SpectroelectrochemicaI Studies of MetaI Deposition and Stripping and of Specific Adsorption on Mercury-Plati num OpticalIy Transparent Electrodes William R. Heineman and Theodore Kuwana Department of Chemistry, The Ohio State Unioersity, Columbus, Ohio 43210 Optical and electrochemical characteristics of the mercury-platinum optically transparent electrode (Hg-Pt OTE), as applied to the deposition of a metal into the thin mercury film, are evaluated. Substantial mercury character can be achieved with film thicknesses of as little as 10 mC of mercury/cm2 (ea. 150 A), as evidenced by the stripping behavior for lead. The use of the Hg-Pt OTE for the evaluation of molar absorptivities of metals dissolved in mercury and the detection of ionic surface excess at the electrode-solution interface is described. Light passing through the Hg-Pt OTE during the diffusion controlled reduction of metal ions is attenuated by the accumulation of electrodeposited metal in the thin mercury film. The rate of this attenuation is related to the molar absorptivity of the metal in mercury. Molar absorptivities for Pb, Cd, TI, and Zn which were determined in this manner are compared with reported values for the bulk metal. The existence of a surface excess of a metal ion can be detected by a perturbation on the transmission absorbance-time curve. This is quantitatively demonstrated for P b B r p . A step change in the applied potential was also found to produce an optical perturbation which i s attributed to the attendant change in the surface concentration of non-electroactive ionic species such as NO,- and Br-. Signal averaging was necessary to resolve the small optical responses involved. Use of the Hg-Pt OTE for stripping analysis is considered.
INA PREVIOUS COMMUNICATION, the preparation and properties of a mercury-platinum optically transparent electrode (Hg-Pt OTE) were described ( I ) . The electrode consisted of a Pt OTE upon which a thin film (ca. 50-500 A) of mercury was electrodeposited. The structure of the Hg-Pt OTE was represented as a thin film of mercury which was bound to the underlying film of platinum substrate by a layer of PtHg, (Cf: Figure 3, ref. 1). The relationship between change in absorbance with the amount of mercury deposited on a Pt OTE was given. Deposition of the mercury film resulted in an increase in hydrogen overvoltage of up to 400 mV, yet optical transparency was retained. The advantage of the increased hydrogen overvoltage of the Hg-Pt OTE was illustrated for the reduction of methyl viologen dication and Cd2+; these reductions being obscured by hydrogen evolution on a Pt OTE with no mercury deposited. Spectral monitoring of the methyl viologen cation radical MV.+ and of amalgamated Cd" were accomplished by direct transmission spectrophotometry through the OTE. Thus, the electrode could be used for spectral observation of products of heterogeneous electron transfer reactions in which these products are either retained in the diffusion layer of the solution or are amalgamated into or deposited on the thin mercury film. The present paper embodies a more detailed study of the optical and electrochemical characteristics of the Hg-Pt OTE, particularly, as applied to the deposition of a metal into the (1) W. R. Heineman and T. Kuwana, ANAL.CHEM., 43,1075 (1971). 1972
thin mercury film. Signal averaged optical changes which are interpreted as the optical observation of adsorbed reactants are included. EXPERIMENTAL
Apparatus. Because of the somewhat fragile nature of the Hg-Pt OTE, it was found useful to employ an electrochemical cell into which deoxygenated solutions could be readily transferred immediately following deposition of the thin mercury film. Such a cell is illustrated in Figure 1 . The cell body was formed by boring a a/8-in.diameter hole through a lucite disk 4.5 cm in diameter and 2.0 cm thick. The rear half of this hole was squared off on the bottom to accommodate a very small stirring bar. Four holes were then drilled from the perimeter of the disk to the central hole. Approximate placement of these holes is shown in Figure 1. A length of 22-gauge platinum wire was pressfitted through to a/4 in. extended into the center cavity. one hole until ca. This auxiliary electrode was curved around the cavity perimeter to avoid interference with the light beam and to provide reasonable current distribution to the OTE. A glass tube with a piece of 22-gauge platinum wire sealed in the end was fitted into a second hole to serve as a salt bridge. (Small cracks which formed at the glass-platinum interface upon cooling maintained solution-solution contact.) The other end of this tube was enlarged to accept a saturated calomel reference electrode. The remaining two holes were fitted with Hamilton valves (one-way Kel-F bodies) for solution transfer. A circular quartz disk was epoxied to the rear face of the cell body. The front face was grooved to accept an O-ring (0.95-cm i.d.) which formed the seal between the lucite body and the OTE. A thin sheet of copper foil sandwiched between the OTE and the cell body provided electrical contact with the conductive film on the OTE. An aluminum disk which could be secured by screws to the lucite body enabled tight compression of the OTE-foil-0-ring-lucite sandwich. It is important that the OTE be tightly pressed to the copper foil in order to minimize resistance. Excess resistance caused negative intercepts for A-t1I2plots. Slight compression of the O-ring gave good definition to the surface area of the OTE and prevented the solution from seeping under the O-ring. The electrode area characteristic of this cell was 0.58 cm2. Electrochemical and optical instrumentation for potential step with concurrent spectral monitoring through an OTE have been reported (2, 3). For the chronoabsorptometric determinations of molar absorptivities in Table I, the duration of the cathodic potential step was one second, and the resulting optical change was traced on a high speed oscillographic strip chart recorder. Accurate measurement of the small absorbance changes attributed to ionic surface excesses was accomplished by repetitive averaging of the optical signal with a PAR TDH-9 Waveform Eductor, During such an experiment, the electrode potential was stepped from an initial value (2) N. Winograd, H. N. Blount, and T. Kuwana, J . Phys. Cliem., 73,3456 (1969). (3) N. Winograd andT. Kuwana, ANAL.CHEM., 43,252 (1971).
ANALYTICAL CHEMISTRY, VOL. 44, NO. 12, OCTOBER 1972
F--I9
-L
+li‘
--f B
A
Figure 1. Cell for preparation and use of Hg-Pt OTE A . Aluminum retaining plate B. OTE C. Copper foil
D. O-ring E. Lucite body
F. Glass salt bridge for reference electrode G. Hamilton valves H. Auxiliary electrode I . Quartz disk at which no Faradaic reaction occurred to a more cathodic potential for an interval of 10 msec and then stepped back to the original value. This step was repeated every 100 msec 1800 times to give the final averaged signal. Since the Waveform Eductor displayed only the correct form of the averaged signal, the absorbance axis had to be calibrated. This calibration was accomplished by calculating the expected slope of the A-t1/2 curve for PbZ+ reduction by means of Equation 1 and a value for €Pbo of 7.2 X 1031.mol-1cm-L. The absorbance axis of the experimental curve was then fitted to give this slope for the A-t1/2 plot. Electronic iR compensation was added as necessary in order to charge the double layer in less than 0.5 msec. The wavelength for all spectral measurements was 589 nm. Reagents. Mercurous solution for the preparation of the Hg-Pt OTE was made by dissolving reagent Hg*(NO&,2H20 (J. T. Baker Chemical Company) in 0.5M acetic acid-sodium acetate buffer. Metal ions were used as the nitrate or chloride salts. Doubly distilled water was used for all solutions. Procedure. Pt OTE substrates were prepared by vapor deposition of platinum on microscope slides (4). Electrodes were selected which exhibited surface resistances of 5-10 Q/sq. Immediately prior to preparation of an Hg-Pt OTE, the Pt OTE was rinsed with 2-propanol and distilled water, blotted dry,and placed in a radio-frequency discharge (Plasma Cleaner, Harrick Scientific Corporation, Ossining N.Y.) for about five minutes. This last procedure sufficiently removed adsorbed organic materials so that a drop of water placed on the surface did not bead. The cleaned Pt OTE was then assembled in the cell which was filled with saturated Hg2(NOJ2 (Baker Analyzed Reagent)-0.3M acetate buffer, pH 4.0. The potential of the Pt OTE was then stepped to 0.00 V 6s. SCE. Some care is necessary to obtain an even deposit of mercury. Maintaining a high concentration of HgZ2+ by using a saturated solution and intermittantly stirring at open circuit to reestablish concentration-distance profiles gave best results. Continual stirring during deposition often gave uneven mercury films because of stagnant volumes in the cell near the bottom of the OTE surface. Uneven deposition was also encountered at lower HgZ2+concentrations. The amount of mercury deposited was determined by either measuring the charge for HgZ2+reduction or monitoring the absorbance change and converting the A to charge via a calibration curve (Figure 1, ref. I). The above procedure gave more evenly deposited films than were previously reported ( I ) as evidenced by linear(4)W.von Benken and T. Kuwana, ANAL. CHEM., 42,1114(1970).
L
I
I
0
l
-0.2
l
I
I
I
- 0.6
I
I
I
L
- 1.0
I
I
P O T E N T I A L , V 9 SCE
Figure 2. Current-potential curves. 0.97mM PbZf-0.1 KNO,, sweep rate 0.025 V/sec. A . ROTE, B. Hg-Pt OTE, 10 mC/cm2,60 min old C. Hg-R OTE, 4 mC/cm2,25 min old D. Hg-Pt OTE, 4 mC/cm2,26 hr old
ity of absorbance-charge plots for charges of up to 15 mC/ cm2. The slope of such a plot was found to be 0.036 absorbance unit/mC/cmz. Uneven deposition of the mercury film exhibited negative deviation in the absorbance-charge plot. After preparation of the Hg-Pt OTE, the mercurous solution was drained and the cavity rinsed thoroughly by directing a stream of distilled water through one of the valves. Solutions to be studied with the Hg-Pt OTE were placed in serum-capped flasks, and a hypodermic needle connected to a nitrogen line was inserted deep into each solution for nitrogen purging. This needle was sheathed in a Teflon tube to prevent reaction with the solution. A second needle was inserted into each cap so that the tip was above the solution. This served as a nitrogen outlet. Prior to transfer of a particular solution, the outlet needle was connected via a thin Teflon tube to the inlet valve of the electrochemical cell containing the Hg-Pt OTE and the cavity was flushed with the nitrogen. To fill the cavity with solution the outlet needle was pushed beneath the solution level, allowing the gas pressure to force solution through the tube and into the cell. The two valves were closed after the cell was full. RESULTS AND DISCUSSION
Mercury Film Thickness. During the preparation of a ’Hg-Pt OTE, it is desirable to deposit sufficient mercury to ensure optimum “mercury character” of the ensuing electrode and yet to minimize the film thickness in order to retain
ANALYTICAL CHEMISTRY, VOL. 44, NO. 12, OCTOBER 1972
1973
optical transparency. The following paragraphs pertain to an evaluation of the relative “mercury character” which corresponds to varying thicknesses of the Hg film. Since the behavior on exposed Pt was observably different than on the Hg film, the experimental probe was the cyclic voltammetry of Pb2+in0.1MKN03. Figure 2 A shows a cyclic voltammogram for Pb2+ on a Pt OTE which was cleaned in a radio frequency discharge immediately prior to its incorporation into the electrochemical cell. The cathodic peak potential came at -0.60 V cs. SCE for the reduction of Pb2+. This potential varied slightly depending upon the number of scans and the scan rate. An ill-defined wave at about -0.90 V was also observable, although its magnitude varied and was not reproducible. During the anodic sweep the electrodeposited Pb” was stripped off in three distinctive potential regions, -0.45, -0.3, and -0.1 V cs. SCE. The potentials and relative intensities of these three stripping peaks varied considerably depending upon scan rate, the number of previous scans, and the cleaning of the Pt OTE. Similar curves have been reported for Pb2+ on bulk platinum (5). Figure 2B shows the voltammogram for Pb2+ which was recorded on a Hg-Pt OTE 60 minutes after formation of the electrode by the electrodeposition of 10 mC/cm2 of mercury. This corresponds to a film thickness of ca. 150 A of mercury. In comparison with Figure 2 A , the cathodic peak potential shifted slightly anodically to -0.58 V. However, the anodic sweep was characterized by one prominent stripping peak as compared to the three peaks obtained on platinum. This well-defined cyclic voltammogram for Pb2+ is strongly suggestive that substantial mercury character had been imparted to the electrode, even with such a thin coverage of mercury. However, the potential of the reduction peak is cathodic of the reported value on pure mercury. The polarographic half-wave potential for Pb2+ in 0.1F N a N 0 3 is -0.382 V cs. SCE (6). It is quite possible that this cathodic shift is caused by the presence of considerable platinum in the very thin mercury film. The presence of an intermetallic compound, PtHg,, has been identified by X-ray diffraction at bulk platinum-mercury interfaces (7). Formation of PtHg, at the interface and amalgamation of some platinum into the overlying mercury film probably occurs simultaneously with mercury deposition. The presence of the reduction peak for Pb2+ on the Hg-Pt OTE at -0.58 V which is so close to its value on the Pt OTE is suggestive that a substantial amount of platinum is present in the mercury film as soon as one hour after electrode formation. As the ratio of platinum to mercury in the “mercury” film increases with time, the thickness of the PtHg, layer may also increase until finally the entire structure of the “mercury” film is the thermodynamically stable PtHg,. Even though the mercury film, because of its extreme thinness, may be considerably contaminated with platinum, the desirable mercury characteristic of a single, well-defined stripping peak for lead is retained as evidence in Figure 2B. Similar cyclics could be routinely obtained on electrodes for as long as three to five days after deposition of the mercury a m . Although the Pt-Hg film may be changing in its ratio of Pt to Hg during this lifetime, the resulting electrode is evidently quite stable. Such electrodes
have been used satisfactorily for spectroelectrochemical measurements for as many as five days. The minimum amount of mercury which could be deposited to convert entirely from “Pt character” to this “Hg-Pt character” was estimated by recording cyclic voltammograms for Pb2+on electrodes with varying amounts of mercury deposited. Criteria for “Hg-Pt character” was a sharp anodic stripping peak at -0.40 V with no observable anodic peaks between +0.1 and -0.25 V which were attributed to “Pt character”. The minimum amount of mercury deposition necessary to achieve this condition was approximately 10 mC/ cm2, which corresponds to a film thickness of about 150 A for uniform coverage. During the preparation of some electrodes, random irregularities in the Pt film substrate or in the electrodepositing procedure caused the formation of accumulations of relatively pure mercury on parts of the electrode. These thicker spots of mercury gave rise to cyclic voltammograms which exhibited partial character of PbZ+ reduction on pure mercury. An example of this behavior is shown in Figure 2C. This voltammogram was obtained on a Hg-Pt OTE (4 mC/cm2 of mercury) 25 minutes after its formation. By comparison with the cathodic sweep in Figure 2B, an additional cathodic wave appeared at -0.47 V us. SCE with its anodic analog at -0.32 V. The potential of this additional wave is close to that reported on pure mercury and is attributed to the accumulation of a spot of relatively pure mercury on the OTE. The remainder of the electrode surface is the thin mercury film containing dissolved platinum which gives rise to the more cathodic set of peaks which are also found in Figure 2B. The size of the more anodic wave could be enhanced by partially covering the Hg-Pt OTE with a drop of mercury and recording a cyclic voltammogram for PbZ+on this composite surface. This resulting increase in size of the first wave (at the expense of the second wave) confirms the suspected origin of the peak at -0.47 V. The size of this peak varied from electrode to electrode and was frequently completely absent. Allowing a Hg-Pt OTE which exhibited this “mercury peak” to age while in contact with a deoxygenated Pb2+ solution produced a noticeable change in the nature of the electrode surface as evidenced by the cyclic voltammetry of Pb2+. Figure 2 0 shows a votammogram which was recorded on the same electrode as in Figure 2C, but on a fresh solution 27 hours later. The cathodic peak at -0.48 V has disappeared and the anodic peak at -0.32 V is barely discernible. The disappearance of this couple is attributed to the accumulation of platinum in what was previously a relatively pure mercury spot as a result of slow amalgamation of the underlying platinum substrate. Chronoabsorptometry. The reduction of a metal ion at the Hg-Pt OTE can be optically followed by chronoabsorptometry. Here the potential of the OTE is stepped to a region where the rate of conversion of M”+ to Mo is controlled by diffusion. The change in transmittance of the OTE resulting from accumulation of Mo in the thin film of mercury is monitored by passing light through the OTE in what has been referred to as a transmission mode experiment. This change in absorbance as a function of time is mathematically described by Equation 1
( 5 ) E. Schmidt, P. Moser, and W. Riesen, Helc. Chim. Acta, 46,
2285 (1963). (6) J. Heyrovsky and J. Kuta, “Principles of Polarography,” Academic Press, New York, N.Y., 1966. (7) G. D. Robbins and C. G. Enke, J . Elecfroanal. Chem., 23, 343 (1969). 1974
where A b p is the observed change in absorbance, ESP is the molar absorptivity of the metal dissolved in the mercury film, C l I n + is the concentration of the metal ion in the solution in moles/l, Dhp+is the metal ion diffusion coefficient in solution
ANALYTICAL CHEMISTRY, VOL. 44, NO. 12, OCTOBER 1972
~
in cm*/sec, and t is the time in seconds. It is important to realize that Equation 1 is valid regardless of whether the absorbing product remains in solution and diffuses away from the electrode, diffuses into the thin film of mercury, or adsorbs or deposits at the electrode-solution interface. The chronoabsorptometry experiment with a Hg-Pt OTE offers a unique means of determining the molar absorptivity, EMO, of a metal dissolved in a thin solution of “mercury” on platinum. It is not necessary to know the thickness of the mercury film nor the actual concentration of Mo dissolved in the film (as would be the case for a conventional Beer’s law determination), because the method is based on the rate of accumulation of Mo in the film as described by the diffusion Equation 1. Since a plot of AM^ us. t l ’ * is a straight line, the value for EMO can be extracted from the slope of the line, C M ~and + D M ~being + determined independently. Molar absorptivities for several metals were determined from chronoabsorptometry experiments on the respective metal ions. From the slopes of plots of A us. t l ’ * , values for EM^ D M n + “ 2 were obtained. Values for EMO were then calculated using known values for the diffusion coefficients (6) of the metal ions in the same electrolyte. Molar absorptivities for Pb, Cd, T1, and Zn as determined in this manner are shown in Table I. Although most metals behave as neutral density filters, absorbing light at all visible energies, some dependence of E M O on wavelength does occur. For the purpose of comparison with other determinations, the values here were all measured at 589 nm. Some dependence of the value of EM^ on the age of the Hg-Pt OTE was found. The results reported here are for determinations on electrodes which were less than five hours old. Cyclic voltammetry on solutions of the different metals being examined showed no evidence of intermetallic compound formation between the metal being deposited and platinum dissolved in the mercury film except for the case of zinc. This is the only one of these metals which has been reported to form a compound with platinum (8). An alternative way of making this measurement involves the simultaneous recording of the Faradaic charge used to reduce Mn+to M”. The relationship between charge and time for chronocoulometry (9) is described by Equation 2
where Q is charge in coulombs per cm2, concentration is in moles/ml, n is the number of electrons added per metal ion, and F is Faraday’s number. Division of Equation 1 by Equation 2 shows that the ratio of the slope of the A-tl’z line to that of the Q-t”z line should be a constant as shown by Equation 3 slope of A-t”2 slope of Q-t l’ 2
E M O (1000)
nF
~~
~
Table I. Molar Absorptivities of Metals. Chronoabsorptometry Metal (Hg-Pt OTE’) Reflectivity Pb 7 . 2 x 103 5 . 8 9 x 103 Cd 6.05 x 103 8 X los T1 Zn Hg
7 . 0 x 103 4 x 103 3 . 5 x 103~
N.A. 3.96 X IOa 6 . 5 5 x 103
a
X = 5890 8; 1. mol-’ cm-l.
c
Determined by absorbance-chargeratio method.
* Hg-Pt OTE less than 5 hours old.
exceedingly thin, such constants have been traditionally obtained from the changes which light undergoes on reflection from a metal. Since partial penetration of light into a metal occurs during reflection, information about absorption constants can be obtained. Molar absorptivities for pure metals can be calculated from these reported optical constants, n, the index of refraction and K , the attenuation index. The attenuation index and index of refraction are related to the absorption coefficient, a,and Beer’s law absorptivity, a, by the following equation (ref. 11,p 1811). 4TnK
a = (1nlO)aC = A0
(4)
where Xo is the wavelength of light used in L.’UCUO expressed in centimeters and C is the concentration of the pure metal in moles/liter. Since the concentration is expressed in moles/ liter, this absorptivity, a, is the same as the molar absorptivity, eMO, in Equation 1. Molar absorptivities calculated from values of nK at 5890 A via Equation 4 are tabulated in Table I for comparison with the chronoabsorptometricallydetermined molar absorptivities. The molar absorptivities calculated from the optical constants for the pure metals are comparable to the molar absorptivities determined by chronoabsorptometry for the metals dissolved in the thin film of the Hg-Pt OTE. Adsorption at the Hg-Pt OTE. A step change in the potential applied to the Hg-Pt OTE produced measurable optical perturbations which are attributed to the attendant change in the amount of ionic surface excess at the electrode/ solution interface. Two types of adsorbing species were investigated: the nonelectroactive anions, Nos- and Br-; and the electroactive anion, PbBrd2-. The equation for chronoabsorptometry can be extended to account for the influence of the change in ionic surface excess on the A-t curve as shown in Equation 5
(3)
The factor of 1000 comes from converting the concentrations to the same units. Since n and F are constants, EMO can be determined from the ratio. The value for E H was ~ ~ obtained in this manner and is listed in Table I. The optical constants for most metals have been reported in the literature (IO). Because a specimen of metal which transmits any appreciable fraction of incident light must be (8) E. Barendrecht in “Electroanalytical Chemistry,” Vol. 11, A. J. Bard, Ed., Marcel Dekker, New York, N. Y., 1967, Chap. 2. (9) F. C. Anson, ANAL.CHEM., 38, 54 (1966). (10) M. Born and E. Wolf, “Principles of Optics,” Pergamon Press, New York, N. Y., 1965, p 621.
Three factors contributing to the A-t curve can be identified. (1) The t”2-dependent increase in absorbance arising from the diffusion-controlled conversion of non-absorbing Mn+ to absorbing M”. The terms in this portion have been previously identified. (2) A step change increase in absorbance caused by the immediate reduction of any surface excess of Mn+.The magnitude of the absorbance change is proportional to the amount of surface excess r h p + , moles/cm2, and the molar absorptivity, EMO of the reduction product which is the species being detected optically. (Although r is ex(11) W. N. Hansen, T. Kuwana, and R. A . Osteryoung, ANAL. CHEM., 38,1810(1966).
ANALYTICAL CHEMISTRY, VOL. 44, NO. 12, OCTOBER 1972
1975
0 0 0 4. -B .. ._
*
0.002
0
-
t5
0 t,
10
msec
Figure 3. Signal averaged absorbance-time curves. Hg-Pt OTE, 15 mC/cm2,potential step -0.30 to -0.87 V us. SCE. 1800 repetitions A . 0.598 m M PbBra2--1.0MNaBr B. 1.OMNaBr C. 0.597 mM Pb2+-l.0M KNOa D. 1.OMKNOs
pressed as moles/cm2, conversion to a “three-dimensional” concentration is necessary in order to use it in conjunction with eb10 whose units are 1. mol-lcm-’. Consequently, Far”+i s multiplied by 1000 cm3/l.) (3) A perturbation in the optical constants of the thin-film electrode and of the adjacent solution caused by changes in the ionic surface excess. This perturbation is identified as A n K . It is apparent that a plot of A us. t1’z is a straight line, the slope of which is proportional to the diffusion-controlled process. The absorbance interA,,, which is a measure of the cept is equal to 1000 EhlOI’yn+ effects of changes in surface concentrations of the species involved. First, consideration is given to the origin and magnitude of An*. Small spectral changes (ca. absorbance unit and less) have been observed by reflection spectroscopy during variations of applied potential to electrodes in contact with solution which contains only supporting electrolyte (12-14). Quantitative correlations have been attempted to explain the potential-dependence of these spectral changes. In the case of the thin gold film OTE, the optical effect is apparently caused primarily by a perturbation in the optical constants of the gold film (13). Both changes in the optical constants of the thin film as well as in the refractive index of the ionic double layer in the adjacent solution have been correlated to changes in absorbance as a function of applied potential for tin oxide electrodes (14). Although the above-mentioned reports deal with reflection spectroscopy, analogous potential-dependent optical changes were observed at the Hg-Pt OTE with transmission spectroscopy. Indeed, these perturbations are sufficiently large to warrant inclusion as a separate term, A,,, in Equation 5. Here A,, represents the composite absorbance change caused by a perturbation in the optical constants, n or K , of the electrode film and of the ionic double layer in the adjacent solution. Magnitudes of A,, were investigated for aqueous solutions of K N 0 3 and NaBr. Curve D in Figure 3 shows the result-
+
(12) T. Takarnura, K. Takamura, and E. Yeager, J . Electroanal. Chem., 29,279 (1971). (13) W. N. Hansen and A. Prostak,Phys. Rev., 174,500(1968). (14) N. Winograd and T. Kuwana, J. Electroanal. Chem., 23, 333 (1969). 1976
ing A-t response for a Hg-Pt OTE in contact with 1.OM K N 0 3 during a 10-msec potential step from -0.30 V to -0.87 V us. SCE. The absorbance rapidly increased ( t < 1 msec) to a constant value of 0.0006 absorbance unit during the step to -0.87 V. Switching the potential back to the initial value of -0.30 V resulted in a return to the original absorbance, This increase in absorbance is attributed to the difference in relative surface excesses of NO3- at the two potentials. Electrochemical studies have shown that the amount of adsorbed NO3-, which is considerable on Hg at -0.30 V, steadily diminishes with increasingly cathodic potentials (15, 16). Thus, stepping the potential to -0.87 V causes substantial diminution of the relative surface excess because of coulombic repulsion of the negatively charged ions by the negatively charged electrode. This rapid desorption of anions from the electrode surface certainly contributes, either directly or indirectly, to the observed absorbance change. The direct contribution is attributed to the change in refractive index of the solution immediately adjacent to the interface as a result of the rearrangement of the double layer upon desorption of the excess of anions in the inner Helmholtz layer. The indirect contributions are the concomitant perturbation on the optical constants of the Hg-Pt film by the change in the number of free electrons (or charge) induced in the electrode at the interface by the adsorbed species and any physical change in the Hg-Pt film due to surface tension differences. Since the light beam traverses both the solution and the entire electrode film, the relative contributions of these two effects are not resolvable by a simple transmission experiment. The size of the absorbance change, A,,, was related to the magnitude of the change in surface excess which accompanied the potential step. Since almost complete desorption occurred at -0.87 V, this change in ionic surface excess was dependent upon the amount of adsorbed material at -0.30 V. Lowering the solution concentration of NOo- which diminished the surface excess of NOs- according to its adsorption isotherm resulted in a decrease in the value for Ant. For example, a perturbation of only 0.0001 a.u. was recorded for 0.1M KNOs as compared to 0.0006 a.u. for 1.0M KN03. Larger values of A,, were obtained for solutions of NaBr than for solutions of KN03. For instance, 1.0M NaBr gave a change of 0.0025 a.u. (curve B, Figure 3). This is attributed to the more extensive adsorption of Br- than NOaon mercury at -0.30 V (15, 16). Potential steps to less cathodic values caused smaller absorbance changes since the extent of desorption was less at the less negative potentials. The results obtained on the Hg-Pt OTE by transmission spectroscopy are consistent with those reported for Br- adsorption on gold electrodes as studied by reflectance (12). The increase in reflectance with anion desorption reported on gold is analogous to the decrease in transmittance observed here for anion desorption on mercury. Although it is not expressed in Equation 5 , some timedependence can exist in the behavior of A n K . In the cases described above, the potential was stepped so that NO3- or Br- was desorbed. This desorption was rapid as evidenced by the immediate increase in absorbance to a limiting value. However, in the situation in which the potential is stepped so that the amount of adsorbed material increases, a measurable time dependence in the A,, term would be expected. The (15) P. Delahay, “Double Layer and Electrode Kinetics,” Inter-
science, New York, N . Y., 1965. (16) D. C . Grahame and B. A, Soderberg, J. Chem. PhYs., 22, 449 (1954).
ANALYTICAL CHEMISTRY, VOL. 44, NO. 12, OCTOBER 1972
rate of absorbance change here is dependent upon the rate of diffusion of ions to the electrode surface until an equilibrium coverage for the new potential is attained. This behavior was observed experimentally for adsorption of NOs- and Br-. It is also apparent that A,, can be positive or negative, depending upon the initial and final potentials. Consider now the situation in which the electroactive species, Pb%+,is added to the solution. If no surface excess of this electroactive species exists, chronoabsorptometry for the reduction of Pb2+ should yield an A - t 1 / 2plot which is a straight line with an intercept equal to AnX. Since no adsorption on mercury of Pb2f in 1.OMKNO, has been reported, this system was examined. The chronoabsorptometric absorbance-time curve for reduction of 0.597mM Pb2+1.OM KNOBis shown as curve C in Figure 3. Curve C in Figure 4 is the resulting plot of A us t 1 i 2 . The intercept is the same as A,, determined on 1.OM KN03 alone, curve D. Thus, it is apparent that at this low concentration of electroactive species, a simple correction for A,, can be applied. Apparently, the discharge of Pb2+ from the electrolyte and acceptance as PbG in the film does not detectably alter the factors contributing to AnK. If a surface excess of the electroactive species, Mn+, exists, the intercept should increase by an amount equal to 1000 e M O r W + , assuming that the excess does not alter AnK. This situation was examined with an aqueous solution of PbBrk2in 1.OM NaBr. A competition now exists between the PbBr42-, which is reported to adsorb strongly on mercury (17), and Br- for adsorption sites on the surface. Results of chronoabsorptometry of 0.598mM PbBrr2--1 .OM NaBr are shown by curves A in Figures 3 and 4. The intercept corresponding to curve A in Figure 4 is greater than the intercept for curve 3 by 0.00013 a.u. This difference represents the contribution to absorbance caused by the specific adsorption of PbBrd2-,assuming that the presence of adsorbed PbBr42- does not influence the value of A,, which was determined on bromide solution alone. The surface excess of PbBrd2-was calculated by means of the relationship AinteroeptA,, = 1000 ePborPbBr,l- to be 2 X mole/cm2 where EPbO = 7.2 X l o 3 1. mol-l cm-I. This value agrees with that determined by chronocoulometry simultaneously on the same electrode. This agreement suggests that in this case A,, is the same with and without PbBr42-. However, this question should be examined further since some interdependence might be expected. The value of rPbBrr2- obtained on the Hg-Pt OTE is less than that determined on a fresh mercury drop, ca. 9 X 10-lGmole/cm2(17). This is attributed either to the influence of dissolved Pt on the Hg-Pt OTE or to partial contamination of the surface by other absorbable species. The latter explanation is quite reasonable, since a time interval of at least 20 minutes occurred between electrode formation and the chronoabsorptometry experiment whereas experiments on pure mercury involved freshly formed mercury drops. The exact magnitudes of the intercepts varied somewhat depending on the thickness and age of the mercury film. Signal averaging of optical perturbations from repetitive potential steps was essential for the accurate observation of the small absorbance changes caused by these surface phenomena. The precision inherent in this signal averaging method is excellent. For example, the intercept of curve A in Figure 4 was calculated by the method of least squares to be 0.00262 absorbance unit with a standard deviation of only i 0.00001 absorbance unit. ..__
(17) R.W. Murray and D. J. Gross, ANAL.CHEM., 38, 392 (1966).
0.004
4:
r
I
t
0.002
0.001
D
01 0
I
I
I
0.02
0.04
0.06
I
0.08
I 0.10
tu”, secl’*
Figure 4. A us. t 1 / 2plots for chronoabsorptometry A. B. C. D.
0.598 mM PbBr42--1.0M NaBr 1.OMNaBr 0.597 m M Pb2+-1.0MKNOB LOMKNOI
In view of the results reported here, optical transmission through a transparent electrode appears promising as a spectral means for quantitatively examining surface excesses of ionic species at an electrode-solution interface. Stripping Analysis. Mercury electrodes have been used extensively for the analysis of low concentrations (usually less than l O + M ) of metal ions in aqueous solution by stripping voltammetry (8). In this analysis, the potential of a mercury electrode is maintained such that the various metal ions are electrodeposited from a strirred solution into the mercury to form an amalgam. This procedure effectively converts a portion of the metal ions in the dilute solution into a more concentrated solution of metal atoms in the mercury electrode. A specific time interval is selected for this concentration step5 to 60 minutes depending upon the concentration level of the metal ions being analyzed. During this reduction, the amalgamated metal atoms diffuse into the Hg electrode under the influence of a concentration gradient. After this electrolysis period, the potential of the electrode is scanned anodically, and the accumulated metal atoms are stripped out of the electrode. The resulting current-voltage curve exhibits current stripping peaks at potentials characteristic of the different metals. The magnitude of a stripping peak is proportional to the concentration of the metal ion being analyzed. A commonly used electrode for this procedure is a hanging mercury drop. The rate of the current decay which immediately follows the appearance of a peak is determined by the rate of diffusion of amalgamated MGto the mercuryaqueous interface. Stripping voltammograms from a mercury drop are characteristically “drawn out” after the peak current is passed because some metal atoms must diffuse from deep within the drop to the interface for oxidation during the stripping step. In an analysis of this type, the maximum inherent sensitivity is approached as the time interval for expulsion (oxidation) of the amalgamated species is minimized. This results in maximum peak heights. Thus, sensitivity can be improved by restricting the depth to which the amalgamated metal atoms can diffuse into the mercury. This end can be achieved by maximizing the surface area of the mercury-solution interface relative to the volume of mercury. Since a spherical mercury drop offers the lowest
ANALYTICAL CHEMISTRY, VOL. 44, NO. 12, OCTOBER 1972
1977
possible ratio of surface area to volume, electrodes consisting of mercury films deposited on a planar platinum substrate have been used with improved sensitivity. In these films, which are typically 1-10 pm thick, the distance from the point innermost in the electrode is substantially reduced and the stripping curves are sharper. Mathematical evaluation of the diffusion equations show that sensitivity improves as the thickness of the film decreases (8, 18). The Hg-Pt OTE represents the extreme evolution in mercury film electrodes. Here the thickness of the mercury film is the minimum amount which is necessary to convert from “platinum character” which gives multiple stripping peaks (Figure 2 4 to “mercury character” which gives a well-defined stripping peak for lead (Figure 2B). This requires as little as 10 mC/cm2 of mercury-a film thickness of ca. 0.015 pm. The time interval for diffusion from the depth of the film to the mercury-solution interface is considerably reduced for this electrode compared to the thicker mercury film electrodes, and stripping curves are consequently sharper. A feature of the Hg-Pt OTE which warrants comment is the total recovery of PbZ+achieved during the stripping procedure at a normal scan rate. Comparison of the charge accumulated during reduction of Pb2+with the charge used for the anodic
stripping indicated complete recovery of lead from the mercury film in cyclic voltammetry with scan rates of up to 0.20 V/sec on 0.97mMPb*+-0.1MKN03. Initial experiments show that the Hg-Pt OTE does exhibit enhanced sensitivity for the stripping analysis of metal ions compared to the commonly used thicker mercury film electrodes. The electrode may prove useful for the analysis of trace amounts of Pb2+ and other metal ions which do not form intermetallic compounds with the platinum which is also dissolved in the mercury film. ACKNOWLEDGMENT
The authors appreciate the helpful suggestions of F.M. Hawkridge concerning the cell design. RECEIVED for review March 15, 1972. Accepted May 30, 1972. This paper was presented in part at the 164th National Meeting, American Chemical Society, New York, N.Y., August 1972. The authors gratefully acknowledge the financial support provided by the National Science Foundation Grants GP31236 (OSU) and GP9306 (CWRU) and the U S . Army Electronics Command, Contract DA AR07-68-C-0278. The authors also acknowledge the Department of Chemistry, Case Western Reserve University, Cleveland, Ohio, where this work was begun.
(18) W. T. de Vries, J. Elecrroanul. Cliem., 9,448 (1965).
Simultaneous Electrochemical and Photometric Monitoring of Intermediates Generated by Flash Photolysis J. I. H. Patterson’ and S. P . Perone Department of Chemistry, Purdue Uniaersity, Lufayette Ind. 47907 This paper describes instrumentation which allows the simultaneous electrochemical and photometric monitoring of transient intermediates generated by flash photolysis. An interface for an on-line digital computer is also described. This system was used to follow the second-order reaction of the ketyl-radical and ketyl-radical-ion generated by the flash photolysis of benzophenone. Because of the simultaneous measurements, it was possible to calculate the value of the molar absorptivity of the ketyl-radical-ion. It was found to be (9.5 & 1.3) x 103M-1 cm-l.
(1) S. P. Perone and J. R. Birk. ANAL.CHEM., 38, 1589 (1966). (2) J. R. Birk and S. P. Perone, ibid.,40, 496 (1968).
by flash photolysis (6). Both of these methods have distinct advantages and limitations. The electrochemical measurements have the disadvantage that they perturb the measured system as a result of electrolysis of electrochemically active species at the electrode. Because of theoretical limitations, it is possible to obtain meaningful kinetic data by continuous monitoring of electrolysis currents ( 2 ) only at times less than the half-life of the reaction if the rate of reaction is other than first-order. Using the time-delay method ( I ) , it is possible to obtain concentration measurements over a much larger time scale; but it is necessary to perform a separate experiment for each datum desired. Moreover, the shortest time at which a measurement can be made is determined by the nature of the solution under study. For highly conducting solutions, no severe limitation is imposed; however, as the conductivity decreases, the time at which the first meaningful measurement can be made increases because of the increased time necessary to charge the double layer. In the limiting case of non-conducting solutions no electrochemical measurement can be made. Electrochemical measurements, however, have the advantage of high and similar sensitivity for reactants and intermediates. The sensitivity is determined mainly by diffusion rates which can easily be deter-
(3) G. L. Kirschner and S. P. Perone, ibid.,44,443 (1972). (4) H. E. Stapelfeldt and S. P. Perone, ibid.,41, 628 (1969). (5) R. A. Jamieson and S . P. Pelone, J . Phys. Cliern., 76, 830 (1972).
(6) G. Porter, “Rates and Mechanisms of Reactions,” S. L. Friess. E. S. Lewis, and A. Weissberger, Ed., 2nd ed., Wiley, New York, N.Y., 1963, Chapter 19.
POTENTIOSTATIC CHRONOAMPEROMETRY has been employed as a method for monitoring intermediates generated by flash photolysis (I-j), and has been applied successfully to the determination of several photochemical mechanisms ( 4 , 5 ) . More commonly, photometric methods introduced by Norrish and Porter have been used to monitor reactions generated Present address. The Milton S. Hershey Medical Center, The Pennsylvania State University. Department of Biological Chemistry. Hershey, Pa. 17033.
1978
ANALYTICAL CHEMISTRY, VOL. 44, NO. 12, OCTOBER 1972
