EPA-R2-73-254 JUNE 1973 Environmental Protection Technology Series Polarographic Determination of NTA National Environmental Research Center Office of Research and Monitoring US. Environmental Protection Agency Corvallis, Oregon 97330 ------- RESEARCH REPORTING SERIES Research reports of the Office of Research and Monitoring, Environmental Protection Agency, have been grouped into five series. These five broad categories were established to facilitate further development and application of environmental technology. Elimination of traditional grouping was consciously planned to foster technology transfer and a maximum interface in related fields. The five series are: 1. Environmental Health Effects Research 2. Environmental Protection Technology 3. Ecological Research 4. Environmental Monitoring 5. Socioeconomic Environmental studies This report has been assigned to the ENVIRONMENTAL PROTECTION TECHNOLOGY series. This series describes research performed to develop and demonstrate instrumentation, equipment and methodology to repair or prevent environmental degradation from point and non-point sources of pollution. This work provides the new or improved technology required for the control and treatment of pollution sources to meet environmental quality standards. ------- EPA-R2-73-254 June 1973 POLAROGRAPHIC DETERMINATION OF NTA by Thomas B. Hoover Southeast Environmental Research Laboratory College Station Road Athens, Georgia 30601 Project 16020 - EWE Program Element 1B1027 NATIONAL ENVIRONMENTAL RESEARCH CENTER OFFICE OF RESEARCH AND MONITORING U.S. ENVIRONMENTAL PROTECTION AGENCY CORVALLIS, OREGON 97330 For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, D.C. 20402 Price 75 cents domestic postpaid or 50 cents QPO Bookstore ------- ABSTRACT Pulse differential polarography was found to be a more sensitive and precise method "than linear sweep voltam- metry for the determination of UTA as the bismuth- complex in natural and waste waters. % Itidium was a less satisfactory complexant. Relative standard deviations of less than 5% were obtained by the dual cell, pulse differential technique at 1 ppm NTA and calibrations were linear from 0.01 to 4 ppm. Copper, added at con- centrations greater than 10-5 M, caused low analytical results for NTA. No other significant interference by metals or complexants was found. An acetate electro- lyte was much better than chloride for the determination of bismuth-NTA complex in sewage-type samples. Recom- mended procedures are given in the Appendix. ii ------- CONTENTS Page Abstract ~~iT~ List of Figures iv List of Tables v Acknowledgments vi Sections I Conclusions and Recommendations 1 II Introduction 2 III Experimental 4 Instrumentation Indicator Choice Indium Complexation Bismuth Complexation IV Results and Discussion 14 Instrument Evaluation Indium Bismuth, Chloride Medium Bismuth, Acetate Medium V References 39 VI Appendix 43 111 ------- FIGURES No. Page 1 Excitation Waveform for Pulse Differential Technique 5 2 Determination of the Indium-NTA Complex in Sewage Effluent by Pulse Differential 6 Polarography 3 Determination of the Bismuth-NTA Complex in Seawater by Pulse Differential Polarography 24 4 Copper and Bismuth-NTA Waves by Linear Sweep Voltammetry 25 5 Effect of Sewage Effluent on the Bismuth-NTA Polarographic Wave in Chloride Electrolyte ^7 6 Calibration Curves for the Determination of the Bismuth-NTA Complex — Chloride Electrolyte 28 7 Determination of the Bismuth-NTA Complex in Distilled Water by Dual Cell Pulse Differen- tial Polarography 34 8 Determination of the Bismuth-NTA Complex in Distilled Water by Dual Cell Pulse Differen- tial Polarography 35 9 Differential Pulse Polarographic Estimation of NTA in Raw and Treated Municipal Sewage 36 IV ------- TABLES No. Page 1 Effect of pH on the Polarography of In-NTA in 0.2 M Sodium Chloride 10 2 Calibration Runs for the Determination of NTA as the Indium Complex ^ 17 3 Recovery of NTA in the Presence of 5 x 10~5 M of Fe, Mn, Cu, or Bi 19 4 Recovery of In-NTA in the Presence of Metals and Complexants 20 5 Calibration Runs for the Determination of NTA as the Bismuth Complex in Chloride Electrolyte 23 6 Calibration Runs for Bi-NTA in Chloride Electrolyte—Sewage Effluent and River Water 29 7 Calibration Runs for the Determination of NTA as the Bismuth Complex in Acetate Electrolyte 31 8 Recovery of Bismuth-NTA in the Presence of Metals and Complexants in Acetate Electrolyte 38 v ------- ACKNOWLEDGMENTS Most of the experimental work was performed by George D. Yager, whose initiative led to the development of the acetate procedure. Sydney H. Young assisted with the studies on indium. VI ------- SECTION I CONCLUSIONS and RECOMMENDATIONS Nitrilotriacetic acid (NTA) was determined with a standard deviation of 3 ]ag/& in natural and waste waters as the bismuth complex using dual-cell pulse differen- tial polarography. * Indium was a less satisfactory complexant than bismuth for the determination of NTA. An acetate electrolyte gave better results in sewage type samples than a chloride electrolyte. The acetate electrolyte is therefore recommended for the polaro- graphic determination of NTA in natural waters and waste waters containing less than 0.1 percent chloride. A chloride electrolyte is recommended for saline samples. Analytical procedures for both are given in the Appendix. Linear sweep (LS) voltammetry was a more rapid analyti- cal technique than pulse differential (PD) polarography/ but was less sensitive. Dual cell PD is therefore recommended for analyses in which the greatest sensi- tivity and precision are required. Dual cell LS is recommended for routine and repetitive analyses when the highest precision is not needed. The effects of possible interference by metal ions and complexants were evaluated. Copper at concentrations greater than about 1 mg/& may lead to low results for NTA. Linear calibration curves for the determination of NTA were obtained for the concentration range of 0.01 to 4.0 mg NTA/I, according to the recommended procedures in the Appendix. ------- SECTION II INTRODUCTION Nitrilotriacetic acid (NTA) was used briefly as a builder in laundry detergents, partially replacing polyphosphate. The high degree of biodegradability of NTA makes its use attractive, especially in regions where detergent phosphates contribute materially to eutrophication. On the other hand, the potential health hazards of NTA in the environment that have led to a moratorium on its use (1) dictate that reliable analy- tical methods should be available for monitoring the build-up of NTA in the environment should commercial use be resumed. Many analytical methods for NTA have been developed, most of which utilize complexation with metals for the separation or determination. Colorimetric methods have been based on the measurement of Fe(III) (2), Cu(II) (3), or Zn(II) (4,5) with appropriate color reagents. Swisher (2) separated NTA as the Fe complex, together with any other soluble Fe complexes, decomposed the complexes, and determined the resulting Fe. The other authors above measured the excess metal ion after reaction with complexants in the sample. Consequently, these methods are not specific for NTA unless elaborate separation procedures are used. The zinc-zincon colori- metric procedure (5) has been automated to provide rapid analyses with relative standard deviations of 2-10% in the concentration range 0.1 to 10 mg NTA/&. However, several metal ions cause negative interference, sewage-type samples give high results, and the method is not applicable to saline waters. The Cu-NTA complex has been employed analytically, with a copper-ion selective electrode as the detector (6,7). Like the colorimetric procedures, this technique lacks specificity but it is especially suitable for continuous monitoring (6) or for complexometric titrations (7). Methods employing gas chromatography (gc) (6,8,9,10) determine NTA directly as a suitably volatile ester, and thus have an important advantage of specificity. On the other hand, the determination of total NTA requires its separation from complexing metals, generally by ion- exchange pretreatment, with the possible loss of NTA ------- (11). Furthermore/ to esterify the NTA, substantially all the water must be removed by evaporation. This step is lengthy and may lead to losses (6). The better of the gc methods require injections of at least 0.2 yg as NTA and give relative standard deviations of 7 to 13% (9,10). Polarographic methods are based on the reduction of a suitable metal ion complexed with NTA. Since the half- wave potentials for the reduction of several complexes of the same ion usually vary appreciably according to the ligand, this procedure is nearly as specific for NTA as are the gc methods. In fact,some of the earliest polarographic studies were made for the purpose of resolving a series of similar chelating agents (12) or for determining NTA as an impurity in EDTA (13,14). The objectives of this project were • to develop and evaluate an analytical method for NTA that would determine 10 yg/£ or more with 15% accuracy in seawater and in waste water, and • to evaluate and compare some of the newer polarographic techniques of high analytical precision, specifically, pulse differential polarography (PD) and linear sweep voltarametry (LS). ------- SECTION III EXPERIMENTAL INSTRUMENTATION The experimental plan was significantly influenced by the type of instrumentation that could be obtained and by the equipment purchase schedule. The first analyti- cal instrument obtained, and work horse for the entire project, was a Princeton Applied Research (P.A.R.) Model 170 Electrochemistry System. It is one of a few general-purpose electroanalytical instruments (15) that offer great versatility in types of experiments that can be conducted and range of parameters that can be controlled. It was selected primarily for its pulse polarography capability, a relatively new technique having great sensitivity. In the pulse differential (PD) mode, the current is sampled and its value stored twice during the life of each drop from the mercury electrode. These measure- ments are taken before and after a square potential pulse of 50 mV is applied to the electrode on top of a continuous ramp potential that may be increasing a few mV during the life of the drop. The pulse excitation wave form is shown schematically in Figure 1. The second current measurement is taken after sufficient time has passed (about 40 msec) to allow the charging current associated with the pulse to decay to an insig- nificant level. The difference between two current measurements on the same drop gives a derivative-type of polarogram (Figure 2) that is nearly independent of random fluctuations in drop size. The 48 msec measure- ment cycle is completed near the end of the drop life, when the drop size is changing slowly. The technique is especially useful for trace determinations because the effect of the residual current is removed automatically in the subtraction step. The P.A.R. Model 172 Mercury Drop Timer is an essential component of the PD system because it provides accurate drop times of 0.5, 1, 2, or 5 seconds precisely syn- chronized with the current measurement. The timer dis- lodges the mercury drop from the capillary using opposing solenoids that give a nearly vibrationless acceleration to the capillary. ------- 151 ^ 55 msec msec TIME Figure 1. Excitation waveform for pulse differential technique ------- INDIUM - NTA IN SEWAGE EFFLUENT TECHNIQUE: SINGLE P D 0.4 ppm NTA -0.7 -0.9 V -0.7 -0.9 V Figure 2. Determination of the indium-NTA complex in sewage effluent by pulse differential polarography 6 ------- A Chemtrix Model SSP-5A Polarographic System was obtained to provide*a more direct comparison with the dual-cell linear-sweep (LS) voltammetric technique used by Afghan and Goulden (16,17). Although the Model 170 provided LS programming, it was not designed for dual cell work, the voltage sweep was not synchronized with the DME, and sweep rates were limited to less than 0.5 V/sec by the X-Y recorder. On the other hand, the Model SSP-5A was designed for dual dropping mercury electrodes and provided sweep rates from 0.05 to 20 V/sec. The current-potential response of the cell was displayed on a Tektronix Type 564B storage oscilloscope and a Tektronix camera system provided a permanent record of the traces (Figure 4). The polarographic programming and amplification circuitry were supplied as two special plug-in modules for the oscilloscope and drop timing was controlled by a solenoid on the electrode stand. Drop times were continuously variable from about 0.5 to 10 sec by an R-C timing cir- cuit. When the solenoid closed it tapped the capillaries to dislodge a drop from each and simultaneously sent a pulse to the programming module to initiate the polaro- graphic sequence. Near the end of the project the P.A.R. Model 174/51 Linear Sweep Module Accessory became available, together with an adapter kit to allow its use with the Model 170. This modification made possible dual cell operation in either PD or LS modes. The output of the LS mode was displayed on the Tektronix oscilloscope, using general- purpose plug-in modules 3A75 and 3B3 Time Base. After some experimentation the cells for use with either instrument were made from 18 x 75 mm test tubes of borosilicate glass or polycarbonate. Platinum leads were sealed in the bottoms to make contact with the mercury anode and plastic caps were made with holes for the capillary, for the nitrogen inlet (2-mm polyethylene tubing), and for the sample introduction. In most experiments the cathode potential was controlled with respect to the mercury pool. When reference electrodes were used they were made of silver wire anodized in a chloride solution. The salt bridge consisted of a 3-mm glass tube, drawn to a tip, plugged with filter paper, and filled with saturated KC1. The reference electrodes were periodically compared with a Beckman SCE or with the reference electrode of a combination glass pH electrode. ------- INDICATOR CHOICE The choice of metal ion to be used as a polarographic indicator for NTA depends somewhat on the objective of the analysis. To determine total NTA in natural or waste waters one should use an ion that forms a suf- ficiently strong complex to displace all other ions that are likely to be present in significant concentra- tions. On the other hand, to determine free NTA in the presence of natural complexes one should select an indicator ion that does not appreciably displace the equilibria among ions naturally present in the sample but forms a sufficiently strong complex with NTA to give an identifiable polarographic wave. The stability constants of complexes of NTA with many metals have been determined in fundamental studies of chelation (14,18,19). A second consideration in the choice of an indicator ion is the resolution that can be obtained in a complex mixture. Since the width of a polarographic wave is inversely proportional to the number of electrons involved in the electrode reaction, a three-electron reduction will give a sharper, more easily identified wave than two-or one-electron reductions. Bi(III) (6,17) and In(III) (20) have been preferred for the determination of total NTA, although methods have also been based on Cd(II) (21,22,23), and Pb(II) (17). INDIUM COMPLEXATION Studies of the indium-NTA complex were made to determine whether NTA could be measured reliably in the presence of a considerable excess of indium. Afghan (24) and Taylor (6) raised an objection to the use of indium because the In-NTA complex has a half-wave potential within 200 mV of that for free indium. A large excess of the reagent therefore tends to obscure the In-NTA wave. Haberman (20) solved this problem by titrating the sample with indium reagent until a free In(III) wave just appeared. He then used this wave as a baseline for the following In-NTA reduction. He showed, however, that up to 2-fold excess In did not change the indicated concentration of the NTA more than 10%. A second problem with excess indium arises from the recognized tendency of indium in many media to show dc polarographic minima (25,26,27) rather than a 8 ------- constant diffusion current. This behavior would cause a downward-sloping baseline for the In-NTA peak in the PD mode of measurement. A stock solution of indium (1 x 10~3 M) was made from In(NOs)3(Alfa inorganics) in 0.5 M NaCl buffered to pH 4.7 with 0.1 M acetate (20). Preliminary trials indicated that flatter baselines for the In-NTA peak were obtained if additional acetate (or tris(hydroxy- methyl)aminomethane) buffer was not added to the sample solutions. Consequently, calibration curves were run in 0.2 M NaCl supporting electrolyte, or in 60% sea- water, with indium stock solution added to give the desired total In concentration (usually about 1 x 10~5 M). The effect of pH was studied in a series of solutions containing 1 ppm NTA (as the acid form, CgHgOeN) and 100% excess In in 0.2 M NaCl. The pH was adjusted with dilute HC1 in 0.5-unit steps from pH 2.5 to pH 4.5. The results, summarized in Table 1, indicate that the widest peak separation and greatest response to In-NTA were obtained at pH 4.0-4.5, where the flattest base- lines were also observed. In another, more limited trial, however, better baselines were found at pH 3.0. In subsequent work, 0.2 M NaCl adjusted to pH 4.0 was adopted as the standard supporting electrolyte. Since the literature contains no data on the polaro- graphic parameters for the reduction of the In-NTA complex, solutions containing 100% excess indium were studied by dc polarography, LS voltammetry, and chronoamperometry. It was assumed, without independent verification, that a 1:1 complex was formed (28). Under the above conditions, approximately equal limiting currents were observed for the free In and for the complex. These studies confirmed that the indium wave in a chloride medium is reversible and diffusion- limited, and involves a 3-electron transfer. The diffusion current constant, I, in the usual units of mA-fc secl/2 mole~lmg~2/3, was 5.9 + 0.5 for artificial drop times of 0.5 to 2 sec and 5.5 +_ 0.5 for natural drop times of 3-8 sec; m values were in the range 0.7- 2.3 mg/sec. DeSesa et al. (29) found a diffusion current constant of 3.7 in an acetate medium. ------- Table 1 Effect of pH on the Polarography of In-NTA in 0.2 M NaCl PH Ei/2(vs Ag,AgCl) AE (rn3+-In-NTA) In-NTA,ip 2.5 -0.72 V 90 mV 0.158 yA 3.0 - .75 120 .254 3.5 - .76 130 .294 4.0 - .79 160 .335 4.5 - .82 190 .312 El/2(In3+) = -0.63 V (Ag^gCl) 10 ------- The In-NTA wave was also diffusion limited and showed a polarographic minimum, as reported by Haberman (20). The reduction apparently was pseudo-reversible, as reported by Staroscik (28) for an acetate electrolyte, since the combination of LS and chronoamperometric data (30) led to anomalously low values for n (less than 0.5 electron). I was markedly dependent on drop time, with values of 3.1 + 0.4 for artificial times of 0.5 to 2 sec and 9.8 + 0.8 for natural times of 2-8 sec. BISMUTH COMPLEXATION The Bi-NTA complex is the basis for an automated analytical method for NTA, using LS (16) and was proven applicable to seawater samples (31). The present work was undertaken primarily to evaluate the PD technique as a means of increasing analytical sensitivity. The Chemtrix instrument was obtained to provide a nearly direct comparison with the Afghan (17) procedure. In the course of the investigation considerable inter- ference was encountered in sewage-type samples and an acetate electrolyte was developed to overcome these problems. The chloride media were based on those used by Afghan (16). The final analysis solution was about 0.1 M in total chloride (from NaCl or HC1), and contained bismuth, diluted in the sample to give 100% excess (using bismuth stock solution — 1 x 10-3 M Bids in 1 M HC1). The final pH was adjusted to about pH2. When interference from iron was suspected 0.02%, NH2OH-HC1 was added. In seawater the added chloride was omitted and the sample was diluted to 60%. The optimum acetate medium contained 0.1 M acetic acid, 1 x 10-5 M Bi(NO3)3 and 0.02% NH2OH.R2S04, at pH 4.0. It was convenient to use a stock solution of 1 x 10~3 M Bi(NC>3)3 in 9'5 M acetic acid plus 0.5 M sodium acetate. The 1:100 dilution of the stock solution in the sample gave the correct pH in unbuffered samples. The polarography of Bi(III) has been studied extensively and was recently reviewed by Bond (32). His major conclusions were that the reversibility of the reduction of the chloride complexes correlated with the degree of complexation in the bulk solution. At sufficiently high chloride concentrations the reduction 11 ------- was completely reversible. Other halide ions provided sufficient complexation in the double-layer for pseudo- reversible reduction but the aquo complex of bismuth was reduced irreversibly because of the slow removal of water from the ion. In the chloride electrolyte used in the present work the reduction of the Bi-Cl complex cannot be observed because it overlaps the mercury oxidation process. NTA, of course, forms a stronger complex with bismuth, dis- placing chloride and giving a separate polarographic wave at a more cathodic potential. No published data are available for the polarographic reduction of bismuth in an acetate medium, or for the formation of bismuth-acetate complexes. When we dis- covered empirically that an acetate electrolyte was advantageous for determining NTA, we undertook some studies to better understand the electrochemical process involved. As a first step we attempted to define the presumed Bi-NTA complex in acetate electrolyte by the method of continuous variations. No stoichiometric relation could be determined. For all bismuth concentrations, includ- ing zero, a polarographic current proportional to the total NTA concentration was observed, but the propor- tionality constant increased with Bi concentration up to a slight excess of the 1:1 molar ratio of BiiNTA. This may parallel the observation of Staroscik and Webs (28) that the polarographic current for In-NTA in an acetate medium decreased with increasing excess NTA, eventually yielding an adsorption wave. We found, in the dc polarography of solutions containing no bismuth or less than 2:5 mole ratio of Bi:NTA, that El/2 = -0.21 to -0.24 v(SCE). The diffusion current "constant", J, was more nearly independent of mercury flow rate when calculated for the total NTA concentra- tion than for bismuth or excess NTA concentrations. The wave shape corresponded to a pseudo-reversible process, with an of 1.5 to 1.9. No ac polarographic wave could be observed in these solutions. NTA appears to facilitate the reduction of bismuth by displacing water from the aquo complex and forming a 1:1 Bi-NTA complex in the double layer. However, the 12 ------- mechanism of the diffusion-limited process in the absence of bismuth is unclear. The effect of pH (between pH 1.5 and 4.5) was studied in solutions containing 1 x 10~5 M Bi and 5 x 10~6 NTA. The chloride medium contained 0.1 M total chloride and the acetate medium was made 0.02 M in acetic acid and 0.08 M in NaClC>4 for an equal ionic strength. The pH was adjusted with NaOH solution. In the chloride electrolyte, in contrast to the report of Afghan and Goulding (17), the peak current recorded in the PD mode increased sharply from pH 1.5 to pH 2.0 and changed little at further increases to pH 4.0, having a slight maximum at pH 2.5. In the acetate medium the peak current did not vary more than 10% over the pH range of 2.0 to 4.5, but showed a broad maximum at pH 3.0. In both systems the peak potential (£"1/2) shifted to more cathodic values with increasing pH. The shift over 2.5 pH units was about 50 mV in the acetate. Another study of four solutions at an ionic strength of 0.1 and pH 3.0 showed that increasing the total acetate concentration from 0.02 to 0.1 M had negligible effect on either the peak current or potential. 13 ------- SECTION IV RESULTS AND DISCUSSION INSTRUMENT EVALUATION The precisions attainable by the various instrumental techniques studied are best compared by reference to Table 7. In this table are presented the calibration results obtained with the optimum chemical and instru- mental conditions. The practical choice for an analytical application involves a balance between speed and convenience on one hand and precision on the other. Dual-cell PD has a distinct advantage in terms of precision and detection limits. However, it is slower than LS and there is relatively less to be gained from automation of the sample preparation. The P.A.R. instrumentation used in this project cost three times as much as the Chemtrix instrument; but the price differential might be sub- stantially reduced by substitution of the P.A.R. Model 174 specialized analytical version for the Model 170. The Model 174 provides single- and dual-cell PD and LS capabilities at a price competitive with other LS instruments. The Chemtrix Model SSP-5A is a compact, special-purpose instrument that provides as wide a range of programming variables as one is likely to need for the LS technique. The major problem in obtaining reproducible results appeared to be in setting the initial potential for the voltage scan. The small meter and manual potentiometer on the instrument made it difficult to set this value much closer than 0.1 V. The setting seemed to be reasonably stable when left untouched for repetitive measurements and it should be a simple matter to provide a fine-tuning adjustment to simplify arbitrary changes. The Chemtrix drop timer and cell stand, in comparison with the P.A.R. model, was crude. The relay-actuated drop knocker was noisy and imparted considerably turbulence to the solution in the cell. Drop times were continuously adjustable by means of a potentio- meter control and were not re-settable without a stop watch. The mercury stand did not provide for varying the height of the mercury column, except for a very 14 ------- small differential adjustment between the cells. Our Chemtrix instrument had a poor repair record. In the first twelve months transistors in the drop timer circuit were replaced three times. The plug-in amplifier module was replaced by the manufacturer four times and components of the amplifier were replaced in the field twice. Eventually, the entire instrument, including oscilloscope, was recalled by the factory for repair and calibration. The great versatility of the Model 170 Electrochemistry System is provided at the cost of a rather bewildering set of controls. For routine measurements of the same type, however, only a few need to be adjusted indivi- dually. The effects of changing the available instru- mental parameters were not investigated fully or systematically, although some such studies have been reported (33). Resolution can be improved at the cost of lengthening the analysis time by decreasing the potential scan rate. The analytical signal can be increased by increasing the pulse amplitude or the time of current sampling, both of which cause distortion and poorer resolution, and by increasing the drop time, which lengthens the analysis time. The major operating problems involved sensitivity of detection. The nanoampere range of the Model 170 was nearly unusable because of a high noise level. In dual cell operation, there was an optimum pre-amplifier gain setting in the 174/51 module and an optimum current range setting on the Model 170. Also, the current offset and electrode balance controls, in dual-cell operation, required painstaking adjustment to avoid amplifier saturation. The Model 170 was very reliable electronically, and no troubles developed in two years. After about three months of operation with the 174/51 dual-cell attachment, problems arose that seemed to involve both components. The routine trouble-shooting checks were inadequate to identify the source of malfunction and the entire instrument was returned to the manufacturer. 15 ------- INDIUM Calibrations A number of calibration runs made under different con- ditions are summarized in Table 2. These include com- parisons of seawater and sodium chloride as supporting electrolytes. Most were made by the PD technique, with a single cell. Each calibration run is internally consistent with respect to operating conditions but different runs are not necessarily comparable in sensi- tivity . The parameters of mercury flow rate and potential scan rate, in particular, were not always the same. The effect of scan rate on the observed current (33) was not fully recognized at the time, but rates of 2 or 5 mV/sec were usually employed. The current is directly proportional to the area of the mercury drops, and a compromise can be made between sensitivity and analysis time most directly by varying the drop time. In these experiments the mercury flow rate was usually in the range 0.5 to 1.0 mg/sec and the drop time was one second. Because of the above variations in experimental condi- tions the calibrations are best compared in terms of the linearity and scatter of the points. Table 2 gives the parameters for least-squares fits to the experimental points by the equation J(nA) = a + b X(ppm NTA) (1) or, in the case of calibration by standard addition, J(ppm found) = a + b X(ppm taken) (2) When the intercept a is enclosed in parentheses its value is smaller than its standard deviation. The table shows the effects of improved technique, since the relative standard deviations were about 10% for the first two runs and decreased to 1% in Run 9. Runs 3 and 4 were a direct comparison of sodium chloride and seawater electrolytes run at the same time; the slope of the calibration curve was anomalously low in sodium chloride. 16 ------- Table 2 Calibration Runs for the Determination of NTA as the Indium Complex n, No. Concentration Std. Dev. Run of Points Range (ppm NTA) a(nA) b (nA/ppm of Fit(nA) Remarks 1 2 3 4 5 6 7 8 9 10 14 9 4 4 6 6 6 5 5 7 0 0 0 0 0 0 0 0 0 0 .2 - .2 - .2 - .2 - .2 - .2 - .2 - .1 - .1 - .1 - 2. 2. 2. 2. 1. 1. 1. 1. 1. 3. 0 0 0 0 2 2 2 0 0 2 (-5) -28 0.09 0.07 (7) (0.02) (0.03 (0.4) (-1.0) (8) 181 354 0.77 0.98 293 0.99 1.00 94 272 152 14 44 0.07 0.08 19 0.05 0.05 1.8 2.8 21 NaCl and seawater, combined , PD Seawater , NaCl , std. Seawater , PD NaCl Na5P NaCl NaCl NaCl NaCl 60% + 1.6 3°10' , std. , std. , LS , PD Sewage PD addn . , PD std. addn. , x 10~5 M PD addn. , PD addn. , PD effluent, LS 11 7 0.1 - 3.2 26 817 40 60% Sewage effluent, PD ------- The same solutions were used in Runs 5, 6, and 7. The method of standard additions was compared for two different operators in Runs 6 and 7, while a calibration curve was computed for the original solutions in Run 5. In this case, the relative standard deviation is lower for the addition technique (5%) than for the calibra- tion curve (6.3%) . Runs 8 and 9 are a direct comparison of the single-cell PD (P.A.R.) and LS (Chemtrix), both in sodium chloride electrolyte. The relative standard deviation/ at 1 mg NTA/&, is 1% for the former and 2% for the latter. In 60% sewage effluent both techniques were much poorer (runs 10 and 11), with relative standard deviations of 5% for PD and 14% for LS. Two chart recordings from Run 11 are shown in Figure 2. Interferences The response to NTA in the range of 0.4 mg/& was not significantly affected by EDTA, citric acid, or NaH2PO4 in varying proportions up to four times the concentration of NTA. The total indium concentration was 2 x 10~5 M in all cases. Table 3 shows recoveries (the fraction, in percent, of the concentration of the added species that was indicated by the analysis) of randomized concentrations of NTA from 0 to 2.0 mg/Jl in the presence of 5 x 10~5 M Fe(III), Mn(II), Cu(II), or Bi(III). Under these conditions bismuth appeared to complex a significant portion of the NTA, leading to low results. The effect of interferences might be mitigated by using the standard addition method of calibration. This approach assumes that a linear response to NTA would be observed, although the interferences might change the slope of the calibration curve. A test is summarized in Table 4, where 5 lag of NTA was added to 10 ml of the original solution. The initial concentrations of NTA were selected in a randomized order in the range 0.2 to 2.0 mg/Jl. The trials of Fe(III) plus hydroxylamine were made at a constant level of NTA. A least-squares fit of the data gave the equation. Y(ppm NTA found) = 0.11 + 0.95 X(ppm NTA taken) (3) with a standard deviation of fit of 0.14 ppm. 18 ------- Table 3 Recovery of NTA in the Presence of 5 x 10 Fe, Mn, Cu, or Bi Fe(III) MnJ11) Cu(II) Bi(III) TA Taken 0 . 8 ppm 1.2 1.4 1.6 Found 0.86 1.18 1.36 1.50 Taken 0.4 1.8 2.0 Found 0.35 1.59 1.96 Taken 0.8 1.0 1.0 1.6 Found 0.69 0.94 0.81 1.61 Taken 0.2 1.4 1.4 1.8 Found 0.02 1.04 0.84 1.62 ------- Table 4 Recovery of In-NTA in the Presence of Metals and Coroplexants L Taken mg/fc) 1.6 1.6 1.6 1.8 0.2 1.8 1.4 0.6 1.2 0.6 1.2 1.0 1.0 1.0 1.0 0.4 1.0 0.8 1.4 NTA Found (rog/A) 1.58 1.55 1.53 2.03 0.49 L98 1.67 0.72 1.16 0.59 .58 1.05 0.81 0.99 1.13 0.43 1.61 0.88 1.14 Interference Cu Cu Zn Zn Mn(II) Mn(II) Bi Fe(II) Fe(II) Fe(III) Fe(III) N2H4-HC1 N2H4+Fe(III) N2H4+Fe(III) N2H4+Fe(III) EDTA EDTA Na5P3Oi0 NasPqOin Concentr, x!06, 8 16 4 8 4 8 2 8 16 8 16* 40 12* 16 32 8 16* 8 16 * More than twice the standard deviation from the least-squares fit. 20 ------- Three solutions, indicated in Table 4, deviated by more than twice the standard deviation and were not included in the calculation. The least-squares equation shows that the standard addition method tended to give high results, especially at lower concentrations, and that the precision was rather poor (14% relative standard deviation). Both results may be a consequence of a non-linear response to NTA in the presence of some or all of the additives. Fe(III) interfered at 1.6 x 10~5 M but hydrazine effec- tively removed the interference. The high result for 1.6 x 10~5 M EDTA is a consequence of the standard- addition technique. After the addition of 0.5 ppm NTA the sum of NTA and EDTA concentrations was in molar excess of the total indium (2 x 10~5 M), causing a lower response for the standard addition. Since Haberman (20) observed that a large excess of indium had to be added to some sewage samples to obtain a polarographic wave for free In, a few titrations were made. Indium solutions of 2 to 8 x 10~^ Af concentra- tion were prepared in three distilled water and three sewage effluent samples (15 ml of sewage effluent in a final volume of 25 ml) and the excess In was titrated with NTA by following the decrease in height of the In(III) wave in dual-cell LS. Although two of the effluent samples had very little excess In (2 x 10~5 M total concentration) the results indicated 1.7 to 1.9 x 10~5 M oomplexed indium, corresponding to an indium-complexing power of the original effluent of 3 x 10-5 M. As noted by Haberman (20) NTA did not release In from this complex in a polarographically determinable form. The mean slopes of the three titrations in distilled water corresponded to 1.7 moles of In removed per mole of NTA. The corresponding value in the effluent samples was 1.4 moles of In per mole of NTA. These numbers probably do not represent a stoichiometric complex, but reflect an accumulation of NTA in the double layer surrounding the mercury drop. Staroscik and Webs (28) observed that, for an increasing excess of NTA, the In-NTA wave decreased in height and eventually was transformed to an ill-defined adsorption wave. 21 ------- BISMUTH, CHLORIDE MEDIUM Calibrations Since an analytical procedure had already been developed (17) and applied to seawater samples (31), relatively little was done in the way of evaluating specific interferences or in modifying the procedure. Calibration runs were made primarily to evaluate the instrumentation, particularly for PD. Five such calibrations, made early in the project, are summarized in Table 5, in which the parameters for least-squares fits of equation 1 are shown in the same format as in Table 4. The most noteworthy point is the great varia- tion in precision. The relative standard deviations range from about 1% (Run 1) to 13% (Run 4). The poor precision is believed to result more from unfamiliarity with the instrument than from the chemical procedure. Typical chart recordings, taken from Run 3, are shown in Figure 3. Interferences Although Afghan et al. (17) found that copper did not interfere in the determination of the Bi-NTA complex and that Bi quantitatively displaced Cu from its NTA complex at 2 x 10"^ M concentration, other workers have not confirmed these results. Gahler (31) found that in the presence of 2 mg Cu/fc, the concentration of NTA indicated by a dc polarographic measurement was high, and Taylor et al. (6) , using the same type of instru- mentation and procedure as Afghan (17) observed an interference in tap water that they believed to result from copper. Figure 4 shows traces obtained by LS of solutions containing NTA and Cu. Curves 1, 2, and 3 correspond to solutions of 2 mg NTA/& and curve 4 to a 1 mg NTA/£ solution. Curves 3 and 4 were obtained with 1 x 10-5 M Cu, Curve 2, with 0.5 x 10~5 M Cu, and Curve 1, with no copper. The test and reference solutions all were 0.1 M in NaCl, and 1 x 10~5 M in Bi at pH 2.0. A better test of the analytical procedure would have been obtained with NTA and Cu, but no Bi, in the appro- priate reference solutions. These results, however, show that the copper peak precedes and interferes with the Bi-NTA wave. There is no indication that copper 22 ------- Table 5 Calibration Runs for the Determination of NTA as the Bismuth Complex in Chloride Electrolyte to U) Run 1 2 3 4 5 n, No. of Points 5 9 8 9 9 Concentration Range (ppm NTA) 0.2 - 1.6 0.2 - 0.2 - 0.2 - 0.2 - 1.8 2.0 2.0 1.0 a(nA) -68 (-18) (-6) (-1) (9) b (nA/ppm) 648 375 374 395 209 Std. Dev. of Fit(nA) Remarks 8 24 17 52 7 NaCl (PD) NaCl (PD) Seawater (PD) NaCl Seawater (LS) ------- BISMUTH - NTA IN SEAWATER TECHNIQUE: SINGLE P D MEDIUM: CHLORIDE 00 nA I.36 ppm NTA 0.7 ppm NTA -0.3 -0.5 V 0 -0.2 -0.4 V Figure 3. Determination of the bismuth-NTA complex in seawater by pulse differential polarography 24 ------- Figure 4. Copper and bismuth-NTA waves by linear sweep voltammetry 25 ------- complexes any NTA at pH 2.0. When the method was tried in waste waters difficulties appeared, particularly in trying to match the sample and reference scans in the LS technique. Figure 5 shows some typical results obtained with the Afghan (17) procedure in an effluent from a small activated-sludge plant serving a residential development in Clarke County, Georgia. The fresh sample was filtered through 0.45 y Millipore but no preservative was added. Figure 5(a) shows the sample (A) and reference (B) scans, as well as the difference, for no added NTA; thus, the only difference between the cells is 1 x 10~5 M Bi added to cell A. Figures 5(b) and 5(c) were taken at a later time but for the same effluent and procedure. Curves 4, 1, and 2, respectively, of Figure 5(b) are the difference traces for 0.5, 1.0, and 2.0 mg NTA/£. In Figure 5 (c) , curves A and B are for sample and reference, respectively, and 1 is the difference, for 0.2 mg NTA/& in A, while Curve 2 is the difference trace for 1.0 mg NTA/£. For more than about one ppm of NTA fairly satisfactory traces and calibration curves were obtained, but at lower concentrations of NTA the problem of matching the cells was more severe. This difficulty in determining low levels of NTA in natural or waste waters is further illustrated in Figure 6, which shows calibration curves obtained in the above sewage effluent and in Oconee River water, by both LS and PD techniques. The slopes of the curves are strikingly different for the two sources, showing that a calibration based on distilled water standards would be highly inaccurate for at least one of the samples. Because of the non-zero intercepts of the curves in Figure 6, the method of standard additions would not be a great improvement and could not be used for low levels of NTA. The parameters of the least- squares equations for these calibrations are given in Table 6. Interference from sulfur compounds in some sewage samples has been reported (17) , and treatment with hydrogen peroxide was recommended. We were not able to achieve any improvement through hydrogen peroxide treatment and found it difficult to remove the excess 26 ------- M^HB^^^^Bl^^^W. f^f mmm Figure 5. Effect of sewage effluent on the bismuth-NTA polarographic wave in chloride electrolyte 27 ------- LS nA 100 50 0.5 1.0 I PD nA 400 200 1.5 ppm NTA Figure 6. Calibration curves for the determination of the bismuth-NTA complex — chloride electrolyte 28 ------- Table 6 Calibration Runs for Bi-NTA in Chloride Electrolyte- Sewage Effluent and River Water rel. std. dev. at 1 ppm NTA Oconee River, PD Technique Y(nA) = (-18±4) + (286±5) X (mg NTA/£), 1.7% Oconee River, LS Technique Y(nA) = (-16±2) + (95±2) X (mg NTA/fc) , 2.4% Sewage Effluent, PD Y(nA) = (-32±7) + (175±8) X (mg NTA/£), 5.3% Sewage Effluent, LS Y(nA) = (-13±3) + C69.0±3.6) X (mg NTA/&), 6.1% 29 ------- peroxide sufficiently. About this time we discovered that an acetate electrolyte gave greatly superior results in waste waters and we discontinued investigation of the chloride medium. BISMUTH, ACETATE MEDIUM Calibrations Calibrations in the acetate medium are summarized in Table 7. The experiment of Run 1 was a balanced, incomplete block design for testing four levels of NTA in four sources of water, in duplicate. The standard deviation of the duplicates was 13 nA, including one pair that differed by 43 nA (the outlier was not discarded). No significant effects were found for the water sources, which included distilled, tap, river water and sewage effluent. Accordingly, the results were pooled for the calibration shown in Run 1. Another similar balanced, complete block design was used to compare the LS and PD techniques using six sources of water: distilled, tap, woodland stream, Oconee River, and effluent from a small activated sludge treatment facility and from a municipal sewage treatment plant. No significant differences due to source were observed in the LS mode, and the pooled results are summarized in Run 2. For the PD results source effects were significant at the 95% level of significance and the data have been subdivided. The potable and natural waters are combined in Ran 3 and results for the two effluents are summarized in Run 4. The effluents gave a somewhat lower response to NTA than the "clean" waters but gave somewhat better precision. A direct comparison of calibration in distilled water and in sewage effluent, both by PD, is shown in Runs 7 and 8. These were made at higher concentrations of NTA and the bismuth concentration was increased to 5 x 10~5 M. Both lines show negative intercepts (a) greater than their standard deviations, corresponding to about 0.2 mg NTA/Jl, but the slopes of the calibration curves are essentially identical and the precision is somewhat better in the effluent. The much lower response factor (2>) in comparison with Runs 1, 3, or 4 results from a change of capillaries, differing by a factor of nearly three in mercury flow rate. 30 ------- u> Table 7 Calibration Runs for the Determination of NTA as the Bismuth Complex in Acetate Electrolyte n, No. Concentration Std. Dev. Run of Points Range (ppm NTA) g(nA) b (nA/ppm) of Fit (nA) Remarks 1 2 3 4 5 6 7 8 24 24 16 8 8 5 4 4 0.05 - 0.2 - 0.2 - 0.2 - 0.05 - 0.2 - 1 1 0.50 1.5 1.5 1.5 0.2 0.8 4 4 (1.5) (0.1) (3) (3) 2.9 -22 -35 -30 313 77.8 339 308 40.6 347 164 162 11 4.5 20 7 0.9 3 19 4 PD; four types of water source LS (Chemtrix) ; six types of water source PD; four types of "clean" water PD; two sewage effluents LS (Chemtrix) ; distilled water PD; distilled water plus NaClO4 PD; distilled water PD; sewage effluent ------- w N) Table 7 (cont'd) Calibration Runs for the Determination of NTA as the Bismuth Complex in Acetate Electrolyte n, No. Concentration Std. Dev. Run of Points Range (ppm NTA) a(nA) b(nA/ppm) of Fit (nA) Remarks 9 10 11 12 13 5 7 10 9 12 0.01 - 0.01 - 0.01 - 0.02 - 0.02 - 0.12 0.1 0.1 2.0 2.0 (-.1) -1.0 0.6 -26 (5) 354 99.0 242.4 669 857 0.6 0.5 0.8 39 23 Dual PD; distilled water LS (PAR) ; distilled water Dual PD; stream water PD; sewage effluent Dual PD; sewage effluent ------- When the dual-cell modification of the P. A. R. Model 170 was obtained a comparison of LS and PD modes on the same instrument was made in Runs 10 and 11. Single-cell and dual-cell PD operation were compared in Runs 12 and 13. The high precision attainable with dual-cell PD is indicated in Runs 9, 11, and 13. Chart records for the complete calibration Run 9 are shown in Figure 7, and for the two lowest concentrations of Run 11 in Figure 8. A demonstration of the presence of NTA in raw and treated municipal sewage is shown in Figure 9. Samples from the Athens Oconee treatment plant were preserved with 1% formaldehyde and filtered through 0.45 u Millipore. The samples were analyzed by the dual-cell PD technique in acetate electrolyte. The raw sewage gave a peak of about 3.0 nAf which increased to 21.5 nA when the cell was spiked with 0.1 mg NTA/£. The effluent sample gave a peak of 2.0 nA, increasing to 19.5 nA when spiked with 0.1 mg NTA/fc. These results indicate 16 and 11 yg/&, respectively, in the cells, or about 30 and 20 yg NTA/£ in the original raw and treated samples. Since the standard deviation of the corre- sponding calibration was equivalent to 4 yg NTA/&, the difference between the samples may not be real, although some degradation of the NTA in the treatment process would be expected. Instrumental problems prevented confirmation or further study of NTA in natural samples. Interferences In preliminary studies of the effects of metal ions on the determination of NTA, Fe(III) and Cu(II) seemed to interfere most significantly. Heating the sample with at least 0.02% of either hydrazine sulfate or hydro- xylamine hydrochloride, as in the chloride procedure (17), effectively removed all interference from iron at concentrations up to 1 x 10~5 M (0.5 ppm). In the presence of 1 x 10"^ M Cu, NTA recoveries were about 80-90% and were quite variable. Although the problem seemed to be associated with low concentrations of chloride ion, it was not completely explained or solved. In the acetate medium, free of chloride, separate waves were observed for Cu(II), Bi(III), and Bi-NTA at half-wave potentials of +0.07, -0.005, and -0.16V (vs. SCE), respectively- In the presence of 33 ------- ppm NTA 0. 12 BISMUTH - NTA CALIBRATION IN DISTILLED WATER TECHNIQUE: DUAL P D MEDIUM: ACETATE 10 nA 0.01 -0.3 -0.5 V (Hg pool) Figure 7. Determination of the bismuth-NTA complex in distilled water by dual cell pulse differen- tial polarography 34 ------- BISMUTH - NTA IN STREAM WATER TECHNIQUE^ DUAL P D MEDIUM: ACETATE I nA 0.02 ppm NTA 0.01 ppm NTA -0.3 .-0.4 V -0.4 -0.5 V Figure 8. Determination of the bismuth-NTA complex in distilled water by dual cell pulse differential polarography 35 ------- Raw Sewage Treated Effluent 21.5 nA 19.5 nA Raw plus 0.1 ppm NTA Effluent plus 0.1 ppm NTA Figure 9. Differential pulse polarographic estimation of NTA in raw and treated municipal sewage 36 ------- 0.02% hydroxylamine hydrochloride (3 x 10 3 M) or 0.1 M sodium chloride, the Cu(II) wave at +0.07V disappeared and was replaced by a very large wave at +0.21V, which may be due to mercury oxidation. These potentials were measured with a mercury-mercurous sulfate reference electrode having a potential of +0.360V vs. SCE. EDTA was tried as a masking agent for copper but had no effect, probably because bismuth was able to displace copper from the EDTA complex. Both copper and chloride might be expected in natural and waste waters at the concentrations used in these studies and the analyst should be aware of the possi- bility of low results for NTA due to copper. The cali- bration results reported in the preceding section, however, did not reveal any problems in applying the method to a variety of water sources. A number of possible interferences were investigated by single-cell PD at NTA concentrations of 0.2, 0.5, and 1.0 mg/&. The average recovery was calculated from the slope of the best calibration curve through the origin for each set of NTA concentrations, relative to standards prepared in the same Oconee River water but without added metals or complexants. The results are summarized in Table 8. Lead, at 1 x 10~4 M concentra- tion, caused a low recovery of NTA but did not inter- fere significantly at 1 x 10~5 M. High concentrations of calcium and magnesium (400 mg/£) also led to low recoveries of NTA but no other interferences were found. Citric acid, at 1 x 10~5 M concentration, was included in several of the preliminary tests with no perceptible effect on the response to NTA. 37 ------- Table 8 Recovery of Bi~NTA in the Presence of Metals and Complexants in Acetate Electrolyte Average Recovery, Interference Percent Fe(III), lxlQ-5 M 101 Fe(IIl) + Mn(II), each lxlQ-5 M 100 Fe(III) + NasPaOio, each lxlO~5 M 97 Mn(II) , 1x10-4 M 303 Pb(II), IxlO-4 M 78 Pb(II), IxlO-5 M 96 Cd(II), 1x10-4 M 98 Zn(II), 1x10-4 M 99 Ca(II), 400 mg/1 (1.0xlO~2 M) 49 Mg(II), 400 mg/1 (1.7xlO~2 M) 72 P043-, 1x10-5 M 103 Na5p3°10f 1x10-4 M 97 38 ------- SECTION V REFERENCES 1. Longbottom/ J. E.. Determination of Nitrilotriacetic Acid by High-Speed Ion Exchange Chromatography. Anal. Chem. 44: 418-420, February 1972. 2. Swisher, R. D., M. M. Crutchfield, and D. W. Caldwell. Biodegradation of Nitrilotriacetate in Activated Sludge. Environ. Sci. Technol. !_: 820- 827, October 1967. 3. Pfeil, B. H. and G. F. Lee. Biodegradation of Nitrilotriacetic Acid in Aerobic Systems. Environ. Sci. Technol. 2: 543-546, July 1968. 4. Thompson, J. E. and J. R. Duthie. The Biodegradability and Treatability of NTA. J. Water Pollution Control Fed. 40; 306-319, February 1968. 5. Methods for Chemical Analysis of Water and Wastes, 1971. Analytical Quality Control Laboratory, Cincinnati, Ohio. Report No. 16020-07/71. Environmental Protection Agency. 1971. 205-215. 6. Taylor, J. K., W. L. Zielinski, Jr., E. J. Maienthal, R. A. Durst, and R. W. Burke. Development of Method for NTA Analysis in Raw Water. National Bureau of Standards. Report EPA-R2-72-057. Environmental Protection Agency. 1972. 19-21. 7. Rechnitz, G. A. and N. C. Kenny. Determination of Nitrilotriacetic Acid (NTA) with Ion-Selective Membrane Electrodes. Anal. Lett. 3_: 509-514, October 1970. 8. Rudling, L. Determination of Nitrilotriacetic Acid. Water Research. 5_: 831-838, October 1971. 9. Chau, Y. K. and M. E. Fox. Nitrilotriacetic Acid in Lake Water. J. Chromatogr. Sci. 9: 271-275, May 1971. 10. Warren, C. B. and E. J. Malec. Quantitative Deter- mination of Nitrilotriacetic Acid and Related Amino- polycarboxylic Acids in Inland Waters. J. Chromatogr. 6±: 219-237, February 1972. 39 ------- 11. Longman, G. F-, M. J. Stiff, and D. K. Gardiner. The Determination of Nitrilotriacetic Acid (NTA) in Sewage and Sewage Effluent. Water Res. 5: 1171-1175, December 1971. ~ 12. Hoyle, W., I. P- Sanderson, and T. S. West. Polarographic Resolution of Mixtures of Complexans. J. Electroanal. Chera. _2: 166-173, 1961. 13. Farrow, R. N. P. and A. G. Hill. A Modified Method for Determining Traces of Nitrilotriacetic Acid in Ethylenediaminetetraacetic Acid. Analyst 90: 241- 242, 1965. 14. Daniel, R. L. and R. B. LeBlanc. Polarographic Determination of Nitrilotriacetic Acid in (Ethylenedinitrilo)tetraacetic Acid. Anal. Chem. 31: 1221-1223, 1959. 15. Ewing, G. W. General-Purpose Electroanalytical Instruments. J. Chem. Ed. 46; A717 et seq., 1969. 16. Afghan, B. K., P. D. Goulden, and J. F. Ryan. Automated Method for Determination of Nitrilotri- acetic Acid in Natural Water, Detergents, and Sewage Samples. Anal. Chem. 44; 354-359, February 1972. 17. Afghan, B. K. and P. D. Goulden. Determination of Trace Quantities of Nitrilotriacetic Acid by Differential Cathode-Ray Polarography. Environ. Sci. Technol. 5_: 601-606, July 1971. 18. Koryta, J. and I. Koessler. The Polarographic Determination of the Stability Constants of the Complexes Formed by Some Heavy Metals with Schwarzenbach's Complexones. Collect. Czech. Chem. Commun. 15; 241-259, 1950. 19. Sillen, L. G. and A. E. Martell. Stability Constants of Metal-Ion Complexes. Special Publ. No. 17. London, The Chemical Society, 1964. 20. Haberman, J. P. Polarographic Determination of Traces of Nitrilotriacetate in Water Samples. Anal. Chem. 43: 63-67, January 1971. 40 ------- 21. Wernet, J. and K. Wahl. Spezifische Bestimmung von Nitrilotriessigsaeure (NTE) in ppm-Konzentrationen in Oberflaechenwasser und Abwasser. (Specific Determination of Nitrilotriacetic Acid (NTA) in the ppm-Range in Surface Water and Sewage). Fresenius1 Z. Anal. Chem. (Berlin). 251; 373-374, 1970. 22. Asplund, J. and E. Wanninen. Polarographic Determination of Low Concentrations of Nitrilotri- acetate in Lake Water. Anal. Lett. 4: 267-275, May 1971. 23. Al-Sulimany, F. and A. Townshend. Polarographic Determination of Tripolyphosphate Ions and of Tripolyphosphate and Nitrilotriacetic Acid in Admixture. Analyst (London). 98; 34-39, 1973. 24. Afghan, B. K. Differential Cathode Ray Polarography for Trace Analysis with Special Refer- ence to NTA and its Complexes with Heavy Metals. Department of Fisheries and Forestry, Canada. (Presented at International Symposium on Identifi- cation and Measurement of Environmental Pollutants. Ottawa. June 15, 1971), 19 p. 25. Bulovova, M. Polarography of Indium. Collect. Czech. Chem. Commun. 19: 1123-1132, 1954. 26. deLevie, R. and A. A. Husovsky. On the Negative Faradaic Admittance in the Region of the Polaro- graphic Minimum of Indium(III) in Aqueous NaSCN Solution. J. Electroanal. Chem. 22; 29-48, 1969. 27. Pospisil, L. and R. deLevie. Thiocyanate Electro- catalysis of the Reduction of Indium(III). J. Electroanal. Chem. 25; 245-255, 1970. 28. Staroscik, R. and K. Webs. Polarographic Behavior of Indium(III) in Nitrilotriactic Acid Solutions. Chem. Anal. (Warsaw). 12; 1275-1281, June 1967; Chem. Abstr. 68; 65177 m, 1968. 29. DeSesa, M. A., D. N. Hume, A. C. Glamm, Jr., and D. D. DeFord. Polarographic Characteristics of Metallic Cations in Acetate Media. Anal. Chem. 25; 983-984, June 1953. 41 ------- 30. Malachesky, P- A. Correlation of Linear Sweep Voltammetric and Chronoamperometric Data for n-Value Determinations. Anal. Chem. 41; 1493-1494, September 1969. 31. Gahler, A. R., Pacific Northwest Water Laboratory, EPAr personal coiraminication. 1971. 32. Bond, A. M. The AC and DC Polarographic Reduction of Bismuth (III) in Acidic Halide and Other Media. Electrochim. Acta. 17; 769-785, April 1972. 33. Christie, J. H., J. Osteryoung, and R. A. Osteryoung. Instrumental Artifacts in Differential Pulse Polarography. Anal. Chem. 45: 210-215, January 1973. 42 ------- SECTION VI APPENDIX RECOMMENDED NTA ANALYSIS PROCEDURES SALINE WATERS - MORE THAN 0.1% CHLORIDE Reagents 1 N Hydrochloric Acid 1 M Sodium or Potass ium Chloride 1 x 10~3 M Bi: Dissolve 50 ± 2 mg Bi(NO3)3-5H2O in 100 ml 1 N HC1. 1% Hydroxylamine Hydrochloride NTA stock solution, 2 g/1. Dissolve 1.434 g H0 in 500 ml distilled water. Procedure Place 5.0 ml of sample in each of two 10-ml volumetric flasks and add 0.2 ml of 1% hydroxylamine hydrochloride to each. Add 0.1 ml of 1 N HC1 to the reference and 0.1 ml of 1 x 10~3 M Bi to. the sample flask. If the original sample contained less than about 0.5% chloride, add 1.0 ml of 1 M chloride solution to each flask and dilute to the mark with distilled water. If the original water sample contained more than about 0.5% chloride do not add additional salt solution, but dilute to the mark with distilled water. Place 3.0 ml of sample and reference solutions in the respective cells, de- aerate for 10 minutes , and run a polarographic scan from about -0.2 to -0.4 V vs. the mercury pool anode. Compare the peak currents with those for standard solutions similarly prepared with freshly diluted NTA stock solution. The procedure is applicable to NTA concentrations from 0.05 to 5.0 mg/1. 43 ------- NATURAL AND WASTEWATERS - LESS THAN 0.1% CHLORIDE Reagents Acetate electrolyte, 9.5 M in acetic acid plus 0.5 M in sodium acetate. 1 x 10-3 M Bi: Dissolve 50 + 2 mg Bi(N03)3-H2O in 100 ml of the acetate eTectrolyte. 1% Hydroxylamine sulfate solution NTA stock solution, same as above. Procedure Place 5.0 ml of sample in each of two 10-ml volumetric flasks, add 0.2 ml of 1% hydroxylamine sulfate solution to each and heat nearly to boiling. Cool to room temperature and add 0.1 ml of acetate electrolyte to the reference and 0.1 ml of 1 x 10~3 M bismuth solution to the sample. Dilute each to the mark with distilled water, mix, and place 3.0 ml in the respective cells. De-aerate for 10 minutes and run a polarographic scan from -0.1 to -0.3 V (vs. SCE). The mercury pool anode can be used as a reference but its potential is apt to vary with different sample sources. It is preferable to use an external reference electrode with a sodium perchlorate salt bridge to both cells. Compare the peak currents with those for standard solutions similarly prepared from freshly diluted NTA stock solution. The procedure is applicable to NTA concentrations from 0.01 to 5.0 mg/1. 4 U. S. GOVERNMENT PRINTING OFFICE : 1973—514-156/365 44 ------- SELECTED WATER RESOURCES ABSTRACTS INPUT TRANSACTION FORM l.jReportffo. 3. Accession No. w 4. Title POLAROGRAPHIC DETERMINATION OF NTA 7. AuthoT(s) Hoover, Thomas B. 5. Report Date 6. 8. Performing Organization Report No. 9. Organization Southeast Environmental Research Laboratory Athens, Georgia 30601 Project No. EPA, 16020-EWE 11. Contract/Grant No. 13 Type ••{ Repoi t and Period Covered 15. Supplementary Notes Report EPA-R2-73-254, June 1973. 9 fig, 8 tab, 33 ref. 16. Abstract Pulse differential polarography was found to be a more sensitive and precise method than linear sweep voltammetry for the determination of NTA as the bismuth complex in natural and waste waters. Indium was a less satisfactory complexant. Relative standard deviations of less than 5% were obtained by the dual cell pulse differential technique at 1 ppm NTA and calibrations were linear from 0.01 to 4 ppm. Copper, added at concentrations greater than 10-5 M, caused low analytical results for NTA. No other significant interference by metals or complexants was found. An acetate electrolyte was much better than chloride for the determination of bismuth-NTA complex in sewage-type samples. Recommended procedures are given in the Appendix. 17a. Descriptors *Analytical Techniques, *Polarographic Analysis, *Nitrilotriacetic Acid, Sea Water, Sewage, Chemical Analysis, Water Analysis 17b. Identifiers *Bismuth-NTA, *Indium-NTA, Linear Sweep Voltammetry, Differential Pulse Polarography lie. COWRR Field & Group 05A 18. Availability 19. Security ~ (Repots) *0. Security C/.ss. (Page) .-..- Abstractor T. B. Hoover Send To: WATER RESOURCES SCIENTIFIC INFORMATION CENTER U.S. DEPARTMENT OF THE INTERIOR WASHINGTON. D. C. 2O24O institution EPA, Southeast Env. Res. Laboratory WRSIC 1O2 iKEV.JUN? 1971) ------- |