Osman Sorkhabi, Scott C. Deskin and Selim M. Senkan

One-color REMPI measurements of ortho-, meta-, and para-xylene have been made over a wavelength range of 258-274 nm using a heated pulsed valve and time-of- flight mass spectrometry. The analysis of the spectra via the S₁ ¬ S₀ vibronic transitions was consistent with those observed in absorption and two-photon resonance studies conducted under ambient conditions. From a temperature analysis of the hot bands in ortho-xylene a barrier to internal rotation of the methyl groups of 400 ± 50 cm^-1 has been calculated.

Resonance-enhanced multiphoton ionization (REMPI) has been proven to be a useful tool in differentiating isomers and isomeric mixtures [1–2]. The REMPI principle may be exploited in a real-time detection scheme where, coupled with a time-of-flight mass spectrometer (TOFMS), one may perform a mass-selective and frequency-selective analysis of a specific chemical species, including isomers [3]. Real-time detection using REMPI has important applications in environmental [4,5] and chemical process monitoring [6]. More recently, REMPI has also been used as a rapid screening tool to evaluate catalyst libraries prepared by combinatorial techniques [7]. However, in all of these applications, a detailed spectroscopic analysis of the system is necessary.

One-photon and two-photon studies of ortho-, meta-, and para-xylene are well known [2, 8–10]. Cooper and co-workers reported vibronically-resolved structure in the absorption spectra for all three isomers but no vibrational modes were assigned [11,12]. In a more recent paper, Bolovinos et al. carried out a full vibronic analysis of the S₁ ¬ S₀ system for all three isomers from their absorption data [13]. Since their measurements were carried out at room temperature, not all of the vibronic features were observed. In this work, we have carried out one-color (1+1) REMPI measurements of jet cooled o-, m-, p-xylenes. An analysis of the vibronic structure as well as the symmetry of the vibrational modes will be presented in this paper.

Previous REMPI studies of xylene have focused on the two-photon excitation of the S₁ ¬ S₀ system. One-color (2+2) REMPI measurement of the S₁ ¬ S₀ system of ortho and meta isomers of xylene was first made by Rava et al. [8]. Blease et al. carried out one color (2+2) REMPI studies of all three isomers of xylene and reported full vibronic analyses for all three isomers [9]. Different selection rules allowed for the observation of features not present in the one-photon spectra. For instance, strong vibronic mixing was observed for the n₁₄ mode, whereas transitions for this mode are strictly forbidden in a one-photon excitation. However, in both one-photon and two-photon studies, in-plane vibrations appear to be prevalent.

In addition to the vibronic analysis, we have carried out a temperature study of the hot bands of o-xylene. Ingham and Strickler have attributed these hot bands to torsional vibrations of the methyl groups and have reported a barrier to internal rotation of 700 ± 50 cm^-1, which is consistent with earlier literature reports [14–19]. However, in all of these studies, the two methyl groups were treated as independent oscillators. In this paper, we will present a direct measurement of the barrier to internal rotation.

The time-of-flight (TOF) apparatus has been described in detail elsewhere [6], thus we will focus on the REMPI technique used to measure the three isomers of xylene. An excimer laser (Lambda Physik Compex 102) operated with a XeF gas mixture was used to pump a dye laser (Laser Analytical Systems LDL 2051). Coumarin 153 dye was used to give a tunable output from 522 to 600 nm. The output of the dye laser was frequency doubled with a BBO crystal and the optimum crystal angle for second harmonic generation was maintained with an autotracker as the dye laser was tuned over the above mentioned wavelength range. One-color (1+1) REMPI measurements of ortho-, para-, and meta-xylene were made over the 258 to 274 nm range using the attenuated UV output of the BBO crystal.

The xylene samples were prepared as ~1% mixtures balanced in helium. The seeded xylene samples were then introduced in the differentially pumped vacuum chamber via a continuously purged pulsed valve operated with a pulse width of 200 ms [20]. Typical back pressures of 700 to 800 Torr were used while the vacuum chamber was maintained at pressures below 10^-5 Torr in the source region and about 10^-6 Torr in the detection region. The pulsed valve was controlled with a pulsed valve driver (General Valve Iota One). Typical repetition frequencies of 10 to 20 Hz were used in the experiments. The molecular beam was intersected with the tunable UV laser output 1.5 to 2 cm downstream from the nozzle. This arrangement allowed for the generation of photoions in the high-density region of the expanding free jet where sample concentrations are high. Positive photoions generated in this manner were collected via extractor plates biased at about -200 to -500 volts and proper ion optics were used to focus the ions into the time-of-flight chamber. The timing between the pulsed valve and the laser were controlled by using a delay/pulse generator (Stanford Research Systems DG 535).

The ion TOF signal was detected by using a microchannel plate detector (R. M. Jordan Company). After amplification with a fast preamplifier (EG&G Ortec VT120), the signal was fed into a 500 MHz digital oscilloscope (LeCroy 9350AM) where it was gated and averaged over 20 to 40 laser shots per data point. A LabView (National Instruments) program was written to relay, display, and store the gated, digitized, and averaged signal from the oscilloscope onto a 486 personal computer.

For the temperature-dependent studies, the nozzle was electrically heated and the stagnant back gas was allowed to reach thermal equilibrium before injection into the vacuum chamber. All of the other parameters were the same as in the other studies.

We have measured one-color, one-photon resonant REMPI-TOF spectra of o-, m-, and p-xylene and these are shown in Figs. 1, 3, and 4. All spectra have been corrected for changes in laser intensity over the range of the Coumarin 153 dye. Also shown in these figures are the assignments for all three isomers. The notation used is that of Wilson and Blease, et al. [9,21]. The S₁ ¬ S₀ vibronic transitions are one photon allowed for all three isomers and, hence, the observed bands are built on the 0⁰₀ origin. Since the methyl groups of xylene impart only a small perturbation on the aromatic ring, the S₁ ¬ S₀ system of the xylenes should possess certain features of the ¹A_1g ¬ ¹B_2u system of benzene [15].

If one considers the effect of the methyl groups on the benzene ring the same as a two-atom substitution, then a C_2v point group can be assigned to o-xylene. In this case, the S₁ ¬ S₀ corresponds to the ¹A₁ ¬ ¹A₁ electronic system. An analysis based on symmetry considerations leads to the vibrational modes with irreducible representations of A₁, B₁, and B₂ as contributing to the observed vibronic structure in Fig. 1. Assignments of the measured (1+1) REMPI spectrum for o-xylene have been made by using this information along with the fundamental frequencies of the vibrational modes of o-xylene [9]. The spectrum is more congested yet well resolved compared to the other isomers. Bolovinos et al. observed the 6a, 1, and 7a vibrational modes as well as several combination bands. However, as seen in Fig. 1, we observed additional structures. The observed vibrational modes and their symmetries are summarized in Table I.

The features to the red of the 0⁰₀ band for o-xylene are due to the interactions of the adjacent methyl groups. Ingham and Strickler have assigned these to the hot bands of the methyl group torsional vibrations [14]. Fig. 2 shows the hot band region for o-xylene. The observed features are in good agreement with those reported by Ingham and Strickler. Table II lists the position of the observed bands as well as their frequency shifts with respect to the origin. Although we were unable to make an assignment for Peak A, we believe it is also a hot band, similar to those observed for o-difluorobenzene [14]. The sensitivity of Peak A on temperature supports this hypothesis. Peaks B, C, and D are believed to be due to 1-1, 2-2, and 3-3 hot bands arising from the torsional vibrations of the methyl groups.

From the analysis of these bands, Ingham and Strickler extracted a value of 700 ± 50 cm^-1 for the barrier to internal rotation of the methyl groups in o-xylene [14]. Other spectroscopic as well as thermodynamic studies yielded similar values for this barrier [15–19]. However, in all of these studies the two methyl groups were treated as independent oscillators. Using such a model, the barrier to rotation in the ground and excited states can be determined from the room temperature vibronic spectrum.

On the other hand, the measurement of the o-xylene REMPI spectra at different temperatures should allow the determination of the ground state barrier directly. The results of these experiments are shown in Fig. 2. As evident from this figure, the well-resolved structure in these bands becomes increasingly diffuse. At 300° C, no structure can be discerned. This can be interpreted as indicating the free rotor nature of the methyl groups at these temperatures. If we take 300° C as the threshold temperature for the conversion to a free rotor, then a barrier to internal rotation of 400 ± 50 cm^-1 for the ground state can be calculated. This value is lower than all of the studies in which an independent oscillator model was used [14-19]. Either our result is in great error or the independent oscillator model oversimplifies the problems and hence leads to an overestimation of the barrier height. On the other hand, if the independent oscillator model is taken to be accurate, a value of 400 cm^-1 for the ground state barrier implies an excited state free rotor which was pointed out explicitly by Ingham and Strickler [14].

Fig. 3 shows the ¹B₂ ¬ ¹A₁ electronic system for m-xylene. Like the ortho isomer, m-xylene also belongs to the C_2v point group and has ¹A₁ as the term symbol for the lower electronic state. However, the upper state term symbol is different, namely ¹B₂. Given this assignment, the vibrational modes with irreducible representations of A₁, A₂, and B₂ should contribute to vibronic transitions for this molecule. The electronic moment for m-xylene lies in the x-y plane (i.e., the molecular plane). Therefore, vibrations giving rise to moments in the x and y directions should participate strongly. Knowing this and the fundamental frequencies of the vibrational modes of m-xylene [9], we have assigned the (1+1) REMPI spectrum. Note that the 6b, 8b, and 18b modes of the B₂ symmetry, 16a mode of the A₂ symmetry, and 1, 6a, 12, 13, 18a, and 19a modes of the A₁ symmetry are observed. All of the observed vibronic transitions belong to those vibrational modes in which nuclear displacement occurs in the x-y plane with the exception of the 16a mode. Similar to o-xylene, there appears to be a strong tendency toward in-plane vibrations. These results are summarized in Table I. Also, the most prominent peaks belong to the M_x and M_y transition moments which is consistent with the above analysis.

Unlike the meta and ortho isomers, p-xylene belongs to the higher symmetry D_2h point group. Therefore, S₁ ¬ S₀ corresponds to the ¹B_2u ¬ ¹A_1g electronic system and vibrational modes of A_g, B_1g, and B_3g symmetry should contribute to the vibronic structure. Of these modes, the B_1g vibrations will result in moments in the y direction (i.e., the electronic moment of the molecule) and hence should participate strongly to the vibronic structure. The 10a mode is of B_1g symmetry and from Fig. 4, belongs to one of the stronger bands. The other strong bands (1 and 7a) belong to A_g symmetry and vibrations of this symmetry should in principle lead to moments in the x direction. Apparently, both M_x and M_y moments contribute to the band intensities for p-xylene even though vibronic transitions giving rise to moments in the x and z directions should not participate strongly [16]. In addition to the bands listed in Table I, several combination bands are observed. Since these bands are forbidden in a one-photon spectrum, they must arise from multiphoton excitations. To assess this, we carried out a power-dependent study, in which the REMPI signal was measured as function of the laser intensity. For the assigned combination bands, the REMPI intensity varied as the fourth power of the laser intensity (compared to a second-power correlation for the one-photon allowed transitions).

One-color (1+1) REMPI measurements of the three isomers of xylene have been reported under jet-cooled conditions. The resulting spectra show good agreement with previous absorption and two-photon resonance ionization experiments conducted under ambient conditions. A brief vibronic analysis has been presented for all of the isomers. For the ortho isomer, a barrier to internal rotation of the methyl groups was calculated directly from a temperature analysis of the hot bands. A study of xylene mixtures, which will prove useful in real-time detection in chemical processing, is in progress in our laboratory and will be reported in a future publication.

a = fundamental frequencies are taken from Ref. 3

b = in-plane vibrational mode

c = out-of-plane vibrational mode

Table II. The observed hot bands for o-xylene. The frequency shifts for the peaks shown in Fig 3 are giving with respect to the 0⁰₀ band.

Fig. 1. One color (1+1) REMPI spectrum of jet cooled o-xylene.

Fig. 2. The expanded region of the o-xylene REMPI spectra recorded at different nozzle temperatures are shown. The observed hot bands are labeled A through E. Table II lists the positions and the frequency shifts of these bands with respect to the 0⁰₀ band.

Fig. 3. One color (1+1) REMPI spectrum of jet cooled m-xylene.

Fig. 4. One color (1+1) REMPI spectrum of jet cooled p-xylene.