c ALYSIS’ B LOSTAT iAN r‘ée U ,. mam ROCHE BASED' mafia-M FOR. ELEC is , MPUTER CG ....yu- . .. . .. w . . . .. . Q . . . o ow n w y A n a I 4, . 4 .‘ .~ u .. . .. ,a .2. ,. V p . .3 . . . g . , \. .. . ‘ III. 1!" 7 r4 (.1 "a z ’\ "\fi ‘5 This is to certify that the thesis entitled COMPUTER-BASED COULOSTATIC SYSTEM FOR ELECTROCHEMICAL ANALYSIS presented by Briah Keith Hahn has been accepted towards fulfillment of the requirements for __P_1L-D_-_ degree in _C_he_mi_§§Ly_ 44% Major professor Date October 9, 1974 4.3.2., 0-7 639 angina av f; 4 Jill“ &_ “'5' UK BINDER-Y INC. ..“:§:.':'.."L':f.‘.2.’.. v :- LT a" i.- ES" T‘fi .- Lt. .. r . u y .p r. e . pl. 5» a» r) r. Bl r- CI .-. . e... ABSTRACT COMPUTER-BASED COULOSTATIC SYSTEM FOR ELECTROCHEMICAL ANALYSIS By Brian Keith Hahn A computer based coulostatic system has been developed and it has been used to apply the coulostatic method to electrochemical measurements at a dropping mercury electrode. The coulostatic system consists of an electrochemical reaction cell, a digitally controlled pulse generator for adding charge to the working electrode, a voltage measurement system to monitor the working electrode potential, a drop fall detector to synchronize the system with the dropping mercury electrode, and a small digital computer to control the operation of the pulse generator and voltage measurements system. Since many ap- plications of the coulostatic method require the addition of pre- cisely known amounts of charge to the electrode, the pulse generator output was characterized for a wide range of pulse sizes. The ac- curacy of the pulse charge content equations developed to describe the pulse generator output was limited primarily by the uncertainty in the measurements used during the calibration procedure. Using a computer instead of hardwired control logic allowed the coulostatic system to be operated in many coulostatic polariza- tion modes without modification of the instrument hardware. This eitra VE'E to a;:iy t' enerfre". or cartrc7 T’e : analysis a re'1al 2.: e1e:t*:: étd 1‘51 1 C'lte {fig SUTEC er: Brian Keith Hahn extra versatility provided by the computer control enables the system to apply the coulostatic method to a wide range of electrochemical experiments when the charge, current, or potential must be measured or controlled. The theoretical potential-time relationship for coulostatic analysis at a dropping mercury electrode was developed. The experi- mental evaluation of this theory showed that corrections for the electrode double layer capacitance are necessary for maximum accuracy and that the plane electrode approximation does not accurately des- cribe the decay rate, even if the electrode growth baseline is mea- sured and removed from the measured decay curve. The coulostatic system was used to perform and evaluate four types of scanning coulostatic analysis: scanning coulostatic analysis (SCA), derivative SCA, anodic stripping SCA, and derivative anodic stripping SCA. The two anodic stripping methods have not been previously reported. The sensitivity of the two derivative methods was limited to 0.4 MM cadmium by measurement noise, while the sensitivity of SCA and anodic stripping SCA was found to be 0.01 ufl_cadmium. COMPUTER-BASED COULOSTATIC SYSTEM FOR ELECTROCHEMICAL ANALYSIS By Brian Keith Hahn A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirement for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1974 Prc‘e; ;t€dan:e. el dark at 9’." interactirr re~zers c‘ I the' .r; P] S. an. EU! ACKNOWLEDGMENT Professor C.G. Enke has my deepest gratitude for his help, guidance, encouragement and friendship during the course of my work at Michigan State University. I thank my fellow group members~in Professor Enke's research group for their friendship, interaction, and assistance. I especially thank Keith Caserta, Tim Kelly, Tim Nieman, and Dr. T.V. Atkinson. I gratefully acknowledge the many interesting interactions I have had with Professor S.R. Crouch and with the members of his research group. I thank my parents, grandparents and sister for their interest in my studies and their unfailing encouragement and moral support. I gratefully acknowledge the support provided by the National Science Foundation Graduate Fellowships and by the American Chemical Society, Division of Analytical Chemistry Fellowship sponsored by Carle Instruments. ii LIST Cl? 7;. LIST OF F' Ln (1 n: J t) .__4.—. TABLE OF CONTENTS Page LIST OF TABLES ...................... vi LIST OF FIGURES ...................... vii CHAPTER I - Introduction ................. 1 CHAPTER 2 - The Coulostatic System ............ 4 Introduction ..................... 4 The Cell ....................... 6 Pulse Generator ................... 8 A/D Converter System ................. 16 Drop Fall Detector .................. l9 Computer ....................... 25 CHAPTER 3 - Pulse Generator Characterization with Compensation for Non-Ideal Pulse Shapes. . . . 30 Introduction ..................... 30 Constraints on the Pulse Size ............ 32 Calibrating the Pulse Generator ........... 36 Calibration for 1.0 mfl Potassium Chloride Solution (Positive Pulses) .............. 37 Digitization Noise in the Charge Measurement ..... 47 The Generalized Pulse Content Description ...... 48 Precision of the Pulse Generator ........... 52 CHAPTER 4 - Emulation of Basic Polarization Techniques with the Coulostatic System .......... 56 Controlled Charge Experiments ............ 56 Controlled Current Experiments ............ 57 Controlled Potential Experiments ........... 58 iii Chapter Page CHAPTER 5 - Theory of Coulostatic Analysis of a Dropping Mercury Electrode .......... 6l Introduction ..................... 6l Potential-Time Relationship for a Drapping Mercury Electrode .................. 62 Initial Conditions .............. g 62 Case I: Reaction diffusion limited before charge addition ................. 65 Case II: No reaction prior to charge addition .................... 65 Evaluation of the Potential-Time Relationship . . . . 68 Applying the potential-time relationships of real decay curves .............. 68 Decay Time Approximations ............ 72 CHAPTER 6 - Scanning Coulostatic Analysis Methods ..... 77 Principles ...................... 77 Scanning Coulostatic Analysis (SCA) ....... 77 Derivative SCA ................. 78 Anodic Stripping SCA .............. 80 Derivative Anodic Stripping SCA ......... 8T Capacitance Correction ............. 81 Software ....................... 84 ICOUL and DCOUL ................. 85 EXPT ...................... 9l PLOT ...................... 93 Experimental Results ................. 95 Effect of the decay measurement parameters . . . 96 Baseline correction ............... 104 iv Chapter Page Supporting electrolyte concentration dependence ................... 107 Useful concentration range ........... 116 APPENDIX A - Real Time Clock ............... 119 APPENDIX B - Program Listings ............... 128 Program DEFS ..................... 129 Subroutine XYSYS ................... 130 Subroutine AXIS ................... 138 Program CALIB .................... 141 Program FIT ..................... 145 Program ICOUL .................... 149 Program DCOUL .................... 155 Overlay PARAM .................... 160 Subroutine EXPT ................... 162 Subroutine PLOT ................... 166 Subroutine SMOTH ................... 173 REFERENCES ........................ 179 LIST OF TABLES Table Page 1 Limits on the Pulse Voltage .............. 32 2 Interelectrode Resistances for Potassium Chloride Solutions .................. 34 3 Experiment Parameters for the Calibration Curves. . . . 110 4 Clock I/O Instructions ................. 121 vi Figure \lO‘U‘I-DwN G) 10 11 12 13 14 15 16 17 18 19 LIST OF FIGURES Block Diagram of the Coulostatic System Electrochemical Cell (a), and Dummy Cell (b) Pulse Generator Pulse Length Clock Bipolar High Voltage Power Supply A/D Converter System Optical Arrangement of the Dr0p Fall Detector Dr0p Fall Detector Monitor Circuit and Waveforms DrOp Fall Detector Flag Circuit Real Time Clock Block Diagram Pulse Output Voltage as a Function of Pulse Height and Cell Resistance (or Concentration) Calibration Region for the 1.0 mM Potassium Chloride Solutions Initial Dialog with CALIB Error in the Theoretical Pulse Charge Content Equation Printout from the Least Squares Fit of the Calibration Data Residual Error Plot from the Least Squares Fit of the Calibration Data Effect of Digitization Noise of the Calibration Data Systematic Error Previously Hidden by the Digitization Noise Precision of the Pulse Generator Output vii Page 12 15 17 20 22 24 27 35 38 40 43 45 46 49 4O 54 Figure 20 21 22 23 24 25 26 27 28 29 3O 31 32 33 34 35 36 37 38 39 40 Decay Rate as a Function of Measurement Time Differential Double Layer Capacitance as a Function of the Electrode Potential Flow chart for ICOUL and DCOUL Initial Dialog with DCOUL Measurement Time Interval Study for SCA (5.0 uM_Cadmium) Measurement Time Interval Study for SCA (0.6 uM_Cadmium) Measurement Time Interval Study for Anodic Stripping SCA Measurement Time Interval Study for Derivative SCA Potential Jump Study for Derivative SCA Measurement Time Interval Study for Derivative Anodic Stripping SCA Potential Jump Study for Derivative Anodic Stripping SCA Baseline Correction of the Decay Rate Cadmium Calibration Curves for SCA Cadmium Calibration Curves for Derivative SCA Cadmium Calibration Curves for Anodic Stripping SCA Cadmium Calibration Curves for Derivative Anodic Stripping SCA SCA Decay Rate Curve from 0.020 pM_Cadmium 612n Decode Card Logic 613n Decode Card Logic Buffer/Counter/Driver Card Logic Control Register Card Logic viii Page 75 83 86 88 98 99 100 102 103 105 106 108 111 112 113 114 117 124 125 126 127 CHAPTER 1 Introduction The basic principle of the coulostatic method is the very rapid addition of charge to the working electrode of an electro- chemical cell and the subsequent measurement of the electrode potential versus time with no net current flow through the cell. Delahay has developed the coulostatic method into an electro- analytical method for trace analysis in the 10'5 to 10.7 molar range (1-6). The principle for a coulostatic analysis is as follows: the potential of the working electrode is held at the foot of the polarographic wave for the oxidation or reduction of the substance being analyzed. The charge is added to the electrode under coulostatic conditions to bring the electrode potential into the plateau region of the polarographic wave where the reaction rate is diffusion limited. The potential-time variations are measured after the charge addition. Delahay has developed theoretical equations to describe the potential-time variations for a variety of stationary electrodes. For diffusion limited reactions at a plane electrode, the potential variations are linear in the square root of the time after the charge addition, and the slope of the decay is proportional to the concentration of the reacting substance. Thus, the concentration can be determined from the slope of a potential versus root-time plot of the decay measurements. The double layer charging current problem which limits the usefulness of most polarographic methods in trace analysis, is 1 2 avoided entirely in coulostatic analysis. In addition, the voltage measurements are made when no net current passes through the cell, which eliminates the effects of the solution resistance during the measurements. This makes the method very attractive for analyses where the supporting electrolyte concentration is very low. Despite these advantages, very few applications of this method have been reported. This can be attributed to the method's lack of selectivity, the need to replot the data in root-time for an accurate analysis, or the relative inconvenience of a stationary electrode compared to a dropping mercury electrode. Recently, several scanning coulostatic analysis methods that greatly improve the selectivity of coulostatic analysis have been reported (7-12). These methods used a dropping mercury electrode and adjusted the experiment parameters so the decay curves measured from successive drops started from a different initial potential, which gives a decay rate versus potential curve. This curve was used much like a conventional polarogram to determine the species present and their concentrations. A small digital computer was used in one scanning coulostatic method to acquire the data and calculate the decay rates, thus simplifying the measurements and improving their accuracy. The scanning coulostatic analysis measurements can be further simplified by using the computer not only to acquire the data and perform the calculations, but also to control the entire experiment sequence. The development and characterization of such a computer- controlled coulostatic system is the subject of this thesis. The 3 computer control added an extra dimension of versatility to the system, permitting it to perform a wide range of electrochemical experiments where the charge, current, or potential must be measured or controlled. The principles of these measurement and control procedures are discussed in subsequent chapters and several of them are applied to perform coulostatic experiments at a dropping mercury electrode. Since the computer is used to calculate the coulostatic decay rates, an accurate description of the potential-time relation- ship for coulostatic analysis of the dropping mercury electrode was also developed and evaluated. The system was used to perform and evaluate four types of scanning coulostatic analyses, two of which had not been previously reported. CHAPTER 2 The Coulostatic System Introduction The coulostatic system (Figure 1) consists primarily of an electrochemical reaction cell with working, reference, and counter electrodes, a pulse generator for adding charge to the cell, a voltage measurement system to monitor the working electrode potential, a drop fall detector to synchronize the system with a dropping mercury elec- trode, and a computer to control the operation of the pulse generator and potential measuring system. Using a computer instead of hard- wired control logic allows the coulostatic system to be Operated in many diverse modes without modifications to the instrument hardware. It also provides real time mathematical analysis of the raw data, reducing it to the significant results, and it finally provides means of outputting these results to the experimenter. The various parts of the coulostatic system are discussed in detail in the following sections. The computer programs needed by the system to perform various functions are discussed in the appropriate chapters on those functions. The interfacing between the computer, a Digital Equipment Corpora- tion PDP 8/1, and the instrument was done via the Heath EU-80lE Com- puter Interface ADD system. This system is described in detail in the Heath manual (13), and in experiments for the Scientific Instru- mentation course offered by the Department of Chemistry of Michigan State University (14). As far as this discussion is concerned, the PDP 8/I input/output (1/0) bus lines can be divided into four basic groups: the buffered accumulator output lines (BAC) which .Emumxm ovumgmo_:ou mg“ mo Eoemmvo xqum ._ mczmwm . moem 41 o\< let moeomemo A“, 444m demo carry data from the computer to the peripheral devices; the accumula- tor input lines (AC in) which are used to transfer data from the peripheral device to the computer; the buffered memory buffer lines (8MB), the middle six bits of which carry the address of the peripheral involved in the I/O action; and a group of timing and control lines. The 8M8 lines are decoded at the peripherals to form a device select (05) signal which becomes a logical 1 whenever that device is addressed. The useful timing and control lines include: IOP1, IOP2, and IOP4, 3 timing pulses which are used to synchronize the transfer of infor- mation between the computer and peripheral during an I/O instruction; the skip line (SKP) which is active during an I/O instruction and can be used to force the computer to skip the next instruction; and the clear accumulator line (CIA) which can be used to clear the accumulator during an I/O instruction before the new data is read into the computer. These various signals are buffered by Heath Computer Interface Buffer and brought to the instrument by the I/O patch cards. Since the mechanics of interfacing using this system are very well described elsewhere (13,14) many of the details of the interconnections between the coulostatic instrument and the computer are shown by block diagram or omitted entirely. All of the logic was implemented using standard transistor-transistor logic (TTL) integrated circuits. A complete list of all the I/O commands used by the system is given in the program DEFS found in Appendix B. The Cell The cell used throughout this study was a standard three-electrode cell (Figure 2a) with a dropping mercury electrode (DME) for the A l DME Pt GAUZE ‘1 \ B Cd 1 1, WORKING ‘ R R ELECTRODE 5 ref REFERENCE ELECTRODE Rf COUNTER ELECTRODE Figure 2. Electrochemical Cell (a), and Dummy Cell (b). working electrode, a cylinder of platinum gauze (area 8 cm2) for the counter electrode, and a saturated calomel electrode (S.C.E.) for the reference electrode. The DME used a 2-5 second capillary and the height of the mercury was adjusted to give nominal drop times of 5.9 and 9.5 seconds (mercury flow rates of 1.62 and .99 mg/sec respectively). The capillary was modified for use with the drop fall detector as is described in the discussion of the detector. The cell also included a fritted glass bubbler through which nitrogen was passed to purge the oxygen from the solution. During an experi- ment, the bubbler could be raised above the solution so the nitrogen continued to flow over the top Of the solution. The nitrogen was passed over BTS deoxygenater from BASF at 200 °C and bubbled through distilled water before it entered the cell. The geometry of the cell was important only in calibrating the pulse generator output (Chapter 3). The dummy cell used in the calibration had to correspond as closely as possible to the cell for proper calibration of the Charge content of the pulse generator output pulse. The dummy cell that corresponds to the real cell is shown in Figure 2b. The solution resistances between the counter electrode and the working and reference electrodes are represented by RS and Rref respectively, while Cd and Rf are the double layer capacitance and the faradaic reaction resistance associated with the working electrode. Pulse Generator The pulse generator (Figure 3) is a digitally controlled constant current source. This current source has four basic elements: 1) a .Lopocmcme amps; .m assure AOAHV Auzmv N no .hhmm .me gooou zeezmu amuse Aezwumzv Nepo H E ' L.> > ompw <5 om“ m~\mmom zoh<4 - fiL mh4o> mu 0 mhmmlmp z<¢a mmc¢m>zoo <\o do om, 5P ‘- uuuu mm=UHH3m Ema egm Switching= SOJTCE or I tne cort' generatar araiifier tnr3ugn : SOdrce, pulse 19' SuffTCIEr' The < Vel‘ter d-C’ to 4 Earn A :5 vol: V1505! re the QEOEr The lZ-bi 10 digitally controlled current source; 2) two digitally controlled i switching elements (the PRAM and FET switches) to turn the current source on and off; 3) a digital pulse duration clock which provides the control signal to the switching elements, causing the current generator to output pulses of known length; and 4) an output buffer amplifier to provide the necessary voltage to drive the current pulses through the cell and to minimize loading effects on the current source. With digital control of both the pulse amplitude and the pulse length (and thus the charge content) the pulse generator is sufficiently versatile to be used in many diverse applications. The current source, consists of a digital-to-analog (D/A) con- verter and a resistor to convert the output voltage of the converter to a current. The converter is a 12-bit Analogic DACPAC MP1812 with 3.15 volt output; the resistors are fast response metal film 0.01% Vishay resistors. The bipolar output of the D/A converter allows the generation Of currents, and thus pulses, of either polarity. The 12-bit resolution (1 part in four thousand) of the converter provides accurate control of the voltage (14.3 mV). The output of the D/A converter is buffered by one of the switching elements (the PRAM) to provide currents of up to :20 mA. Either one of the two resistors can be chosen by the range control and FET switches to perform the voltage-to-current conversion. The 2500 resistor gives a :20 mA output range. For short pulses with high charge content, while the 5K9 resistor with its 14.0 mA range, is used to produce high resolution pulses of low Charge content. The switching elements serve several functions, the primary one being to turn the current source on and Off to produce a current 11 pulse. The first switch element is a Harris HA-2405 four channel programmable amplifier (PRAM). The PRAM is an operational amplifier with four identical input stages, any one of which may be connected to the output stage by an on-chip digital control network. Only two of the inputs are used; one is connected to the output of the D/A converter and the other to ground. The PRAM is wired as a voltage follower, so that the output voltage is the same as the selected input voltage. In addition to switching the D/A converter voltage on and off (by switching to the grounded input) the PRAM also buffers the D/A converter output to a :20 mA maximum output current. A further level of switching is provided by the FET switches, which turn the current output from the resistors on and off. These switches are Siliconix DG 152AP dual SPST FET switches with driver Circuits. These switches are also used to control which current range (resistor) is used. Thus, the output current pulse is formed by simultaneously switching the PRAM to the D/A converter input and the FET switch for the appropriate current range on. The pulse is terminated by turning the switch off and switching the PRAM to the grounded input. The output current during the off state is kept as low as possible by switching the voltage source to zero so that the voltage across the FET switches is no greater than a few millivolts. The pulse length is controlled by the pulse length clock, which supplies the logic signal to turn the two switches on for a precise length of time. The pulse length clock, shown with partial waveforms in Figure 4, is a specially modified variable time clock (15). The presettable counter is used as a variable modulus counter to multiply the oscillator's period by the counters modulus, m. When the counter 12 .xoo_o gamed; omP=a d1: 3 .omo :tL: mmezaoo . m4m cop: Lopoawm Excel :3 ? +71 Ike L>oo Noonzu +3 H 80. +1F .m «cameo 15 .xraazm Luzon mumppo> cop: Lopoapm .m mcam_u Noanzm x00. .6 >09 A >02. x o. xv.“ + x. .vs utnov L\ov ( gt. Ex on. m x 3. ..o >oo so. . .9m 33.: 4 o: + Q . A ll Llovm +71 28 av +7 V ELJ $71 16 base circuits of these transistors absorb most of the voltage drop needed to regulate the output, while the amplifier driver transistors provide fine control of the voltage. Two crowbar circuits auto- matically turn the supply off if the output voltage exceeds 1120 volts. They also prevent the supply from operating if the 115 volt supply for the operational amplifiers is not present from the ADD modules. A/D Converter System The analog-to-digital (A/D) converter system measures the potential between the working and reference electrodes, amplifies it, then digitizes it and transmits the result to the computer (Figure 6). High speed data acquisition systems which as this are a standard component in many computerized instruments and consist of an input buffer, amplifier, a sample and hold to hold the voltage during the digitization, and an A/D converter. In this case however, the high voltage pulses (up to 1100 volts) available from the pulse generator necessitate a few additions to the standard approach if the system is expected to survive and operate at maximum speed. Protection against the high input voltages is easily accomplished by placing 10 volt zener diodes between the input and ground. This limits the input voltage to the amplifier to 110.6 volts, which is within the safe operating range for most Operational amplifiers. The first amplifier stage uses the voltage follower with gain con- figuration to amplify the signal while retaining the highest possible input impedance for the reference electrode. The gain determining resistor, RG’ is mounted in connectors atop the card, so that the gain can easily be changed . when needed. The overall gain Of the 17 .Emumam eupem>coo o\< .m mcamwu 2H u< 33% EM 1 - m; o._ - 932% I f m: , A4 l..- :25 .m: 0.. EM thm-~P _ m=pzoo o\< . > 2 > 2 8 So: + , Y + 3 main 1 m ax op 18 amplifier is (RG + 10KO)/lOKQ; a typical value for RG is SOKO giving an amplification factor of 6.00. This amplification is necessary if the full resolution Of the A/D converter is to be utilized. However, the amplification causes another problem during and immediately after a pulse: the ten volt input will drive the amplifier output to its limit. With many Operational amplifiers, recovery from such an overload takes milliseconds - far too long when data must be acquired on a microsecond time scale. This problem has been minimized by using operational amplifiers with a very short overload recovery time and by limiting the input to the sample and hold to 110.6 volts with the zener diodes and a resistor at the output of the amplifier. The second amplifier buffers the low impedance input of the sample and hold, so the output resistor will not affect measured voltage. The amplifiers are Analog Devices model 430, chosen for their FET- input, high common mode rejection ratio and bandwidth, and short overload recovery time. The sample and hold, an Intronics F3201, tracks the voltage continuously holding it only while the A/D converter digitizes it, eliminating the time normally required for the sample and hold to acquire a new voltage level before each conversion. The A/D con- verter is an Analog Devices ADC-12U 12-bit successive approximation analog-to-digital converter with an input range of 0.0 to 10 volts and a specified conversion time of 10 microseconds, although the unit used in these experiments finishes in 7.5 microseconds. The conversion sequence starts upon the receipt of the appropriate I/O command. The leading edge of this command triggers a monostable, producing a 1.0 usecond start pulse. The leading edge Of this pulse 19 Clears the converter and sets the converter's status line. This status line is used to switch the sample and hold to the hold mode. The trailing edge of the start pulse starts the A/D conversion process. When the conversion is completed, the A/D converter status drops to zero, returning the sample and hold to the tracking mode. This status signal can be interogated by the computer through an I/O skip instruction. Drop Fall Detector The drop fall detector synchronizes the start of an experiment with the beginning of the electrode's life. The literature methods for detecting the drop fall were unsatisfactory because they either required electrical contact with the electrode, or the detector alignment was very critical. As a result, a new drop fall detector was developed for use in this system (17). The drop fall detector is a reflecting type optical detector which has the advantage of having no electrical interaction with the cell; yet it requires no critical Optical alignment. The drop fall detector monitors the amount of light reflected by the drOp as it grows. When the drop detaches from the electrode, the amount of light reflected by the drop falls to zero. This sudden change in light level is used to produce the drop fall signal. The detector has two basic parts: an Optical system to monitor the reflected light from the drop and an electronic monitor to measure the light and produce the drop fall signal. The optical system is shown in Figure 7. The drop is illuminated from the side by a prefocused "penlight“ lamp mounted outside Of 20 OPAQUE , MOUNTING COLLAR , PHOTO I -FET I DETECTOR : 1 ______ .I ' I 1 FIBER OPTIC , 1 CLEAR / I n—CAPILLARY EPOXY ' CEMENT : : CELL TOP I F ’ :3 ‘ll . I .. .._11 I I I PENLIGHT : BULB To A i .. we DC 2% MERCURY POWER DROP SUPPLY OPAQUE «CELL TAPE” Figure 7. Optical Arrangement of the DrOp Fall Detector. 21 the cell. The lamp's position is adjusted to provide a maximum amount of light at the tip Of the capillary. A piece of tape was placed on the side of the cell to reduce stray reflections from the pool Of mercury that eventually accumulates at the bottom of the cell and the cell was covered with a loose shroud of black felt to reduce interference from room lighting. Part of the light which is reflected up the capillary by the mercury drop is diverted out Of the capillary by a fiber Optic light guide. For maximum light pick-up, the fiber optics are mounted in a notch ground into the side of the capillary above the top Of the cell. The protective outer tubing was removed from the end Of the pipe to permit greater flexibility in aligning the fibers and the end of the fiber bundle was polished to reduce reflections and stray light pick up. Optical contact was further improved by using a clear epoxy cement to secure the fiber bundle in the notch. The fiber Optic light guide was a 64-fiber 10 mil plastic fiber Optic light guide (available from Edmund Scientific CO., Barrington, NJ 08007). The other end of the fiber Optic guide was polished and firmly mounted in front Of the photo-FET in the monitor module. The drop fall monitor (Figure 8) measures the light from the fiber Optic input and takes the first derivative Of the light level. The photo-FET light detector, a Teledyne Crystalonics FFlO8, is powered by two zener diode voltage supplies to reduce power supply fluctuation. The bias voltage applied to the gate of the FET (through the l megohm resistor) is adjusted with the bias potentiometer to give maximum detector sensitivity. In the monitor used here, the Optimum sensitivity occurred when the Output Of the photo-FET was 22 .msgomm>m3 sup; ppaue_u couvcoz coguwuma Fpom coco 39:0 .363 ” 53> LoamLOQEOO_ BEBE/x II _ ma: _ III >np- vm—z _ I.. comm OE. _ M342. _ I 4 882“ BE. - _ .r .7 e< _ FQO" xnd I sn+. xnn .m weaned n 2053:9330 825:0“. mmmxo> 88300 23 about 3.5 volts. The voltage follower is used to buffer the output Of the photo-FET so that the differentiation stage will not affect the Operation Of the FET. The buffered output is then differentiated by an Operational amplifier differentiator (A2) and amplified by the inverting amplifier (A3). This amplifies the signal change due to drop fall to a level where it can be easily and reliably detected by a comparator. The comparator compares the amplified output of the differentiator with the voltage from the trigger level poten- tiometer. The trigger level is adjusted to a level just below the peak voltage produced at drop fall. When the drOp fall occurs, the input voltage to the comparator exceeds the reference voltage and the output Of the comparator changes polarity, providing the output pulse. Hysteresis is introduced into the comparator through the resistor feedback network to reduce the possibility of multiple comparator triggering from the noise on the input signal. The output transistor converts the comparator output to a logic level signal that is used to set the drop fall flag. The entire circuit for the monitor was fit inside a small grounded aluminum box and mounted near the electrochemical cell. Adjustment of the drop fall detector was facilitated by displaying the output Of the detector and its flag on two lights of the Binary Information Module (BIM) that is part Of the Computer Interfacing ADD. In addition, one Of the momentary contact switches was used for manually clearing the flag as shown in Figure 9. The detector was aligned by first aiming the lamp so the light intensity is highest at the tip Of the capillary. A piece of white paper placed on the opposite side Of the cell tO act as a screen for the shadow 24 .pmzuewu mofim eouumpmo .Pmm coca .m meaawm mamou S<=znv _m .Eocmopa xooFm xuopo we.» Poem .oF mesmwm ,g.nu;‘ mmZxo A"_O_.m_2m IU._.