Reprinted from THE JOURN AL OF CHEMICAL PHYSICS, Vol. 48, No.2, 749-753, IS Janua,ry 1968 Printed in U. S. A.

Spectroscopy of EDA Complexes at High Pressures. VI. Absorption and Fluorescence of Crystalline TNB and TNF Complexes*

A. H . KADHIM AND H. W. OFFEN KADH AH67 0765

Department of Chemistry, University of California, Santa Barbara, California

(Received 11 September 1967)

Several aromatic hydrocarbons are complexed with trinitrobenzene (TNB) and trinitrofluorenone (TNF) in the crystalline state and subjected to increasing pressures (0-30 kbar). The charge transfer absorption shows a ",1200 em- I red shift, while the fluorescence red shift may vary between 200-1700 em-I at 25 kbar. In the same complexes the Stokes shift decreases at higher pressures. This decrease can be understood in terms of charge-transfer forces and potential energy curves appropriate for weak EDA complexes.

INTRODUCTION

One fruitful approach to the understanding of charge­transfer (CT) interaction in weak molecular complexes (EDA complexes) has been crystal spectroscopy at low temperatures. In this connection, single crystals of anthracene-trinitrobenzene have received considerable attention.l-4 Few other spectroscopic data5•8 are available on single crystals of EDA complexes contain­ing sym-trinitrobenzene (TNB) or 2,4,7-trinitro­fluorenone (TNF) as the acceptors and aromatic hydrocarbons as the donors. The crystal structure, which has been detennined for a ntlmber of EDA com­plexes in this group,7 as well as thermodynamic data8

support the interpretation that the molecular crystal consists of weak 1: 1 molecular pairs, so that to a first approximation we can discuss the crystal spectra in terms of the isolated molecular complexes.

This work uses pressure perturbations to study charge-transfer interactions in crystalline molecular complexes. Since the present high-pressure apparatus9

is unable to provide purely hydrostatic pressures, this technique is limited to polycrystalline materials, which for light transmission are best dispersed in alkali halides.10 The sodium chloride pressed pellet technique has been discussed in Part V of this series.lI Absorption and fluorescence spectra of TNB and TNF complexes in

• This work has been supported in part by the U.S. Office of Naval Research.

1 S. K. Lower, R. M. Hochstrasser, and C. Reid, Mol. Phys. 4, 161 (1961).

2 R. M. Hochstrasser, S. K. Lower, and C. Reid, J. Chern. Phys. 41, 1073 (1964).

a R. M. Hochstrasser, S. K. Lower, and C. Reid, ]. Mol. Spectry. 15, 257 (1965).

4 J . Tanaka and K. Yoshihara, Bull. Chern. Soc. Japan 38, 739 (1965).

r; S. K. Lower, Ph.D. thesis, University of British ColUDlbia, 1963.

6 H. Kuroda, T . Kunii, S. Hiroma, and H. Akamatu, J. Mol. Spectry. 22, 60 (1967).

7 D. S. Brown, S. C. Wallwork, and A. Wilson, Acta Cryst. 17, 168 (1964); S. C. Wallwork, J. Chern. Soc. 1961, 494; S. C. Wallwork, ibid. 7, 648 (1964).

8 D. Hammick and H. Hutchison, J . Chern. Soc. 1955, 89. I H. W. Offen, J . Chern. Phys. 42, 430 (1965). 10 E. A. Chandross and J. Ferguson, J. Chern. Phys. 45, 3564

(1966) . 11 A. H. Kadhim and H. W. Offen, (unpublished).

fused salt pellets12- 15 have been reported. High-pressure studies of this class of compounds have been limited to anthracene-TNB.9 Such features as the vibrational structure in CT bands, the lower intensities and the blue shift of the La band of complexed anthracene, which were observed at atmospheric pressure in pellets,9 have been confirmed by single-crystal data.1-3 This is sur­prising when the limitations of the pellet technique are considered and suggest that the observed pressure ef­fects on CT transitions in polycrystalline material are in one-to-one correspondence to the effects that would be found for single crystals. Luminescence intensities are very sensitive to crystal imperfections and are excluded from the previous statement. CT fluores­cence at high pressures was first reported in Part IIll6 for TCPA complexes in plastics. The high-pressurt ab­sorption17 and fluorescence18 techniques, as well as the pellet preparations,lI have been described.

PRESSURE EFFECT ON THE STOKES SHIFT

The pressure shifts in the CT absorption of EDA complexes in solid solutions or in the crystalline form have been discussed.ll •19 At present we are concerned with the pressure influence on the Stokes shift, i.e., the separation of the absorption (VCTa) and fluorescence (vd) band maxima at higher pressures. One-compo­nent 71' systems in solid solution show a difference in the pressure dependence of the 0, 0 band in the long-wave­length absorption !::.pG and in the reverse fluorescence f:.vI. 18 •20 This observation can be rationalized in terms of the solvation energies appropriate for the equilibrium

12 J. Czekalla, A. Schmillen, and K. J. Mager, Z. Elektrochem. 61,1053 (1957); 63, 623 (1959).

18 B. L. Van Duuren and C. E. Bardi, Anal. Chern. 35, 2198 (1963) .

14 M. J. S. Dewar and A. R. Lepley, J. Am. Chern. Soc. 83, 4509 (1961); A. R. Lepley, ibid. 84, 3577 (1962).

15 H. Kuroda, K. Yoshihara, and H. Akamatu, Bull. Chern. Soc. Japan 35, 1604 (1962).

iSH. W. Offen and J. F. Studebaker, J. Chern. Phys. 47,253 (1967) .

17 H. W. Offen and A. H. Kadhim, J. Chern. Phys. 45, 269 (1966) .

18 H. W. Offen and R. R. Eliason, J. Chern. Phys. 43, 4096 (1965) .

11 H. W. Offen and T. T. Nakashima, J. Chern. Phys. 47, ~ (1967) •

JO M. Nicol, J. Arn. Opt. Soc. 55,1176 (1965). 749

750 A. H. KADHIM AND H. W. OFFEN

u ./Co

~f(p)

FIG. 1. Potential-energy curves for weak EDA complexes showing the effect of pressure on transition energies.

initial electronic state of the solute. In general, the solvation energy of the excited state s' is not the same for the absorption and emission because in the former process the solute is found in the Franck- Condon state, from which it relaxes to an equilibrium configuration during the lifetime of the excited state and before emis­sion to the Franck-Condon ground state. Therefore, Av"=A(s,," -s,,') and Av' =A(s/'-s/) , and a pressure dependence of the Stokes shift is expected, provided the difference in solvation energies for the two processes is appreciable. For the usual case, where the polariza­bility and induced-dipole moment are greater in the excited 71'* state than in the ground state, solvent or dielectric effects predict

(1)

The arguments for the increasing Stokes shift for the 0, 0 band can be extended to the respective band maxima, if the potential energy curves are similar for the two states. Experiment confirms Eq. (1) for anthra­cene20 and pyrene.18

In analogy to discussions of vapor-solution shifts,21 the relative effect of pressure on the CT absorption and fluorescence band maxima, i.e., Stokes shift, can be dis­cussed in terms of potential-energy curves22 appropriate for the electronic energy of the molecular complex and the surrounding solvent in its equilibrium configuration. Figure 1 shows a diagram for the if;N and if;E states of the complex with the solvent cage appropriate for if;N. A similar pair of curves, vertically and unequally displaced, exists when the solvent cage for the CT state (if;E) is considered. For simplicity, these are not included in Fig. 1. The transitions are indicated for 0

21 J. Prochorow and A. Tramer, J. Chern. Phys. 44, 4545 (1966) .

22 S. P. McGlynn and J. D. Boggus, J. Am. Chern. Soc. 80, 5096 (1958).

kbar and some higher pressure P. It is seen that ARDA! < ARDA" due to the higher compressibility of the ground state. Both absorption and fluorescence spectra are displaced red for weak molecular complexes. The fact that RDA changes less in fluorescence than in ab­sorption at a given isobar as well as the difference in the slopes of the final states (dUEldR>dUNldR) con­tribute to

(2)

Then increasing pressure produces a stronger complex with a smaller Stokes shift.23

The pressure shift for the CT absorption of a two­component weak complex in solution has been analyzed in terms of three contributions to the shift,U

AVCT"=A"(E1- Eo) +2k2A"[S'l/(E1 - Eo) J +A.,(S"_S') , (3)

where the subscripts identify the dative bond if;l(D+A-) and no bond structure if;o(D· •• A), respectively;24 S is the overlap of the donor and acceptor orbitals involved in the CT process, and 1 <k<20. Trotter26 has attrib­uted the large vapor-solution red shift in the absorp­tion of EDA comp exes to the high CT "bond" com­pressibilityand successfully predicted the magnitude of the shift. It is then anticipated that changes in the first two terms of Eq. (3) predominate over the dielectric solvation of the molecular pair. For weak EDA com­plexes the magnitude of the second term is only about 5% relative to the first term, but

A(E1- Eo)r-vARDA <0 and

A[ S'l/ (E1 - Eo) }--exp ( - 2ARDA/ ARDA) > 0

so that the blue-shift contribution from the second term representing CT resonance forces may reduce the over­all red shift by a significant amount. This is especially true at the shorter R characteristic of the CT state (if;E) , i.e.,

A/[S'l1 (E1- Eo) J> A"[S'l/ (E1- Eo) J> O.

These inequalities are also applicable to the crystalline complex.2 •6 It follows from the more sensitive R de­pendence of the CT forces that the Stokes shift be­comes smaller at higher pressures which is consistent with the above considerations of potential-energy surfaces.

CRYSTAL SPECTRA AT 1 ATM

The CT absorption and fluorescence bands of several crystalline TNB and TNF complexes are illustrated in Figs. 2-4. The broad CT absorption bands show some

23 H. M. Rosenberg and E. C. Eirnutis, J. Phys. Chern. 70, 3494 (1966).

2' R. S. Mulliken, J. Am. Chern. Soc. 74, 811 (1952); J. Phys. Chern. 56, 801 (1952).

1& P. J. Trotter, J. Am. Chern. Soc. 88,5721 (1966).

SPECTROSCOPY OF EDA COMPLEXES AT HIGH PRESSURES. VI 751

vibrational structure in the case of TNB complexes, in accord with previous findings in pellets.lO,15

The CT-band maxima of TNF complexes occur at slightly longer wavelengths because TNF is a better acceptor than TNB, in agreement with theory.24 The electron affinities of TNB and TNF are estimated to be 0.7 and 1.0 eV, respectively.26 The CT bands are ill defined for TNF complexes and appear almost as shoulders submerged under higher-intensity intra­molecular transitions. Figure 3 illustrates the absorp­tion spectrum of indole-TNF which yields the best resolved spectrum among the studied TNF complexes. Indole is considered to be a 7f' donor,27 as are the other aromatic hydrocarbons.

Crystalline TNB and TNF complexes are similar to other CT complexes in that they fluoresce very weakly at room temperature.12 ,13 The large Stokes shift is at­tributed to the difference in equilibrium RDA for the ground and CT state and the very large binding energy in the o.PE relative to the o.PN state. The fluorescence spectra are broad, structureless, and possess similar bandwidths to those found in absorption. The TNF complexes have the higher fluorescence quantum yield, as seen from Fig. 4, where the fluorescence spectra of fluorene-TNB and fluorene-TNF are compared under similar experimental conditions.

HIGH-PRESSURE SPECTRAL SHIFTS

The spectral changes brought about by high pressures are illustrated in Figs. 2-4. The red shifts at higher pres-

ANTHRACENE - TNB

1.6

1.2

'" u z Z ..: 8! -;

'" 0 z V> .8

V> III ::; ..: -<

O.B

.4

0 .4

400 500 600

WAVELENGTH (mil)

FIG. 2. The absorption and fluorescence spectra of crystalline anthracene-TNB at different pressures. The unmarked curves were recorded at atmospheric pressure.

26 G. Briegleb, Ang. Chern. Int!. Ed. 3, 617 (1964). 'lJ. P. Green and J. P. Malrieu, Proc. Nat!. Acad. Sci. 54,

659 (1965).

'" u ~ III 0:: o V> III ..:

400 500

WAVELENGTH (mu)

600

FIG. 3. The absorption spectra of the indole-TNF complex crystal at three isobars.

sures are summarized in Tables I and II. The absorp­tion red shifts of these complexes are larger than those observed for TCNE complexes.1l This is expected from previous considerations because stronger complexes show a smaller net shift since the stronger CT interac­tions [asf(R) ] oppose the general red shift. Prochorow and Tramer21 estimate for other complexes dUEl dR"" 10 cm-I/mA at the equilibrium RDAN. If the slope of the excited-state potential-energy curve is similar for the present complexes, it is predicted that ARDAN"" 0.12 A at 25 kbar to explain the magnitude of the red shift. For an isotropically compressible, generalized molecular crystal28 ARDAN",,0.14 A at 20 kbar. Hence, the potential-energy curve of the CT state appears to be very similar for many 7f', 7f' complexes and to be independent of the state of aggregation.

It is seen that for those complexes which have distinct absorption and fluorescence bands, Avar' < AvCT4 , as predicted from Eq. (2) and in agreement with the results from plastic solutions.IG The data indicate that AVCT4 values are similar in magnitude for the crystalline complexes (Table I ) , but that Ave'll range in values from - 200 cm-l to -1700 cm-1 at 25 kbar (Table II) . This may be understood in terms of the potential­energy diagram (Fig. 1). The slope dUgldR is steep at typical RDA for the Franck-Condon absorption transition, so that the absorption frequency is not as susceptible to small differences in the complex orienta­tion and its interaction with the environment for related

28 G. A. Samara and H. G. Drickamer, J. Chern. Phys. 37, 474 (1962).

752 A. H. KADHIM AND H. W. OFFEN

TABLE 1. Absorption shifts t.vCT" and absorbance ratios at 25 kbar.-

EDA crystal

Anthracene-TNB 9-Methylanthracene-TNB 9,10-Dimethylanthracene-TNB 9-Phenylanthracene-TNB Pyrene-TNB Pyrene-TNF Benzidine-TNB Indole-TNB Indole-TNF

-1264 -1130 -1228 -1226 -1180 -1080 -1500 -750 -1200

1.6 1.6 1.8 1.3 2.0 1.4 1.8 1.3 2.0

• The estimated error is :1=100 cm-1 in t.VCTa at 25 kbar and :1=0.2 in A (25)/A (0).

complexes at high pressures. The same is not true for the emission process because the shallow ground-state potential curve may differ considerably from complex to complex. The size and orientation of the donor rela­tive to the acceptor appear important factors in deter­mining the magnitude of ~vc"'. The fluorescence red shift appears to be largest for complexes in which the molecular size of the two components are similar. A smaller pressure dependence and a smaller blue-shift contribution of the CT forces is expected for those com­plexes in which the orbital overlap is already favorable at 1 atm. In stilbene and substituted anthracene com­plexes, pressure may affect the resonance interaction (302/(E1-Eo) considerably upon compression, thereby depressing the UN(R) curve at R<RDAeq

• Propor­tionately, greater CT interactions in these complexes results in a smaller net shift at higher pressures. The larger fluorescence red shift of TNF complexes relative to TNB complexes probably arises from greater solvent stabilization of the CT state for the carbonyl-containing acceptor.

TABLE II. Fluorescence shifts and intensity ratios at 23 kbar.-

EDA crystal t.vd (em-I) 1(23) /1 (0)

Anthracene-TNB -850 0.08 9-Methylanthracene-TNB -530 0.43 9,10-Dimethylanthracene-TNB -280 0.45 9-Phenylan thracene-TNB -520 0.40 Pyrene-TNB -350 0.30 Pyrene-TNF -S70b O.04b Hexamethylbenzene-TNF -420 0.50 Naphthalene-TNB -350 Extremely

weak Naphthalene-TNF -1130 0.12 Stilbene-TNB -210 0.28 Stilbene-TNF -400b 0.35b Phenanthrene-TNF -SlOb 0.10b FhlOrene-TNB -560 0.14 Fluorene-TNF -1710 0.05 Chrysene-TNB -520 0 . 15 Chrysene-TNF -430b 0.25b

- The estimated error is :1=150 cm-1 in t.vd at 23 kbar and :1=0.10 in 1(23)/1(0).

b The pressure is 17 kbar.

INTENSITIES AT HIGH PRESSURES

It is seen from Table I that the absorbance of the crystalline complexes increases by 1.3 to 2.0 at 25 kbar. This intensification has been observed so far for all weak complexes, in solid solution as well as in the crystalline state. It is easily explained by an increased orbital overlap at smaller RDA .26 For ~RDAN'"'-'0.12, A (P)/ A (O)'"'-'exp( -2~RDAN)'"'-'1.3. This value repre­sents a lower limit because other terms in the oscillator strength expression also change under pressure, includ­ing orientation and relative size. Further, the extent of intensity stealing from the donor, if present, may be very susceptible to lattice deformations produced by high pressures. The bandwidth is a characteristic of the isolated complex and pressure has little effect on it even in the crystalline state. This is another reason why the absorption of weak EDA-complex crystals can be

>­... in z "' ...

20 FLUORENE-TNB

'!; 1.0

500 600 700 500 600 700

WAVELENGTH (m")

FIG. 4. CT fluorescence of crystalline fluorene-TNB and fluorene-TNF complexes at 1 atm and 23 kbar.

discussed from the viewpoint of solid solutions, i.e., the relative slopes of the potential energy surfaces are not changed by environmental perturbations.

Although more light is absorbed at high pressures, the fluorescence intensity is sharply decreased. This ob­servation was established to be a pressure and not a photochemical effect, although the intensity did not return completely to its value before pressure cycling. No effort was made to remove oxygen from the pellets. Fluorescence is normally not very sensitive to oxygen, but it may have some influence on intermolecular CT processes.

A decrease in the fluorescence intensity was also ob­served for TePA complexes in polymethylrnethacry­late.le In other words, the smaller quantum yield at higher pressures may arise from intra-intermolecular processes in a "solution" environment as well as from crystalline-field phenomena. This view is supported by the difference of CT absorption and fluorescence polari­zation ratios2.4 which can be interpreted in terms of the closer approach of the donor and acceptor molecule

S P E C T R 0 S COP Y 0 FED A COM P LEX E SAT HI G H PRE S SUR E S. V I 753

after excitation· and the formation of a localized exciton within the deformed crystalline lattice.z Both higher temperatures2 and higher pressures diminish the quan­tum yield for CT fluorescence. Hochstrasser, Lower, and Reid2 consider the sensitive temperature depend­ence as evidence for efficient quenching of the localized exciton by lattice motions. It appears probable that phonons in solid solutions as well as crystalline media participate in radiationless processes more efficiently at higher pressures. This is not true for aromatic hydro­carbons to which the rigid-lattice model is applicable. Increasing pressures do not diminish the fluorescence intensity of aromatics until excimer-like luminescence appears at higher pressures.29 •M Whether the quenching

til P. F. Jones and M. Nicol, J . Chern. Phys. 43, 3759 (1965) . 30 H. W. Offen, J. Chern. Phys. 44, 699 (1966).

of CT fluorescence occurs via increased intersystem crossing or directly to the ground state is to be answered by future work.

In summary, the results of pressure perturbations on CT spectra are consistent with the prevalent view about potential energy curves of and charge-transfer processes in weak EDA complexes. The strong pressure depend­ence of CT forces, first postulated by Mulliken24 in his valence-bond description of molecular complexes, has been confirmed. The Stokes shift is observed to decrease at higher pressures because the stronger CT forces at shorter distances give a larger blue-shift contribution to the over-all red shift for CT spectra of weak com­plexes. Higher pressures yield stronger ground-state complexes and weaker excited-state complexes which permit more efficient fluorescence quenching.