J. Phys. Chem. 1981,85,263-268
263
Studies on Spin-Trapped Radicals in ?-Irradiated Aqueous Solutions of L-Prollne and trans-4-Hydroxy-~-prolineby High-Performance Liquid Chromatography and ESR Spectroscopy Nobuhlro Suzukl, Kelsuke Maklno,’ Fumlo Morlya, Soujl Rokushlka, and Hlroyukl Hatano Laboratory of Radiation Chemistry, Department of Chemistry, Faculty of Sclence, Kyoto University, Sakyo-ku, Kyoto, &pan (Received: May 29, 1980; In Final Form: September 4, 1980)
The short-lived radicals generated in y-irradiated aqueous solutions of either L-proline (I) or trans-4hydroxy-L-proline(11)were investigated by spin trapping with 2-methyl-2-nitrosopropaneand subsequently by high-performanceliquid chromatography combined with ESR spectroscopy. Four ESR spectra of the spin adducts could be obtained from I and two from 11, which were separated individually. The stability differences between these spin adducts are discussed in relation to the structures of the adducts.
Introduction Absorption of energy from ionization radiation passing through water results in some reactive intermediates. The principal reactive intermediates are the hydroxyl radical (-OH) and the hydrated electron (eaq-). They react with biomolecules such as amino acids and peptides in aqueous solutions to generate short-lived free radicals. The reaction of -OHwith an amino acid, L-alanine for instance, causes H abstraction as shown by eq 1,1-3and the attack of ea< H
*OH
+
H 3 N t b c m -
k
-
7
c
H,N+--c--coo-
(1)
’ H2 on it results in a reductive deamination reaction (eq 2).H H
eq- t
I H++-C-COOI
-
H
I
H~N+-C-C
I
./o‘0-
-
CH3 H NH,
I I
f *C-COO-
(2)
CH3
A large number G. :ate constants for the reactions of .OH and e,, with various materials have been e~timated.~-lO The free radicals generated in this way are so short-lived that most of them may not be observed by conventional ESR measurements. As one of the most powerful and simplest methods for the investigation of such short-lived radicals, spin trapping has recently been introduced in various fields of freeradical chemistry.”-l6 Spin trapping is a method which ~~~~
(1) Taniguchi, H.; Fukui, K.; Ohnishi, S.; Hatano, H.; Hasegawa, H.; Maruyama, T. J. Phys. Chem. 1968, 72, 1926. (2) Neta, P.; Fessenden, R. W. J. Phys. Chem. 1971, 75, 738. (3) Neta, P.; Simic, M.; Hayon, E. J. Phys. Chem. 1970, 74, 1214. (4) Garrison, W. M. Radiat. Res. Reu. 1972,3,305. (5) Neta, P.; Fessenden, R. W. J. Phys. Chem. 1970, 74, 2263. (6) Tal, Y.; Faraggi, M. Radiat. Res. 1975, 62, 337. (7) Anbar, M.; Bambenek, M.; Ross, A. B. Natl. Stand. Ref. Data Ser., Natl. Bur. Stand. 1973, 43. (8) Dorfman, L. M.; Adams, G. E. Natl. Stand. Ref. Data Ser., Natl. Bur. Stand. 1973,46. (9) Klapper, M. H.; Faraggi, M. Q. Reu. Biophys. 1979, 12, 465. (IO) Rao, P. S.; Simic, M.; Hayon, E. J. Phys. Chem. 1975,79,1260. (11) Janzen, E. G.; Blackburn, B. J. J. Am. Chem. SOC.1969,91,4481. (12) Lagercrantz, C.; Forshult, S. Nature (London) 1968,218, 1247. (13) Evans, C. A. Aldrichim. Acta 1979, 12, 23. (14) Harbour, J. R.; Bolton, J. R. Biochem. Biophys. Res. Commun. 1975, 64, 803.
0022-3654/81/2085-0263$01 .OO/O
utilizes a diamagnetic compound (the spin trap) that reacts with a free radical giving rise to a relatively stable free radical (the spin adduct). The spin adduct is ESR observable and from its ESR parameters the parent shortlived radical can be identified. Usual spin traps are either nitroso compounds or nitr~nes.’~-~OAmong them, 2methyl-2-nitrosopropane (MNP) has been used mainly in investigations of short-lived radicals in aqueous solutions of biomolecules21-26because it is appreciably soluble in water and because its spin adduct exhibit more informative ESR spectra than those of other spin traps. The reaction of the spin trap (tBuN=O) with a short-lived radical (.R) is represented by eq 3. The spin adduct PBuN(O.)-R) tB~N=O
+ .R
-
tB~N(O.)-R
(3)
exhibits a primary 14Ntriplet and some secondary splittings, from which the parent short-lived radical can be identified. When this useful method of spin trapping with MNP is used, two essential problems have been encountered. One is that ESR spectra obtained from y-irradiated aqueous solutions of amino acids containing MNP are too complicated to analyze completely, owing to the overlap of spectra of the several spin adducts present. The other is due to the spectra of the spin adducts produced by the reaction of MNP with short-lived radicals from MNP. These two problems have been solved in previous investigations by a new method of high-performance liquid chromatography combined with ESR s p e c t r o s ~ o p y , ~ ~ - ~ ~ (15) Janzen, E. G. “Free Radicals in Biology”; Pryor, W. A., Ed.; Academic Press; London, 1980; Vol. IV, p 115. (16) McCaA’P. B.; Noguchi, T.; Fong, K.-L.;Lai, E. K.; Poyer, J. L. “Free Radicals in Biology”;Pryor, W. A., Ed.; Academic Press: London, 1980; Vol. IV, p 155. (17) Janzen, E. G. Acc. Chem. Res. 1971,4,31. (18) Terabe, S.; Kununa, K.; Konaka, R. J.Chem. SOC.,Perkin Trans.
19773.1252. (19) Lagercrantz, C. J.Phys. Chem. 1971, 75, 3466. (20) Sargent, F. P.; Gardy, E. M. J. Phys. Chem. 1976, 80,854. (21) Rustgi, S.; Joshi, A,; Riesz, P.; Friedberg, F. Int. J. Radiat. Biol. Related Stud. Phys., Chem. Med. 1977, 32, 533. (22) Rustgi, S.; Joshi, A.; Moss, H.; Riesz, P. Znt. J. Radiat. Biol. Related Stud. Phys., Chem. Med. 1977, 31, 415. (23) Riesz, P.; Rustgi, S. Radiat. Phys. Chem. 1979,13, 21. (24) Taniguchi, H.; Hatano, H. Chem. Lett. 1974,531. (25) Taniguchi, H.; Hatano, H. Chem. Lett. 1975, 9. (26) Lagercrantz, C. J. Am. Chem. SOC.1973,95, 220. (27) Makino, K. J. Phys. Chern. 1979,83,2520. Makino, K.; Hatano, H. Chem. Lett. 1979, 119. (28) Makino, K. J. Phys. Chem. 1980,84,1012. Makino, K.; Suzuki, N.; Moriya, F.; Rokushika, S.; Hatano, H. Chem. Lett. 1979, 676. 2 -
I
0 1981 American Chemical Society
264
The Journal of Physical Chemistty, Vol. 85, No. 4 1981
which we have proposed. In the investigations of L-valine, L-isoleucine, and L-leucine, the method has proved to be powerful for the separation and subsequent ESR identification of various kinds of spin adducts. In this study, the method is applied to y-irradiated aqueous solutions of either L-proline or trans-ChydroxyL-proline. L-Proline and its derivatives differ from other amino acids by having imino groups and relatively rigid rings (pyrrolidine rings). Because prolines play highly specific roles in giving characteristic structures to proteins, great interest exists in the y radiolysis of proteins in order to understand the radiation effects on the structures of proteins. Concerning the investigations of the y radiolysis of aqueous solutions of proteins, a few papers have appeared in which short-lived radicals produced from amino acids were investigated by spin trapping with MNP.21p22 However, the spin adducts were not separated nor was the computer simulation analysis of the spectra undertaken. Therefore, this investigation was undertaken in order to identify the short-lived radicals produced in the y-irradiated aqueous solutions of L-proline and ita derivatives trans-4-hydroxy-~-prolinein more detail. In this study, four spin adducts from L-proline and two from trans-4-hydroxy-~-prolinecould be separated and identified. Experimental Section MNP used in this study was synthesized as previously d e ~ c r i b e d .L-Proline ~~ and trans-4-hydroxy-~prolinewere purchased from Nakarai Chemicals, Kyoto, and used without further purification. MNP (50 mg) was added to 10 cm3 of water without degassing, and the heterogeneous solutions were stirred for 1.5 h at 45 "C until they became homogeneous. Either (2 mmol) was added L-proline or trans-4-hydroxy-~-proline to 10 cm3 of the prepared aqueous MNP solution. The sample solutions were irradiated at ice temperature with 6oCoy rays at a dose rate of 6.0 X lo6 rd/h to a total dose of ca. 3.0 X lo5 rd. Quantities (1.0 cm3)of the irradiated solutions were injected into the six-port sample injector of a high-performance liquid chromatograph (HLC-803 of Toyo Soda) equipped with an ESR spectrometer (JEOL, Model PE-3X) operated at 100-kHz modulation in the X-band. A chromatographic column (IEX-21OSC of Toyo-Soda 3/s in. i.d. and 60 cm long) was attached to the chromatograph. A quartz flow cell (the practical part being 0.7 mm i.d. and 5.0 cm long), which was fixed in the ESR cavity, was connected to the exit of the column with ca. 0.25-mm i.d. Teflon tubing. Chromatographic conditions were as follows: pressure, ca. 70 kg/cm2; flow rate, 0.2 mL/min; temperature, ca. 25 "C (air-conditioned room temperature). For separation of the adducts from both prolines, gradient elution was adopted. The column was equilibrated with the new eluant for at least 1.0 h, and after the gradient elutions the gradients were reversed and the column was then equilibrated with the intial mobile phase. In this study, a 90-min gradient (X4mode, Gradient Device GE-2 of Toyo-Soda) was used: 0.2 M NaH2P04-Na2HP04 buffer (pH 6.8) to 0.2 M NaOH-Na2HP04 buffer (pH 11.5). The gradient was started at the same time as the injection of the sample solution. Before ESR measurements, the separated fractions were deoxygenated by bubbling nitrogen gas for 1min through the solution for the purpose of decreasing the line broad(29) Makino, K.J.Phys. Chem. 1980,84, 1016. Makino, K.;Suzuki, N.; Moriya, F.; Rokushika, S.;Hatano, H. Anal. Lett. 1980, 13, 301. (30) Makino, K.J. Phys. Chem. 1980,84, 1968. (31) Stowell, J. J. Org. Chem. 1971, 36, 3055.
Suzuki et ai.
'
100'
4-4
10 G
Flgure 1. ESR spectra of y-inadiated aqueous solutions of (a) L-proiine and (b) trans-4-hydroxy-~-proiinecontaining MNP.
ening due to molecular oxygen. Some ESR spectra were measured in D20. For the preparation of D20 solutions of the adducts, the collected fractions were lyophilized and then D20 was added (99.75%, Merck). The microwave power was 15 dB (6 mW). The magnetic field was fixed at the positions indicated by the vertical arrows in parts a and b of Figure 1 and the magnetic-field modulation was applied at a high amplitude of 10 G to cover a wide range during the chromatographic separation. The hyperfine coupling constanta (hfc constants) from the ESR spectra obtained in this way were measured with Mn2+in MgO as a reference. All experiments were carried out in the dark. Results and Discussion Figure l a shows a typical ESR spectrum obtained from aqueous L-proline solutions containing MNP just after y irradiation. The spectrum is complicated by the overlapping of several signals from the spin adducts which are produced by reaction of MNP with short-lived radicals from both L-proline and MNP. Chromatographic separation of the same solution gave a chromatogram as shown in Figure 2a. Six peaks labeled A-F appeared there. No analyzable ESR signals could be obtained in the fraction from peak A. ESR spectra obtained from the fractions around peak D consisted of more than two spectra, so satisfactory identification could not be made in this study. Parts a 4 of Figure 3 show ESR spectra obtained from the fractions at peaks B, C, E, and F, respectively. The adducts produced by the self-trapping of MNP (except for the adduct from tert-butylnitrosohydroxylamine32)are found to be eluted so slowly that their peaks are not seen in the chromatogram. In the spectrum of Figure 3a obtained from the fraction giving peak B, each of the primary triplet due to 14Nof the nitroxide group splits into 2 X 3 secondary lines due (32) Makino, K.;Suzuki, N.; Moriya, F.; Rokushika, S.;Hatano, H. Anal. Lett. 1980, 13, 311. Radiat. Res. In press.
L-Proline Spin-Trapped Adducts
I 0
The Journal of
Physical Chemistty, Vol. 85, No. 3,
1981 265
. 100
50
150
Elution Time ( rnin
)
CI
100
sDectra obtained from the fractions giving peaks (a) (c)E, and (d) F.
Flaure 3. ESR
(b)
B,(b) C,
0
50
100
Elution Time
150 (
min I
Chromatogram of y-irradiated aqueous solutions of (a) proline and (b) frans-4-hydroxy-~-proline containing MNP. Flgure 2.
to a 0hydrogen and two equivalent y hydrogens. The hfc H 5.0 constants of the adduct at pH 7 are UN = 15.6 G, U ~ = G, and uyH = 0.5 G. In a basic solution (pH ll),the values are almost unchanged, U N = 15.7 G, U ~ H= 4.9 G, and a7y = 0.5 G, implying that the amino group of the adduct is distant from the nitroxide center. The spin adduct is found to be very stable in a basic solution whereas other spin adducts found here which maintain rigid pyrrolidine rings are relatively unstable, suggesting that the adduct might not have a rigid ring. Consequently, the spectrum can be assigned to adduct I. Since the spin adduct, which
Flgure 4. MI = 0 components of
I
COOT = +B"-N-
the ESR spectra from the fraction
giving peak C measured (a)at pH 7 in H20, (b) at pH 11 in H20, (c) at pH 7 in D20,and (d) at pH 11 in D20.
I
0.
I is produced by the deamination reaction due to the attack of eaq'-on an amino acid, has been known to be excluded when it is chromatographed using a cation-exchange column with neutral or basic eluants27~2g~30 and since the adduct found here is not excluded under these conditions, it seems possible that the NH3+group is not detached from the molecule, supporting the assignment given in I. The adduct is produced by the reaction of MNP with a short-lived radical generated by the attack of e, - which opens the pyrrolidine ring. The adduct has been &served previously.21 The spectra from the fraction giving peak C were measured at both neutral and alkaline pHs in both H 2 0 and DzO, as shown in Figure 4. Parts a-d of Figure 4 show
the MI= 0 component of each spectrum. A t pH 7, the spectrum consists of 13 lines in HzO whereas there are 12 lines in DzO. On the other hand, at pH 11, no difference between the spectra in HzO and D,O was observed. The spectral change depending on the solvents indicates that the spectrum at pH 7 in HzO involves extra hyperfine couplings due to at least one of the two exchangeable hydrogens of the imino group, implying that the imino group lies near the nitroxide center. Consequently, the spectra can be assigned to spin adducts I1 and 111. The
600-
boo-
(in H,O) I1
(in D,O) 111
266
The Journal of Physlcal Chemistty, Vol. 85, No. 3, 1981
Suzukl et al.
Figure 7. ESR spectra from the fractlon giving peak B’ measured In H,O at (a) pH 7 and (b) pH 11.
1
5G
I
Flgure 5. MI = 0 components of ESR spectra from the fraction giving peak E measured (a) at pH 7 in H,O, (b) at pH 7 in D,O, (c) at pH 11 in H20, and (d) pH 11 in D20.
100
50
Flgure 6. ESR spectra from the fraction giving peak F measured at pH 7 In (a) H20 and (b) D20, and thelr MI = 0 components.
adduct is unstable above pH 10. The spectra obtained from the fractions at peaks E and F are shown in Figures 5 and 6, respectively. When they were measured in DzO,no remarkable differences in both spectra were seen. This fact indicates that imino groups in both molecules lie relatively distant from the nitroxide centers, which are located on either C (3) or C(4) as described in IV and V. In alkaline solutions, the adduct in
oxy1 and carboxyl groups, can distort the pyrrolidine ring under some conditions, it seems probable that the carboxyl group of adduct F is located closer to the nitroxide center than that of adduct E. From this consideration, tentatively the spectrum obtained from the fraction at peak E is assigned to adduct IV and that from the fraction at peak F is assigned to adduct V. Adduct V, in which the carboxyl group is adjacent to the nitroxide center, is expected to be less stable than adduct IV. As described above, four individual spectra could be obtained from the separated fractions of y-irradiated aqueous L-proline solutions containing MNP. Analysis of each spectrum revealed the production of radicals caused by H abstraction from C(3) and C(4) of L-proline. These radicals have not been detected in the previous investigations nor have the spectra of the adducts been observed. It is also found that the attack of e,; on L-proline results in opening of the pyrrolidine ring, which is of interest in considering radiation effects on the structural change in proteins, because it is well-known that rigid pyrrolidine rings prevent the formation of an a-helix in some proteins. In order to study the stability of spin adducts from L-proline in more detail, we investigated aqueous solutions of trans-4-hydroxy-~-proline by the same method. This imino acid contains the hyroxyl group in addition to the pyrrolidine ring. The ESR spectrum of the y-irradiated sample solution is shown in Figure lb, and the chromatogram obtained from the same sample solution is in Figure 2b. Four peaks labeled A’-D’ appeared there. The fraction at peak A’ did not give any signals, and the fraction at peak C’ did not contain a sufficient amount of the spin adduct for analysis. Figure 7 shows the spectra obtained from the fraction at peak B’. The spectra consisting of the triple double triplets can be assigned to spin adduct VI. The adduct OH
‘\/A IV
‘&(
V
fraction E (adduct E) is found to be relatively stable in comparison with that in fraction F (adduct F). The spectrum of adduct F decays immediately in an alkaline solution. Since bulky groups, for instance tert-butylnitr-
H
”
H HNH;
YT coo-
VI
is produced by the reaction of MNP with the short-lived
The Journal of Physical Chemistry, Vol. 85, No. 3, 1981 207
L-Proline Spin-Trapped Adducts
TABLE I: Hfc Constants for the Individual Spin Adducts from L -Proline and trans-4-Hydroxy.~-proline spin adduct peak solvent pH hfc constants, G L-Proline B 7 U N = 15.6, U ~ = H,O H 5.0, U ~ = H 0.5 H,O 11 U N = 15.7, U ~ = H 4.9, U ~ H =0.5 B C H,O 7 a~=16.0 H,O 11 a ~ = 1 6 . 2 C D,O 7 ~ ~ = 1 5 . 9 C C D,O 11 U N = 16.2 E H,O 7 ~ ~ ~ 1 6 . 0 H,O 11 a ~ = 1 6 . 3 E D,O 7 a~=16.1 E D,O 11 ~ ~ = 1 6 . 4 E F H,O 7 a~=16.0 D,O 7 a~=15.9 F trans-4-Hydroxy-L -proline B‘ H,O 7 U N = 15.7, a p =~ 4.9 B‘ H,O 11 ON = 15.7, U ~ H= 4.6 D H,O 7 ~ ~ = 1 5 . 4
D
Flgure 8.
“ 10G ESR spectra from the fraction givlng peak D‘ measured at
pH 7 in (a) H20 and (b) D20.
radical due to the opening of the pyrrolidine ring caused by the attack of ea;. Figure 8 shows the spectra obtained from the fraction a t peak D’. The lines were sharpened in D20 compared with those in H20,implying that the imino group lies close to the nitroxide center. Consequently, the spectra can be assigned to adducts VI1 and VIII.
TC---c i
i
coo(in H,O)
Loo(in D,O)
VI1 VI11 The adducts produced by H abstraction from C(3) was not found here. The adducts formed by H abstraction from C(4) decayed too fast to be observed by conventional measurement. These facts might reveal structural characteristics related to the stability of the adducts. Since trans-4-hydroxy- proli line has an extra hydroxyl group compared with L-proline, it seems probable that the spin adducts from the imino acid having the nitroxide center on C(3) or C(4) might not be formed readily because of steric hindrance from the neighboring bulky groups, or that, even if they are formed, they might not be stable enough to be observed owing to such hindrance. These facts and considerations support the assignment made from the spectra of the adducts having the nitroxide center on the C(3) and C(4) positions of L-proline. Some of the ESR spectra of the adducts from both imino acids were found to be complicated. One of the reasons for this is considered below. As shown in structure I, free rotation around the C(2) and C(3) bond will give two = uyH.J, from equivalent y hydrogens in the adduct which a relatively simple spectrum will be obtained. On the other hand, y hydrogens in adduct I1 might be nonequivalent owihg to the prevention of free rotation in the rigid pyrrolidine ring, giving different hfc constants for the and making the spectrum two y hydrogens (aysl# complex as shown in Figure 4a. In addition, when one hydrogen is abstracted from one of the three methylene
D,O
7
UN=
15.5
groups of L-proline, the resulting spin adducts have additional chiral centers, for example, the adducts in reaction H
\/” 1-7
H k 6 H H i H z 2 H\C MNP H/
H/
\p/ cooI
H
\*
T
7
/‘-6 \Cy“z+ TH H
LOO-
/” P\
\* H\
or H,C\p”z+
H H H
(4)
COO-
4 which have chiral centers indicated by asterisks. Three sets of diastereomeric spin adducts might be produced by such H abstraction from each carbon, C(3), C(4), and C(5). The coexistence of such diastereomeric adducts is probably another reason for the complication of the spectra, because each member of a set of diastereomeric adduds might have slightly different hfc constants. These interpretations are also supported by the observation described below. Comparing the spectra shown in Figure 3a with those in Figure 7, we found that the notable distortions of the MI= +1 components appear in Figure 7,which is indicated by small arrows, and that broadening of the lines appears more remarkably in the M I= -1 components in Figure 7 than those in Figure 3a. Since the adducts giving the spectra in Figure 7 can have the two chiral carbons as shown in structure VI with asterisks, the distortion of the spectra can be elucidated by the difference in the hfc constants of the diastereomericadducts. For diastereomeric adducts having rigid pyrrolidine rings, the difference between their hfc constants might be more remarkable. Consequently, it seems very difficult to obtain hfc constants for all nuclei of the separated spin adducts by computer simulation analysis, becuase some spectra found here consist of several spectra due to isomers having slightly different hfc constants. Therefore, only the hfc constants due to 14Nnuclei of the nitroxide groups have been obtained for some spin adducts. The values are listed in Table I. In order to support the assignments made here more definitely, we will perform similar experiments with chloro or bromo derivatives of L-proline. Useful information about the coexistence of diastereomeric adducts may be obtained from the experiments, because the reaction of e, with these compounds results in the specific production of adducts having nitroxide groups on the halogenated carbons. Some stereochemical information may also be
268
J. Phys. Chem. 1981, 85, 268-272
obtained from the investigation of &-4-hydroxy-~-proline by comparing the data with those from trans-4-hydroxyL-proline. Results of these investigations will appear in the future.
Acknowledgment. The authors thank to the Toyo Soda Co. Ltd. for supplying the high-performance liquid chromatograph, HLC-803, and the cation exchanging column, IEX-21OSC.
Photodeposition of Palladium and Platinum onto Titanium Dioxide Single Crystals Hlroshl Yoneyama,' Naritoshl Nlshimura, and Hldeo Tamura Department of Applied Chemisfty, Faculty of Engineerlng, Osaka University, Yamadakami, Suita, Osaka 565, Japan (Received: July 3, 1980; In Final Form: September 4, 1980)
The photosynthetic deposition of palladium and platinum from their chloride solutions onto TiOzsingle crystals of 1-mm thickness was studied. When the front face of the crystals was illuminated and the back was kept in the dark in contact with the plating bath, the deposition took place preferentially on the dark face except for the case of a nonreduced crystal in which the deposition occurred on both faces with comparable rates. However, if the back face was covered with an insulating wax to block any electrical contact to the bath, the deposition occurred on the illuminated front face with a rate comparable to that obtained when the back was in contact with the plating bath. The rate of photodeposition was markedly affected by the doping densities, and the highest rate was achieved at a critical value at which the maximum quantum yield for the photosensitized oxidation of water is obtained.
Introduction It has been revealed from electrochemical studies on heterogeneous reactions at semiconductor photocatalysts that the working mechanism of the photocatalyst has a close resemblance to that of photoelectrochemical cells in which a semiconductor is used as a photoelectrode.lI2 For example, a heterogeneous reaction on an illuminated ntype semiconductor catalyst is composed of a photosensitized oxidation process and an electrochemicalreduction process which acts as a counter part of the former proIllumination is required, without doubt, for the former process to occur but not so for the latter, leading to the idea that surface sites of photocatalysts which work effectively for the former process are illuminated faces of the catalyst, while those for the latter process are nonilluminated faces. The validity of such an idea was demonstrated in photocatalytic and photosynthetic deposition of copper on Ti02 and W 0 3 powders.6 If the coexistence of illuminated and nonilluminated faces in the catalyst surface is essential for any heterogeneous reaction to occur, no reaction should take place under illumination of the entire surface of the catalyst, because there is no room for any electrochemical reduction to occur in this case. However, it has already been noticed in several heterogeneous reactions on illuminated Ti02 single crystal catalysts such as the oxidation of cyanide? hydroquinon,8 and methanolg that these reactions occur (1)M. Miyake, H.Yoneyama, and H. Tamura, Bull. Chem. Soc. Jpn., 50,1492 (1977). (2)A.J. Bard, J. Photochem., 10,59(1979);Science, 207,139(1980), and references cited therein. (3)H.Yoneyama, Y. Toyoguchi, and H. Tamura, J. Phys. Chem., 76, 3460 (1972). (4)F. Mollers, H.J. Tolle, and R. Memming, J. Electrochem. Soc., 121,1160 (1974). (5)M. S.Wrighton, P. T. Wolczanski, and A. B. Ellis, J. Solid State Chem., 22, 17 (1977). (6)H. Reiche, W. W. Dunn, and A. J. Bard, J.Phys. Chem., 83,2248 (1979). (7)K. Kogo, H.Yoneyama, and H. Tamura, J.Phys. Chem., 90,1705 (1980). 0022-3654/81/2085-0268$01.OO/O
even if the entire surface of the catalyst is illuminated. Therefore, it seems meaningful to elucidate the roles of the illuminated and nonilluminated faces of the catalyst surface for heterogeneous reactions. This is one purpose of the present study. Until now, little attention has been paid to effects of the doping density of photocatalysts on the rate of heterogeneous reactions. However, it has already been shown theoreticallylOJ1 and experimentally12 that the doping density has a marked effect on the quantum yield of photoelectrochemical reactions. Considering that there are close similarities between heterogeneous reactions on semiconductor photocatalysts and photoelectrochemical cell reactions, the doping level of the photocatalysts must have an important role in determining the rate of heterogeneous reactions. This is another purpose of the present study. The reaction systems chosen here are the photosynthetic deposition of palladium4 and platinum13 onto Ti02, in which the metal deposition process is accompanied by the photoassisted oxidation of water; electrons in the conduction band participate in the deposition process, and positive holes in the valence band are consumed to oxidize water to oxygen. Experimental Techniques Ti02 single crystals of 1-mm thickness cut perpendicularly to the c axis were used in the present study. Both end faces were successibly polished with emery papers, 0.3and 0.05-pm alumina. Afterwards, the crystals were washed in an ultrasonic bath, followed by immersion in (8) M. Miyake, H. Yoneyama, and H. Tamura, Electrochim. Acta, 22, 319 (1977). (9)M. Miyake, H.Yoneyama, and H. Tamura, J. Catal., 68,22(1979). (10)H.Gerischer, J. Electroanal. Chem., 58, 263 (1975). (11)R. H.Wilson, J. Appl. Phys., 48,4292 (1977). (12)H.Tamura, H.Yoneyama, C. Iwakura, H. Sakamoto, and S. Murakami, J.Electroanal. Chem., EO, 357 (1977). (13)B. Kraeutler and A. J. Bard, J. Am. Chem. Soc., 100,4317(1978).
0 1981 American Chemical Society
