J. Phys. Chem. 1992,96,6585-6592 in accord with the barrier for FSSF (24 kcal/m01)~~ than the very small value (1.5 kcal/mol)2 calculated for the &-trans isomerism of S4. The observation of both isomers growing on matrix annealing requires a barrier separating the two structures that ex4 s the S2-S2binding energy. Reaction Mecbinisnrs. In the superheater experiments at 700 "C the spcctrum showed mainly S2and a weak S4 band centered at 530 nm. Decrease of the S4 band on annealing indicates that the reaction S2 S2 S4 involving ground state S2molecules does not occur in solid argon. The red S4 band and the S3band intensities also increased together as the superheater temperature was increased. On the other hand, the red S4 band was observed with relatively higher intensity in the discharge experiments, and it increased on annealing; discharge experiments are known to produce sulfur atoms.' These observations suggest that the superheater also produces S atoms and that S3is formed from the combination reaction S2 S S3and the structured red band belongs to the species formed from the reaction S + S3 S4 in solid argon. The earlier formation of the structured red S4band on annealing solid krypton samples3containing sulfur is proposed to arise from the S + S3 reaction as well. This suggests that addition of S to S3in the out-of-plane branched-ring orientation is less inhibited by steric effects than for the in-plane cis orientation. Although S3and both S4 isomers clearly exist in superheated gaseous sulfur,6* a considerable fraction of the S3and S4 observed in the present matrix studies must be due to matrix combination reactions involving sulfur atoms.
+
-
+
-
-
Conclusions Distinctly different electronic absorptions have been observed for S2, S3,and two different S4 clusters as a function of sulfur concentration in matrix discharge and superheater experiments. The S4 isomers give rise to broad 5 18-nm and structured 560660-nm absorptions, which exhibit reversible photochemical isomerism on selective irradiations. Infrared bands at 661.6 and 642.4 cm-I exhibit identical photochemical isomerism with the same photolysis treatment. Correlation of ab initio electronic energy, structure, and vibrational spectra calculations2with the observed visible and infrared absorption spectra with mixed sulfur isotopic multiplets' characterizes cis-planar and branched-ring nonplanar S4 structural isomers. This study provides strong evidence for a cis-planar s branched-ring photoisomerism for S4,
6585
which requires less atomic motion than the conventional cis s trans isomerization, and suggests that the branched-ring isomer may be involved in other photochemical rearrangements. Similar electronic spectra and photochemical rearrangements have been observed for Se4 and Te4, which will be reported in later papers. Acknowledgment. We gratefully acknowledge financial support from NSF Grant CHE 88-20764 and helpful conversations with H. F. Schaefer, K. Raghavachari, and C. Trindle.
References and Notes (1) Brabson, G. D.; Mielke, Z.; Andrews, L. J . Phys. Chem. 1991,95,79. (2) Quelch, G. E.; Schaefer, H. F., 111; Marsden, C. J. J. Am. Chem. Soc. 1990, 112, 8719. (3) Raghavachari, K.; Rohlfing, C. M.; Binkley, J. S.J. Chem. Phys. 1990, 93,5862. Raghavachari, K., personal communication of more recent calcu-
lations which find the cis isomer slightly lower in energy than the rectangular isomer, 1992. (4) Von Niessen, W. J . Chem. Phys. 1991, 95, 8301. (5) Meyer, B.; Oommen, T. V.; Jensen. D. J . Phys. Chem. 1971, 75,912. Meyer, B.; Stroyer-Hansen, T.; Oommen, T. V. J . Mol. Spectrosc. 1972,42,
335. (6) Steudel, R.; Jensen, D.; Godel, P. Eer. Bunsen-Ges. Phys. Chem. 1988, 92, 118. (7) Lenaine, P.; Piquenard, E.; Corset, J.; Jensen, D.; Steudel, R. Eer. Bunsen-Ges. Phys. Chem. 1988,92,859. ( 8 ) Billmers, R. I.; Smith, A. L. J . Phys. Chem. 1991, 95, 4242. (9) Rice, J. E.; Amos, R. D.; Handy, N. C.; Lee, T. J.; Schaefer, H. F. J . Chem. Phys. 1986,85,963. (IO) Andrews, L.; Spiker, R. C., Jr. J. Phys. Chem. 1972, 76, 3208. (11) Andrews, L.; Mielke, Z. J . Phys. Chem. 1990, 94, 2348. (12) Brewer, L.; Brabson, G. D.; Meyer, B. J. Chem. Phys. 1965,42,1385. (1 3) Mills, K. C. Thermodynamic Data /or Inorganic Sulfides, Selenides and Tellurides; Butterworth: London, 1974. (14) Dupuis, M.; Rhys, J.; King, H. F. J . Chem. Phys. 1976, 65, 1 1 . (15) Dupuis, M.; Watts, J. D.; Villar, H. 0.;Hurst, G. J. B. Computer Phys. Commun. 1989,52, 415. (16) Hassanzadeh, P.; Andrews, L. J . Am. Chem. Soc. 1992, 114, 83. (17) Greenough, K. F.; Duncan, A. B. F. J. Am. Chem. Soc. 1%1,83,555.
(18) Lenain, P.; Piquenard, E.; Lesne, J. L.; Corset, J. J . Mol. Stmcr. 1986, 142, 355. (19) Kelsall, B. J.; Andrews, L.; Schwarz, H. J . Phys. Chem. 1983, 87, 1413. (20) Dunkin, I. R.; Andrews, L.; Lurito, J. T.; Kelsall, B. J. J . Phys. Chem. 1985, 89, 2528. (21) Andrews, L.; Dunkin, I. R.; Kelsall, B. J.; Lurito, J. T. J . Phys. Chem. 1985. 89. 821. (22) Andrews, L.; Lurito, J. T. Tetrahedron 1986, 42, 6343. (23) Marsden, C. J.; Oberhammer, H.; Losking, 0.;Willner, H. J . Mol. Srruct. 1989, 193, 233.
A Surface-Enhanced Raman Spectroelectrochemical Study of a Number of p-Amino-Substituted Tetraphenylporphyrins in Aprotic Media Charles M. Rosten, Ronald L. Birke,* and John R. Lombardi Department of Chemistry, The City College, City University of New York, New York, New York 10031 (Received: February 14, 1992)
The surface-enhanced Raman scattering (SERS) spectra of five paminesubstituted tetraphenylporphyrins were investigated at a silver electrode in the aprotic solvent acetonitrile. Good quality spectra were obtained for both the neutral porphyrin and its first reduction product, the radical anion. Fluorescence interference from the porphyrin was completely quenched by the surface. SERS spectra allowed for the identification of two geometries of the adsorbate. Cyclic voltammetry and UV-vis thin-layer spectroeiectrochemistry were used to substantiate information obtained by SERS. The utility of SERS as a technique for identifying electrochemical products in a nonaqueous environment, where elucidation of the electrode mechanism is simpler and radical anions are more stable, is demonstrated.
6586 The Journal of Physical Chemistry, Vol. 96, No. 16, 1992
Hosten et al.
1548
\
Rz
Figure 1. Structure of p-amino-substituted TPP and the labeling of the carbon atoms in the macrocycle. TPP: R,,R2, R,, and R4 = H; AOOO: R1 NH2, R2, R3 and R4 = H; AAOO: R I and R2 = NH2, R3 and R4 = H; AOAO: R, and R3= NH2, R2 and R4 = H; AAAO: R I , R2, R3= NH2, R4 = H; AAAA: R,,R2,R3 and R4 = NH2.
although providing information on molecular structure which could not be obtained electrochemically, do not offer the degree of specificity which is necessary for the complete characterization of electrochemically generated porphyrin intermediates. Since its discovery by Fleischmann et al.,s surface-enhanced Raman spectroscopy (SERS) has been used"' to characterize the structure and orientation of electrochemicallygenerated species on roughened Ag, Au, and Cu electrodes. The large enhancement of the SERS spectra along with the high degree of specificity offered by Raman spectroscopy makes SERS an ideal tool for the characterization of electrochemically generated porphyrin species. SERS has rarely been coupled with electrochemicaltechniques in the study of porphyrin electrochemistry in nonaqueous media because of the difficulties in preparing SERS active electrodes in nonaqueous solvents. There have been, however, some SERS studies of compounds in nonaqueous solvents, e.g., pyridine in dimethylformamides and tri~(2,2'-bipyridine)ruthenium(II)~in acetonitrile, and one study of a porphyrin compoundlo also in acetonitrile. In the latter case, Sanchez and Spiro'O studied the transition from low-spin Fe(I1) to high-spin Fe(II1) in the dimethyl ester of hemin. In the present paper we have studied the electrochemical and spectroscopic properties of a number of p-amino-substituted tetraphenylporphyrin (TPP) compounds in acetonitrileusing SERS spectroscopy. The results show that at -0.6 V an adsorption change occurs, resulting in a more firmly bound porphyrin adsorbate. The orientation of the adsorbate is established, and spectroscopic evidence is presented to show that expansion of the porphyrin ring aromatic system occurs to include the phenyl substituents. Experimental Section The derivatives of tetraphenylporphyrin used in this study incorporated p-amino substitution at one or more of the phenyl groups (Figure 1). The labeling of specific carbon atoms on the macrocycle is given in the figure, and the nomenclature used is given in the figure caption. The porphyrins were a gift from Prof. C. Guo of the Beijing Institute of Technology and were synthesized" by condensation of pyrrole, benzaldehyde, and parasubstituted benzaldehyde at a different molar ratio in organic solvents. The aminoporphyrin with the desired substitution was isolated from the product mixture by chromatographicseparation and recrystallization from solution.'' Porphyrin composition and structure were confiied by elemental analysis, mass spectrometry, IR, and NMR methods.I2 The p-amino-substituted porphyrins studied had increasing numbers of amino substituents, from one to four. The porphyrin with one p-amino substituent is labeled A000, where "A" represents an amino group and "On a hydrogen atom. When two amino
400
1000
1604
RAMAN SHIFT IN WAVENUMBERS
Figure 2. SERS spectra of five p-amino-substituted TPP compounds in acetonitrile on a silver electrode at -0.4 V with 488-nm laser excitation: (a) AAOO, (b) AOAO, (c) A000, (d) AAAA, and (e) AAAO. See text for definitions.
groups are added to the molecule, they can be added to adjacent phenyl groups, resulting in the amino groups being cis to each other or, on opposite phenyl groups, trans to each other. The cis compound is labeled AAOO, while the trans compound is labeled AOAO. The three-aminesubstituted compound is labeled AAAO and the four-substituted compound, AAAA. The four-aminosubstituted porphyrin, meso-tetrakis(4-aminophenyl)porphyrin, is referred to as TAPP in this paper. Dilute solutions, approximately 1 X M, of the porphyrin in acetonitrile were used for all SERS spectra and cyclic voltammograms. Optima grade acetonitrile (Fisher products) was dried over 4-A molecular sieva prior to use. Tetrabutylammonium perchlorate (TBAP) was recrystallized from pentane/ethyl acetate, dried under vacuum, and stored in a desiccator prior to use. Solutions of the porphyrin were degassed using nitrogen which was first passed through a drying agent and then acetonitrilebefore passing through the porphyrin solution. A Spectra Physics Model 164 argon ion laser was used for 488-nm excitation. A Cary Model 11 spectrometer was used for the visible spectroelectrochemistry. Cyclic voltammograms were recorded using a Model 175 universal programmer and a Model 173 potentiostat/galvanostat, both EG&G Princeton Applied Research. SERS and Raman spectra were recorded with a Spex Model 1401 double monochromator with a wavenumber resolution of about 4 cm-I. Photon counting detection was used, and the intensities were recorded digitally and are presented here unsmoothed. The sample cell for SERS consisted of a 99.999% pure silver working electrode, a platinum counter electrode, and a saturated calomel electrode (SCE) as the reference. All potentials are reported relative to the SCE. A bridge was used to separate the reference electrode from the rest of the electrochemical system. The oxidation reduction cycle (ORC)pretreatment was performed using a 2-s pulse from 0 to +O.5 V and then returning the electrode to 0.0 v.
p-Amino-Substituted Tetraphenylporphyrins
The Journal of Physical Chemistry, Vol. 96, No. 16, 1992 6587
1010
I
,,.L
A Figure 4. SERS of TAPP in acetonitrile with 488-nm laser excitation on a silver electrode. (a) 0.0, (b) -0.4, (c) -0.6, (d) -0.7,and (e) -1.1
V. the order of 10-6-10-7 M, to be used in some of our investigations. The potential dependence of the SERS spectra of the five pamino-substituted TPP compounds used shows that for four of the compounds, AAAA, AAAO, AAOO, and AOAO, spectra were obtainable at potentials as negative as -1.4 V. For the compound A 0 0 0 , spectra were obtained only up to a potential of -0.5 V. At more negative potentials no spectra are obtained, suggesting that the molecule is desorbed from the electrode surface. This facile desorption is a reflection of the strength of the interaction between the porphyrin molecule and the electrode surface. The four porphyrins AAAA, AAAO, AAOO, and AOAO all showed similar spectroscopic and electrochemical behavior. The results which are presented in this paper are for compound AAAA, meso-tetrakis(4-aminopheny1)porphyrin (TAPP). Its behavior is characteristic of that of the other three porphyrins studied. Metal Incorporation. Cottonl5 in her SERS study of the water-soluble porphyrin tetrasodium meso-tetrakis(4-sulfonatopheny1)porphine (TSPP) observed bands at 1346,425, and 355 cm-’ which were not present in the solution spectra. She characterized the bands at 1346 and 425 cm-’ as being indicative of metal incorporation into the porphyrin ring. An examination of the potential-dependent spectra of TAPP between 0.0 and -1.1 V (Figure 4) shows peaks at 418 and 1340 cm-l. These peaks are equivalent to those found by Cotton, and their absence from the spectra when the ORC was performed ex situ is further indication that they are due to metal incorporation. The potential dependence spectra of TAPP show that the intensities of the 418- and 1340-cm-’ bands greatly increase at -0.6 V. This could be due to a closer interaction between the incorporated metal and the electrode surface. It is clear, however, that metal incorporationdoes occur for TAPP and that this incorporation occurs during the anodization step when the potential of the electrode is pulsed to +SO0 mV. The 355-cm-I band observed by Cotton in the SERS of TSPP was not observed in our study. Cotton in examining the spectrum of AgTSPP also failed to observe this band. Itabashi16also studied
6588 The Journal of Physical Chemistry, Vol. 96, No. 16, 1992
Hosten et al.
TABLE I: Assignment of the SERS Bands for TAPP (in cm-l) TSPP (ref 15) TAPP (at -0.4 V) assignment 425 418 metal sensitive 670 porphyrin deformation 922 solvent 886 962 (a) phenyl 1008 1002 4cu-G) 1080 1078 (a) 6 (Cp-H) C,-Ph 1174 1232 1236 (a) v(C,-N) aromatic C-N 1274 metal sensitive 1346 1342 (a) u(C,-N) + 6(Cp-H) 1362 1356 (a) solvent 1364 unassigned 1486 1555 1548 (a) V(C,-cnl) 1597 phenyl 1598 (a)
I
- 1.0 15 microamp mwoamp
+irdicatasscan direction direchon
1010
- 0.4 I .1.2
v
POTENTIAL vs SCE
Figure 6. Cyclic voltammogram of TAPP dissolved in acetonitrile with TBAP as the supporting electrolyte. Potential range from 0.0 to -1.2 V. Scan rate 100 mV/s. The polished Ag electrode has an area of 1.6 mm2.
- 1.0 v
I
25 microam
F
+idcalm
w n dirrctioi
F
I9
0
RAMAN SHIFT IN WAVE NUMBERS
Fgrre 5. SERS spectra of TAPP in acetonitrile on a silver electrode with 488-nm laser excitation at (a) -0.5, (b) -0.6, and (c) -0.7 V.
TSPP, and his work again failed to show this 355-cm-l band. Because the 355-cm-' vibration was observed only for TSPP in H20, Cotton assigned this band as being due to a ligand mode involving water or the sulfonato group of another TSPP molecule. Itabashi also observed metal incorporation at -0.4 V for TSPP in 0.05 M sulfuric acid solution. No such incorporation was observed at -0.4 V in our study. Spectrorcopy and Electrochemistry. Table I lists the bands observed by Cotton in her study of TSPP, the bands observed for TAPP in our study, and assignments for the Raman bands for TAPP. The bands labeled (a) are assigned from Cotton's study of TSPP. The effect of the potential on the SERS spectrum of TAPP is given in Figure 4, which shows that for potentials at or more negative than -0.6 V (Figure 4d) there is a significant increase in the intensity of the spectrum. Figure 5 shows portions of the potential-dependent SEW spectra in the region around 900-1 100 and 1500-1600 cm-I of porphyrin AAAA on a roughened Ag electrode at potentials of -0.5, -0.6, and -0.7 V. At -0.5 V a singlet peak is observed at 1002 cm-I. When the potential is changed to -0.6 V two peaks are observed, one at 1002 cm-l and the other at 1010 cm-I. Similarly, at -0.5 V one peak is observed
- I 0.ov
- 0.4 V .I .2v
POTENTIAL vs SCE Figure 7. CV of TAPP dissolved in acetonitrile with TBAP as the supporting electrolyte and a potential range from 0.0 to -1.2 V on a polished Ag electrode and a scan rate of 500 mV/s.
at 1548 cm-' while at -0.6 V two peaks are observed, one at 1534 cm-'and one at 1548 cm-l. As the potential is made mote negative both the 1002- and 1548-cm-' bands decrease in intensity while the 1010- and 1534-cm-I bands increase in intensity. To assist in characterizing the process observed spectmaqically at -0.6 V, cyclic voltammetry (CV) was performed on TAPP in acetonitrile. The CV curves were obtained at scan rates ranging from SO to 500 mV/s. Figures 6 and 7 for silver and Figure 8 for gold electrode surfaces show representative CV curves at either 100 or 500 mV/s. An early study of the polarographic reduction of porphyrins in nonaqueous medial8 showed that, for HzTPP and for a number of its metalated analogues, the first reduction peak occurred at -1.05 V for HzTPP and potentials as negative as -1.35 V for MgTPP. All of the derivatives of TPP showed a prewave having a peak potential of -0.7 V for TPP and -0.67 V for MgTPP. No discussion was given in the report18 as to the reason for this
p-Amino-Substituted Tetraphenylporphyrins
The Journal of Physical Chemistry, Vol. 96, No. 16, 1992 6589 422
422
.%%
li
450
5 mini
.0.07
I
400
l
l
500
WAVELENGTH Inm)
-1.2v
0.ov
POTENTIAL vs SCE
Figure 8. CV of TAPP in acetonitrile with TBAP as the supporting electrolyte on a polished gold electrode (scan rate 100 mV/s, electrode area 1.6 mm2).
prewave except to say that the prewave was not due to an impurity and its peak height did not increase linearly with porphyrin concentration. An examination of the cyclic voltammogram of TAPP at a scan rate of 100 mV/s (Figure 6) shows two cathodic peaks at -0.6 and -1 -0V, and on the anodic scan two small peaks are observed at -0.3 and -0.4 V. The CV scans do not show well-defined oxidation peaks. Other studies of porhyrins in nonaqueous media have shown a similar lack of reversibility which is believed to be due to the reaction of the electrochemically generated products with small traces of water from the solvent. The small anodic peak observed in the CV at -0.4 V is probably due to the reoxidation of the product formed at -1.0 V. The cyclic voltammogram peak currents for the processes occurring at -0.6 and -1 .O V behave differently with changing scan rate. A plot of ipvs the square root of the scan rate for the -1.0-V peak is linear with a linear correlation coefficient of 0.983. No such linearity is obtained for a similar plot of i, for the -0.6-V peak. A plot of peak current vs scan rate is close to linearity for this peak with a linear correlation coefficient of 0.949. These results indicate that the process that is occurring at -0.6 V is an adsorption-controlled process. Similar results were obtained when the working electrode was switched from Ag to Au, with the only difference being that the positions of the peak potentials were shifted to more positive values by approximately 200 mV on the gold electrode (Figure 8). Further confinnation of the non-Faradaic nature of the process Occurring at -0.6 V on Ag was obtained by performing thin-layer visible spectroelcctrochemistryof TAPP in an acetonitrile solution. A tiuec-elcctrode electrochemical system was used. This consisted of a gold minigrid working electrode having a transmission of 60%, a platinum counter electrode, and a saturated calomel electrode. Figure 9 shows the potential-dependent spectra obtained for the 400-500-nm region with TAPP. At 0.00 V (Figure 9a) the
Figure 9. Potential-dependent absorption spectrum from a thin-layer spectroelectrochemical cell.
spectrum of TAPP showed a peak at 422 nm. The potential of the system was changed in steps of -0.2 V. No changes were observed in the spectra until -0.8 V (Figure 9b). At this potential two Soret bands were observed, one at 422 nm and the other at 450 nm. As the time at -0.8 V was increased (Figure 9c) the 422-nm band decreased in intensity while the 450-nm band increased in intensity. Wilson and co-w~rkers'~ using TPP in dimethylformamide (DMF)have shown that a 418-nm band is due to TPP while a 448-nm band is due to the radical anion. From this result we conclude that the spectrum obtained at -0.8 V on gold, equivalent to the -1.0 V on silver, is that of the radical anion. No change in the UV-vis spectra was observed at -0.4 V which corresponds to the -0.6-V peak on the silver electrode. This confirms that the process occurring at -0.6 V is not the first reduction of the porphyrin. The UV-vis spectra conclusively show that first reduction product of TAPP occurs at -0.8 V on Au which is equivalent to the -1.0 V on silver. Kadish et a1.20 studying nickel meso-tetrakis( l-methylpyridinium-4yl)porphine in DMF observed two peaks in the &ret region of the UV-vis spectrum of the porphyrin. One peak was observed at 440 nm and the other at 420 nm. They assigned one of the peaks (420 nm) to the monomer and the other (440 nm) to the dimer. In an attempt to establish whether TAPP aggregates in solution and, therefore, whether the process occurring at -0.6 V was the separation of the aggregate into the monomer, UV-vis spectra were observed over a range of concentrations. All showed one Soret band at 422 nm and a Beer's law plot (Figure 10) was linear, with a linear correlation coefficient of 0.989, indicating that there was no aggregation of TAPP in acetonitrile solution and that the process occurring at -0.6 V was not a dissociation of the aggregate into the monomer. The shape of the -0.6-V peak in the CV along with the results presented before suggests that the process occurring at -0.6 V is an adsorption process. Weaverz1studied the SERS of benzene adsorbed on gold electrode and observed a single peak at 973 cm-' at an electrbde potential of -0.5 V. He assigned this as the ring
6590 The Journal of Physical Chemistry, Vol. 96, No. 16, 1992
W
.I 0
10
20
CONCENTRATION IN M x 10-5
Figure 10. Plot of absorbance against concentration for TAPP dissolved in acetonitrile. TABLE II: Comparison of the Normal Raman Spectrum of TAPP witb Its SERS Spectrum (488-nm Excitation, Raman Shift in cm-', TAPP Dissolved in Acetonitrile) SERS SERS normal normal A Raman (-0.4V) A (-0.4V) Raman 2 1274 1274 0 920 922 12 1314 1326 12 962 974 18 1552 1550 2 1020 1002 1598 6 1078 12 1604 1090
breathing mode. When the potential was shifted to more positive values, this peak split into two peaks. Since there was no Faradaic electrochemical process occurring at this potential, he explained this split as beiig due to the presence of two adsorbate geometries at the electrode surface. The appearance in our study of two resolvable pairs of peaks, 1002-1010 and 1534-1548 cm-' at -0.6 V, suggests the two binding geometries. The peaks at 1010 and 1534 cm-l can be associated with one bound surface species and those at 1002 and 1548 cm-' with another bound species. The increase of intensity in the 1010- and 1534-cm-' SERS spectral bands as the potential moves negative is evidence that these bands represent an adsorbate which is in closer contact with the electrode and thus more tightly bound. The spectral and electrochemical results allow for the conclusion that the process occurring at -0.6 V is an adsorption process resulting in an adsorbate which is more tightly bound to the electrode. Also, the spectrum obtained at -1.0 V is established as being that of TAPP radical anion. The spectrum at -1.0 V shows a number of differences when compared to the spectrum at more positive potentials. First there is a significant decrease in intensity of the spectrum. Also, the band at 1010 cm-' at -0.9 V broadens and shifts to 1002 m-',suggesting that the less firmly bound adsorbate is last. Finally, the band at 1272 cm-'decreases in intensity while the bands at 1342 and 1356 cm-' are no longer present. Mentation of the Adsorbed PorpByrin. Because TAPP is highly colored, there is some absorption at the wavelengths used in this study. It also fluoresces at longer wavelengths, in the red. The solution normal Raman spectrum was obtained by passing the 488-nm laser line through the sample and reflecting it from a polished silver electrode to the entrance slits of the monochromator. Table I1 shows a comparison of the solution Raman and the SERS spectrum at -0.4 V. The intense peak at 922 cm-'in the solution Raman is due to the solvent, and it is found at about 920 cm-' in the SERS spectrum. A comparison of the intensity of the solvent peak in both the solution Raman and the SERS spectrum shows that it is greatly reduced in the SERS spectrum. This illustrates that the SERS of the adsorbate is enhanced to a greater extent than the solvent. The 974-cm-' band in the solution Raman is the 962-cm-' band in the SERS downshifted by 12 cm-I. The 102Oar-' solution Raman band is observed at 1002 cm-'in SERS downshifted by 18 cm-' while the 109onn-' solution Raman band
Hosten et al. is the 1078-cm-I SERS band downshifted by 12 cm-I. A comparison of the intensities of the 974- and 1020-cm-I solution Raman bands with their SERS counterparts shows that in the solution Raman the 974-cm-' band is more intense than the 1020-cm-' band, while in the SERS the reverse is true. Other bands at longer frequencies show smaller shifts of 6 cm-' or less, and these shifts are considered insignificant. The 1274-cm-' band is unchanged for both the solution Raman and the SERS, while the 1550-cm-' solution Raman band is virtually the same in the SERS and the 1598-cm-I solution Raman is downshifted by 6 cm-I. The strongly downshifted bands in the SERS spectrum are at 962, 1002, and 1078 cm-I. The 962-cm-I band we assign as a phenyl vibration. CottonI5in her study of TSPP observed the phenyl vibration at 886 cm-'. Burke et al." observed the phenyl vibration at 890 cm-' in Fe(TPP)Cl and (FeTPP(Im)2)Cl. Because of the absence of a band around 890 cm-' in the SERS spectra of TAPP and the fact that the band observed at 962 cm-' in TAPP is close to the 890-cm-I band, we assign it to a phenyl vibration. The 1002-cm-' band is assigned as a v(Ca-C,) and the 1078-cm-I is a G(Cp-H).15 A downshift of a stretching mode strongly indicates adsorptive interaction of one of the atomic participants involved in the stretching mode to a coordinating metal. These downshifts suggest a flat orientation of the molecule. Of the possible orientations that TAPP can assume when it is adsorbed on the electrode surface we will consider two orientations: one with the molecular plane parallel to the electrode surface and the other with the molecular plane perpendicular to the electrode surface. A perpendicular orientation of the molecule would result in a strong interaction between the amino group and the surface. This would manifest itself by a downshift in the amino vibration. Examination of the 1600-4500-cm-' region failed to detect the amino vibration. A number of porphyrin studies15J6of sulfonato and carboxyl TPP have also failed to observe the vibrations due to the substituents on the phenyl ring. A perpendicular orientation should also affect the aromatic carbon to nitrogen vibration, and a downshift of this vibration should be expected.22 In the assignment of the Raman spectra of aniline,22the (C-N) stretch is listed as occurring at 1278 cm-I. Examination of the NR spectra shows a moderate band at 1274 cm-l while the S E W spectra show a weak band at 1274 cm-l which increases in intensity and becomes well-defined at -0.6 V. This band we assign as the aromatic C-N vibration, and the absence of any appreciable downshift (only 2 cm-' is observed) suggests that bonding between the electrode and the porphyrin does not involve this vibration. This rules out the possibility that the molecule is oriented perpendicular to the electrode surface. The above evidence is consistent with other results23regarding the geometry of the porphyrin molecule. The r-conjugated system of the porphyrin is considered to consist of the 16-atom, 18-relectron ring that excludes the &carbons.24 The (C&) bond has been shown to have approximate doublebond character.24 In the 16-atom, 18-r-electron ring system there is a flat bonding about the interior carbon atoms. The pyrrole ring is tilted about 6.6°,25resulting in the (Cp-C,) not being coplanar with the rest of the molecule. The shifts observed in the SERS spectra for the v(Ca-C,) and the b(CgH) vibrations are consistent with this molecular geometry. Also, because the aniline group is at an angle of between 40° and 60° to the rest of the molecule the (C,-Ph) and the aromatic (C-N) vibrations should not be in close contact with the surface, and it is expected that no downshift in these vibrations should be observed if the porphyrin molecule is adsorbed flat. The absence of any shift in the 1274-cm-I band further supports this assumption. SERS from a Roughened Dried Electrode. In an attempt to obtain information regarding the role of the solvent in the orientation of the TAPP molecules on a silver electrode, spectra were obtained from a roughened dried electrode (DE). The electrode was first roughened in situ. It was then removed from the solution and left to dry in air. Intense well-resolved spectra were obtained from this roughened dried electrode, and the spectrum is presented in Figure 11. An examination of the 900-1600-cm-' region of the spectrum shows
The Journal of Physical Chemistry, Vol. 96, No. 16, 1992 6591
p-Amino-Substituted Tetraphenylporphyrins
1536
h
1080
1000
1200
1400
1600
cml
RAMAN SHIFT IN WAVENUMBERS
Figure 11. SERS from a silver electrode roughened in an acetonitrile solution of TAPP and left to dry (488-nmlaser excitation). TABLE III: Comparison of the S E W of TAPP on a Roughened Dried Electrode (DE) Against an in Situ Electrode at 0.0 and -0.6 V (488-nm Excitation An Electrode)" in situ SERS in situ SERS 0.0 V -0.6 V DE 0.0 V -0.6 V DE
982 1016 1080 1184 1238 a
962 1002 1078 1174 1236
962 1010 1078 1174 1234
1322 1344 1482 1536 1602
1326 1342 1486 1550 1598
1324 1340 1484 1534 1586
All values are in cm-I.
that the 922-cm-' band which is present in the solution spectrum is absent in the spectrum from the dried electrode. This band is the solvent peak, and its absence indicates that all the solvent has evaporated from the electrode. Therefore, the spectrum observed from the roughened dried elmode represents interactions between the electrode and the TAPP molecules without solvent interference. A comparison of the spectra from the DE with the in situ potential-dependent SERS spectra of TAPP (Table 111) indicates that the spectrum from the DE is similar to that of the in situ SERS spectrum at -0.6 V. The band at 1016 cm-' on the DE is equivalent to the lOlO-cm-' in situ band. The 1184-cm-' band which is weak in the in situ spectra at potentials more positive than -0.6 V increases in intensity when the potential is made more negative. This vibration is intense in the DE spectrum. Also,the 1536-cm-' band from the DE is equivalent to the 1534-cm-' band which is found at -0.6 V in the in situ spectra. The excellent agreement between the two sets of spectra, except for the 1586cm-' band, is further evidence that the process occurring at -0.6 V is a reorientation of the molecule and that the spectrum obtained at -0.6 V is that of the more strongly adsorbed porphyrin molecule with solvent molecules displaced. There are also some differences between the DE spectrum and the in situ SERS spectra. The 962cm-I band in the in situ SERS spectra is observed at 982 cm-l on the DE, and the band found at 1274 cm-I in the in situ SERS spectra is absent from the DE spectrum. Also,the 1174-an-' band of the in situ SERS spectrum is upshifted by 10 cm-l to 1184 cm-' on the DE. The 982-cm-' band has been assigned as a phenyl vibration. We assign the 1274-m--' band as an aromatic (C-N) stretch based on the observation by Shindo of a band at 1278 cm-I in the SERS spectrum of anilineZ2which he identified as the aromatic (C-N)
vibration. In Shindo's study this band at 1278 cm-'was observed in the solution spectra of aniline, but when the molecule was adsorbed flat onto the electrode surface this band disappeared. Shindo used this to characterize the orientation of the molecule on the electrode surface. The presence of the 1274-cm-' band in the in situ SERS spectra indicates that the aniline portion of the molecule is not adsorbed flat on the electrode surface. X-ray analysisZSof triclinic TPP shows that the phenyl groups are rotated by about 60-63O for free base TPP and are nearly perpendicular to the porphyrin ring in Ru(C0)TPP. For TAPP adsorbed parallel to the electrode surface, assuming no molecular distortion, the phenyl groups would not be parallel to the surface but would be at an angle, the size of which would depend on the value of the dihedral angle between the phenyl substituent and the porphyrin ring. Calculations of the phenyl ring rotation angle in tetraphenylporphyrins show that the minimum-energy configuration occurs when the angle between the porphyrin plane and the phenyl ring is 44O.% There is, therefore, a wide range of values for the dihedral angle when the molecule is in solution. The absence of the 1274-cm-' band in the DE spectra indicates that the aniline portion of the molecule has become more coplanar with the porphyrin ring and has changed its orientation. The estimated activation energies for this rotation range from 15.6 kcal/mol for TiO(p-PrTPP) to 18.6 kcal/mol for Ru(CO)(p-iPrTPP)(4,5-dimethylpyridazine).27 Phenyl rotation requires some distortion of the porphyrin ring, and the ease with which coplanarity is achieved for different porphyrins is a reflection of the differences in the ease of distortion of the complex. Rotation of the phenyl groups into coplanarity with the porphyrin ring should result in some form of resonance interaction between the conjugated ring system of the porphyrin and the phenyl substituent. This effect has received some investigation.28 However, the extent and importance of this *-delocalization are still unresolved. Molecular orbital c a l c ~ l a t i o n focus s ~ ~ on the 16-atom, 18-*-electron ring system and ignore any contribution due to delocalizationof the *-electron ring system over the phenyl substituents. There is experimental evidence which has been interpreted as indicating delocalization of the *-electron system work ' done on linear into the phenyl substituents. An e ~ a m p l e ~is. ~ force energy correlations. However, other explanations2which ignore *-delocalization have been presented to explain the experimental data. Because of its specificity, resonance Raman spectroscopy has been used to identify r-delocalization. Early studies on protoporphyrin and its derivatives could not unequivocally identify the C=C stretching frequency and so could not demonstrate whether the vinyl group was part of the delocalized *-system. W. H. Fuchsmans3studied the resonance Raman spectroscopy of three tetrakis(phalopheny1)porphines and their derivatives. He observed halogen-sensitive bands, of medium intensity, in the 1050llOO-cm-l region of the spectrum. These bands were assigned as phenyl ring modes containing significant C-X stretching character. They are strong in the IR but very weak in the RR of the solid porphyrin dication, solid neutral porphyrin, and the solid Cu(I1) porphyrin. The authors attribute the presence of these bands and their increase in intensity in solution as an indication of delocalization of the *-electron system. Fu~hsman'~has suggested that even a dihedral angle of over 60° would not preclude *-overlap between the phenyl groups and the porphyrin aromatic system. These conclusions are strongly denounced by
theoretician^.^^ To further investigate the effect of phenyl rotation and the possible extension of the *-system to include the phenyl rings, we focus on the C,-Ph vibration. There has been some debate as ~ ~ has assigned to the exact assignment of this v i b r a t i ~ n .Spiro a band at 1182 cm-' in a nickel porphyrin compound to the C,-Ph ~ i b r a t i o n .A~ ~band at 1187 cm-' in (FeTPP)20 has also been assigned as the C,-Ph vibration." We assign the vibration found at 1174 cm-l in the in situ SERS spectra of TAPP as the C,-Ph vibration. Rotation of the phenyl ring into coplanarity with the lQatom, 18-*-electron aromatic system should result in electron donation to this bond, resulting in an increase in its bond order
6592 The Journal of Physical Chemistry, Vol. 96, No. 16, 1992
which should express itself by a shift to higher frequency of the C,-Ph vibration. Cotton36observed a similar effect in the study of 4,4’-bipyridine, and C. Takahashi” observed a similar upshift in the study of the biphenyl negative ion. Examination of our spectra shows that the C,-Ph vibration is observed at 1174 an-’in the in situ SERS spectra and at 1184 cm-’ for the DE spectra. Although the upshift of 10 cm-I is relatively small, it is still significant and along with the upshift of 20 cm-I for the phenyl vibration is additional proof of rotation of the phenyl ring and an indication of some degree of increased overlap between the *-electron system of the porphyrin aromatic system and the phenyl groups. It appears that when TAPP is adsorbed onto the electrode in the presence of solvent molecules the solvent molecules are coadsorbed with the porphyrin. When the electrode is dried, the solvent molecules evaporate and the TAPP molecules are left directly adsorbed unto the electrode surface without the interaction of solvent molecules. This results in a more tightly bound TAPP which forces rotation of the phenyl groups so that the phenyl substituent and the &atom, 18-r-electron porphyrin core become either coplanar or closer to coplanarity.
Conclusion The results presented in this study have illustrated the application of SERS to the investigation of acetonitrile-soluble porphyrins at a silver electrode. The results indicate that vibrational spectra can be used to determine the orientation and the mode of interaction between the adsorbed porphyrin molecules and the electrode surface. The SERS spectra show that the process which is observed at -0.6 V in cyclic voltammetry is due to a change in the geometry of the adsorbed porphyrin. The result is a more tightly bound adsorbate whose molecular axis is oriented parallel to the electrode surface. The spectra also indicate that, in the absence of coadsorbed solvent, adsorption of the dry porphyrin results in a rotation of the phenyl rings into coplanarity or near coplanarity with the 16-atom, 18-*-electron porphyrin skeleton resulting in an extension of the porphyrin r-conjugation system to include the phenyl substituents. The study illustrates that S E W can be utilized to obtain information regarding adsorbate/surface interaction that would be difficult to obtain by other analytical methods. Acknowledgment. We are indebted to the PSC-BHE Research Award Program of the City University of New York (RF 662275, 668274, and 669276), the National Institutes of Health MBRS program (GM-0818), and the NSF (CHE-8711638) for financial
Hosten et al. assistance. We thank Prof. C. Guo for providing us with the porphyrin compounds and for stimulating discussions.
References and Notes (1) The Porphyrins; Dolphin, D., Ed.; Academic Press: New York, 1975. Academic (2) Hewson, W.New D.; York, Hager,1975; L. D.vel. In The 7. Porphyrins; Dolphin, D., Ed.; (3) Kadish, K. M.; Araullo, C.; Maiy, G. B.; Sazon, D.; Barbe, J. M.; Guilard. R. Inora. Chem. 1989. 28. 2528. (4) Kadish, Kr M.; Sazon, D.; Li, Y.M.; Saoiabi, A.; Ferhart, M.; Guilard, R. Inorg. Chem. 1989, 27, 1198. (5) Fleischman, M.; Hendra, P. J.; McQuillan, A. J. Chem. Phys. Lett. 1974, 26, 163. (6) Takahashi, M.; Sakai, Y.; Fujita, M.; Ito, M. Sur/: Sci. 1986,176, 351. (7) Gao, P.; Weaver, M. J . Phys. Chem. 1985, 89, 5040. (8) Hutchinson, K.; McQuillan, A. J.; Hater, R. E. Chem. Phys. Lett. 1983, 98, 27. (9) Stacy, A. M.; Van Duyne, R. P. Chem. Phys. Lett. 1983,102, 365. (10) Sanchez, L.; Spiro, T. G. J . Phys. Chem. 1985, 89, 763. (1 1) Hong, X.; Xin, Z.; Yangbo, F.; Guo, C. Chin. J . Lumin. 1988, 9,268. (12) Guo, C.; Hong, X. Unpublished results. (1 3) Birke, R. L.; Lu, T.; Lombardi, J. R. In Techniquesfor Electrodes and Electrochemical Processes; Varma, R., Selman, J. R., Eds.; Wiley and Sons: New York, 1991. (14) Xu,J.; Lombardi, J.; Birke, R. L. J . Am. Chem. Soc. 1987,109,5645. (15) Cottom, T. M.; Schultz, S.G.;Van Dyne, R. P. J. Am. Chem. Soc. 1982, 104,6528. (16) Itabashi, M.; Kato, K.; Itoh, K. Chem. Phys. Lett. 1983, 97, 528. (17) Burke, J. M.; Kincaid, J. R.; Spiro, T. G.J . Am. Chem. SOC.1978, 100,6077. (18) Itabashi, M.; Kato, K.; Itoh, K. Chem. Phys. Lett. 1983, 97, 528. (19) Lanese, J. G.;Wilson, G.S.J . Electrochem. SOC.1972, 119, 1039. (20) Kadish, C. M.; et al. Inorg. Chem. 1988, 22, 686. (21) Gao, P.; Weaver, M. J . Phys. Chem. 1985, 89, 5040. (22) Shindo, H. J . Chem. Soc., Faraday Trans. 1972, 119, 39. (23) Gouterman, M. In The Porphyrins;Dolphin, D., Ed.; Academic Ress: New York, 1975. (24) Gouterman, M. Ann. N.Y. Acad. Sci. 1973, 206, 70. (25) Silvers, S.A.; Tulinsky, A. J . Am. Chem. Soc. 1964, 86, 927. (26) Wolberg, A. J . Mol. Struct. 1974, 21, 61. (27) Eaton, S.S.;Eaton, G.R.; Holm, R. H. J . Organomet. Chem. 1972, 39, 179. (28) Fuchsman, W.; Smith, Q.R.; Stein, M. J . Am. Chem. Soc. 1977,99, 4190. (29) Gouterman, E. M. Ann. N.Y. Acad. Sci. 1973, 206, 70. (30) Meot-Ner, M.; Adler, A. D. J . Am. Chem. SOC.1972, 94, 4763. (31) Meot-Ner, M.; Adler, A. D. J . Am. Chem. SOC.1975, 97, 5107. (32) Schneider, W. Struct. Bonding 1975, 23, 123. (33) Fuchsman, W. H.; Smith, R.; Stein, M. M. J . Am. Chem. Soc. 1977, 99, 4190. (34) George, P. Chem. Rev. 1975, 75, 85. (35) Stein, P.; Ulman, A.; Spiro, T. G.J. Phys. Chem. 1984,88,369-374. (36) Spiro, T. G.;Li, X . Y. In Biological Applications ofRaman Spectroscopy; Wiley: New York, 1987; Vol. 111. (37) Cotton, T. M.; Van Duyne, M. Chem. Phys. Lett. 1984, 106, 491. (38) Takahashi, T.; Maeda, S.Chem. Phys. Lett. 1974, 24, 584.
