1536
Val. 66
NOTES
mercaptan (B.M.) by the n.m.r. techniquea i s described. The 1i.m.r. spectrum of neutral R.M. consists of a doublet due to the methylene group, a triplet due to the thiol group, arid a single line due to the aromatic protons. The fipectra of the methylene a d thiolic group show a second-order splitting, since the spin-spin interaction is not negligible in comparison to the chemical shifts of the two groups. The addition of both acid and base to neutral B.M. causes the resonance lines of the methylene and the thiol to broaden. These changes in the line shape can be interpreted in terms of the average lifetime, 7, of the thiolic hydrogen between successive exchanges. In practice only the methylene doublet was used.4 Pure B.M. was purchased from Fluka A. G. and was distilled in vacuum under nitrogen. No degassing of the samples was carried out before use. Acidic solutions of B.M. were prepared by dissolving anhydrous silver perchlorate in neutral B.M. The resulting precipitate of silver salt was removed by several centrifugations and the titers of the supernatants were determined by titration against an aqueous-ethanol solution of NriOH using thymol blue as indicator. Basic solutions of B.M. were prepared by the addition of known quantities of Triton B (benzyltrimethylammonium hydroxide) solution in methanol to neutral B.M. These solutions contained, therefore, certain amounts of CHaOH, about 0.5% by volume for the most basic solution studied. N.m.r. tipectra of these solutions were recorded and the line shape interpreted to give values for the specific rate of exchange of the thiolic hydrogen, 117. (For definition of specific rate see ref. 5 ) .
The experimental results show that 1 / increases ~ linearly with the acid concentration, being practically zero for neutral B.M. This behavior can be explained in terms of the reaction
kI Ce,HSCHzSHz' CaIT6CH2SH + CsHsCL-12SH C&H&€LSI3~ (1)
+
+
f
The slope of the linc describing the dcpeiidencc of 1 / ~ upon the acid concentration gives ICI. It was found that
I c ~ = 1.35 X 102(M-'
SCC.-')
'l'h rtwlts for thc basic rmge show tlial, Lhe addition of basc, up t o 1.2 X loe3 M , does not affect the spectra of B.M. I:urther addition of basc results in t h t broadelling of the lirics of thc spectra. 1,'. cnlculattd from the line shapc is liiicarly dependeiit on the basc concentration in excess of 1.2 X M . The exchange reaction in basic solution is lC2
C&I&I&S-
+ CtjHsCH2SII + C&I&IIZSH + CsRsCfIzS-
(2)
There are however two experimental difficulties in obtaining the rate constant of this reaction. (a) Basic solutions of B.M. are unstable. This is (3) J. A. Pople, JV. G. Schneider, a n d H. J. Bernstein,"Highresolution Nuclear Magnetic Resonance," nloGraw-Hill Book Go., Ino., New York, N. Y., 1959, p. 454. (4) A. Loewenatein a n d S. Meiboom, J . Chem. Phys., 27, 1067
(1987). (5) Z,Luz, D. Gill, a n d 9. Meiboom, ibid., 80, 1640 (1959).
manifested by changes with timr of the line shape of the resonances of both the thiol and methylene groups in basic solutions of B.M. Triton B solutions mere found to be more stable than other buses tried (CE,ONa and C6€IaCB2Sli). (b) The fact that the addition of less than loA3111 of base does not affect the rate of exchange can be attributfd to the presence of a small amount of acidic impurities ( M ) in the original B.M. This assumption is supported by experiments with B.M. that was first kept for several hours over anhydrous I1&03. In this case the intercept was much closer to zero but on the other hand these solutions were more unstable and therefore could not be studied quantitatively. The slope of 1,'. as a function of base concentration above 1O-3M of base is used as an approximate measure of the rate constant of reaction 2 , giving k2 6 X lo5 M-lsec.-l. The above interpretation of the experimental results could in principle be checked by conductance measurements in B.M. solutions; however, such measurements were not carried out.
-
PHOTO-INDUCED BINDING OF "FLUORESCEIN. DYES-TO%N~-OXIDET BY GERALD OSTERAND MARKWASSERMAN~ Department o j Chemzstry, Polytechnzc Instztute o j Brooklyn, Bi o o b l p , *v. Y . Rocezved J a n u a r y 6 , 1061
During the course of studies in the binding of dyes to solid substrates we noticed that for fluorescein and its halogenated derivatives the rate of binding is considerably increased if the dye is illumiiiated by visible light. The effect is particularly noticeable with zinc oxide powder as the rubstrate although, as will be demonstrated, the effect does not arise from the photochemical properties of the substrate itself. It is the purpose of the present paper to describe some of the kinetic aspects of this unusual reaction. I n particular, we are concerned with which excited state of the dye is involved in the reaction. Thc kinetics of photoreduction of fluorescein and its halogenated derivative^^,^ have established that the Iong-lived (probably triplet) state of the excited dye is respoiisible for the electron abstraction from the mild reducing agent present during the photoreduction. As mill be seen, the same excited state is involved in photobinding to solid surfaces. Those substances which retard photoreduction also retard photobinding and in the same quantitative manner. Experimental M~terials.-Fluorescein and its halogenated drrivativcs mere sodium salts obtained from Fisher Scientific (Kational Aniline, histological grade). Most of the quantitative studies were carried out with rose bengal (2',4',5',7'-tetraiodo-3,4,5,6-tetrachlorofluorescein). (1) Presented a t the New York-Sew Jersey Regional Mecting of the American Chemical Society, January 22, 1962. Supported i n part b y the United States Air Force through the Air Force Cambridge Contract No. 19(604)-3065. (2) On the National Science Foundation Summer Chemistry Program of the Polytechnic Institute of Brooklyn while a senior a t the Brooklyn Technical High School. (3) G. Oster a n d A. H. Adelman, J . Am. Chem. Soc., 78, 013 (1956). (4) A. H. Adelman a n d G. Oster, ibid., 78, 3977 (1956).
August, 1962
NOTES
The aim oxides employed were of two types: one is Analytical Reagenti from Mallinckrodt Chemical Company and the other is Florence Green Seal 8 obtained from the New Jersey Zinc, Company. This latter material is made by heating zinc oxide in the presence of zinc vapor. With the use of wire meshes only particules between mesh sizes No. 100 to No. 200 where chosen (approximate mean diameter of sinc oxide particles 110 p ) . Aluminum oxide (“activated alumina”) was obtained from Alcoa. All other chemicals were Eastman C.P. grade. The nitrogen employed is prepurified Airco and further purified to remove any traces of oxygen by passing through a chromous chloride solution. Procedure.-All dye solutions were made up in 0.1 M phosphate buffer a t pH 7.0. The solid substrate powder was introduced into the cylindrical reaction cell (10 cm. high and 5 cm. diameter) fitted with a polyethylene cap to allow nitrogen t o be bubbled through the suspension. The bubbling was carried out 15 min. prior to and during the irradiation so as to agitate the suspension as well as t o flush the system free of oxygen. The cell was illuminated with a 500-watt tungsten lamp slide projector provided with the appropriate filters. Using a green interference filter (Baird Associates) which is 80% transmitting a t its peak (539 mp) i t wa8 found with a calibrated thermopile (Eppley Laboratories) that the intensity of light falling on the cell is 9.2 X einsteins/cm.*/sec. After illumination of the system for a particular period, the samples were centrifuged a t 3000 r.p.m. for 30 min. so as to obtain a clear, non-turbid supernatant above the colored ZnO sediment. The optical density of the carefully removed supernatant wm determined in a Carey Model 14 spectrophotometer. I n one experiment the system was illuminated with a medium pressure mercury lamp (General Electric AH4) fitted with EL Wood’s glass filter to isolate the 365 mp line. All the above solutions were in 0.1 M phosphate buffer at p H 7.0.
mitting filters) including erythrosin B, phloxine 13, eosin Y, and fluorescein along with its dihalogenated derivatives. On the other hand, no member of the thiazine, acridine, os azo families of dyes of the many examined were found to give the photobinding effect. Bubbling a 50-50 mixture of oxygen and nitrogen through the illuminated solution resulted in a decreased optical density, but this is due entirely to irreversible photooxidation of the dye since the result was independent of the presence of zinc oxide, Ultraviolet irradiation with 365 mp where the dye does not absorb appreciably gave no reaction. This was the case for both types of zinc oxide. When an aqueous suspension of zinc oxide in water is illuminated with ultraviolet radiation in the presence of oxygen, hydrogen peroxide is produced.6 However, in our case, the absence of oxygen, due to the constant flushing with nitrogen, prevents this reaction from occurring. The zinc oxide powders were subjected to high vacuum with heating to remove gases and then used for the dye studies. It was found that the rate of binding was unaffected by such treatment. Aluminum oxide served equally well as zinc oxide for the photobinding studies. Illumination with orange light of rose bengal and of zinc oxide under nitrogen prior to their mixture, produced no effect unless the mixture was further illuminated. Discussion The photobinding effect does not involve the photoexcitation of the solid substrate. Although certain zinc oxides are photoconductors, their spectral sensitivity6J generally lies at shorter wave lengths than the light employed to excite rose bengal. Although the zinc oxide with excess zinc could be sensitive to blue light, the pure zinc oxide would be sensitive only to ultraviolet light. Furthermore it is unlikely that alumina would be sensitive to visible light, yet it was also a suitable substrate. Illumination with ultraviolet light undoubtedly excites both types of zinc oxide, but does not produce photobinding. The photoconduction of zinc oxides can be sensitized by Photobinding is apparently unrelated to this phenomenon in that in our case the dye is not initially adsorbed to the solid. Furthermore, dye-sensitized photoconductivity can be carried out with dyes such as methylene blue and acriflavine which we found exhibit no photobinding. The retardation of photobinding with small concentrations of potassium iodide and especially of p-phenylenediamine clearly indicates that the longlived excited species of the dye is involved in the r e a ~ t i o n . ~The , ~ Stern-Volmer quenching con-
Results I n the concentrations employed (below M) all the dyes obey Beer’s law. In the absence of light, rose bengal and the other fluorescein dyes do not adsorb to zinc oxide over the time periods (up to 4 hr.) involved iu the present work. With orange light (obtained with a Corning No. 3486 filter which cuts off light below 520 mp) there is a considerable photobinding effect as manifested by a decrease in optical density of the supernatant of the centrifuged sample. Resuspending the dyed solid particles and heating the system a t 100’ for 1 hr. gave practically 100% recovery of the adsorbed dye, the dye being unchanged in spectrum. Under green monochromatic light at 539 mp (the dye absorbs maximally at 548 mp with a molar extiiiction coefficient of 8.5 x lo4) it wv&sestablished that the quantum yield for disappearance of the dye, averaged over 1 hr. of illumination, is 1.2 X low4. This value is only approximate since the powder (Ainalytical Reagent grade ZnO, 1.5 mg./ ml.) scatters light and considerably complicates the photometry of the system. The rate of photobinding is proportional to the zinc oxide content up to about 1.5 mg./ml., aftcr which it appears to be insensitive to the concentration of powder present. Trace amounts (of the order of M ) of potassium iodide and p-phenylenediamine retard the reaction. The rate is reduced to one-half its normal value in the presence of 1.0 X lo-* M KI. Less than one-hundredth this concentration of pphenylenediamine was equally effective. Qualitatively similar results t o those described above were obtained for other members of the fluorescein family (using the appropriate trans-
1537
(5) M. C. Markham and K. J. Laidler, J . Phys. Chem., 57, 363 (1963). ( 6 ) P. Miller in “Semiconducting Materials,” Butterworths, London, 1951. (7) R. Arneth quoted by G. Heiland, E. Mollwo, and F. Storkman in Vol. 8 of “Solid State Physics,” F. Seitz and D. Turnbull, Ed., Academic Press, New York, N. Y . ,1959. (8) E. K. Putseiko and A. N. Terenin, Do& Akad. Nauk SSSR,90, 1005 (1953). (9) C. J. Young and II. G. Greig, R.C.A. Rea., 15, 469 (1954).
1538
stant for the long-lived excited state was found to be 1.0 X lo7l./mole. The lifetime is computed to be 1.5 X sec. assuming that every diffusional encounter (6.6 X 10Y/sec./l. for a 1 ,1/1 solution) between quencher and excited molecule is effective. It is significant that the concentrations of retarders for this reaction are close to those for the retardation of the photoreduction of the fluorescein-type dyes. It appears that the triplet species of these dyes have a strong affinity for the solid substrates. Thioniiie dyes in the long-lived state undergo certain photochemical reactions when bound to specific high polymers.1° In the present case, however, the long-lived species wanders through the solution until it encounters a solid particle; failing such an encounter during 1 msec. it drops to the ground state. This may account for the low quantum vield of photobiiiding for rose bengal. This is approximately l/aoo that of the maximum (extrapolated to infinite reducing agent concentration) for photoreduction. Increasing the amount of zinc oxide should increase the chance of such an encounter and hence, increase the rate of the reaction. However, the system is unaffected by concentrations of the powder beyond 1.5 mg. per ml. due to the light scattering by the suspended substrate which decreases the amount of light allowed into the solution. For zinc oxide powder of this mesh size, the surface area is estimated from extrapolation of published data on ziiic oxide powderll t o be about 100 cm.Z/g. For a 10-ml. solution of rose bengal at 1 0 F M in the presence of 15 mg. of zinc oxide, complete binding takes place with prolonged illumination. That is, 4 X 1OI6 molecules are bound per cm.2 of surface. Since the dye molecules ocit is obvious cupy an area of the order of 100 that niultilayer adsorption occurs. Apparently the light-excited dye has an af€iiiity for solids coated with adsorbed dye as well as for the original solid substralx i. (10) K Wotherspoon a n d G Ostet, J . Am Chem (1957). (11) C W Siller, zbid., 66, 431 (1943).
Sol
, 79, 3992
IlhDIOLk-SIS OF TOLUENE : JIECHANISXL OF FORMATIOK OF BESZYL RADICALS BY 1,. H. GALE,B. E. GORDON,G. STEINBERG, AND C. D. WAGSER Shell
'Vol. 66
NOTES
022
Compantj, Ma.il%nez Caldornm und Shell DeLdopment Company, Ernes y u d i e , Cahfool nzn
radica1s.l When isolated molecules of toluene in the gas phase are ionized by energetic elertrons in a maws spectrometer, the principal fragment ion is the cyclic tropylium ion, C7H7+, as shown by %!eyerson and Rylander.4 (They found by isotopic means that the probability of loss of the single hydrogen to form the ion is the same for all types of hydrogen on toluene, and that all carbon atoms in the C7H,+ ion are equivalent.) Although a t 25' some of the benzyl radicals arr undoubtedly formed by hydrogen abstracatioii by tolyl radicals,l they also must be formed by a dirwt process. They may be formed directly by detachment of an alpha hydrogen froni an excited toluene molecule, as in (a), or by rearratigenieiit of Ihc tropylium ion on neutralization by an electron, as
in (b). 111 the present study, t,oluene-a-C14 was irradiated with 3 Mev. electrons in order to differentiate between these pat'hs of reaction. Experimental A sample of two microcuries of t01uene-cu-C~~ (Research Specialties Go., Richmond, California) was diluted with 71 g. of research grade toluene, and placed in a circulation system containing an irradiation cell.6 The cell was cooled to -30" and degassed by successive addition and removal of dry, oxygen-free nitrogen while the liquid wa,s circulated through the cell. The irradiation was conducted with a 45-painp. 3-1Iev. electron beam, collimated by a 1 /2 in. 0.d. pipe before striking the thin window of the cell. A i irradiat,ion timc of 4800 see. gave an est>imateddose of 8 X lo8 rad. Toluene was separated from the heavier material by distillation, and the dimer fraction was recovered from t'hc residue by several gas chromatographic separations o r il. 6-mm. diameter X 1.5 n:. silicone column at 170". The recovered fraction (1 51 mg.) was diluted with 299 nig. of bibenzyl (Eastman Nodak White Label) and oxidized by chromic acid. The resulting benzoic acid was purified by sublimation, yielding 307 mg. of material ( 5 1 % ) . llost, ol t.his (274 mg.) was degraded by the Schmidt reaction t>oaniline (8656) and carbon dioxide (8275), which in turn wcrc convert,ed to acetanilide and barium carbonatc.
TABLE I SPECIFIC ACTIV~TIEB OF SCHMIDT D E C ~ ~ A D A TPIE OO S U~CW Starting material
Products
Bibenzyl-benzyltoluene-bitolyl
Benzoic acid Acetanilide Barium carbonate Benzoic acid Acetanilide Barium carbonate
Toluene
IZecezbed January 11, 1962
Toluene in the liquid phase reacts under ionizing radiation to give hydrogen and polymeric products. The Clinierip materials have been shown by HoigiiB , i i d Gaumann to consist of bihenxpl, tht. h i u y l toluenes, and the six bitolyls. Bridge2 a i d Chiltoir and Porter3 has-e shown that ionizing radiation acting on toluene in a solid matrix a t low temperature generates benzyl radicals, and the bibenzyl appears to be formed by rombinatiou of the benzyl ( I ) J Jloigiid and 1' G&*ninrirn,Jiclv C h 7 n i l d n , 44, 2141 (1Uhl) (21 N. I'i Ui,dge, Iature, 186,30 (1960) ( 3 ) H. T. Chiltoil and 6 Porter, J. Pitus Chern , 63, 904 ilSSD)
Nc./mole
0.42 i.0.03 o.ooo=koo.oo15 0.37 =k0 . 0 1 2.47 0.000f0.0051
2.21
Results and Discussion lmizoic :wid wits derir-cd solely froin LIIC id)eiizyl, siirtae thr, tliiiieric acids From the other 7'11tk
dimers were discarded. All of the activity was found in the carboxyl carbon. The reduced specific: activity in the carboil dioxide may be due to dilution of the sample by atmospheric carbon dioxide (4) S. Meyersoiu a n d P. X. Rylander, , I . Am. C h r m . S n c . , 79, 812 (3957);d . Chenz. Phys., 9 7 , !)01 (i!457). (j) Tlic agparatiis was dcscribod prcvioualy: C. 11. WUIICU../. I'iiys. CRem., 64, 231 (1960).
