DSVELOPiMENT OF PORTABLE ELECTROCHEMICAL TRANSDUCERS FOR THE DETECTION OF SULFUR DIOXIDE AND OXIDES OF NITROGEN CONTRACT NUMBER CPA 22-69-118 FINAL REPORT PERIOD COVERED: JUNE 1969 TO AUGUST 1970 AUTHORS RAMESH CHAND ROLAND V. MARCOTE SEPTEMBER, 1970 PREPARED FOR DEPARTMENT OF HEALTH, EDUCATION, AND WELFARE DURHAM, NORTH CAROLINA DYNASCIEKCES CORPORATION INSTRUMENT SYSTEMS DIVISION ------- NSTRii^EHT SYSTEM CORPORATION TABLE OF CONTENTS Section Page 1.0 INTRODUCTION . . . . 1 2.0 GENERAL DISCUSSION . . 2 2.1 Sweep Voltaimietry 2 2.1.1 Instrumentation. . . . '. 2 2.1.2 Experimental Procedure 5 2.2 The Electrochemical Transducer. 5 3.0 EXPERIMENTAL 11 3.1 Sv/eep Voltammctry Studies 11 . 3.2 Experimental S02 and NOX Transducers 17 3.2.1 Evaluation of Semipermeable Membranes 17 3.2.2 Transducers Employing Aqueous Sulfuric Acid Electrolyte. 19 3.2.3 Transducers Employing KPF6 in Propylene Carbonate as Electrolyte 20 3.2.4 Transducers Employing Dryed KPFg and Cl^COOH in Propylene Carbonate as Electrolyte. . 22 3.2.5 Transducers Employing Other Electrolytes 23 3.3 Selected NOX and S02 Transducers 25 3.4 Sulfur Dioxide Scrubbing Systems 26 3.4.1 Electro-oxidation Systems 26 3.4.2 Wet Scrubbing Systems 27 4.0 PROTOTYPE MONITORS PERFORMANCE DATA 30 4.1 Sensitivity 30 4.1.1 Sensitivity and Operational Temperature 30 4.1.2 Sensitivity Ranges 30 4.2 Selectivity 30 4.3 Response and Recovery Times 34 4.4 Stability of NOX and S02 Prototype Monitors 34 4.5 Transducers Life Data 38 4.6 Temperature Control and Compensation 38 ------- RPORATIOIM i LIST OF FIGURES - \ * - ' ' - s Figures . Page j .••''' ^ 1 Oxidation of N02, NO, and S02> in l.ON HUSO, . | at a Gold Sensing Electrode 2 • j 2 Cyclic Voltammetric Instrumentation 4 i 3 Sweep Voltammetric Cell 6 ! 4 Electrochemical Transducer (Schematic) 8 , i 5 Potenti.ostatic Control of an Electrochemical Transducer . . 9 6 Linear Response of an NOX Prototype Monitor 31 : 7 Characteristic Response of Monitor SS-330 35 ------- ORPQRATIQN LIST OF TABLES Table" • Page 1 Sweep Voltaimnetry Evaluations 12 2 Sweep Voltammetry Evaluations .... 13 3 Sweep Voltammetric Evaluations of Potential Interferences 15 4 Sweep Voltammetric Evaluations of Potential Interferences 16 5 Selective Membranes 18 6 Evaluation of NOV Transducers Employing KPF.--PC .... 21 x o 7 Evaluation of Various Organic Electrolytes.. 24 8 The Effect of Temperature Upon Transducers Sensitivity 32 9 Evaluation of Various Interferences . . 33 10 Stability of NO and S00 Prototype Monitors 37 X £* ------- ORPO RATION t> . "l.O INTRODUCTION o ' ' This is the Final Report on the Dept. of Health, Education and Welfare Contract CPA-22-69-118, covering the period from June 1969 to August 1970. The objective of this effort was to develop and provide portable electro- chemical transducers for the detection of sulfur dioxide and oxides of nitrogen in the industrial stack. The v?ork carried out involved the screening of different aqueous as well as nonaqueous electrolytes for the purpose of developing selective sulfur dioxide and oxides of nitrogen transducers. Though many organic electrolytes were found to be feasible based on the sweep voltammetry studies, they were all abandoned from consideration in the final design because the transducers employing these' electrolytes gave very limited useful life, slow response time and erratic behavior. A dilute solution of sulfuric acid (IN H2SO^) was selected as the electrolyte for both sulfur dioxide and oxides of nitrogen transducers. The selectivity of the sulfur dioxide transducer was obtained by subjecting its sensing electrode to a fixed potential and by the use of polyethylene as the gas permeation membrane. The total interference from 1000 ppm NO and 100 ppm N0£ was found to be about 5% full scale on a 0-1000 ppm S02 transducer. The selectivity of the oxides of nitrogen transducer could not be obtained in the presence of sulfur dioxide and a scrubber was designed to remove S0? without appreciable affecting NOV concentration. The scrubber consisting of X aqueous solution of I^CO-j, KHCO^j KNOo and KNO~ was found to absorb less than 1% of 2000 ppm of NO and less than 20% of 100 ppm N02- However, considering the low concentration of NOo in the stack, the overall effect on the NO concentration is insignificant. ------- SYSTEMS Bi a2.0 .-GENERAL DISCUSSION ..___. •'.."' ' • • 2.1 JWEEP VQLTAMMETRY " . " • The oxidation and reduction potentials of different electroactive gases _ in different aqueous and nonaqueous electrolytes were determined using- sweep voltammetry . This technique, which permits the scanning of an entire voltage range in a matter of seconds, not only shows the potentials at which reactions occur, but also their rates and efficiencies in terms of current generated. Moreover, information regarding electrolyte stability and electrode kinetics also result from sweep voltammetric studies. The existence of an electrode reaction is indicated by a peak in the voltage- current curve traced out by an x-y recorder. The potential at which the . peak occurs is related to the specific electrode reaction, whereas the psak height (current) is proportional to the concentration and diffusion coefficient of the reacting species. The efficiency of the electrode reaction, in terms of electrochemical reversibility, is also reflected in the peak height and the shift of peak potential. The purpose of this study was to find an electrolyte which showed selective oxidation peaks for either SO, or NO , or £* ft. which showed their respective oxidation peaks at substantially different values. It was understood at this stage that both NOX and S02 being highly F;' electroactive are likely to offer the most interference with each other. As K an example, Figure 1 shows the voltammograms for S02, NO and N02 obtained in \ IN H2SO/,. electrolyte. • " 2.1.1 INSTRUMENTATION The instrumentation employed for the cyclic voltammetric measurements is schematically shown in Figure 2. Basic to the method is a potentiostat which maintains the voltage between the reference-working electrode couple at a value equal to the sum of the input reference voltages represented by the dc reference and function generator values. The potentiostat accomplishes this by controlling the current between the working and counter- electrodes. The potentiostat senses the difference between the reference- working electrode voltages and the sum of the reference input voltages by. means of a high gain amplifier system, and feeds the output through the working electrode - counterelectrode couple. This amplifier has high input and low output impadance, and a rise-time less than 10 l sec. The system permits the measurement of current-voltage-time relationships under 2 ------- POTENTIOSTAT [REFERENCE VOLTAGE REFERENCE ELECTRODE WORKING ELECTROD FUNCTION GENERATOR f POWER ' ' AMPLIFIER •PC-Y RECORDER COUNTERELECTRODE .CELL' ------- well-defined conditions. No voltage change is possible at the reference .electrode, and the measurements reflect only the voltage changes at the working electrode-electrolyte interface independent.of ohmic losses through the electrolyte and polarization at the counterelectrode. The potentiostat was.designed and built at this laboratory. Employing solid state operational amplifiers and a Harrison type 6824A current amplifier, the instrument is capable of an output of ±50 volts at 1.2 amp. The function generator is a Hewlett Packard Model 3300, with a Hewlett Packard Model 3302A Trigger, phase lock plug-in unit. The two-chamber sweep voltammetry cell (Figure 3) employs three electrodes. j These electrodes are the working, counter and reference electrodes. The » two chambers are connected to each other by an electrolyte bridge. The J working or sensing electrode was a 1.27cm x 0.127cm piece of gold mesh 99.9% .> pure. The counterelectrode was a platinum (99.9%) mesh, and the reference jj electrode was cadmium/cadmium fluoride (Cd/CdF2) obtained by direct fluorination i> of 99.9% pure cadmium wire. ji 2.1.2 EXPERIMENTAL PROCEDURE . |( The sweep voltammetry cell was cleaned and dried. The gold electrode was . jj cleaned by immersion in hot 50% aqueous HN03 followed by water rinse, immersion jj in hot 20% aqueous KOH, water rinse, another brief immersion in the 50% HNO_3 j and rinsed with distilled water. This electrode was further rinsed with the electrolyte. The Cd/CdF7 reference electrode was cleaned by brief immersion in diluted HN03 followed by rinses with distilled water and the electrolyte. The platinum counterelectrode was anealed at a flame before using it. Individual gases were bubbled through the electrolyte for three minutes, and sweep curves were recorded on the third sweep cycle. The scanning range was -0.5 volts to -KL.25 volts at a sweeping rate of 40 mv/second. 2.2 The electrochemical transducer is a sealed faradaic device in which the electro-oxidation or electro-reduction of absorbed gas molecules at a sensing electrode results in a current directly proportional to the partial pressure of the gas. ------- WORKING ELECTRODE LEAD REFERENCE ELECTRODE \ GAS INLET TUBE (TEFLON) <™»GAS INLET PORT . -COUNTER - ELECTRODE LEAD -WORKING ELECTRODE PLATINUM GAUZE COUNTER- ELECTRODE ------- BJlillll CORPOfiATION INSTRUMENT SYSTEMS DiViSiO;] ' Figure 4 shows a schematic drawing of an electrochemical transducer. The pollutant molecules diffuse through the semi-permeable membrane, dissolve in the film of electrolyte, and diffuse to the surface of the sensing electrode where they undergo electro-oxidation or electro-reduction, depending upon the nature of the gas molecules and the transducer operational mode. The counterelectrode material has an oxidation potential higher than that of the pollutant. When the two electrodes are joined together externally through a small resistor, the sensing electrode which is a polarizable electrode assumes the potential of the counterelectrode less the iR drop across the resistor. The pollutant molecules, upon reaching the sensing electrode encounter a potential higher than required for their electro-oxidation. The net result is lowering of the potential of the sensing electrode proportional to the concentration of the pollutant molecules and equivalent release of electrons from the sensing to the counter electrode. Electrochemically, while the pollutant is being oxidized at the sensing electrode, a corresponding reduction reaction occurs at the counterelectrode. Because the current is diffusion limited, the resulting electrical current is directly proportional to the pollutant partial pressure. The magnitude of the concentration gradient at the sensing electrode is directly proportional to the gas concentration in the electrolyte layer (Pick's law of diffusion), and therefore to the partial pressure of the pollutant in the sample. The diffusion current is expressed by: nFADC x e_ Where i is the current in amps, n is the number of exchanged electrons par mole of pollutant, F is the Faraday constant (96500 coulombs), D is the f\ diffusion coefficient of the gas in cm /sec., C is the concentration of the gas dissolved in the electrolyte layer (moles per cm-*) and £ is the thickness of the diffusion layer in cms. The potential exerted upon the gold electrode by the counterelectrode was regulated by use of potentiostatic control. This is schematically shown in Figure 5. A stable D.C. power source, such as a mercury battery across a potentiometer was used to lower or increase the effective potential on the sensing electrode. We have used a lead dioxide counterelectrode for the sensors employing aqueous sulfuric acid electrolytes and a cadmium ------- THIN FILM ELECTROLYTE SENSING ELECTRODE / ''/. BULK ELECTROLYTE 77T7// '777/T7 / / / / / // / f / < / // COUNTERELECTRODE / // TO AMP ------- FIGURE 5 SCHEMATIC SH.Ovr.tHG POTEHTIOSTATIC CONTROL OF ELECTRO- /' ". . . CHEMICAL TRANSDUCER TO OBTAIN S02 SELECTIVITY SENSING ELECTRODE OUNTER- ELECTRODE M.35 VOLTS TRANSDUCER -o -o ------- CORPORATION The lead dioxide electrodes were made by the anodic electroplating on platinum mesh (52 mesh size, Englehard Industries). The plating solution consisted of 350 gms/liter of Pb(NO~)0 and 2 gms/liter of Cu(NOo)9% H.O. 3 f. ~> / 2. The plating was carred out at 15 amp/ft current density and at a solution temperature of 120°F. The cadmium fluoride electrodes were prepared by mixing 85% CdF9 (Research Inorganic, 98%), 5% carbon (Research Inorganic, 99.9997=) and 10% polypropylene poxtfder (Hercules Co.). About 6 gms of this mixture was evenly spread on both sides of a platinum mesh and pressed to 3,000 psig for three minutes at 110°C. ------- CORPORATION ,3.0 EXPERIMENTAL 3.1 In the evaluation of sweep voltanmograms the most useful parameters are the potential and current flow values at which the oxidation and reduction of the gas being studies occurs. • These values are generally referred to as the peak potentials. The peak potential values represent the minimum voltage required to obtain the maximum current f'low. The primary objective of these sweep voltammetry studies was the discovery of electrolytes with selective response for NOX and S0?. With this purpose in mind, a number of electrolytic systems were characterized. The oxidation and reduction potentials of NO, N0?, S0_ and other electroactive species which could interfere with NOX and S02 were obtained. Tables 1 and 2 show the oxidation and reduction,potential of various gaseous pollutants in each of the electrolytes studied. The selective oxidation or reduction of a pollutant in the presence of other electroactive speci.es is obtained if either its oxidation potential is lower, or its reduction potential is higher than those of the interfering gases • (Stockholm Convention). The electro-oxidation of sulfur dioxide in the i: I ' presence of oxygen and oxides of nitrogen clearly exemplifies the selective '• * i ! oxidation of pollutants. The electrochemical half-cell reactions, in ij IN H^SO,, for these gases are shown below. The potential values (E°) S i throughout this report are in volts relative to the normal hydrogen v electrode (NHE) p 298° ' Volts •p & -~00 » S02+2H20 —* S04 44H+42e~ 40.17 N024H 0 —»? NO "+2H+4e~ 40.80 N0421I20 -^» NO "44li+4-3e~ 40.96 , 4 { i } The sweep voltammetry studies indicated the existence of several electrolytes j. potentially suitable for the selective sensing of NO and SO . The most £ x i £ promising systems were for SO- IN 1I_SO,, and for NO undistilled propylene ? carbonate saturated with undried KPF, . *' • l> ? ------- • TABLE 1 SWEEP VOLTAMMETRY EVALUATIONS Electrolyte IN H SO, aqueous IN H_SO, aqueous Saturated with Tri- nitrobenzene 7.8 N H2S04 36 N H2S04 ' Distilled propylene Carbonate -0.75M Dry KPF, Undistilled propylene carbonate -0 .7511 Undried KPF, o Undistilled propylene carbonate -0.75M Undried KPF, +10% H00 D / Peak Potentials, Volts (NKE) OXIDATION NO 0.94 1.20 0.94 1.14 0.96 N0£ 1.10 1.30 0.97' 1.25 1 . 20 j 0.67 1.55 0.66 1.07 0.85 1.35 1.10 1.70 0.70 1.26 1.00 1.50 1.03 so2 0.85 1.16 1.00 NH3 NPO NPO NPO ! 0.75 1.55 0.78 •1.20 1.30 1.09 NPO -0.17 ' NPO NPO REDUCTION NO NPO NPO NPO 0.60 - 0.56 . ' 0.75 NPO N02 NPO NPO NPO 0.90 0.56 0.70 0.83 so2 NPO NPO NPO NPO -0.50 NPO NPO NH3 0.45 0.50 l.OS NPO -0.50 -0.85 0.15 -0.83 ------- SWEEP VOLTAMMETRY EVALUATIONS Undistilled propylene carbonate saturated with undried KPF, Distilled gaircna- butyrolactone -0.75M dry KPFg Distilled gamma- butyrolactone -0.75M undried KPF- 5 Distilled gamma - butyro lac tone -0.75M undried KPFg +10% H.,0 N,N-Dimethyl Formamide -0.75M dry XPFg Distilled Formamide -0.75M dry KPFg Distilled Formamide -0.75M undried KPF, +10% H20 Glycerol - H»0 (1:1) saturated wiuh KPF- 6 Peak Potentials, Volts (NKE) OXIDATION NO 0.86 0.77 1.07 0.98 1.00 0.77 0.97 0.95 1.18 0.87 NPO K02 0.72 0.77 0.94 0.99 1.00 1,25 1.03 0.70 1.04 so2 1.25 0.88 0.30 1.00 0.87 0.55 1.02 0.90 0.90 0.65 NH3 -0.11 NPO NPO NPO NPO NPO NPO NPO REDUCTION NO 0.73 0.41 0.98 NPO NPO NPO NPO 0.62 N°2 0.55 0.75 0.52 0.82 0.85 1.08 NPO NPO NPO so2 -0.44 -0.43 -0.50 NPO -0.30 NPO NPO -0.45 NPO NPO NPO NPO NPO . . NPO NPO NPO. ------- CORPORATION • Sulfur dioxide in IN aqueous H_SO, has an oxidation peak at 0.85 volts. » This potential is lower than those of NO and N0_. Both, NO and N02> have two oxidation peaks at 0.94 and 1.20, and at 1.10 and 1.30 respectively. Therefore, IN aqueous HJ50, can be employed for the selective sensing of S02 (Figure 1). . ' ' In 7.8 N aqueous H_SO, and 36 N H?SO, nitric oxide undergoes electro- oxidation at a lover potential than sulfur dioxide and nitrogen dioxide. These electrolytes could be utilized in an NO monitor, but they are not suitable for use in either NO or S00 monitors. Jt ^ Nitric oxide, in N -H«SO, aqueous electrolyte saturated with trinitrobenzene, was found to have an oxidation potential of 0.94 V. Another oxidation peak was observed at: 1.14V. Nitrogen dioxide showed an oxidation potential of 0.97V and sulfur dioxide of 1.16. This electrolyte is thus of possible use in an NO sensor. A moderate oxidation maximum current flow for 00 was X . £. observed at 1.4V. Three large reduction peaks were observed at 0,0 volts for II , 0 and CO (Tables 3 and 4). The 0.75M KP1;' in undistillecl f ormamide ,with 10% H-O by weight had a potential use in an S0? monitor. Sulfur dioxide in this electrolyte is oxidized at a lower potential (0.85V) than NO and N0?. Formic acid and H«S are oxidized at 0.85V and 0.35V respectively and thus they would interfere. Small reduction maximums were measured at -1.0V for formalde- • hyde, H90, 0., and CO. A moderate current flow was observed for !!„ at £• £• £., -1.0V. Large reduction maximums were measured for C09 and E S at -1.0 volts and for S00 at -0.6V. The 0.75M dryed KPF, in distilled formamide 3 6 also was found to be suitable for the selective sensing of S09 (Table 2). Gamma-butyrolactone (distilled) - 0.75M dryed KPF,. electrolyte appeared to be useful for an N02 monitor. The addition of 10% of HO by weight to this electrolyte causes a reversion of its selectivity. In other words, it becomes a selective electrolyte for S09. The N, N-dimethyl formamide - 0.75M KPF- electrolyte was found to be selective for sulfur dioxide, S09 (0.55V) being oxidized below NO and NO.. A small ^ ^ oxidation peak for formic acid was observed at 1.24V, and two large peak potentials for H?S at -0.21 volts and 1.40 volts were observed. Formic acid reduces at -0.80 volts giving a large current flow. Sulfur trioxide and H«S gave large reduction peaks at -0.937 and -0.3V respectively (Table 3). ------- TABLE 3 SWEEP VOLTAMMETRIC EVALUATION OF POTENTIAL INTERFERENCES Electrolyte Forrnamide (undistilled) -0.75M KPF, +107. (by weight) HO & -Butyrolactone (Distilled) -0.75HXPF, +107, (by -weight) H20 Propylsne Carbonate (undistilled) saturated with KPF, IN K2S04 saturated with Trinitrobenzene Spectro-grade N,N-Dimethyl Forrnamide -0.75M KPF, 6 Peak Potentials, Volts (NEE) OXIDATION YJL~~' 0.35 NPO -0.21 1.40 NPO NPO o? | so3 co NPO NPO 1.19 NPO NPO NPO NPO 1 NPO j NPO NPO ! ! NPO -1.40 NPO . NPO NPO NPO ' ' REDUCTION E2S -1.0 NPO -0.93 H2° -1.0 NPO NPO NPO °2 -1.0 0.0 NPO 0.0 v*PO so3 -0.6 NPO -0.8 -0.3 ,co -1.0 NPO NPO 0.0 NPO ------- TABLE 4 SWEEP VOLTAMMETRIC EVALUATION OF POTENTIAL INTERFERENCES Electrolyte IN H.SO^ saturated with Trinitrobenzene Spectro-Grade N,N- Dimethyl Formsraide - 0.75M KPF, 0 Formamide (Undistilled) -0.75M KPF6 +10% (by weight) H20 <£ -Butyrolactone (Distilled) -0.75M KPF, +10% (by weight) H20 Propylene Carbonate (Undis- tilled) saturated with KPF, 0 Peak Potentials, Volts (NKE) OXIDATION N2 NPO NPO +1.8 NPO NPO co2 — NPO:. NPO NPO NPO HCOOK — — 1.24 0.85 NPO NPO H£CO — NFO NFO NPO NPO H2 NPO NPO NPO NPO NPO REDUCTION N2 NPO NFO -0.5 -1.5 NPO co2 — NPO -1.0 NPO - NPO HCOOH •* • -0.80 NPO 0.0 -1.0 H2CO — NPO -1.0 NPO NPO H2 . 0.0 NPO -1.0 -0.75 NPO ------- fffiWiffliK CORPORATION ,'Voltammo grams of undis tilled propylene carbonate saturated with KPF, showed * the possibility of utilizing this electrolyte in NO transducers, both NO X • • (0.86V) and N02 (0.72V) being oxidized at a lower potential than S02 (1.25V). Sweep voltammetric curves of C02, H-, H2S, H_0, 0 ,.. SO-, CO, formic acid and formaldehyde indicated that these gases would not interfere with the oxidation of NO and N0. • 3.2 EXPERIMENTAL SO AND NO TRANSDUCERS 3.2.1 EVALUATION OF SEMIPERMEABLE MEMBRANES The concept of selective permeation, which dates back to the nineteenth century, is based on the phenomenon that gases diffuse at a different rate through certain polymeric films. A membrane or polymeric film, through which a gas diffuses at a rate several times larger in order of magnitude than another gas, is said to be selective for the former. From the above discussion, the importance of semipermeable membranes in the development of S09 and NO selective sensors seems evident. Electro- £• X chemical transducers were assembled employing different semipermeable membranes, PbO_ and Au electrodes, and IN aqueous H_SO, electrolyte. The net selectivity (potentiostatic control was not employed) of these permeation membranes toward N02, NO and SO- are shown in Table 5. These studies revealed the existence of several semipermeable membranes highly suitable for SO- transducers, they are: MEM 213, polyethylene, cellophane, 'poly-vinylchloride and j>ellicon. Unfortunately not a single NO selective membrane was found. x The polyvinylchloride and cellophane membranes are absolutely impermeable to NO and N0_, but they present severe drawbacks. The polyvinylchloride membrane has a sluggish response. The cellophane membrane is prepared by imbibing the natural polymer, cellulose, in glycerol, glucose or polyhydric alcohol. These compounds, which act as plasticizers, were found to slowly hydrolyze in acidic solutions. Once the plasticizer is totally removed, this membrane becomes brittle and loses its selectivity towards S0_. The pellicon membrane, which responds almost equally to NO- and S0?, could be employed in either S0« or NO- transducers. The utilization of this membrane would require potentiostatic control in an SO- transducer and would require an SO- scrubber in an NO- sensor. Pellicon hydrolyses in diluted ------- • . TABLE 5 . • • SELECTIVE MEMBRAKES MEMBRANES' Teflon, 1/4 mil MEM. 213, 1.2 mils Polyethylene, 1.2 rails . Polyethylene, 0.3 mils Polyvinylchloride, 1.5 mils Tedlar, 1.2 mile Saran Wrap, 1 mil Pellicon, 0.5 mils Zitex 12-137B, 2 mils Cellophane, 1 mil. . Polyethylene and Pelicon Double Membrane, 1.5 mils Selectivity Ratios so2 0.83 1.00 1.00 1.00 1.00 0.0 0.0 0.94 1.00 1.00 1.00 NO 0.73 0.08 0.08 0.14 0.0 0.0 0.0 0.0 0.30 0.0 0.0 K02 1.0 0.39 0.03 0.0 0.0 0.0 0.0 1.0 0.69 0.0 0.0 V- "\ | ------- 'JORPORATION /solutions of strong acids (H-SO , ) , but it is very stable in diluted and even reasonably concentrated solutions of weak acids (CHiGOCl!) . Transducers J employing IN acetic acid electrolyte were assembled and tested. It was found that IN CH^COOH can successfully replace the IN H-SO, electrolyte. ' -J £. H- The polyethylene and MEM 213 (a silicone-polycarbonate copolymer) membranes are the most suitable for S0_ transducers. MEM 213 responds to S0_ at a slower rate than the polyethylene membrane, and its sensitivity is temperature dependent. Sensors employing polyethylene membranes require potentiostatic control. By adjusting the applied potential upon the gold sensing electrode to 0.550V, the N0_ interference is totally eliminated. The interference from NO cannot, at this potential, be eliminated, 'but it is very small (2 to 5%). At an applied potential of 0.533V the NC>2 is reduced and its interference is of the order of -2 to -5%, while that of the NO is 0.4 to 1%. 3.2.2 TRANSDUCERS EMPLOYING AQUEOUS SULFURIC ACID ELECTROLYTE - 1 - 2 Transducers employing 10 N, 10 N, IN, 25% and 50% aqueous II SO, were f-~ M* assembled and tested. In all cases Pb09 counter and Au sensing electrodes were used. No substantial differences -- selectivity, response and response time wise -- were observed for S0?, N0_ and NO in the first four aqueous H9SO. electrolytes. But in the fifth one, 50% aqueous H SO. , £> *\ • £ T- sluggish responses were observed. Electrolytes composed of IN H9SO, with various additives were tested with £. *"T the objective of improving the N09:NO ratio, the selectivity for NO , and £•* • 2C the response time. Attempts were made to reduce the S0? solubility by employing the principle of "common ion effect". The ideal ion for this purpose would have been SO,-, because it affects the first order reaction: Unfortunately, this cannot be accomplished because, if S0_ was added to the electrolyte, it would also undergo electro-oxidation. Therefore, efforts were made to reduce the S0? solubility by employing the same principle, common ion effect, upon the second order reaction. = + 4ll+ + 2e~ This was done by employing electrolytes composed of K0SO. or Na SO in . 2424 ------- CORPORATION IN II' SO. at concentration ranging from 5% to saturation. The addition of sulfate ion did not have any effect upon the sensors selectivity. Electrolytes composed of IN H^SO, saturated with three nitroderivates of benzene, C,HCN00, D-C..H, (N0_)0 and C,H0(N00) 0 were evaluated. The C,HCNO ' O _> 2 OH/i^ O J £. O • O J /. and the CfiH.>(N(V)- exhibited neither positive, nor'negative effects upon selectivities, sensitivities and response times. The IN H9SO. saturated £. ^T with p-C,H. (NO,.") „ electrolyte showed a positive effect upon the response O *T £* f~ times of oxides of nitrogen, and made the NO^rNO selectivity ratio almost equal to one. Polyalcohols are known to inhibit the oxidation of sulfur dioxide. For this reason, sensors utilizing 5% and 10% of glycerol and mannitol in IN H-SO^ respectively were evaluated. Glycerol showed a medium detrimental effect upon the S0~ and NO- responses, and a very large negative effect on the NO response. 3.2.3 TRANSDUCERS EMPLOYING KPF6 IN PROPYLENE CARBONATE AS ELECTROLYTE Transducers utilizing various concentrations of potassium hexafluorophosphate in propylene carbonate, CdF? counterelectrode and Au sensing electrode were assembled and tested. 'Dr3' gaseous mixtures of 2000 pprn NO, 2000 ppm NO- and 2000 ppm SO., were used in the evaluation of these sensors. The range of applied voltage was +1.45V to -1-1.60V anodic to the CdF~ counter- electrode . The experimental work performed employing these transducers corroborated the results obtained from the sweep voltam.aie.tric studies, i.e., electrolytes composed of PC-KPF, are selective for NO . No interferences from CO , O X tL formic acid, formaldehyde, II , HpS, 0?, CO and NH~ were observed. A small response to SO^ was observed. Table 6 shows the selectivity ratios of these sensors under various conditions, for S0_, NO and N0_. The effect of electrolyte additives to the electrolyte comprising 0.25M KPF, in propylene carbonate were evaluated. The addition of 1.3% (by volume) nitrobenzene and 10% HO to the 0.25X KPF in P.C. resulted in a faster response (2.5 times faster) to N0_. But the backgi-ound current was increased from 7.8uA to 84.5uA for nitrogen, and the response to NO and NO decreased from 23uA to 2.5uA approximately. The addition of 1.3% HK03 to the 0.25M KPF, in P.C., 1.3% nitrobenzene, and 10% II 0 electrolyte, raised the background current to approximately 125 uA and the NO and NO response time increased to ------- TABLE 6 . EVALUATION OF NO- TRANSDUCERS EMPLOYING K?F,-PC x 6 ELECTROLYTE Saturated KPFg in Propylene Carbonate Saturated KPF, in Propylene Carbonate 0.5M KPF, in Propylene Carbonate (160°F) 0.25H KPF in Propylene Carbonate MEMBRANE 1/4 mil Teflon 1/4 mil Teflon 1/4 mil Teflon 1/4 mil Teflon POTENTIAL +1.560 +1.45 +1.53 +1.59 CURRENT FLOWS (ua)/2000 ppai N2 11.0 1.36 27.8 7.8 so2 12.65 1.4 27.8. 7.83 NO 23.85 10.3 83.2 21.8 NO 2 27.30 17.2 60.4 23.9 SELECTIVITY • ' RATIOS S°2 9.S 0.46 0 .21 NO 76.7 100 100 87.0 N02 100 92.5 58.9 ------- CORPORATION .Electrolytic systems employing KPF, in P.C., despite their selectivity for NO , present severe drawbacks, the most important being response and recovery 2t times. The average time to reach 90% response was 5 minutes for NO and 30/40 minutes for NO . The average time required to reach 90% recovery was 1 hour and 1.. 3 hours respectively. These sensors also require long periods of time to achieve stable responses. This time may vary from a few hours to several days. .The other drawback of this electrolyte is that it loses its selectivity for NO when H90 is present. X £+ 3.2.4 f^. ^COOH IN PROPYLENE CARBONATE A single experiment performed utilizing P.C. saturated with KPF, and doped o with CELCOOH revealed a system with apparently large potentialities. The J addition of CH'COOH caused a reversion of the NO-NO selectivity, i.e., the systems became moire sensitive to NO than to NO,.,. It also results in improved response and recovery times. A considerable effort was placed upon the development of an NO sensor JC utilizing the KPFg-PC-CH COOH electrolytic system. In the course of these studies, Pt and Au sensing electrodes and Ag?S, CdF~ and PbO,, counterelectrodes were evaluated. Approximately thirty sensors using four different concentrations of CiUCOOH in 0.75M KPF in propylene carbonate were exhaustively tested. The acetic acid concentrations were: three drops per 30 mis. of electrolyte, 2% by weight, 5% by weight and 10% by weight. Previous findings, i.e., reversed NO,:NO ratio and faster response times, were corroborated; but the number of difficulties encountered were sizable. Among them ware the following: a. Response times two to three times slower than in NTUSO, electrolyte sensors . b. Large zero and span drift. c. Background current twice as large as the span current for 1000 ppm of NO. d. Sensor exhibited erratic behavior. e. None of the sensors assembled were operational for periods longer than two weeks. No reasons for this behavior were observed. f. Sensors were very noisy. ------- CORPORATION •Due to time shortage, the study of the KPF^-PC-CH^CQOH electrolytic system was abandoned. ' • . 3.2.5 TRANSDUCERS EMPLOYING OTHER ELECTROLYTES Table 7 summarizes the electrolytic systems studied in this phase. The transducers, in most cases, had gold sensing and cadmium fluoride counter- electrodes. Gaseous dry mixtures, in the 2000 ppm range, of NO, N0_ and S0_ were employed. Sweep voltammotrie studies indicated the possibility of employing N, N dimethyl formamide, 0.75M KPF, electrolyte in SO, sensors. Sensors were evaluated at applied anodic to counterelectrode voltages of 4-0.75V, +1.00V, +1.10V and -1-1.20 volts; but no S0? selectivity was observed. Apparently, the oxidation peak values for NO, NO, and S0? differ but their oxidation curves overlap. Nitrogen dioxide was reduced at voltages of +1.10V or lower. At 1.10V the current flows were 30ua for 1987 ppm NO, 26 ua for 2054 ppm S02 and 24 ua for N?. Transducers were assembled using the electrolyte gamma-butyrolactone 0.75M KPFg +10% H2<3. Teflon. (1/4 mil), pellicon (0.5 mil), polyethylene (1 mil), and MEM 213 (1 mil) were used as membrane materials. Nitrogen dioxide was reduced and nitrogen oxide was oxidized at potentials lower than 0.93V. This makes the electrolyte unsuitable for use in NO sensors. Different permeation membranes were used in the evaluation of gamma- butyrolactone -0.75M KPF, +10% H_0 electrolyte. The pellicon membrane dissolved in the electrolyte. The silicone membrane became brittle and disintegrated. A transducer assembled using a 1/4 mil teflon membrane . performed well but there was too much interference from nitric oxide. Another transducer using 1 mil polyethylene as a membrane showed good selectivity towards S0? in the presence of NO and N0?. However, rapid loss of electrolyte during life data evaluations at 160°F resulted in salt crystallization under the 0.2-0.5 mil polyethylene membrane causing the transducer to become inoperative. Transducers were also assembled utilizing 0.375M KPF, gamma-butyrolactone •1-5% 11JD as electrolyte. Decreasing the salt concentration reduced crystallisation as electrolyte was lost, but the response current and response rate to NO were significantly decreased. The use of 0.2-0.5 mil X • polyethylene, 2 mil polyvinyl chloride membranes, and 1/2 mil teflon ------- TABLE 7 EVALUATION OF VARIOUS ORGANIC ELECTROLYTES Electrolyte Formaruide -0.75M KPF, + 10% H20 N, N- Dimethyl Formamide -0.75M KPF, D Propylene Carbonate Saturated with- KPF, o % Butyrolactone -0.75M KPF6 4-10% H20 .375M KPF, in "o Bu ty- ro lactone-^5% H£0 * .375H KPF, in ^ Buty- rolactone +5% EJD •>''>' Transducer so2 so2 NO X so2 so2 S02 Potential +1.05 •fl.lOv +1 . 60v -rO.92 +0.92 +0.92 so2- . +100% +100% 0.48% +100% 63.5% 100% NO 0 +300% +100% +12.3% 100% 0 N02 +0.935% Large Reduction 92.5% -3% -5.15% 0 -•> With I/2m Teflon membrane ------- U irillWl. ^OiitiMiit'l!:^ COKPGRATIOiM membranes did not substantially affect the entent of the response to S02* In the evaluations of transducers using formamide -0.75M KPF,. 4-10% H_0, the voltage applied to the gold working electrode was varied from -KL.OOV to 41.10V to achieve a large response to S09 without interference from NO and N02< At -1-1. 05V the current flow was 61.5 ua for 1941 ppm SO , 33 ua for 2011 ppm NO, 33.6 ua for 2410 ppm N02, and 33 ua for dry nitrogen. The interference was 0% for KO and 0.90% for NO . At applied voltages of +1.00V and balow, N0? is reduced, but at potentials of -i-1.05 and above it is oxidized. This electrolyte was considered unsuitable because of the large background current . 3.3 SELECTED NO AND S00 TRANSDUCERS " X 2. ,-.— ——— The transducer utilizing l.ON HJSO, electrolyte, gold sensing and lead dioxide counterelcctrodes, and a 1 mil polyethylene membrane was selected for employment in the S0~ prototype monitor. The transducer utilizing aqueous l.ON H^SO, was selected over the transducer utilizing formamide -0.75M KPF,. -M0% HJD primarily because of the latter 's small response to SOy in comparison to background current. The transducer utilizing gamma - butyrolactone -0.75M KPF, -!-10% H?0 was unsatisfactory primarily due to the rapid loss of electrolyte and the crystallization of salt. The 1 mil polyethylene membrane material was selected over other materials because of its stability, selectivity for S0« and permeation characteristics. Sweep voltaianetry studied revealed the existence of only one electrolyte, KPF,. in propylene carbonate, suitable for use in NO selective transducers. 6 l L J ' x . Further evaluation of this electrolyte, with and without additives employing the actual electrochemical sensors, showed the existence of severe drawbacks. The overcoming of the problems encountered with this transducer would most likely require a considerable length of time. Studies performed employing a wide variety of simepermeable membranes did not uncover the existence of NO selective membranes. Therefore, because of lack of alterna x ' tives, the NO transducer to be used in the prototype monitor will employ X IN II^SO, , gold sensing electrode, lead dioxide counterelcctrodes and 0.25 mils teflon semipcrmeable membrane. This sensor is not selective for NO • *•• ancl therefore a scrubber for the removal of S0_ is a necessary part of the NO sensing system. : X • ------- COHPORAYICKO , '3.4 SULFUR DIOXIDE .SCRUBBING SYSTEMS Once the necessity of an SO scrubber was established, various SO scrubbing systems were designed and evaluated. Scrubbing is a basic chemical engineering unit process which consists in the absorption of gases-" in liquid or solid phases. The process is favorably affected by turbulence, high inter- 1 facial surface areas, and high diffusion coef f iciencies . In order to have jj efficient absorption the gas must either react irreversibly with, or be f very soluble in the absorption liquid or solid. The studied scrubbing systems I may be grouped in tv?o categories, electro-oxidation and wet scrubbing. The I I latter clascif ication, wet scrubbing, can also be divided into two sub- groups, oxidants and alkaline absorbants . v 3. A.I ELECTRO- OXIDATION SYSTEMS I The possibility of scrubbing SO by electro-oxidation was considered because I sulfur dioxide and nitrogen oxides can be readily oxidized to sulfuric acid \ V and nitric acid respectively. i The reactions take place on a solid electrode and the principles involved jj are similar to the gas diffusion fuel cells. Porous electrodes are used to C fi provide large contact areas between the gaseous fuel (in this case, pollutants), f: the liquid electrolyte and the solid electrode. However, there are some important differences. Hydrogen as a fuel in a typical H?-0 fuel cell requires a platinum surface which needs to be rejuvenated often. On the other hand, we have abserved that NO and S09 may be readily electro-oxidized on a variety X £» of surfaces including porous nickel, carbon, gold and platinum. Also, unlike in a fuel cell, the cathode may be contained as an integral part of the system. Thus, in place of the 0_ electrode, we could have a strong oxidizing lead • dioxide electrode. The reaction taking place at the lead dioxide electrode is given below: . . Pb02 H- SO^," + 4H+ + 2e - PbSO^ + 2^0 1.685 Volts The reaction product, PbSO,, can be converted back to Pb09 in a fashion similar to that of a lead-acid battery. ------- CORPORATION . '3.4 -SULFUR DIOXIDE SCRUBBING SYSTEMS Once the necessity of an S0» scrubber was established, various SO. scrubbing systems were designed and evaluated. Scrubbing is a .basic chemical engineering unit process which consists in the absorption of gases"' in liquid or solid phases. The process is favorably affected by turbulence, high inter- facial surface areas, and high diffusion coefficiencies . In order to have efficient absorption the gas must either react irreversibly with, or be very soluble in the absorption liquid or solid. The studied scrubbing systems may be grouped in two categories, electro- oxidation and wet scrubbing. The latter classification, wet scrubbing, can also be divided into two sub- groups, oxidants and alkaline absorbants . 3.4.1 ELECTRO- OXIDATION SYSTEMS The possibility of scrubbing S0_ by electro-oxidation was considered because sulfur dioxide and nitrogen oxides can be readily oxidized to sulfuric acid and nitric acid respectively . The reactions take place on a solid electrode and the principles involved 5 are similar to the gas diffusion fuel cells. Porous electrodes are used to [.. I; provide large contact areas between the gaseous fuel (in this case, pollutants), i $ the liquid electrolyte and the solid electrode. However, there are some || important differences. Hydrogen as a fuel in a typical H^-O fuel cell requires I t a platinum surface which needs to be rejuvenated often. On the other hand, we \ have abserved that NO and SO- may be readily electro-oxidized on a variety | X £•* t of surfaces including porous nickel, carbon, gold and platinum. Also, unlike i in a fuel cell, the cathode may be contained as an integral part of the system. jj Thus, in place of the 0_ electrode, we could have a strong oxidizing lead i. dioxide electrode. The reaction taking place at the lead dioxide electrode ' ' [< £ is given below: - ? Pb02 H- SO, -f 4H -h 2e = PbSO, + 21^0 1.685 Volts The reaction product, PbSO, , can be converted back to PbO in a fashion similar to that of a lead-acid battery. ------- FTiP'Irl'lfi : ; ______ .JEE^^Hiii^EiSi CORPOKATIOPJ The electro-oxidation cell was assembled using a porous nickel screen as the electrode. The electrode was made hydrophobic on one side by a thin teflon coating. The cathode used, was a lead dioxide electrode obtained from a lead-acid battery. Dilute sulfuric acid was used as the electrolyte. The cell was assembled such that the gas sample could be passed over the hydrophobic side of the electrode and not come into direct contact with the electrolyte. Applied potentials were varied between 0.0V and 0.4V. Sulfur dioxide was most efficiently scrubbed at an applied voltage of 0.0V. Oxides of nitrogen (NO and N09) were also electro-oxidized at this potential. The S09 removal efficiency was of the order of 60-70 percent. This electro- chemical oxidation system could most likely be made more efficient and selective. The system could be made selective by employing a more selective membrane, by the addition of nitrate and nitrite ions, or by finding a mean to keep the potential difference bet\?een the electrodes selective and efficient for S0~. The complete development of this scrubbing system would require an effort beyond the scope of the original contract. 3.4.2 WET SCRUBBING SYSTEMS 3.4.2.1 The __Fe (OH).- iJKaNO3_OxMSH1L.S_cr The ferric ion selectively oxidizes SO,, in the presence of oxides of nitrogen according to the following half-cell reactions: +3 - +2 Fe -f e -:.--—>- Fe 0.77 NO + 2H20 ^r±: NO " 4- 4H+ + 3e" 0.96 N02 -I- HO ^=ZZ N03" + 2H"'" + e- 0.80 S0? + 2H20 ?L=.-£. S04"2 + 411* + 2e" 0.17 The absorption and ulterior oxidation of S0? occurs according to the following reactions; -I- 2Fe (OH)3 ^rz±: 2Fe (OH)2 H- Fe (Oil) =ivrr±: . 2\i, + Fe ------- Y-w; IKSTRIIM SYSTEMS BSVISiCfi tiil!:© __. __ .——-—— — COKPORATSON 'The overall reaction can be. summarized by: 2Fe(QU), -I- SO 5> 2LI 0 + FeSO. -I- Fe(OH), J £» • £• r - ^£ The addition of NaNO~ is necessary because of the large solubility of N0~. Sodium nitrate and nitrite decrease the solubility of N0? by means of the "common ion effect" principle.. A large nitrate ion concentration causes the NOp solvation reaction to shift to the left; hence causing the N0? solubility to decrease accordingly with: .: : J> + j> „ H+shift H,0: -I- ;N-N ;s=5r 1L.O -N + NO -£——!? HONO + UNO . 0 D ^0 . The scrubbing solution was prepared by slowly adding Fe (SO,)„ (100 g) to a solution containing KaOlI and 250 g of NaNO~ in 400 mis of water. The scrubber container was a standard, glass bubbler with 1/2-inch diameter disperser. This scrubbing solution removed 2% of 2000 ppm KO, 30% of 257 ppm NO and was 100% efficient in the removal of 2000 ppm SO at a flow rate of 2 SCFH. The major drawback of this scrubbing solution is its bulkiness, i.e, for one week operation at 2 SCFH of 2000 ppm SO , 400 g of Fe (SO,)0 in two liters of water would be necessary. 3.4.2.2 The_H 0 7^NaNO jOx idant JSc^uJbb ing_Svj;J: £m /C. £. J) This sc7:ubbing solution was composed of 250g of NaNO», 320 mis of II 0 and 80 mis of 30% HO. Sodium nitrate is also used in this solution to reduce the solubility of N0_. The gaseous mixtures employed in the evaluation of this scruber were 2000 ppm NO, 2000 ppm SO and 257 ppm NO . The SO removal was 100%, the NO's 4% and the N0» removal 12%. The main drawback of the H_0?-NaNO_ scrubber is that its decomposition would be catalyzed by organics present in the stack gas. 3.4.2.3 lhe_Na CO«1NaNO»_.Alkaline Scrubber 2 ~> ~ ~~> A scrubbing solution containing 112g of Na CO and 320g of NaNO,, in 500 mis of water was evaluated employing 2000 ppm NO, 1049 ppm NO , 100 ppm NO and 2000 ppm S0». The SO was totally removed. This scrubber also absorbed 2% of the NO, 13.7% of the 100 ppm NO and 18% of the 1049 ppm NO . It must b f. 2 noted that this scrubber does not oxidize the S0_, but merely neutralizes the sulfurous acid 'formed by the dissolution of S0? in water. ------- -h Na + OH" -h NaHCO, NalIC0 -I- Oil H2co3 ^ H20 + CO H- 2XaOH 2H0 This scrubber was selected, with some modifications, for use with the NO prototyps monitors. 3.4.2.3 The_KrCOoHKHGO[,-KNOoiKNO A Ucalina_Scrubber This scrubbing solution is basically like the aforementioned one. The NaJX),. was substituted by K CO,, because of the latters larger solubility; this will extend the useful life of the scrubber. In addition, the solution also contains KKCO^ for the. purpose, of reducing the absorption of the C0_ contained in stack gases. Because experimental data showed that the addition of KNO,, reduces the absorption of N0_, this chemical is also contained in the scrubbing solution. The composition of the scrubber, par 100 gr of I water, is; 10.6 grs. KHC03> 75.7 grs. K^O , 3.2 grs. KNO , and 8.17 grs. KNO . ' Laboratory tests have shows that this scrubber will remove 2000 ppm S0_n_ completely while absorbing less than 1% of 2000 ppm NO and less than 20% of 100 ppm N0?. Because of the low NO concentration (with respect to NO) in stack gases, the overall effect on the NO concentration is insignificant. ------- ' ' . iriSTjUJ?,;Ef!T SYSTEMS DP«'ISiD:i i ._.--. ....._ -i— r-i-_ ._ .. . -._... -••_-.. .. .. .-— ___ -.-u.r--.f-TT - - — axu- _LI.--.- • i~ ' . ..... " ' -------------- --------- - — ~»...a-mJ»..i».MD«mi^.. ••.-ma .• -*...s^.i ..• ......... aiia.im^ -ORPORATIOIM 4.0 PROTOTYPE MONITORS PERFOR^IAHCE DATA The S02 and NO. transducers response is linear throughout the 0 to 5000 ppm concentration range. Figure 6 shows a plotting of NO concentrations versus X the sensor's output and versus the amplified output, thereby illustrating the linear response of an NX- 130 transducer. Similar results are obtained with S0_ transducers. 4.1.1 SENSITIVITY AND OPERATIONAL TEMPERATURES Table 8 shows the effect of temperature upon the sensitivity of NO and SO • X transducers, for various pollutant concentrations, and at three operational temperatures -- 70°F, 120°F and 160°F. The values shown in Table 8 are mean valuar, obtained! from long- exposure data of a large number of transducers It is interesting to note that although the background (zero) currents increase at a rate larger than those of the spaa currents, the transducers sensitivity increases, with increasing temperatures. 4.1.2 SENSITIVITY RANGES Originally the NO and S00 prototype analysers were to have three ranges of x / sensitivity: 0-50, 0-1000 and 0-3000 parts per million. This is equivalent to span ratios of 1:30:60. Two analyzers, SS-330 and NX- 130, capable of sensing currents of the order of 150ua, without saturating the operational amplifier, and having the aforementioned span ratios were evaluated. It was found that at the lower range (0-50 ppm) the instruments zero and span pots, had poor resolution. It was therefore, necessary to increase the lower range from 0-50 ppm -to 0-100 ppm, the new span ratios being 1:10:30. 4.2 .iELj^rVTrY The effect of various potential interferences was determined employing the final version of the S0~ and N0_ prototype analyzers (Table 9). The selectivity ratios of various permeation membranes for NO, NO and S0_ ware shown in Table 5. Four different pollutants were found to interfere with the selective sensing ------- 0 1000 2000 3000 4000 5000 PPM NO Figure 6. Response of KO Prototype-Monitor X ------- TABLE 8 TIB! EFFECT OF TEMPERATURE UPON TRANSDUCERS SENSITIVITY STACK MONITOR SS-330 NX- 130 POLLUTANT CONCENTRATION 20 ppm SO „ .'. 300 ppm SO 1000 ppm SO 2000 ppm SO N2 50 ppm NO „ 2000 ppm NO CURRT 70°F 0.15 1.75 6.95 15.8 0.03 0.7 2.84 1000 ppm NO | 12.10 2000 ppm NO N2 23.0 0.03 :NT, .MIC ROAM 120°F __ 7.3 25 48 1.2 -- 21.4 44 . 1 1.0 PERES _ 160°F . __ 13 45 95 8- '- 1.8 6,5 32.7 64.0 5.5 ------- 'TABLE 9 EVALUATION OF VARIOUS INTERFERENCES MONITOR NX130 SS330 . CO 121ppm 0.0 0.0 266ppm 45.0 ' 30.2 lOOppm 0.0 0.0 ] Q 2 217. 0.0 0.0 INTERFERENCES PERCENT* HCOOH ; H CO 600ppni ; 500?pn: 0.0 0.0 0.0 0.0 C°2 m- , 0.0 0.0 * / 548ppm 2.00 3.34 E2 0.0 0.0 • . 03 ! 10% SOOppsd : H,0 £ 0.0 0.0 0.0 0.0 * • Monitor J NX130 respect 1000 ppn NO at 70°F ^ SS330 respect 1000 ppm S0? st 70°F ------- ^ J5SI5HIE!! CqitPORATIOM control (Figure 5) the interference from 300 ppm N09 is almost wholly eliminated (0-0.5%), and the interference from 2000 ppm NO reduced to 7%. The interference from 548 ppm SO- and 266 ppm E9S were found to be 3.37o and 30.2% respectively. All of the aforementioned degrees of interference are expressed with respect to 1000 ppm of S09. Among these interferences only that of II93 may be said to be significant. Should it become desirable, the H9S 'can be readily and selectively absorbed in a column containing silica gel impregnated with Ag SO.. Ammonia interferes with the selective absorption (1) -of 1I_S from SO -H S gaseous mixtures. ' It is important to bear in mind that 11 S-SQ mixtures are not stable because of the following probable reactions: US 4- 5SO 4- HO rrr——>• 2S 0 "2 4- 411* 8H2S -I- 4S02 -rs=±^ SO -I- 4lI20 4- Polythionic Acids The NO (KX-130) analyzeris, despite of the results shown in Table 9, free of interference from S0», 11 ~S and S0_ because this analyzer employs a scrubber which absorbs SO,., SO,, and H^S with an efficiency of 100%. 4. 3 RESPONSE AHD RECOVERY TIMES Figures 7 and 8 illustrate the response and recovery times characteristic of S00 (Model SS-330) and NO (Model NX-130) respectively. The S09 transducer £•• X ' f~ has a response time of 90% of full scale (F.S.) in less than forty-eight seconds and a recovery time of 90% of F.S. in less than eighteen seconds. The NO electrochemical transducers have a response time of 90% F.S. in x • five to twelve seconds for NO, and thirty to forty seconds for NO,.,. The recovery times (90% F.S.) for these two pollutants are twenty-four and forty seconds respectively. 4-4 ST^BILITY_pj;_KO AMD .-SO PROTOTYPE JHNITORS The zero and span drifts per day of NO and SO selective analyzers were measured at temperatures of 40°F, 100°F and 160°F (Table 10). The values shown in this table were obtained under the most adverse conditions, i.e., no temperature control was employed. As it may be expected, the instrument's drift increases with increasing temperature. Still, even at 160°F the daily (1) ------- FIGURE 7 CHARACTERISTIC RESPONSE OF MONITOR SS-330 TO VARIOUS SO CONCENTRATIONS CHART SPEED: 1 INCH/HOUR 29 PPM SO,, II' 10 PPM 2000 PPM 2 .LOW.: LOW „_ .....HIGH. A.P.M. RANGE ------- FIGURE 8 CHARACTERISTIC RESPONSE OF MONITOR NX-130 TO NO AND NO,, 2000 KO HIGH 2000 PPM •NO,, CHART SPEED: 1 INCH/HOUR 20 PPM NO LOW LOW A.P.M.HANG IS ------- TABLE 10 STABILITY OF NO^ AND S02 PROTOTYPE MONITORS :-: "." " ~~ ' ' MONITOR NX- 130 . • SS-330 TEMPERATURE "F 40 100 160 40 100 160 ±%, CURRENI ZERO 0.15 0.50 0.9 0.23 0.61 0.95 DRIFT /DAY * SPAN 0 . 24 0.51 1.3 0.5 0.61 1.2 'Relative to 340 PPM SO -Full Scale, in the instruments' most sensitive range. ------- CORPORATION drifts do not exceed ±1.3%. At temperatures below 110°F. (the most widely vised temperature) the instruments' drift is well below ±17o of full scale per day. With the use of temperature control, the span and zero drifts are well below ±0.5% F.S. per day. 4.5 TRWSDUCERS LIFE DATA The transducer's life is a direct function of the diffusion rate of the electrolyte through the permeation membrane, and therefore also a function of the pollutants' flow rate through the sensor. The same liquid electrolyte diffuses at different rates through different'membrane materials. The evaporation rate of IN H-SO^ electrolyte through 0.25 mil Teflon and one mil polyethylene membranes were determined at two different temperatures, 70°F and 160°F. The electrolyte losses were found to be 0.04 grs/day at 70°F, and 0.25 grs/day at 160°F for the polyethylene membrane. Through the . 0.25 mil Teflon membrane, the losses were 0.26 grs/day at 70°F and 1.86 grs/day at 160°F. The gas flow through the cells was one S.C.F.H. It is necessary to emphasize that thase. values were obtained utilizing dry gases. Under standard conditions, the losses of electrolyte would be substantially smaller. 4•6 TEMPERATURE COMPENSATION AND CONTROL Because of the large effect of temperature upon the sensors output (Section 4.1.1), the necessity of temperature compensation or control is evident. Attempts were made to temperature compensate the sensors output by means of a thermistor and associated electronics. Various different thermistors, positioned in the top half of the sensors, were evaluated. The best results obtained were with GA45P1 (50K, Fenwal Electronics Inc.) thermistor. This thermistor and associated electronics were found to effectively compensate (±87, of full scale) for temperature changes within the 40°F to 120°F range. Since larger accuracy and stability were desirable, a temperature controlled f K environment for the sensor was designed. The heater and sensor assembly are f discussed in detail in the Appendix. The heater and associated electronics : effectively control (±0.5CF) the temperature of the sensor within the 40° to | 140°F range. This small temperature variation (±0.5°F) to which the sensor is is 3 subject to, has no significant influence upon the sensors output. / ------- COPPOrtATION APPENDIX Figure 1A shows a front view of the newly developed - second generation - Dyriasciences Air Pollution Monitor. Front, top and side schematic views of the. Monitor are shown in Drawing No. 620432. Four of these instruments (two NX-130 and'two SS-330) were supplied to H.E.W. as part of the Contract CPA-22-69-118. The monitors are compact and portable, and they were designed to continuously measure the concentration of different pollutants in a gas stream. The pollutants that can be measured are sulfur dioxide and oxides of nitrogen. The outstanding .feature of the monitor is that any one of the six sensors developed by Dynasciences can be incorporated in the same monitor. _ The model numbers, ranges, and the recommended applications of the different sensors are as follows: MODKL 1. NX--130 SP 2. SS-330 SP GAS MEASURED ""Total oxides of nitrogen Sulfur dioxide .RANGES 0-100, 0-1000 and 0-3000 ppm 0-100, 0-1000 and 0-3000 ppm APPLICATION Industrial stacks, automobile exhaust, process control, development studies Industrial stacks, process control, development studies "Requires scrubber if S0? present. Table 1A shows the specifications for Monitors NX- 130 and SS-330. Each sensor is a sealed unit and no addition of any electrolyte is ever required. The sensor operating life may be as long as one year or more depending upon the condition of the sample. Generally spsaking, hot, dry and dirty samples will shorten the sensor life. The sensor may be easily replaced with a new sensor as described below. Drawing 672020 and 672021 show the design and specifications of the transducers polypropylene body. ------- WIJASOTS AIR POLLUTION MONITOR BAMPLB FLOW 40 . EQ/ BO • •'• -.., . mo OF RANGE SYSTEM HEATER ZEHO RANOB SPAN POWER POWER „__ Figure 1 -A. Figure 2-A. ------- COHPORATIOW TABLE 1A MONITORS SS-330 AND MX- 3.30 SPECIFICATIONS SENSITIVITY RANGES; DETECTION LIMIT: RESPONSE: RESPONSE AND RECOVERY TIMES; STABILITY: SPECIFICITY: TEMPERATURE CONTROL: OUTPUT: SENSOR LIFE: POWER SUPPLY: MONITOR WEIGHT: FLOWRATE: 0-100,0-1000 and 0-3000 parts per million •2% of full scale Linear response over the entire range of the pollutant. Overall accuracy: ±2% full scale .with the use of integral meter, ±1% full scale with the use of external potentiometrie recorder 907o of full scale in less than 60 seconds Better than ±1% of full scale per day No response to N2, 0~> CO, COo, water vapor or hydrocarbons About 120°F ±1%F, adjustable (see Text) Meter readout and 0-10 mv output for recorder One year at 70°F, moist gases 105 to 125V 60 Hz 5 VA required for powering solid state amplifier or 12 V.D.C. Less than 15 Ibs. Keep the flow rate within the green indicator on the flow metc-.r ------- A| -5.EE DON f^JtfM IOAU&70 \-'ff ,"ArL-t*i VIEW B' B OPTION CHART -I SENSOR AMD HEATER (MODEL £OO) -Z. SENSOR AND HEATER (MODEL4OO) -3 SENSOR MODEL 2OO -4 LESS SENSOR AND HEATER(5TOCK OMLY) -5 LESS SENSORCSTOCK ONLY> COVER MOUKITING SCREWS-2 PX.ACE.SIREF) S2 REOO lEO'APART TMRO \TEM IA\ -CHASSIS MOUMT SCREWS-2 PLACESCREF) FOP PARTS LIST SEE PL620432 VIEW A-A mcr tan I uso> o u HR»J ft-18-fo f nppnDATinM ' MON ITOR,POLLUTION, AIR ------- •®-\ K-®«- ; A PARTS LISTS FOR DASM NUMBERS NOT DRAWN; INrORMATION ^"^ FOR REF ONLY. /§> ARROWS INDICATE DIRECTION OF FLOW ^ HEM 16 IS REPLACEMENT LAMP FOR ITEMS 135118. A, ITEM 45 IS 7/16 LONG AND CENTERED ON'S' POSITION. P^ITEM 143 USED w ITH HEATER, ITEM 88USEDVW/O HEATER ------- ,, cor:rjor.ATioi\! • SE-Ng Ojj, REPLA CEMENT " The instrument is designed to be maintenance free. Tha sensor must never be taken apart. A fault in the sensor usually shows up as slower response . time and/or inability to calibrate the instrument. The faulty sensor may be easily replaced with a new sensor in the manner described below. The newly replaced sensor may take 6 to 8 hours to stabilize before it is ready for use. Before replacing the sensor, make sure that the slide switch located near the battery on the chassis is in the proper position. The use of models Nil- 210 and NR-230 for NO- monitoring requires the slide switch moved towards the front of the monitor. All other models require slide, switch in the opposite direction as marked. The procedure . for sensor replacement is as follows: Step 1 - Disconnect 115V input power cord. Step 2 - Remove (2) screws, Item 65, and remove cover/rear panel assembly. Step 3 - Disconnect (2) tube fittings, Item 152. Step l\ - Unplug electrical leads running from sensor to chassis. Step 5 - Remove insulator, Item 130. Step 6 - Remove (2) screws, Item 150, which hold cover, Item 141, to enclosure, Item 140. Step 7 - Remove cover/sensor assembly 602580 and replace with new assembly. Step 8 - Reverse Steps 1 through 7. NOTE: Fittings, Item 152, shall be firmly Jhand tightened only. HEATER A heater assembly with the appropriate electronics is provided on all the monitors. The purpose of the heater, as it was previously stated, is to maintain a constant temperature environment for the sensor, thereby improving the stability and accuracy of the system. Figures 2A and 3A show "the sensor and heater enclosure in different stages of assembly. Drawing 620432 schematically shows the sensor and heater assembly. A-5 1 ------- ; . ______ inSTngOJT$Vg£gD;yi?g COJU'ORATIOW 5 "The heater can be switched on by pushing the HEATER POWER switch on the front panel. . About 3 1/2 to 4 hours are required for the sensor to come to equilibrium at a factory-set temperature of about 120°F. Laboratory tests have shown that heating the sensor at temperatures higher than 120°F result in shorting its operating life. Continuous heating to 160°F for 4 hours resulted in permanent damage to the sensor. The sensor may be operated at 140°F for a short duration of up to 8 hours at a time, without damaging the sensor. To readjust the temperature set-point, proceed as follows: Remove the printed circuit board from its receptacle. Using an Ohmmeter, check the resistance between Points A and B (see Drawing No. 602517, Printed Wiring Assembly) and adjust pot R41 to obtain 25K ohm resistance. Install the printed circuit board back in its receptacle. For the applications where the ambient sample and room temperatures are fairly constant or where the monitor is required to be. used for a spot test, the heater is not required, thus avoiding the delay for the sensor to come to equilibrium. The heater is proportionally controlled thus eliminating the possible inter- ference from an On/Off action. The HEATER red light on the front panel indicates the level of power to the heater. In tha beginning when the HEATER POWER is switched on, the HEATER light is very bright. However, within the first two hours it starts dimming and stays fairly dim for the continuous operation. CAUTI.ON - If the HEATER light remains bright for more than two hours, turn the power off and check for possible malfunction in the heater circuit. The two white leads from the heater should be firmly plugged in the white jacks on the chassis which supply the power to the heater. Check the resistance across the two green leads after removing them from the chassis. These loads are joined to a thermistor which should read about 100 Kohrn resistance at room temperature and about 45 Kohm resistance at the controlled temperature of 120°F. A broken thermistor would show an infinite resistance avid would keep the heater power on all the time. ------- s , i • • . • The meter reading is directly proportional to the concentration of the pollutant. In addition, a 10 mv recorder output is provided in the back of • the monitor for continuous recording. The output jack is available on the back of the instrument next to the sample fitting. 115/230 volts and 100 volts, 50 or CO Hz powered from the three-prong receptacle on the back of the monitor. Also a 12 V.D.C. capability is provided. The 12 V.D.C. input jack is provided on the back of the instru- ment next to the fuse. A size "D" mercury battery is used to bias the sensor. Its voltage at all ' I times should remain 1.35V ±10%. The battery voltage may be checked by * holding the RANGE switch on the front panel to the BATT TEST position. [ Replace the battery if the output on the meter shows less than 90%. full •, L scale deflection. . f ?' I: .SCRUBBER I' For the applications where the analysis of NO using NX-130 is desired in JT the presence of SO,., a liquid scrubber is provided for the removal of SO > before the sample is passed to the monitor. The scrubbing solution is especially prepared to remove SO- without affecting the NO concentration. A cartridge containing glass wool is attached on the exit of the scrubber to trap axiy liquid carry-over. SAMPLE _ C01TDI TIONIJgG AND gRESSURE AD.JITS THEM TS , The gas sample should be free of suspended material and the moisture content of the sample should be such that it does not condense in the sensor. Except for this restriction, it is actually beneficial in terms of life- time of the sensor to use moist gases or gases with a dew point slightly below the operating temperature of the sensor (relative humidity of 40 to 80% is recommended). The temperature of the sample should be kept less than 100°1;1 and preferably at room temperature. Higher sample temperatures to 140°F may be used for a short duration of up to 8 hours. Longer exposure ------- CORPORATION be ;passed over the sensor by connecting it to the input 1/4" polypropylene i » - tube fitting provided at the right side when facing the back of the instru- ment. The sample passes through the sensor, flowmeter and exits from the fitting on the left side which should be connected to the vent line. The instrument reading is independent of the flow rate. However, it is suggested that to get the optimum response time and a long operating life, a flow rate of 0.5 to 2.0 SCi7H be kept steady throughout the run. If the flow indicator is kept in the blue area, the flow rate will be satisfactory. The sample should be preferably "pushed" rather than "pulled" through the sensor. If it is desired to pull the sample by creating a suction at the exit, care must be taken to assure that the vacuum created in the sensor is not more than 6" of H90. i ' £ The sensor is designed for operation at atmospheric pressure. It can be . operated "in line4." at pressures slightly above or below atmospheric, but it is of u^ino£t_im]>Drtance that any pressure changes occur slowly in order to prevent damage to the sensor. j ! Since the sensor measures the partial pressure of the gas being monitored, | it must be realized that a correction should be made if the sensor has been 5 calibrated at one pressure and is used at another. For example, calibration of the sensor at 740 mm Hg of ambient atmospheric pressure and 1009 ppm of NO gas results in a full scale meter reading equivalent to 740 mm Hg. J\ "• If the total "in line" pressure is changed to say 700 mm Hg, and the ratio of molecules of NO gas to total gas is kept constant at 1000 ppm, then the new reading will by ~ — of full scale. The instrument may be calibrated in any RANGE position. However, for maximum accuracy, calibrate in the RANGE most commonly used. Turn on power to the instrument by pushing in the SYSTEM POWER switch. Set the RANGE switch to the desired position. Turn the SPAN knob completely clockwise. Wail about 15 minutes before proceeding with the calibration. If the heater is used, wait 4 hours before, proceeding. Flow the zero gas (N9 or NO , and SO, free air) through the instrument and wait for 5 minutes *~ 5t /- Then bring the reading to zero by adjusting the ZERO knob. Lock the ZERO knob. Pass the calibration gas over the sensor. Wait: 5 minutes. Adjust the SPAN knob so that the meter reading corresponds to the concentration ------- rill ____ l?ISTnijLM^5fE^ COFIPOKATIOW in the calibrated sample. Lock the SPAN knob. The instrument is designed to be essentially drift free and will retain its calibration for a long time. However, for optimum results, calibrate the instrument at regular intervals. TjlE ELECTRONIC PACKAGE The following material was prepared by Mr. J. W. Schmidt, Supervisor of Laboratory Instruments. During system operation the cell output current, is converted to a proportional voltage by the Current Amplifier. This voltage signal is then scaled by the span control and amplified by the scaling amplifier. The scaling amplifier gain is determined by the range control (High - Keel - Low) which provides a proper output for full scale meter deflection and recorder output. .The current amplifier input is provided by the cell (MTJ.) and bias network composed of resistors RGO through RG4 and battery (B) . The function of the bias network is to provide the proper cell bias for system application required. The front panel zero control circuit consists of pot R66, resistors R3, R4 and R5 and 6.2V zeners VR, and VR?. This provides a ± current to offset residual cell currents. Q, , R- and R are creating a short circuit to cell currents when the power is in the off state. This maintains the cell in a ready state for rapid stabilization. The current amplifier is basically a low noise FET input amplifier which utilizes the ua 741 for high open loop gain and complimentary output drivers Q. and Gv for linearity. -Potentiometer R,„ and resistors R and *~r *J i. -J i.\J R.„ are used to zero the amplifier during initial alignment. The amplifier gain is determined by RT, , RI and R . R and R are used to modify the temperature characteristic of thermistor RT . RT and R are physically located in the cell to obtain temperature tracking gain which will provide the best system accuracy during heater warm-up. The basic gain equation is: Vo = Iin RF ------- UNLESS OTHEHWISE SPCCIFIED INTPR DWG P£R MIL-STU-100 OiM.AflE IN INCHES DO NOT SC*t£ DRAWING DIM. TO BE MLT AFTER FINISH STANDARD HOLES PFR AHD1038 HSTWUIKT SlTTtWS BYIPOW CHATSWOB' SCHEMATIC D1AGE.A/Y\ MOX MOMITOE. kPPROVED HR.W 4 - 19 - TO 4i. G6024& ------- II ^^.ZVPO 7.FOR ASSY SEE DWG N0.62043E. ^^COMPONENTS IN THIS AREA DELETED ON UNITS W/0 HEATER. C23 IS REPLACED WITH IOO MF,^^5^25 v CAPACITOR. ^^P3 WILL MATE WITH 0 3, J4, J5,J £,J~7 OR J8 DEPENDING ON APPLICATION. GROUND POINT AT PINS C.D.&E OFCONMECTORSOKI CHASSIS ADJACENT TO PIN C . --... -.. GROUND POINT AT PINS 5 S. 10 OF TRANSFORMER AND ON CHASSIS ADJACENT TO TRANSFORMER. " RT'S>RT£ ARE LOCATED irl CELL MT I. RT3& RT<5 MAY BE REQUIRED FOR TEMPERATURE COMPENSATION ON UNITS Y7/O HEATER. OPTION POR MOUNTING THEM ON CIRCUIT BOARD SHALL BE PROVIDED. 2. CaPACITANCt VAl_UE6 AKE IMjjF, IO'/., IOO V I. ftiliToK VALUES AE6 IN OHM1,±l?i, Vtl<*jf/ NOTES: UIILESS OTMESWISE SPECIFIED ------- Rl RT] lin, :•.;• the cell, current Rf = R, 4- «~T~i7^"~ ' ' J. L KI T Jxi, The scaling ampli.fj.er AR_ is an inverting operational amplifier with four selectable gain settings determined by the range switch. The basic gain equiation is: . Rf . • - Av ~ — - Rin where Rf is R 0 for all ranges and Rin composed of one of the series resistors R.(7> R-.O or R,, and the equivalent source resistances. This source resistance is the effective resistance of the span not R,0 and 1 68 , any parallel resistance such as R, , or R, ,_• I J 16 45 I To maintain an accurate raster reading, the meter is electrically contained '. in the feedback loop and the system is therefore immuned to changes in the \ t mater resistance as a function of time and temperature. t Overload protection for the circuit is provided by the diode network of CR.. through CR and R9r|. This limits the amplifier to ±1.4V. • F The recorder output is developed by the resistive divider R ' and R9o- f, rf— <— t*J I This provides a lOmv full scale output with a low output impedance . £ Initial ^ero adjust for AR0 is set by R n . • i; /L /.L v: (.' Temperature control of the cell heater is determined by resistive bridge ij amplifier AR_ and associated driver transistors R through R . | i The resistive bridge is composed of resistors R/0> R/?> ^YiV ^9 (thermistor) [ and potentiometer R , , . Bridge unbalance is sensed by. amplifier AR which proportionally changes the output voltage to the heater. At 25CC the bridge unbalance causes pin 3 to be positive with respsct to pin 2. The output at pin 6 is then at a maximum positive voltage. This voltage is buffered by emitter followers Q., and Q which control the high current /, O pass transistors Q . As the heat increases, the thermistor RT (located on the cell) v?j'.ll continue to decrease in valxie until the bridge becomes balanced. The high gain of the amplifier maintains the heater at a high input current for fast warm-up and then maintains a proportional control within a few tenths of a degree. Circuit protection is obtained by the heater power switch S_ and diodes A-11 CR and CR. . . ------- C O ft ij ORATION The heater light DS3 is directly in parallel with the heater blankets ana 'has a brilliance which is proportionate to the applied voltage. Monitor pov?er can be provided by cither 115 VAC line voltage or +12 TOC. The -I-12VDC operation is dependent-on the PSI DC to DC converter. This is_ a special module which provides ±22V of partially filter voltage in parallel with the rectified AC line voltage. The filtered DC voltages are regulated by integrated circuit regulators Z. and Z . The regulated ±15 VDC is used to virtually eliminate the effects of line voltage and. frequency. They also provide short circuit protection to preve.nt propagation failures. Initial adjustment is set by potentiometer R _ for the H-15VDC and R 3 for the -15VDC. ------- |