Anal. Chem. 1988, 58, 1959-1964
(27) Stromberg, A. G.; Zhikharev, Yu. N. Zavod. Lab. 1965, 37, 1185. (28) Anderson, J. E.; Taiiman, D. E.; Chesney, D. J.; Anderson, J. L. Anal. Chem. 1978. 50. 1051. (29) Sevclk, A. Col/ect. Czech. Chem. Commun. 1948, 13, 349. (30) Randles, J. E. B. Trans. Faraday SOC. 1948, 4 4 , 327. (31) Levich, 8. Dlscuss. Faraday SOC. 1947, 1 , 37. (32) Orinberg, A. A. An Introduction to the Chemistry of Complex Com~wnds; Leach, J. R., Busch, D. H., Trimble, R. F., Eds.; Addlson-Wesley: Reading, MA, 1962. (33) Watt, G. W.; Cunningham, J. A. J . Electrochem. SOC. 1983, 110,
(36) (37) (38) (39) (40) (41) (42)
1959
Gllman, S. J . Phys. Chem. 1963, 6 7 , 78. Gilman, S. J . Phys. Chem. 1964, 68, 2098, 2112. Giiman, S. J . Electroanal. Chem. 1964, 7, 382. Gilman, S. Ektrochim. Acta 1964, 9 , 1025. Vrana, 0.; Klelnwachter, V.; Brabec, V. Talanta 1983, 30, 288. Bobteisky, M.; Eisenstadter,J. Bull. SOC. Chim. Fr. 1957, 708. Rozencwelg, M.; Von Hoff, D. D.; Abeie, R.; Muggia, F. M. In Cancer Chemotherapy Annual 2 ; Pinedo, H. M., Ed.; Eisevier: New York, 1980.
716. .
(34) Basoio, F.; Pearson, R. G. Mechanism of Inorganic Reactions, 2nd ed.; Wiley: New York, 1967. (35) Gllman, S. J . Phys. Chem. 1962, 66, 2657.
RECEIVED for review August 26, 1985. Resubmitted April 1, 1986. Accepted April 1, 1986.
Precision in Current Responses and Half-Wave Potentials for Some Common Pulse Polarographic Methods Leif Nyholm and Gunnar Wikmark* Department of Analytical Chemistry, Uppsala University, P.O. Box 531, S- 751 21 Uppsala, Sweden
An automated microcomputer-controlled polarographic analyzer has been used for the Investigation of the precision in differential pulse polarography (DPP), alternate drop differential pulse polarography (ADPP), normal pulse polarography (NPP), pseudoderhratlvenormal pulse polarography (PDNPP), and direct current polarography (DCP) In three solutions 3 X loa, and 3 X lo-' M Pb2+, reContaining 7 X spectively. For these solutions, normal pulse polarography gives the best relative precision In the current response and differential pulse polarography gives the most precise halfwave potentials. Standard devlatlons of less than 1 mV are obtalned for the hal-wave potential determined by DPP in the 7X M Pb2+ solution.
In a previous paper ( I ) , we investigated the precision in the current response of differential pulse polarography (DPP) and alternate drop differential pulse polarography (ADPP). It was shown that DPP gives better or equally good results as ADPP in half the recording time. A theory for the precision in ADPP and D P P based on the proportionality between noise and the faradaic current was presented. This theory predicts normal pulse polarography (NPP) to have better relative precision than DPP, when normal values of drop time and pulse time are used. The aim of the present work was, hence, to verify this and also to investigate the precision in the half-wave potential determined by the methods mentioned. Precise half-wave potentials are of utmost importance in the determination of complexation in solutions. A precision of 0.1 mV would be desirable if the stability constants should be accurately evaluated ( 2 ) ,but normally the reproducibility is not reported to be better than within f0.5 mV (2).In order to cover the practical concentration range for the methods, solutions of 7 X lo-' to 3 x lo4 M Pb2+have been studied. Systems containing 0.2 mM Pb2+,Cd2+,Cu2+,or T1+ were studied with fundamental harmonic ac polarography by Bond and Hefter (3) and the summit potentials were reported to be within 0.6 mV. In a study on the direct measurements of the reversible half-wave potential by second harmonic ac polarography and voltammetry, Bond and Smith (4) obtained a precision in the half-wave potential of h0.2 mV for a 0.2mM Pb2+solution. Better results can be obtained by application 0003-2700/86/0358-1959.$01.50/0
of computer-controlled instruments, when data can be recorded into the computer memory and numerical methods can be applied for the evaluation of the current responses and half-wave potentials. Thus Zollinger et al. ( 5 ) reported half-wave potentials with standard deviations of 0.3 mV, in a dc polarographic study on 80 p M solutions of Cd2+with computerized instrumentation and evaluation procedures. Bond and Grabaric (6) noted standard deviations in the DPP peak potential of fl mV and f2 mV for 1 p M and 0.1 pM solutions of Cd2+,respectively, in a study on computerized background correction techniques. Eliason and Parker (7) showed that high precision can be obtained for LSV peak potentials by use of analogue differentiation of the currentvoltage curves. The authors reported repeatabilities of f0.03 mV in the determination of the LSV peak potential for the reduction of 1.00 mM 1,l-diphenylethylene in N,N-dimethylformamide containing methanol. In another work Eliason and Parker (8) obtained a repeatability and a reproducibility of f0.2 mV in the reduction of perylene in acetonitrile. A second-order polynomial was fitted to the LSV peak to obtain the peak values of current and potential. EXPERIMENTAL SECTION Instrumentation. The automated microcomputer-controlled polarographic analyzer and the software used have been previously described ( I ) . The same DME capillary with drawnout tip (9) was used in all experiments. The DME had a mercury flow of 1.19 mg/s when measured in the supporting electrolyte at open circuit, at a column height of 450 mm. A platinum wire served as counter electrode and an Ag/AgCl(satd) electrode with a bridge containing 0.1 M NaN03 and 0.01 M "OB was used as reference electrode. All measurements were performed in a cell thermostated to 25.0 0.1 "C in a thermostated room. A dummy cell according to Figure 1 was constructed for the evaluation of the noise of the instrument. The resistances and capacitance were chosen to approximately equal the actual electrode system resistances and capacitances. If not otherwised stated, the instrumental parameters were as follows: drop time, 0.5 s; pulse duratioe, 60 ms; pulse height, -50 mV. The potential interval scanned was -100 to -500 mV and the scan rate was -1.92 mV/s. No filter was used when recording the polarograms. Chemicals and Solutions. All chemicals were of analytical grade quality. The water was deionized, distilled, and filtered through a Milli Q filtration system. The mercury and nitrogen were purified as described by Baecklund et al. (1). The same
*
0 1986 American Chemical Society
1980
ANALYTICAL CHEMISTRY, VOL. 58, R"
NO. 9, AUGUST 1986 Table I. 0.2609 mM Pb*+in Supporting Electrolytea
sol
COUNTER ELECTRODE
IlnA
e
day
REFERENCE ELECTRODE
drop
R ref
I
l
WORKING ELECTRODE
Figure 1. Dummy cell. Rref= 1 ki2 is the emulated resistance of the reference electrode, R,,, R , = 168 fl is the emulated solution resistance, R , = 68 fl is the uncompensated solution resistance, R,,
+
= 10 Mi2 emulates the charge current because of drop growth, and C,,= 1.0 pF is the emulated double layer capacitance.
standard solution with a nominal concentration of 0.2609 mM Pb2+was used throughout the whole work. Through dilution of the standard solution solutions with nominal concentrations of 2.610 pM and 0.6528 pM were obtained. All solutions were 0.1 M in sodium nitrate and 0.01 M in nitric acid. The background currents were obtained from measurements on solutions containing only sodium nitrate and nitric acid. Each sample analyzed consisted of 25 mL of solution to which 100 p L of a 0.1% Triton X-100 solution had been added. The samples were deaerated with nitrogen for 750 s initially and for an additional 30 s between the recordings. At the highest concentration of lead, the experiments consisted of seven to ten DPP and NPP polarograms alternately recorded on the same sample. At the lower concentrations ADPP polarograms were also recorded. The number of polarograms recorded each day at these concentrations was eight to ten. Evaluation Procedures. The evaluation procedures have been described by Baecklund et al. ( I ) . For the evaluation of the current responses three methods were employed, namely, the point of maximum current, the integrated current response, and the result of a nonlinear least-squares fitting of theoretical polarogram. As in the former work ( I ) , the last method gave the smallest standard deviations and, hence, only the results of this evaluation procedure have been presented. In addition to the evaluation of the current responses, the nonlinear least-squares routine was also used for determination of the half-wave potentials. The DPP half-wave potentials were calculated from the peak potentials by use of the equation given by Christie and Osteryoung (IO). Dc polarograms were obtained from the DPP samplings before the pulse. In a recent paper, Anderson and Bond (11) described a technique in software for the evaluation of pseudoderivative normal pulse polarograms (PDNPP) from NPP data. They concluded that PDNPP should give results similar to DPP (without the dc term (12))for reversible and irreversible reactions. Their technique has been used in the present work. NPP data points separated by 50 mV were sequeqtially subtracted to give the PDNPP polarogram. It should thus be noted that the DPP and DCP as well as the NPP and PDNPP polarograms are not independent and some covariation is to be expected. Further details on the Experimental Section have been given in the previous work ( I ) .
1 2 3
1 5 6
"
1-7
DPP
NPP
DCP
PDNPP
10 10 7 10 10 10 9
16740.14 16780.39 1676 0.32 1673 0.35 1702 0.33 1692 0.34 1680 0.25
23030.18 23000.22 2301 0.16 23000.28 2338 0.24 2318 0.14 2303 0.11
10860.27 10990.25 1098 0.30 10850.29 1116 0.34 1096 0.19
16950.58 16990.26 1606 0.41 16930.33 1724 0.30 1707 0.37 1698 0.47
7 (66)
1682 0.63 (0.31)
2309 0.61 (0.20)
1097 1.01 (0.28)
1702 0.63 (0.40)
.V
"Parameters are given in the text. The figures following the current values are the relative standard deviations in percent. The values within parentheses are the square roots of the sum of the weighted squares of the standard deviations of each day's experiments expressed in percent of the mean current response. N is the number of determinations. When a background polarogram is subtracted, the total variance will also contain a variance due to the sum of the current sampling variances for the background solution. Thus, an equation for the total variance is obtained u2 = UFl2 (rFZ2 CAI2 'TA22 (rBF12 OBF? UBAl2 + UBAS2 (4) Subscript "B" indicates noise originating from the background measurements. At high concentrations, the terms due to the absolute noise can be neglected as well as the terms due to the faradaic noise of the background. Expression 4 then reduces to
+
+
(rHgh2
=
u2 =
UF2
+ UA2
(1)
The term uF2 is due to a random noise, which is proportional to the faradaic current (this kind of noise is therefore called "faradaic noise") and uA2is a variance due to "absolute noise", which is assumed to be random and independent of the faradaic current present. Assuming that the current samplings are independent, an expression for the variance due to faradaic or absolute noise is obtained as a sum of the current sampling variances OF2 'TA2
=
UFl2 UA12
+ (TFz2 + UA22
(2)
(3)
UF12
+
+ (rFz2
+
(5)
In this situation, the ratios between the relative standard deviations for the methods can be calculated as described by Baecklund et al. ( I ) . At low concentrations, the absolute noise dominates and the faradaic terms can be neglected. The influence of the absolute noise will be equal in all methods and for all samplings since its variance originates from noise of the capacitive current and from instrumental noise. Thus, the following is valid at sufficiently low concentrations: 6A12
= UA2'
z uBAI2
(6)
Hence, the ratios between the absolute standard deviations for the different methods will depend only on the number of samplings made in the different techniques, since each sampling gives an equal contribution to the total variance. For DCP and NPP, the expression for the variance of the current will be response a t low concentrations, qOw2, uLow2
RESULTS AND DISCUSSION C u r r e n t Responses. As discussed in the previous work ( I ) , the variance u2 in the current response can be considered to be a sum of two terms
+
+
26A1'
= UA2
(7)
while the corresponding expression for DPP, ADPP, and PDNPP is 6L02
4 U ~ 1 'z 2uA2
(8)
Since the current responses are different for DCP and NPP, and for DPP and ADPP, the relative standard deviations Will, however, be different for the different methods. The results from the measurements of the current responses a t the different concentrations of lead in the supporting electrolyte are presented in the Tables 1-111. The data show that the repeatability of NPP is superior to that of the other methods, despite the sloping base line of the NPP polarogram~ seen in Figure 2. This base line does not limit the precision, as it is as reproducible as the base lines for DPP or ADPP. The reproducibility (the day-to-day variation) seems to be independent of the method chosen. This indicates that the variations mainly are due to changes in parameters which do
ANALYTICAL CHEMISTRY, VOL. 58, NO. 9, AUGUST 1986 DPP
Table 11. 2.610 pM PbZ+in Supporting ElectrolyteQ day 1 2 3 4 5 6
1-6
N 10 10 10 8 10 10
DPP
NPP
IlnA DCP
17.1 0.75 15.4 0.69 15.6 0.81 15.80.71 16.6 0.98 16.4 0.95
25.8 0.50 23.8 0.53 24.0 0.75 24.0 0.82 24.5 0.41 24.0 1.04
13.0 1.23 12.3 1.18 12.4 1.80 12.2 1.21 12.2 0.94 11.2 2.06
20nA
PDNPP
ADPP
17.3 1.22 17.4 1.16 17.8 1.10 17.8 0.74 17.4 0.55
8.2 1.17 7.7 0.69 7.8 2.60 8.0 1.50 8.0 2.80 7.8 2.60
6 16.2 4.07 24.4 3.07 12.2 4.76 17.5 1.37 7.9 2.32 (58) (0.83) (0.70) (1.44) (1.00) (2.07)
4.
aParameters are given in the text. The figures following the current values are the relative standard deviations in percent. The values within parentheses are the square root of the sum of the weighted squares of the standard deviations of each day's experimenta expressed in percent of the mean current response. N is the number of determinations.
20nA
Afler background correction
-200
N
DPP
NPP
I/nA DCP
1 2 3 4
10 10 10 8
4.5 3.7 4.2 3.4 4.5 3.5 4.5 4.1
5.5 2.2 5.5 1.0 5.8 1.2 5.9 2.0
2.4 2.2 2.9 2.7
6.1 6.8 7.9 4.2
1-4
4 (38)
4.4 3.4 (3.7)
5.7 3.6 (1.7)
2.6 12.2 (6.4)
-LOO
-300
Eh?
-500
ADPP
Table 111. 0.6528 p M Pb2+in Supporting Electrolytea day
1961
PDNPP
ADPP
5.5 3.9 5.1 5.1 5.4 2.8 5.0 4.0
2.1 4.0 1.7 4.0 2.1 5.2 2.0 6.3
5.2 4.5 (4.0)
2.0 9.6 (4.9)
oParameters are given in the text. The figures following the current values are the relative standard deviations in percent. The values within parentheses are the square roots of the sum of the weighted squares of the standard deviations of each day's experimenta expressed in percent of the mean current response. N is the number of determinations.
i
*OnA
b -300
w
Before background -500 correction E/mV
- 300 -LOO Figure 2. Differential pulse polarography (DPP), normal pulse polarography (NPP), and alternate drop differential pulse polarography (ADPP) on 0.6528 1M Pb2+ In 0.1 M NaNO,. The parameters were as follows: pulse height, -50 mV; drop time, 0.5 s; pulse duration, 60 ms; the scan rate, -1.92 mV/s. ment with the theoretical ratios for the current responses. The DPP and ADPP current responses are however somewhat low, resulting in high NPP/DPP and NPP /ADPP ratios. Similar deviations have been observed previously (1, 10). Employing the calculated theoretical quotients of the current responses in Table IV, quotients for the expected standard deviations have been calculated for either faradaic or absolute noise (Table V). For example, the NPP current response is theoretically 1.33 times bigger than the DPP response, with the parameters used (Table IV). Nevertheless, for faradaic noise we expect DPP and NPP to have the same relative standard deviation (RSD), since NPP has a larger current response but also a proportionally larger absolute standard deviation. For DPP the contribution from the first sampling to the total variance can be neglected for faradaic noise, since the DPP peak appears A E / 2 mV more anodic than the DCP half-wave potential (AI3 is the pulse amplitude) U ~ p p= KiNpp = 1.33Ki~pp (9)
not depend on the method, for example, capillary characteristics and contamination. The theoretical quotients between the current responses of the different methods have been calculated from equations given in the literature (10,13). In Table IV these quotients are compared with the experimental quotients obtained from the slopes of the calibration curves. The experimental factors were calculated by dividing the slope for NPP by the slopes for the other methods. The calibration lines were calculated from the mean current responses for each day, found in Tables 1-111. For the calculation a weighted linear least-squares program (Id),with weights inversely proportional to the variances of the current responses, was used. As seen in Table IV all calibration lines are linear and have small intercepts. The standard deviations of the experimental factors were estimated, assuming no covariation between the slopes, by the Gauss approximation equation, as follows: U2(NPP/Y) =Z (UK,NPP~/KY') + (u~,y~K~pp~/Ky~) where ?C is the variance, K is the slope and Y indicates DPP, NPP, DCP, PDNPP, or ADPP, respectively. The ratios between the slopes of the different methods are in good agree-
Background 1.
20nA
RSDNpp = uNpp/iNpp = u ~ ~ ~ / 1 . =3 K3 i (10) ~ ~ ~ (11) RSDDPP = uDPP/iDPP = iDPPK/iDPP = K u is the standard deviation and the constant K is assumed to
be equal for all methods.
Table IV. Slopes and Intercepts for DPP, NPP, DCP, PDNPP, and ADPP" DPP
NPP
DCP
PDNPP
ADPP
N 17 17 16 16 10 slope, nAlpM 6.3 (0.08) 9.0 (0.1) 4.4 (0.1) 6.5 (0.04) 3.1 (0.1) intercept, nA 0.23 (0.1) -0.20 (0.1) -0.33 (0.1) 1.01 (0.08) -0.05 (0.1) exptl factor 1.43 (0.02) 1.00 (0.02) 2.04 (0.05) 1.38 (0.02) 2.90 (0.1) theoretical factor 1.33 1.oo 2.05 1.33 2.60 'Parameters are given in the text. Numbers within parentheses are standard deviations. The standard deviations for the experimental factors have been estimated by use of the Gauss approximation equation as described in the text.
1962
ANALYTICAL CHEMISTRY, VOL. 58, NO. 9, AUGUST 1986
Table V. Relative Standard Deviation Ratiosa “faradaic noise”
DPP/NPP DCP/NPP PDNPP/DPP ADPP/DPP
1.0 1.0 1.2 2.2
Table VI. Dummy Celln
0.2609 2.610 0.6528 mM pM pM “absolute Pb Pb Pb noise’‘ 1.5 1.4 1.3
1.2 2.1
1.2 2.5
2.2 3.8 1.1 1.3
IJnA day 1
1.9 2.1 1.0 1.9
2 3 4 5
a The theoretical ratios for the two types of noise have been calculated with the parameters given in the text.
At very low concentrations, where the absolute noise is dominant, the theoretical ratio between the absolute standard deviations (DPP/NPP) should be bigger than unity. This is due to the fact that two samplings are made in DPP but only one in NPP. The variance in DPP should hence be twice as big as in NPP (eq 6,7, and 8) and the theoretical ratio between the absolute standard deviations is consequently 2 l l 2 . The theoretical relative standard deviation for DPP is thus RSDDpp 2 1 / 2 . 1 . 3 3 ~ N p p / i N p p x 1.9RSDNpp (12) The ratios for the other methods have been calculated in analogy with the DPP/NPP ratios as described above. DCP, NPP, and DPP will all have equal relative precision in situations where the faradaic noise is limiting. ADPP should give larger relative standard deviations than DPP at all concentrations (1). Since NPP,gives higher limiting currents than DCP, the PDNPP current samplings are bigger than the samplings in DPP, while the differential current responses are approximately equal. The relative standard deviation in PDNPP should therefore be bigger than in DPP for faradaic noise but equal for absolute noise. In cases where a low concentration is measured and the absolute noise has the greatest influence, N P P will give the lowest relative standard deviation. The superiority of N P P in this case is due to the single current sampling and the hgher current response obtained. DPP and PDNPP should have equal precision for this type of noise, since two samplings are made in both methods. In DCP only one current sampling is made and thus the absolute standard deviation is calculated to be smaller than for DPP and PDNPP. The current response of DCP is however smaller and hence the resulting theoretical relative standard deviation turns out to be equal to that of DPP and PDNPP, with the instrumental parameters used. The experimental ratios between the relative standard deviations presented in Table V are in good agreement with the ratios calculated for the main influence of faradaic or absolute noise. As seen in Tables 1-111, the day to day variations are significantly higher than the variations between the runs, especially at the lower concentrations. The standard addition technique could offer a way to reduce the day to day variations, since better precision is obtained when the solution is not removed from the DME. It seems, however, that the evaluation procedures (15-1 7) introduce uncertainties that diminish the applicability of the method. These aspects will be discussed in a future paper. As mentioned before, the absolute noise consists of noise from the capacitive current and the electronics in the instrument. When the background current is measured, only absolute noise would be present. By use of a dummy cell, thus avoiding influence of variations in drop size etc., an estimation of the instrumental noise can be obtained. In order to study this noise, some measurements were performed on the dummy cell. The values found for the DPP/NPP standard deviation ratio were 0.9, while the ADPP/DPP ratio was 1.6 (Table VI). The former value is lower than the theoretical ratio 1.4 and the latter value is
1-5
DPP
NPP
ADPP
10 10 10 10 10
5.5 0.27 5.6 0.21 5.5 0.23 5.5 0.20 5.7 0.19
25.9 0.37 26.3 0.30 26.1 0.18 26.1 0.19 26.0 0.18
0.42 0.27 0.07 0.61 0.31 0.19 0.41 0.26 0.33 0.24
5 (50)
5.6 0.10 (0.22)
26.1 0.15 (0.26)
0.31 0.14 (0.35)
N
‘Parameters are given in the text. Figures following the current values are the standard deviations. The values within parentheses are the square roots of the sums of the weighted squares of the standard deviations of each day’s experiments. N is the number of determinations. Table VII. Supporting ElectrolyteD IfnA
day
N
DPP
NPP
ADPP
1 2 3 4 5 6
5 8 5 8 3 10
13.3 0.3 13.2 0.3 13.0 0.1 12.8 0.2 12.8 0.1 12.8 0.2
30.1 0.1 30.8 0.4 30.0 0.4 29.7 0.5 29.7 0.4
0.76 0.2 0.51 0.3 0.57 0.3 0.71 0.2 0.34 0.2 0.45 0.4
1-6
6 (39)
13.0 0.2 (0.2)
30.1 0.4 (0.4)
0.56 0.2 (0.3)
nParameters are given in the text. Figures following the current values are the standard deviations. The values within parentheses are the square roots of the sums of the weighted squares of the standard deviations of each day’s experiments. N is the number of determinations. Table VIII. 0.2609 mM Pb2+in Supporting Electrolyten E/mV day
1 2 3 4 5 6 7 1-7
N 10 10 7 10 10 10 9
DPP -340.93 -340.99 -340.97 -341.12 -341.07 -340.78 -340.71
0.07 0.08 0.04 0.04 0.04 0.05 0.05
NPP -341.82 -341.94 -341.88 -341.99 -341.92 -341.64 -341.62
0.11 0.08 0.12 0.08 0.07 0.09 0.11
DCP -340.39 -340.16 -340.12 -340.64 -340.18
0.15 0.12 0.19 0.05 0.15
PDNPP
-341.91 -342.05 -342.02 -342.18 -342.08 -342.77 -339.77 0.18 -341.75
0.06 0.08 0.09 0.04 0.06 0.06 0.11
7 -340.94 0.15 -341.83 0.15 -340.21 0.29 -342.11 0.32 (66) (0.06) (0.09) (0.14) (0.07)
“Parameters are given in the text. The figures following the potential values are standard deviations. The values within parentheses are the square roots of the sums of the weighted squares of the standard deviations of each day’s experiments. N is the number of determinations. higher than the unity value predicted by the theory for absolute noise. The uncertainties in these measurements are, however, rather pronounced. In the calculation of the theoretical ratios we further assumed that there were no covariations between the current samplings in DPP and ADPP. Since the time between the current samplings is longer in ADPP, the covariance is likely to be less in ADPP. A lower covariation in ADPP would be even more likely for measurements with the DME since the current samplings in ADPP are performed on different drops. With covariation between the current samplings the DPP/NPP ratio would approach unity while the ADPP/DPP ratio would exceed unity. The values obtained in the dummy cell measurements presented in Table VI should be compared with the results
ANALYTICAL CHEMISTRY, VOL. 58, NO. 9, AUGUST 1986
1963
Table IX. 2.610 WMPb2+in Supporting Electrolytea day
N
1 2 3 4 5 6
10 10 10 8 10 10
1-6
6 (58)
~
DPP
-339.43 -339.93 -340.18 -340.11 -340.43 -340.06
EImV DCP
NPP 0.2 0.2 0.1 0.4 0.2 0.2
-340.02 0.3 (0.2)
-339.92 -339.98 -340.44 -339.74 -340.78 -340.27
0.3 0.2 0.4 0.6 0.3 0.3
-340.19 0.4 (0.4)
-332.76 -332.41 -334.78 -333.86 -336.08 -339.83
0.9 0.6 0.7 0.9 0.4 0.6
-334.95 2.7 (0.7)
PDNPP -342.30 -341.63 -342.04 -341.65 -341.29
0.5 0.6 0.4 0.3 0.2
-341.78 0.4 (0.4)
ADPP -341.23 -341.00 -341.46 -341.87 -341.98 -341.13
0.6 0.8 0.5 0.4 0.9 0.7
-341.45 0.4 (0.7)
'Parameters are given in the text. The figures following the potential values are standard deviations. The values within parentheses are the square roots of the sums of the weighted squares of the standard deviations of each day's experiments. N is the number of determinations. Table X. 0.6528 pM Pb2+in Supporting Electrolytea day
N
1 2 3 4
10 10 10 8
1-4
4 (38)
DPP -341.06 -340.82 -339.64 -339.15
EImV DCP
NPP 0.6 0.8 0.6 0.6
-340.17 0.9 (0.7)
-341.48 -344.74 -340.07 -338.23
0.9 1.1 0.7 1.3
-341.13 2.8 (1.0)
-344.08 -346.71 -335.80 -336.44
2.8 3.6 2.9 3.4
-340.76 g.5 (3.2)
PDNPP -342.08 -344.71 -341.97 -339.97
0.7 1.0 0.6 1.0
-342.18 1.9 (0.8)
ADPP -340.53 -340.20 -341.68 -341.91
0.7 1.2 0.7 1.1
-341.08 0.8 (0.9)
'Parameters are given in the text. The figures following the potential values are standard deviations. The values within parentheses are the square roots of the sums of the weighted squares of the standard deviations of each day's experiments. N is the number of determinations
.
of the measurements on the background solution in Table VII. Theoretically NPP should have the lowest noise, as discussed before, but this is not the case either in the background or the dummy cell measurements. As seen in Table VII, DPP gives the smallest standard deviation in the current response for the supporting electrolyte while NPP gives the largest value. This is also in accordance with the measurements on the dummy cell. The similarities between the results from the different measurements indicate that it is the instrumental noise rather than the capacitive current noise that determines the lowest absolute noise level. Half-WavePotentials. Eliason and Parker (7,8) obtained better precision in the LSV peak potential after differentiation of the current potential curves. This is in analogy with our results. The nonlinear least-squares refinement gives lower standard deviations for the peak-shaped polarograms (DPP, ADPP, or PDNPP) than for the sigmoid-shaped ones (NPP or DCP). The half-wave potentials are clearly best evaluated by DPP in the concentration range investigated. The evaluated half-wave potentials at the different concentrations are shown in Tables VIII-X. At a concentration of 0.2609 mM Pb2+a repeatability (within-run standard deviation) of 0.06 mV is obtainable in DPP while the corresponding reproducibility (between-run standard deviation) is 0.15 mV. It is still possible to detect the half-wave potential with a precision of better than 1 mV a t a concentration of 0.6528 p M Pb2+. To obtain better precision of the half-wave potential in the NPP mode, the software evaluation of PDNPP polarograms from NPP polarograms was introduced. As shown in Tables WI-X, the PDNPP evaluation actually gives better precision in the half-wave potential, although not quite as good as that obtained from the DPP polarograms. This makes it possible to obtain high precision in both current response and half-wave potential from only NPP measurements by employing NPP and PDNPP evaluations. At the highest concentration there is a difference between the half-wave potentials evaluated from the DPP and NPP
polarograms. We have no explanation for this but Galvez (18) found that sphericity effects in amalgam forming systems lead to different half-wave potentials for DPP and NPP. This is also deduced from a paper by Oldham and Bond (19). However, these treatments predict DPP to have more negative half-wave potentials than NPP and the opposite is exhibited by the data presented in this paper. CONCLUSIONS
NPP is the pulse method with the smallest relative standard deviation for solutions containing 0.3 mM to 0.7 pM Pb2+when the background can be accurately recorded. The results thus confirm the presumptions in a previous paper by Baecklund et al. (I). The use of NPP instead of DPP for precise measurements of current responses is also favored by the fact that NPP is less sensitive to changes in reversibility and to adsorption effects, because of its limiting current plateau. Evidently, the reproducibility of the background and the current response determines the analytical applicability of the method rather than the shape of the background as suggested by Christie et al. (13). ACKNOWLEDGMENT The authors gratefully acknowledge valuable discussions with Ake Olin. LITERATURE CITED (1) (2)
Baecklund, P.; Nyholm, L.; Wlkmark, G. Anal. Chem. 1984, 56, 1209. Crow, D. R. polarography of Metal Complexes; Academic Press: New York, 1969.
(3) Bond, A. M.; Hefler. 0. J . flectfoanal. Chem. 1972, 3 4 , 227. (4) Bond, A. M.; Smith, D. E. Anal. Chem. 1974, 46, 1946. (5) Zolilnger, D. Ph.; Bos, M.; Van Veen-Blaauw, A. M. W.; Van Der Linden, W. E. Anal. Chlm. Acta 1985. 167, 89. (6) Bond, A. M.; Grabarlc. B. S. Anal. Chem. 1979, 5 1 , 337. (7) Eliason, R.; Parker, V. D. J . flectroanal. Chem. 1984, 165. 21. (8) Eliason, R.; Parker, V. D. J . flectroanal. Chem. 1984, 170, 347. (9) Meites, L. Polarographic Techniques, 2nd ed.; Wiiey: New York, 1965. (10) Parry, E. P.; Osteryoung,R. A. Anal. Chem. lB85, 3 7 , 1634. ( 1 1 ) Anderson, J. E.; Bond, A. M. Anal. Chem. 1981, 53. 504. (12) Christie, J. H.; Osteryoung, R. A. J . flectroanal. Chem. 1974, 49, 301.
1964
Anal. Chem. 1986, 58, 1964-1969
(13) Christie, J. H.; Jackson, L. L.; Osteryoung, R. A. Anal. Chem. 1976, 48, 242. (14) Garden, J. S.;Mitchell, D. G.; Milis. W. N. Anal. Chem. 1980. 52,
(17) Miller, J. C.; Miller, J. N. Statistics for Analytical Chemlsfry; Wiiey: New York, 1984. (18) Galvez, J. Anal. Chsm. 1985, 57, 585. (19) Oldham. K. B.; Bond, A. M. J . Electroanal. Chem. 1983, 758, 193.
2310.
(15) Larsen, I. L.; Hartmann, N. A,; Wagner, J. J. Anal. Chem. 1973, 45,
1511. (16) Franke, J. P.; de Zeeuw, R. A,; Hakkert, R. Anal. Chem. 1978, 50, 1374.
RECEIVED
for review August 27, N85. Accepted March 20,
1986.
Polarographic Reduction and Determination of 1,2=Dibromoethanein Aqueous Solutions Robert Tokoro,l Renata Bilewicz? and Janet Osteryoung*
Department of Chemistry, State University of New York at Buffalo, Buffalo,New York 14214
The electrochdcal reduction of 1,P-dlbromoethane (EDB) In aqueous solutlon has been studied with dc, normal pulse, and reverse pulse polarography, vdtammetry, and coulometry. The reductlon mechanlsm Involves slow two-electron transfer wlth fast d l o p l a c W of bromide leading to formation of ethene. Resutts of detalled RPP lnvestlgatlons In neutral and slightly acid solutions suggest that the flnal produds are formed through a carbanbn Intermediate. EDB can be detemdned dkectty from the llmlthg redudion current or lndkectly from the llmltkrg curren! lor the anodk oxklatbn of mercury In the presence ofthe redudkn product, bromkle. The detection Ihtt Is ca. 1 @A. The method Is used to determine the OdUMltty of EDB In water, whlch was found to be 3.1 mg/g of H,O.
In recent years there has been concern for the environmental and toxicological effects of 1,2-dibromoethane (EDB). As it was found to cause cancer, birth defects, and genetic disorders in test animals, recommended acceptable levels in grain and food have been established (I). Widespread use of EDB as a fumigant and a scavenger added to leaded gasoline focused the attention on its determination in foodstuff (2-4) and polluted atmosphere (5-7). A number of analytical procedures have been presented with the determinative step via gas chromatography coupled with electron capture detection, electrolytic conductivity, or plasma emission spectrometry (8, 9). Of these methods the electron capture detector yields the best detection limits. Difficulties arising in this technique are associated with the long times used in the chrnmatographic steps and interferences from solvent impurities and coextractables. Though retention times may be reduced, it is difficult to perform a sufficiently rapid analysis with good resolution from the solvent front ( 2 ) . The purity of solvent seems to be critical as the impurity peaks are close to EDB retention times (10)and coextractables may interfere with the confirmation by mass spectrometry (8). The lack of any electrochemical procedure for the determination of EDB prompted the present work. Electrochemical Permanent address: Instituto de Quimica USP B-843,Caixa postal 20780, Sa0 Paul0 SP OlOOO, Brazil. Permanent address: Department of Chemistry, University of Warsaw, 02093 Warsaw, Pasteura 1, Poland. 0003-2700/86/0358-1964$01.50/0
studies of EDB were first carried out by von Stackelberg (11). Meites (12) and Feoktistov (13) have reported a few studies of EDB since that time. Our work was performed to study the electrochemical behavior of 1,2-dibromoethanein aqueous solutions and then to establish the best electrochemical conditions for analytical purposes. EXPERIMENTAL SECTION Reagents. 1,2-Dibromoethane (Fisher Scientific Co. E 173 Certified) was of 99.5% purity (density 2.179Z4%g/cm3). Ethanol was supplied by US. Industrial Chemicals Co. (U.S.P. specifications). Distilled water was purified further by a Millipore MilliQ system (resistivity > 100 Mi2 cm). Sodium bromide (Mallinckrodt Analytical Reagent), tetraethylammonium perchlorate (TEAP, Eastman Kodak), tetrabutylammonium perchlorate (TBAP, Eastman Kodak), sodium chloride (Baker Analyzed Reagent), sodium perchlorate, purified (Fisher Scientific), and lithium perchlorate, anhydrous (Baker reagent analyzed), were used as supporting electrolytes. Catalyst R-3-11 (Chemical Dynamics Corp.) was used to remove oxygen from argon (99.998%, Liquid Carbonic). Perchloric acid (Fisher Scientific) and mercury (Bethlehem Apparatus) were used as supplied. Instruments and Accessories. The following were used: EG&G PARC Model 174 polarographic analyzer and conversion of module (14,151 with EG&G PARC Model 174/70 drop timer assembly 174/70; Tacussel Electronique Polaropulse PRGB with homemade converter to PARC 174/70; EG&G PARC Model 173 potentiostat/galvanostat with Model 179 digital coulometer; Metrohm Calomel reference electrode, EA404, NaCl 3 M; Eppendorf micropipeta, 10-100 rL. In some normal pulse (NP) and reverse pulse (RP) polarographic experiments a computer-controlled pulse-voltammetric instrument based upon a Digital Equipment Corp. PDP-8/e minicomputer and homemade interface were used. The system has been described elsewhere (16). Gaseous products were identified by use of gas chromatography with a Perkin-Elmer SIGMA 3-B chromatograph and a 180 cm X 3 mm column packed with 20% Apiezon L on 60-80 mesh Chrom-W (column inlet temperature 150 "C, column temperature 40 "C, detector temperature 200 "C). Procedures. All potentials are reported vs. the 3 M NaCl calomel electrode. Measurements were carried out on two dropping mercury electrodes with the following characteristics: (1)natural drop time ( t d ) = 7.20 s, flow rate ( m )= 0.9 mg/s, (2) t d = 6.17 s, m = 1.14 mg/s. Voltammetric experiments were carried out on a HMDE, Kemula and Kublik type (17) with surface area 0.024 cm2. A Luggin capillary with Vycor tip containing 0.1 M TEAP was used to avoid contamination of the working solution with NaCl from the reference electrode. The solutions were thermostated at 25 0.2 "C. Oxygen was removed from solutions by sparging with purified argon (18). Pure EDB
*
0 1986 American Chemical Society