962
Anal. Chem. 1907, 59, 962-964
be obtained experimentally. Such a potential is high enough to construct the plot up to the inflection point (0.264 V as determined by square wave voltammetry). The second half of the S-shaped plot was estimated as explained above. The slopes for the plateau region of the plot corresponding to pulse potentials of 0.408, 0.428, and 0.468 V were used to calculate an average electrochemical area using the Cottrell equation. To avoid interferences from adsorption, the potential pulse width was 250 ms. Extending the pulse duration up to 1 s did not significantly change the slope values, however. In calculation of the slope of Anson’s plots, the BAS-100 ignores the first 20% of collected data and the slope measurement is then carried out in a time window where the effect of adsorption on the determination of the electrochemical area is of no consequence (10). Correlation coefficients were between 0.9970 and 0.9999. With the approach just outlined, untreated CP-01 surfaces showed electrochemical areas of 0.19 f 0.01 cm2(0.21 cm2geometrical area) compared with electrochemical areas of 0.43 f 0.01 cm2 after surfactant treatment. The removal of the film of pasting ingredient seems to increase the number of carbon centers available for electron exchange at the surface of the electrode. Increases in electrochemical area have been observed when polished electrodes based on Kel-F as inert ingredient were roughened and when the ratio of graphite to Kel-F was increased in the paste composition (11).
Solutions of hexacyanoferrate(II1) in acidic media (HC1+ 1.0 M KC1) of pH 1.5-3.0 did not show much improvement in current by treatment with surfactant solutions 0.1% (w/v) for 10 min. Increasing the surfactant concentration to 1% (w/v) and the treatment time to a t least 30 min, however, resulted again in an improved response of the treated surfaces.
LITERATURE CITED (1) Adams, R . N. Electrochemistry at Solid Electrodes; Marcel Dekker: New York, 1969; p 26. (2) Kissinger, P. T.; Refshauge, C.: Dreiling, R.; Adams, R. N. Anal. Lett. 1973, 6 , 465-477. (3) Albahadily, F. N.; Mottola, H. A. J. Chem. Educ. 1986, 6 3 , 271-273. (4) Rice, M. E.; Galus, 2.; Adams, R. N. J. Nectroanal. Chem. 1983, 143, 69-102. ( 5 ) Ravlchandran,R.; Baldwin, R. P. Anal. Chem. 1984, 56, 1744-1747. (6) Engstrom, R. C. Anal. Chem. 1982, 5 4 , 2310-2314. (7) Urbaniczky, C.; Lundstrom, K. J. Nectroanal. Chem. 1984, 176, 169-1 62. (8) Operationllnstallatlon /Maintenance Manual for the BAS- 100 Electrochemical Analyzer; Bioanalytical Systems, Inc.: West Lafayette, IN, 1983; Appendix E, pp E-1 and E-2. (9) Lindquist, J. J. Electroanal. Chem. 1974, 52, 37-46. (10) Rusllng, J. F.; Brooks, M. Y. Anal. Chem. 1984, 56, 2147-2153. (1 1) Anderson, J. E.; Tallman, D. E.; Chesney, D. J.; Anderson, J. L. Anal. Chem. 1978, 50, 1051-1056.
RECEIVED for review April 24, 1986. Resubmitted August 25, 1986. Accepted December 2, 1986. This research has been supported by the Office of Basic Energy Sciences, U.S. Department of Energy (Grant No. DE-FG05-85ER13346).
Rotating Ring-Disk Electrode Method for the Quantitative Determination of Nitrate Ions in Aqueous Solutions in the Submicromolar Range Xuekun Xing and Daniel A. Scherson* Case Center for Electrochemical Sciences and T h e Department of Chemistry, Case Western Reserve University, Cleveland, Ohio 44106
A rotating rlng-disk technique has been developed for the analytical determlnatlon of nitrate Ions In acid medla. I t Is based on the measurement of the ring currents associated with the oxldatlon of nitrite Ions generated by the reduction of nitrate Ions on a gold-dlsk electrode covered by a layer of underpotentlaliy deposited cadmlum. The ring currents were found to be proportlonal to the amount of nitrate In solution for concentrationsIn the mlllhnolar range down to 20-30 ppb.
Two classes of electrochemically based techniques have been employed in the quantitative analysis of nitrate ions in aqueous media (1-19). The first is based on polarographic measurements conducted in the presence of certain polyvalent cations such as Zr02+, La3+, Yb3+, V02+,and Ce3+, which appear to promote the rates for nitrate reduction (1-15), whereas the second relies on cyclic voltammetry or dynamic polarization curves utilizing solid electrode materials capable of catalyzing the reduction of NO3- (16-19). Davenport and Johnson (16,17), for instance, using a rotating cadmium-disk electrode obtained linear plots for the current as a function of the concentration of nitrate in the range of M. In related studies, Bodini and Sawyer (18) observed that the
reduction of NO3- was catalyzed by the simultaneous bulk deposition of cadmium and copper in stagnant solutions by using a pyrolytic graphite electrode as a substrate. Despite the fact that the residual currents in this case were found to be quite substantial, these authors were able to construct linear plots of the peak currents obtained from cyclic voltammetry as a function of [NO;] even down to concentrations of about 60 ppb. Sherwood and Johnson (19) extended these studies to include rotating Cd-disk electrodes that had been coated with a metallic Cu layer prior to the analytical determination and obtained good resolution and sensitivity for [NO,-] as low as M. One disadvantage associated with some of these nonpolarographic methods is that the potentials involved are negative enough for the measured currents to contain contributions from other processes, such as hydrogen evolution and bulk deposition of metal species, for which account is difficult to be made especially for low nitrate concentrations. The method to be presented in this communication is based on the ability of underpotentially deposited cadmium to electrocatalyze the reduction of nitrate ions (20). The actual determination of [NO,-] relies on the quantitative oxidation of nitrite ions generated by the reduction of nitrate on a rotating gold-disk electrode, on a concentric gold-ring electrode polarized at potentials at which the residual currents are
0003-2700/87/0359-0962$01.50/00 1987 American Chemical Society
ANALYTICAL CHEMISTRY, VOL. 59, NO. 7, APRIL 1, 1987
I
I
f
Ed/V v5 SCE
-07 -06 -05 -04 -03 -02
Figure 1. Polarization curves for the reductlon on nitrate on a Au rotatingdisk electrode (RDE) in a 0.1 M NaCIO,, 1 X M Cd(CIO,),, and 5 X lo-, M KNO, solution containing (a) 0.5, (b) 0.6,(c) 0.7,(d) 0.8,(e) 0.9,and (f) 1.0 mM HCIO,. Conditions: rotation rate, f = 900 rpm; scan rate, u = 10 mV-s-'; disk area, 0.196 cm2.
lo3
,
l
,
I
,
I
,
l
.
I
963
.
Figure 2. Polarization curves for the reduction of nitrate on a Au RDE in a 0.1 M NaCIO,, 1 X lo-, M Cd(C104)2, and 1 X lo-, M HCIO, solutlon containing (a) 20, (b) 30, (c) 40, (d) 50, and (e) 75 pM NaNO,, where f = 1600 rpm and other conditions are the same as those given in Figure 1.
essentially zero. This technique has made it possible not only to accurately detect nitrate in the submicromolar range but also to obtain linear plots of the ring-limiting current vs. [NO;] spanning over 3 orders of magnitude in concentration.
EXPERIMENTAL SECTION A Au-disk Au-ring electrode designed and manufactured in this laboratory was used in all the measurements. The surface was polished by using sandpaper of different grades followed by 1-, 0.3-, and 0.05-pm alumina powder. The collection efficiency evaluated from experiments involving the ferrous/ferric redox couple yielded a value of 0.38 in excellent agreement with that predicted based upon the actual geometry of the electrode (rl = 0.250, r2 = 0.275, and r3 = 0.350 cm) (21). The chemicals used were ACS Reagent Grade except for HClO.,, which was of U1trapure quality (J. T. Baker). A solution of 0.1 M NaC10, and 0.001 M HClO, was used as the supporting electrolyte in all the measurements. The NaNOS,KNOB,and NaN02 were dried at 100 OC for 2 h before preparing the 0.1 M stock solutions. The cadmium perchlorate was added to the electrolyte from a 0.1 M Cd(C104)2stock solution. Ultrapure water was obtained by reverse osmosis followed by distillation in a modified Gilmont system using a nitrogen counter flow to purge volatile impurities. The electrochemical experiments were conducted in a conventional three-compartment cell with a gold foil as a counter and a standard calomel as a reference electrode, respectively. The counter electrode was isolated from the main cell compartment by means of a fine frit whereas the reference electrode was connected through a salt bridge filled with the supporting electrolyte and a Luggin capillary. A dual potentiostat with a built-in triangular signal generator (Pine RDE 3, Pine Instruments Co.) and an X-Y-Y recorder (Yokogawa Model 3036) were used in all the experiments. All measurements were conducted at room temperature.
RESULTS AND DISCUSSION The electrocatalytic effects for the reduction of nitrate associated with the underpotential deposition of Cd on either Au or Ag electrodes have been described in a recent communication (20). Additional studies conducted in this laboratory have indicated that the forced convection diffusion limiting current, iL, for the reduction of nitrate is proportional to the concentrations of protons in solution for [NO3-] >> [H+] (see Figure 1). In the case of [H+] >> [NO3-], however, no well-defined limiting currents were observed as shown in Figure 2. Nevertheless, a linear relationship was obtained between the currents measured at -0.6 V and [NO3-]. These observations appear to be consistent with the results reported earlier by Pletcher and Poorabedi (22),who obtained nitrate reduction limiting currents that were proportional to [NO3-]
E,/ V vs SCE
1-06 -0.4 -02
0
02 04 0.6 08
1.0
Figure 3. Dynamic polarization curves for the ring electrode of a rotating Au (disk)/Au (ring) electrode (RRDE). Collection efficiency (N) = 0.38 in a 0.1 M NaCIO,, 1 X M Cd(CIO,),, 1X M HCIO,, M KNO, for the Au disk polarized at (a) 0.00 V (---) and and 2 X (b) -0.62 V vs. SCE (-). Other conditions are the same as those given in Figure 1.
lo3
lo3
lo-,
for [H+] >> [NO3-] on Cu electrodes, and those reported by Mayer et al. (23)for silver surfaces in which case the limiting currents associated with the same redox process were found to be proportional to [H+] for [NO3-] >> [H+]. Although the results shown in Figure 2 provide in principle a means for the quantitative determination of nitrate in solution, the occurrence of residual currents associated presumably with hydrogen evolution and/or the reduction of adventitious oxygen would make the data analysis particularly unreliable in the case of very low nitrate concentrations. One possible means of overcoming these difficulties is by monitoring the concentration of the nitrate reduction product by using a ringdisk electrode arrangement. Figure 3 gives the results of an experiment in which the Au-ring electrode was scanned in a wide potential range while the Au-disk electrode was polarized at 0.0 V (curve a) and -0.62 V vs. SCE (curve b), a voltage negative enough for the reduction of nitrate to occur. As can be clearly seen, the product of the reduction undergoes reoxidation at about +0.70 V vs. SCE, reaching a well-defined diffusion limiting current at more positive potentials. The main product of the electrochemical reduction of nitrate under the conditions of this experiment based on the analysis of the results shown in Figures 1 and 2 appears to be nitrite. Additional evidence in support of this view was provided by rotating-disk electrode experiments involving nitrite-containing solutions in the same supporting electrdyte in which case the onset for the oxidation was found in essentially the same potential range as that shown in curve b, Figure 3. Figure 4 gives the results of rotating ring-disk measurements conducted for two different nitrate concentrations (-,
964
ANALYTICAL CHEMISTRY, VOL. 59, NO. 7, APRIL 1, 1987 I
I
Fa6 I
I
I
-04 I
I
I
/
-02 I
I
I
i
0 I
i
I
I
02 I
I
TQL?J~A
0 I
I
I
I
I
i
l
I
I
Ed/V vs SCE
Figure 4. Ring-disk polarization curves for nitrate reduction on an Au/Au RRDE electrode, N = 0.38, in a 0.1 M NaCIO,, 1 X lo-, M HCIO, solution (- - -), and this same solution containing 0.1 mM Cd(CIO,), 0.1 mM Cd(CIO,), and 2 pM NaNo, (-), and 0.1 mM Cd(CIO,), and 6 pM NaNO, (- -), where f = 1600 rpm, u = 10 mV s-', E , = 1.0 V vs. SCE.
-
(..a),
-
2 X lo4 M; - -, 6 X M). The ring was polarized in this case at 1.0 V vs. SCE, a potential at which the oxidation of nitrite is expected to occur under pure diffusion control (24, 25). As shown in the figure, very well-defined ring-limiting currents were observed in contrast to the corresponding disk currents for which no plateau could be easily identified. The dashed and dotted curves in Figure 4 were obtained in the absence and presence of Cd(C10& in the supporting electrolyte, respectively. A series of rotating ring-disk experiments were then performed in order to measure the oxidation currents at the ring electrode, i,, as a function of the concentration of nitrate in solution. The results are shown in Figure 5 spanning in each case an order of magnitude in [NO,-]. The plots of i, vs. [NO,-] for a rotation rate of 1600 rpm were found to be extremely linear in the concentration range between 1 x lo4 and 1 X M (slope, S = 101 f 2 A.mol-l.cm3; intercept, I = 0; corr coeff, CC = 0.9998). At concentrations of nitrate M a similar plot also yielded in the range 1 X to 5 X a straight line with essentially identical parameters except that S was found to be 96 f 2 A.mol-l.cm3 (see Figure 5C). At even higher nitrate concentrations a lower rotation rate was employed in order to decrease the flux of electroactive species to the electrode surface and thus achieve complete convective diffusion control a t the same potential as that used for lower nitrate concentrations. This is clearly shown in Figure 5C for experiments conducted at 900rpm for which the relative slope ratio obtained experimentally was precisely that expected based on the square root dependence of the diffusion controlled limiting current ( S = 7 5 f 2 A.mol-'.crn3, I = 0, CC = 0.9998). It may be interesting to note that the Cd2+base electrolyte appears to contain about 20-30 ppb of adventitious nitrate as judged by the magnitude of the ring currents observed in the absence of intentionally added nitrate to the solution (see dotted curve in Figure 4). Several anions and cations were found to interfere with the analytical method presented. The addition of chloride or sulfate ions to the solution, for instance, resulted in a decrease in the magnitude of the disk and, consequently, of the ring current. This is most probably due to changes in the underpotential deposition isotherms due to specific adsorption
Figure 5. Ring currents, i,, vs. E , curves for an Au/Au RRDE, N = 0.38, in a 0.1 M NaCIO,, 1 X lo-, M Cd(CIO,),,and 1 X M HCIO, solution containing: {A) (a) 0, (b) 2, (c) 4,(d) 6,(e) 8, and (f) 10 pM NaNO,; {B)(a) 20, (b) 30, (c) 40, (d) 50, and (e) 75 pM NaNO,; {C] (a) 0.1,(b) 0.2, (c) 0.3, (d) 0.4,(e) 0.5,and (f) 0.6 mM NaNO,; u = 10 mV s-'; E , = 1.0 V vs SCE. Also shown in C are the i , vs. [NO,-] calibration plots obtained at 1600 rpm (0)and 900 rpm (m).
of the anions. The reduction of nitrate was also depressed in the presence of TI+in the electrolyte, whereas Pb2+totally suppressed the electrocatalytic effects.
LITERATURE CITED (1) Kolthoff, I. M.; Harris, W. E.; Matsuyama, G. J . A m . Cbem. Soc. 1944, 6 6 , 1782. (2) Collat, J. W.; Lingane, J. J. J . Am. Cbem. Soc. 1954, 7 6 , 4214. (3) Johnson, M. G.;Robinson, R . J. Anal. Cbem. 1952, 2 4 , 366. (4) Mechelynck Ph.; Mechelynck, C. Anal. Cbim. Acta 1959, 2 1 , 432. (5) Wharton. H. W. J . Elecfroanal. Chem. 1985, 9 , 134. (6) Kolthoff, I. M.; Hodara, I. J . Electroanal. Cbem. 1963, 5 , 2. (7) Lundquist, G. L.; Washinger, G.;Cox, J. A. Anal. Cbem. 1975, 4 7 , 319. (8) Boese, S.W.; Archer, V. S.; O'Laughlin, J. W. Anal. Cbem. 1977, 4 9 , 479. (9) Cox A,; Brajter, A. Electrocbim. Acta 1979, 2 4 , 517. (IO) Timofava, 2. N. Elektrokbimiya 1979, 15, 876. (11) Zhdanov, S. I. Russ. J . f b y s . Cbem. (Engl. Transl.) 1983, 3 7 , 200. (12) Liu Shou-jung; Zhdanov, S. I . Russ. J . Pbys. Cbem. (Engl. Transl.) 1983, 3 7 , 943. (13) Kvaratskheliya, R. K.; Gabriadze-Machavariani, T. Sh. Elektrokhimiya 1980, 16, 600. (14) Knoeck, J. Anal. Cbem. 1969, 4 1 , 2069. (15) Wang, E.; Lin, X. J . Elecffoanai. Cbem. Interfacial Electrochem. l.Q - -8-,2 . 136. . - - , 311. - . .. (16)Davenport. R. J.; Johnson, D. C. Anal. Cbem. 1973, 4 5 , 1979. (17) Davenport R. J.; Johnson, D. C. Anal. Cbem. 1974, 4 6 , 1971. (18) Bodini M. E.; Sawyer, D. T.Anal. Chem. 1977, 4 9 , 485. (19) Sherwood, G.A., Jr.; Johnson, D. C. AnalCbim. Acta 1981, 729, 87. (20) Xing, X.; Scherson, D. A. J . Elecfroanal. Cbem. Interfacial Elecfrocbem. 1988. 799, 485. (21) Albery, W. J.; Hitchman, M. L. Ring-Disc Electrodes; Clarendon: Oxford, U.K., 1971;p 156. (22) Pletcher, D.; Poorabedi, 2 . Electrocbim. Acta 1979, 2 4 . 1253. (23) Mayer. C.;Juttner. K.; Lorenz, W. J. J . Appl. Electrochem. 1979, 9 . 161. (24) Harrar, J. E. Anal. Cbem. 1971, 4 3 , 143. (25) Guidelli, R.; Pergola, F.; Raspi. G. Anal. Cbem. 1972, 4 4 , 745.
RECEIVED for review September 29,1986. Accepted December 8,1986. Partial support for this work was provided by an IBM Faculty Development Award to D.S.