Exciplex photophysics. II. Fluorescence quenching of substituted

Dec 1, 1974 - Exciplex photophysics. II. Fluorescence quenching of substituted anthracenes by substituted 1,1-diphenylethylenes. William R. Ware, J. D...
0 downloads 0 Views 497KB Size
7861

Exciplex Photophysics. 11. Fluorescence Quenching of Substituted Anthracenes by Substituted 1,l -Diphenylethylenes William R. Ware,* J. David Holmes, and Donald R. Arnold' Contribution No. I I 4 f r o m the Photochemistry Unit, Department of Chemistry, University of Western Ontario, London, Ontario, Canada N6A 3K7. Received July 18. I974

Abstract: The fluorescence quenching of substituted anthracenes by substituted 1 , l -diphenylethylenes in benzene and acetonitrile solvent at room temperature is reported. The quenching efficiency is observed to vary strongly with the donor-acceptor properties of the interacting pair, with no quenching seen ( k q < IO7) in some cases and diffusion-controlled quenching observed in others. In the systems studied. the olefin is generally assumed to be the donor, but in one case probably becomes the acceptor. Both fluorescence lifetime and steady-state quenching measurements are reported and in one case exciplex emission and two-component decay curves a r e observed and analyzed. The kinetics a r e discussed in terms of the standard excimer kinetic model.

This paper* describes a study of the fluorescence quenching of several anthracene derivatives by a series of substituted 1 ,I-diphenylethylene derivatives. The anthracenes and olefins were selected to span a considerable range of donoracceptor properties of the pairs, as deduced by the usual electron donating-withdrawing considerations. The following questions were considered of interest. (a) Is exciplex emission observable? (b) Are there discrepancies between the steady-state and transient quenching rate parameters which would indicate exciplexes of lifetime considerably in excess of the duration of an encounter between noninteracting species? (c) Does the quenching efficiency follow the donor-acceptor properties of the quenching pairs? (d) What is the effect of nonpolar vs. polar solvents on the fluorescence quenching? This information would provide a firm foundation for the interpretation of the photochemical behavior of these systems which is now under investigation. The following anthracenes were used: anthracene, (A); 9,1O-dicyanoanthracene, (DCA); 9,lO-dimethoxyanthracene, ( D M E O A ) . As quenchers the following olefins were used: 4,4'-dimethoxy- 1,l -diphenylethylene, (DMEODPE); 4,4'-dimethyl- 1 , l -diphenylethylene, (DMDPE); and 4,4'dinitro- 1 ,I-diphenylethylene, (DNDPE). Experiments with unsubstituted 1,l -diphenylethylene proved difficult because of the instability of this olefin during the time required to measure fluorescence lifetimes and steady-state quenching constants. In all cases, the primary excited species was either anthracene or one of its derivatives. Benzene and acetonitrile were selected as solvents for reasons of solubility and polarity.

I. Experimental Section Fluorescence lifetimes were measured with the single photon counting technique3 which employed a gated nanosecond flash lamp and conventional electronics. The excitation system consisted of a nanosecond lamp and a Spex Minimate mnochromator. Filters were used to isolate the fluorescence. The technical details of the single photon counting technique have been described Steady-state measurements were made with an intensity ratio instrument described by Ware and Lewis.5 The search for exciplex emission was done with a fluorimeter consisting of a Heath monochromator and a Spex Minimate monocromator (excitation), a 150-W xenon high-pressure lamp, and a low-noise 1 P28 photomultiplier. The sensitivity of this system was such that fluorescence with a quantum yield could be detected for systems absorbing IO-50% of the incident light. Care was taken to ensure that the measurements were not influenced by thermal or photochemical decomposition of the so-

lutes, by solvent fluorescence, or by fluorescence of impurities in the solutes. The anthracene and substituted anthracenes were obtained commercially. The synthesis and characterization of the olefins will be discussed elsewhere.

11. Results and Discussion With one exception, all the anthracenes examined exhib-. ited single exponential decay in both the presence and absence of quencher. The decay kinetics followed simple pseudo-first-order kinetics,3 i.e.

1 / ~=

(1/7.)[Q]=o +

(k,)t,[QI

(1)

where

(k,)tr = A(1/7)/A[Q]

(2 )

The exception was D C A quenched by D M E D P E in benzene which exhibited a double exponential decay. Studies of decay curves as a function of quencher concentration for this system yielded sufficient variation in the components, which were similar in magnitude, so as to permit approximate analysis for the individual rate constants. Values for ( k J t , are given in Tables I and 11. In all cases studied, fluorescence spectra at several quencher concentrations were examined for evidence of exciplex emission. The only case where exciplex emission was detected was the DCA-DMEDPE system. Steady-state quenching experiments gave linear SternVolmer plots. Values of K s", i.e.

are also given in Tables 1 and 11. For all systems examined where quenching was observed to occur, no discrepancy between (kq)ss and ( k & , was found within experimental error. Values for (kq)crwere consistently higher when the solvent was acetonitrile as compared to benzene. As can be seen from Tables I and 11, in one case the ratio of rate constants i n the two solvents is somewhat greater than that explicable in terms of relative diffusion rate (?(acetonitrile)/ ?(benzene) = 0.571, implying a solvent effect due to polaritY. The quenching constants in Tables I and I1 range from 1 O8 to I O 'OM- sec-'. Thus for a number of the quencherfluorophor systems we observe rate constants considerably below that for diffusion-controlled quenching. These data can be rationalized on the basis of the following simple scheme

Ware, Holmes, Arnold

'

/ Quenching of Anthracenes by 1,1 -Diphenylethylenes

7862 Table 1. Quenching Data for Various Aromatic-Olefin Pairs in N2-Saturated Benzene at 25" . ...

_.______~

Fluorophor ~

~

~

~

9. IO-Dicyaiioanthracene

Queiicher

-

9. 10-Dimethoxyanthracene

M-' sec-l

_~________

_________.

Diinethoxqdiphenyletli~lene Dimethyldiplien~lethylene Dinitrodiphenylethyleii~ Dimethoxydiphenylet hylene Dimetli~ldipheii).lethyleiie Diiiitrodiplienyletliylene Dimetlioxydiplienylethq lene

Anthracene

nsec

TO,

--____ (k,itr x 10-9.

12.4 12.4 12.4 3.6 3.6 .3 . 6 13.3 13.3 13.3

Ditneth~ldiphenplethylene Dinitrodiplieii?leth! lene

-

( k ~ . x. 10-9. M-1 Fec - *

~

k.,:X,.~___

9.8

IO 4.5

3.4 1.3 0.33

3.1

0.70

1.4 0.33

0.034

(1

h NQtd

NQ 17

0 15

h

h

NQ NQ I6

NQ NQ 3.0

" Two-component decay. ', Spectral overlap makes measurements impossible. Based on h n = 1 X 10'0 !W1sec-I. Calculated from ( k & N Q = no quenching.

Table TI. Quenching Data for Various Aromatic -Olefin Pairs in N?-Saturated Acetonitrile a t 25' _______~~~

______

__

Quencher

Fluorophor

.-

_ _ _ _ _ _ _ _ ~ - ~

9. I 0-Dicyanoantliracene

~

___

__

(kqltr x 10-9. (kq).%x 10-9, 7".

ixec

______-

-

Dimethoxydiplien4letliylene

Dinietliyldiphen yletlirlene

15.2 15.2 15 2 5.4 5.4 5.4 11.2 11 2 11.2

M-l set.-' I5 13

sec-'

k,,, k i(1

____

17 13

R'

1.7 3 1

Diiiitrodiphenyletliylene Dimethox ydiphenrleth ylene 3.4 ? 7 0 28 1.6 0 51 0.55 0.058 I .7 Dimetli yldiplien yletlirlene Dinitrodiplienqlethqlene b 1; h 9. IO-Dimethoxyanthraceiie Dimethoxydip henyleth ylene NQd NQ NQ Dimetli~ldiplien~lcth~leiie NQ NQ NQ Dinitrodiphen) lethylene _ _ _ _ ~ __ ________~_ --Based on !in= I X IO'" > k4 and if k , = 10'o-lO'l sec-', then the quenching efficiency will be high and the behavior will approach that of a diffusion-controlled process. The same situation prevails if both k , and k4 are 1 0 1 o - l O ~sec-I. ~ If k4 >> k , and if k4 = 101o-lO1' sec-', then quenching will be inefficient. Either k q or k , must be very large compared to lo9 sec-' in order that one not see a discrepancy between the steady-state and transient quenching constants (unless k 3 is considerably below the diffusion-controlled limit). Only in one case do we see a two-component decay. Otherwise, all the combinations examined exhibit the limiting behavior; ;.e., the exciplex, if present, probably has a lifetime not greater than 100 p s t x 6 In both Tables I and 11 the excited molecule-quencher systems are arranged in order of their donor-acceptor properties. Consider first DCA. It is quenched by all three olefins but with decreasing efficiency (as judged by k 4 ) as one goes from the best to the worst donor. If one now decreases the acceptor property of the aromatic hydrocarbon by going to anthracene, the quenching efficiency is less than with DCA and any of the olefins and furthermore appears to again decrease as one goes from the best to the worst donor. Unfortunately, it was not possible to study A and D N D P A because of unfavorable spectral overlap. Finally, a further decrease in the acceptor tendency of the aromatic in the case of D M E O A results in the absence of quenching ( k q < lo7 M - I sec-I) by the best donor or by D M E D P E . However, D N D P E quenches D M E O A a t a diffusion-controlled rate! A reasonable conclusion is that in this case the aroJournal of the American Chemical Society

/ 96.26 /

matic and olefin have changed roles, with the olefin becoming the acceptor. This view is consistent with the observation that no quenching is seen with either A or D C A and DNDPE. It must be emphasized that the simple scheme given above requires five rate constants, and only two ( k l and k 2 ) are directly accessible. Also, k may well be the sum of two or more rate constants representing several channels out of the exciplex other than feedback (step represented by k 4 ) to regenerate A*. The single exponential decay in the presence of quencher, the absence of exciplex emission, and the agreement between steady-state and transient measurements limit the information available from fluorescence measurements. The individual constants k3, k4. and k , cannot be determined without additional assumptions. T h e system DCA-DMDPE was unique in exhibiting emission that we attribute to a n exciplex. In Figure 1 a r e shown two time-resolved spectra for this system in benzene a t 25'. These time-resolved spectra were taken with the single photon counting t e c h n i q ~ e . The ~ . ~ single channel analyzer associated with the A D C of the multichannel pulse height analyzer was used to select a portion of the decay curve which was then subjected to spectral analysis. Curve B represents the spectrum a t the maximum of the decay curve and was obtained with a narrow time window of a few nanoseconds. Curve A was taken with a much wider time window which spanned the tail of the decay curve where there should be more exciplex emission present. T h e difference between these two curves is also shown, and corresponds approximately to the difference spectrum obtained from steady-state emission spectra for this system in the presence and absence of quencher. This time-resolved spectrum provides rather convincing evidence for an exciplex as a n intermediate in the fluorescence quenching in this case and in turn strengthens the assumption that in all of the systems studied the intermediate is an exciplex, presumably with charge-transfer character. As mentioned above, the system DCA-DMDPE exhibited a two-component decay for the DCA*. This implies a

December 25, 1974

,

7863

9,:.

,2L'2

233

334

C'5

1

.'3'. .M:

Figure 2. Plots of the decay constants associated with C I exp(-XIf) t C l exp(-Xrt) for 9.10-dicyanoanthracene quenched by 4,4'-dimethyl],I-diphenylethylene in ben7ene at room temperature. Note that the X I intercept is a measured quantity (the reciprocal unquenched lifetime).

.Navelength

[nv)

Figure 1. Nanosecond time-resolved fluorescence spectrum of 9.1 O-dicyanoanthracene quenched by 4,4'-dimethyl- I , I -diphenylethylene in benzene a t room temperature. Curve A: measured with a wide time window with emission occurring in the tail of the fluorescence decay curve. Curve B: measured at the peak of the fluorescence decay curve with a narrow time window. Curve C : difference spectrum attributed to the exciplex.

substantial decrease in the lifetime of the exciplex such that feedback ( k 4 ) is now an important process relative to the processes labeled k , and k3 [Q]. An attempt was made to resolve thse two-component curves for a series of quencher concentrations using a grid search program for the four parameters in [A*] = cle-*lt

+

(41

c2e-*Zt

+

The results a r e shown in Figure 2. A plot of XI A 2 was linear with approximately the scatter shown in Figure 2. Since2,*

X i + X, = ki

+

k,

+

k3[Q]

+ kd

+

k,

(5)

+

this plot should have a slope of k3 and an intercept of k 1 k2 k4 k , . The intercept of the plot of A 2 can be used,x k2, to obtain a point for the inalong with the known k 1 X2, a procedure which is helpful in obtaining tercept of XI the best estimate for the slope and thus k3. This approach k , and k3 and gave k3 = 9 X lo9 M - ' sec-I. From k4 the equation

+ +

+

+

+

(6

one obtains k4 N k , E 4 X 10' sec-I. Unfortunately the D C A - D M D P E system was the only one which exhibited exciplex emission and a two-component decay curve. It is of interest with respect to the other systems studied to investigate the implications of two constraints that can be applied: (a) since there is no discrepancy between the steady-state and transient quenching conk , must be large6 probably stants, in these systems k4 greater than 1 O 1 O sec-I; (b) assume k3 to be approximately diffusion controlled. From eq 6 one obtains

+

k,/k, = (k,/k,) - 1

(7 1

and one can estimate k 4 / k , from eq 7 and the assumed value for k3, provided k , > k3 by a considerable amount. On the other hand, if k , E kdiff,then it is perhaps not unreasonable to assume that k , is in fact approximately equal to k3. which implies k4 mixtures. A more recent infrared investigation of the condensed products of a xenon-chlorine glow discharge by Nelson and Pimente14 identified XeClz and assigned a band a t 3 I 3 c m - ' to q .In one of the early Raman matrix-isolation studies. Boa1 and Ozins produced XeClz using the same synthetic approach and identified V I a t 253 cm-I. Vatrix-Raman studies of OF2 and ClrO i n this laboratoryh.' have shown that laser photodecomposition of the precursor gives rise to new chemical species, namely O F , C10, and CI-CIO, which were directly observed by laser Raman spectroscopy. The matrix moderates photodecomposition by the laser and allows the observation of Raman spectra of the precursor and photolysis product. Accordingly. we attempted a study of inert gas dihalides using laser photolysis of matrix-isolated halogens in the inert gas matrix. For compounds of limited stability, like Xe(:ll. and the new mixed halogen species XeCIF. this was a particularly useful approach.

Experimental Section .I.he experimental techniques, equipment. and manometric procedures have been described in previous papers from this laboratory.a.l' Most experiments involved a 2% halogen-rare gas mixture: f o r onc investigation. an Ar:Xe:F2 ratio of 100:l:l was used. Halogens s e r e purified as follows: F2 (Allied) was passed slowlq through t; \t;iinless-steel I: tube precooled with liquid Nz: CI. (Mathcson) *as condensed with liquid K2 and outgassed: OF:. CIF, and CIF.3 (Ozark-Mahoning) were used without further purification. Rb"CI (99% enriched) was reacted with concentrated H N O 3 to produce 'iClz. The gaseous product was passed through Drierite, condensed, and outgassed. Prior 10 sample preparation of fluorine-containing species, the \t:iinlc\\-\tccl vacuum line and sample can were passivated overnight a i t h 30 50 Torr of the appropriate compound. Sample deposition M ~ 3Smniol/hr for the Raman experiments and varied from ; i n initi'il rate oi' 1 mmol/hr to a final rate of 2 mmol/hr for i n frared \tudies. G:ise\ u e r e deposited on a tilted copper &edge ( R a m a n ) or a Csl window (infrared) maintained at 16'K. I n some

Journal of'the American Chemical Society

of the present experiments, the inert gas-halogen mixture was subjected to microwave discharge prior to deposition. The microwave discharge (Burdick Corp., Milton, Wis., Model M W 200, 2450 M H z , 375 W ) was maintained through the flowing gas stream using a standard Evenson-Broida cavity on a quartz sprayon nozzle (2-mm orifice). The cavity was placed 6 in. from the copper block; the glow extended through the orifice to within 1 cm of the copper block. Raman spectra shifted 100-1200 cm-I from the laser source were recorded on a Spex Ramalog obtained by irradiation with the 4519, 4765, 4880, 4965, 5017, and 5145 %, argon ion and 3564 %, krypton ion laser lines using 50 to 150 m W of power at the sample unless otherwise specified. Dielectric filters were employed to prevent unwanted plasma emission lines from appearing in the spectra, and band calibration was accomplished by superimposing emission lines on the spectrum during the running scan by replacing the dielectric filters with a neutral density filter. Infrared spectra were recorded on a Beckman IR-12, and calibrations were made relative to HzO rotational lines present in the scans. Precision of band positions in infrared and Raman spectra is f l cm-I or better (as listed). Slit widths were 500 p for Raman surveys and varied from 3 cm-' at 300 cni-.l to 1 cm-I at 800 cm-' during infrared scans.

Results The experimental observations will be presented in turn for the several noble gas-halogen mixtures studied. XeFt Raman. Laser irradiation of 2% F? in Xe yielded a Raman signal a t 5 12.4 f 0.5 cm-I, which is shown in Figure I . This signal is in excellent agreement with the gasphase assignment 514.5 cm-] to vI of XeF2 which is listed i n Table I.Io Survey scans with blue (4579, 4765, and 4880 A) excitation showed the XeF2 band present immediately after sample exposure to the laser beam. Also, the signal intensity exhibited little change with prolonged irradiation from the beam. Green (5145 A) excitation initially showed no signal in the spectrum, but 10 min of photolysis was sufficient to produce a weak band at the frequency shift appropriate to XeF2. No signal was discernible f o r Fz, normally found about 890 cm-I. Streng and Strengl I produced XeF2 from sunlight irradiation of Xe:OF2 mixtures, a reaction that was duplicated by laser photolysis of 2% OF2 in Xe condensed a t 16'K. The resultant XeF2 band ( 5 12.5 f 1 .O cm-I) was the strongest observed in this series of experiments and was even

/ 96:26 / December 25, 1974