- Email: [email protected]
Laser two-photon ionization of organic molecules in dielectric liquids
CHEMICAL PHYSICS LETTERS
Voltime 80. number3
LASER TWO-PHOTON IONIZATION OF ORGANIC MOLECULES
15 June 1981
IN DIELECTRIC
LIQUIDS *
K. SIOMOS’, G. KOUROUKLIS and L-G. CHRISTOPHOROU* Atomic. Molecular. and High Voltage Physics Group, Health and Safety Research Division, Oak Ridge Natior;al Laboratory. Oak Ridge, Tennessee 37830, Recewed
26 Februaq
1981;
USA in final Form 14 March 1981
The iotiation thxeshold of fluoranthene and of TMPD in rz-pentane was determmed by laser two-photon ionization (TPI) and found to be 450 c- 0.05 and 3 88 + O-05 eV respectmely. For both molecules the TPI spectra show distinct structure due to autoionization. For fluoranthene the TPI spectrum suggests that the molecule dissociates via the fist excited singlet state.
1_ Introduction
Recently [l] ,we descrrbed a two-photon ionization (TPI) technique suitable for liquid-phase studies, which we employed to determine the photolonization threshold 1: of solute molecules embedded in dielectric liquids_ While one-photon ionization is limited due to the absorption of the host medium, two-photon transitions are at an energy twice that of the individual photons absorbed, and thus ionization thresholds of molecules in liquids as high as = 10 eV can be reached with frequency-tunable dye lasers. The TPI technique is not only suitable for accurate measurement of Ii but also for obtaining information on molecular states embedded in the ionization continuum. The 1; is lower than the corresponding ionization threshold I& of a molecule in the gas phase (see, e.g. ref. 121 and references therein). Distinct peaks can appear in the TPI spectrum when the energy of a molecular state in the ionization continuum is resonant with the two-photon energy, assuming the parity of this state is the same as that of the ground state for molecules with a center of sym-
* Research sponsored by the Office of Health and Environmental Research, US Department of EnerOT, under contract W-740Seng-26 with the Union Carbide Corporation_ * Aisn at Department of Physics, The University of Tennessee, Knoxville, Tennessee 37916, US.4.
504
metry [3] and thus parts of a molecular spectrum that do not avail themselves of one-photon spectroscopy can be investigated [ 11 _In addition, the TPI technique can be employed to probe photophysical processes resulting from two-photon transitions which do not lead to ionization [4] _ In this paper, results are reported and discussed on photoionization of fluoranthene and N,N,N’,N’-tetramethyl-p-phenylenediamine (TMPD) in liquid rr-pentane _Fluoranthene is a non-altemant conjugated hydrocarbon of environmental interest and with “peculiar” absorption and fluorescence characteristics [S-9]. TMPD has a low ionization onset, and for th& reason it has been investigated by various methods whose results can be compared with those of the TPI technique. The potential of the TPI technique and the basic significance of the TPI data for molecules in liquid media are indicated.
2. Experimental The experimental apparatus has been described [ 11. The beam from a frequency-tunable nitrogenlaser-pumped dye laser (Molectron UV lOOO/DL 200) is focused on the photoconductivlty cell inducing two-photon transitions which result in ionization of the solute. The TPI signal from a parallel-plate chargesensitive detector (maintained at apotential difference
Volume 80, number 3
15 June 1981
CHEMICAL PHk!SICS LETTERS
[I] for further discussion) are plotted as a function
of 3 kV corresponding to an electric field of 15 kV cm-l) after linear amplification is averaged using a box-car integrator. The dye laser beam leaving the cell is monitored by a photodiode of known spectral responsivity and is sampled by a second boxcar integrator. The output signals of both integrators after digitization are fed into a PDP-10 computer for analysis_ The TPI signal after intensity correction (division by the square of the light intensity after correction for the spectral responsivity of the photodiode; see ref.
Of %aserDue to the low 1: value of TMPD, it was possible to measure the one-photon ionization (OPI) spectrum for TMPD in n-pentane near the ionization threshold. For this, the apparatus was slightly modified to accommodate a frequency-doubling unit. The OPI current fcr ThWD in n-pentane near the threshold was measured as a function of the second harmonic of the dye laser generated in an angle-tuned KDPcrystal.
:86
-PI-THRESHCLO [4_s0t0_0sleV
1
\ I
I
380
400
I
420
1
440
LASER
I
460
I
I
480
WAVELENGTH
500
I
520
I
540
5 Q
(nml
Fig_ 1. Two-photon ionization(TPI) currentas a function of the Iaserwavelength(curve 1). The sectionsof this spectrumabove 440 nm (curve2) and 500 nm (curve 3) are ma@fied by factors of 17 and 86, respectively.The arrowindicatesthe position of
the photoionization threshold.
505
Volume 80, number 3
CHEMICAL
PHYSICS
The liquid n-pentane used in this study was Phillips Research Grade (9998%) and was purified both chemicaily and under vacuum before use [I] _ Fluoranthene and TMpD were of a quoted purity 9999 and 98%, respectively_ They were used without fur*her purification. Au measurements were made under vacuum and at room temperature_
of fluorauthene
3. The TPX spectrum
in n-pentane
The TPf spectrum of =5 X 10d6 mof dme3 sofutions of fluoranthene in PI-pentane is shown in fig_ 1. The photocurrent is plotted as a function of Alaser in the spectral region 366 to ~560 run which corresponds to two-photon transitions in the range 188 to e280 nm. The spectrum presented comprises ten individual corrected spectra which were combined by
Table 1 Positions (in cm-l)
15 June 1981
normalization in the respective regions of overlap. The TPI spectrum shows three important features: (I) The photocurrent declines sharply for A,,,,, > 420 run and then reaches asymptotically a ‘stero levei” as the wavelength increases. (2) Distinct structure is observed over the entire energy range investigated. (3) The photocurrent declines for A,,,,, < 414 MI. Featuxe (1) was utilued to determine the 1: of fluoranthene in n-pentane. The photocurrent reaches asymptotically a “zero lever” [l] at 551 .O k 6.2 nm. This corresponds to an energy for a two-photon transition of 4.50 rt 0.05 eV, which we identify with the 1: of fluoranthene in tr-pentane.
This structure is attributed to autoionization of discrete states embedded in the ionization continuum
of the maxima in the one-photon absorption and TPL spectra of fhxoranthene in n-pentane
One-photon absorption [ 111
36742 +
80
j 37390 * 80 1 38105 + 80 38531 r 39500 40684 41205 42474
ii i *
80 100 130 100 50
43450 f 400 44705 *
80
45400 + 130 45842 + 130 46980 r 200
TPI two-photon resonant tmnsrtions (present work) 36742 +
sitionswasnot possible. b, Theseare the positions of the
One-photon absorption a)
38105 38531 38789 39500 40684 41205 42105 42842 43474 43579 43940 44705 44974 4.5168 45421 45868 46816 47079 47289
T?I one-photon resonant transitions (present work)
SO
37390 5 80
+ SO t 130 i 130 + 100 t 180 f 80 i 100 + 50 f 80 -c 50 + 100 f 80 c 50 c 80 i 100 -c 130 & 50 * 80 -F 130
a) The one-photon absorptron spectrum in the region =24000-27000
506
LETTERS
24140 r
30 (48280) b,
24384 f 100 (48768)b)
-
24655 +
SO (493ll)b)
25129 c
80
(sozss)b)
25368 i 150
(5o737P)
25618 r ISO (51237)b) 26105 r 150 (s22ll)b) 26265 + 150 (sz3l)b) 26592 + 150 (,c3184)b)
cm-l
is diffuse [7] and thus assignment of
maxima in the TPI spectrum due to one-photon resonant two-photon ionization.
vibrationaltran-
Volume 80, number 3
CHEMKAL
and reached by two-photon absorption. The fluoranthene molecule belongs to the CZv [lo] symmetry group. Members of this group do not possess a center of symmetry and thus all excited states are both oneand two-photon allowed [3] _ The correspondence between the *maxima in the one-photon absorption and the TPI spectra seen in table 1 is consistent with this. Although several of the peaks in the TPI spectrum coincide with O-O transitions of one-photon absorption bands (see table 2), the TPI spectrum above a 42100 cm-r is more clearly resolved than the onephoton absorption spectrum_ It should, however, be kept in mind that in the photoabsorption spectrum one observes the disappearance of photons from the excitation beam independently of the ensurng processes, while in TPI we probe only the two-photon absorption as monitored by the ionization channel.
Table 2 Positions (in cm-r)
3.2. Decline in the TPIspectrum >48300
It is seen from fig. 1 that at hisser = 421 nm (* 23750 cm-l) corresponding to the onset of the first absorption band of fluoranthene, the TPI signal increases sharply as expected for one-photon resonant two-photon transitions. The TPI signal, however, unexpectedly decreases for hl,,,,<414nm. This dechne in the TPI spectrum (fig. 1) was not observed for the other molecules - pyrene [I] and TMPD - studied by us using the same technique. It is known [4,12,13] that the long (=lWg s) hfetimes of intermediate states, as opposed to the short (= lo-l5 s) lifetimes of virtual states, not only contribute to the probability for upward transitions but also allow for other processes which under favorable time and energy conditions compete effectively with multiphoton ionization_ Since the decline in the TPI
Positions of O-O transitions one-photon absorption theory
2 3 4 5 6 7 8
af energies
cm-l
of the O-O transitions of the posstble electronic states of fluoranthene
Possible excited electronic statea)
1
15 June 1981
PHYSICS LETTERS
26778 [S] 2660 [S] 28311c) [5] 2S400c) [S] 3169Sc) [5] 318OOc) [S] 36135c) [5] 36000c) [S] 40974c) [S] 43200 [8] 41377 [S] 40400 [S] 43474 [5] 45200 [S] 43474 [S] - 46100 [S]
TPI (present work) experiment ~26000 [5] 24670 [7] 278006) [S] 2784Oc) [7] 3osooc) [5] 3091oc) [7] 348OOc) [5] 34s4oc) [7] 382OOc) [S] 38170 [S] 40000 [S] 39530 [7] 42400 [S] 42460 [7] 43450
[S]
30 (u’ = o) ‘4 80 (Y’ = 3)b)
24140 24655
+ +
38105
+ SO
39500
+ 100
42105
f 100
43940
* 100
a) Although for the fust, fifth, and sixth electronic states the energy positions assigned to the respective O-O transitxons [5,7,S J are jn agreement with the corresponding resonances in the TPI spectrum, a srnaRshift is observed in the O-O transitions of the seventh and eighth states which ma$ indicate that the two-photon character of these states is quite significant. b) Our analysis suggests that the S 1,~ + Su,u transition energy should be 24140 cm-l. The energy 24670 cm” associated with this transition in ref. [7] shouid correspond rather to the Slg~+ Soi t&m&ion. C) These
states cannot be seen in the TPI spectrum because they lie outside the energy range of observation (see fg. 1). 507
Volume 80, number 3
CHEMICAL PHYSICS LETTERS 31s
310
-_^ i0 ’
15 June 1981 7 3L
320
I
I
*
I
--TPI-SPECTRLIK -tip I -SPECTRUM
OP! -ThRESHOU 13_83’0_021e\
1 &cl
I 625
I 630
I 635
I 640
I 645
EXCITATION WAVELENGTH Inml
I 65
2
I
400
450 Lf3SER
500 bJR\IELENGTH
550
600
f
[nml
Fig. 2. Two-photon iomzation (TPI) current rtsa function of the laser wavelength (curve 1). The section of this spectrum above 400 nm (curve 2) is magmfied b; a factor of 6,.7 The arrow indicates the position of the photoionization threshold. The literature value of 4.89 eV 1161 would hzve corresponded to iqaser = 507 nm. Insert. The one-photon ionization (OPJ) current (dashed line) as a function of the rasr \vdVekngth (upper s&e). The arrow rndicates the OPI threshold_ For comparison, the TPI current (solid line) as a function of the laser wavelength (lower scale) is also presented.
current sets in when TPI occurs via the one-photon resonant S, manifold and since autoionization usual.ly is faster than dissociation, it seems reasonable to attribute the decrease in the T’PI signal above 48300 cm-l to molecular dissociation from the S,, state * , a process competing with two-photon ionization via this state. The molecule probably starts to predissociate
’ Fluoranthene, although planar in the ground state So, is considered [6] non-planar m Sr _ Also, the absorption and fluorescence spectra of fhroranthene are diffuse [6-g], due perhaps to the non-planarity of Sr and the associated change of the bond lengths which calcnlations suggest must be considerable, particularly for the single C-C bonds of the fivemember ring of fhroranthene. Both factors would lower the energy required for molecular dissociation from S,,,.
Volume 80, number 3
CHEMICAL
PHYSICS
15 June 1981
LETTERS
at energies just higher *han the energy (24140 cm-l) of SI,-, , where two-photon ionization still competes effectively with dissociation. As the photon energy increases, direct dissociation seems to predominate, so that the TPI current goes virtually to “‘zero” at energies of the one-photon resonant transition of -27320 cm-l (=339 eV)+ wluch corresponds to a position above S1,8. Studies using tunable pulsed lasers in the picosecond range would undoubtedly elucidate the competitive processes of multiphoton ionization and molecular dissociation_
Table 3 Positions (in cm-‘) of the observed maxima in the TPI spectmm of TMPD inn-pentane (two-photon resonant transitions)
4. TPI and OPI spectra of TMPD in n-pentane
4.2. OPI spectnm
4-J. TPI spectrum
To ensure that the 3.88 eV value for the 1; of TMPD in n-pentane determined from the TPI spectrum is not influenced by any higher-order multiphoton process **, we measured (as described in section 2) the OPI spectrum of TMPD from 310 to 325 run. This spectrum is shown by the dashed line in the insert of fig_ 2 where, for comparison, the TPI spectrum (solid line) is also presented_ The OPI current decreases smoothly with increasing Alaser, and it asymptotically reaches a “zero level” at 324 + 3 ML This “onset” corresponds to an 1: value of 3 -83 + 0.02 eV, in agreement with the TPI 1: value of 3 -88 f 0.05 eV_ Very interestingly, the present 1L value for TMPD in n-pentane is = 1 eV lower than the literature value of 4.89 eV [16] _ The latter value would have corresponded to Alaser = 507 nm in the TPI spectrum_
The TPI spectrum of =S X 10v6 mol dm-3 solutions of TMPD in rz-pentane in the Alaser range from 260 to ~650 nm (corresponding to two-photon transitions from 180 to a:325 nm) is shown in fig. 2. The TPI spectrum comprises 15 individual corrected spectra, combined by normalization in the respective regions of overlap. Unlike fluoranthene, the TPI current for TMPD shows a monotonic slow increase with decreasing Alaser until, at Alaser = 425 run (= 47000 cm-l), it rises very steeply. As mentioned earlier, such sharp rise is expected when one-photon resonant two-photon ionization is possrble. On the long-wavelength side, the photocurrent approaches a “zero level” at Alaser = 639.1 k 4.1 run. This corresponds to a two-photon energy of 3.88 i- 0.05 eV which we associate with the It of TMPD in liquid n-pentane. The TPI spectrum of TMPD is rich in structure. The positions of the peaks in the TPI spectrum are given in table 3 _The one-photon absorption spectrum [ 1 l] shows two wide bands with maxima at = 38010 cm-1 and half-widths of z 4600 cm-1 and a49530 cm-l. The TMPD molecule belongs to the D,n symmetry group [15] and transitions from the totally symmetric (even-parity) ground state to upper evenparity states are one-photon forbidden, but two-photon allowed. Therefore, as in the case of pyrene [ 11, we ascribe the observed resonances in the TPI spectrum to two-photon excitation of one-photon forbidden autoionizing states at energies >3 1259 cm-I _ *Dissociation energies of = 3.6 eV have been reported 1141 for the single bonds of benzyl-benzyltype molecules.
31451 31778 32353 32745 33791 34673 35229 35948 37026 37487
f f + + t + + f r f
100 30 100 320 180 100 160 160 100 100
38039 38804 40621 42582 44444 46373 48183 48598 49466 49771
5. The photoionization
r 200 c 130 f 260 + 160 + 160 2 200 f 100 2 80 -A100 f 80
threshold of molecules in
liquid media
The accurate determination of the ionization threshold of molecules in liquid media is extremely important in efforts to link the properties of isolated molecules with those in the liquid phase and in efforts to assess the effect of the medium on charge-separated states. For this reason it has been the subject of many studies over the last two decades (see, e-g_ ref. [2] and references therein). Despite these efforts, basic
** The intensity dependence of the TPI signal near the photoiomzation threshold could not be measured because of the weak signal [l] _
509
Volume 80, number 3
CHEMICAL PHYSICS L.E-iTERS
questions remain concerning both the accurate measurement of Ii and the effect of the medium on it _
Concerning the measurement of IL, attention should be drawn to the much lower values ofit obtained by us compared with those obtained by others [2,16,17] _ As in any measurement of a threshold quantity, the value determined is a function of the sensitivity of the method employed_ We believe that our ability to detect very small signals (see ref. [l] and figs. 1 and 2) enables us to locate Zi more accurately_ This advantage of our method removes the need of fitting assumed “threshold-law” functions to the data “close” to the threshold to infer It, as has been practised by others [16-19]_ Analyses based on threshold laws are beset with serious problems: lack of knowledge of the energy region over which the assumed threshold law is valid (usually this region is of the order of a few meV), lack of accurate knowledge of the angular momentum of the electron prior to ejection, which determines the threshold behavior itself [20] ; and the serious assumption that the ionizatron ISdirect. Clearly the present work shows that photoionization is strongly influenced by autoionization. Concerning the effect of the medium on the photoionization threshold, attention is drawn to the following recent findings: (1) The 1: ofpyrene [if, TMPD, and fluoranthene m n-pentane as measured with the present technique are, respectiveIy,4.80 + 0_02,3.88 rfr0.05, and 4.50 t 0.05 eV (i-e_ 2_61,2_87, and 322 eV lower than the corresponding photoionization thresholds in the gas phase**). These differences are much larger than those inferred from other data on a number of molecules in non-polar liquids (see, e-g_ refs. [2,16,17]). (2) Recent studies [24,25] using strongly electronattaching molecules at high concentrations to probe the photoionization process for solute molecules in dielectric liquids have shown that the value of 1: depends on the time evolution of the photoior-ization and electron attachment processes. It was inferred from those studies that the photoejected electron is captured by the electron attaching additive at times of= 10-14 s. (3) Dynamical studies of electron solvation, following photoionization in liquid water using laser pico** The 1i v&es [%.--331 are uncertainby Iessthan =02eV. 510
15 June 1981
second techniques [26], indicated that the visible absorption spectrum of solvated electrons appears at times <3X lo--13 S. The aforementioned recent findings seem to suggest that the photoionization process for a solute molecule in a liquid medium is completed at times*% 1 O-13 s independently of the nature of the medium _They also suggest that sub-picosecond dynamical effects of the medium on the photoiomzation process are probably responsible for the lowering of 1; compared with IA _
Acknowledgement The authors express their thanks to Dr. M .B. Robin (Bell Laboratories) for stimulating discussions_ +9 Studies of the photoionization threshold of molecules in polar liquids using the TPI technrque are in progress.
References [ 11 K;.Siomos and L.G. Christophorou, Chem. Phys. Letters 72 (1980) 43. [2 J L. Kevan and B.C. Webster, eds., Electron-solvent and anion-solvent interactions (Eisevier, Amsterdam, 1976). [3] W L. Peticolas, Ann. Rev Phys.Chem. 18 (1967) 233. [4] 1SI.B.Robin, Appl. Opt. 19 (1980) 3941. [Sl E. Herlbronner.J_P.Weber, J_ Michl and R. Zahradnik, Theoret. Chim. Acta 6 (1966) 141. 1‘31 LB. Berlrnan, H-0. Wu-th and OJ. Steingraber, J. Am. Chem. Sot. 90 (1968) 566.
171 J. MichI, J. Mol. Spectry. 30 (1969) 66. 181 J. Kolc, E.W. Thulstrup and I_ MichI, J. Am. Chem.
Sot. 96 (1974) 7188. K. Sxomos and L.C. Christophorou, Chem. Phys. Letters., to be submitted for publicationIlO1 E.W_ Thuistrup and J.H. Eggers, Chem. Phys. Letters 1 (1968) 690. 1111 G. Kourouklis, K. Siomos and LG. Christophorou, J. Mol. Spectry., to be published. WI W.M_ Jackson and C.S. Lin, Intern. J. Chem. Kinetics 10 (1978) 945_ f131 P-A. Freedman, Can. J. Phys. 55 (1977) X387_ f141 K-B. Eisenthal, W.L. Peticolas and K-E_ Rieckhoff, J. Chem. Phys. 44 (1966) 4492. [IsI M _ Goeppert-Mayer and K.J. McCaUum, Rev. Mod. Phys. 14 (1942) 248.
PI
1161 R-AA.Holroyd and R-L. Russell, 3. Phys. Chem. 78 (1974) 2128.
Volume 80, number 3
CHEMICAL
1171 A. Bernas, D. Grand and E. Amouyal, J. Phys. Chem. 84 (1980) 1259. [18] W.F. Schmidt and W. DBidissen, Z. Naturforsch. 33a (1978) 1393. [ 191 U. Sowada and R-A. Hohoyd, J. Chem. Phys. 70 (1979) 3586. 1201 H. Hotop and WC. Lineberger, J. Chem. Phys. 58 (1973) 2379. [21] W.F. Frey, R.N. Compton,W.T. Naff and H.C. SchweinIer, Intern. J. Mass Spectrom. Ion Phys. 12 (1973)
PHYSICS LETTERS
ISJune
[22]
R. Roschi and W. Schmidt, Tetrahedron Letters (1972) 2577. [23] Y. Nakato, M. Ozaki. A. Egawa and H. Tsubomura, Chem. Phys. Letters 9 (1971) 615 [24] K. Siomos, G. Kouroukiis, L-G Christophorou and J.G. Carter, Radiat. Phys. Chem. 15 (1980) 313. [25] K. Siomos, G. KouroukIis, L.G. Christophorou and J.G. Carter, Rddiat. PhysChem., to be published. [26] J.M. Wiesenfeid and EP. Ippen, Chem. Phys. Letters 73 (1980) 47.
19.
511
Voltime 80. number3
LASER TWO-PHOTON IONIZATION OF ORGANIC MOLECULES
15 June 1981
IN DIELECTRIC
LIQUIDS *
K. SIOMOS’, G. KOUROUKLIS and L-G. CHRISTOPHOROU* Atomic. Molecular. and High Voltage Physics Group, Health and Safety Research Division, Oak Ridge Natior;al Laboratory. Oak Ridge, Tennessee 37830, Recewed
26 Februaq
1981;
USA in final Form 14 March 1981
The iotiation thxeshold of fluoranthene and of TMPD in rz-pentane was determmed by laser two-photon ionization (TPI) and found to be 450 c- 0.05 and 3 88 + O-05 eV respectmely. For both molecules the TPI spectra show distinct structure due to autoionization. For fluoranthene the TPI spectrum suggests that the molecule dissociates via the fist excited singlet state.
1_ Introduction
Recently [l] ,we descrrbed a two-photon ionization (TPI) technique suitable for liquid-phase studies, which we employed to determine the photolonization threshold 1: of solute molecules embedded in dielectric liquids_ While one-photon ionization is limited due to the absorption of the host medium, two-photon transitions are at an energy twice that of the individual photons absorbed, and thus ionization thresholds of molecules in liquids as high as = 10 eV can be reached with frequency-tunable dye lasers. The TPI technique is not only suitable for accurate measurement of Ii but also for obtaining information on molecular states embedded in the ionization continuum. The 1; is lower than the corresponding ionization threshold I& of a molecule in the gas phase (see, e.g. ref. 121 and references therein). Distinct peaks can appear in the TPI spectrum when the energy of a molecular state in the ionization continuum is resonant with the two-photon energy, assuming the parity of this state is the same as that of the ground state for molecules with a center of sym-
* Research sponsored by the Office of Health and Environmental Research, US Department of EnerOT, under contract W-740Seng-26 with the Union Carbide Corporation_ * Aisn at Department of Physics, The University of Tennessee, Knoxville, Tennessee 37916, US.4.
504
metry [3] and thus parts of a molecular spectrum that do not avail themselves of one-photon spectroscopy can be investigated [ 11 _In addition, the TPI technique can be employed to probe photophysical processes resulting from two-photon transitions which do not lead to ionization [4] _ In this paper, results are reported and discussed on photoionization of fluoranthene and N,N,N’,N’-tetramethyl-p-phenylenediamine (TMPD) in liquid rr-pentane _Fluoranthene is a non-altemant conjugated hydrocarbon of environmental interest and with “peculiar” absorption and fluorescence characteristics [S-9]. TMPD has a low ionization onset, and for th& reason it has been investigated by various methods whose results can be compared with those of the TPI technique. The potential of the TPI technique and the basic significance of the TPI data for molecules in liquid media are indicated.
2. Experimental The experimental apparatus has been described [ 11. The beam from a frequency-tunable nitrogenlaser-pumped dye laser (Molectron UV lOOO/DL 200) is focused on the photoconductivlty cell inducing two-photon transitions which result in ionization of the solute. The TPI signal from a parallel-plate chargesensitive detector (maintained at apotential difference
Volume 80, number 3
15 June 1981
CHEMICAL PHk!SICS LETTERS
[I] for further discussion) are plotted as a function
of 3 kV corresponding to an electric field of 15 kV cm-l) after linear amplification is averaged using a box-car integrator. The dye laser beam leaving the cell is monitored by a photodiode of known spectral responsivity and is sampled by a second boxcar integrator. The output signals of both integrators after digitization are fed into a PDP-10 computer for analysis_ The TPI signal after intensity correction (division by the square of the light intensity after correction for the spectral responsivity of the photodiode; see ref.
Of %aserDue to the low 1: value of TMPD, it was possible to measure the one-photon ionization (OPI) spectrum for TMPD in n-pentane near the ionization threshold. For this, the apparatus was slightly modified to accommodate a frequency-doubling unit. The OPI current fcr ThWD in n-pentane near the threshold was measured as a function of the second harmonic of the dye laser generated in an angle-tuned KDPcrystal.
:86
-PI-THRESHCLO [4_s0t0_0sleV
1
\ I
I
380
400
I
420
1
440
LASER
I
460
I
I
480
WAVELENGTH
500
I
520
I
540
5 Q
(nml
Fig_ 1. Two-photon ionization(TPI) currentas a function of the Iaserwavelength(curve 1). The sectionsof this spectrumabove 440 nm (curve2) and 500 nm (curve 3) are ma@fied by factors of 17 and 86, respectively.The arrowindicatesthe position of
the photoionization threshold.
505
Volume 80, number 3
CHEMICAL
PHYSICS
The liquid n-pentane used in this study was Phillips Research Grade (9998%) and was purified both chemicaily and under vacuum before use [I] _ Fluoranthene and TMpD were of a quoted purity 9999 and 98%, respectively_ They were used without fur*her purification. Au measurements were made under vacuum and at room temperature_
of fluorauthene
3. The TPX spectrum
in n-pentane
The TPf spectrum of =5 X 10d6 mof dme3 sofutions of fluoranthene in PI-pentane is shown in fig_ 1. The photocurrent is plotted as a function of Alaser in the spectral region 366 to ~560 run which corresponds to two-photon transitions in the range 188 to e280 nm. The spectrum presented comprises ten individual corrected spectra which were combined by
Table 1 Positions (in cm-l)
15 June 1981
normalization in the respective regions of overlap. The TPI spectrum shows three important features: (I) The photocurrent declines sharply for A,,,,, > 420 run and then reaches asymptotically a ‘stero levei” as the wavelength increases. (2) Distinct structure is observed over the entire energy range investigated. (3) The photocurrent declines for A,,,,, < 414 MI. Featuxe (1) was utilued to determine the 1: of fluoranthene in n-pentane. The photocurrent reaches asymptotically a “zero lever” [l] at 551 .O k 6.2 nm. This corresponds to an energy for a two-photon transition of 4.50 rt 0.05 eV, which we identify with the 1: of fluoranthene in tr-pentane.
This structure is attributed to autoionization of discrete states embedded in the ionization continuum
of the maxima in the one-photon absorption and TPL spectra of fhxoranthene in n-pentane
One-photon absorption [ 111
36742 +
80
j 37390 * 80 1 38105 + 80 38531 r 39500 40684 41205 42474
ii i *
80 100 130 100 50
43450 f 400 44705 *
80
45400 + 130 45842 + 130 46980 r 200
TPI two-photon resonant tmnsrtions (present work) 36742 +
sitionswasnot possible. b, Theseare the positions of the
One-photon absorption a)
38105 38531 38789 39500 40684 41205 42105 42842 43474 43579 43940 44705 44974 4.5168 45421 45868 46816 47079 47289
T?I one-photon resonant transitions (present work)
SO
37390 5 80
+ SO t 130 i 130 + 100 t 180 f 80 i 100 + 50 f 80 -c 50 + 100 f 80 c 50 c 80 i 100 -c 130 & 50 * 80 -F 130
a) The one-photon absorptron spectrum in the region =24000-27000
506
LETTERS
24140 r
30 (48280) b,
24384 f 100 (48768)b)
-
24655 +
SO (493ll)b)
25129 c
80
(sozss)b)
25368 i 150
(5o737P)
25618 r ISO (51237)b) 26105 r 150 (s22ll)b) 26265 + 150 (sz3l)b) 26592 + 150 (,c3184)b)
cm-l
is diffuse [7] and thus assignment of
maxima in the TPI spectrum due to one-photon resonant two-photon ionization.
vibrationaltran-
Volume 80, number 3
CHEMKAL
and reached by two-photon absorption. The fluoranthene molecule belongs to the CZv [lo] symmetry group. Members of this group do not possess a center of symmetry and thus all excited states are both oneand two-photon allowed [3] _ The correspondence between the *maxima in the one-photon absorption and the TPI spectra seen in table 1 is consistent with this. Although several of the peaks in the TPI spectrum coincide with O-O transitions of one-photon absorption bands (see table 2), the TPI spectrum above a 42100 cm-r is more clearly resolved than the onephoton absorption spectrum_ It should, however, be kept in mind that in the photoabsorption spectrum one observes the disappearance of photons from the excitation beam independently of the ensurng processes, while in TPI we probe only the two-photon absorption as monitored by the ionization channel.
Table 2 Positions (in cm-r)
3.2. Decline in the TPIspectrum >48300
It is seen from fig. 1 that at hisser = 421 nm (* 23750 cm-l) corresponding to the onset of the first absorption band of fluoranthene, the TPI signal increases sharply as expected for one-photon resonant two-photon transitions. The TPI signal, however, unexpectedly decreases for hl,,,,<414nm. This dechne in the TPI spectrum (fig. 1) was not observed for the other molecules - pyrene [I] and TMPD - studied by us using the same technique. It is known [4,12,13] that the long (=lWg s) hfetimes of intermediate states, as opposed to the short (= lo-l5 s) lifetimes of virtual states, not only contribute to the probability for upward transitions but also allow for other processes which under favorable time and energy conditions compete effectively with multiphoton ionization_ Since the decline in the TPI
Positions of O-O transitions one-photon absorption theory
2 3 4 5 6 7 8
af energies
cm-l
of the O-O transitions of the posstble electronic states of fluoranthene
Possible excited electronic statea)
1
15 June 1981
PHYSICS LETTERS
26778 [S] 2660 [S] 28311c) [5] 2S400c) [S] 3169Sc) [5] 318OOc) [S] 36135c) [5] 36000c) [S] 40974c) [S] 43200 [8] 41377 [S] 40400 [S] 43474 [5] 45200 [S] 43474 [S] - 46100 [S]
TPI (present work) experiment ~26000 [5] 24670 [7] 278006) [S] 2784Oc) [7] 3osooc) [5] 3091oc) [7] 348OOc) [5] 34s4oc) [7] 382OOc) [S] 38170 [S] 40000 [S] 39530 [7] 42400 [S] 42460 [7] 43450
[S]
30 (u’ = o) ‘4 80 (Y’ = 3)b)
24140 24655
+ +
38105
+ SO
39500
+ 100
42105
f 100
43940
* 100
a) Although for the fust, fifth, and sixth electronic states the energy positions assigned to the respective O-O transitxons [5,7,S J are jn agreement with the corresponding resonances in the TPI spectrum, a srnaRshift is observed in the O-O transitions of the seventh and eighth states which ma$ indicate that the two-photon character of these states is quite significant. b) Our analysis suggests that the S 1,~ + Su,u transition energy should be 24140 cm-l. The energy 24670 cm” associated with this transition in ref. [7] shouid correspond rather to the Slg~+ Soi t&m&ion. C) These
states cannot be seen in the TPI spectrum because they lie outside the energy range of observation (see fg. 1). 507
Volume 80, number 3
CHEMICAL PHYSICS LETTERS 31s
310
-_^ i0 ’
15 June 1981 7 3L
320
I
I
*
I
--TPI-SPECTRLIK -tip I -SPECTRUM
OP! -ThRESHOU 13_83’0_021e\
1 &cl
I 625
I 630
I 635
I 640
I 645
EXCITATION WAVELENGTH Inml
I 65
2
I
400
450 Lf3SER
500 bJR\IELENGTH
550
600
f
[nml
Fig. 2. Two-photon iomzation (TPI) current rtsa function of the laser wavelength (curve 1). The section of this spectrum above 400 nm (curve 2) is magmfied b; a factor of 6,.7 The arrow indicates the position of the photoionization threshold. The literature value of 4.89 eV 1161 would hzve corresponded to iqaser = 507 nm. Insert. The one-photon ionization (OPJ) current (dashed line) as a function of the rasr \vdVekngth (upper s&e). The arrow rndicates the OPI threshold_ For comparison, the TPI current (solid line) as a function of the laser wavelength (lower scale) is also presented.
current sets in when TPI occurs via the one-photon resonant S, manifold and since autoionization usual.ly is faster than dissociation, it seems reasonable to attribute the decrease in the T’PI signal above 48300 cm-l to molecular dissociation from the S,, state * , a process competing with two-photon ionization via this state. The molecule probably starts to predissociate
’ Fluoranthene, although planar in the ground state So, is considered [6] non-planar m Sr _ Also, the absorption and fluorescence spectra of fhroranthene are diffuse [6-g], due perhaps to the non-planarity of Sr and the associated change of the bond lengths which calcnlations suggest must be considerable, particularly for the single C-C bonds of the fivemember ring of fhroranthene. Both factors would lower the energy required for molecular dissociation from S,,,.
Volume 80, number 3
CHEMICAL
PHYSICS
15 June 1981
LETTERS
at energies just higher *han the energy (24140 cm-l) of SI,-, , where two-photon ionization still competes effectively with dissociation. As the photon energy increases, direct dissociation seems to predominate, so that the TPI current goes virtually to “‘zero” at energies of the one-photon resonant transition of -27320 cm-l (=339 eV)+ wluch corresponds to a position above S1,8. Studies using tunable pulsed lasers in the picosecond range would undoubtedly elucidate the competitive processes of multiphoton ionization and molecular dissociation_
Table 3 Positions (in cm-‘) of the observed maxima in the TPI spectmm of TMPD inn-pentane (two-photon resonant transitions)
4. TPI and OPI spectra of TMPD in n-pentane
4.2. OPI spectnm
4-J. TPI spectrum
To ensure that the 3.88 eV value for the 1; of TMPD in n-pentane determined from the TPI spectrum is not influenced by any higher-order multiphoton process **, we measured (as described in section 2) the OPI spectrum of TMPD from 310 to 325 run. This spectrum is shown by the dashed line in the insert of fig_ 2 where, for comparison, the TPI spectrum (solid line) is also presented_ The OPI current decreases smoothly with increasing Alaser, and it asymptotically reaches a “zero level” at 324 + 3 ML This “onset” corresponds to an 1: value of 3 -83 + 0.02 eV, in agreement with the TPI 1: value of 3 -88 f 0.05 eV_ Very interestingly, the present 1L value for TMPD in n-pentane is = 1 eV lower than the literature value of 4.89 eV [16] _ The latter value would have corresponded to Alaser = 507 nm in the TPI spectrum_
The TPI spectrum of =S X 10v6 mol dm-3 solutions of TMPD in rz-pentane in the Alaser range from 260 to ~650 nm (corresponding to two-photon transitions from 180 to a:325 nm) is shown in fig. 2. The TPI spectrum comprises 15 individual corrected spectra, combined by normalization in the respective regions of overlap. Unlike fluoranthene, the TPI current for TMPD shows a monotonic slow increase with decreasing Alaser until, at Alaser = 425 run (= 47000 cm-l), it rises very steeply. As mentioned earlier, such sharp rise is expected when one-photon resonant two-photon ionization is possrble. On the long-wavelength side, the photocurrent approaches a “zero level” at Alaser = 639.1 k 4.1 run. This corresponds to a two-photon energy of 3.88 i- 0.05 eV which we associate with the It of TMPD in liquid n-pentane. The TPI spectrum of TMPD is rich in structure. The positions of the peaks in the TPI spectrum are given in table 3 _The one-photon absorption spectrum [ 1 l] shows two wide bands with maxima at = 38010 cm-1 and half-widths of z 4600 cm-1 and a49530 cm-l. The TMPD molecule belongs to the D,n symmetry group [15] and transitions from the totally symmetric (even-parity) ground state to upper evenparity states are one-photon forbidden, but two-photon allowed. Therefore, as in the case of pyrene [ 11, we ascribe the observed resonances in the TPI spectrum to two-photon excitation of one-photon forbidden autoionizing states at energies >3 1259 cm-I _ *Dissociation energies of = 3.6 eV have been reported 1141 for the single bonds of benzyl-benzyltype molecules.
31451 31778 32353 32745 33791 34673 35229 35948 37026 37487
f f + + t + + f r f
100 30 100 320 180 100 160 160 100 100
38039 38804 40621 42582 44444 46373 48183 48598 49466 49771
5. The photoionization
r 200 c 130 f 260 + 160 + 160 2 200 f 100 2 80 -A100 f 80
threshold of molecules in
liquid media
The accurate determination of the ionization threshold of molecules in liquid media is extremely important in efforts to link the properties of isolated molecules with those in the liquid phase and in efforts to assess the effect of the medium on charge-separated states. For this reason it has been the subject of many studies over the last two decades (see, e-g_ ref. [2] and references therein). Despite these efforts, basic
** The intensity dependence of the TPI signal near the photoiomzation threshold could not be measured because of the weak signal [l] _
509
Volume 80, number 3
CHEMICAL PHYSICS L.E-iTERS
questions remain concerning both the accurate measurement of Ii and the effect of the medium on it _
Concerning the measurement of IL, attention should be drawn to the much lower values ofit obtained by us compared with those obtained by others [2,16,17] _ As in any measurement of a threshold quantity, the value determined is a function of the sensitivity of the method employed_ We believe that our ability to detect very small signals (see ref. [l] and figs. 1 and 2) enables us to locate Zi more accurately_ This advantage of our method removes the need of fitting assumed “threshold-law” functions to the data “close” to the threshold to infer It, as has been practised by others [16-19]_ Analyses based on threshold laws are beset with serious problems: lack of knowledge of the energy region over which the assumed threshold law is valid (usually this region is of the order of a few meV), lack of accurate knowledge of the angular momentum of the electron prior to ejection, which determines the threshold behavior itself [20] ; and the serious assumption that the ionizatron ISdirect. Clearly the present work shows that photoionization is strongly influenced by autoionization. Concerning the effect of the medium on the photoionization threshold, attention is drawn to the following recent findings: (1) The 1: ofpyrene [if, TMPD, and fluoranthene m n-pentane as measured with the present technique are, respectiveIy,4.80 + 0_02,3.88 rfr0.05, and 4.50 t 0.05 eV (i-e_ 2_61,2_87, and 322 eV lower than the corresponding photoionization thresholds in the gas phase**). These differences are much larger than those inferred from other data on a number of molecules in non-polar liquids (see, e-g_ refs. [2,16,17]). (2) Recent studies [24,25] using strongly electronattaching molecules at high concentrations to probe the photoionization process for solute molecules in dielectric liquids have shown that the value of 1: depends on the time evolution of the photoior-ization and electron attachment processes. It was inferred from those studies that the photoejected electron is captured by the electron attaching additive at times of= 10-14 s. (3) Dynamical studies of electron solvation, following photoionization in liquid water using laser pico** The 1i v&es [%.--331 are uncertainby Iessthan =02eV. 510
15 June 1981
second techniques [26], indicated that the visible absorption spectrum of solvated electrons appears at times <3X lo--13 S. The aforementioned recent findings seem to suggest that the photoionization process for a solute molecule in a liquid medium is completed at times*% 1 O-13 s independently of the nature of the medium _They also suggest that sub-picosecond dynamical effects of the medium on the photoiomzation process are probably responsible for the lowering of 1; compared with IA _
Acknowledgement The authors express their thanks to Dr. M .B. Robin (Bell Laboratories) for stimulating discussions_ +9 Studies of the photoionization threshold of molecules in polar liquids using the TPI technrque are in progress.
References [ 11 K;.Siomos and L.G. Christophorou, Chem. Phys. Letters 72 (1980) 43. [2 J L. Kevan and B.C. Webster, eds., Electron-solvent and anion-solvent interactions (Eisevier, Amsterdam, 1976). [3] W L. Peticolas, Ann. Rev Phys.Chem. 18 (1967) 233. [4] 1SI.B.Robin, Appl. Opt. 19 (1980) 3941. [Sl E. Herlbronner.J_P.Weber, J_ Michl and R. Zahradnik, Theoret. Chim. Acta 6 (1966) 141. 1‘31 LB. Berlrnan, H-0. Wu-th and OJ. Steingraber, J. Am. Chem. Sot. 90 (1968) 566.
171 J. MichI, J. Mol. Spectry. 30 (1969) 66. 181 J. Kolc, E.W. Thulstrup and I_ MichI, J. Am. Chem.
Sot. 96 (1974) 7188. K. Sxomos and L.C. Christophorou, Chem. Phys. Letters., to be submitted for publicationIlO1 E.W_ Thuistrup and J.H. Eggers, Chem. Phys. Letters 1 (1968) 690. 1111 G. Kourouklis, K. Siomos and LG. Christophorou, J. Mol. Spectry., to be published. WI W.M_ Jackson and C.S. Lin, Intern. J. Chem. Kinetics 10 (1978) 945_ f131 P-A. Freedman, Can. J. Phys. 55 (1977) X387_ f141 K-B. Eisenthal, W.L. Peticolas and K-E_ Rieckhoff, J. Chem. Phys. 44 (1966) 4492. [IsI M _ Goeppert-Mayer and K.J. McCaUum, Rev. Mod. Phys. 14 (1942) 248.
PI
1161 R-AA.Holroyd and R-L. Russell, 3. Phys. Chem. 78 (1974) 2128.
Volume 80, number 3
CHEMICAL
1171 A. Bernas, D. Grand and E. Amouyal, J. Phys. Chem. 84 (1980) 1259. [18] W.F. Schmidt and W. DBidissen, Z. Naturforsch. 33a (1978) 1393. [ 191 U. Sowada and R-A. Hohoyd, J. Chem. Phys. 70 (1979) 3586. 1201 H. Hotop and WC. Lineberger, J. Chem. Phys. 58 (1973) 2379. [21] W.F. Frey, R.N. Compton,W.T. Naff and H.C. SchweinIer, Intern. J. Mass Spectrom. Ion Phys. 12 (1973)
PHYSICS LETTERS
ISJune
[22]
R. Roschi and W. Schmidt, Tetrahedron Letters (1972) 2577. [23] Y. Nakato, M. Ozaki. A. Egawa and H. Tsubomura, Chem. Phys. Letters 9 (1971) 615 [24] K. Siomos, G. Kouroukiis, L-G Christophorou and J.G. Carter, Radiat. Phys. Chem. 15 (1980) 313. [25] K. Siomos, G. KouroukIis, L.G. Christophorou and J.G. Carter, Rddiat. PhysChem., to be published. [26] J.M. Wiesenfeid and EP. Ippen, Chem. Phys. Letters 73 (1980) 47.
19.
511