1264
Journal o f t h e American Chemical Society
(35) Hallock, S. A. Ph.D. Dissertation, The Ohio State University, 1974, p 90. (36) Gray, H. 6.; W i g , E.; Wojcicki, A,; Farona, M. Can. J. Chem. 1963, 41, 1281-1208. (37) Blakney. 6 . G.: Allen, W. F. lnorg. Chem. 1971, 10, 2763-2770. (38) Bamford, C. H.; Burley, J. W.; Coldbeck, M. J. Chem. SOC.,Dalton Trans. 1972, 1846-1852. (39) Wrighton, M. Chem. Rev. 1974, 74, 401-430. (40) While this process has not been studied in detail, UV spectral analysis indicates that the products are not those of Mn, CO)lo or [Mn(C0)4Br]2,the latter being the reported product in benzenej8 (41) Waltz, W. L.; Akhtar, S. S.; Eager, R. L. Can. J. Chem. 1973, 51, 25252529. (42) Owing to the strong near UV absorption of Mnz(CO)lo solutions, it was not feasible to observe at wavelengths below ca. 440 nm. (43) Dorfman, L. M.; Adams, G. E. Natl. Stand. Ref. Data Ser., Natl. Bur. Stand. 1973, No. 46. (44) Fieldhouse. S. A.; Fullam, 6. W.; Neilson, G. W.; Symons, M. C. R. J. Chem. SOC.,Dalton Trans. 1974, 567-569. (45) Data for the Mn(CO)@ case was that taken from the fourth to the eighth
(46) (47) (48)
(49) (50) (51) (52)
/
100:23
/
November 8,1978
pulse because, as in the case of Mn2(CO)lo, the decay rate decreased slightly from the first to the third pulse. Adams, G. E.: Michael, B. D.; Willson, R. L. Adv. Chem. Ser. 1968, 81, 289-308. -.. ... In contrast to the transient behavior in the cases of Mn2(CO)lo and Mn(C0)sBr. the decay rate of the transient for the MII(CO)~I system was relatively insensitive to the number of irradiation pulses. Kidd and Brown' have proposed that M!I(CO)~. can undergo rapid loss of CO: under our conditions, the decay of Mn(CO)5. is second order. In addition, Hudson and co-worker$ have proposed that Mn(CO)S. can react with Mnz(CO)lo by electron transfer; however, our results indicate that if this reaction can occur it is much slower than reaction 6 under our conditions. Von Hartel, H.; Polanyi, M. 2.fhys. Chem. (FrankfurtamMain) 1930, 716, 97-138. Anderson, 0. P.; Symons, M. C. R. J. Chem. SOC.,Chem. Commun. 1972, 1020-1021. Reference 12b, p 124. Elian, M.; Hoffmann, R . lnorg. Chem. 1975, 14, 1058-1076.
Preparation of Chemically Derivatized Platinum and Gold Electrode Surfaces. Synthesis, Characterization, and Surface Attachment of Trichlorosilylferrocene, ( 1,l '-Ferrocenediyl)dichlorosilane, and 1,l '-Bis( triethoxysily1)ferrocene Mark S. Wrighton," Michael C. Palazzotto, Andrew B. Bocarsly, Jeffrey M. Bolts, Alan B. Fischer, and Louis Nadjo Contribution f r o m the Department of Ckemistry, Massachusetts Institute of Technology, Cambridge, Massachusetts 021 39. Receiced May 1.5. 1978
Abstract: The synthesis and characterization of three ferrocene-centered, hydrolytically unstable, surface derivatizing reagents and their attachment to pretreated (anodized) Pt and Au electrode surfaces are described. Trichlorosilylferrocene ( I ) has been isolated from the reaction of Sic14 and lithioferrocene; ( 1 ,l'-ferrocenediy1)dichlorosilane (11) has been isolated from the reaction of Sic14 and 1,l'-dilithioferrocene; and I , 1'-bis(triethoxysily1)ferrocene (111) has been isolated from reaction of CISi(OEt)3 with 1,l'-dilithioferrocene. The species I, 11, and 111 have been fully characterized by 1H N M R , mass, and UV-vis spectra and elemental analyses. All are moisture sensitive and are capable of derivatizing anodized Pt surfaces. Detailed studies for derivatization of anodized Au using I I are described. I n many respects the properties of derivatized Au electrodes parallel those for derivatized Pt. Such derivatized electrodes exhibit persistent cyclic voltammetric waves at a potential expected for an electroactive ferrocene derivative. Greater than monolayer coverages a r e found in each case, as determined by the integration of the cyclic waves. The cyclic voltammetric parameters a r e as expected for a reversible, one-electron, surface-attached electroactive system except that the peak widths are broader than theoretical. This is attributed to chemically distinct ferrocene centers resulting from the oligomerization of the derivatizing reagent during the derivatization procedure.
In a recent preliminary account we outlined the chemical derivatization of anodized Pt electrode surfaces using trichlorosilyferrocene (I), which resulted in the persistent attachment of polymeric, reversibly electroactive ferrocene.' Also, we have shown that (1 ,I'-ferrocenediy1)dichlorosilane (11) is capable of derivatizing n-type semiconducting Si photoelectrodes,* illustrating the first example of photoelectroactive surface-attached species. These studies have added to a growing literature of chemically derivatized electrode surfaces .3-' Elaboration of the number and types of chemical derivatizing procedures is seemingly a necessary step in the illustration and understanding of fundamental and practical consequences of the resulting surfaces. I n this article we wish to report the synthesis and characterization of I, 11, and 1,l'bis(triethoxysi1yl)ferrocene (111) and their use as derivatizing reagents for anodized Pt electrode surfaces. W e also report the first derivatized Au electrodes using I1 as the derivatizing reagent, and outline a procedure for the surface pretreatment of Au that leads to successful derivatization. The structures 0002-7863/78/ 1500-7264$01.OO/O
of the three derivatizing reagents are shown below. A qualitative comparison of the behavior of the resulting derivatized surfaces is made.
Results and Discussion 1. Preparation of Compounds. The compounds presented here are, by design, extremely moisture sensitive. Therefore, care must be taken i n the exclusion of moisture in their preparation and subsequent handling. It should be noted that compounds I and I1 seem to react with glass over a period of time even when kept a t low temperatures (-10 "C). Because of this problem, in order to obtain satisfactory analysis, the samples were purified immediately before being sent for analysis, sealed in plastic ampules under dry nitrogen, and analyzed as quickly as possible. This slow decomposition also requires that the compounds be purified before use in electrochemical studies. All three complexes which we have prepared have been characterized by elemental analyses and the results are satisfactory. The mass spectral characterization provides a quick 0 1978 American Chemical Society
Wrighton et al.
1 Synthesis of
7265
Trichlorosilylferrocene
+
I
\
I
Fe
I
SiC1,
II
I
Fe
I11
way to determine whether I, 11, or I l l are present, since these complexes all give good parent peaks. For the complexes containing Cl's the parent peak region is complex and diagnostic owing to the characteristic distribution of 35Cland 37CI. ' H N M R and electronic spectral data have also been used to characterize I, 11, and 111. The low-energy absorption spectral features of the complexes prepared here are quite similar to those for ferrocene itself, which exhibits absorption maxima a t X 440 nm, t = 90, and 325 nm, t =50.'* Complex I1 exhibits a slightly red-shifted first absorption a t -470 nm and the shift is likely a consequence of distortion of the +-CsH5 ring systems from perfectly parallel; such has been found in the 1 ,l'-(ferrocenediy1)diphenylsilane analogue and it too has a red-shifted first absorption compared to f e r r 0 ~ e n e . IThe ~ first electronic absorption band has been attributed to a ligand field (d-d) transition and, consequently, should be sensitive to the geometry and nature of the immediate coordination sphere of the Fe.I8 Similarity of the low-energy region of the electronic spectra of 1-111 and ferrocene itself is in accord with an expectation that the electrochemical oxidation would occur at a similar potential for each of the species. Indeed, many simply substituted ferrocenes have Eo values in the vicinity of +0.4 V vs. SCE.20Complexes I and I1 are so moisture sensitive that solution electrochemical behavior is ambiguous. But complex 111 is qualitatively less reactive in solution, and its cyclic voltammetry in CH3CN/O.l M (n-Bu4N)C104 shows reversible waves establishing an Eo of +0.47 f 0.02 V vs. SCE. 2. Preparation and Characterization of Derivatized Electrodes. a. Surface Pretreatment. Pt electrode surfaces susceptible to derivatization with 1-111 have been prepared either by the procedure outlined in the literature13aor by a slight modification. For Pt disk electrodes the surfaces were hand polished to a mirror finish, anodized a t 1.9 V vs. S C E for 5 min in 0.5 M HzSO4, and then cycled between the 0 2 and H2 evolution potentials in 0.5 M H2S04 until the cyclic voltammograms were constant (-2 h). The electrode was then held a t 1.1 V vs. S C E in that medium until the current declined to < 1 yA/cm2. Such a procedure presumably gives a hydrated Pt oxide surface capable of reacting with the Si-CI or Si-OEt bonds of 1-111. Finding that the hand-polishing procedure was too labor intensive, we attempted the use of a procedure such that Pt foil electrodes could be prepared without polishing. Pt foil was cleansed in concentrated "03 and then was anodized, cycled, and anodized as in the procedure for the Pt disk electrodes. Qualitatively, comparable results have been obtained with both types of Pt surfaces (vide infra). Exposed Au surfaces were cleaned by dipping into concentrated H N 0 3 for 5 min a t 298 K. The electrode was then held
+
+
a t a potential of 1.9 V vs. S C E for 15 s in 0.5 M H2SO4 in HzO. This was followed by cycling the potential (linear sweep, 100 mV/s) from 0.0 to 1.0 V vs. SCE until the cyclic voltammogram remained unchanged, 1-3 h. Finally, the potential was held a t 1.9 V vs. S C E for 15 s. The electrode was then washed with distilled H2O and then with acetone and allowed to dry in air. b. Derivatizing Procedure. Pt surfaces to be derivatized were exposed to isooctane solutions of 1-111 under Ar as described in the Experimental Section. Au was derivatized with I 1 in a similar fashion. The procedures actually used (reaction time, temperature, concentration of derivatizing reagent) were arrived a t by trial and error. The one thing that is reproducible is that one can usually detect electroactive material after the derivatization, but the persistence, coverage, reversibility, etc., seem to depend on subtle factors not yet elucidated in the derivatizing procedure. Even carrying two electrodes through all procedures in parallel does not always yield electrodes having the same coverage of electroactive material. c. Characterization of Derivatized Electrodes. Pt electrodes derivatized with 1-111 have been characterized by cyclic voltammetry in CH3CN solutions of 0.1 M (n-Bu4N)C104 containing no deliberately added electroactive species. While all three derivatizing reagents do yield surface-attached electroactive material, there are some qualitative differences which are interesting. The data in Figures 1-4 and Table I are a collection of what we now regard as typical characteristics of Pt electrodes derivatized with 1-111. Similar characterization of Au derivatized with I1 is provided by the data in Figures 5-9 and Table 11. First, that electroactive material is attached to the Pt or Au is unequivocally established by the observation of cyclic voltammetric waves in CH3CN solutions containing no electroactive material; nonderivatized electrodes show no such waves in the potential window open in CH3CN. The position of the anodic and cathodic peaks, E,, and Ered, respectively, for Pt electrodes derivatized with 1-111 are listed in Table I. That the surface-attached species are electroactive ferrocene derivatives is in accord with peak positions in the range 0.42-0.62 V vs. SCE. Surface attachment of reversibly electroactive species generally has not been found to effect large changes in the redox It is known that simply substituted ferrocene derivatives all have fairly similar ferrocene itself is measured to be +0.40 f 0.02 V vs. S C E at a nonderivatized Pt electrode in CH3CN/O.1 M (n-Bu4N)C104, in our hands. The derivatizing reagents bearing only one Si atom, I and 11, result in cyclic peaks a t a more cathodic potential than from electrodes derivatized with 111. The Eo for 111 is more positive than for ferrocene itself, but unfortunately Eo's for I and I 1 were not measurable. For either derivatized Au or Pt electrodes, the area under the anodic and cathodic peaks is the same, revealing that the redox processes are chemically reversible. Several other characteristics point to kinetically reversible behavior as well. First, note that the peak to peak separations are below 59 mV in a number of cases a t scan rates of 100 mV/s. At lower scan rates the peak to peak separation often approaches zero, as expected for a reversible surface-attached redox couple.21 Further, the peak current is directly proportional to scan rate up to 500 mV/s and the peak to peak separation does not increase markedly a t scan rates up to 500 mV/s. For solution species the peak current is predictedZ2and observed to be proportional to (scan rate)'/2. Figures 1,2, and 4 illustrate such information for Pt electrodes derivatized with I1 and 111; similar data have already been published for Pt surfaces derivatized with I.' The data for Au derivatized with I1 are given in Figures 5 and 6. One feature of the cyclic voltammetric waves is not in accord with the theory for a reversible, one-electron, redox couple
+
+
Journal of the American Chemical Society
7266
/
100:23
/ Nocember 8, 1978
Table 1. Electrochemical Data for Derivatized Pt Electrodeso
derivatizing reagent
expt
I
1 2 3 4 5 6 7 1 2 3 4 5 6 7 8 9
II
IO Ill
1 2 3 4 5 6 7
v
EonV vs. S C E
vs. S C E
+0.52 +0.48 +0.56 +0.54 +0.55 $0.53 $0.55 +0.48 +0.47 +0.52 +0.53 +0.52 f0.53 $0.49 +0.48 +0.55 +0.53 +0.61 +0.62 +0.58 +0.60 +0.59 +0.62 +0.61
f0.50 +0.42 +0.53 +0.44 +0.43 +0.45 f0.46 +0.44 +0.44 +0.49 +0.45 +0.44 f0.47 +0.44 +0.43 $0.53 +0.48 +0.59 +0.58 +0.55 +0.57 +0.57 +0.58 +0.60
Ered,
coverageb x 1010, mollcrn2
AEilz, b , r rnV
3.9 27. I 60.0 82.5 49.7 177 286 9.3 15.0 28.6 28.8 29.3 33.6 33.7 34.5 52.2 57.1 3.1 4.8 9.3 14.6 16.8 19.9 23.1
400 300 340 300 240 240 270 250 220 260 225 380 350 220 240 240 270 380 540 380 320 370 390 310
aData presented here are from the 100 mV/s scan. bCalculated from the anodic wave. 'Peak width at half height.
I
I
I
I
-
