- Email: [email protected]
Electronic and vibrational states of carbonyl compounds—I
Spectrochimica Acta,1958,Vol. 12,pp. 78 to 8;. Pergamon PressLtd.,London
Electronic and vibrational states of carbonyl compounds-I Electronic states of camphorquinone R. A. FORD* and Miss F. PARRY? Research Department, British Nylon Spinners Limited, Pontypool, England (Received 28 October1957) Abstract-The absorption spectrum of camphorquinone has been recorded in a series of solvents at room temperature. The emission spectrum of crystalline camphorquinone has been recorded at 77°K and the lifetime of this emission in a rigid glass at 77°K has been observed to be 7.1 x 1e3 sec. The two main regions of absorption in the near ultra-violet and visible, Systems I and II, are assigned to the lA, -+ II’B, and ‘A, --f IIB, transitions respectively. It is suggested that the very weak visible absorption, System III, is lA, --+ 13B, and the sharp yellow emission either 13B,+ lA, or 13A,+ lA,. Solvent effects on Systems I and II are discussed in some detail since System I shows a small “red shift” in polar solvents although it is n + x in type. Modified rules for the effects of solvents on n -+ n transitions are suggested. 1.
hh0auOti0n
OF LATE, considerable
study has been made of the weak, long wavelength transitions in carbonyl compounds. In particular formaldehyde [l, 2, 3, 41, glyoxal [5] and biacetyl[6, 7, 81, have received detailed analyses. The main characteristics of these transitions, which involve the excitation of an electron from a ‘lone pair’ (2~) orbital localized on the oxygen atom to an extended, anti-bonding n MO., have been successfully accounted for by the simple LCAO. MO. theory developed by MULLIKEN and MCMURRY [9, lo], but it has not been found possible to extend this to include the related vibrational states [S, 111. Camphorquinone [I] owes its yellow colour to a weak absorption system extending from 20,000-25,000 cm-l which is similar in origin to the bands that cause glyoxal and biacetyl to be coloured. But since glyoxal and biacetyl are both normally in their planar trans-configurations [12], symmetry C,,, whereas camphorquinone is locked into the planar c&-configuration by the rigid ring system (intercarbonyl angle 8 = O-1 0” [ 131) it was felt that a comparison of these spectra might reveal additional information about this class of a-diketones. Since the intensity of electronic transitions in polyatomic molecules is controlled by the local symmetry, we shall consider the ground state of camphorquinone to be C,,. The infra-red and ultra-violet spectra of camphorquinone have been recorded previously by LEONARD and MADER [ 131 and by ALDER et al. [ 141, but we have been unable to find any previous analysis. In this investigation the absorption spectrum in a series of solvents at room temperature and the emission spectrum of the crystal at 77°K have been recorded. The lifetime of the emission at 77°K has been measured using a mechanical phosphoroscope. * Present address, Mullard Research Laboratories, Salfords, Redhill, Surrey, England.
t On vacation study from Dept. of Physics, University of Birmingham.
78
Electronic and vibrational states of carbonyl compounds-1
2. Experimental Camphorquinone, kindly prepared by Dr. A. G. PETO in these laboratories, was purified by recrystallization from petroleum ether followed by vacuum sublimation, and stored in the dark. The emission spectrum of the crystal at 77°K was measured on a Bellingham and Stanley wavelength spectrometer using a glass prism. The dispersion was 1.4 A/mm at 5000 A, and 2.7 A/ mm at 6000 A. The spectra were recorded photoelectrically with the sample only partially immersed in the liquid refrigerant to prevent interference from bubbles. A medium pressure mercury arc, chopped at 13 countslsec, was used for excitation with a Wood’s glass filter to select the 4047 A and 3660 b lines. This filter effectively removed the intense 4358 A, 5461 A and 5780 A source lines. Internal calibration of the records was obtained by an electrical ticker system which was subsequently calibrated against a Zn-Cd-Mg arc. The period of scanning from 5000-6000 A was 25 min. Lifetime measurements, in a rigid glass at 77”K, were made with a mechanical phosphoroscope consisting of a pair of semi-circular sectors mounted 180” out of phase on a common shaft. The detector was a photon-multiplier which was coupled to a double beam C.R.O. for recording purposes. Calibration of these records was obtained by feeding a 50 countslsec signal to the other C.R.O. trace. In this way lifetimes down to ri10e4 set could be resolved using a synchronous motor to turn the phosphoroscope sectors at 50 rev/set. Absorption spectra were recorded at room temperature using a Unicam S.P. 500 with 10, 40, and 100 mm cells. Eastman-Kodak “Spectra” grade n-hexane and methanol were used without further purification for this work. Analar acetic acid was purified by fractional distillation using a 10 in. column packed with & in. pentane and methyl cyclohexane used for the dia. glass helices. The 2,2,4trimethyl rigid glass were also Eastman-Kodak “Spectra” grade, again used without further purification.
3. Results Absorption
spectrum
Solutions of camphorquinone in n-hexane show three regions of absorption; System I extending from 29,000-40,000 cm-i and System II extending from 20,000-25,000 cm-l, which are shown in Fig. 1, and System III, a very weak transition, extending from 17,000-19,000 cm-l shown in Fig. 2. The similarity to the spectrum of biacetyl in n-hexane [14] is most striking. System I. The most notable feature of this transition is the well resolved vibrational structure of seven bands. The two longest wavelength bands have been labelled a and B as they appear from their spacing to be separate from the five higher energy components which are labelled a, b, c, d and e. Details of the separations and a tentative partial analysis are given in Table 1. This transition is very broad (half peak band width = 8000 cm-l) and the intensity increases towards higher energies. The low intensity oscillator strength, suggests that the transition is partially forbidden. 6,38 x lo”, System II. This transition also shows some vibrational structure but the detail is insuflicient for any analysis. It is much sharper than System I, its half-peak 79
R. A. FORD and Miss F. PARRY
band-width being 2500 cm-l and shows no degradation to the red. However the similarity of its oscillator strength, 4.2 x 10-4, to that of System I suggests that both transitions are of the same type. System III. This is a very weak transition only evident as a shoulder on the Comphorquinone (in N’ Hexace) 40
Fig.
1. Absorption spectrum of camphorquinone in n-hexane.
Since it is analogous in position to a weak steep onset of the System II absorption. singlet-triplet transition recently detected in solutions of biacetyl [S], the effect of perturbation by a “heavy atom” solvent [ 15, 161 was studied using et,hyl iodide. Some evidence of the expected enhancement is shown Fig. 2. Table 1. System I of camphorquinone
-
Band
-
v cm-l
29,500 31,250 33,000 34,480 35,680 37,050 38,530
Y -
in whexane Assignment
33,000
- 3500 -1750 0 1480 2680 4050 5530
2(1750) 1750 IA, - II lB,
-
Solvent e:Sfects
Systems I and II have been studied in n-hexane, diethyl ether, methanol and acetic acid as shown in Figs. 3 and 4. On going to more polar solvents the vibrational structure of System I is retained although it becomes more diffuse. especially in acetic acid. In ether and methanol small red-shifts are observed accompanied by a_small intensity increase. In 50 per cent acetic acid a clear intensity increase is observed, but the considerable change of band shape makes it difficult to be definite about the band shift. However it does appear that a small red-shift has occurred here also. Solvent effects on the corresponding transition of biacetyl do not seem to have been investigated [ 171. 80
Electrvnic
and vibrationa, states of carbonyl compounds-I
Fig. 9. Long wavelength absorption spectrum of camphorquinonei$df$ n-hexane, (b) ethyl
Fig. 4. System II of camphorquinone in different solvents; (a) n-hexane, (b) diethyl ether, (c) methanol, (d) 50 per oent acetic acid.
o”++
40 000
Fig. 3. System I of carn~ho~qu~o~e in different solvents; fa) n-hexfme, (b) diethyl ether, (c) methanol, (d) 50 per cent acetic acid.
R. A. FORD end Miss F. PARRY
System II shows larger solvent effects than System I and what vibrational structure exists in n-hexane is completely lost in methanol and acetic acid. In these solvents the transition shows considerable blue-shifts the magnitudes of which are given in Table 2. Table 2. Effects of solvents on System II of camphorquinone Solvent
Band centre (cm-l)
n-hexane diethyl ether methanol acetic acid
21,100 21,200 21,500 21,900
Shift (cm-l)
1
0 100 400 800
In ether and acetic acid there is little intensity change, but the intensity falls in methanol. Since the readings in methanol changed slowly with time, some of this reduction is certainly due to slow tautomeric [18] or chemical changes. Emission spectrum Crystals of camphorquinone at 77°K show a very sharp yellow emission extending from 16,500-18,000 cm-l (half-peak band-width 600 cm-l), which is shown in Fig. 5. There is some evidence of five small vibrational peaks showing separations
t cm-i’
Fig. 5. Emission spectrum of Cry8tallkWcamphorquinone at WK.
of 35-40 cm-l, but as the details varied between specimens, this may be an artifact. No evidence of larger vibrational separations in this transition or of any other emission region was found. The lifetime of the yellow emission measured in a rigid glass consisting of 8 parts methyl-cyclohexane to 1 part 2,2,4_trimethyl pentane at 77’K was 7-l x 1O-3 sec. The lifetime of biacetyl measured at the same time for comparison was 5.5 x 1O-3 sec. This is in reasonable agreement with MCCLURE’S [19] value of 2.25 x 1O-3 set obtained in E.P.A. at 77’K.
4. Discussion Assignment
of electronic transitions
The similarity between the absorption spectra of biacetyl and camphorquinone in n-hexane is such as to make a comparison between them profitable. SIDMAN 82
Electronic and vibrational states of carbonyl compounds-I
and MCCLURE [S] have recently set down the one electron L.C.A.O. M.0.s and Details of the related orbitals for camtheir derived electronic states for biacetyl. phorquinone are given in Table 3. The transitions to be expected from L.C.A.O. M.O. theory and the electronic states derived from them are given in Table 4. Table 3. Electronic orbit& L.C.A.O.
M.0.s.t
and symmetries for camphorquinone
(one electron)
77’2
(X0, + xc1 + q + X0,) b&J1+ xc1 - gc*- X0,)
n1
(Yol + YOJ
“2
(Yo,
-
Yo*)
773
(X&
-
xc1
-
q
+
XOJ
x4
PO1
-
xc1
+
xc2
-
X0,)
Tl
-j
+
LB,
+
-
)
A,
Tz
!_ 7 X and Y represent atomic 2p wave functions which are symmetric about the X and Y molecular axes respectively.
The contour of System II of camphorquinone in n-hexane shows a remarkable similarity to the transition in biacetyl extending from 21,000-27,000 cm-l. Both show three peaks; two closely spaced long wavelength components and a third at a larger separation to the blue. For sometime it was held that two electronic transitions were required to explain this region in biacetyl [20, 211, but SIDMAN and Table 4. Electronic transitions and excited states of camphorquinone
Transitions
Orbital configuration
I II I II I II
Electronic states
I’B,, 13B, IIlA 2, I13A 2 I’A,, 13A2 IIrB,, I13B, I’B,, 13B2 IPA,, IPA,
[6] have shown with a better vibrational analysis that there is no evidence Accordingly we shall consider for more than one transition being involved. System II to be a single electronic transition. As with biacetyl, the transition shows considerable “blue shifts” on going into more polar solvents, the values of which were given in Table 2. Clearly it should be assigned to an n --+ VTprocess of which the two lowest-energy singlet transitions are n, -+ rr3 and n2 + n3. In the absence of definite information on the origins of the absorption and emission transitions, a comparison of the intensity with biacetyl suggests that the transition is allowed by symmetry. In this case it is assigned ‘A, -+ I IB,. The low intensity is then due to the small overlap between the ground state and excited state wavefunctions.
MCCLURE
83
R. A.
FORD
and Miss F.
PARRY
The vibrational fine structure in System I is much more extensive than in System II, which suggests that the upper state of System I may be unstable. This agrees with the observation that solutions of camphorquinone undergo photochemical decomposition on illumination in this region. The bands labelled cc and /? appear at separations of 1750 cm-’ to the red of band a. These separations are almost exactly the size of the 1753 cm-l ground state carbonyl stretching frequency of camphorquinone. Therefore the simplest explanation is that u and B are “hot” bands associated with the 1 -+ 0 and 2 -+ 0 transitions from a ground state in which this frequency is excited. In this case band a must be the O-O band. It is not however one of the most prominent bands, the intensity of the transition increasing towards higher energies. This indicates that the dimensions of the upper and lower states must differ considerably. The effect of solvents on System I is of particular interest since it shows a small “red shift” on going into more polar solvents. This is usually considered conclusive evidence that the transition cannot be n + 7~in type [22]. If this is so the most likely transition is 7~~-+ ~a which is an allowed transition to a IlB, state. This should have a high intensity but the intensity of System I is low and very similar to that of System II. which is an n --f n process. L.C.A.O. M.O. theory predicts two n + n transitions in this region n, + ra and n2 ---f rr4 of which n2 -+ 7~~to an II’B, upper state is allowed by symmetry. The intensity of this may be expected to be in the region of that observed, and SIDMAN and MCCLURE [6] have assigned the corresponding absorption region in biacetyl extending from 31,000-42,000 cm-l to the analogous lA, + II1AU transition. We believe this assignment to be correct and the red-shift of this transition in polar solvents to be an exception to the n-rr blue-shift rule of MCCONNELL. We shall discuss the reasons for this in a later section on solvent effects. Accordingly we assign System I in camphorquinone to an lA, -+ IIlB, transition. Clear cut evidence on this point could be obtained by using polarized light since the 7~+ 7r process will be Y-polarized but the n --t 7~ process X-axis polarized. We must now consider System III in relation to the yellow emission transition since the small enhancement of System III in ethyl iodide and the lifetime of 7.1 x 1O-3 set for the emission show that both are intercombination processes. SID&~ANand MCCLURE [6] have found this region in biacetyl to be complex. They suggest that the weak visible absorption, which has recently been detected in solution by FORSTER [8], is lA, -+ 13A, whilst the strong green emission is 13B, -+ lAg, radiationless conversion from the 13A, - 13B, states being of very high effiThus with the limited data available for camphorquinone it would be ciency. unwise to attempt anything but a tentative assignment. However, we can safely assume that we are dealing with n -+ n processes in which case the absorption transition is almost certainly the allowed IA, ---t 13B, transition. The emission may be 13B, - IA, or PA, - IA,. It is worth noting that, as with biacetyl, the separation between the singlet and triplet states arising from the first excited n + T configuration is about 2500-3000 cm-l, which is much greater than the value predicted by REID [23] from overlap considerations. Also biacetyl and camphorquinone resemble each other in emitting mainly phosphorescence (triplet + singlet) whereas with glyoxal the predominant emission is fluorescence (singlet -+ 84
Electronic
and vibrational
states of carbonyl
compounds-I
singlet). This is probably due to the two CH, groups in biacetyl and the rigid ring system in camphorquinone lowering the frequency of, and therefore increasing, the population of excited torsional and bending states (of the intercarbonyl C-C bond) in their respective molecules. Since the slopes of the potential energy curves of the ground state and the first excited singlet or triplet states in these molecules, plotted as a function of twist about the C-C bond, are opposite in sign, [24] perturbations of this kind will de&ease the overlap and paired character of the n-electrons thereby mixing a certain amount of triplet character into the ground state wavefunction [%I. Solvent
efects
MCC'ONNELL [22] has proposed the rule that “electronic transitions in molecules which have low intensities, because the ground and excited state charge distributions do not overlap strongly, are likely to give blue-shift bands.” The combination of these intensity and solvent shift criteria has become standard practice However with camphorquinone, System I shows a for assigning n -3 7r transitions. small red-shift in polar solvents although the transition is almost certainly n -+ 7r in type. We believe that the reason for this is solvent interact’ion modifying the n-electron resonance energy of the solute molecule. Jf w-e consider the two carbonyl groups of camphorquinone separately their 7r M.0.s may be written:
Ground state: Excited
41 = aXcl + bXP1 r&* = aXcl -
state:
bX$
$2 = aXcz + bX$ +2* = aXcz -
bXo2
Our reason for choosing this description is that changes in the n-electron resonance energy due to deviations from planarity are most likely to result from changes in the n-electron distribution over the intercarbonyl C-C bond, since the coupling here is wea,ker than in the C=O bonds. Then the splitting between the IlB, and II lB, states of the molecule will be the energy difference between the orbitals
V’?r4= (A* Since (4,. +& and (+1*, &*)
+ A*)
and
yn3 = (A*
are pairs of degenerate
-A*)
orbitals
where 8 = J+,*$,* d T and b is the resonance integral. Because of the (1 & 8) term in these energy values, En4 should exceed E,* by a lesser amount than En3 lies below it. This is exactly what is found with camphorquinone, the splitting between Systems I and II being clearly asymmetric with respect to the 29.500 cm-l transition in acetaldehyde. The magnitude of this splitting will depend on the value of the overlap integral 8, which will be reduced by any deviations from planarity since S = a0 cos 13,where 8 is the intercarbonyl angle and a0 is some function of 8. The effect of such reductions in splitting on the spectrum of camphorquinone will be to shift System I more slowly to the red than System II will shift to the blue as is in fact observed. This is shown diagrammatically in Fig. 6. 85
33. A. FORD
and Miss F.
PARRY
The possible ways in whioh solvent interaction could produce a perturbation of this type axe either by mass coupling through interaotion at the oxygen atoms which would flatten the potential curve for twist around the intercarbonyl G-C bond of the ground state thereby increasing the population of excited torsional states at room temperature, or by steric hindrance between solvent molecules
Fig. 6, XXectronic orbital energies of camphorquinone as a function of twist around the iutercarbouyl C-C bond.
~n~~a~t~~g with the adjacent carbonyl groups. Bore work is required before a clear decision between these can be made. However from the difference between camphorquinone and glyoxal, where both the analogous transitions show blue shifts in polar solvents [22], it appears that steric hindrance may be the cause. This will assume importance only for the cis configuration. The observed shifts of it -+ r transitions on going into polar solvents are of course the summation of a series of effoots. In most oases the blue-shift due to the stabilization of the n-electrons in the ground state is the predominant effect. However, where solvent interaction can modify the v electron resonance energy of the solute, exceptions to this may occur. We may summarize the effects to be expected in this case as follows. i
(a) Excitation to ~~-orbitA with antinoae in the region of the overlap change, (b) Excitation to rr-orbital with node in the regi#n of the overlap change.
!
! ’ IWd-Shift~
1
blue-shift
/ ; “blue-shift
1
red-shift
Electronic and vibrational states of cerbonyl compounds-I ~4ek-nourZedgementhe
authors wish to thank British Nylon
Spinners Limited
for permission
to publish this work.
References [l] [2] [3] [4] [5] [6] [7] [8] [9] [lo] [Ill [12]
[13] [la] [I51 [16]
[17] [18] [19] [20] [21] [22] [23]
[24] [25]
BRAND J. C. D. J. Chem. Sot. 1956 868. BRAND J. C. D. and REED R. I. J. Chem. Sot. 1957 2386. WALSH A. D. J. Chem. Sot. 1953 2306. REID C. and COHEN A. D. J. Chem. Phy8. 1966 24 85. BRAND J. C. D. Trans. Faraday Sot. 1954 50 431. SIDMAN J. W. and MCCLURE D. S. J. Amer. Chem. Sot. 1955 77 6461, 6471. MIYAZAWA T. J. C&m. Sot. Japan 1953 74 743. FORSTER L. S. J. Chem. Phys. 1957 26 1761. MCMURRY H. L. and MULLIKEN R. S. Proc. Nut. Ad. Sk., Wash. 1940 26 312. MCMURRY H. L. J. Chem. Phys. 1941 9 231, 241. CANNON C. G. and FORD R. A. European Molecular Spectroscopy Conference, Freiburgim-Breisgau, July 1957. J. Amer. Chem. Sot. 1939 61 3520. LUVALLE J. E. and SCHOMAKERV. LEONARD N. J. and MADER P. M. J. Amer. Chem. Sot. 1950 72 5388. ALDER K. et al. Liebigs Ann. 1955 593 23. KASHA M. J. Chem. Phya. 1952 20 71. PLATT J. R. J. Opt. Sot. Amer. 1953 43 252. HOLMAN R., LUNDBERG W. and BURR G. J. Amer. Chem. Sot. 1945 67 1669. MODIANO J. and PARIAUD J. C. Bull. Sot. Chim. BY. 1952 19 642. MCCLURE D. S. J. Chem. Phys. 1949 17 905. Lqwrs G. N. and KASHA M. J. Amer. Chem. Sot. 1945 87 994. FORSTER L. S. J. Amer. Chem. Sot. 1955 77 1417. MCCONNELL H. J. Chem. Phys. 1952 20 700. REID C. J. Chem. Phys. 1953 21 1906. MULLIKEN R. S. end ROOTHAAN C. C. J. Chem. Rev. 1947 41 219. MCCONNELL H. J. Chew. Phys. 1952 20 1043.
87
Electronic and vibrational states of carbonyl compounds-I Electronic states of camphorquinone R. A. FORD* and Miss F. PARRY? Research Department, British Nylon Spinners Limited, Pontypool, England (Received 28 October1957) Abstract-The absorption spectrum of camphorquinone has been recorded in a series of solvents at room temperature. The emission spectrum of crystalline camphorquinone has been recorded at 77°K and the lifetime of this emission in a rigid glass at 77°K has been observed to be 7.1 x 1e3 sec. The two main regions of absorption in the near ultra-violet and visible, Systems I and II, are assigned to the lA, -+ II’B, and ‘A, --f IIB, transitions respectively. It is suggested that the very weak visible absorption, System III, is lA, --+ 13B, and the sharp yellow emission either 13B,+ lA, or 13A,+ lA,. Solvent effects on Systems I and II are discussed in some detail since System I shows a small “red shift” in polar solvents although it is n + x in type. Modified rules for the effects of solvents on n -+ n transitions are suggested. 1.
hh0auOti0n
OF LATE, considerable
study has been made of the weak, long wavelength transitions in carbonyl compounds. In particular formaldehyde [l, 2, 3, 41, glyoxal [5] and biacetyl[6, 7, 81, have received detailed analyses. The main characteristics of these transitions, which involve the excitation of an electron from a ‘lone pair’ (2~) orbital localized on the oxygen atom to an extended, anti-bonding n MO., have been successfully accounted for by the simple LCAO. MO. theory developed by MULLIKEN and MCMURRY [9, lo], but it has not been found possible to extend this to include the related vibrational states [S, 111. Camphorquinone [I] owes its yellow colour to a weak absorption system extending from 20,000-25,000 cm-l which is similar in origin to the bands that cause glyoxal and biacetyl to be coloured. But since glyoxal and biacetyl are both normally in their planar trans-configurations [12], symmetry C,,, whereas camphorquinone is locked into the planar c&-configuration by the rigid ring system (intercarbonyl angle 8 = O-1 0” [ 131) it was felt that a comparison of these spectra might reveal additional information about this class of a-diketones. Since the intensity of electronic transitions in polyatomic molecules is controlled by the local symmetry, we shall consider the ground state of camphorquinone to be C,,. The infra-red and ultra-violet spectra of camphorquinone have been recorded previously by LEONARD and MADER [ 131 and by ALDER et al. [ 141, but we have been unable to find any previous analysis. In this investigation the absorption spectrum in a series of solvents at room temperature and the emission spectrum of the crystal at 77°K have been recorded. The lifetime of the emission at 77°K has been measured using a mechanical phosphoroscope. * Present address, Mullard Research Laboratories, Salfords, Redhill, Surrey, England.
t On vacation study from Dept. of Physics, University of Birmingham.
78
Electronic and vibrational states of carbonyl compounds-1
2. Experimental Camphorquinone, kindly prepared by Dr. A. G. PETO in these laboratories, was purified by recrystallization from petroleum ether followed by vacuum sublimation, and stored in the dark. The emission spectrum of the crystal at 77°K was measured on a Bellingham and Stanley wavelength spectrometer using a glass prism. The dispersion was 1.4 A/mm at 5000 A, and 2.7 A/ mm at 6000 A. The spectra were recorded photoelectrically with the sample only partially immersed in the liquid refrigerant to prevent interference from bubbles. A medium pressure mercury arc, chopped at 13 countslsec, was used for excitation with a Wood’s glass filter to select the 4047 A and 3660 b lines. This filter effectively removed the intense 4358 A, 5461 A and 5780 A source lines. Internal calibration of the records was obtained by an electrical ticker system which was subsequently calibrated against a Zn-Cd-Mg arc. The period of scanning from 5000-6000 A was 25 min. Lifetime measurements, in a rigid glass at 77”K, were made with a mechanical phosphoroscope consisting of a pair of semi-circular sectors mounted 180” out of phase on a common shaft. The detector was a photon-multiplier which was coupled to a double beam C.R.O. for recording purposes. Calibration of these records was obtained by feeding a 50 countslsec signal to the other C.R.O. trace. In this way lifetimes down to ri10e4 set could be resolved using a synchronous motor to turn the phosphoroscope sectors at 50 rev/set. Absorption spectra were recorded at room temperature using a Unicam S.P. 500 with 10, 40, and 100 mm cells. Eastman-Kodak “Spectra” grade n-hexane and methanol were used without further purification for this work. Analar acetic acid was purified by fractional distillation using a 10 in. column packed with & in. pentane and methyl cyclohexane used for the dia. glass helices. The 2,2,4trimethyl rigid glass were also Eastman-Kodak “Spectra” grade, again used without further purification.
3. Results Absorption
spectrum
Solutions of camphorquinone in n-hexane show three regions of absorption; System I extending from 29,000-40,000 cm-i and System II extending from 20,000-25,000 cm-l, which are shown in Fig. 1, and System III, a very weak transition, extending from 17,000-19,000 cm-l shown in Fig. 2. The similarity to the spectrum of biacetyl in n-hexane [14] is most striking. System I. The most notable feature of this transition is the well resolved vibrational structure of seven bands. The two longest wavelength bands have been labelled a and B as they appear from their spacing to be separate from the five higher energy components which are labelled a, b, c, d and e. Details of the separations and a tentative partial analysis are given in Table 1. This transition is very broad (half peak band width = 8000 cm-l) and the intensity increases towards higher energies. The low intensity oscillator strength, suggests that the transition is partially forbidden. 6,38 x lo”, System II. This transition also shows some vibrational structure but the detail is insuflicient for any analysis. It is much sharper than System I, its half-peak 79
R. A. FORD and Miss F. PARRY
band-width being 2500 cm-l and shows no degradation to the red. However the similarity of its oscillator strength, 4.2 x 10-4, to that of System I suggests that both transitions are of the same type. System III. This is a very weak transition only evident as a shoulder on the Comphorquinone (in N’ Hexace) 40
Fig.
1. Absorption spectrum of camphorquinone in n-hexane.
Since it is analogous in position to a weak steep onset of the System II absorption. singlet-triplet transition recently detected in solutions of biacetyl [S], the effect of perturbation by a “heavy atom” solvent [ 15, 161 was studied using et,hyl iodide. Some evidence of the expected enhancement is shown Fig. 2. Table 1. System I of camphorquinone
-
Band
-
v cm-l
29,500 31,250 33,000 34,480 35,680 37,050 38,530
Y -
in whexane Assignment
33,000
- 3500 -1750 0 1480 2680 4050 5530
2(1750) 1750 IA, - II lB,
-
Solvent e:Sfects
Systems I and II have been studied in n-hexane, diethyl ether, methanol and acetic acid as shown in Figs. 3 and 4. On going to more polar solvents the vibrational structure of System I is retained although it becomes more diffuse. especially in acetic acid. In ether and methanol small red-shifts are observed accompanied by a_small intensity increase. In 50 per cent acetic acid a clear intensity increase is observed, but the considerable change of band shape makes it difficult to be definite about the band shift. However it does appear that a small red-shift has occurred here also. Solvent effects on the corresponding transition of biacetyl do not seem to have been investigated [ 171. 80
Electrvnic
and vibrationa, states of carbonyl compounds-I
Fig. 9. Long wavelength absorption spectrum of camphorquinonei$df$ n-hexane, (b) ethyl
Fig. 4. System II of camphorquinone in different solvents; (a) n-hexane, (b) diethyl ether, (c) methanol, (d) 50 per oent acetic acid.
o”++
40 000
Fig. 3. System I of carn~ho~qu~o~e in different solvents; fa) n-hexfme, (b) diethyl ether, (c) methanol, (d) 50 per cent acetic acid.
R. A. FORD end Miss F. PARRY
System II shows larger solvent effects than System I and what vibrational structure exists in n-hexane is completely lost in methanol and acetic acid. In these solvents the transition shows considerable blue-shifts the magnitudes of which are given in Table 2. Table 2. Effects of solvents on System II of camphorquinone Solvent
Band centre (cm-l)
n-hexane diethyl ether methanol acetic acid
21,100 21,200 21,500 21,900
Shift (cm-l)
1
0 100 400 800
In ether and acetic acid there is little intensity change, but the intensity falls in methanol. Since the readings in methanol changed slowly with time, some of this reduction is certainly due to slow tautomeric [18] or chemical changes. Emission spectrum Crystals of camphorquinone at 77°K show a very sharp yellow emission extending from 16,500-18,000 cm-l (half-peak band-width 600 cm-l), which is shown in Fig. 5. There is some evidence of five small vibrational peaks showing separations
t cm-i’
Fig. 5. Emission spectrum of Cry8tallkWcamphorquinone at WK.
of 35-40 cm-l, but as the details varied between specimens, this may be an artifact. No evidence of larger vibrational separations in this transition or of any other emission region was found. The lifetime of the yellow emission measured in a rigid glass consisting of 8 parts methyl-cyclohexane to 1 part 2,2,4_trimethyl pentane at 77’K was 7-l x 1O-3 sec. The lifetime of biacetyl measured at the same time for comparison was 5.5 x 1O-3 sec. This is in reasonable agreement with MCCLURE’S [19] value of 2.25 x 1O-3 set obtained in E.P.A. at 77’K.
4. Discussion Assignment
of electronic transitions
The similarity between the absorption spectra of biacetyl and camphorquinone in n-hexane is such as to make a comparison between them profitable. SIDMAN 82
Electronic and vibrational states of carbonyl compounds-I
and MCCLURE [S] have recently set down the one electron L.C.A.O. M.0.s and Details of the related orbitals for camtheir derived electronic states for biacetyl. phorquinone are given in Table 3. The transitions to be expected from L.C.A.O. M.O. theory and the electronic states derived from them are given in Table 4. Table 3. Electronic orbit& L.C.A.O.
M.0.s.t
and symmetries for camphorquinone
(one electron)
77’2
(X0, + xc1 + q + X0,) b&J1+ xc1 - gc*- X0,)
n1
(Yol + YOJ
“2
(Yo,
-
Yo*)
773
(X&
-
xc1
-
q
+
XOJ
x4
PO1
-
xc1
+
xc2
-
X0,)
Tl
-j
+
LB,
+
-
)
A,
Tz
!_ 7 X and Y represent atomic 2p wave functions which are symmetric about the X and Y molecular axes respectively.
The contour of System II of camphorquinone in n-hexane shows a remarkable similarity to the transition in biacetyl extending from 21,000-27,000 cm-l. Both show three peaks; two closely spaced long wavelength components and a third at a larger separation to the blue. For sometime it was held that two electronic transitions were required to explain this region in biacetyl [20, 211, but SIDMAN and Table 4. Electronic transitions and excited states of camphorquinone
Transitions
Orbital configuration
I II I II I II
Electronic states
I’B,, 13B, IIlA 2, I13A 2 I’A,, 13A2 IIrB,, I13B, I’B,, 13B2 IPA,, IPA,
[6] have shown with a better vibrational analysis that there is no evidence Accordingly we shall consider for more than one transition being involved. System II to be a single electronic transition. As with biacetyl, the transition shows considerable “blue shifts” on going into more polar solvents, the values of which were given in Table 2. Clearly it should be assigned to an n --+ VTprocess of which the two lowest-energy singlet transitions are n, -+ rr3 and n2 + n3. In the absence of definite information on the origins of the absorption and emission transitions, a comparison of the intensity with biacetyl suggests that the transition is allowed by symmetry. In this case it is assigned ‘A, -+ I IB,. The low intensity is then due to the small overlap between the ground state and excited state wavefunctions.
MCCLURE
83
R. A.
FORD
and Miss F.
PARRY
The vibrational fine structure in System I is much more extensive than in System II, which suggests that the upper state of System I may be unstable. This agrees with the observation that solutions of camphorquinone undergo photochemical decomposition on illumination in this region. The bands labelled cc and /? appear at separations of 1750 cm-’ to the red of band a. These separations are almost exactly the size of the 1753 cm-l ground state carbonyl stretching frequency of camphorquinone. Therefore the simplest explanation is that u and B are “hot” bands associated with the 1 -+ 0 and 2 -+ 0 transitions from a ground state in which this frequency is excited. In this case band a must be the O-O band. It is not however one of the most prominent bands, the intensity of the transition increasing towards higher energies. This indicates that the dimensions of the upper and lower states must differ considerably. The effect of solvents on System I is of particular interest since it shows a small “red shift” on going into more polar solvents. This is usually considered conclusive evidence that the transition cannot be n + 7~in type [22]. If this is so the most likely transition is 7~~-+ ~a which is an allowed transition to a IlB, state. This should have a high intensity but the intensity of System I is low and very similar to that of System II. which is an n --f n process. L.C.A.O. M.O. theory predicts two n + n transitions in this region n, + ra and n2 ---f rr4 of which n2 -+ 7~~to an II’B, upper state is allowed by symmetry. The intensity of this may be expected to be in the region of that observed, and SIDMAN and MCCLURE [6] have assigned the corresponding absorption region in biacetyl extending from 31,000-42,000 cm-l to the analogous lA, + II1AU transition. We believe this assignment to be correct and the red-shift of this transition in polar solvents to be an exception to the n-rr blue-shift rule of MCCONNELL. We shall discuss the reasons for this in a later section on solvent effects. Accordingly we assign System I in camphorquinone to an lA, -+ IIlB, transition. Clear cut evidence on this point could be obtained by using polarized light since the 7~+ 7r process will be Y-polarized but the n --t 7~ process X-axis polarized. We must now consider System III in relation to the yellow emission transition since the small enhancement of System III in ethyl iodide and the lifetime of 7.1 x 1O-3 set for the emission show that both are intercombination processes. SID&~ANand MCCLURE [6] have found this region in biacetyl to be complex. They suggest that the weak visible absorption, which has recently been detected in solution by FORSTER [8], is lA, -+ 13A, whilst the strong green emission is 13B, -+ lAg, radiationless conversion from the 13A, - 13B, states being of very high effiThus with the limited data available for camphorquinone it would be ciency. unwise to attempt anything but a tentative assignment. However, we can safely assume that we are dealing with n -+ n processes in which case the absorption transition is almost certainly the allowed IA, ---t 13B, transition. The emission may be 13B, - IA, or PA, - IA,. It is worth noting that, as with biacetyl, the separation between the singlet and triplet states arising from the first excited n + T configuration is about 2500-3000 cm-l, which is much greater than the value predicted by REID [23] from overlap considerations. Also biacetyl and camphorquinone resemble each other in emitting mainly phosphorescence (triplet + singlet) whereas with glyoxal the predominant emission is fluorescence (singlet -+ 84
Electronic
and vibrational
states of carbonyl
compounds-I
singlet). This is probably due to the two CH, groups in biacetyl and the rigid ring system in camphorquinone lowering the frequency of, and therefore increasing, the population of excited torsional and bending states (of the intercarbonyl C-C bond) in their respective molecules. Since the slopes of the potential energy curves of the ground state and the first excited singlet or triplet states in these molecules, plotted as a function of twist about the C-C bond, are opposite in sign, [24] perturbations of this kind will de&ease the overlap and paired character of the n-electrons thereby mixing a certain amount of triplet character into the ground state wavefunction [%I. Solvent
efects
MCC'ONNELL [22] has proposed the rule that “electronic transitions in molecules which have low intensities, because the ground and excited state charge distributions do not overlap strongly, are likely to give blue-shift bands.” The combination of these intensity and solvent shift criteria has become standard practice However with camphorquinone, System I shows a for assigning n -3 7r transitions. small red-shift in polar solvents although the transition is almost certainly n -+ 7r in type. We believe that the reason for this is solvent interact’ion modifying the n-electron resonance energy of the solute molecule. Jf w-e consider the two carbonyl groups of camphorquinone separately their 7r M.0.s may be written:
Ground state: Excited
41 = aXcl + bXP1 r&* = aXcl -
state:
bX$
$2 = aXcz + bX$ +2* = aXcz -
bXo2
Our reason for choosing this description is that changes in the n-electron resonance energy due to deviations from planarity are most likely to result from changes in the n-electron distribution over the intercarbonyl C-C bond, since the coupling here is wea,ker than in the C=O bonds. Then the splitting between the IlB, and II lB, states of the molecule will be the energy difference between the orbitals
V’?r4= (A* Since (4,. +& and (+1*, &*)
+ A*)
and
yn3 = (A*
are pairs of degenerate
-A*)
orbitals
where 8 = J+,*$,* d T and b is the resonance integral. Because of the (1 & 8) term in these energy values, En4 should exceed E,* by a lesser amount than En3 lies below it. This is exactly what is found with camphorquinone, the splitting between Systems I and II being clearly asymmetric with respect to the 29.500 cm-l transition in acetaldehyde. The magnitude of this splitting will depend on the value of the overlap integral 8, which will be reduced by any deviations from planarity since S = a0 cos 13,where 8 is the intercarbonyl angle and a0 is some function of 8. The effect of such reductions in splitting on the spectrum of camphorquinone will be to shift System I more slowly to the red than System II will shift to the blue as is in fact observed. This is shown diagrammatically in Fig. 6. 85
33. A. FORD
and Miss F.
PARRY
The possible ways in whioh solvent interaction could produce a perturbation of this type axe either by mass coupling through interaotion at the oxygen atoms which would flatten the potential curve for twist around the intercarbonyl G-C bond of the ground state thereby increasing the population of excited torsional states at room temperature, or by steric hindrance between solvent molecules
Fig. 6, XXectronic orbital energies of camphorquinone as a function of twist around the iutercarbouyl C-C bond.
~n~~a~t~~g with the adjacent carbonyl groups. Bore work is required before a clear decision between these can be made. However from the difference between camphorquinone and glyoxal, where both the analogous transitions show blue shifts in polar solvents [22], it appears that steric hindrance may be the cause. This will assume importance only for the cis configuration. The observed shifts of it -+ r transitions on going into polar solvents are of course the summation of a series of effoots. In most oases the blue-shift due to the stabilization of the n-electrons in the ground state is the predominant effect. However, where solvent interaction can modify the v electron resonance energy of the solute, exceptions to this may occur. We may summarize the effects to be expected in this case as follows. i
(a) Excitation to ~~-orbitA with antinoae in the region of the overlap change, (b) Excitation to rr-orbital with node in the regi#n of the overlap change.
!
! ’ IWd-Shift~
1
blue-shift
/ ; “blue-shift
1
red-shift
Electronic and vibrational states of cerbonyl compounds-I ~4ek-nourZedgementhe
authors wish to thank British Nylon
Spinners Limited
for permission
to publish this work.
References [l] [2] [3] [4] [5] [6] [7] [8] [9] [lo] [Ill [12]
[13] [la] [I51 [16]
[17] [18] [19] [20] [21] [22] [23]
[24] [25]
BRAND J. C. D. J. Chem. Sot. 1956 868. BRAND J. C. D. and REED R. I. J. Chem. Sot. 1957 2386. WALSH A. D. J. Chem. Sot. 1953 2306. REID C. and COHEN A. D. J. Chem. Phy8. 1966 24 85. BRAND J. C. D. Trans. Faraday Sot. 1954 50 431. SIDMAN J. W. and MCCLURE D. S. J. Amer. Chem. Sot. 1955 77 6461, 6471. MIYAZAWA T. J. C&m. Sot. Japan 1953 74 743. FORSTER L. S. J. Chem. Phys. 1957 26 1761. MCMURRY H. L. and MULLIKEN R. S. Proc. Nut. Ad. Sk., Wash. 1940 26 312. MCMURRY H. L. J. Chem. Phys. 1941 9 231, 241. CANNON C. G. and FORD R. A. European Molecular Spectroscopy Conference, Freiburgim-Breisgau, July 1957. J. Amer. Chem. Sot. 1939 61 3520. LUVALLE J. E. and SCHOMAKERV. LEONARD N. J. and MADER P. M. J. Amer. Chem. Sot. 1950 72 5388. ALDER K. et al. Liebigs Ann. 1955 593 23. KASHA M. J. Chem. Phya. 1952 20 71. PLATT J. R. J. Opt. Sot. Amer. 1953 43 252. HOLMAN R., LUNDBERG W. and BURR G. J. Amer. Chem. Sot. 1945 67 1669. MODIANO J. and PARIAUD J. C. Bull. Sot. Chim. BY. 1952 19 642. MCCLURE D. S. J. Chem. Phys. 1949 17 905. Lqwrs G. N. and KASHA M. J. Amer. Chem. Sot. 1945 87 994. FORSTER L. S. J. Amer. Chem. Sot. 1955 77 1417. MCCONNELL H. J. Chem. Phys. 1952 20 700. REID C. J. Chem. Phys. 1953 21 1906. MULLIKEN R. S. end ROOTHAAN C. C. J. Chem. Rev. 1947 41 219. MCCONNELL H. J. Chew. Phys. 1952 20 1043.
87