Anal. Chem. 1980, 5 2 , 1013-1020
vitamin B12ain 0.5 M Na2S04.0.04 M phosphate buffer (pH 5.15) on gold and Hg-Au wire electrodes. The gold wire working electrode shows the well defined, irreversible Co(III)/Co(II) curves around 0.0 V. This wave is not as well defined a t the Hg-Au electrode, but the totally irreversible Co(II)/Co(I) wave a t -0.8 V is easily seen. This thin-layer electrochemical behavior of vitamin B12acorresponds exactly to that described in the literature (13, 20, 21). In this case less than 0.2 pg of BIzawas employed in the experiment.
LITERATURE CITED (1) Caja, J.; Czerwinski, A.: Mark, H. B., Jr. Anal. Chem. 1979, 57, 1328. Hubbard, A. T.; Anson, F. C. Anal. Chem. 1964, 36, 723. McClure, J. E.; Maricle, D. L. Anal. Chem. 1967, 39, 236. Christensen, C. R.; Anson, F. C. Anal. Chem. 1963, 35, 205. Sheafer, J. C.: Peters, D. G. Anal. Chem. 1970, 4 2 , 430. Anderson, L. B.: Reilley, C. N. J . Nectroanal. Chem. 1965, 70, 295. Propst, R . C. Anal. Chem. 1971, 4 3 , 994. Murray, R. W.; Heineman, W. R.; O'Dorn, G. W. Anal. Chem. 1967, 39,
(2) (3) (4) (5) (6) (7) (8)
1066. (9) Hubbard, A. T. Crif. Rev. Anal. Chem. 1973. 3 , 201.
1013
(IO) LeCuire, J. M.; Pillet, Y . J . Electroanal. Chem. 1978, 97, 99. (11) Tom, G. M.; Hubbard, A. T. Anal. Chem. 1971, 4 3 , 671. (12) Meyer, M. L.; &Angelis, T. P.: Heineman, W. R. Anal. Chem. 1977, 49, 602. (13) Kenyhercz, T. M.; Mark, H. B., Jr. J . Hectrochem. SOC. 1976, 123, 1656. (14) Will, F. G.; Knorr, C. A. Z.Hecfrochem. 1960, 6 4 , 270. (15) Ferro, C. M.; Calandra, A. J.; Arvia, A. J. J . Electroanal. Chem. 1974, 5 0 , 403. (16) Nowak. R. J.; Kutner, W.: Mark, H. B., Jr. J . Electrochem. SOC.1978, 125, 232. (17) DeAngelis, T. P.: Heineman. W. R . J , Chem. Educ. 1976, 53, 594. (18) "Handbook of Chemistry and Physics"; CRC Press: Cleveland, Ohio, 1976: 57th Ed., D-141. (19) Harrar, J. E.: Shain, I . Anal. Chem. 1966, 38, 1148. (20) Kenyhercz, T. M.; DeAngeiis, T. P.; Norris, B. J.; Heineman, W. R.; Mark, H. B., Jr. J . Am. Chem. SOC.1976, 98, 2469. (21) Lexa, D.: Saveant, J. M.; Zickier, J. M. J . Am. Chem SOC.1977, 107, 2786.
for review December 3, 19'79. Accepted March 17, 1980. This research was supported in part by National Science Foundation Grants NSF CHE 76-04321 and CHE 77-04399. RECEIVED
Determination of Nickel by Differential Pulse Polarography at a Dropping Mercury Electrode Carmen J. Flora and Evert Nieboer Department of Chemistry, Laurentian University, Sudbury, Ontario, Canada, P3E 2C6
Dimethylglyoxime (DMG) sensitized the differential pulse polarographic determination of NiZf. The detection limit was 2 ppb and the linear response in the concentration range 0-85 ppb could be extended to ppm levels by simply omitting DMG. Actual analyses of tap and lake water and plant tissue are reported. Characterization of the electroactive process included an examination of the degree of reversibility and the effect on the peak current of pH, buffer composition, DMG concentration, and the presence of other metal ions.
Electrothermal atomic absorption spectroscopy is being proposed by the IUPAC Subcommittee on Environmental and Occupational Toxicology of Nickel as the standard technique for trace nickel determinations in biological samples (1, 2 ) . Other than atomic absorption (AA) procedures ( 1 , 3, 4 ) , relatively few independent ultramicrogram methods for nickel are available. We have been successful in developing a sensitive differential pulse polarographic ( D P P ) procedure for the determination of nanogram quantities of this metal. The basic methodology of this polarographic approach is outlined in this report, as well as its application to the analysis of tap a n d lake water. T h e merits of this new method are further demonstrated by the evaluation of nickel concentrations in nickel uptake studies involving living plant tissue. Reference will also be made t o its application to the analysis of human blood and urine. Previously characterized polarographic techniques suitable for the determination of submicrogram levels of nickel lack sensitivity ( 5 ) or suffer from serious interferences, in addition t o being time consuming (6). Furthermore, the standard anodic stripping voltammetric procedures a t a mercury electrode employed (7) for other metals (e.g., Cd, Pb, Cu, Zn) are not applicable to nickel because of the irreversibility of 0003-2700/80/0352-1013$01 OO/O
the NiZf/Ni couple. While investigating derivative polarography of nickel at a dropping mercury electrode in ammoniacal tartrate and citrate buffers, we discovered t h a t the addition of butane-2,3-dione dioxime (common name is dimethylglyoxime, DMG) enhanced the peak current by a factor of about fifteen. This enhancement resulted in excellent peak characteristics and afforded a detection limit of 2 ppb. On completion of the characterization of this DMG sensitized reaction, we became aware that a group of Russian researchers had previously noted this enhancement phenomenon (8, 9 and references therein). However, their reports lack detail on instrumentation, instrumental parameters, and experimental procedures. Equally important is that these earlier references appear to have been overlooked by analysts interested in trace nickel analysis (e.g., 10, 1 1 ) . The current work is intended to rectify these deficiencies and illustrates that the differential pulse polarographic analysis of nickel described complements even the most sensit,ive AA methods.
EXPERIMENTAL Reagents. Double-distilled deionized water (DDI) was pre-
pared by demineralizing laboratory distilled water (Corning LD-2a Demineralizer) and distilling it in a Corning Mega-Pure still. Standard nickel solutions were prepared from NiC'12.6H20(BDH AnalaR) and were standardized against EDTA (BDH AnalaR, disodium salt dihydrate) using murexide indicator (12). Mineral acids (HCl and HN03) were of Baker Analyzed grade. Ammoniacal tartrate and citrate buffers were prepared from aqueous ammonia (CIL Reagent, 28-29%), (+)-tartaricacid (BDH Anal&) and citric acid monohydrate (Baker Analyzed),respectively. DMG (BDH AnalaR) reagent solutions in 95% ethanol usually were 1% w / v . All other chemicals were reagent grade. Used mercury (Fisher ACS Reagent) was cleaned by passing a fine spray (gravity fed) through a 1-m column of 1 M "0,. This purification step was repeated nine times using fresh acid after every third run. Further purification was achieved by bubbling prepurified O2 for 12 h through mercury which was covered with 1070NaOH and C 1980 American Chemical Society
1014
ANALYTICAL CHEMISTRY, VOL. 52, NO. 7 , JUNE 1980
Table I. Polarograph Operating Parameters instrument control scan rate scan direction initial potentia 1 modulation amplitude operation mode range current range output offset display direction drop time low pass filter N, purge time
setting 5 mVs-l
negative
-0.70 v 50 mV
differential pulse -1.5
v
0.5-2 p A 0 FF positive 0.5 s 0 FF 5 min
was agitated vigorously by stirring. This oxygenation process was repeated with 10% "OB. The purified mercury was then washed copiously with DDI and separated from it by the use of a separatory funnel. It was stored until required under an atmosphere of prepurified Nz in a Nalgene bottle. A p p a r a t u s a n d I n s t r u m e n t a l Detail. Polarograms were recorded with a PAR 174A Polarographic Analyzer outfitted with a Houston Omnigraphic 2000 Recorder and the standard PAR dropping mercury electrode cell assembly. The latter included a Model 174/70 drop timer, a calomel reference electrode with bridge, a Pt counter electrode, and a Nz purge tube. The most commonly used instrument settings are shown in Table I. The purging N2 gas was passed through a scrubbing assembly consisting of 2 or 3 gas washing bottles fitted with fritted glass disks of medium porosity. The first wash bottle contained 0.1 M Cr2+in 2.4 M HC1 over zinc amalgam (13). The second bottle contained DDI water. When a large number of measurements were required on the same solution, it was advantageous to insert a third bottle containing the analyte buffer. All AA measurements were made using a Perkin-Elmer model 303 atomic absorption spectrophotometer, following the standard procedures prescribed by the manufacturer (14). Measurements of p H were performed with a Radiometer model pHM52 meter standardized against two buffers of p H 7 and 9. All essential glassware and plastic reagent bottles were washed with 50% H N 0 3 and thoroughly rinsed with DDI prior to use. Eppendorf pipets were used for the routine additions of microliter quantities of reagents. Characterization of t h e Enhancement Phenomenon. E f f e c t of p H and B u f f e r Concentration The pH dependence of the peak current was investigated by adjusting the pH of 10-mL aliquots of 0.1 M (NHJ,Cit/O.l M NH3 with strong base (10M NaOH) or strong acid (12 M HC1). The pH adjustments were carried out directly in the polarographic cell, and a fresh solution was used for each individual peak current measurement. The p H range examined was 9 f 2. Each sample contained 50 ppb M in DMG, corresponding to Ni (0.85 pM) and was 1.29 X a 150-pL addition of a 1% w/v solution. This experiment was repeated with the analogous ammoniacal tartrate buffer. The effect of buffer concentration was examined for a range of 1:l (NH4)3Cit/NH3buffers with both constituents a t concentrations of 0.05, 0.10, 0.15, 0.20, or 0.25 M. The Ni and DMG concentrations were 25 ppb (0.425 pM) and 4.31 X M, respectively. Dependence on D M G Concentration. Polarograms were recorded for 12 samples in 0.1 M (NH4)3Cit/0.1M NH, each at pH 9.2 and containing 50 ppb Ni but varying amounts of DMG (4.3 X to 6.9 X lo4 M). Similar DMG titrations were carried out in 0.25 M (NH4),Cit/0.25 M NH3 and 0.1 M (NH,),Tart/O.l M "3.
Magnitude and SpeccfzcLtj of Enhancement. The enhancement of the Ni signal (100 ppb) due to DMG (4.31 X 10 M) was compared for 3 buffer systems. All 3 buffers were 0.1 M NH, with the other constituent (also 0.1 M) being (NH4)2Tart,(NH,),Cit, or NH4Cl. The simultaneous presence of comparable amounts (50 ppb) of Fe3+,Cu2+,Zn2+,Pb2+,and Co2+was studied in the 0.1 M ammoniacal tartrate buffer. In the same medium, the individual presence of up to 1 ppm Fe3+or Zn2+was investigated. Eualuatcon of n. Current-sampled (Tast) dc polarograms were recorded for 100-ppb Ni solutions in the presence of DMG for
both the ammoniacal citrate and tartrate buffer systems. The number of electrons ( n )involved in the electroactive process were evaluated from plots of the applied potential vs. log [ i / ( ~ ,-, i~) ] . For comparison, n was also calculated from the line width at half the peak height ( W , 2) obtained from D P P polarograms. Analytical Methods a n d Applications. Calibration Procedures. Calibration curves (peak height vs. concentration) in the presence of DMG were established up to 100 ppb of Ni. The DMG concentration in the 0.1 M ammoniacal tartrate system was 4.31 X lo-' M (50 pL of 0.1% DMG in 10 mL), while it was 6.03 X M (35 pL of 1% DMG in 5 mL) or 1.29 X M (150 pL of 1%DMG in 10 mL) in the corresponding citrate system. For these experiments, a gas scrubbing bottle containing the appropriate buffer was inserted into the N2 purification train between the DDI scrubbing bottle and the sample cell. In the absence of DMG, no measurable signal occurred in the citrate buffer. In contrast, well-defined signals were observed in the tartrate system a t Ni concentrations exceeding 50 ppb. Standard curves were examined in the concentration range 50 to 1500 ppb Ni. Analysis of S u d b u r y L a k e and T a p Water. Tap water was collected directly from the laboratory tap after running it for 15 min. Lake water was collected from Ramsey Lake which is situated near the Laurentian University campus and which is heavily used for recreation and is about 5 km from the Copper Cliff nickel smelting facilities. Both tap and lake water samples were stored in glass containers (pre-rinsed with 50% HN03), refrigerated overnight, and analyzed the next day by both AA spectrophotometry and DPP. In preparation for AA analysis, a known weight (-500 g) of the water samples was placed in a 1-L round bottom flask. T o prevent hydrolysis (especially of Fe3+),250 pL of 12 M HC1 was added. Sample volumes were subsequently reduced by a factor of about 5 and analyzed directly by AA. D P P analysis of the samples consisted of diluting 5 to 50 mL in a volumetric flask with 0.1 M (NH,),Cit/O.l M NH3 buffer. Aliquots (10 mL) were analyzed after the addition of 50 pL of 1% DMG. Standard addition methods were employed to quantify the peak currents. Samples were spiked with 25 pL of 10 ppm Ni to calculate the Ni concentrations as follows: CN, = . ipu
i*L.C,
+ (ig - i J V
where il = sample peak current; i2 = peak current after spiking; v = spiking volume; V = original sample volume; C , = concentration of the added standard; and Csi = concentration of Ni in the sample. Nickel C'ptake by Plant Tissues. Lichen samples (Umbilicaria rnuhlenbergii) were collected near Nairn, 50 km west of Sudbury, Ontario. In the laboratory, lichen material was sampled as described elsewhere (e.g., 15, 16). Briefly, samples were thoroughly washed in DDI to remove debris. Sixteen 5-mm disks were cut from a single thallus. Eight disks were incubated (submerged) for 1 h in 4 mL of 0.01 M NiCl, contained in a 20-mL glass scintillation vial, which was agitated gently by a wrist-action shaker. The remaining 8 disks (the control samples) were similarly incubated in 4 mL of DDI. Each batch of disks was then isolated by vacuum filtration, rinsed on the filter with DDI, and then returned to the shaker for 0.5 h in 4 mL of DDI. After filtration and washing on the filter with DDI, the disks were left to air dry on filter paper (ca. 24 "C and R.H. of 30%). Subsequently, 4 disks from each set of 8 were moistened, and dissected by scalpel under a microscope to remove the upper algal layer ( 1 7 ) . The average weight of each disk was 10 mg. The tissue removed by dissection averaged 20% of the total weight. Sample and control disks (dissected and undissected) were then incubated individually in 0.1 M HC1 for 1 h on the wrist-action shaker. In this HC1-stripping step, the Ni taken up in the Ni-incubation phase was displaced by an ion-exchange process (15, 18, 19). Individual disks were separated by filtration accompanied by washing with DDI, and the accumulated filtrate was evaporated to dryness on a hotplate. The residues from the 8 control samples (those initially incubated in DDI) were taken up in 15 mL of 0.1 M (NH,),Cit/O.l M NH3. Five-mL aliquots of these samples were analyzed for Ni in the presence of 25 pL of 1% DMG. Quantification of the Ni peak was accomplished by the method of standard additions using 25
ANALYTICAL CHEMISTRY, VOL. 52, NO. 7, JUNE 1980 075,
I
,,
I
-I 0 25
- . ..
1..
4,050
0.60 -
I
i
c'
I
a $
1-
0 700
1015
0.40-
0.20 -
f
f
m
d
800
900
1000
11 00
40
M1
CONCENTRATION OF DMG ( M x 10'1
PH
Figure 1. Dependence of signal height on pH. Experimental conditions: 50 pbb Ni and 150 pL 1 % DMG in 10 mL of ammoniacal citrate buffer, and an instrument sensitivity of 1 FA. Peak currents were corrected
for titrant volume differences pL of 10 ppm Ni t o spike the samples. Residues isolated from the lichen disks pretreated in the Ni solutions were dissolved in 10 to 20 mL of 0.1 M (NH.JzTart/O.l M NH3. Because of the presence of considerable amounts of Ni, the enhanced sensitivity resulting from the addition of DMG was not required. In this case, spiking was achieved by adding 10-pL aliquots of 100 ppm Ni to 5 mL of sample. After the HCl stripping treatment, the disks were isolated by filtration, washed with DDI, air dried, and finally weighed after drying for 24 h at 80 O C . All lichen Ni levels are expressed in units of pg Ni/g oven dry weight of lichen.
RESULTS AND DISCUSSION Characterization of the Enhancement Phenomenon. E f f e c t of p H , B u f f e r , and DMG Concentration. It is evident from the data in Figure 1 that the peak current in the presence of a large excess of DMG (reagent/Ni = 1520) was independent of p H for values >9.2. In the analogous tartrate buffer, pH independence occurred at pH >8. A more complex p H dependence was observed in the tartrate buffer for a small excess of DMG (reagent/Ni = 85). In this case, the peak current increased between p H 7 and 8, reached a maximum near p H 8.7, and decreased between pH 9.2 and 10.4. Vinogradova and Prokhorova (8) have reported a similar parabolic p H response in a NH4Cl/NH3 buffer. T h e buffer concentration had an appreciable effect on the peak current. For a 0.1 M increase in both (",)&it and NH,, the peak currents measured at constant p H decreased by 0.1 FA, corresponding to a 40% reduction in the Ni peak on increasing the buffer concentration from 0.10 to 0.20 M. As the data in Figure 2 illustrate, the peak current (i,) showed a strong dependence on the DMG concentration, [DMG]. A double reciprocal plot of the data in Figure 2 was found to be linear and conformed to the expression (see legend to Figure 2 for statistical parameters):
where i, is the peak current at any value of c; i, is the maximum observable peak current; d is a constant (units of M-2); and c represents the bulk concentration (M units) of DMG. This expression may be rearranged to give Equation 3:
(3) In this form, the DMG dependence resembles a Langmuir adsorption isotherm ((20);p 288 of (21);p 57 of (22)). However, as explained in a subsequent section, this agreement does not provide a priori evidence for the occurrence of adsorption.
Figure 2. DMG adsorption isotherm. The solid curve was calculated using the following statistical parameters for the linear plot corresponding to i;' v s [DMGI-' (see Equation 2): rn = 6.48 X l o - @PA-' M2 ( r = 0.999) and b = 1.1 1 FA-'. Experimental conditions: 50 ppb Ni (0.85 pM) in 0.1 M (NH,),Ci/O. 1 M NH, at pH 9.2. Each data point corresponds to a separate sample solution
-0.85
.1.05 -1.25 APPLIED VOLTAGE. V lversus SCEI
.1.45
Figure 3. Magnitude and specificity of the DMG enhancement. Differential pulse polarograms were recorded at a sensitivity of 1 p A and corres ond to: (a) 0.1 M (NH,),Tart/O. 1 M NH,; (b) buffer and 50 ppb of Fe3', Ni2+, CU", Zn", and Pb2+; (c) buffer, metal ions, and DMG (43 FM)
Equilibrium considerations illustrate that the right-hand side of Equation 3 corresponds to the fraction of Ni2+present in solution as NiA,, with A representing the dimethylglyoximate monoanion. At the experimental p H , the bulk of the DMG (-95%) remained protonated and thus c [HA]. Empirically, values of /T1,'denote the concentrations of DMG a t which half the maximum peak current is achieved. For 0.1 M ammoniacal citrate and tartrate buffers a t p H 9.2 with 50 ppb Ni, these half-height current Concentrations were 2.4 X M and 1.6 X M, respectively. The analogous DMG concentration for 25 ppb Ni in 0.25 M (NH4),Cit/0.25 M NH, was 7.7 x M. Calculations using experimental values of p in Equation 3 indicate that the ratio iJim = 0.99 requires a 2840-fold excess of DMG (calculated for 50 ppb Ni) for the 0.1 M (NH,),Cit/O.l M NH3 buffer, while a 180-fold excess is needed for the analogous tartrate buffer. From a solubility point of view, the former excess was considered unsuitable. The maximum reagent excess selected for routine use was 1520 a t a Ni level of 50 ppb, which corresponded to an addition of 150 pL of 1% DMG per 10 mL of sample and a calculated value of ip/im = 0.96.
1016
ANALYTICAL CHEMISTRY, VOL. 52, NO. 7, JUNE 1980
M a g n i t u d e a n d S p e c i f i c i t y of E n h a n c e m e n t . The differential pulse polarogram of the buffer (curve a, Figure 3) shows the presence of Zn2+and Fe"' contamination. (Proper cleaning of glassware and careful selection of chemicals would circumvent the presence of such undesirable impurities.) Compared to t h e Zn2+a n d Fe3+peaks in polarogram b, the Ni2+ peak is considerably broader. This is consistent with the irreversibility of the Ni'+/Ni couple. T h e addition of the DMG (curve c, Figure 3) selectively enhanced the Ni2+peak. Neither was any enhancement observed for the Cu2+or PI?+ peaks, which occurred (not shown) at more positive potentials. I n 0.1 M (NHJ2Tart/0.1 M NH3, the DMG enhancement of the Ni2+peak was 15.0 at a DMG/Ni2+ ratio of 250 compared to 12.2 in 0.1 M NH4Cl/0.1 M NH3 No observable Ni2+peak was obtained in the absence of DMG in 0.1 M (NH,)~$it/O.l M NH,. Interestingly, the observed peak currents at pH 9.2 in the presence of DMG were in the ratio of 1.3:1.2:1.0 for 0.1 M NH,Cl/O.l M NH,, 0.1 M (NH&Tart/O.l M NH:,, and 0.1 M (NH4)&it/O.1 M NH3, respectively. I I I A separation of 150 mV was observed for the DMG sen-0.80 -090 -1.00 -1.10 sitized peak currents of Ni2+ ( E , = -~0.99V SCE) and Co2+ APPLIED VOLTAGE, V (versus SCEI ( E , = -1.14 V SCE) in ammoniacal citrate buffer. The peak heights were comparable for both metal ions. Consequently, b the determination of nanogram quantities of Co2+should he feasible and may be expected to be similar in sensitivity and precision to the Ni analysis. The DMG-sensitized reduction of Co2+ and its analytical applications are currently being investigated in our laboratory. T h e peak potentials in ammoniacal tartrate buffer for Zn'+ ( E , = -1.18 V SCE) and Fe3+ ( E , = -1.42 V SCE),and of course those for Pb2+and Cu2+,were far enough removed from the Ni2+ potential ( E r > = -0.99 V SCE) to circumvent peak area overlap. For example, the presence of 100 to 950 p p b Zn2+ did not alter the observed peak current (0.350 k 0.013 p A for 25 p p b Ni2+;total of 6 measurements). Similarly, the presence of up to 670 ppb Fe3+had no measurable effect (0.622 f 0.015 p A for 50 ppb Ni2+;4 measurements). Reuersibility. Plots of the applied potential vs. l ~ g [ i / ' ( i , ~ , ~ , - i)] a t pH 9.2 and [DMG]/[Ni2+]= 250 had slopes of 28.5 mV ( r = 0.999, N = 6; 0.1 M citrate buffer) and -27.3 m l ' ( r = 0.999, N = 7; 0.1 M tartrate buffer). For a 2-electron reduction process at ambient temperature (ca. 24 "C), the theoretical gradient is -29.5 mV. ( N refers to the number of -0.70 .0.80 4).W -1.00 -1.10 data points and r is the correlation coefficient.) Consequently, APPLIED VOLTAGE, V (versus SCE) the experimental value of n was 2.1 1 f 0.04. The agreement Figure 4. Typical differential pulse polarograms for a series of Ni2+ (within 5.5%) of the observed slopes with the theoretical concentrations in the presence (a) and absence (b) of DMG. Expervalues, and to a lesser extent the high degree of linearity, may imental conditions: (a) as described in footnote a , Table 111; (b) as be taken as an indication that the electrode process effectively given for data set 5, Table I1 is reversible or simulates reversibility ((23);p p 129, 205 of (21)). concentration in the range 360 to 740 ppb Ni giving an average T h e involvement of protons in a reversible reaction is value of 90 f 4 mV ( L V = 50 mV). In contrast, t h e DMGcharacterized by a linear relationship between E,:?, the sensitized Ni peak yielded values of W1,* for AV = 50 mV half-wave potential of the dc wave, and pH. For D P P powhich were independent of concentration: 47.1 f 0.7 mV in larograms, E , = E l I 2- LV/2 with L V the modulation amthe concentration range tested, namely 20 to 180 p p b Ni plitude (potential pulse). Such pH dependence was observed ([DMG] = 6.03 X M). Similar Wl12values (45 f 1 mV) for E , in both ammoniacal citrate and tartrate buffers were observed in the presence of DMG for L V values down ([DMG]/[Ni*+]= 1520). At ambient temperature (ca. 24 "C) to 5 mV. Consequently when DMG is present, the magnitude a slope of -31.9 mV ( r = 0.999, N = 13) was observed in the of WIl2 approximates t h e limiting, pulse-amplitude indeformer medium, while in the case of the latter rn = -31.5 mV pendent, values calculated by Parry and Osteryoung ( 2 4 ) for ( r = 0.999, N = 12). T h e theoretical value at 24 "C of the a reversible system. The expected value is 45.2 mV for n = slope, -2.303 p R T / n F = - - 0 . 0 5 9 0 ( p / n ) ,is -29.5 mV for n = 2. Previous studies (25-27) have shown that the reduction 2. Consequently the experimental value of p , the number of of Xi2+a t a mercury electrode is irreversible, with the addition protons, is 1.07 (or p = 1.00 when n = 1.86). These data of the first electron being rate determining (27, 28). Interprovide further confirmation that the electrode process may estingly, the average value of W1,* (90 mV) observed in the be characterized as Nernstian. absence of DMG does correspond to t h e Parry-Osteryoung limiting value for n = 1 ( 2 4 , 28). Finally it is interesting to compare the effect of DMG on Analytical Applications. Calibration Procedure, S e n the observed line-width at half-height, Wli2. In the absence sitzaity, a n d O p t i m u m Conditions. The set of polarograms of DMG, W1,* values decrease systematically with increased (I
~
ANALYTICAL CHEMISTRY, VOL. 52, NO. 7, JUNE
1017
1980
Table 11. Summary of Regressional Parameters for Calibration Curves no. of
data set
bufferQ
1 '
0.1 M (NH,),Cit/O.l M NH,
2
0.1 M (NH,),Cit/O.l M NH,
3 4 5d
0.1 M (NH,),Tart/O.l M NH, 0.1 M (NH,),Tart/O.l M NH, 0.1 M (NH,),Tart/O.l M NH,
concentration slope, m [DMGI ( M ) range (ppb) (ccA/ppb)" 6.0 x 10-4 0-95 0.00444 0.00426 0-151 1.3 X 0-76 0.0124 0-1 1 4 0.0119 4.3 X 0-74 0.0152 0 0-740 3.00 x 10-4 0 0-1500 5.79 x
intercept, b (PA) d.00673 0.0158 0.0161 0.0342 0.0178 -1.72 x 1 0 - 3 -9.39 x
data corr. points, coefi., r N 1.000 1.000 1.000
5 8
0.999 0.9!39
12 6 14 6
i.ooo
1.000
8
Least squares fit of the data was achieved by y-residual minimization. pH = 9.2. Data are given in Table I11 (see Corresponding polarograms are exhibited in Figure 4b. Different capillaries were used in seis 4 and 5. also Figure 4a). Table I V . Analysis of Sudbury Tap arid Lake Water
Table 111. Calibration Curve Linearitya
sample
observed peak current,
no.
PA
1
0.095 0.173 0.258 0.346 0.429 0.500
2 3 4 5 6 7 8
9
experimental nickel concn., PPb 19.1
0.581
38.2 57.2 76.1 95.0 114 132
0.650 0.721
151 17 0
calculated nickel concn ppbb" 19.9 37.5 56.6 76.4 95.1 111
129 145 161
Experimental conditions: 0 . 1 M (NH,),Cit/O.l M NH,, pH 9.2; [DMG] = 6.0 X 10.' M ; 1 0 - p L aliquots of a 10.0 ppm Ni" solution were added consecutively t o 5 mL of buffer in the polarographic cell. Based o n the least squares fit (y-residual minimization) to the data f o r samples 1-5, inclusive: y = 0 . 0 0 4 4 4 ~+ 0.00673 with r = 1.000 (see Table 11). shown in Figure 4a demonstrates that in the presence of DMG excellent peak characteristics were observed. The flat base line facilitated peak current measurements. In contrast, in the absence of DMG, base-line currents were more prominent and somewhat less reproducible (see Figure 4b). It appears t h a t unfavorable charging or capacitance currents are responsible for this (see, for example, the discussion by Bond and Grabaric (29)). T h e improvement in residual currents in t h e presence of DMG suggests a reduction in electrode double-layer capacitance by adsorption of this reagent on the electrode surface (pp 60, 76 of (22); (30)). Regressional parameters for a number of calibration curves are summarized in Table 11. Generally a high degree of linearity was observed as indicated by the near-perfect values of the correlation coefficients. In addition the y-intercept values were also very close to zero, corresponding to 51.5 ppb in the case of the DMG-sensitized reaction (but see comments below) and 1 1 6 ppb for the conventional D P P procedure (data sets 4 and 5 , Table 11). However, in the presence of DMG a slight curvature could be discerned. The data in Table I11 illustrate that deviations from linearity became significant for Ni2+ concentrations exceeding 85 ppb. Similarly, the y-intercept values reported in Table I1 (see data sets 1 and 2) increased by a factor of 2 when data points beyond the linear range were included in the statistical analysis. The analytical application of the DMG-sensitized reaction is therefore limited. However, Ni2+measurements at concentrations >85 ppb were readily obtained in the absence of DMG. Although the upper experimental concentration range was normally limited to 750 ppb Ni2+,and to 1500 ppb in data set 5, Table 11, there is no obvious experimental reason why it could not be extended to even larger concentrations. It is deduced from the
water sample
collection date
nickel concentrat ion, no. _ _ _ _ _ _PPb _ ~ of AAa DPPa samples
2 July 7, 1978 227 t 2 b 231 i O b 2 266 2 1' 268 O b July 7 , 1978 358 t 7 c 333 i 8c 8 tap Dec. 1 2 , 1978 AA, flame atomic absorption spectrophotometry; DPP, differential pulse polarography. See text for experimental detail. i average deviation I standard deviation. lake
tap
__
4
___-
high linearity and insignificant >-intercepts that the addition method summarized in Equation 1 is valid for both procedures. Examination of the slope values reported in Table I1 reveals that the sensitivity of the D P P procedure is dependent upon two factors: (1)DMG concentration; and (2) the buffer type. A 2.2-fold increase in DMG concentration resulted in a 2.8-fold increase in sensitivity (compare data sets 1 and 2, Table 11). Interestingly, a linear response was observed in both cases. Maximum sensitivity was observed in the tartrate buffer system where the DMG requirements were much lower (data set 3, Table 11). Although the DMG-induced enhancement factor of 15 was observed in the ammoniacal tartrate buffer a t a [DMG]/[Ni2'] ratio of 250 (vide supra), the slopes of the calibration curves in the presence of DMG were as much a t 50 times larger as in its absence (compare data sets 3 and 4, Table 11). In our view, the ammoniacal citrate buffer has certain attractive features, even though the DMG enhancement is greater in the corresponding tartrate buffer. The former medium was more suitable for dissolving residues obtained from the ashing or digestion of biological tissues. Furthermore, background currents were generally smaller and more reproducible. The optimum conditions for the citrate medium were: buffer concentration of 0.1 M, pH 29.2, and [DMG] 24.3 X to 1.3 X 10 M. The same conditions may be employed in the tartrate medium, although the optimum pH \ d u e may be lowered (pH > 8 ) . Analysis of Sudbur? Lake a n d Tiip Water. T h e results of the analysis are summarized in Table IV. Agreement between the D P P and AA results are excellent for the July 7 . 1978, data and somewhat less satisfactory for i he remaining set. Low absorbance readings were involved in the AA measurements, and thus the slightly higher AA results are likely not significant. The D P P analysis requires no pretreatment of the samples and is an extremely rapid, sensitive, reproducible, and convenient method of water analysis. The high contents of Ni in Sudbury tap and lake water may be related to Ni emissions associated with the local smelting and refining of the metal (31).
1018
ANALYTICAL CHEMISTRY, VOL. 5 2 , NO. 7, JUNE 1980
Table V. Uptake of Nickel by Different Regions of the Thallus of the Lichen Umbilicaria muhlenbergii
sample set a 1 2 3 4
sample treatment Ni2+nickel content incuba- dissec- of zone, Wg/g tion oven-dry wt tionC -
+ +
-
+
-
+
1 3 ? 3e 12* 5f 790 t 6 0 e
640
I
SOf
method of analysisd DPPiDMG DPP/DMG DPP DPP
Each set consisted of 4 individual disks; each disk was analyzed separately. -, treatment omitted; - , treatment carried out. Weight removed as algal zone was 1 9 f 2% of total weight. DPP: differential pulse polarography; DMG: dimethylglyoxime sensitized reaction. e Corresponds t o Ni content of intact disk. f Corresponds t o Ni content of fungal zone.
on twice the magnitude of the observed background current. T h e reproducibility reported in Table IV (see also Table V) concurs with these precision estimates. The inherent lack of precision associated with adding microliter aliquots of Ni standards in the spiking step contributes significantly to t h e U-values reported for the unknowns.
The Electroactive Process in the Presence of DMG. Equations 4 to 9 describe the various processes and relationships t h a t are deduced empirically to be important for the ammoniacal buffers employed.
+ 2HA
Ni2+(compl)
a
Nickel Uptake by Plant Tissues. The uptake of Ni2+from aqueous solutions by the lichen Umbilicaria muhlenbergii has been examined in detail by Nieboer et al. (18, 19, 32). Mass balance and charge balance studies, as well as the kinetics, pH, and temperature dependence of Ni2+uptake, all support an extracellular ion-exchange mechanism. Lichens consist of two organisms, a fungus and a n alga, which co-exist in a symbiotic partnership. More specifically, Umbilicaria muhlenbergii is a foliose (leaf-like) lichen with the algal cells forming a relatively thin surface layer which is covered by a protective layer known as t h e upper cortex. Although the overall mechanism of Ni2+uptake is well known, it is uncertain as to whether metal ions such as Ni2+associate preferentially with one of the symbionts, ( 1 6 , 3 3 , 3 4 ) . The new data summarized in Table V demonstrate that in the lichen examined, 66% of the Ni2+taken up was associated with the fungal zone a n d 34% with the algal zone. Furthermore, the uptake capacity of the fungal layer was considerably smaller (640 f 50 pg/g; see Table V) t h a n that of the algal zone (1430 pg/g; calculated from the experimental fungal and whole-lichen capacities and the weight fractions). T h e background Ni2+ contents of just over 10 yglg, although based on small sample weights (a single disk of 10 mg), are in good agreement with values determined on 2-g samples by X-ray fluorescence for lichen material collected at the periphery of the Sudbury pollution zone (15). Since dissection of lichen disks is a difficult task, the Ni2+uptake experiments had to be limited to a few small disks. Nevertheless, the sensitivity of the DMG-sensitized Ni2+reduction was sufficient to evaluate the natural and low Ni backgrounds in these. The analysis of the Ni*+ incubated disks was more conveniently achieved by the less sensitive DMG-free D P P procedure. T h e lichen application summarized above illustrates that the newly devised D P P Ni methods are suitable to determine both nanogram and microgram quantities of Ni. The small tissue requirement, and the ability to circumvent laborious dilution by the judicious use of DMG are both very attractive features. Applications to other biological analytical problems such as Ni in human blood and urine are being investigated (vide infra). Precision a n d Detection Limit. The standard deviation, u, for the measurement of currents of replicate samples in the citrate medium containing 25 or 50 p p b Ni and DMG was 0.014 pA (N= 10). Consequently, the lowest limit of detection may be estimated as 2 u l m or 3u/m corresponding to confidence limits of 95.4% and 99.7%, respectively; in these expressions m is the slope of the calibration curves as summarized in Table 11. Thus, the lowest limit of detection using DMG is between 2 to 3 ppb. For comparison, the detection limit in the absence of DMG is estimated at 50 ppb, based
-
K
(Equation 4) =
NiA,(aq)
+ 2H+
[NiAdaq)l WI2
[ Ni'+(compl)] [HA]
[Ni2+(compl)]= CNi- [NiA,(aq)]
(4) (5)
(6)
Equation 5 may be rearranged to yield Equations 7 and 9 after replacement of the term [Ni2+(compl)]by the right-hand side of Equation 6.
In these expressions, Ni*+(compl)denotes the Ni2+complexed with the buffer components (NH3and citrate or tartrate); CKl the total concentration of Ni2+in all forms; and as before HA the neutral protonated DMG molecule and NiA, the bis(dimethylglyoximato)nickel(II) complex. T h e similarity of Equations 3 and 7 suggests that the coefficient /3 may be defined in terms of known equilibrium constants and experimental concentrations. Evaluations of the Ni2+(compl)to NiA,(aq) conversion ratio p' (Equation 9) for the data sets 1-3 in Table I1 using the experimental DMG concentrations and the observed p's (see discussion of Equation 3) yielded the following values: 0.85 (data set 1, Table 11) and 0.96 (data set 2) for the citrate buffer; and 0.86 for the tartrate system (data set 3). Assessment of h from the experimental /3 values requires t h e identity and stability of the species Ni2+(compl)in Equation 4. Appropriate equilibrium calculations yielded the best agreement between the experimental and predicted estimates of the formation constant K,,, for Ni2+(compl)when it was assigned the formula NiL(NH3)4,with L denoting the citrate (3-1 or tartrate (2-) anion. All cases corresponding to NiL(NH3)x=1.6, as well as Ni(NH&*+ and NIL, were examined. The experimental estimates of log K,,, were 11.4 (citrate) and 9.5 (tartrate), compared to 11.3and 9.0, respectively, calculated from 4 log K(NiNH3) + log K 1 (NiL), with log K = 116 log Ps = 1.5 (see below) and log K, = 5.4 (citrate) and 3.0 (tartrate). In these computations and related equilibrium considerations, the following equilibrium constants ( 3 5 , 3 6 ) were employed for the protonation of ligands and their interactions with Ni2+ (at 20 or 25 "C and p = 0.1 M): ammonia, pK, = 9.3 and log p6 = 9.1 (at p = 2 M); DMG, pK,, = 10.5, pKa2 = 11.9 and log B2 = 17.2; citrate, log K 1 = 5.4; and tartrate, log K1 = 3.0 (37) and log p2 = 5.4. T h e process summarized in Equation 4 explains the p H dependence depicted in Figure 1. A plot of log i, vs. p H for the first three data points a t the foot of the sharply rising portion had a slope of 2.0, indicating the release of 2 protons in the intrinsic chemical reaction. Independence of p H of the
ANALYTICAL CHEMISTRY, VOL. 52, NO. 7, JUNE 1980
formation of NiA, occurred a t p H values (9.2-11) where a significant proportion of HA was present as A-. The more complicated p H dependence reported for a relatively low [DMG] in t h e tartrate medium reflects the competition between the buffer components (especially NH,) and DMG for the Ni2+. Thus, the observed current reached a maximum at intermediate p H values (8.7-9.2) and decreased thereafter owing to the increase in [NH,]. The lack of a Ni2+signal in the absence of DMG in the citrate medium is consistent with the more efficient complexing of citrate compared to tartrate. Similarly, the peak current ratios 1.3 (NH4C1/NH3):1.2 (Tart/NH3):l.0 (Cit/NH3) reflect the ability of the buffer constituents to compete with DMG for the available Ni2+. Reaction 10 is consistent with the various experimental observations characterizing the electroactive process and the equilibrium considerations outlined above. NiA,
+ H+ + 2e- ii products
(10)
One mole of protons is consumed in the 2-electron process as dictated by the p H dependence of E,. Evaluations of n and line widths support the conclusion that this reaction may be characterized as reversible. Evidence in support of the postulate that the electroactive species is NiA, and that furthermore, it is adsorbed on the electrode surface, may be summarized as follows. (i) T h e equilibrium considerations outlined above confirm that NiA2 is the dominant species in solution. (ii) NiA, has limited solubility: NiA2(s) e NiA,(aq) with K(solubi1ity) = lo4 (35). (iii) The independence of E , of buffer concentration implies that Ni2+ is not the electroactive species (see, for example (38)). (iv) Base-line flatness in the presence of both DMG and Ni2+ is compatible with adsorption decreasing residual currents ( p p 60, 76 of (22)). (v) Current enhancement like that observed is often induced by adsorption (30,39-41). (vi) And finally, the narrow concentration range for which calibration curves were linear in the presence of DMG (see Tables I1 and 111) is consistent with adsorption. Langmuir-type adsorption isotherms exhibiting the mathematical form as in Equation 3 are linear only if the denominator has a value of unity. This condition is normally met only at low solution concentrations of the electroactive species. Reversible reduction of absorbed metal complexes at a dropping mercury electrode is known. Thus, Kalvoda et al. (39) examined the reduction of a sparingly soluble complex of Cd2+ in the presence of both thiocyanate and hexamethylenetetramine. In comparison to the reduction of uncomplexed Cd2+,the shift in half-wave potential of the adsorbed complex was slight and the signal enhancement large as in the present case (cf., Figures 4a and 4b). Furthermore, Anson and Rodgers (40)suggest that adsorption may enhance the rate of a n electrode process. Prokhorova et al. ( 8 , 9, 42) have interpreted the DMG enhancement as a catalytic hydrogen wave. They assumed that the reduction of the H + was catalyzed by NiA,. In their model, the protonated form of NiA, is reduced to NiA2H which decomposes bimolecularly to yield H, and regenerates NiA,. For the pH-independent portion of the current curve in Figure 1, E , continued to shift as a function of pH. This lack of current dependence on p H suggests that a protonation step involving NiA, is likely not rate-limiting, even though a proton is consumed in the overall electrode process (Equation 10). Furthermore, the absence of prewaves (43- 45) and postwaves ((46; chapter 18 of (21))supports the notion that we are not dealing with a rate-determining catalytic process. Prokhorova et al. (8, 9,42) used mostly indirect evidence to conclude that H2is released at the mercury working electrode. Nevertheless it is conceivable that the reduced compound, perhaps NiA2H(see Equation lo), could be oxidized rapidly by H+ yielding NiA2 and H2. Rapid regeneration of the electroactive species NiA2 could he responsible for the effective reversibility of
1019
reaction 10. Zero-valent Ni compounds have been implicated as good catalysts in polarographic processes ( ( 4 6 ) ;p 424 of (21)).
CONCLUSIONS T h e detection limit of 2 ppb (at a confidence interval of 95%) and observed standard deviations of the same magnitude for the DMG-sensitized procedure are comparable to those possible with modern electrothermal AA analysis of Ni ( 1 , 4 , 47, 48). Detection limits and precision are virtually limited by the ability to avoid contamination. T h e proposed D P P methods are especially suited for routine analysis of water and other aqueous samples. Especially advantageous is the flexibility to extend the analytical concentration range from ng/mL to pg/mL by simply omitting DMG. I n AA, this switch would involve the more complex replacement of the graphite furnace with the traditional flame accessory. Of course, modern electrothermal AA equipment is computerized and can handle large numbers of samples because of automation. The advent of the laboratory minicomputer should also lead to many innovations and improved performance in D P P (29). And finally, preliminary results ( ( 4 9 ) ;Nieboer et al. in ( 2 ) )have shown that more complicated samples such as human urine and blood requiring oxidative wet digestion or ashing may also be analyzed for trace nickel levels by the DMG-sensitized D P P procedure.
ACKNOWLEDGMENT T h e authors are indebted to David H. S.Richardson for dissecting the lichen samples and for his support of the project. Technical help was provided by P a t Lavoie and Dilva Pado-
LITERATURE CITED D. B. Adams, S. S. Brown, F. W. Sunderman, Jr., and H. Zachariasen, Clin. Cbem., 24, 862 (1978). F. W. Sunderman, Jr., Ann. Clin. Lab. Sci., 8, 491, 495 (1978). S. Nomoto and F. W. Sunderman, Jr., Clin. Chem., 18, 477 (1970). T. K. Jan and D. R. Young, Anal. Chem., 50, 1250 (1978). D. D. Gilbert, Anal. Cbem.. 37, 1102 (1965). M. M. Nicholson, Anal. Cbem., 32, 1058 (1960). J. T. Kinard, J . Environ. Sci. Healtb, A12 ( l o ) , 531 (1977). E. N. Vinogradova and G. V. Prokhorova, Zh. Anal. Kbim., 23, 1666 (1968). V. V. Astaf'eva, G. V. Prokhorova, and R. M. F. Saiikhdzhanova, Zb. Anal. Khim., 31, 260 (1976). F. W. Sunderman, Jr., F. Coulston, G. I. Eichhorn, J. A. Fellows. E. Mastromatteo. H. T. Reno, and M. H. Samitz, in "Nickel", (A Report of the Committee on Medical and Biological Effects of Environmental Pollutants.), National Academy of Sciences, U.S.A., Washington, D.C. 1975. E. Nieboer and A. Cecutti, in "Effects of Nickel on the Canadian Environment", T. C. Hutchinson, Ed., National Research Council of Canada, Ottawa, Ontario, 1980. in press. T. S. West, "Compiexometry with EDTA and Related Reagents", 3rd ed., BDH Chemicals Limited, Poole, England, 1969, p 198. L. Meites, "Polarographic Techniques", 2nd ed., John Wiley and Sons, New York, 1965, p 89. Perkin-Elmer Corp., "Analytical Methods for Atomic Absorption Spectrophotometry", Perkin-Elmer, Norwalk, Conn., 1976. F. D. Tomassini, K. J. Puckett, E. Nieboer, D. H. S. Richardson, and B. Grace, Can. J . Bot., 54, 1591 (1976). E. Nieboer, D. H. S. Richardson, P. Lavoie, and D. Padovan, New fbyfol., 82, 621 (1979). D. C. Smith, Symp. SOC. General Microbiol., 13, 3.1 (1963). E. Nieboer, P. Lavoie, R. L. P. Sasseville, K . J. Puckett, and D. H. S. Richardson, Can. J . Bot., 54, 720 (1976). E. Nieboer, K. J. Puckett, and B. Grace, Can. J . Bot., 54, 724 (1976). S. Gbsstone, "Physical Chemistry", 2nd ed., D. Van Nostrand, Princeton, N.J., 1946, p 1198. J. Heyrovskq and J. Kota, "Principles of Polarography", Publishing House of the Czechoslovak Academy of Sciences, Prague. Czechoslovakia, 1966. Stal'G. Mairanovskii. "Catalytic and Kinetic Waves in Polarography", Plenum Press, New York. 1968. A. M. Bond and G. Hefter, J . Necfroanal. Cbem., 42, 1 (1973). E. P. Parry and R. A. Osteryoung, Anal. Cbem., 37, 1634 (1965). H. B. Mark, Jr., J . Necboanal. Cbem., 8 , 253 (1964). J. Dandoy and L. Gierst, J . Electroanal. Cbem., 2, 116 (1961). M. Goto and K. 6.Oldham. Anal. Chem., 48, 1671 (1976). J. W. Diliard, J. J. O'Dea, and R. A. Osteryoung. Anal. Cbem.. 51, 115 (1979). A. M. Bond and B. S. Grabaric, Anal. Chlm. Acta, 101, 309 (1978). A. P. Brown and F. C. Anson, Anal. Chem., 49, 1589 (1977).
1020
Anal. Chem. 1980, 52, 1020-1024
(31) N. Conroy, K. Hawley, W. Keller, and C. Lafrance, in "Proceedings of the First Speclalty Symposium on Atmospheric Contribution to the Chemistry of Lake Waters'' Int. Assoc. Great Lakes Research, Sept. 28-Oct. 1, 1975, pp 146-165. (32) E. Nieboer, D. H. S. Richardson, and F. D. Tornassini. &yologist, 81. 226 (1978). (33) D. H. S. Richardson and E. Nieboer in "Cellular Interactions in Symbiotic and Parasitic Associations", C. B. Cook, P. W. Pappas, and E. D. Rudolph, Eds. Proceedings, Fifth Annual Colloquium, College of Biological Sciences, Ohio State University, Ohio State University Press, Columbus, Ohia. 1980. (34) D H.' S.-ichardson, E. Nieboer, P. Lavoie, and D. Padovan, New Phflol , 82. 633 11979). (35) A. 'E, MaAell and R. M. Smith, "Critical Stability Constants" (4 volumes), Plenum Press, New York and London, 1977. (36) L. G. Sill& and A. E. Martell, Stability Constants of Metal-ion Complexes", Chem. SOC.(London) Spec. Pub/., No. 17 (1964): Suppl. No. 1, Chem. SOC.(London) Spec. Pub/., No. 25 (1971). (37) N. E. Milad, S. E. Morsi, S.T. Soliman, and L. M. N. Seleem, Egypt. J . Chem. 1SDec. Issue. Pub. 1974). 101-6 (1973): Chem. Abstr.. 81. . 422279 '(1'974), and 62, 145909f(1975). G. A. Heath and G. Hefter, J , Electroanal. Chem., 84, 295 (1977). R. Kalvoda, W. Anstine, and M. Heyrovskf, Anal. Chirn. Acta, 50, 93 (1970). F. C. Anson and R. S. Rodgers, J . Electroanal. Chern., 47, 287 (1973). F. C. Anson, J. 8.Flanagan, K. Takahashi, and A . Yamada, J . Electroanal. Chem., 67, 253 (1976). I.
(42) (43) (44) (45) (46) (471 .
I
(48) (49)
G. V .
Prokhorova, L. K. Shpigun, and E. N. Vinogradova, Zh. Anal. Khim., 27, 780 (1972). H. B. Mark. Jr., and C. N. Reilley, J . Electroanal. Chern., 4, 189 (1962). H. B. Mark, Jr., and C. N. Reiliey, Anal. Chern., 35, 195 (1963). H. 6.Mark, Jr., J . Electroanal. Chem.. 7 , 276 (1964). E. Itabashi. J . Electroanal. Chem., 60, 285 (1975). H. Zachariasen. I. Andersen. C. Kostd. and R. Barton. Clin. Chern.. 21. 562 (1975). D. Mikac-Devib, F. W. Sunderman, Jr., and S. Nomoto, Clin Chern., 23, 948 (1977). C. J. Flora, "Determination of Nanogram Quantities of Nickel by Differential Pulse Polarography at a Dropping Mercury Electrode and Selected Applications", M.Sc. Thesis, Laurentian University, May 1979.
RECEIVED for review April 23, 1979. Resubmitted February 15, 1980. Accepted February 15, 1980. Financial assistance is gratefully acknowledged from Falconbridge Nickel Mines Limited, Falconbridge, Ontario, and the Inco/United Steelworkers Joint Occupational Health Committee, Copper Cliff, Ontario. One of the authors (C.J.F.)records his appreciation for the award of a National Research Council of Canada Postgraduate Scholarship.
Ammonia Electrode with Immobilized Nitrifying Bacteria Motohiko Hikuma, Tatsuru Kubo, and Takeo Yasuda Central Research Laboratories, Ajinomoto Co., Inc., 7 Suzuki-cho, Kawasaki-ku, Kawasaki, 210, Japan
Isao Karube" and Shuichi Suzuki Research Laboratory of Resources Utilization, Tokyo Institute of Technology, Nagatsuta-cho, Midori-ku, Yokohama, 227, Japan
The ammonia electrode consisted of immobilized nitrifying bacteria and an oxygen electrode. When a sample solution was pumped into the flow system of the sensor, the electrode current decreased until a steady state after 8 min. Measurement could be also made with a 3min pumping period with lfflle loss of sensitivity. Calibration plot of the current difference vs. concentration of ammonia was linear up to 1.3 mg/L. The sensor was applied to wastewater and good agreement was obtained with the conventional method (f6YO). The sensor could be used for more than 2 weeks and 1400 assays.
T h e determination of ammonia in a sample solution is important in various fields such as medical, environmental and industrial process analyses. An ammonia-gas electrode consisting of a combined glass electrode and a gas permeable membrane is usually used for this purpose. In this case, the determination must be performed under strong alkaline conditions (above p H 11). Furthermore, volatile compounds such as amines sometimes interfere in the determination of ammonia. Recently many microbial sensors consisting of immobilized microorganisms and an oxygen probe have been developed (1-8). As previously reported, the concentration of substrate can be determined from the respiration activity of the immobilized microorganisms which is directly measured by the oxygen probe attached ( I , 6). A nitrifying bacterium Nitrosomonas sp., utilizes ammonia as a sole source of energy and oxygen is consumed by the respiration as follows; 2NH3 + 302
Nitrosornonas sp.
2HN02
+ 2H20
0003-2700/80/0352-1020$01 .oo/o
Therefore, ammonia may be determined by the microbial sensor using immobilized nitrifying bacteria such as N i t r o somonas europaea and a n oxygen probe. In this study, nitrifying bacteria were immobilized by being entrapped between two membranes, a porous acetylcellulose membrane and the Teflon membrane of the oxygen electrode. T h e microbial sensor was applied t o the determination of ammonia in wastewaters. EXPERIMENTAL Materials. Chloramphenicol was purchased from Sankyo Co. Other reagents were commercially available analytical reagents or laboratory grade materials. The wastewaters were obtained from a fermentation factory. Culture of Microorganisms. Activated sludges containing nitrifying bacteria were obtained from a fermentation factory. About 200 mL of the activated sludge (MLSS 5000 ppm) was inoculated into 2.5 L of the culture medium containing 0.0670 (NHJ2S04,0.05% K2HP04,50 ppm MgS04,4 ppm CaCl,, 20 ppm FeC13,and 25 g of powdered CaC03 (30-pm average diameter, a support for nitrifying bacteria) (9). The mixture was placed in a vessel as shown in Figure 1 and the bacteria were cultured under aerobic conditions (1/2 VVM aeration) for more than four months at room temperature (20-30 "C). The pH of the culture medium was controlled at 8.2 by adding 1 N Na2C03. The fresh culture medium of the same composition described above except for 0.6% (iVHJ2S04employed was fed at a rate of 200 mL/day. On the other hand, nitrifying bacterium, Nitrosomonas europaea ATCC 19718, was cultured in a 2-L flask containing 500 mL of culture medium (pH 8.2-8.4) for 25 days at 25-30 "C. The culture medium employed was the same composition described above except for 0.3% (NH,),S04 employed (sterilized at 120 "C for 30 min). Then, 0.05 ppm cresol red was added as a pH indicator. Calcium and magnesium salts were sterilized separately to avoid precipF 1980 American Chemical Society
