2634
J. Phys. Chem. 1980, 84, 2634-2641
drogen-bonded imidazole molecules on the uranyl ion in the UNI crystal, and a similar effect of the noncoordinated lattice water molecules in the UNH complex can also be postulated. The inherently greater luminescence associated with the UNI system relative to that of UNH does suggest that some amount of energy transfer may be occurring. At the present time, it is not possible to partition the amounts of the contributions from these several potential effects. Acknowledgment. This work was supported by a grant from the Research Corporation to H.G.B. through the Cottrell Research Program (No. 8926). One of the authors (D.L.P.) wishes to thank the Miller Institute for Basic Research, University of California, Berkeley, for a Miller Research Fellowship which sponsored the initial synthetic portion of this research. The authors also thank Professor Paul N. Schatz for a most helpful discussion and the Department of Energy for support of this work under Contract No. W-7405-ENG-48.
References and Notes (1) National Science Foundation Postdoctoral Fellow, 1976-77; Miller
Fellow, 1977-79. Rabinowitch, E.; Belford, R. L. ”Spectroscopy and Photochemistry of Uranyl Compounds”; Pergamon Press: Oxford, 1964. Burrows, H. D.; Kemp, T. J. Chem. SOC. Rev. 1974, 3 , 139. Jorgensen, C. K.; Reisfeld, R. Chem. Phys. Lett. 1975, 35, 441. Perry, D. L.; Ruben, H.; Templeton, D. H.; Zalkin, A. Inorg. Chem. 1980, 79, 1067. Taylor, J. C.; Mueller, M. H. Acta Crystallogr. 1965, 79, 536. (a) Conn, G. K. T.; Wu, G. K. Trans. Faraday SOC.1938, 34, 1483. (b) Jones L. H.; Penneman, R. A. J . Chem. Phys. 1953, 21, 542. (c) Jones L. H. J . Chem. Phys. 1955, 23, 2105. The uranlum-uranium bond distance of 3.927 A in di-p-aquo-bis[doxobis(nitrato)uanium(VI)] dllmidazde is the shwtest reported length between two uanyl ion centers, and it thus affordsthe best opportunliy to date for detecting metal-metal coupllng in the UO ’; system. Two other bridging uranyl dlmers (which possess bridging hydroxy groups instead of water) that have quite similar bonding are the [(NO&U02(OH)2U02(H20)3]~Hz0e and CI(Hz0)3U02(OH)~U0z(H,0)3CI’o complexes which have uranlum-uranium distances of 3.939 and 3.944 A, respectively. Perrin, A. Acta Crystallorgr.,Sect. B 1976, 32, 1658. Aberg, M. Acta Chem. Scand. 1969, 23, 791.
Optical and Magnetic Properties of Substituted Benzophenones with Lowest %7r* States G. P. M. van der Velden, E. de Boer,* and W. S. Veeman Department of Physical chemistry, University of Nvmegen, Toernoolveld, 6525 Ed Nvmegen, The Netherlands (Received: February 25, 1980)
Phosphorescence and optically detected magnetic resonance (ODMR) experiments are reported for the lowest excited triplet state of some benzophenone derivatives, Le., 4-benzoylbiphenyl,4,4’-diphenylbenzophenone, and 4,4’-dibenzoylbiphenyl. Moreover, we studied two compounds similar in structure, l-benzoylnaphthalene and 1,5-dibenzoylnaphthalene. All of these compounds have been studied in single-crystal form and have a lowest m* triplet state. The contribution of the magnetic dipole-dipole interactions to the zero-field splitting is determined and compared with data for benzaldehydes and acetophenones. An isotope effect has been seen in the ODMR spectrum of the triplet state of 4,4’-bis(4-~hlorophenyl)benzophenone, which appeared to be present in 4,4’-diphenylbenzophenone.
Introduction In the present study benzophenone-like systems have been investigated for which it was expected that the character of the lowest triplet state would be no longer, as in benzophenone, nr*. The state character is partly determined by the nature of the substituent attached to benzophenone (BP). Substituents which promote rr* lowest triplet states are amino, diethylamino, dimethylamino, and phenyl. We have chosen phenyl groups containing substituents because amino-, (dimethylamino)-,and (diethy1amino)benzophenone are too sensitive for solvent effects, display multiple emission,l and are difficult to purify, and moreover no crystallographic data are available for these compounds at present. The phenyl-containing groups have been substituted solely at the para position(s) of BP. Examples are 4-benzoylbiphenyl (BB), 4,4‘-diphenylbenzophenone (DBP), and 4,4’-dibenzoylbiphenyl (DB). Moreover, we have studied two compounds similar in structure to BB and DB except that the biphenyl group has been replaced by a naphthyl group. These two compounds are l-benzoylnaphthalene (1-BN) and 1,5-dibenzoylnaphthalene (1,5-DBN). All mentioned structures have been depicted in Figure 1,together with the structure of 2-benzoylnaphthalene (2-BN). The triplet state of BB has been studied in the gas phase,2 in solution at room temperat~re,~ and in matrices 0022-3654/80/2084-2634$0 1.OO/O
at 77 KS4 Its naphthyl analogue (1-BN) has also been studied frequently in matrices at 77 K.516 The crystal structures of the compounds considered are unknown, except those of DB and 1,5-DBN which have been solved recently in the crystallographic department of our f a c ~ l t y . ’ Hardly ~ ~ ~ ~ any spectroscopic data are available for these two compounds? The systems, BB, DB, 1-BN, and 1,5-DBN, have been studied by us in neat crystal form. 4,4’-Bis(4-~hlorophenyl)benzophenone (CBP) or its monochloro analogue appeared to be present as a radiative impurity in 4,4’-diphenylbenzophenone.Phosphorescence and ODMR experiments carried out on extensively zonerefined material clearly indicate that the emission belongs to CBP, which apparently cannot be removed by zone melting. The above-mentioned systems have been investigated with the help of phosphorescence emission (PE), ODMR, and MIDP experiments. We have found that the To states of these five systems are best described as a?r* states. The contribution of the magnetic dipole-dipole interaction (&) to the zero-field splitting (ZFS) has been determined by calculating the spin-orbit coupling (SOC) contribution. Subsequently the values of Dss have been compared with data for benzaldehydes and acetophenones. It was found that a description of DB as a “double 0 1980 American Chemical Society
The Journal of Physical Chemistry, Vol. 84, No. 20, 1980 2835
Properties of Substituted Benzophenones
r.?? z
I
v p ry 0
1-Benzoylnaphthalene (1-BN) and 1,5-dibenzoylnaphthalene (1,5-DBN) (both K and K chemicals, specified purity between 95 and 99%) were recrystallized (1-BN: ethanol (lx),acetone (1X); 1,5-DBN: glacial acetic acid (2X), ethanol (2X), acetone (2X), sublimed (2X)) and zone melted along 65 passes. Crystals were grown via the Bridgman technique. The zero-field ODMR spectrometer is of a common design (ref 9 and references therein). The microwaves are supplied by a H P 8620 C sweep oscillator with the appropriate plug-ins for 0.2-0.5 and 1.0-6.5 GHz and are amplified by a H P amplifier (0.2-0.5 GHz, 10 W; 1.0-6.5 GHz, 1 W). The sample is irradiated by a 200-W highpressure Hg lamp via liquid and glass filters.
z
U
I
O
E
IU
Q
Results and Analysis General Remarks. The results of experiments on the triplet state of CBP in DBP will be discussed in full detail. Similar experiments have been performed on the triplet state of BB, DB, 1-BN, and 1,5-DBN. The results for these five systems have been tabulated in Tables 1-111 and concern X-traps excluding CBP in DBP. No exciton emission has been seen in these systems. 3.2. 4,4‘-Ris(4-~hlorophenyl) benzophenone (CBP) in 4,4’-Diphenylbenzophenone (DBP). The PE spectrum at 4.2 K of the DBP crystals is shown in Figure 2 and is as well resolved as those from the other biphenyl ketones. The PE spectrum at 1.2 K is identical with the spectrum at 4.2 K. Clearly two C-0 bands can be discerned at 21079 and 21061 cm-l. Utilizing the IR data for DBP, we endeavored a vibrational analysis (see Table I) which was not completely gratifying. Distinct differences were noted, for instance, between the C=O stretch (1658/1659 vs. 1645 cm-l) and the C-C stretch vibrations (1613/1610 vs. 1605 cm-l). This suggests that both emissions do not arise from DBP but from an impurity triplet state. Because the 0-0 bands satisfactorily agree with the 0-0 band for DBP in a glass at 77 K,15 voo = 21230 cm-l, this impurity must be a molecule structurally related to DBP. From the analysis of the ODMR spectrum as discussed below, it is immediately apparent that we have to attribute the emission to a chlorinated molecule closely related to DBP. Using IR data,16 we can trace the position of the chlorine substituent(s). The vibration at 795/793 cm-l is characteristic for a para mono- or disubstituted aromatic system; the 1658/1659 cm-l vibration points to the presence of an aromatic carbonyl group. The 1289/1283 cm-l vibration characterizes an R-C (=O)-R’ fragment, whereas the vibration at 1111/1105 cm-’ indicates that R and R’ must contain at least a phenyl group. On account of these
ry 0 PI[
Figure 1. Moleicular structures and zf axes: (I) benzophenone (BP); (11) 4-benzoylbl~phenyl(66); (111) I-benzoylnaphthalene (I-EN); (IV) 2-benzoylnaphthalene (2-EN); (V) 1,5dlbenzoyinaphthalene(1 ,5-DBN); (VI) 4,4’dibenzciylbiphenyl (DB); (VII) 4,4’diphenyibenzophenone (DBP).
molecule” of BP is not applicable. Finally, an isotope effect has been seen in the ODMR spectrum of the triplet state of CPB.
Experimental Section 4-Benzoylbiphenyl (BB) (Aldrichl) (>96% purity) was recrystallized from ethanol p.A (3X) and zone melted (120 passes). 4,4’-Diphenylbenzophenone (DBP) (K and K chemicals, ICN) (purity between 95 and 99%) was recrystallized from glacial acetic acid (3X) and acetone (lx),sublimed (2X), and zone melted (120x1. 4,4’-Dibenzoylbiphenyl (DB) was synthesized via a Friedel-Crafts acylation, starting from biphenyl, benzoyl chloride, and aluminum trichloride. The crude product was recrystallized from nitrobenzene, acetone, and glacial acetic acid (each 2 times). DB was sublimed two times and zone melted along 70 passes.
1900
IO00
5100
Figure 2. Phosphorescence emission spectrum of 4,4’-bis(4-~hlorophenyl)benzophenone at 4.2 K.
I200
2636
The Journal of Physical Chemistry, Vol. 84, No. 20, 1980
van der Veiden et ai.
TABLE I: Analysis of PE Spectra for CBP, BB, DB, and 1,S-DBN ( T =1.2 K) peak
A, a
1 1'
4744 4748 4753 4762 4783 4786 4790 4794 4804 4809 4847 4849 4856 4859 4902 4906 4930 4934 4964 4968 5008 5011 5020 5024 5038 5042 5053 5056 5137 5141 51 49 5 1 54
21079 21061 21039 20999 20907 20894 20877 20860 20816 20794 20631 20623 20593 20580 20400 20383 20284 20268 20145 20129 19968 19956 19920 19904 19849 19833 19790 19778 19466 19451 19421 19402
4708 4718 4729 4746 4866 4884 4921 4936 4949 4973 5006 5029 5041 5092 5102
21240 21195 21146 21070 20551 2047 5 20321 20259 20206 20109 19976 19885 19837 19639 19600
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
4945 4950 4955 4963 497 1 4976 4987 4992 5005 5015 5018 5092 5033 5037 5112 5155 5185 5240 5245 5273 5278 5368
20222 20202 20181 20149 20116 20096 20052 20032 19980 19940 19928 19884 19868 19853 19561 19399 19286 19683 19065 18964 18946 18629
20 41 73 106 126 170 190 242 282 294 338 3 54 367 661 823 936 1139 1157 1258 1276 1593
23 24
5374 5382
18609 18580
1614 1642
2 3 4 4' 5 5' 6 6' 7 7' 8 8' 9 9' 10 10'
11 11' 12 12' 13 13' 14 14' 15 15' 16 16' 17 17'
1 2 3 4 5 6 7 8 9 10
11 12 13 14 15
1
v,
cm-Ia
A, cm-Ib
IR, cm-"
CBP 0 0 40 80 172 167 202 201 263 267 448 438 486 481 679 678 795 793 934 932
vaa vaa
2x40
C=O bend C= 0 bend 172+40 167t40 445 445 690 690 786 786 93 2 932
ringdist ip ringdist ip C-H bend C-H bend C-H bend C-H bend
op op op op
? ? 1150 1150
C-H bend ip C-H bend ip 1111+3x40
1286 1286 1605 1605 1645 1645
C-H bend (1,4) C-H bend (1,4) C - C stretch C - C stretch C=O stretch C=O stretch
1105t3x40
BB 0 45 94 170 689 765 919 981 1034 1131 1264 1355 1403 1601 1640
r(CCC) r(CCC) r(CCC) r(CCC) 448+40 438+40
1111 1105 1159 1157 1230 1228 1289 1283 1613 1610 1658 1659
assignment
Von
lattice 2x45 689 765 919 980 1022 1127 1270 1340 1400 1601 1642
DB 0
C= 0 bending (1,4) ring dest ipc C-H bend 0p3 C-H bend op C-H bend op 981tlx45
C-H bend ip (1,4) bend or 1131+3X45 1264+94
C-C stretch C-C stretch C=O stretch Vno
lattice lattice (2x20) lattice 106+20
C= 0 bending 1 7 0 t 2 0 C=O bending 170+73 242t41
ring dist op 294+41 242+73+41 294+73 2X294+73 8 28 936 1136 1152
C-H bend o p C-H bend op
1592 1596 1606 1644
C-C stretch
936 +73+ 126
C-H bend ip
1157 t 2X 4 1 + 20 12 58+ 20
C-C stretch C=O stretch
The Journal of Physical Chemistty, Vol. 84, No. 20, 1980 2637
Properties of Substituted Benzophenones
TABLE I: (C!ontinued) peak A,
a
v , crn-'"
IR, cm-IC
A , cm-Ib
assignment
1,B-DBN
1
12 13 14 15 16
4910 5079 5095 5106 5123 5128 5143 5164 5214 5246 5279 5470 5482 5494 5513 5531
20367 19689 19626 19584 19520 19500 19444 19365 19179 19062 18942 18282 18242 18202 18139 18080
63 105 169 189 245 324 510 627 747 1407 1447 1487 1550 1609
17
5552
18011
1678
2 3 4 5 6 7 8 9 10
11
a
Uncorrectled for vacuum.
A =
It)
voo
0
- vool.
--
24108
23233
0 42 155
0
A , cm-I
415 170
21061 21079 0 40 172
689 765 919
679 795 934
1131 1264 160'1 1640
1159 1289 1613 1658
21240
0
20222
0 41 171)
334 705 727 920
743
823 936
1011 1154 1287 1611 1650
1283 1611
622 760 1402 1445 1500 1597 1668 1662
C=O stretch
Spectra recorded at 298 K in CsI.
TABLE 11: Comparison of 0-0, Phonon, and Vibrational Bands in the I?hosphorescence of Some Carbonyl Compounds and Biphenyllo ( T = 1.2 IK) CBP DB B BB systems BP v o o , cm-l
voo
lattice lattice C= 0 bending 3x63,105+63 169t63 r(CCC) 3 24 + 189 ringdist ip C-H bending C-C stretch C-C stretch C-C stretch, 2x747 1487t63 C-C stretch
1157 1614 1642
arguments we conclude that the unknown species must contain the fragment below:
0
The absence of vibrations due to a 1,Cdisubstituted phenyl ring (between 600 and 650 cm-l) excludes 4,4'-dichlorobenzophenono. The 0-0 bands and the zfs parameters (see below) which point to a species geometrically related to DBP underline this conclusion."
However, vibrations are present which could be attributed to a para-substituted biphenyl compound, namely, at 486/481 cm-'. Therefore we assume that the emission is due to 4,4'-bis(4-chlorophenyl)benzophenone or its monochloro analogue. Although we cannot discriminate between these two systems, we use the symbol CBP throughout this section. The vibrations at 4481483 and 263/267 cm-' can then be ascribed to in-plane bending vibrations of a C-C1 bond in a phenyl ring. The presence of CBP in DBP is not surprising if one considers the synthesis of the latter. If the starting material, biphenyl, is contaminated by 4-chlorobiphenyl, CBP is produced.lsePb As already indicated, we have used IR data of DBP for the analysis of the P E spectrum. From the above-mentioned facts, it follows that it would be better to use the IR data of 4,4'-bis(4-chlorophenyl)benzophenone or its monochloro analogue. Unfortunately, these are not known. However, from the IR spectra of benzophenone, 4chlorobenzophenone, and 4,4'-dich1orobenzophenone,leit can be inferred that chlorine substitution hardly alters the IR frequencies, so that it might be anticipated that the IR spectra of CBP and DBP have large similarities. Hence it is by no means surprising that the vibrational analysis of the CBP phosphorescence spectrum shows a resemblance with IR data of DBP. Large similarities also occur in the vibrational structures and the position of the 0-0 bands of the triplet states of benzophenone and 4,4'-dichlorobenzophenone. The 0-0 bands lie at 24108 and 24181 cm-', respectively (X-trap emission). Large similarities also exist between the magnetic properties of the triplet states of benzophenone and 4,4'-dich10robenzophenone.l~
TABLE 111: Zero-Field and Kinetic Parameters for the Triplet States of Biphenyl, Benzophenone, and Some Biphenyl and Naphthyl Ketones parameters BPa BB CBP DB
-
AB^
MBC Bd 1-BN 2-BNe 1.5-DBN a
Reference 11.
-
c1m-I k x , k y , k z , s-' rav, ms kxr, kyr, kzr -0.1581, 0.0211 36, 54, 625 4.2 6, 9, 8 5 -0.0948, 0.0088 2.4, 2.7, 7.4 240 -0.0939, 0.0077 2.5, 2.9, 22.0 110 27, 9, 64 -0.0886, O.OY.06 2.1, 2.7, 16.7 140 6, 22, 72 -0.0954, 0.0082 230 -0.0978, 0.0074 370 +0.1095, - 0.0036 4200 - 0.0879, 0.0062 1.1, 1.5, 7.7 290 -0.095, 0.026 -0.0850, 0.01.74 4.8, 1.1,7.6 222 49.5, 1, 49.5 Reference 4. Reference 12. Reference 13. e Reference 14. D e x p , Eexpr
Px0, pya, Pz0 38, 39, 23
0.1, 1.9, 98.0 0.3, 3.0, 96.7
2638
The Journal of Physical Chemistty, Vol. 84, No. 20, 1980
DUMA intensity
van der Velden et al.
TABLE IV: Calculated Energies and Measured Frequencies for 4,4'-Bis(4-chloropheny1)benzophenone
I
transition
1 2 3 4 5 6 7
8 9
io
11
12
calcd energya
Z-X+6Q'+qZQ't3/,p z-x+3/4p
z-x
t
01
+ 3/4y
measured frequency
3045
Z-X-6Q'-qzQ'+~+3/4y Z - Y + 6Q' + aZQ' + (Y + 1'lzP 2610, 2618 Z - Y + 3/4p t '147 2583 z - Y + 2a t 3/4p + 'I4y Z - Y - 6 9 ' - q2Q' + CY + l'/ay 2556,2549 Y-X+6Q'+qZQ'-3/4y Y-x-31 462 Y - x- 01 :'3/,p Y - X - 6Q' - q2Q' - CY - 3/4p
a 01 = Axx2/(Z- Y ) ,P = AXXz/(Z- Y + 6Q'), y = A,,'/ (Z - Y- 6Q'), Q' = e2qQ/12. 2519 2556
2583
2610 2618
MHZ
Flgure 3. ODMR spectrum of the 2-Y transition of 4,4'-bis(4-chloropheny1)benzophenoneshowing the satellite lines. Spectrum (a) has been recorded with low and spectrum (b) with maximum microwave power.
A t this point it is meaningful to unravel the multiple emission in CBP. The 0-0bands are separated by 18 cm-'. From Figure 2 and Table I we see that all vibrational bands occur as twins. Only two, probably phonon, bands appear single. Furthermore a good correspondence is seen between the vibrational structures of the two emissions in the PE spectrum. Also ODMR experiments show that both emissions give rise to the same principal zero-field (zf) transitions. Thus we conclude that both emissions are due to a CBP molecule. As hot band emission can be excluded, the emission must come from two CBP guest molecules having different trap depths. In the case of 4,4'-dimethoxybenzophenone, where also a double emission was found,lg it was proposed that two symmetry inequivalent X-traps were responsible for it. This could also be the case for the CBP/DBP mixed crystal. However, for 4,4'-dimethoxybenzophenone,the two zero-field patterns were quite different, which is not so in the present case. Another attractive explanation for the different trap depths could be the presence of two chlorine isotopes, 36Cl (75%) and 37Cl(25%). Following the latter explanation, the presence of one chlorine atom in the emitting species would lead to an intensity ratio of the emission spectra of 3:l. This is certainly not the case, as the experimental ratio shows. The presence of two chlorine atoms, as in CBP, will then give rise to three emission spectra, because of three possible isotopic combinations, namely, (36Cl,35Cl), (36Cl,37Cl),and (37Cl,37Cl)with respective intensity ratio 9:6:1. The emission spectrum shows only two subspectra, which, however, is not decisive to reject this explanation due to the presence of a broad phonon structure. On the basis of the intensity pattern we may then assign the 0-0 band at 21079 cm-' to (35C1,35C1)CBP and the 0-0 band at 21061 cm-' to (36C1,37C1)CBP.The undetected 0-0 band for (37C1,37C1)CBP would then be hidden under the phonon band 2 at 21039 cm-' (see Figure 2). As will be described below, different chlorine isotopes give rise to different ODMR signals. To solve the question whether the double emission is due to an isotope effect, we tried to relate the and 37Clsatellites in the ODMR spectra (see Figure 3) with one of the emissions but the signal-to-noise ratio was not sufficient to take these socalled action spectra.20 Subsequently we would like to discuss the magnitude of the "isotope shift". Yamauchi and Pratt have found for
the 3na* state of benzophenone-dlo an isotope shift of 32 cm-' and for benzophenone, in which the carbon and oxygen atom of the CO group had been replaced by I3C and l80, an isotope shift of, respectively, 9 and 11 cm-lS2* However, it is striking that such isotope effects have not been reported for the 3n7r* states of 4,4'-dichlorobenzophenone,174,4'-dichloro- and 4,4'-dibromobenzophenone in 4,4'-dibromodiphenyl etherz2and 3,3'-dibromobenzophenone2aand also not for the 3aa* states of l-chloronaphthalene in durene%and 8-chloroquinoline in durene.= Therefore we abandon also this explanation. In view of the arguments given above, we prefer to ascribe the double emission to two CBP molecules at different DBP sites. A similar situation (trap depths differing 13 cm-l, equal zfs parameters) has been found for the 3 7 r ~ * state of 8-chloroquinoline in durene.26 Three zf lines were observed at 462,2583, and 3045 MHz by using the slow passage technique and low microwave power. Clearly they belong to one triplet state. The zfs parameters were calculated and tabulated in Table I11 with analogous values for 4-acetylbiphenyl (AB), 4-(y-methylvalery1)biphenyl (MB), benzophenone, and biphenyl (B). The signs of the zfs parameters have not been determined. Therefore we assume that Dexp< 0, as in acetophenone.26 The signs for D,, of B and BP are known1'J8 and given in Table 111. For a triplet system containing a nucleus with I = 3 / z (Cl), we calculated the possible zf transitions, in the same way as has been done for the triplet state of 4,4'-diexchlorobenzophenone and 4,4'-diiodoben~ophenone~~*~ cept that here only perturbation theory is used. Apart from allowed transitions (AmI = 0), also so-called "forbidden" transitions are expected, for which the electron and nuclear spin change simultaneously. In Table IV the results of this calculation are given. Figure 3a shows the Z-Y transition at 2583 MHz. It is gratifying that on increasing the microwave power additional peaks are observed (Figure 3b). The intensity ratio of the satellite lines at 2618 and 2610 and at 2549 and 2556 MHz is 3:l. If the nucleus is C1, this is in excellent agreement with the isotopic ratio between %C1and 37Cl. Furthermore this means, that the lines at 2618 and 2549 MHz can be attributed to a 35Clisotope and the lines at 2610 and 2556 MHz to a 37Cl isotope. Further analysis of the satellite lines supports this conclusion, From the difference of the two transition frequencies 5 and 8, neglecting the differences between a, p, and y (&'
