO. OF BYTE STORE THE ACTUAL BYTE COUNT MOVE TO THE DEVICE HANDLER PREPULSE POTENTIAL. SLOPE 00000000000000000 000 000 201 903 904 -178- SLOPE.FTN MINCHEN WANG OCT.29 1977 SUBROUTINE CALLED LINFIT =’) 102 FORMAT(I2) 103 FORMAT(’$NO. OF DATA POINTS IN A FILE {***)=’) 104 FORMAT(I3) 105 FORMAT(’$ FILE NAME ’1 125 FORMAT(’$NEW FILE NAME ’) 106 FORMAT(20A1) END CAPF3 0000 00000000000000000000000 00000 - 180 - DIFFERENTIAL CAPACITANCE MEASUREMENT CAPF3.FTN MINCHEN WANG SEP 5 77 SUBROUTINE CALLED: DCAP(CAPM3.MAC).LINFIT(LINFIB.FTN) A MODIFFICATION OF CAPF2.FTN TO ALLOW ENTERING OF PULSE HIGH AND LENGTH TWO DATA FILES ARE CREATED :ORIGINAL FILE AND FITTED EXPOENTIAL DECAY DATA FILE . CAP=GIV LOG V=LOG V(T=O)-E(NF/RT)(I/CAP)TJ DIMENSION IV(100).VI(100).IT(100).TIME(100) DIMENSION VJ(100).SIGMAY(100) BYTE FNAME(40).TODAY(9).AA(2) EGUIVALENCE (IFLAG.AA(1)) MAXIMUN CURENT=3.5 M AMP 1 MAMP HAS A D/A SETTING OF 1600 WRITE(6.116) READ(6.117)CURNT VO=-(2.0*CURNT-5000.0)/2.4414 JPH=IFIXBUFFERHEAD ISET ADDRESS FOR LINE BUFFER BUFHD .RLSE #LINBK MOV (SP)+.R0 MOV (SP)+.R4 RTS R5 MOV R5.-(SP) 3SAVE R5 MOV #8UFHD.R5 3POINT TO BUFFER HEADER MOV R1.6(R5) 3STORE ADDRESS OF LINE HEAD CLR R0 3DATA COUNT INC R0 TSTB (R1)+ 3COUNTING THE NO. OF DATA BNE COUNTI MOV R0.4(R5) MOV #LINBK.R0 .INIT R0 .WRITE R0.R5 .WAIT R0 .RLSE R0 3STORE THE ACTUAL COUNT IMOVE TO THE DEVICE HANDLER CHRONO LINBK: BUFHD: M51: TMP: M52: TMPI: COUNT: RET: DATAHD: N1: TBASE: PL: IVO: ICHAN: NDATA: - 197.. MOV (SP)+.R5 RTS R5 .EVEN .WORD FINISH .WORD 0000 .RADSO IKB/ .BYTE 1.0 .RAD50 lKB/ .WORD 6 .BYTE 6.1 .WORD 6 .WORD O .BYTE 15.12 .ASCII IDAC OUT .BLKB 6 .WORD 0 .BYTE 15.12 .ASCII IDAC OUT .BLKB .WORD .WORD .WORD .WORD .WORD .WORD .WORD .WORD .WORD .WORD 145 .END LINT ll \ ll \ 0080000000 0 DEPOSF 000000000000000000000 00 000 11 12 13 20 21 -198- DEPOSF.FTN MINCHEN WANG APRIL 20 1978 A MODIFICATION VERSION OF THE CURRENT REVERSAL EXP. SUBROUTINE CALLED DEPOSM.MAC Q.DEP=2.4 MA*1000 USEC. ENTER FILE NAME AS FL:DEPOSA.01 2.4 UCOUL NEGATIVE CHARGE IS DUMPED ON THE WORKING ELECTRODE. STRIPPING IS STARTED AFTER A PRESELECTED WAITTING PERIOD. DIMENSION A(260).TIME(260).IV(260) BYTE FNAME(20).LINE(80).TODAY(10).CNTR(2) EGUIVALENCE (JFLAG.LINE(1)).(FNAME(11).CNTR) FIVE DATA FILES ARE CREATED WRITE(6.11) READ(6.12)AMP FORMAT(’$AMPLIFICATION FACTOR ’) FORMAT(F10.5) WRITE(6.13) FORMAT(’$ REF. ELECTRODE POTENTIAL VS. R.H.E ’) READ(6.12)REF WRITE(6.20) FORMAT(’$IR DROP.MV. PER 100 UA CURRENT ’) READ(6.21)BSE FORMAT(F15.6) CATHO=~2400.0 FCTR=2.4 WRITE(6.125) READ(6.106)FNAME WRITE(6.106)FNAME NPTS=1OO WAIT FOR 10**ITWAIT USEC.-1000USEC. ANODI=-CATHO/FCTR BSE=-BSE*ANODI/100.0 VCATHOt-(CATHO*2.0-5000.0)/2.4414 VANODI=-(ANODI*2.0-5000.0)/2.4414 IANODI=IFIX(VANODI) DEPOSF 99 111 113 110 115 112 114 126 129 133 102 104 105 125 106 107 108 109 -‘|99- ICATHO=IFIX(VCATHO) IFCTR=IFIX(FCTR) ITWAIT=3 II=1 CONTINUE CALL DEPOS(IV(1).ICATHO.IANODI.ITWAIT) DO 1 I=1.NPTS IV(I)=IV(I)-2048 A(I)=FLOAT(IV(I))/(AMP*O.4096) CONTINUE ENCODE(4.109.CNTR)II CALL ASSIGN(4.FNAME.20.IERR) WRITE(4.111)FNAME FORMAT(’3FNAME:'4OA1) CALL DATE(TODAY) WRITE(4.113)TODAY FORMAT(’3 '.20A1) WRITE(4.110)CATHO.ANODI FORMAT('3 CATHODIC CURRENT='F15.6. 1’ ANODIC CURRENT=’F15.6) WRITE(4.112)BSE WRITE(4.115)ITWAIT FORMAT(’3WAITING FOR 10**’I2’ USEC. - 1 MSEC. FORMAT(’3IR DROP IS ’.F15.5.’MV. PER 100 UA’) WRITE(4.114) FORMAT(’3 MUCOUL. MV. VS. R.H.E.') DO 126 I=1.100 VOLT=A(I)+BSE+REF CHARGE=FLOAT(I)*ANODI*0.04 WRITE(4.129)CHARGE.VOLT CONTINUE FORMAT(’RD’.2F15.5) WRITE(4.133) FORMAT(’ED’) END FILE 4 FORMAT(I2) FORMAT(I3) FORMAT(’$ FILE NAME ’) FORMAT(’$NEW FILE NAME ’) FORMAT(2OA1) ' FORMAT(' NEXT FILE ’20A1) FORMAT('$NEW CHAR. ’) ITWAIT=ITWAIT+1 II=II+1 FORMAT(I2) IF (ITWAIT.LT.8) GOTO 99 END ’) DEPOSM .0 -0 .0 .0 DEPOS: T7: T8: CONT: CONT1: CHAN1: A1: A2: - 200 - .TITLE .IDENT .PSECT .NLIST .GLOBL DEPOSITION IA/ PARAME.RW.I.REL TTM.BEX.MEB DEPOS APPLY A SHORT CURRENT PULSE FOLLOWED BY A WAITING PERIOD THEN. A STRIPPING PULSE IS ISSUED .MCALL .INIT..WRITE.. WAIT..EXIT..READ..RLSE .MCALL .BIN20..F4DEF..O2BIN .F4DEF 0 .PSECT MOV MOV MOV MOV MOV MOV MOV MOV MOV CMP BMI BNE T7 MOV #6.TMP BR CONT1 CMP WAIT.#7 BNE T8 MOV #22.TMP BR CONT1 MOV #100.TMP BR CONT1 MOV WAIT.TWAIT MOV TWAIT.CKCSR 2(R5).BFADDR BFADDR.R2 R5.RET 04(R5).PH1 26(R5).PH2 010(R5).WAIT #47.PGCNTR #2.TMP #144.NDATA WAIT.#6 CONT MOV #1750.PL1 MOV #6200.PL2 BIS #14000.CKCSR SETA HEIGHT.PHI SETA LENGTH.PL1 SETN TIMEBS.0 SETN CHANLD.4000 SENT PULSE MOV Q1.CKINIT BIT #100.CKCSR BEG A1 BIS #200.CKINIT BIT “100.CKCSR 5050‘...» CATHODIC CURRENT WAITING PERIOD 40 U SEC. 2 * TWAIT 100 POINTS 1.8 S 6.4 S. PL=1000 US. PL2= 3200 US. DOWN. PRESET . 1 MSE. TB CLEAR DM DIVIDER DEPOSM A3: A4: TRANS: A6: MORE: FINISH: WAIT: PHI: PL1: PL2: PH2: RET: TMP: N1: TWAIT: ICHAN: NDATA: BFADDR: - 201 - BEG A2 BIS #200.CKINIT DEC TMP BNE A2 SETA HEIGHT.PH2 SETA LENGTH.PL2 MOV #0.CKCSR MOV #14000.CKCSR RSTCK A3 SENT PULSE BIS #100.CKINIT SENT CONVRT BIT #40000.CKCSR BEQ A4 DATAAQ ADREAD.(R2)+ DEC NDATA BEQ TRANS BIS #100.CKINIT BR A3 MOV 07.177566 MOV BFADDR.R2 BIS #1.CKINIT COM (R2) BIC #170000.(R2)+ INC NDATA CMP #144.NDATA BPL A6 BIS #200.CKINIT BIT #100.CKCSR BNE MORE RTS R5 .EXIT .WORD .WORD .WORD .WORD .WORD .WORD .WORD .WORD .WORD .WORD .WORD .WORD .END DEPOS ObOOOOOOOOOO ETREAF 0000000000000 00 0000 99 11 12 13 111 113 112 114 - 202 - ETREAF.FTN MINCHEN WANG FEB 14 1978 THIS PROGRAM CALCULATE THE CELL IR FROM THE EXTRAPOLATED POTENTIAL VALUE . DOSE IR CORRECTION DIMENSION A(210).TIME(210).IV(210) BYTE FNAME(20).LINE(80).TODAY(10).CNTR(1) EGUIVALENCE (JFLAG.LINE(1)).(FNAME(9).CNTR) DATA FILE CREATED WRITE(6.125) READ(6.106)FNAME WRITE(6.106)FNAME NPTS=200 CALL ETREAT(IV(1)) WRITE(6.11) READ(6.12)AMP FORMAT(’$AMPLIFICATION FACTOR ') FORMAT(F10.5) WRITE(6.13) FORMAT(’$ REF. ELECTRODE POTENTIAL VS. READ(6.12)REF DO 1 I=1.NPTS IV(I)=IV(I)-2048 A(I)=FLOAT(IV(I))/(AMP*O.4096) CONTINUE BSE=(A(100)-A(101))/2.0 CALL ASSIGN(4.FNAME.20.IERR) WRITE(4.111)FNAME FORMAT(’3FNAME:'4OA1) CALL DATE(TODAY) WRITE(4.113)TODAY FORMAT(’3 ’.20A1) WRITE(4.112)BSE FORMAT(’3IR DROP IS ’.F15.5.’MV.’) WRITE(4.114) FORMAT('3 MUCOUL. MV.’) DO 126 I=1.100 R.H.E ’) ETREAF 126 127 129 133 102 104 105 125 106 107 108 109 ..203- VOLT=A(I)-BSE+REF CHARGE=FLOAT(I)*50.0 WRITE(4.129)CHARGE.VOLT CONTINUE DO 127 I=101.NPTS VOLT=A(I)+BSE+REF CHARGE=FLOAT(I)*50.0 WRITE(4.129)CHARGE.VOLT CONTINUE FORMAT(’RD’.2F15.5) WRITE(4.133) FORMAT(’ED’) END FILE 4 FORMAT(I2) FORMAT(IS) FORMAT(’$ FILE NAME ’) FORMAT(’$NEW FILE NAME ’) FORMAT(20A1) FORMAT(’ NEXT FILE ’20A1) FORMAT(’$NEW CHAR. ’) FORMAT(1A1) GO TO 99 END ETREMA - 00 .0 .0 \o 5. .0 i ETREAT: CHAN2: POINV: CKA1: START: POLINV: AGAIN: CONV: -204- .TITLE ELECTROCHEMICAL TREATMENT .IDENT /A1/ .PSECT ETREAT.RW.I.REL .NLIST TTM.BEX.MEB .GLOBL ETREAT USAGE: THIS SUBROUTINE IS CALLED BY ETREAT.FTN PURPOSE: DELIVER 250. UAMP ALTERNATING CURRENT STEPS FOR 51 CYCLES. PERIOD IS 0.2 5. FEB 16 78 .MCALL .INIT..WRITE..WAIT..EXIT..READ..RLSE .MCALL .BIN20..F4DEF..O2BIN .F4DEF 0 .PSECT .=.+100 MOV 2(R5).R2 MOV R2.DATALK MOV R5.RET MOV #144.CYCLE MOV #5.TBASE 3 100 MS TIME BASE MOV “16011.0UT ilERO BUFFER OA SETA TIMEBS.TBASE 3PULSE TIME BASE.10 M SEC SETN CHANLD.2000 32 CHANNEL SETA LENGTH.PL MOV #3.CKCSR 3MS. FOR TIME BASE RSTCK A1 BIS #200.CKINIT 3CLEAR DM FLAG SENT PULSE NOP SETA HEIGHT.IVO MOV #144.COUNT BIT 3100.CKCSR 3 TIME YET? BEG CKA1 - BIS #200.CKINIT DEC COUNT BNE CKA1 COM IVO BIC #170000.IVO DEC CYCLE BEQ START JMP POINV MOV DATALK.R2 MOV #2.INDEX SETA HEIGHT.IVO MOV 0144.COUNT MOV #4.NDATA SENT CONVRT FLAG DONE.AA ETREMA CKA2: LAST: MORE: FINISH: INPUT: 3 READER: PUTR1: COUNT1: - 205 - DATAAQ ADREAD.-(SP) ADD (SP)+.(R2) DEC NDATA BNE CONV BIT #100.CKCSR BEG CKA2 ASR (R2)+ BIS #200.CKINIT DEC COUNT BNE AGAIN DEC INDEX BEQ LAST COM IVO BIC #170000.IVO MOV #144.COUNT JMP POLINV SETN TIMEBS.O MOV RET.R5 MOV 0310.COUNT MOV DATALK.R2 ASR (R2) COM (R2) BIC #170000.(R2)+ DEC COUNT BNE MORE RTS R5 .EXIT MOV RO.-(SP) MOV #BUFHD.R0 MOV R4.6(RO) MOV NUMW.(R0) .INIT GLINBK 3 TIME YET? 3INVERT VOLTAGE 3SAVE R0 3 R0->BUFFERHEAD 35ET ADDRESS FOR LINE BUFFER 35ET MAX WORD COUNT . READ “LINBKfliBUFHD .WAIT #LINBK .RLSE #LINBK MOV (SP)+.R0 MOV (SP)+.R4 RTS R5 MOV R5.-(SP) MOV #BUFHD.R5 MOV R1.6(R5) CLR R0 INC R0 TSTB (R1)+ BNE COUNT1 MOV R0.4(R5) MOV QLINBK.RO .INIT R0 .WRITE R0.R5 .WAIT R0 . SAVE R5 3 POINT TO BUFFER HEADER . STORE ADDRESS OF LINE . HEADER 3 DATA COUNT 3 COUNTING THE NO. OF DATA 3 STORE THE ACTUAL COUNT 3 MOVE TO THE DEVICE HANDLER ETREMA LINBK: BUFHD: M51: TMP: M52: TMPI: COUNT: RET: DATAHD: CYCLE: INDEX: N1: TBASE: PL: IVO: ICHAN: DATALK: NDATA: .RLSE R0 MOV (SP)+.R5 RTS R5 .EVEN .WORD FINISH .WORD 0000 .RAD50 IKB/ .BYTE 1.0 .RAD50 IKB/ .WORD 6 .BYTE 6.1 .WORD 6 .WORD 0 .BYTE 15.12 .ASCII IDAC OUT .BLKB 6 .WORD 0 .BYTE 15.12 .ASCII IDAC OUT .BLKB .WORD .WORD .WORD .WORD .WORD .WORD .WORD .WORD .WORD .WORD .WORD .WORD .WORD .END ETREAT «50000000000000~ -206- -207- PLATINIZE he - ‘o \o - So ETREAT: E1: CHAN1: POINV: CKA1: START: POLINV: AGAIN: CKA2: .TITLE PLATINIZATION .IDENT IBll .PSECT PLATINZ.RW.I.REL .NLIST TTM.BEX.MEB .GLOBL ETREAT EACH RUN CONSISTS OF 20480 CYCLE. 30 MS. CURRNET STEPS MAY 78 .MCALL .INIT..WRITE..QAIT..EXIT..READ..RLSE .MCALL .BIN20..F4DEF..D2BIN .F4DEF 0 .PSECT WRITE M51 READ NOCYC .D2BIN “NOCYC MOV (SP)+.CYCCNT MOV (SP)+.JUNK MOV #50000.CYCLE MOV #5.TBASE MOV #16011.0UT SETA TIMEBS.TBASE MOV #2314.IVO SETN‘CHANLD.4000 SETA LENGTH.PL MOV #3.CKCSR 3 RSTCK A1 BIS #200.CKINIT 3 SENT PULSE NOP SETA HEIGHT.IVO MOV #17.COUNT BIT #100.CKCSR 3 BEQ CKA1 BIS #200.CKINIT DEC COUNT BNE CKA1 COM IVO BIC #170000.IVO DEC CYCLE BEQ START JMP POINV MOV #2.INDEX SETA HEIGHT.IVO MOV #17.COUNT NOP BIT #100.CKCSR 3 BEQ CKA2 MQoMQoM§ofio CLEAR STAKE 20480 CYCLES 100 MS TIME BASE ZERO BUFFER OA PULSE TIME BASE.10 M SEC 1000 UAMP ON CHANNEL 1 CHANNEL 1 MS. FOR TIME BASE CLEAR DM FLAG TIME YET? TIME YET? PLATINIZE LAST: FINISH: INPUT: READER: PUTR1: COUNT1: LINBK: BUFHD: BIS DEC BNE DEC BEQ COM BIC MOV JMP #200.CKINIT COUNT AGAIN INDEX LAST ~208- IVO 3 INVERT VOLTAGE #170000.IVO 0144.COUNT POLINV SETN TIMEBS.O DEC BEQ JMP CYCCNT FINISH E1 .EXIT MOV MOV MOV MOV RO.-(SP) #BUFHD.R0 R4.6(R0) NUMW.(RO) .INIT #LINBK .READ #LINBK. .WAIT #LINBK .RLSE #LINBK MOV MOV RTS MOV MOV MOV CLR INC (SP)+.R0 (SP)+.R4 R5 R5.-(SP) #BUFHD.R5 R1.6(R5) R0 R0 TSTB (R1)+ BNE COUNT1 MOV R0.4(R5) MOV #LINBK.R0 .INIT R0 .WRITE R0.R5 .WAIT R0 .RLSE R0 MOV (SP)+.R5 RTS R5 .EVEN .WORD FINISH .WORD 0000 .RAD50 lKB/ .BYTE 1.0 .RADSO IKB/ .WORD 6 .BYTE 6.1 .WORD 6 ..‘0..~0 BUFHD i SAVE R0 RO-DBUFFERHEAD SET ADDRESS FOR LINE BUFFER SET MAX WORD COUNT SAVE R5 POINT TO BUFFER HEADER STORE CONTENT OF R1 3BYTE COUNT i i i COUNTING THE NO. OF BYTES STORE THE ACTUAL COUNT MOVE TO THE DEVICE HANDLER PLATINIZE M51: COUNT: RET: DATAHD: CYCLE: INDEX: N1: TBASE: PL: IVO: ICHAN: DATALK: NDATA: NOCYC: JUNK: CYCCNT: -209- .WORD O .BYTE .ASCII INO. OF REPEAT.(DECIMAL)/ .EVEN .WORD .WORD .WORD .WORD .WORD .WORD .WORD .WORD .WORD .WORD .WORD .WORD .WORD .BLKB .WORD .WORD .END ETREAT 15.12 OOO~¥IOOOV00000000 STRIPF 00000000000000000 00000 207 209 200 99 100 201 203 204 205 FORMAT(‘SIR DROP PER 100UA.=’) - 210 - STRIPF.FTN FEB.13 78 WANG VERSION: A AMP AMPLIFICATION FACTOR IS TO BE READ IN SUBROUTINE CALLED: LINT(FROM STRIPM.MAC) DIMENSION JVOLT(500) DOUBLE PRECISION VO BYTE FNAME(40).TODAY(9) WRITE(6.115) READ(6.114)AMP WRITE(6.207) READ(6.208)REF FORMAT(‘SPOTENTIAL OF THE REF. FORMAT(F15.5) WRITE(6.200) READ(6.201)ICHAN FORMAT(‘SCHANNEL NO.= ') WRITE(6.99) FORMAT(‘SFILE NAME: ’) READ(6.100)FNAME FORMAT(4OA1) FORMAT(I1) WRITE(6.203) READ(6.204)CURNT FORMAT(‘SCURRENT.UA. = ’) FORMAT(F10.2) VO=CURNT*2.*10**(ICHAN-1) VO=-(VO-5000.0)/2.4414 2.441MV IS ONE BIT SIZE ELECTRODE ’) COMPLEMENTARY VALUES OF VOLTAGE SETTING ARE LOADED IVO=VO WRITE(6.205) READ(6.206)VDROP VDROP=VDROP*CURNT/100.0 206 FORMAT(F15.5) STRIPF 104 110 151 108 109 103 114 115 -211- JVOLT(1) IS THE FIRST HUMP POTENTIAL 850.0MV IS ASSUMED JVOLT(1)=(850.0-REF+VDROP)*AMP/2.441+2048.0 CALL LINT(JVOLT(1).ICHAN.IVO) CALL ASSIGN(4.FNAME.40.IERR) CALL DATE(TODAY) WRITE(4.104)TODAY FORMAT(’3 ’20A1) WRITE(4.110) FORMAT(‘3 DATA FROM STRIP.LDA ') TIME=FLOAT(ICHAN)*65.536+FLOAT(IVO)/1000.0 CHARGE=CURNT*TIME WRITE(4.151)CURNT FORMAT(‘iCURRENT USED UA..'F15.0) WRITE(4.108)CHARGE FORMAT(’CHARGE. UCOUL.’F15.5) JVOLT(1)=JVOLT(1)-2048 JJ=JVOLT(1)/(AMP*O.4096) VOLT=FLOAT