Anal. Chem. 1907, 59, 2334-2339
2334
T a b l e 11. Lithium Assay in S e r u m Samples by E l e c t r o d e I n c o r p o r a t i n g PVC-2 M e m b r a n e C o m b i n e d w i t h D i a l y s i s Membrane
Li+ concentration, mM serum sample
actual
founda
coeff of var, 7%
1 2 3 4 5
0.32
0.35 0.38 0.83 1.08 1.61 3.08
4
6
0.40 0.80 1.00 1.50 3.00
6 4 5 7 3
n M e a n of four repeat measurements. (Measurement o f a set of the six samples was repeated.)
assay is rather reliable. In conclusion, the Li+-selective electrode incorporating the PVC-2 membrane in combination with the dialysis membrane seems useful for serum Li+ assay in spite of its somewhat prolonged response time. Registry No. 1, 91539-72-9; 2, 106868-21-7; 3, 106868-32-0; NPOE, 37682-29-4; NPPE, 2216-12-8; FPNPE, 93974-08-4; DOS, 122-62-3;TEHP, 78-42-2;TOPO, 78-50-2;DOPP, 1754-47-8;PVC, 9002-86-2; KTpClPB, 14680-77-4; Li, 7439-93-2.
LITERATURE CITED (1) Guggi, M.; Fiedler, U.; Pretsch, E.; Simon, W. Anal. Lett. 1975, 8 , 857-866. (2) Zhukov, A. F.; Erne, D.; Ammann, D.; Guggi, M.; Pretsch, E.; Simon, W. Anal. Chim. Acta 1981. 131, 117-122.
(3) Metzger, E.; Ammann, D.; Asper, R.; Simon, W. Anal. Chem. 1988, 58, 132-135. (4) Gadzekpo, V. P. Y.; Hungerford, J. M.; Kadry, A. M.; Ibrahim, Y. A,; Christian, G. D. Anal. Chem. 1985, 57, 493-495. (5) Gadzekpo, V. P. Y.; Hungerford, J. M.; Kadry, A. M.; Ibrahim, Y . A,; Xie. R. Y.; Christian, G. D. Anal. Chem. 1988, 58, 1948-1953. (6) Gadzekpo, V. P. Y.; Moody, G. J.; Thomas, J. D. R. Analyst(London) 1985, 170. 1381-1385. (7) Aalmo, K. M.; Krane, J. Acta Chem. Scand., Ser. A 1982, A36, 227-234. (8) Olsher, U. J . A m . Chem. Soc. 1982, 104, 4006-4007. (9) Gadzekpo. V. P. Y.; Christian, G. D. Anal. Lett. 1983, 16, 1371- 1380. (IO) Kitazawa, S. Kimura, K.; Yano, H.; Shono, T. J . Am. Chem. Soc. 1904, 106, 6978-6983. (11) Kitazawa, S.;Kimura, K.; Yano, H.; Shono, T. Analyst (London) 1985, 170, 295-299. (12) Kimura, K.; Yano, H.; Kitazawa, S.;Shono, T. J . Chem. Soc., Perkin Trans. 2 1986, 1945-1951. (13) Xie, R. Y.; Christian, G. D. Analyst (London) 1987, 112, 61-64. (14) Gadzekpo, V. P. Y.: Moody, G. J.; Thomas, J. D. R. Anal. R o c . (London) 1986, 2 3 , 62-64. (15) Allen, C. F. H.; Gates, J. W. Organic Synthesis; Wiley: New York, 1955; Collect Vol. 111, pp 140-141. 16) Brewster, R. Q.; Groening, T. G. Organic Synthesis: Wiley: New York, 1943; Collect Vol. 11, pp 445-446. 17) Ryba, 0.; Petranek, J. Collect. Czech. Chem. Commun. 1984, 4 9 , 2371-2375. 18) Cassaretto, F. P.; McLafferty, J. J.; Moor, C. E. Anal. Chim. Acta 1965, 32, 376-380. 19) "Recommendations for Nomenclature of Ion-Selective Electrodes" Pure Appl. Chem. 1976, 4 8 , 129-132. (20) Imato, T.; Katahira, M.; Ishibashi, N. Anal. Chlm. Acta 1984, 765, 285-289. (21) Gadzekpo, V. P. Y.: Moody, G. J.; Thomas, J. D. R. Analyst (London) 1986, 711, 567-570.
RECEIVED for review March 2, 1987. Accepted June 1, 1987.
Direct Electron Transfer Reactions of Cytochrome c at Silver Electrodes David E. Reed and Fred M. Hawkridge* Department of Chemistry, Virginia Commonwealth University, Richmond, Virginia 23284
The dlrect, heterogeneous electron transfer reactions between horse heart cytochrome c and silver electrodes have been shown to be stable for periods of time exceedlng 12 h. The kinetics of these reactions are quasi-reversible at polished silver surfaces and at ekctrochemically roughened silver surfaces. These results demonstrate that neither electrode surface modification nor the inclusion of mediators is necessary to study the electron transfer reactions of cytochrome c at silver electrodes.
Direct electron transfer reactions of cytochrome c a t bare silver electrodes have been reported to proceed by slow, irreversible heterogeneous kinetics ( I , 2). The slow ra'tes of electron transfer have been attributed to electrode fouling by irreversible adsorption of protein ( 3 , 4 ) . Strong irreversible adsorption of cytochrome c has been indicated by surfaceenhanced Raman scattering (SERS) measurements a t this surface ( I , & 5-7). In several cases, prior modification of the silver surface ( 2 , 5 )by the adsorption of surface promoters has facilitated electron transfer reactions between cytochrome c and silver electrodes ( I , 2). In our studies, quasi-reversible electron transfer kinetics have been obtained for the reaction 0003-2700/87/0359-2334$0 1.50/0
of cytochrome c a t silver electrodes without the presence of mediators, modifiers, or surface promoters. This was accomplished by using cytochrome c samples that had been chromatographically purified, but not lyophilized, prior to the voltammetric study. In a previous report (8) attention had been drawn to the effects of sample purity on the voltammetric response of cytochrome c at solid electrodes. In that report it was clear that impurities found in high-quality commercial samples of cytochrome c (9) could affect the heterogeneous electron transfer kinetics a t electrodes. In a more recent report ( I O ) , it was indicated that the process of lyophilization (this is how the protein is usually stored after purification) denatured a small amount of the ferricytochrome c. This denatured form of cytochrome c does not return to its native state when reintroduced into aqueous media as indicated by various chromatographic bands with mobilities both slower and faster than the native protein. Therefore, these recent results have led us to consider the importance of sample purification as it applies to the direct heterogeneous electron transfer reactions of cytochrome c at bare silver electrodes. In this report, cyclic voltammetry (CV) ( I I , 1 2 ) , derivative cyclic voltabsorptometry (DCVA) (8, 13-15),and single potential step chronoabsorptometry (SPS/CA) (16,17) have been 1987 American Chemical Society
ANALYTICAL CHEMISTRY, VOL. 59, NO. 19, OCTOBER 1, 1987
used t o study the heterogeneous electron transfer kinetic parameters for the reaction of cytochrome c at bare silver electrodes. These studies were conducted on smooth silver electrodes with varying concentrations of chromatographically purified cytochrome c solutions in both 0.05 M Na2S04 electrolyte and 0.05 M Tris/cacodylic acid buffer, pH 7.0. The former electrolyte was chosen because it is often used in SERS experiments at this metal surface. For a rough silver electrode surface, DCVA experiments indicated a less reversible behavior in comparison with a smooth silver electrode surface under the same experimental conditions. In all experiments performed with samples of cytochrome c that had been chromatographically purified just prior t o use, reproducible quasi-reversible voltammetric responses were obtained for periods of time exceeding 12 h. By contrast, if a purified sample of cytochrome c is then lyophilized, impersistent and irreversible DCVA responses are obtained. The results from this study clearly show that direct electron transfer between cytochrome c and silver electrodes occurs at quasi-reversible rates when samples are highly purified and never lyophilized. Samples that are chromatographically purified and then lyophilized exhibit irreversible electron transfer kinetics. T h e electrode surface is apparently fouled by strong irreversible adsorption of a denatured form of cytochrome c t h a t is produced by the lyophilization process. Silver is one of a growing list of electrodes that shows reproducible, quasi-reversible rates of heterogeneous electron transfer without the need for surface modification or mediators (4).
EXPERIMENTAL SECTION Materials a n d Solution Preparation. Horse heart cytochrome c (type VI from Sigma Chemical Co.) was chromatographically purified as previously described (8, 9). Phosphate ions were removed from the central portion of the purified cytochrome c sample by chromatography on a desalting column (Bio-Rad P-6DG) with water as the eluent. Solutions of cytochrome c in 0.05 M Na2S04and in 0.05 M Tris/cacodylic acid buffer pH 7.0 were prepared by adding an appropriate amount of the corresponding electrolyte to this aqueous purified cytochrome c solution. All Na2S04solutions used in this work had a measured pH of 6.3. Cytochrome c concentrations were determined with a Beckman Acta MVII spectrophotometer at 550 nm using a reduced minus oxidized difference molar absorptivity, Ac = 21 100 M-' cm-' (18). Absorbance spectra were recorded for fully reduced cytochrome c by addition of solid sodium dithionite (ACS grade) to the aqueous purified ferricytochrome c sample. Tris(hydroxymethy1)aminomethane (Sigma, Trizma Base, reagent grade) and sodium sulfate (Fisher Scientific Corp., ACS grade) were used as received. Cacodylic acid (Sigma, 98% pure) was recrystallized twice from 2-propanol. All solutions were prepared in water purified with a Milli-RO-4/Milli-Q system (Millipore Corp.) which exhibited a resistivity of 18 MQ on delivery. Samples were deoxygenated prior to use by passing water-saturated nitrogen over the surface of the solution while stirring gently overnight at 4 OC in a sealed serum bottle. Nitrogen used in each experiment was deoxygenated prior to water saturation by passage over hot copper turnings (ca. 500 "C). Polycrystalline silver foil, 1.0 mm thick (Alfa, 99.999%), was used as the working electrode. Prior to each experiment the silver electrode was polished to a mirror finish with successively finer grades of alumina: Fisher 0.3, 0.1, and 0.05 pm. The electrode was then sonicated in copious amounts of Milli-Q water. Immediately after sonication, the silver electrode was preconditioned by continuous voltammetric cycling from +0.380 to +0.100 V vs. NHE in deaerated electrolyte for ca. 30 min a t a scan rate of 20 mV/s. Background CV and SPSJCA responses were then acquired for electrolyte alone under the same experimental conditions that would later be employed for the cytochrome c sample a t that same electrode in a given electrolyte. Solutions of cytochrome c were anaerobically transferred by a glass syringe from the sealed serum bottle to the cell. The cell was kept deoxygenated after injection and during experiments by a blanket of nitrogen.
U
II
2335
a
Yb
e
f 9
i
0
k
I I
Figure 1. Spectroelectrochemical cell design: (a) bifurcated fiber-optic light guide: (b) quartz tube: (c) nitrogen blanket purge tube: (d) cell cover plate; (e) reference electrode; (f) platinum auxiliary electrode: (9) quartz disk window; (h) plastic cell body: (i) O-ring: (j) brass shim: (k) silver electrode: (I) retainer plate.
Cell Construction. The spectroelectrochemical cell, shown in Figure 1, is similar to the design of Pyun and Park (19). Specular reflectance from the silver electrode surface was obtained with a glass fiber optic bundle (bifurcated, 12 in., Catalog No. 22-0285, Ealing Optical Co.). The combined end of the parallel pair of the fiber optic bundle was inserted inside a quartz tube which had been sealed at one end by a flat quartz disk. This prevented the contamination that would result from liquid seepage between the glass fibers. The open end of the quartz tube was epoxyed to the top of the cell cover plate so that the sealed flat end of the tube was positioned ca. 0.6 mm above the silver electrode surface. This maintained a distance which ensured that semiinfinite linear diffusion conditions would exist during each experiment. Wrapping Teflon tape around the combined end of the fiber optic bundle allowed the ferruled end to fit snugly inside the quartz tube and firmly against the sealed quartz window a t the end of the tube. This arrangement made it easy to remove the fiber optic bundle from the cell for safe storage between experiments. The silver electrode was mounted on the bottom of the cell with a retainer plate. Connection to the working electrode was made by a brass shim isolated from the solution by an O-ring seal. The geometric area of the working electrode, determined by the inside diameter of the O-ring, was ca. 1.23 cm2. The cell cover plate also contained a nitrogen blanketlpurge tube, a sample injection port, a reference electrode, and a platinum auxiliary electrode. The reference electrode probe tip was positioned ca. 2 mm above the working electrode. Apparatus. Optical monitoring was performed with a single-beam spectrometer configuration (20). A 63 mm focal length lens focused the source image (quartz halogen) on one arm of the fiber optic bundle. The other arm of the fiber optic bundle was positioned directly in front of a photomultiplier tube. The electrode potential was controlled with a locally constructed potentiostat of conventional three-electrode design (21). The potentiostat and spectrometer signals were both transferred to an integrating 16-bit analog-to-digital (A/D) converter interfaced to a dedicated UNC ZM80 based microcomputer (22). The data acquisition, instrument control, and subsequent data processing were directed by the UNC microcomputer with programs written in BASIC. All calculations were performed with this system.
2336
ANALYTICAL CHEMISTRY, VOL. 59, NO. 19, OCTOBER 1, 1987
For both CV and DCVA, data were acquired at a sampling rate of ca. 1 point/mV. In both of these techniques, an analog fourth-order Butterworth active filter with a cutoff frequency of ca. 50 Hz was placed just before the integrating A/D. The rate of sampling for an SPS/CA experiment was set a t 20 points/s. Changes in the reflectivity of the metal-electrolyte interface occurred when potential step perturbations were applied to the electrode yielding an initial decrease in absorbance. McCreery et al. (23)have reported a similar effect when using an external reflection geometry a t a platinum fiber electrode. This effect has been attributed to changes in electroreflectance (24,25) in combination with the reduction of a surface oxide film when the electrode is stepped to more negative potentials. Therefore, background corrected SPS/CA results were obtained by taking the difference between SPS/CA results acquired for cytochrome c solutions and solutions containing buffer alone. These background-corrected files were then used to determine heterogeneous electron transfer kinetic parameters. For a number of reasons, the 550-nm absorbance maximum for ferrocytochrome c was monitored instead of the more robust maximum at 416 nm. Absorption by the metal surface (26),the sharp drop-off in the transmittance of the glass fiber optics used in this work, and the weaker source intensity a t 416 nm were considered. Moreover, a wider slit width can be used at the broader 550-nm absorption maximum compared with the sharp maximum at 416 nm. Special concern for the signal-to-noise ratio is belabored here because of the signal corruption that occurs upon taking a derivative of the optical signal in DCVA (8, 13-15). Although the 416-nm difference molar absorptivity, Ac = 57 000 M-' cm-' ( I @ , is almost three times larger than the difference molar absorptivity at 550 nm, the latter has been used here for the reasons given above. Electrochemical roughening of the silver electrode surface followed a previously described procedure ( I ) . It consisted of an oxidation reduction cycle (ORC) pretreatment with a double potential step waveform from -0.370 V to +0.680 V in 0.1 M Na2S04. The current passed in the oxidation step was digitally integrated using the computer and then the potential was returned to -0.370 V after ca. 25 mC/cm2 of anodization charge had passed. The cell was then rinsed and a degassed cytochrome c solution was injected into the cell with a glass syringe as described earlier in the Experimental Section. The Ag/AgCl reference electrodes were calibrated against saturated quinhydrone (Eastman) solutions of known pH (27) and found to be 0.229 (f0.005) V vs. the normal hydrogen electrode (NHE). All potentials are reported vs. the NHE and all experiments were performed a t room temperature, 22 (h2) "C. The diffusion coefficient of cytochrome c used in all digital simulations and kinetic analyses was 1.1x lo4 cm2/s (28). The formal potential (EO')used in each calculation was determined from the midpoint between the forward and reverse peak potentials taken from slow scan DCVA experiments for each experimental condition. Digital simulation algorithms for CV, DCVA, and SPS/CA responses have been previously reported (13, 15-17). Formal heterogeneous electron transfer rate constants ( k o l s , h ) were calculated from DCVA peak potential separations after the method of Nicholson (11). The electrochemical transfer coefficients (a)were determined from (Epl2- EP)values taken from the forward potential scans of DCVA experiments (29). For comparison, these kinetic parameters were also determined from SPS/CA experiments (17). The formal heterogeneous electron transfer rate constant was calculated from the intercept of a plot of log k,, vs. overpotential (0)and the electrochemical transfer coefficient was calculated from the slope of these plots.
RESULTS AND DISCUSSION Figure 2 shows first-scan DCVA responses a t 1.04 mV/s for two differently prepared ferricytochrome c samples in 0.05 M Na2S04at a smooth polycrystalline silver electrode. T h e DCVA response for chromatographically purified samples, Figure 2A, is highly reproducible and profoundly different from the response for samples that have been lyophilized after chromatographic purification, Figure 2B. The DCVA response shown in Figure 2A was reproducible for periods of time exceeding 12 h, for different samples, and for different elec-
0350
0 200
0 050
-0100
- 0 250
E ( V v s NHE) Figure 2. Derivative cyclic voltabsorptometry of cytochrome c at a smooth silver electrode in 0.05 M Na,SO,: (A) 199 pM chromatographically purified cytochrome c ; (B) 98 pM chromatographically purified and then lyophilized cytochrome c sample.
t
-
1
i
/ 02
04
01
03
E ( V v s NHE)
Figure 3. Cyclic voltammetry of a smooth silver electrode in degassed 0.05 M Na,SO, electrolyte: (a) background cyclic scan of electrolyte before adding cytochrome c (solid line);(b) background cyclic scan of electrolyte alone after adding 200 p M of purifii/lyophilized cytochrome
c , performing 10 scans, followed by rinsing cell and adding the degassed electrolyte (dashed line).
trodes. However, the DCVA responses for the purified and then lyophilized sample, as shown in Figure 2B, decayed with time leading t o no observable response after ca. 30 min. These results strongly suggest that lyophilization produces denatured form(s) of cytochrome c that strongly and irreversibly adsorb on the silver surface blocking electron transfer. This is consistent with a report (10) which stated that lyophilization of chromatographically purified ferricytochrome c gave a small amount of denatured material. New bands were evident about the native band when lyophilized samples were redissolved and subjected t o ion exchange chromatography. Especially convincing evidence for the irreversible adsorption of a denatured form of cytochrome c on a smooth silver electrode is shown in Figure 3 with background CVs at 10.4 mV/s. Figure 3a shows the CV response obtained for a freshly cleaned silver electrode that had been cycled in deaerated electrolyte for ca. 30 min. Figure 3b indicates the CV response obtained for that exact same silver electrode and electrolyte system after performing 10 cyclic scans with a lyophilized sample of cytochrome c followed by thoroughly rinsing the cell and adding the same deaerated electrolyte. The dramatic
ANALYTICAL CHEMISTRY, VOL. I
I
59,NO. 19,OCTOBER 1, 1987
2337
1
I
- 20 t 3
6
-E W V
z
a
m [L
0
ln
m
uu 04
03
02
01
04
03
02
01
6
L
10 I
tI 1
E ( V v s NHE)
E ( V v s NHE)
Figure 4. Derivative cyclic voltabsorptometry and cyclic voltammetry of 199 p M chromatographically purified cytochrome c in 0.05 M
Na,SO, at a smooth silver electrode: (A) background subtracted CV; = 1.5 X (B)DCVA. Circles indicate simulated responses for k lo3 cm/s and 01 = 0.55. Potential scan rates in mV/s are as follows: (a)
10.40;(b) 5.20;(c) 2.04;(d) 1.04.
reduction in the capacitive charging current after addition of the lyophilized cytochrome c, Figure 3b, suggests strong irreversible adsorption of a contaminating component generated by the lyophilization procedure on the electrode surface. This behavior is absent for cytochrome c samples that are studied immediately following chromatographic purification. The identity of the contaminating component, caused by the lyophilization step, is not known. Ion exchange chromatography (9) of lyophilized samples that had been previously purified shows three bands (10) that are similar to bands observed in initial purification procedures. These bands have been ascribed t o deaminated and polymeric forms of cytochrome c (9). The deamidated forms have shorter retention times while polymeric forms have very long retention times. When a sample of the polymeric form is added to a solution of purified cytochrome c, the electron transfer kinetics rapidly decay (30) while addition of the deaminated form has no observable effect.
Determination of Heterogeneous Electron Transfer Kinetic Parameters for Purified Cytochrome c. Upon the establishment of conditions t h a t provide reproducible electrochemical responses between bare silver electrodes and purified cytochrome c samples, heterogeneous electron transfer kinetic parameters were determined for this system. CV and DCVA results for chromatographically purified cytochrome c samples are shown in Figure 4. Figure 4A shows the background-subtracted CVs obtained for the protein together with simulated responses based on Butler-Volmer theory (31-33). The cathodic peak current increased linearly with the square root of scan rate in the range only up to 10 mV/s. The lack of agreement between experiments and simulated responses on the reverse anodic sweep is probably due to errors caused by shifts in the silver oxidation process. The useful limit of voltammetry due to dissolution of silver from the electrode surface is approximately 50 mV more positive than the formal potential for cytochrome c. Therefore, small changes in the dissolution current with potential for the silver electrode in the presence of cytochrome c can cause errors in background-subtracted CVs in the anodic region. Since the anodic dissolution of silver is a surface process, errors at anodic potentials are expected to be more pronounced a t higher scan rates when small changes occur between the background current obtained before and after cytochrome c is added. This effect is more evident in the CVs at the higher scan rates as shown in Figure 4A. The species selectivity and the freedom from interferences due to nonfaradaic and other faradaic reactions is a widely
0l
1
I
1
l
I
l
0
1
2
3
4
5
TIME(sec)
Figure 5. Single potential step chronoabsorptometry of cytochrome c , same sample as described in Figure 4. Circles indicate simulated cm/s and cy = 0.52. The initial responses for kofs,h = 9.1 X potential before each step was +380 mV vs. NHE. Step potentials in mV are as follows: (a) -370; (b) +155; (c) 4-181;(d) 4-215;(e) +240; (f) +260.
recognized advantage of spectroelectrochemistry (34). DCVA results that were obtained under the same experimental conditions as described in Figure 4A are shown in Figure 4B. The DCVA experimental results agreed with simulated responses a t all applied potentials. DCVA also provides larger peak responses as the scan rate is decreased, the reverse of that which is observed in CV (13). This means that for quasi-reversible systems, DCVA has an important analytical advantage over CV. Figure 4B also shows that the DCVA technique is not sensitive to the dissolution of the silver electrode surface, a t least within this particular potential excursion. With increasing applied potential (i.e., more positive) the dissolution becomes quite evident in the optical response and therefore limits the anodic potential for this particular electrode. As shown in Figure 4B the DCVA responses are symmetrical. A midpoint potential value ( E O ’ ) of +0.250 V vs. NHE was determined by using the formula Eo’= (Epa+ Ep,)/2. This potential agrees with the literature value of +0.260 V vs. NHE (35). The formal heterogeneous electron transfer rate constant determined from the scan rate dependence of the DCVA responses, e.g., Figure 4B, was calculated cm/s from the method of Nicholson to be 1.5 (f0.4) X (11). The electrochemical transfer coefficient (cy) was calculated to be 0.55 (*0.05) using the method of Matsuda and Ayabe (29) from the forward potential scans of the DCVA responses. The reaction of cytochrome c a t polycrystalline silver electrodes is defined as quasi-reversible based on the criteria of Matsuda and Ayabe (29) using a diffusion coefficient (Do)of 1.1 X lo4 cmz/s (28). The heterogeneous electron transfer kinetic parameters for the reaction of cytochrome c a t silver electrodes have also been determined by single potential step chronoabsorptometry (SPS/CA) (16, 17). This method is directly analogous to chronocoulometry (36) but again the advantages of species selectivity and specificity are provided by the optical probe. SPS/CA methods for the determination of heterogeneous electron transfer kinetic parameters have been reported for irreversible (16) and quasi-reversible (17) cases. Figure 5 shows SPS/CA results obtained on the same cytochrome c sample and electrode that was used to obtain the CV and DCVA results shown in Figure 4. From these data ho’s,hwas determined to be 9.1 (+0.3) X cm/s with an cy
2338
ANALYTICAL CHEMISTRY, VOL. 59, NO. 19, OCTOBER 1, 1987
Table I. Heterogeneous Electron Transfer Kinetic Parameters for Chromatographically Purified Cytochrome c at a Smooth Silver Electrode cytochrome c,
(@MI
ko',,h,cm/s
199" 199" 91Q 9 10 181h 181h
1.5 (zt0.4)c X 91 ( ~ 0 . 4 x) 10-4 6.0 ( f 0 . 4 ) X 4.7 (10.2) x 10-4 2.8 (f0.4) X 3.7 (fO.l) x 10-4 1.6 (f0.4)X 1.6 (f0.4) X 2.2 ( M . 6 ) X
98h 51h
cy
0.55 (It0.02)c 0.52 (ztO.02) 0.52 (10.03) 0.52 (h0.02) 0.69 (A0.02) 0.65 (AO.01) 0.67 (10.07) 0.67 (10.12) 0.71 (zt0.08)
technique DCVAd
SPS/CA' DCVA SPS/CA
DCVA SPS/CA DCVA
SPSjCA DCVA
a Solutions prepared in 0.05 M Na2S04. Solutions prepared in 0.05 M Tris/cacodylic acid buffer, pH 7.0. CParentheses contain standard deviations based on four t o five different scan rates or
LL.04
potential step values for the same sample. Derivative cyclic voltabsorptometry. Single potential step chromabsorptometrv.
value calculated of 0.52 (f0.05). These values were then used to simulate the SPS/CA responses shown in Figure 5 (circles) and the agreement with experimental responses is good. The reaction of cytochrome c a t bare silver electrodes is a simple one electron transfer Butler-Volmer system (31-33) based on both experimental approaches, DCVA and SPS/CA. Table I summarizes the heterogeneous electron transfer kinetic parameters determined for the reaction of different purified cytochrome c samples a t freshly polished silver electrode surfaces. All of the kinetic parameters that are reported for the various concentrations of cytochrome c in both Na2S04electrolyte and Tris/cacodylic acid buffer conform to the criteria set forth by Matsuda and Ayabe (29) for quasi-reversible electron transfer systems. As indicated in Table I the agreement between DCVA and SPS/CA experiments is good. Reproducible results were obtained for more than 12 h with very little difference in electron transfer kinetics between experiments. The differences in kinetic parameters that are evident in this table are probably due to the use of a newly polished silver electrode surface in each experiment and new sample solutions. However, these differences would not prevent studies of solution environmental effects on the relative rates of heterogeneous electron transfer. The work presented to this point has dealt with smooth silver electrode surfaces. The reason for using silver as an electrode is that this is one of the surfaces which can increase the Raman scattering cross section by some 6 orders of magnitude for molecules adsorbed on or near its surface. This technique, surface-enhanced Raman spectroscopy (SERS), has been extensively used to study adsorbed cytochrome c ( 1 , 2 , 5-7). In order to obtain such large enhancement effects, the silver surface must be roughened. We plan to use SERS for studying the vibrational properties of adsorbed cytochrome c molecules in the near future, so an examination of the heterogeneous electron transfer kinetic parameters for this type of surface was studied as well. DCVA results for a 284 pM cytochrome c solution obtained a t a silver electrode roughened by a typical oxidation reduction cycle (ORC) pretreatment are shown in Figure 6 with a scan rate of 1.10 mV/s. A midpoint potential value of +0.262 V vs. NHE was determined for this electrode reaction. This reaction is also quasi-reversible with a ko'+ of 1.8 x lo-* cm/s and an (Y of 0.65.
CONCLUSIONS The critical result from this study is that the commonly used procedure of lyophilization produces denatured forms of cytochrome c that have a profound negative effect on the
03
02
01
E ( V v s NHE)
Figure 6. Derivative cyclic voltabsorptometry of 284 pM purified cytochrome c at a roughened silver electrode in 0.05 M Tris/cacodylic acid buffer, pH 7.0. Circles indicate simulated responses for ko's,h = 1.8 X cm/s and a = 0.65.
direct electron transfer kinetics of cytochrome c a t silver electrodes. The potential for using SERS to study the electron transfer reactions of heme proteins a t clean silver electrode surfaces is now clear. Registry No. Ag, 7440-22-4;Na2S04, 7757-82-6; cytochrome e , 9007-43-6; Tris, 77-86-1; cacodylic acid, 75-60-5.
LITERATURE CITED (1) Cotton, T. M.; Schultz, S. G.; van Duyne, R. P. J . Am. Chem. Sac. 1980, 102, 7960-7962. (2) Cotton, T. M.; Kaddi, D. Iorga, D. J . A m . Chem. Soc. 1983, 105, 7462-7464. .- . .. (3) Chao, S.; Robbins, J. L.; Wrighton, M. S.J . A m . Chem. Soc. 1983, 105, 181-188. (4) Bowden, E. F.; Hawkridge, F. M.; Biount, H. N. J . Nectroanal. Chem. 1984, 161, 355-376. (5) Taniguchi, I.; Iseki, M.; Yamaguchi, H.; Yasukouchi, K. J . Hectroanal. Chem. 1984. 175. 341-348. (6) Smulevich, G.; Spiro, T. G J . Phys. Chem. 1985, 89, 5168-5173. (7) Hiidebrandt, P.; Stockburger, M. J . Phys. Chem. 1986, 9 0 , 6017-6024. (8) Bowden, E. F.; Hawkridge, F. M.; Chiebowshi, J. F.; Bancroft, E. E.; Thorpe, C.; Blount, H. N. J . Am. Chem. Soc. 1982. 104. 7641-7644. (9) Brautigan, D. L.; Ferguson-Miller, S.; Margoliash, E. Methods Enzymol. 1978, 5 3 0 , 131-132. (IO) Kolier, K. 8.; Hawkridge, F. M. J . A m . Chem. Soc. 1985, 107, 7412-741 7 (11) Nicholson;.R. S. Anal. Chem. 1965, 3 7 , 1351-1355. (12) Sevcik, A. Collect. Czech. Chem. Common. 1948, 13, 349-377. (13) Bancroft, E. E.; Sidwell, J. S.;Blount, H. N. Anal. Chem. 1981, 5 3 , 1390-1394. (14) Bancroft, E. E.; Blount, H. N.; Hawkridge, F. M. Biochem. Biophys. Res. Cornmun. 1981, 101, 1331-1336. (15) Bancroft, E. E.; Blount, H. N.; Hawkridge, F. M. I n Nectrochemicaland Spectrochemical Studies of Siobgical Redox Components ; Kadish, K. M., Ed.; Adv. Chem. Ser. No. 201; American Chemical Society: Washington, DC, 1982; pp 23-49. (16) Albertson, D. E.;Blount, H. N.; Hawkridge, F. M. Anal. Chem. 1979, 5 .1 . ., 556-560 - - - - - -.
(17) Bancroft, E. E.; Blount, H. N.; Hawkridge, F. M. Anal. Chem. 1981, 5 3 , 1862-1866. (18) Van Buuren, K. J. H.; Van Gelder, 8. F.; Wilting, J.; Braams, R. Biochim. Blophys. Acta 1974, 333, 421-429. (19) Pyun, C. H.; Park, S. M. Anal. Chem. 1986, 58, 251-256. (20) Evans, J. F.; Blount, H. N. J . A m . Chem. Soc. 1978, 700, 4 19 1-4196. (21) Bard, A. J.; Faulkner, L. R. Nectrochemical Methods; Wiley: New York, 1960; pp 563-567. (22) Woodard, F. E.; Woodward, W. S.; Reilley, C. N. Anal. Chem. 1981, 5 3 , 1251A- 1266A. (23) Robinson, R. S.;McCurdy, C. W.; McCreery, R. L. Anal. Cbem. 1982, 5 4 , 2356-2361. (24) Kolb. D. M.; Kotz, R. Surf. Sci. 1977, 6 4 , 96-108. (25) Furtak, T. E.;Lynch, D. W. J . Elecfroanal. Chem. 1977, 7 9 , 1-17. (26) Kissinger, P. T.; Reilley, C. N. Anal. Chem. 1970, 4 2 , 12-15. (27) Hill, R. In Modern Methods of Plant Analysis; Peach, K., Tracey, M. V., Ed.; Springer-Verlag: New York, 1956; Voi 1, p 393. (28) Theorell, H. Biochem. Z . 1936, 285, 207-218.
Anal. Chem. 1987,59, 2339-2344 (29) Matsuda, H.;Ayabe, Y. 2.Hektrochem. 1955, 59, 494-503.
(30) Koller, K. B., Virginia Commonweakh University, Richmond, VA, 1986, unpublished work. (31) Butler, J. A. V. Trans. Faraday SOC. 1924, 19, 729-733. 132) -, Butler. J. A. V. Trans. Faradav SOC. 1924. 19. 734-739. (33) Erdey-Gruz, T.; Volmer, M. Z.’Phys. Chem., A h . A 1930, 150, 203. (34) Armstrong, N. R.: Vanderborgh, N. E.; Quinn, R. K. J . Phys. Chem. 1978, 80, 2740-2745. (35) Cusanovich, M. A. I n Bioorganic Chemistry; van Tamelen, E. E., Ed.;
.
2339
Academic: New York, 1978; Vol. IV, pp 117-145. (36) Christie, J. H.; Lauer, G.; Osteryoung, R. A. J. Electroanal. Chem. 1984, 7, 60-72.
~
RECEIVED for review March 3, 1987. Accepted June 11, 1987. The authors gratefully acknowledge the support of this work by the National Science Foundation (CHE85-20270).
Transient Response of an Enzyme Electrode Sensor for Glucose Pius H. S. Tse and David A. Gough* Department of Applied Mechanics and Engineering Sciences, Bioengineering Group, University of California, S a n Diego, La Jolla, California 92093
An analysis Is presented of the translent response of an enzyme electrode glucose sensor to substrate concentratlon changes. Parallel dlffuslon of substrates Is assumed. The results show that the response Is monotonlc and relatlvely simple when glucose concentration alone is changed. The time to cornpletlon of the transient becomes shorter most notably with reduced membrane thlckness and increased catalytlc actlvlty of the immobilized enzyme. When the concentrations of both substrates are changed simultaneously however, the translent current may undergo an Inflection before reachlng steady state.
The “enzyme electrode” principle has been proposed as the basis of a sensor for glucose ( I , 2). Several previous analyses of this sensor have focused on the steady-state response to glucose concentration and electrochemical transients (3-5). The transient response to concentration changes is also important in many applications but has received relatively little attention. In one configuration (3), the glucose sensor is based on a membrane containing the immobilized enzymes glucose oxidase and catalase, which promote the following overall reaction: glucose
-
+ 1/202 gluconic acid
(1)
The membrane is placed over an electrochemical sensor that is sensitive t o oxygen concentration. Glucose and oxygen diffuse into the membrane, the above reaction occurs, and excess oxygen not consumed by the reaction is detected by the oxygen electrode as a glucose-modulated, oxygen-dependent current, igmo. This current is subtracted from the current of a second oxygen electrode without the enzymes, io,which indicates the background oxygen concentration, and a glucose-dependent difference current, i,, results (5). The complete system is composed of a membrane-covered glucose electrode, an oxygen reference electrode, and an appropriate computational method for determining the glucose-dependent difference current. A comprehensive analysis of the transient response should take into account the response of the glucose electrode to both glucose and oxygen concentration changes and the simultaneous response of the oxygen reference electrode t o oxygen. The analysis should indicate an unambiguous means of
monitoring glucose concentration changes in the presence of oxygen concentration variations, and the transient response should converge with time t o the steady-state response. The results should also be cast in a form that is convenient for sensor design. Several previous analyses of the transient response have been carried out. In one case (6),numerical methods were used to solve a partial differential equation for the conservation of glucose alone and solutions were qualitatively compared to experimental results. In other cases, linear analyses (7,8) and implicit computational methods (9) were used to solve a simplified version of the equation. All previous studies have considered only changes in glucose concentration and are therefore of limited utility to describe sensor operation under conditions in which oxygen is variable or present a t relatively low concentration. There have also been few attempts to compare theoretical predictions with experimental results, and the role of the oxygen reference electrode has not been incorporated. We have described elsewhere a novel enzyme electrode design in which oxygen diffuses into the enzyme region from two directions but glucose enters from only one (10). This design was suggested as a simple means of assuring a sufficient supply of oxygen for glucose-limited operation under physiologic conditions. The steady-state response of this sensor was analyzed (10) and found t o be analogous to that of the conventional one-dimensional sensor when scaled by a dimensionless geometrical parameter. Similar analogies between the two designs are expected for the transient response. An analysis of the transient response of the one-dimensional sensor is therefore important not only for direct application but also as a basis for comparison of subsequent analyses of the more complex two-dimensional sensor. A model based on parallel diffusion is proposed here. The effects of various sensor design parameters on the response are described and comparisons of the model predictions to experimental observations are made. Our objective is to understand the factors that affect the transient response as the basis for the design of sensors to meet specific monitoring requirements.
THEORETICAL SECTION The system considered here is a homogeneous, enzymecontaining membrane place adjacent to the electrode surface. The membrane contains immobilized glucose oxidase and catalase, and the overall reaction is represented by
0003-2700/87/0359-2339$01.50/00 1987 American Chemical Society
