2456
J. N.PITTS, JR.,H. W. JOHNSOK, Jit., A N D T. Iiuwa~.4
Vol. 66
STRUCTURAL EFFECTS IK THE PHOTOCHEPIIICAL PROCESSES OF KETONES I N SOLUTIOS BY J. K. PITTS,JH.,H.
w.JOHNSON, JR.,
AND
T. KUWANA]
University of California, Riuerside, California Receized M a y P6, 1962
The behavior of a series of hydroxy, methoxy, amino, and chlorobenzophenones in the photopinacolization reaction has been studied using product isolation, ultraviolet spectral changes, elertrochemical methods, phosphorescence and fluorescence twhniques, and electron spin resonanre spectroscopy. The results and cwrrelations derived from these and published studies :ire discussed. Electron spin resonance evidence for long-lived triplets of wrt,ain para-substituted benzophenones is presented and correlated with their unusually long phosphorescent lifetimes and their unusually weak H atom alistrarting power in the classic photopinacolization reaction. These phenomena are tentatively explained on the basis that in these ketones the lowest lying triplet state is R,A* in nature rather than n,T* as is the case for benzophenone. A preliminary report on a similar study in the anthraquinone system is included,
Introduction In this scheme, So indicates a ground state singlet, The photochemical reduction of ketones to S1the first excited singlet, and T1the lowest excited pinacols or benzohydrols was first reported by triplet state. Species I has been called the ketyl Ciamician and Silber.2 I n early work the reaction radical.12j14 was studied to determine its synthetic usefulness The transfer of energy between photoexcited and scope. Certain diary1 ketones, aryl alkyl benzophenone triplets and naphthalene in the solid ketones, and dialkyl ketones3 have been shown state was first noted by Terenin and Ermolaev.I6 to give pinacols in varying yield. I n certain cases Energy transfer in solution between photoexcited the reaction proceeds in the reverse d i r e ~ t i o n . ~benzophenone and biacetyl (and other molecules) The hydrogen source in the pinacol reaction usually was first reported by Backstrom and Sandros.16 has been an alcohol, but other sources have been Hammond, Turro, and Fischer13 utilized such employed.5-8 The reactions of photoactivated solut,ion phase energy transfer in an ingenious carbonyl compounds with other substrates have synthetic method while Moore and Ket'chum debeen reviewed.9910 termined the efficiency of the t'riplet-triplet transfer There now seems to be substantial agreement behween benzophenone and naphthalene in benthat the first steps in the photopinacol reaction zene solution.17 sequence Porter and Wilkinson14 studied the flash photolysis of benzophenone in alcohol both in the presence and in the absence of naphthalene. They confirmed the presence of a short-lived triplet of beczophenone which apparently absorbed a t the same wave 1engt.h as the ketyl radical (I), and showed that the triplet apparently is the chemically active species in the photopinacol reaction, since quenching of the triplet with naphthalene inhibited the pinacolization. The lifetime of the (1) Presented a t the Symposium on Reversible Photochemical triplet measured by Porter and Wilkinsonl* was Processes, Duke University, Durham. North Carolina, April,, 1962. in good agreement with that of Backstrom and (2) G. Ci:imicinn ami 1'. Silber, B e r . , 33, 2911 11900); 34, 1541 (1 901). Sandros.16 Hammond, et aZ.,13 arrived a t the (3) C. Weizmann, E'. Bergmann, and Y. Hirshberg, J . A m . Chem. same conclusion by a study of the reaction of Soc., 60, 1530 (1938): 1;. Bergmann and T.Hirshberg, ibid., 66, 1429 benzophenone with benzohydrol. Schenck,18 and i1943). (4) A. Schonberg and A. hfustafa, J. Chem. SOC.,67 (1944). Pit,ts, Leteinger, et aZ.,lz independenOly found (5) G A. Hammond, W. P. Baker, and W. R . Moore, J . A m . Chem. evidence for another species in the reaction chain Soc., 83, 2793 (1961). and proposed possible structures. The work of (6) A. Mustafa, A'nture, 162, 866 (1918). (7) K. Primmerer. H. Hahn, F. Johne, and H. Kehlen, Ber., 6SB, Franzen'g should be mentioned here, as well as the 867 (1942). observations of Bachmann on the color developed (8) E. .I. Bow-en and E. L. .2. E. de la Praudiere, J . Chrm. S o r . , 1603 (1934). during the reduction of benzophenones t'o benxo19) A . Srtionberp and A. Mustafa. C h r m . Rev., 40, 181 (1947). hydrols in alkaline solution.20 (IO) C . I{. Masson, V. Boekelheide, and W. A , Noyes, Jr., in -1. Weissberger, E d . , "Techniques of Organic Chemistry," Vol. 11, 2nd Ed., Interscience Pnblishers. Inc., New York, N. Y., 1956, p. 257 ti. i l l ) H. L. J. Biickstrom, 2. physik. Chem., 26B, 99 (1934). (12) J. N . Pitts, Jr., R. L. Letsinger, R. P. Taylor, J. >I. Patterson, G. Reektenwild, and R. B. Martin, J . A m . Chem. S O C . ,81, 1068 (1959). 113) (a) 0. Y. Hammond, and IY. N. Moore, ibid., 81, 6334 (1959); (b) \Y. R . Moore, 0. S. Hammond. and R. P. Foss, ibid., 83, 2789 (1961); ( e ) G. 8. Hammond, W.P. Baker, and W. Jt. Moore, ibid., 83, 2795 (1961); (d) G. T.. Hammond, N. J. Tiirro, and A. Fischer, ibid., 83, 4674 fl961). (14) C I . Porter and F. Wilkinson, Trans. Faradar] S o c . , 67, 1686 (1961).
(15) A . hT.Terenin and V. I. Ermolaev, Trans. Fnradau Sor., 62, 1042 (19c56). 116) H. L. J. Biickstrom and K. Sandros, Acta Chem. Scand., 14, 48 (1960): ibid.. 12, 823 (1958). (17) IT. 31. Moore and .\I. Ketchuni. .I. Am. Chem. SOC.,84, 1368 (1962). (18) Q. 0. Bchenck, W. Meder, and &I. Pape, Proc. I n t e r n . Conf. PeacGjuZ Uses A t . Eneryy, Geneva, 29, 352 (1958); Plenary Lecture, 5th International Symposium on t'ree Radicals. I'ppsala, Sweden. 1961.
(19) V. Franzen, Ann.,633, 1 11960). (20) W.E, Bnchmann, J . A m . Chem. Soc., 66, 391 (1933).
Der., l!W
P H O T O C H E M I C A L P R O C E S S E S OF I ‘ h T O U E S I V SOLUTIOV
2457
TABLE I PHOTOCHEMICAL BEHAVIOR OF SUBSTITUTED BENZOPHENONES Compciund
Pinacol formed
Benzophenone ----Derivai,ives-ortho
OH OH OH XH? NHCHj (:O?H
Plioaphorescence mean lifetime, aec.
Remarks
0.05-0.08
(CeH&COH. e.s.r. obs.nJ38
n.0.n n.o.
No triplet, or radical e.8.r. signnl cletmt,ed
Yes2 Other groups
p-OH P-OCHI p-OCHs; p’-CH:,
N O X O
NO S O
Yes Yesd
meta
KH2
TJniisual 1)ehavior”
para
CsH,
Yes No
0 . 3 2 i 0.01 Triplet e.5.r. obs. .40, 0.3c Triplet e.s.r. obs. N(CH3). p’-N(CHa)Z NO20 .27 f 0.02 Triplet e.8.r. obs. hT(CHj)? Yes .35 f 0.03 Triplet e.8.r. obs. OCH3 p’-OCHa Yes 1 3 m ,065 =t0.007 KOtriplet e.s.r. n.0. = not observed, either not present or too short to measure. Phosphorescence lifetimes of less than 0.01 sec. could not be detected with theoequipment used. * No change noted using Pyrex cell; in quartz cell with unfiltered mercury arc the absorbance at ‘2450A. (major peak) disappeared completely. c Mean lifetime of 0.3 sec. determined in rigid 2-proponal n t 77°K.: in rigid EPA mean lifetime is of same order of magnitu$e; in benzene at 77°K. mean lifetime was about 0.09 see. Phosphorescence maximum occurred at approximately 4800 4. d Forms phthalide dimer. ”2
Effect of Substituents : Pertinent Literature l f u c h qualitative information has been collected concerning the effect of substitution on the photoreduction of benzophenones, and it seems appropriate t o summarize the litmeratwereports found for certain simple benzophenones. Alkylated Benzophenones.-Pinacols mere obtained from m- and p-methylbenzophenones and p,p-dimethylbenzophenone.21 o-Methylbenzophenone gave rearrangement to III.2? ,C6H5
t i ~ e l y . S~o~ other reports of successful photopinacolization were found. Other Derivatives of Simple Benzophenones.p-Methoxy and p,p‘-dimethoxybenzophenonesgave photopinacols13~21;p-cyanobenzophenone photoreduced13; o-benzoylbenzoic acid photoredu~ed~4,~b; m-phenylbenzophenone photoreducedz6; o-phenylbenzophenone failed to photoreduce. Conflicting reports were found with p-phenylbenzophenone3~20~21; in the present work it was found to photoreduce slowly. Both monoacetyl naphthalenes and 1-benzoylnaphthalene failed to photoreduce in the presence of a l c ~ h o l s . ~ P
a;;*oH I11
Halogenated Benzophenones.-These halogenated benzophenones gave photopinacols : o-, m-, and p - chlorobenzophenones; o,o:,p?p’ - tetrachlorobenzophenone; o,m and p,p -difluorobenzophenones.21 Nitrobenzophen0nes.-No reports of successful photopinacolization were found. Aminobenzophenones.-No reports of successful photopinacolizat’ion found. p,p’-Bis-(dimethylamino)-benzophenone has been reported t’o be unreactive. 2o Hydroxybenzophenones.-o -Hydroxy - p - methoxybenzophenone does not phot’opinacolize, mhereas the opdimethoxy derivative does. 2 3 0-Hydroxy and o,p-dihydroxybenzophenone produced very low yields of pinacol with quantum yields of 7 X and 2 X mole/einstein, respec121) B. Boeseken and
E. Cohen, d k a d . A m s t e r d a m VersZ., 23, 778
(1920). ( 2 2 ) N. C. Y a n g a n d C . Rivas. J. A n . Chem. S O C .83, , 2218 (1961). (23) J. N. l’itts, Jr., and R . Martin, Abstract, Report t o Petroleum Rese‘hrcIi Funil. 278-R, 19.59.
Results and Discussion
Benzophenones.-Tables I and I1 present data concerning the formation of isolated products, measurement of lifetimes of any observed phosphorescence, detection of intermediates by electron spin resonance (e.s.r.) techniques, and photopotential measurements in the photochemical reduction of some substituted benzophenones. Table I11 reports preliminary approximate quantum yields in the photopinacol reactions of some substituted benzophenones. Changes in photopotential appear to correlate well with “go-no-go” photopinacolization properties of substituted benzophenones and anthraquinones. This is reasonable since the change in photopotential reflects the presence of a reduced species (with respect to the parent ketone), e.g., free radicals, as pointed out by Surash and Hercules, who first obtained reliable photopotentials in (24) A. Beckett and G . Potter, Sheffield University, private communication. (25) D. B. Lemaye, J. Uniz;. B o m b a y , 1, part 2 , 52 (1932): Chem. Abstr., 27, 2097 (1932). (26) H. H. H a t t , A . Pilgrim and E. F. 111. Rt~plienson,.I. Clirm. Soc., -178 11041).
J. N. PITTS,JR., H. W. JOHNSON, JR., AND T.KTJWANA
2458
1701.
66
TABLEI1 theoretical aspects of the electrochemical and PHOTOPOTENTIAL MEASUREMENTS OF SOME BENZOPHE-photochemical processes in the irradiated solutions. NONES
Pinacol formed
Compound
Yes
+
0.31 0.12
P-OH No p-OCH, p’-CH* NO
-
-
-
-
Benzophenone -Derivativesortho
OH OH
Photopotential.a volts MeOH EtOH IPA
Other
No
”2
”CHI
-
-
-
No
para
No NHz No0 N(CHa)2 p‘-N(CHa)z Yes 0.05 N(CH3)x c1 p’-Cl Yes20 0.10 OCHI p’-OCHs Yes*1918 0.10 CHI p’-CHs Yesa,zl Br Yesz* 0 (+) indicates a change in the photopotential after 5min. irradiation greater than 10 mv., ( - ) indicates a change of 10 mv. or less. Numbers in the table give the magnitude of the change in volts. Without exception a negative photopotential ( P t electrode vis. 8.c.e.) was obtained when a photochemical reaction occurred.
+” +”
+ +
-
+ +
TABLE 111 APPROXIMATE QUANTUM YIELDSFOR PINACOL REDUCTION OF SUBSTITUTED BENZOPHENONES~ Substituent
P-Cb%
Concentration of ketone
0.02 21f 0.02 M EtOH 10-3 111 lo-* M
0
0
acetoneb
ketone’
The prime purpose of this investigation was to check (by a variety of techniques) the effect of substituent groups and location of substitution on the photoreduction of a series of benzophenone derivatives. The following facts are evident from Table I. ortho-Substitution by certain functional groups has a pronounced effect on the “go-no-go” photoproperties of benzophenone derivakives. It has been reported that o-hydroxybenzophenone did not fluoresce and was a “stabilizer” for the prevention of undesirable aging processes caused by ultraviolet light.30 More recently it was shown in these Laboratories that while the o-hydroxy-p-methoxybenzophenone did not photopinacolize, o,pdimethoxybenzophenone went readily.24 Subsequently, Yang22 demonstrated that an intramolecular H atom transfer, “photoenolization,” occurred in irradiation of o-methylbenzophenone. A similar photoenolization process may be responsible, a t least in part, for the lack of reactivity found with o-amino and o-methylaminobenzophenones. Photoenolization is facilitated by the six membered ring available through internal hydrogen bonding, i e . C,H-, I
0.07 0.2 0.07 0.2 No obs. change 0.9 2.1
Evidently, replacement of amino, hydroxy, or methyl by methoxy or carboxy restores the intermolecular hydrogen atom abstracting power of the benzophenone. These groups cannot form a (CH*)~N 10-4 M d 3 x 10-3 stable photoenol of structure similar to I11 and m-NOs M 0.070 f intermolecular abstraction from the solvent beData of Dr. Graham Black. Quantum yields measured comes the predominant course of the reaction. using Cary Model 14 spectrophotometer to follow benzoWhile this explanatioii originally seemed reasonphenone concentrations, vapor phase chromatography on calibrated columns to follow acetone concentrations. Used able, there is evidence that effects other than steric Xe/Hg arc operated a t 460 w., filtered to allow only 3660 b. (or intramolecular hydrogen bonding or photoenolradiation. Light intensity varied from 3.54 to 2.8 X 101’ ization) can be important also. The compounds, quanta/min. from run to run. b Quantum yield for acetone p-aminobenzophenone and p,p’-bis-(dimethylformation. Quantum yield for benzophenone disappearance. d Acetone appeared to be formed, but sensitivity of amino)-benzophenone, have been studied. These instrument did not permit quantitative measurement. derivatives do not photopinacolize (noted earlier20 * Very approximate value. No apparent change in ketone for the latter compound); neither do they show concentration, but acetone did appear to be formed. spectral or photopotential changes. However, systems of this type. 27 Subsequent photopolaro- p-dimethylaminobenzophenone gives a small yield graphic28 and photochronopotentiometriczgstudies of the photopinacol and a slight change in the photohave verified the significance of photopotential potential. The p-aminobenzophenone has been studied the changes in such ketone systems. I n this paper the most thoroughly and, while certain of the results empirical changes in the photopotentials will be portions of the data simply noted without further speculation on the will be reported will be summarized here. (27) J. J. Surash and D. M. Hercules, Abstracts of 138th National Ten and twenty-four-hr. irradiations a t room Meeting of American Chemical Society, New York, N . Y.. September, temperature of degassed 10-2 M solutions of p1960; J. J. Surash, ”Studies on Photo-Induced Luminescence and aminobenzophenone in isopropyl alcohol using a Electrode Potentials,” Ph.D. Thesis, Lehigh University. Bethlehem, medium intensity mercury lamp (ca. 5 X Pa., 1960. (28) H. Berg, 2. Elecktrochm., 64, 1104 (1960); Nalurwiss., 47, einstein/sec.) produced no significant permanent 320 (1960); H. Berg, Collection Czech. Chem. Comm., 26, 3404 (1960); changes as ascertained by analysis of isolated H. Berg, Noturwias., 48, 100 (1961); H. Berg and H. Schweiss. pNHz p,p’-diOCHs p,p’-di-
0
f
Monotsh. (Berlin), 9, 546 (1960); H. Berg and H. Schweiss, Naturwiss., 47, 513 (1960). (29) T. Kuwana, J. N. Pitts, Jr., and A. Marchetti, Abstracts of Papers, 140th National Meeting of American Chemical Society, Chicago, 1960, p. 29-8, No. 78.
(30) J. €I. Chaudet and J. W. Tamblyn, Abstracts of Papers. 138th National Meeting of American Chemical Society, New York, N. Y., 1960, p. 19-T. (31) I,. Piette, J . Sharp, T. ICuwana, and J. N. Pitts, Jr., J. Chem. Phya., 36, 3094 (1962).
PHOTOCHEMICAL PROCESSES OF KETOKES IN SOLUTION
Dec., 1062
solid, gas-liquid chromatography of the liquid phase, and ultraviolet spectral and photopotential analyses of the reaction mixture. Low temperature (77°K.) electron spin resonance studies were conducted in conjunction with Dr. L. Piette of Varian Associates to see if radicals of the type I V or V could be detected (to determine whether a reversible hydrogen atom abstraction was the cause of the failure to photopinacolize). Upon irradiating this solid solution (lo+ ill p-aminobenzophenone in IPA) directly in the elecotron spin resonance cavity with light above 3000 A. (Pyrex filter), no electron spin resonance (at ca. 3000 gauss) attributable to a p-aminobenzophenone free radical was detected. However, strong phosphores-
OH
V
11‘
2459
triplet mas observed (0.32 sec. mean lifetime), but BachmannZ0had reported isolation of a photopinacol from this reaction. Others reported that the reaction failed. An approximate quantum yield measurement (Table 111) indicated that the photopinacol reaction did proceed, but with a quantum yield of 0.2 mole/einstein, or 10% of the efficiency of the photopinacol reduction of benzophenone. Possibly the low quantum yield is responsible for the previous inconsistent observations. These observations coupled with recent results of Ermolaev and Terenin1h35for long-lived phosphorescence states suggest a possible explanation for the lack of H atom abstracting ability on the part of para-substituted benzophenones which do not pinacolize. Ermolaev and T e r e n i r ~studied ~~ the absorption and phosphorescence spectrum of p-phenylbenzophenone (and other derivatives) in rigid solvents a t 77’K. They pointed out that while its first absorption band was similar to that of unsubstituted benzophenone, and presumably resulted from an n a* transition, its phosphorescence spectrum resembled instead that of p-hydroxybiphenyl, which presumably is due to a transition from the lowest lying excited state, the a,a* triplet, to the ground state. Thus, they suggest an intramolecular transfer of energy between triplet states of pphenylbenzophenone. (Intermolecular tripletsinglet transfer of energy is well established.) The sequence of events being
-
cence was noted during irradiation in the rigid medium a t 77”:K. and an electron spin resonance signal due to the triplet state of p-aminobenzophenone was observed a t ca. 1500 gauss.31 This electron spin resonance is attributable to transitions between the highest and lowest sublevels and corresponds to A M , = 2.32 The mean life of the triplet state of p-aminobenzophenone in EPA measured by both electron spin resonance and phosphorescence decay (first order) was of the order of 0.3 So h ~+ i Si ---+ Tn,,*+TT,+-+- So hvz sec. Long-lived phosphorescence and electron spin where Tn,,* and Tff+ are triplets with the latter resonance triplet signals were also observed for p- lying a t a lower level and reached by an internal dimethylamino-) p,p’-bis-(dimethylamino)-,and p- conversion process from the Tn,T* triplet ( T n , r * phenylbenzophenones. The phosphorescence mean is formed in the usual way by rapid intersystem lifetimes of these compounds are listed in Table I. crossing from the first excited singlet, S1.) While the identification of a long-lived triplet The failure of certain p-substituted benzophestate in these systems was of interest per se, an nones to photopinacolize now seems reasonable if important correlation was that the long-lived it is assumed that the differences in the electronic triplets appeared to be observed with compounds structures of its two triplet states are reflected in which also gave low quantum yields (or no re- their different chemical reactivities. . .that is, action a t all) in the photopinacol reaction. Thus, H atom abstraction is possible by the n,n* triplet a “normal” benzophenone such as p,p’-dimethoxy- of benzophenone and its derivatives because of the benzophenone has a quantum yield of 2.1 com- localization of the excitation (a non-bonding elecpared with 2.0 for benzophenone (presumably, the tron on the 0 atom) but in the x,n* triplet, the 2.1 is in error because of the approximate measure- greater delocalization renders relatively ineffective ments) and has a triplet state with a mean lifetime the H atom abstracting “power” of the 0 atom in of 0.06 sec. (Table I). the carbonyl group. On the other hand, p-aminobenzophenone had R While this is speculative, it is interesting to note triplet whose mean lifetime (Table I) was 0.40 that P-acetonaphone behaves in a similar fashion, sec., and had a quantum yield which was too small that is, its absorption appears to be n + r * but to be observed in the photopinacol reaction. The its phosphorescence seems to originate from a a,n* data for p,p‘-bis-dimethylaminobenzophenonewere triplet state (both in mean lifetime and character similar to those of p-aminobenzophenone : quan- of spectrum.35 This correlates well with the tum yield in photopinacol 10-3; triplet lifetime results of independent photochemical studies of 0.27 sec. With p-phenylbenzophenone a long-lived Hammond and leer maker^^^ for this compound. (32) Triplet resonance corresponding t o AM, = 2 was first observed They found irradiation of P-acetonaphthone in in rigid glass media by van der Waals a n d de Grootaa by irradiating optically active 2-octanol led to no racemization naphthalene in glycerol a t liquid Nz temperatures. More recently, of the alcohol so that the failure of this compound Farmer, Gardner, and MoDowella‘ have used similar techniques t o is not due to reversal of the hydrogen transfer study the exchange of energy between the triplet state of benzophenone and naphthalene in rigid EPA solution a t 77’K. reaction. Instead they suggest that the “lowest
+
+
(33) J. H. van der Waals and M. S. de Groot, M o l . Phye., 2 , 233
(1959): {bid., 3, 190 (1960).
(34) J. B. Farmer, C. L. Gardner, and C. A. Monowell, J . Chem. P h y s . , 34, 1058 (1961).
(35) V. Errnolaev and A. Terenin. J . chzm. p h y s , 698 (1958). (36) G. S. Hammond and P. A. Leermakers. J . Am. Chem. Soc., 84, 207 (1963).
J. N. T'TTTS,
24GO
Square Pyrex
I
\
JR.,TI. MT. JOHNSON,JTL, .ZUD T. K'TTWANA
0
1
2
w
'I
Figure 1.
t'riplet of reactive aldehydes and ketones have the n,n* configurat,ion, and those of the unreactive groups have the n,n* configuration." With o-aminobenxophenone no electron spin resonance signal corresponding to a radical or a triplet state was observed. Thus the triplet + triplet interconversion is not an attractive explanation for the failure of this benzophenone to pinacolize. The photoenolization is the most probable explanation in these compounds. Anthraquinones.-Table IV presents the results
YOl. GG
benzophenones. Substitution in the 1 or 5 position of the ring by a N H or OH group (see Table IV and references 39-41) inhibits photoreaction. The 2,6-dihydroxyanthraquinonehas been rep0rted3~14~ to be non-photoreactive. Photoenolization may be important while in certein cases an intramolecular energy transfer process may occur to yield a a,n* state which is photochemically unreactive. Abrahamson, et al.,42have suggested the latter explanation to rationalize their results with anthraquinones substituted by (basic) groups capable of donating an electron pair (NH, OH, C1, Br). However, apparently their results are not in agreement with ours in respect to the chloro derivatives (see Table IV). Further work is necessary before a definitive statement can be made about substituents and their positional effects on the photoreactivity of anthraquinones. We have discussed this in detail with Dr. F. Wilkinson (Oxford University) who has studied the fluorescence spectra of some substituted anthraquinones. His results have been presented verballya3 and will be published shortly. Experimental
Macro scale photolysis for product isolation was con. ducted using irradiation from a Hanovia #73A-10 lamp The irradiation cell used was constructed from 37 X 200 mm. Pyrex tubing with appropriate ground glass connections for vacuum degassing. All samples were carefully degassed (10-3 mm., oil diffusion pump vacuum system, liquid Nu traps to prevent sample contamination) prior to photolysis. Products were isolated by crystallization, vacuum distillation, or column chromatograph . Infrared spectra were obtained using the 8erkin-Elmer T.4BLE Iv Model 221G infrared spectrometer. Vapor phase chromatoPHOTOCHEMICAL BEHAVIOR OF SOME SUBSTITUTED ANTHRA- grams were used to analyze for liquid phase photolysis products or to check solvent purity. The Cary Model 14 QUINONES spectrophotometer was used for obtaining all ultravioletCompound Photoreaction Photopotential," volts visible spectra. The electron spin resonance experiments IP.4 BIeOH E t O H were performed on the Varian Model V-4500 spectrometer. Photopotentials and spectra of photolyzed solutions were Anthraquinoneb Yes39 0.6 0.28 obtained by utilizin a special cell shown in Fig. 1. The Derivatives 5-cm. diameter cell 8a.s a depth of 3 cm. and has the proLittle, if any 1,8-diOH visions for measuring the solution ultraviolet-visible absorption by tipping the cell to fill the 1-cm.e Pyrex side-arm which 1,5-diOH Little, if any 0.01 - 0.02 fits the cell compartment of the Ca spectrophotometer. 1,4-diOH Little, if any .01 0.02 A newer cell with a 1-cm.2 quartz s i ? k r m has been con.3 1-c1 structed for lower wave length work. Photopotentials .3 1,8diCl were measured using the 3-mm .z platinum indicator electrode and a probe-type calomel reference electrode. The a Photopotentials have same significance as in Table 11. *Radical e.8.r. signal noted in alkaline ethanol a t room latter electrode is secured to the cell through a clamped 0temperature when irradiated. I n neutral alcohol the e.s.1. ring sealed joint. Contamination arising from stopcock radical signal did not appear until the temperature was grease is reduced by the use of this type of seal. The photolowered to 77°K. during the irradiation. For previous work potentials were monitored using a Model 412A HewlettPackard VTVM (input impedance 200 megohm) whose see references 37 and 38. output was fed to a 10-mv. Varian G-11 recorder. An to date of studies in the anthraquinone series, auxiliary O-ring joint with a 1-cm.2 platinum and calomel electrode attached can be used for conducting additional both published and recent experiment'al data. electrochemical experiments, such as chronopotentiometry, Again the correspondence between the appear- etc. To ensure against leaka e, picein wax was used to ance of a significant photopot'ential and gross outer seal the electrodes to t i e glass wall. This cell is chemical react'ion as detected by product isolat'ion particularly valuable for aiding the study of photochemical is to be noted. Anthraquinone also gave an elec- systems where product isolation and analysis are difficult
+ -
-
+ +
tron spin resonance signal characteristic of a free radical when irradiated in alkaline ethanol or when irradiated in neutral ethanol a t 77°K.37138 The effect of structure on reactivity in this series appears to follow closely that observed for the (37) J. H. Sharp, T. Kuwana, .4. Osborne, and J. N. Pitts, Jr., (1962). (38) K. Kuwata and K. Hirota, Bull. Chem. Soc. Japan, 34, 458 (1961).
Ciiem. Ind. (London), 508
(39) N . K. Bridge and G. Porter, Proc. Roy. Sac. (London), 244,259 (1958). (40) J. L. Bolland and H. R. Cooper, ibid., A226, 405 (1954); H. R. Cooper, unpubliwhed results. (41) G. 0.Schenck and G. Koltzenburg, X a t u w i s s . , 41, 452 (1954). (42) E. W. Abrahamson, I. Panik, a n d K. V. Sarkanen, Proceedings of the Second Cellulose Conference, Cellulose Research Institute, Syracuse, New York, hley. 1959. (43) F. Wilkinson, Symposium on Reversible Photochemical Processes, Duke University, remarks presented from the floor, April, 1962.
(e.g., soluble photolysis products that react rapidly when exposed to oxygen in the air). The ketone-alcohol solutions (for photopotential, spectra, or electron spin resonance work) were photolyzed using either an Osram HBQ-200 or PEK-109 high pressure mercury arc lamp. The radiation was rendered approximately monochromatic in the 3130 or 3660 A. region by appropriate was used for Corning glass filters. Ferrioxdate actinometry. Phosphorescence measuremetits were made on vacuum degassed, sealed-tu be samples in rigid ethanol (unless otherwise stated) a t liquid Nz temperatures. Irradiation from the PEK-109 lamp was filtered to allow transmission a t 3150-4000 A. which then passed through a rapid closing shutter system to the sample. At right angles to the excitation light, the phouphorescence was monitored by a photomultiplier tube (filtered t o allow transmission only above which fed an input signal to an Bmerican Instru4400 i.) ment Go. photomultiplier microphotometer whose output si nal was recorded on a Sanborn Model 154-100 recorder (&art speed: 100 mm. per second). Matheson, Coleman and Bell reagent grade alcohols were purified by the following procedure: to a 2-1. round-bottom flask containing ca. 200 ml. of alcohol were added 5 g. of granular magnesium metal and 0.5 g. of iodine crystals. The solution was refluxed for 1 hr., the remainder of the alcohol added (1 to 1.5 I.), refluxed for an additional 3-4 hr., and then distilled from a 60-cm. glass helices packed column. Only the middle fractions were collected and used. (44) C. A. Parker and C. G. Harchard, Proc. Roy. Soc. (London), A220, 101 (1953); i b z d , 8236, 518 (1956); "Photochemistry in the Liquid and Solid States," John Wiley a n d Sons, Inc., New Yrrk, N. I-.,1960,p. 41.
The ketones used in this study were known compounds prepared by reported methods.",* They were purified by crystallization, sublimation, or column chromatography. The samples of the chloro- and hydroxy-substituted anthraquinones were kindly supplied by Dr. E. .J. Rowen and Dr. D. Seaman, Oxford University.
Acknowledgment. -The authors wish to acI;nowledqc the c.oiitrihutions of Mr. ,J. H. Sharp, N r . T. C. Li, A h . 0. Paez, and Drs. 1'. West and G. Black to this paper. lye are indebted to Drs. S. I. Chan and L. Piette for their helpful suggestions concerning the electron spin resonance aspects of the work, and to Dr. F. Wilkinson of the Physical Chemistry Laboratories, Oxford University, for interesting discussions of the problem. This research was supported in part by a grant from the Petroleum Research Fund administered by the American Chemical Society. Grateful acknowledgment is hereby made to the donors of this fund. The authors also gratefully acknowledge partial support of this work by grant No. 4P-109 from the Public Health Service, Air Force Contract, AF 19(604)-8096, from the Geophysics Research Directorate of the Air Force Cambridge Research Laboratories, and a grant from the Yational Science Foundation. (4.7) P. Graminaticakis, Cornpt. rend., 236, 546 (19.52). (46) J. P. Cordner and K. H. Parisacker, J . Chem. Soc., 102 (1953).
R.WD STRUCTURE AND TRANSPORT OF HOLES AND ELECTRONS IN HOMOLOGS OF ANTHRACENE1 BY G. D. THAXTON, R. c. JARNAGIN, Physics and Chemistry Departments, The University of Narth Carolina, Chapel Hill, North Carolina AND
M. SILVER
Army Research Ofice (Durham), Durham, North Carolina, and Physics Department, University of North Carolina, Chapel Hi!l, North Carolina Received M a y 85. 196.9
Calculations of t'he band structure and of the mobility of excess holes and of excess electrons in homologs of anthracene have been coniplcted. Following LeBlanc, the tight binding approximation w m used and applied to naphthalene, tetracene, and pentacme. Calculated mobility tensors and band widths indicate the mobility properties of excess charge carriers in all four molecu1:ir crystals to be much alike, Experimental results for the crystals other than anthracene are not yet available.
Introduction
A theoretical treatment of the band structure and transport, of excess holes and of excess electrons in anthracene has recently been reported by LeBlanc.2 LeBlanc applied the tight binding approximation to construct crystal wave functions in order to describe the motion of excess charge carriers. The crystal wave functions are formed from linear combinations of molecular orbitals constructed within the Huckel approximations from Slater 2p, atomic orbitals.s Linear combinations of molecular orbitals for the highest bonding state describe the band for the excess hole, and linear combinations of the molecular orbitals for the lowest anti-bonding stake describe the band for the excess (1) Partially supported b y the Army Research Office (Durham) and the National Science Foundation. (2) 0. H. LeBlsnc,
