4507
J. Phys. Chem. 1989, 93, 4507-451 1
distinguish primary photoprocesses occurring on the surface directly, and secondary photoprocesses occurring during the same laser pulse, but in a species already removing from the surface (compare Figure 1 of ref 23), no speculations on the source of the differences for the kinetic temperature of the atomic and molecular species will be given here. One important question still to be answered, in order to get a further understanding of those processes relevant for excimer laser induced deposition of aluminum films from AI-alkyl precursors, is the following: Does whatever is removed from the surface have any correlation with what is left behind? And what-if any-is this correlation? A combination of laser diagnostic techniques may be a solution to this topic. Work along this line is in progress.
was suggested earlier2' that AI-C bond formation may occur on the 250 OC film for TMA. We detected the A1CH3 molecules at room temperature. Since the AICH3 molecules are reactive radicals and do have carbons tightly connected to AI atoms, this may be one important reason for the fact that TMA is not well suited as an organometallic precursor of high-purity aluminum films. The high laser fluence used in the adsorbed phase when compared to the gas phase, together with a slope considerably higher than 1 (up to the order of 10) when the fluence dependence of the signal S (compare Table 1) is checked, indicates the possibility of a fluence threshold. The slope n seems to be higher for 308 nm than for 248 nm. Shortening of the excimer photolysis laser pulse22 will allow us to test this hypothesi~.~' The velocity distributions of the A1 atoms and of the A1CH3 molecules can be deduced from the data shown in Figures 6 and 7 for 248 nm, and similar distributions were obtained for other wavelengths. The velocity distributions of the Al' and AICH,' signal are not the fit to the Maxwell-Boltzmann distribution, shown in Figure 9 for AI' signal as an example. The value for the most probable velocity for AI atoms is VA, = 700 f 130 m/s. The most probable velocity of the A1CH3 molecules is found to be about VAICH,= 280 f 40 m/s, considerably lower than the value for the AI atoms. From Figure 9, it is seen that there are more fast species here compared to the Maxwell-Boltzmann distribution, probably due to the photodissociation. The corresponding kinetic = 0.069 f 0.025 eV and Ekn,A1c~,= 0.017 f energies are 0.005 eV compared to the kinetic energy at room temperature of E = 0.025 eV and a laser photon energy on the order of 5 eV. The kinetic temperature of the A1 atoms is more than twice that of the substrate at room temperature. Since, with pulse durations for the photolysis laser on the order of nanoseconds, it is hard to
Conclusion Neutral UV excimer laser photoproducts from adsorbate/ substrate interfaces have been detected by single-shot tunable dye laser mass spectroscopy. For the Alalkyls TMA, TEA, and TIBA and the substrates n-type Si( 100) and Si02(quartz), aluminum atoms A1 and the molecules A1W and AICH3 were detected with abundances depending on the aluminum alkyl adsorbates. Marked differences between gas phase and the adsorbed phase are observed for 308 nm, where only the adsorbed molecules release neutral fragments. The relative abundance of the carbon-metal-containing fragment molecule AICH3 increases considerably when TMA is adsorbed to quartz compared to the case of free TMA molecules.
Acknowledgment. We thank D. Gudlin for measuring N M R spectra of several Al-alkyls for us, R. Larciprete for an important discussion regarding the observation of A1CH3 molecules, F. P. Schafer for support, K. Muller for technical assistance, and, for financial support, SFB 93 (Photochemie mit Lasern, C2+C15) and BMFT (No. 13N 5398/7). Registry No. TMA, 75-24-1; TEA, 97-93-8; TIBA, 100-99-2; Si, 7440-21-3; SO2, 14808-60-7;AI, 7429-90-5;AIH, 13967-22-1;AICH,, 76392-49-9.
(22) Szatmari, S.;Schafer, F. P.; Muller-Horsche, E.; Miickenheim, W. Opt. Commun. 1987, 63, 305. (23) Kuper, S.; Stuke, M. Appl. Phys. 1987, 844, 199.
Determination of the Phosphorescence Quantum Yield of Singlet Molecular Oxygen ('Ag) in Five Different Solvents R. Schmidt,* K. Seikel, and H.-D. Brauer Institut fur Physikalische und Theoretische Chemie. Universitat Frankfurt, Niederurseler Hang, D 6000 Frankfurt am Main, FRG (Received: July 6, 1988; In Final Form: November 2, 1988)
-
The quantum yield of singlet oxygen ( ' 0 2 ) 'A,(u=O) 'Z;(u=O) phosphorescence was determined in acetonitrile, chloroform, carbon disulfide, carbon tetrachloride, and Freon 113 relative to the respective emission in benzene, using the known IO2 phosphorescence quantum yield in benzene as standard. Quantum yields were not found to depend on sensitizer (dicyanoanthracene, rubicene, tetraphenylporphine) but to depend strongly on solvent. The '0, phosphorescencequantum yields are surprisingly large. The maximum value measured is &(Freon 113) = 0.15. The emission quantum yields correlate linearly with IO2 lifetimes for all solvents, including benzene. Consequently the rate constant of '0, phosphorescence is independent of solvent. It amounts to k , = 1.3 s-I. Thus the radiative rate constant is approximately 5000 times larger in liquid solution than for an isolated '02molecule.
Introduction The lifetime of IAg singlet oxygen (IO2) is limited essentially by collisional radiationless deactivation by solvent molecules in most liquids.1,2 The mechanism of the deactivation process can be described as an electronic to vibrational energy transfer from IO2 to a single oscillator of the energy-accepting solvent molecule. The deactivation rate depends over a range of 4 orders (1) Hurst, J . R.; Schuster, G. B. J . Am. Chem. SOC.1983, 105, 5756. (2) Schmidt, R.; Brauer, H.-D. J . Am. Chem. SOC.1987, 109, 6976.
0022-3654/89/2093-4507$01.50/0
of magnitude on the nature of solvent.2 In contrast, the rate of radiative deactivation of IO2 appears to be much less influenced by the solvent. By calculating the oscillator strength of the 0, 'Z;(v=O) I$(u=O) transition from absorption spectra, Long and Kearns found the radiative lifetime (q.)of '0, for a series of perhalogenated solvents to be solvent i n d e ~ e n d e n t . ~In accordance with this result Krasnovsky determined by direct observation of the IO2 emission via a phos-
-
(3) Long, C.; Kearns, D. R. J. Chem. Phys. 1973, 5 9 , 5729
0 1989 American Chemical Society
4508
Schmidt et al.
The Journal of Physical Chemistry, Vol. 93, No. 1 I , 1989
TABLE I: Sensitizer Fluorescence Quantum Yields QF(X) in Different Air-Saturated Solvents' solvent sensitizer DCA
RUB TPP
C6H6
0.80 f 0.06 0.21 f 0.01 0.063 f 0.004 (0.075)
CH,CN 0.69 f 0.04
0.17 f 0.01 0.060 f 0.008
CHCI,
cs2
CCI4
Freon 113
0.78 f 0.03 0.14 f 0.01 0.052 f 0.005
0.70 f 0.07 0.13 f 0.02 0.058 f 0.003
0.80 f 0.04 0.14 f 0.01 0.057 f 0.008
0.82 f 0.04 0.15 f 0.02 0.048 i 0.002
'The standards a r e given in the Experimental Section. T h e number in parentheses is from ref 19
phoroscope also solvent independence of T p 4 However, recently several studies on IO2 have been published revealing apparently a moderate dependence of the phosphorescence rate constant on The largest effect was reported by Scurlock and Ogilby claiming an increase in the radiative deactivation rate by a factor of 25 in going from 1, I , 1-trifluoroethanol to carbon disulfide as ~olvent.~ The easiest method for the calculation of the radiative lifetime 7p(X) in a solvent X demands the determinations of the lifetime 7A(X) and the quantum yield Qp(X) of '0, phosphorescence, according to eq 1. Reliable values of 7&(X)have been obtained
TP(X) = ~ A ( X ) / Q P ( X )
(1)
previously for a large number of solvents by the time-resolved spectroscopic m e t h ~ d . l , ~In~ a~ preceding ,~ paper one of us determined the IO2 phosphorescence quantum yield in benzene by comparison of the corrected emission spectra of sensitizer fluorescence and IO2 phosphorescence recorded from the same sample. Using three different sensitizers a value of Qp(B) = (4.7 i 1.7) X was obtained.I0 Since continuous excitation leads generally to more accurate quantum yields than pulsed excitation we decided to employ the conventional stationary technique also for the investigation of the discussed solvent dependence of kp(X) = I / T ~ ( X ) . Using the benzene value as standard we determined IO2 phosphorescence quantum yields in five different solvents. In these liquids '02 lifetime varies from 3 1 to about 100 000 ys.
Experimental Section 9,10-Dicyanoanthracene (DCA) from Kodak was recrystallized from ethanol. Benz[a]indeno( 1,2,3-hi)aceanthrylene, rubicene (RUB), was synthesized and purified following literature procedures." Tetraphenylporphine (TPP) was purchased from Alrich and used without further purification. Mesodiphenylhelianthrene (MDH) was prepared according to a procedure given earlier." The fluorescence standards 9,10-diphenylanthracene (DPA) and cresyl violet (CV) from Aldrich and rhodamine 6G (RG) "Lambdapure" from Lambda Physik were used as purchased. Benzene (C&), carbon tetrachloride (CCI,), chloroform (CHC13),acetonitrile (CH,CN), and carbon disulfide (CS,) were from the Uvasol series of Merck; Freon 113 (CF2C1CC1,F), spectroscopic grade, was supplied by Aldrich. Small amounts of acid were removed from CHC13 and CH,CN by column chromatography with basic A1203from Woelm. Freon 113, CC14,and CS, were purified by the same method using neutral AI,O, (Woelm) and C6H6was not purified additionally. All solutions were air saturated and the experimental temperature was about 22 OC. Spectra were recorded with a PE 555 spectrophotometer or a 650-40 fluorescence spectrometer, both from Perkin Elmer. Photochemical quantum yields have been obtained with an instrument described earlier.I3 The simultaneous recording of (4) Krasnovsky, A. A., Jr. Chem. Phys. Lett. 1981.81, 443. ( 5 ) Scurlock, R. D.;Ogilby, P. R. J . Phys. Chem. 1987, 91, 4599. (6) Gorman, A. A.; Hamblett, I.; Lambert, C.; Prescott, A. L.; Rodgers, M. A. J.; Spence, H. M . J . A m . Chem. SOC.1987, 109, 3091. ( 7 ) Losev, A. P.; Byteva, 1. M.; Gurinovich, C. P. Chem. Phys. Lett. 1988, 143, 127.
(8) Ogilby, P. R.; Foote, C. S. J . A m . Chem. SOC.1982, 104, 2069. (9) Rodgers, M . A . J. J . A m . Chem. SOC.1983, 105, 6201. (IO) Schmidt, R. Chem. Phys. Left. 1988, 151, 369. ( I I ) Clar, E.; Willicks, W . J . Chem. SOC.1958, 942. (12) Acs, A.; Schmidt, R.; Brauer, H.-D. Photochem. Photobiol. 1983, 38, 527.
5
/
481
u
/
l
tJ
i
1
I
1 0 0 0 0 0 ~TPP [ M I
Figure 1. Ratio of T P P fluorescence over IO2 phosphorescence in dependence of [TPP] in air-saturated CCI,; A,,, = 405.
emission spectra of sensitizer and '0, was performed with a home-built emission spectrometer which is equipped with a germanium diode and covers the broad range from 400 to 1600 nm. This instrument is described in detail in ref 2. Fluorescence quantum yields QF have been calculated according to the procedure described by Parker and Real4 from the corrected fluorescence spectra. As standards served air-saturated solutions of DPA (cyclohexane, QF = 0.7315 for DCA, RG (methanol, QF = 0.8616) for RUB, and CV (methanol, QF = 0.5516)for TPP.
Results IO2 is most conveniently produced in solution by sensitization. However, sensitizers (S) can act as quenchers of IO2 too.17J8 In weakly deactivating solvents like CCI4 sensitizers as for example M significantly to IO2 TPP contribute already at [SI = deactivation.2 To exclude a reduction of Qp(X) caused by sensitizer quenching we therefore determined the '0, phosphorescence emission in dependence of [SI. Variation of [S] leads to changes of the local emission distribution in the sample cell, which are the same for sensitizer fluorescence and IO2 phosphorescence. Therefore, the ratio of intensities of sensitizer fluorescence ZF(X), measured at an emission wavelength in a maximum where reabsorption does not disturb and IO2 phosphorescence Zp(X), determined at the maximum of emission, depends no more on excitation and emission geometries and consequently not on the solvent refractive index. In the [SI range, where the quantum yields QF(X) of fluorescence and Qa(X) of IO2 formation by the sensitizer are constant, the intensity ratio depends linearly on [SI as is shown by eq 2. Here C represents a solvent-independent
- c Q F ( X ) ( ~ P ( X+) kD(X) + kQs(x)[sl) (2) IPW) Q a W kp(X) proportionality constant differing only from sensitizer to sensitizer and kD(X) and kQs(X) are the rate constants of radiationless deactivation of '02by the solvent and by the sensitizer. The designation (X) indicates the solvent X. --
Drews, W.; Schmidt, R.; Brauer, H.-D. J . Photochem. 1976, 6, 391. Parker, C. A.; Rees, W. T. Analyst 1960, 85, 587. Heinrich, G.; Schoof, S.; Gusten, H. J . Photochem. 1974/75, 3, 315. Olmstedt, J. J . Phys. Chem. 1979, 83, 2581. Foote, C. S.; Ching, Ta-Yen J . A m . Chem. SOC.1975, 97, 6209. (18) Krasnovsky, A. A,, Jr. Phorochem. Photobiol. 1979, 29, 29. ( 1 9) Rossbroich, G.; Garcia, N. A,; Braslavsky, S. E. J . Photochem. 1985,
(13) (14) (15) (16) (17) 3 1 , 37.
Phosphorescence Quantum Yield of Singlet Oxygen TABLE 11: Sensitizer Quantum Yields QA(X) of
C6H6 0.19 f 0.04 0.48 f 0.10 0.31 f 0.06 0.67 f 0.13 (0.58 f 0.06)d
sensitizer DCA~ DCAC RUBC
Formation in Different Air-Saturated Solvents'
CH3CN 0.23 f 0.05 0.76 f 0.15 0.18 f 0.04 0.50 f 0.10
~
TPP'
'02
The Journal of Physical Chemistry, Vol. 93, No. 11, 1989 4509
solvent CHCI, 0.09 f 0.02 0.16 f 0.03 0.31 f 0.06 0.50 f 0.10
cs2
0.09 f 0.02 0.17 f 0.02 0.46 0.05 0.51 f 0.05
*
CC14 0.13 f 0.03 0.16 f 0.02 0.30 f 0.03 0.54 f 0.05
'Mean values of at least three experiments. *Determined by physical methods. CDeterminedchemically.
Freon 113 0.09 f 0.02 0.14 f 0.02 0.23 f 0.02 0.41 f 0.04
dFrom ref 19.
I
150
Q z 2
2
B
9
3
2\
120
-
90 -
-
-
-
I
-
60 -
aI
2 3-
e
?
9
0 -
0' 0
I
10000
I
I
20000
30000
40000
I 50000
l/iMDHI l/[MI Figure 3. Reciprocal quantum yield of MDH photooxygenation sensitized by TPP in dependence of the reciprocal MDH concentration. Solvent, air-saturated benzene; Xi,, = 647 nm.
0
represents the emission ratio for [SI = 0 any static or dynamic reduction of Q< (X) caused by the sensitizer is excluded. To obtain Q
