The structure of 4-methylphenol and its water cluster revealed by rotationally resolved UV spectroscopy using a genetic algorithm approach

THE JOURNAL OF CHEMICAL PHYSICS 123, 044304 共2005兲 The structure of 4-methylphenol and its water cluster revealed by rotationally resolved UV spectroscopy using a genetic algorithm approach Grzegorz Myszkiewicz and W. Leo Meertsa兲 Molecular and Biophysics Group, Institute for Molecules and Materials, Radbound University Nijmegen, P.O. Box 9010, NL-6500 GL Nijmegen, The Netherlands Christian Ratzer and Michael Schmittb兲 Institut für Physikalische Chemie, Heinrich-Heine-Universität 40225 Düsseldorf, Germany 共Received 24 March 2005; accepted 27 May 2005; published online 2 August 2005兲 The structure of 4-methylphenol 共p-cresol兲 and its binary water cluster have been elucidated by rotationally resolved laser-induced fluorescence spectroscopy. The electronic origins of the monomer and the cluster are split into four sub-bands by the internal rotation of the methyl group and of the hydroxy group in case of the monomer, and the water moiety in case of the cluster. From the rotational constants of the monomer the structure in the S1 state could be determined to be distorted quinoidally. The structure of the p-cresol-water cluster is determined to be trans linear, with a O–O hydrogen bond length of 290 pm in the electronic ground state and of 285 pm in the electronically excited state. The S1-state lifetime of p-cresol, p-cresol-d1, and the binary water cluster have been determined to be 1.6, 9.7, and 3.8 ns, respectively. © 2005 American Institute of Physics. 关DOI: 10.1063/1.1961615兴 I. INTRODUCTION In the present study we present the determination of the geometry of p-cresol and its binary water cluster in the ground state and the electronically excited S1 state by means of rotationally resolved electronic spectroscopy. The determination of excited state structures of aromatic molecules gives insight into the changes of the electronic properties of the chromophore. In a series of investigations, we determined the structural changes of some para-disubstituted aromatic molecules such as p-fluorophenol,1 p-cyanophenol,2 and p-methylstyrene.3 In all these cases a quinoidal distortion along the a axis of the molecule is found, whereas for singly substituted benzenes a symmetric distortion 共benzenoid兲 of the aromatic ring is found.4 Changes of the electronic properties are also reflected in the change of the torsional barriers of symmetric 共e.g., CH3兲 or asymmetric tops 共e.g., OH兲, which are directly attached to the benzene ring. In recent publications it was shown how the different electronic properties of the fluorine and the cyanogroup influence the barrier to hydroxy torsion in both electronic states of p-fluorophenol and p-cyanophenol.1,2 4-methylphenol possesses two feasible large amplitude motions: the methyl rotation with a mixed V3 / V6 potential and the twofold hydroxy rotation. The molecular symmetry group that describes both motions is G12. In the p-cresol-water cluster, the OH torsion is quenched; instead an additional splitting due to the torsional motion of the water moiety with respect to the water symmetry axis is observed. This barrier is expected from the similar phenola兲 FAX: ⫹33 24 356 3311; electronic mail: leo.meerts@science.ru.nl FAX: ⫹49 211 81 15195; electronic mail: mschmitt@uni-duesseldorf.de b兲 0021-9606/2005/123共4兲/044304/7/$22.50 water system to be quite low. Thus, we have the case of twolow barrier torsional motions, which will split the origin into four sub-bands, an A共3兲-E pair due to the threefold motion and an A共2兲-B pair due to the twofold rotation. We use the superscripts 共2兲 and 共3兲 to distinguish the different A sublevels from the two- and threefold internal rotation from the A rotational constant. Since the rotational constants of this cluster are quite small, the overlapping of four subspectra leads to a very-congested overall spectrum in which it is difficult, if not impossible, to identify single rovibronic lines. However, the assignment of well-resolved lines to specific quantum numbers of the transitions is an indispensable prerequisite for an assigned fit of the molecular parameters. We have shown recently that in these cases of many overlapping lines, the use of a genetic algorithm approach for automated assignment of the spectra simplifies the interpretation and in many cases only makes it possible.5 4-methylphenol has found considerable interest as a chromophore and simple molecular model of the aromatic amino acid tyrosine.6 Proton transfer, electron transfer, and hydrogen transfer are important processes in clusters of tyrosine. E.g., in the water oxidizing complex of photosystem II, tyrosine is oxidized and reduced by hydrogen atom transfer involving tyrosine, water, and histidine in the active center.7 The electronic ground state of p-cresol has been investigated by infrared8,9 and Raman spectroscopy,6,10 by stimulated emission dip spectroscopy,11 and by dispersed fluorescence spectroscopy.12 Recently Lin et al. presented a massanalyzed threshold ionization spectrum of the p-cresol cation.13 The nature of conformational preferences in p-cresol has been studied theoretically by Richardson et al.14 They determined the most stable conformation of p-cresol on 123, 044304-1 © 2005 American Institute of Physics 044304-2 Myszkiewicz et al. J. Chem. Phys. 123, 044304 共2005兲 beam by an imaging optics setup consisting of a concave mirror and two plano-convex lenses. Data acquisition is performed and synchronized by means of a home-built real time 19 LINUX program. III. THEORY OF THE INTERNAL ROTATION FIG. 1. The most stable conformation and atomic numbering of p-cresol. MP2 level of theory to be eclipsed with respect to the methyl-group orientation, and the OH group on the same side of the aromatic ring as the eclipsed methyl hydrogen 共syn兲, c.f. Fig. 1. Several p-cresol-water clusters with up to four water molecules have been studied using dispersed fluorescence and- resonant two-photon ionization 共R2PI兲 spectroscopy in the region of the intermolecular vibrations.15,16 From the vibrational analysis it is deduced that the binary cluster is hydrogen bound via a trans-linear H bond as in the similar phenol-water system.17 II. EXPERIMENTAL SETUP The experimental setup for high-resolution laser-induced fluorescence 共LIF兲 is described in detail elsewhere.18 Briefly, the molecular-beam machine consists of three differentially pumped vacuum chambers that are linearly connected by two skimmers with orifice diameters of 1 mm 共first chamber兲 and 2 mm 共second chamber兲 for the reduction of the Doppler width to 25 MHz. The expansion chamber is evacuated by a 8000-l / s oil diffusion pump 共Leybold DI 8000兲, which is backed by a 250-m3 / h roots blower pump 共Saskia RPS 250兲 and a 65-m3 / h rotary pump 共Leybold D65B兲. The second chamber serves as a buffer chamber and is pumped by a 400-l / s turbo-molecular pump 共Leybold Turbovac 361兲, backed by a 40-m3 / h rotary pump 共Leybold D40B兲, maintaining a chamber pressure below 1 ⫻ 10−5 mbar. The third chamber is pumped by a 145-l / s turbo-molecular pump 共Leybold Turbovac 151兲 through a liquid-nitrogen trap and backed by a 16-m3 / h rotary pump 共Leybold D16B兲 resulting in a vacuum better than 1 ⫻ 10−6 mbar. The molecular beam is crossed at right angles with the laser beam 360 mm downstream from the nozzle. The laser system consists of a ring dye laser 共Coherent 899-21兲, which is pumped with 6 W of the 514-nm line of an Ar+-ion laser 共Coherent Innova 100兲. This light is coupled into an external-folded ring cavity 共Spectra Physics兲 for second-harmonic generation 共SHG兲. Typical UV powers are 10– 40 mW. The laser-induced fluorescence is collected perpendicular to the plane defined by the laser and molecular The coupling between the internal rotation of the methyl group and the overall rotation of the molecule and the coupling of the OH and the water torsion with the overall rotation are treated in the formalism of the principal axis method 共PAM兲.20,21 This is exact in the first 共symmetric-top兲 case, but only an approximation for the twofold 共asymmetric兲 top. Nevertheless as the hydrogen atom of the OH group and in the water moiety is light, and the inertial axes remain virtually unchanged during rotation of the top, the application of the PAM leads to reliable results. The Hamiltonian for the rotation-torsion interaction can be written as Ĥ = ĤR + ĤT + ĤRT 共1兲 with the rigid rotational Hamiltonian given by ĤR = AJ2a + BJ2a + CJ2c . 共2兲 The torsional Hamiltonian for the mixed V3 / V6 potential 共methyl torsion兲 is given by 2 ĤT = Fp2 + 1 兺 V3n共1 − cos 3n␣兲, 2 n=1 共3兲 and for a twofold V2 potential 共OH and water torsion兲 by 2 1 ĤT = F␳ + 兺 V2n共1 − cos 2n␣兲. 2 n=1 2 共4兲 F is the torsional constant and defined by F= h 8␲2rI␣ 共5兲 with r=1− 兺 g=a,b,c ␭2gI␣ . Ig 共6兲 Here the ␭g 共g = a , b , c兲 are the direction cosines between the inertial axes and the axis of internal rotation. I␣ is the moment of inertia of the internal rotor with respect to its symmetry axis and Ig are the principal moments of inertia of the entire molecule. In what follows, the treatment for the n = 2 and n = 3 potentials is identical. The coupling of internal and overall rotation is described by ĤRT:20,22 共2兲 ĤRT = FW␷共1兲 ␴ 共␳aJa + ␳bJb + ␳cJc兲 + FW␷␴ 共␳aJa + ␳bJb + ␳ cJ c兲 2 , 共7兲 where the first-order perturbation coefficients W␷共1兲 ␴ are zero 共2兲 共2兲 共3兲 for the nondegenerate A , A , and B levels. W␷␴ is nonzero for all levels. 044304-3 4-methylphenol-water cluster J. Chem. Phys. 123, 044304 共2005兲 The coefficients ␳g with g = a , b , c are defined as ␳a = I␣/Ia cos ␩ , 共8兲 ␳b = I␣/Ib sin ␩ , with ␩ as the angle between the projection vector of the n-fold rotor axis on the ab plane and the a axis. The nth-order perturbation coefficients W␷共n兲 ␴ in Eq. 共7兲 are defined by20 W␷共0兲 ␴ = E ␷␴ , F 共9兲 W␷共1兲 ␴ = − 2具␷, ␴兩p兩␷, ␴典, W␷共2兲 ␴ = 1 + 4F 兺 ␷⬘ 共10兲 兩具␷, ␴兩p兩␷⬘, ␴典兩2 , E ␷␴ − E ␷⬘␴ 共11兲 where 兩␷ , ␴典 are eigenfunctions of Eqs. 共3兲 and 共4兲, and E␷␴ are the respective eigenvalues with ␷ as the torsional state index. FIG. 2. 共a兲 Experimental and fitted rovibronic spectrum of the electronic origin of 4-methylphenol. 共b兲 10-GHz zoom of the 共a兲 spectrum. 共c兲 Inset showing three pairs of rovibronic transitions split by the twofold hydroxy group rotation in cresol. For the sake of clarity only the strongest transitions are shown and the splitting is marked with the horizontal lines. The solid line represents the experiment and the dashed line, the fit. All frequencies are given relative to the electronic origin at 35 331.257共10兲 cm−1. IV. THE GENETIC ALGORITHMS The experimental spectra were assigned automatically using a genetic algorithm-based fit described in detail in Refs. 5 and 23. We used the genetic algorithm 共GA兲 library 24 PGAPACK version 1.0, which can run on parallel processors. The calculations were performed on eight processors of a SUN UltraSPARC 333 MHz and on a 2.6-GHz personal computer 共PC兲 with two processors under LINUX. The genetic algorithm uses concepts copied from natural reproduction and selection processes. For a detailed description of the GA the reader is referred to the original literature on evolutionary or genetic algorithms.25–27 In the present work a population of 300 was used and the calculation was completely converged after 500 generations. The other parameters which control the genetic algorithm convergency are similar as in Ref. 5. The cost function used to describe the quality of the fit was defined as C fg = 100共1 − F fg兲 with the fitness function F fg: F fg = 共f,g兲 . 储f储储g储 共12兲 f and g represent the experimental and calculated spectra, respectively, and the inner product 共f , g兲 is defined with the metric W, which has the matrix elements Wij = w共兩j − i兩兲 = w共r兲 as 共f,g兲 = fTWg, 共13兲 and the norm of f as 储f储 = 冑共f , f兲 and similar for g. For w共r兲 we used a triangle function23 with a width of the base of ⌬w: w共r兲 = 再 1 1 − 兩r兩共 21 ⌬w兲 for 兩r兩 艋 2 ⌬w 0 otherwise. 冎 共14兲 Since the GA performs an intensity fit of the complete spectrum, much better information on the transition dipole moment 共TDM兲 orientation and linewidth parameters are gathered than from an intensity fit to a few individual lines. Thus, the GA results in improved values for the in plane angle ␪ of the TDM and also of the Lorentzian width of the individual rovibronic lines. Therefore, excited state lifetimes can be determined with a much higher accuracy than in lineshape fits of individual rovibronic lines. V. RESULTS AND DISCUSSION A. 4-methylphenol Figure 2 shows the rotationally resolved LIF spectrum of the electronic origin of p-cresol. The spectrum is a pure b-type spectrum. It is split into two components, A共3兲 and E by the internal rotation of the methyl group 共E␷,0 − E␷,±1兲. Their respective origins, which are 8218 MHz apart, are marked arrows. In two recent publications3,28 we showed how a combined fit to high-resolution rovibronic data and lowresolution torsional transitions can be used to improve the determination of the barriers and torsional constants in both electronic states. A, B, C, V3, V6, F, and ␩ were fitted for each electronic state to the rotationally resolved origin and to the torsional transitions from Ref. 29. These torsional transitions are obtained from the zero-order perturbation coefficient, c.f. Eq. 共9兲. The rovibronic spectrum was fitted using the genetic algorithm 共GA兲 approach, described in Ref. 5. The parameters that determine the intensity and line form are the rotational temperatures, the Lorentzian and Gaussian linewidths, and the projections of the transition dipole moment on the inertial axes. They can be fitted most favorably using the GA algorithm. The Lorentzian contribution to the linewidth of the Voigt profile with a fixed Doppler width of 25 MHz is determined to be 100± 10 MHz, equivalent to a S1-state lifetime of 1.6± 0.2 ns. The results of the fits are compiled in Tables I and II. 044304-4 J. Chem. Phys. 123, 044304 共2005兲 Myszkiewicz et al. TABLE I. Molecular parameters of 4-methylphenol and 4 -methyl关7D兴phenol from the GA fit. The ground-state rotational constants of 4-methylphenol have been kept fixed at the microwave values from Ref. 30. 4-methylphenol S1 S0 A B C V3 V6 F ␯0 ␯0共E兲 − ␯0共A共3兲兲 ␯0共A共2兲兲 − ␯0共B兲 共MHz兲 共MHz兲 共MHz兲 共MHz兲 共MHz兲 共cm−1兲 共cm−1兲 共MHz兲 共MHz兲 5494.570 5154.6共12兲 1456.963 1470.51共67兲 1160.200 1153.62共41兲 18.00共25兲 7.99共34兲 −13.8共45兲 −24.7共43兲 5.224共98兲 5.108共98兲 35 331.257共10兲 8218共16兲 90共13兲 4-methyl关7D兴phenol A B C V3 V6 F ␯0 ␯0共E兲 − ␯0共A共3兲兲 共MHz兲 共MHz兲 共MHz兲 共MHz兲 共MHz兲 共cm−1兲 共cm−1兲 共MHz兲 5442.16共94兲 5110.04共94兲 1417.59共37兲 1431.03共37兲 1133.45共27兲 1127.46共27兲 18.00共25兲 7.99共34兲 −13.8共45兲 −24.7共43兲 5.170共7兲 5.066共7兲 35 326.729共10兲 8656.2共50兲 Closer inspection of the spectrum shows that each of the rovibronic transitions is further split into two close lying sub-bands 关c.f. Fig. 2共c兲兴 due to the hindered –OH torsion. As a twofold torsion leads only to nondegenerate A共2兲 共␴ = 0兲 and B 共␴ = 1兲 levels, the first-order perturbation coefficients from Eq. 共10兲 are zero. The second-order perturbation coefficients which are nonzero 关Eq. 共11兲兴 are quadratic in the angular momentum and can therefore be incorporated in the rotational constants. Since the OH-torsional barrier is very high, the perturbations are small and both sub-bands can be fitted with the same set of effective rotational constants and an origin shift between the A共2兲 and B bands of 90 MHz. Determination of barriers and torsional constants in both electronic states demands more information from other spectroscopic techniques. In the microwave spectrum a splitting of the b-type transitions of 175 MHz has been observed.30 The selection rule for b-type transitions is ⌬␴ = ± 1, so that the splitting between the subtorsional levels amounts to 87.5 MHz. The frequency of the pure torsional transition has TABLE II. Torsional transitions of 4-methylphenol. The electronic transitions are labeled by the m quantum number and the symmetry of the subtorsional level ␴: m␴共S1兲 ← m␴共S0兲 for absorption bands and m␴共S1兲 → m␴共S0兲 for emission bands. p-cresol Exp.29 Fit. Diff. 1e ← 1e 2e ← 1e 3a1 ← 0a1 4e ← 1e 5e ← 1e 1e → 2e 0a1 → 3a1 1e → 4e 8218共16兲 15.1共5兲 51.7共5兲 77共2兲 125共2兲 18.5共10兲 52.5共10兲 80.0共20兲 8214 15.4 52.7 77.9 123.6 17.7 51.9 80.1 −4 +0.3 +1.0 +0.9 −1.4 −0.8 −0.6 +0.1 MHz cm−1 cm−1 cm−1 cm−1 cm−1 cm−1 cm−1 FIG. 3. Rovibronic spectrum of the electronic origin of 4-methyl关7D兴phenol. Frequencies are given relative to the electronic origin at 35 326.729共10兲 cm−1. been determined by IR spectroscopy to be 294 cm−1.9 The first overtone of the torsional vibration in the electronically excited state is determined to be 1199 cm−1. Using these transitions and fixing the torsional constant F at 690 GHz 共the value in phenol兲, we are able to fit the ground-state torsional V2 barrier to 1130 cm−1 and the excited state barrier to 4395 cm−1. Both values are slightly lower than the corresponding values in phenol 共1215 and 4710 cm−1, respectively兲.17 B. 4-methyl†7-D‡phenol Figure 3 shows the rotationally resolved LIF spectrum of the electronic origin of 4-methyl关7-D兴phenol. The spectrum is a pure b-type spectrum. Due to the heavier OD group, the twofold torsion is quenched in this system and the molecular symmetry group for the description of the molecule is G6. Consequently, the origin band consists only of a split pair of A共3兲 and E bands with a frequency spacing of 8656 MHz. Table I gives the molecular parameters that have been determined using the GA program. The lifetime for 4-methyl关7-D兴phenol is determined to be 9.7± 0.6 ns from a Lorentzian contribution to the total linewidth of 16.4± 1 MHz. The A共3兲E subtorsional splitting is very similar to the splitting in 4-methylphenol, also indicating similar potential barriers. As no torsional bands of 4-methyl关7-D兴phenol have been measured not enough data are available to fit the barriers to methyl rotation in this molecule. Nevertheless, since the effect of hydroxy deuteration on the methyl barrier is supposed to be small, we set barriers and torsional constants equal to those of 4-methylphenol. C. 4-methylphenol „H2O…1 Figure 4 presents the rotationally resolved electronic spectrum of the origin of the p-cresol-water cluster at 34 972.873 cm−1. The origin is split into an A共3兲 and an E sub-band due to the methyl torsion with a subtorsional splitting of 5459 MHz and into an A共2兲 and B sub-band due to the water torsion with a subtorsional splitting of 29 428 MHz. 044304-5 4-methylphenol-water cluster J. Chem. Phys. 123, 044304 共2005兲 FIG. 4. Rovibronic spectrum of the electronic origin of 4-methylphenol water. Frequencies are given relative to the electronic origin at 34 972.873共10兲 cm−1. The angle of the transition dipole moment with the a axis is determined from the relative intensities of a and b bands to be 76.70°. The rotation of 13.3° with respect to the monomer is in good agreement with a value of 20°, that is predicted by rotation of the inertial axis upon cluster formation on the basis of the ab initio structure. As in the case of phenol, the subtorsional structure due to the twofold water torsion can be fit using a rigid-rotor model with different rotational constants for both components. Similar to the phenol-water case, the B and C rotational constants of the 共␴ = 0兲 and 共␴ = 1兲 components are equal within their uncertainties. The A constants of the subtorsional components differ by about 11 MHz, indicating that the torsional axis of the water torsion is more or less parallel to the main inertial a axis of the cluster. This structure can be described as a trans-linear hydrogen bond configuration, with the water moiety acting as proton acceptor, like in phenol water. A spin statistics of 1:3 for the ratio of 共␴ = 0兲 to 共␴ = 1兲 further supports a structure with the water moiety symmetric with respect to the aromatic plane. This analysis clearly shows the convenience of the GA analysis of such a complex spectrum. Although by eye no periodicity can be observed at first glance, the GA is capable of easily finding the two splittings from the methyl torsion 共5459 MHz兲 and the water torsion 共29 428 MHz兲. Of course, the application of the GA-based-automated technique still requires the choice of the “correct” Hamiltonian for the problem under consideration. An easy way of finding spectral splittings 共and thus obtaining information about the required Hamiltonian兲 is to perform an autocorrelation of the spectrum. This method has been shown by Helm et al. to be helpful in the analysis of the phenol-water cluster spectrum31 and later by Remmers et al. in the analysis of the tunneling spectrum of the benzoic acid dimer.32 Nevertheless, if one component of the complete spectrum amounts to more than 50% of the total intensity, the GA is capable of fitting this part of the spectrum, although it might be concealed in the dense spectrum of other component共s兲 and would therefore be inaccessible to a classical analysis using line-positionassigned fits. From the difference of the partly fitted spectrum to the experimental one, the missing components can easily be recognized and subsequently fitted. In order to compute the geometrical rotational constants from the torsionally perturbed constants of Table III, we have to estimate the V2 potential energy barrier and the torsional constant F for the water torsion in both electronic states. However, only three pieces of information are available for the determination of these four parameters: the 共␴ = 0兲 / 共␴ = 1兲 subtorsional splitting of 29 428 MHz, and the difference of the A rotational constants for 共␴ = 0兲 and 共␴ = 1兲 in the electronic ground state and in the electronically excited state. Fixing F⬙ of the ground state to the value determined for phenol water28 共14.813 cm−1兲 共the ground-state acidities of phenol and cresol are similar, therefore one would expect similar torsional constants for the water moiety兲, a groundstate barrier V⬙2 of 182.6 cm−1, an excited state barrier V⬘2 of 125.3 cm−1, and a torsional constant F⬘ in the excited state of 14.9 cm−1 are calculated. These values are very close to the torsional barriers of phenol water. D. Determination of the structures The program pKrFit 共Ref. 4兲 was used to determine the structure of 4-methylphenol in the S0 and S1 states from the rotational constants of the two isotopomers described above. Due to the limited number of isotopomers in this study, we performed a fit to the r0 structure, which completely neglects the vibrational contributions from the different isotopomers. TABLE III. Molecular parameters of 4-methylphenol water. S0 ␴=0 A B C V3 V6 F ␩ ␯0 ␯0共A共2兲兲 − ␯0共B兲 ␯0共A共3兲兲 − ␯0共E兲 共MHz兲 共MHz兲 共MHz兲 共MHz兲 共MHz兲 共cm−1兲 共°兲 共cm−1兲 共MHz兲 共MHz兲 S1 ␴=1 3663.48共244兲 3652.19共53兲 765.60共59兲 765.98共31兲 637.00共48兲 637.71共31兲 18.00共25兲 −13.8共45兲 5.41共7兲 19.1共10兲 ␴=0 ␴=1 3590.87共244兲 774.78共59兲 640.76共48兲 3576.37共53兲 774.35共31兲 641.61共31兲 7.99共34兲 −24.7共43兲 5.21共7兲 19.1共10兲 34 971.891共10兲 34 972.873共10兲 29 428共7兲 5459共6兲 044304-6 Myszkiewicz et al. J. Chem. Phys. 123, 044304 共2005兲 TABLE IV. Structural r0 parameters of p-cresol. For atomic numbering see Fig. 1. All values are in pm. S0 B1共C1C2兲 B2共C2C3兲 B3共C3C4兲 B4共CO兲 B5共CCmethyl兲 139.9共1兲 140.5共3兲 138.9共1兲 138.3共6兲 151.3共6兲 S1 B1共C1C2兲 B2共C2C3兲 B3共C3C4兲 B4共CO兲 B5共CCmethyl兲 142.2共10兲 138.1共10兲 144.3共4兲 134.5共11兲 150.3共2兲 A simple model for the geometry has been adopted, that is given in Fig. 1. All aromatic CH bonds are set to the same value 共107.7 pm for the ground state and 107.4 pm for the excited state兲, the CH bonds of the methyl group are fixed at 109.0 pm. The OH bond lengths from the structural fits of phenol have been taken as constants 共96 pm in the S0 and 99 pm in the S1 states兲. Opposing CC bonds are set equal. Table IV gives the results for the fits of the S0 and the S1 structure. The aromatic ring expands upon electronic excitation quinoidally, as has been found for other paradisubstituted aromatic compounds. The CO and the CCmethyl bond lengths decrease, which are in agreement with a quinoidal distortion of the ring system. The relative shortening of these bands cannot be determined accurately, as both bonds nearly coincide with the main inertial a axis. Because the deuteration is at a position close to this axis, both bond lengths are strongly correlated. The results present therefore only one possible combination of the bond length decreases. From the above determined S0 and S1 structures of the cresol moiety, we fitted the structure of the p-cresol water to the geometric rotational constants. All geometry parameters in p-cresol have been kept fixed at the monomer values, the geometry of the water moiety has also been kept fixed. We imposed the symmetry constraint on the fit: The H atoms of the water moiety are symmetric with respect to the aromatic plane. Table V shows the results. The O–O hydrogen bond length decreases by 5.0 pm upon electronic excitation, imaging the increased acidity of p-cresol upon electronic excitation, while the OOC angle and the HOOC dihedral, describing the orientation of the water TABLE V. Structural r0 parameters of p-cresol water. The atomic numbering refers to Fig. 1, the subscript w refers to atoms of the water moiety. The dihedral angle is defined between the H1 atom of the water moiety, the O atom of water, the O7 atom of cresol, and the C1 atom of cresol. The second dihedral angle is determined by the symmetry constraint. S0 S1 r共OwO7兲 共pm兲 a共OwO7C1兲 共°兲 a共Hw1OwO7C1兲 共°兲 a共Hw2OwO7C1兲 共°兲 290.0共2兲 114.4共12兲 120共47兲 −120共47兲 r共OwO7兲 共pm兲 a共OwO7C1兲 共°兲 a共Hw1OwO7C1兲 共°兲 a共Hw2OwO7C1兲 共°兲 285.0共2兲 113.9共20兲 116共16兲 −116共16兲 FIG. 5. Top and side views of the ground and excited state structures of p-cresol water. For all other geometry parameters see Table V. moiety, nearly stay constant. Two different views of the ground and excited state structure are shown in Fig. 5. VI. CONCLUSIONS The value for the torsional barrier due to the internal rotation of the hydroxy group in p-cresol was found to be 1130 cm−1 in the ground state and 4395 cm−1 in the excited state. These values can be compared to the V2 barriers of other p-disubstituted phenols and of phenol itself, given in Table VI. Relative to phenol, para substitution with fluorine 共⫹mesomeric effect兲 and methyl 共⫹inductive effect兲 leads to a decrease of the torsional barrier in both electronic states, while the cyano group 共⫺mesomeric effect兲 increases the barrier. Electron accepting groups 共as cyano兲 stabilize a partial quinoidal structure of the p-substituted phenol with some double bond characters of the CO bond, while electron donating groups destabilize it. The lifetime of 4-methyl关7-D兴phenol determined from the Lorentz contribution to the Voigt profile 共9.7 ns兲 is substantially longer than the respective 4-methylphenol lifetime of 1.6 ns. A similar increase of the S1-state lifetime was found for phenol and 关7D兴phenol and was explained by a smaller probability for tunneling of the OD species through the barrier which separates the 1␲␴* from the 1␲␲* surface.33 The V3 and V6 barriers of the methyl-group torsion are low, as is typical for molecules with G12 symmetry. Barriers determined from 4-methylphenol and 4-methyl关7-D兴phenol are equal within the uncertainty, showing that the two torsional motions are mainly decoupled. TABLE VI. OH-torsional barriers of phenol, p-fluorophenol, p-cyanophenol, and p-methylphenol 共p-cresol兲. All values are in cm−1. V2 barrier p-fluorophenol p-methylphenol phenol p-cyanophenol S0 S1 References 1006 1130 1215 1420 1819 4395 4710 ⬎5000 1 This work 17 2 044304-7 4-methylphenol-water cluster From the rotational constants of both isotopomers, the structural change upon electronic excitation could be determined. As it is typical for para-disubstituted aromatics, the ring expands quinoidally upon excitation, while the two bonds in the para position decrease. The decrease of the CO bond length mirrors the shift of electron density from the oxygen atom to the aromatic ring, which takes place and is the reason for the increased acidity of phenols upon electronic excitation. The rotationally resolved electronic spectrum of the origin of the water cluster is split into an A共3兲E pair due to the methyl-group rotation and into an A共2兲B pair due to the torsional motion of the water moiety. The S1 lifetime of the p-cresol-water cluster determined from the Lorentz contribution to the Voigt profile is 42± 5 MHz, equivalent to a S1-state lifetime of 3.8± 0.5 ns, which is much shorter than the lifetime of the similar phenol-water cluster 共15 ns兲. The large increase of the lifetime by going from phenol to the phenol-water cluster was attributed by Sobolewski and Domcke to a removal of the conical intersection of the 1␲␴* surface with the ground state. Furthermore, the 1␲␴* surface shifts to higher energies, and develops a minimum at the hydrogen-transferred geometry of the cluster.33 The longer lifetimes of deuterated phenol and the phenol-water cluster compared to phenol have therefore different explanations 共smaller tunneling rate versus removal of a conical intersection兲. For the explanation of the lifetimes, it was also discussed that a rapid internal conversion takes place with the OH stretching vibration as an accepting mode.34,35 Both deuteration and complexation with water lower the stretching frequency of the OH vibration, reducing its ability to act as an accepting mode. In this picture the lifetime of the p-cresol-water cluster should be equally increased as the one of deuterated p-cresol. However, we found a much shorter lifetime for the cresol-water cluster than for the deuterated cresol, strongly favoring the lifetime model of Sobolewski and Domcke.33 The structure of the p-cresol-water cluster could be determined to be trans-linearly hydrogen bound, with cresol as proton donor like in the similar phenol-water cluster. The OO hydrogen bond length could be determined to be 290 pm in the ground state and to 285 pm in the electronically excited state. This decrease in hydrogen bond length is a consequence of the increased acidity of p-cresol in the S1 state, which leads to a stronger hydrogen bond. Again, the decrease of the hydrogen bond length of 5 pm is very similar to the corresponding value in phenol water 共4 pm兲. ACKNOWLEDGMENTS The financial support of the Deutsche Forschungsgemeinschaft 共SCHM 1043/9-3兲 is gratefully acknowledged. J. Chem. Phys. 123, 044304 共2005兲 One of the authors 共M.S.兲 would like to thank the Nordrheinwestfälische Akademie der Wissenschaften for a grant which made this work possible. 1 C. Ratzer, M. Nispel, and M. Schmitt, Phys. Chem. Chem. Phys. 5, 812 共2002兲. 2 J. Küpper, M. Schmitt, and K. Kleinermanns, Phys. Chem. Chem. Phys. 4, 4634 共2002兲. 3 M. Schmitt, C. Ratzer, C. Jacoby, and W. L. Meerts, J. Mol. Struct. 742, 123 共2005兲. 4 C. Ratzer, J. Küpper, D. Spangenberg, and M. Schmitt, Chem. Phys. 283, 153 共2002兲. 5 W. L. Meerts, M. Schmitt, and G. Groenenboom, Can. J. Chem. 82, 804 共2004兲. 6 Z. Arp, D. Autrey, J. Laane, S. A. Overman, and G. J. Thomas, Biochemistry 40, 2522 共2001兲. 7 M. Blomberg and P. Siegbahn, Mol. Phys. 101, 323 共2001兲. 8 G. Varsanyi, Assignments for Vibrational Spectra of 700 Benzene Derivatives 共Wiley, New York, 1974兲. 9 R. J. Jacobsen, Spectrochim. Acta 21, 433 共1965兲. 10 J. Laane, K. Haller, S. Sakurai, K. Morris, D. Autrey, Z. Arp, W. Chiang, and A. Combs, J. Mol. Struct. 650, 57 共2003兲. 11 T. Ebata and M. Ito, J. Phys. Chem. 96, 3224 共1992兲. 12 K. Song and J. M. Hayes, J. Mol. Spectrosc. 134, 82 共1989兲. 13 J. L. Lin, C. Li, and W. B. Tzeng, J. Chem. Phys. 120, 10513 共2004兲. 14 P. R. Richardson, M. A. Chapman, D. C. Wilson, S. P. Bates, and A. C. Jones, J. Chem. Phys. 4, 4910 共2002兲. 15 M. Pohl, M. Schmitt, K. Wolf, and K. Kleinermanns, J. Chem. Phys. 94, 1717 共1991兲. 16 M. Pohl, M. Schmitt, and K. Kleinermanns, Chem. Phys. Lett. 177, 252 共1991兲. 17 G. Berden, W. L. Meerts, M. Schmitt, and K. Kleinermanns, J. Chem. Phys. 104, 972 共1996兲. 18 M. Schmitt, J. Küpper, D. Spangenberg, and A. Westphal, Chem. Phys. 254, 349 共2000兲. 19 J. Küpper, Ph.D. thesis, Heinrich-Heine-Universität, Düsseldorf, 2000. 20 W. Gordy and R. L. Cook, Microwave Molecular Spectra, 3rd ed. 共Wiley, New York, 1984兲. 21 C. C. Lin and J. D. Swalen, Rev. Mod. Phys. 31, 841 共1959兲. 22 D. R. Herschbach, J. Chem. Phys. 31, 91 共1959兲. 23 J. A. Hageman, R. Wehrens, R. de Gelder, W. L. Meerts, and L. M. C. Buydens, J. Chem. Phys. 113, 7955 共2000兲. 24 D. Levine, PGAPACK V1.0, PGAPACK can be obtained via anonymous ftp from ftp://ftp.mcs.anl.gov/pub/pgapack/ pgapack.tar.z., 1996. 25 J. H. Holland, Adaption in Natural and Artificial Systems 共University of Michigan Press, Ann-Arbor, MI, 1975兲. 26 D. E. Goldberg, Genetic Algorithms in Search, Optimisation and Machine Learning 共Addison-Wesley, Reading, Massachusetts, 1989兲. 27 I. Rechenberg, Evolutionsstrategie—Optimierung Technischer Systeme Nach Prinzipien der Biologischen Evolution 共Frommann-Holzboog, Stuttgart, 1973兲. 28 C. Jacoby and M. Schmitt, ChemPhysChem 5, 1686 共2004兲. 29 M. Schmitt, Ph.D. thesis, Ruprecht-Karl-Universität, Heidelberg, 1992. 30 A. Hellweg, in Mikrowellenspektroskopische Untersuchungen zur intramolekularen Dynamik von p-Toluidin, p-Kresol und p-Thiokresol, edited by Günter Mainz 共Verlag Gýnter Mainz, Aachen, 2003兲. 31 R. M. Helm, H. P. Vogel, and H. J. Neusser, J. Chem. Phys. 108, 4496 共1998兲. 32 K. Remmers, W. L. Meerts, and I. Ozier, J. Chem. Phys. 112, 10890 共2000兲. 33 A. L. Sobolewski and W. Domcke, J. Phys. Chem. A 105, 9275 共2001兲. 34 A. Sur and P. M. Johnson, J. Chem. Phys. 84, 1206 共1986兲. 35 R. J. Lipert and S. D. Colson, J. Phys. Chem. 93, 135 共1989兲.