Article pubs.acs.org/JPCA
Studies of Structural Isomers o‑, m‑, and p‑Fluorophenylacetylene by Two-Color Resonant Two-Photon Mass-Analyzed Threshold Ionization Spectroscopy Vidya S. Shivatare,†,‡ Aniket Kundu,§ G. Naresh Patwari,§ and Wen Bih Tzeng*,† †
Institute of Atomic and Molecular Sciences, Academia Sinica, P.O. Box 23-166, 1 Section 4, Roosevelt Road, Taipei 10617, Taiwan Taiwan International Graduate Program, Department of Chemistry, National Tsing Hua University and Academia Sinica, 128 Academia Road, Sec. 2, Nankang, Taipei 115, Taiwan § Department of Chemistry, Indian Institute of Technology Bombay, Powai, Mumbai 400 076, India ‡
ABSTRACT: We report the vibrational spectra of o-fluorophenylacetylene (OFPA), mfluorophenylacetylene (MFPA), and p-fluorophenylacetylene (PFPA) in the electronically excited S1 and cationic ground D0 states. These new data show that the relative location of the fluorine atom with respect to the acetylenic group can influence the transition energy and molecular vibration. The adiabatic ionization energies of these structural isomers follow the order: PFPA < OFPA < MFPA. It is found that the molecular geometries of these molecules in the D0 state resemble those in the S1 state. Detailed spectral analysis suggests that the in-plane ring deformation vibrations are slightly “harder” in the D0 state than the corresponding ones in the S1 state.
1. INTRODUCTION Phenylacetylene (PA) and its fluorinated derivatives play an important role in the field of liquid-crystal technology, electronics, photonics, and nonlinear optics.1−3 In recent years, structural isomers of monosubstituted PA have become the subject of growing interest because they possess different molecular properties and chemical reactivity resulting from relative location of the substituent on the ring. The frequencies of normal vibrations of this molecule in the ground S0 state have been well-documented in the literature.4 The S1 ← S0 electronic transition of PA was investigated by UV absorption, one-photon and two-photon excitation laser-induced fluorescence (LIF), and resonant two-photon ionization (R2PI) spectroscopy.5−10 Johnson and coworkers completed detailed vibronic analysis of this first electronic transition by using photoelectron imaging and spectral intensities derived from the electronic structure calculations.11 They also performed the two-color pump−probe delayed photoionization and excitedstate photoelectron spectroscopic experiments to investigate the formation of very long-lived species of PA upon the excitation to the electronically excited S1 state.12 The ionic properties of PA have been studied by laser threshold photoelectron, mass-analyzed threshold ionization (MATI), and photoinduced Rydberg ionization spectroscopic methods.13−15 Castleman Jr. and Patwari and their coworkers used resonance-enhanced multiphoton ionization mass spectroscopy (REMPI MS), infrared spectroscopy, and high-level ab initio calculations to investigate various properties of molecular complexes and clusters of PA.16−21 © 2014 American Chemical Society
Fluoro substitution on PA can alter the electron density, leading to a slight change on the transition energy, molecular geometry, and vibrations. Maity et al.22,23 applied excitation LIF, fluorescence dip infrared, and infrared-optical doubleresonance spectroscopy and computational methods to investigate various hydrogen-bonded complexes of o-fluorophenylacetylene (OFPA) and p-fluorophenylacetylene (PFPA). The vertical ionization energies (IEs) of m-fluorophenylacetylene (MFPA) and PFPA have been reported on the basis of the vapor-phase UV photoelectron experiments.24 To the best of our knowledge, information about the active vibrations of OFPA, MFPA, and PFPA in the electronically excited and cationic states is still not available in the literature. One can combine the molecular beam, mass spectrometry, and laser spectroscopy techniques to obtain experimental data, which reveal many molecular properties.25,26 The LIF27−29 and zero-kinetic energy (ZEKE) photoelectron30−34 spectroscopic techniques that detect either photons or electrons are powerful tools to record the vibronic and cation spectra with high sensitivity. As alternative approaches, the REMPI35−37 and MATI 38−47 methods that detect ions give the same spectroscopic data. Because of having mass information, they are suitable for the studies of molecular isomers to eliminate Special Issue: A. W. Castleman, Jr. Festschrift Received: January 30, 2014 Revised: March 31, 2014 Published: April 7, 2014 8277
dx.doi.org/10.1021/jp501059c | J. Phys. Chem. A 2014, 118, 8277−8286
The Journal of Physical Chemistry A
Article
perpendicularly with the molecular beam at 50 mm downstream from the nozzle orifice. During the experiments, the gas expansion, laser−molecular beam interaction, and ion detection regions were maintained at a pressure of about 1 × 10−3, 1 × 10−5, and 1 × 10−6 Pa (i.e., 1 × 10−5, 1 × 10−7, and 1 × 10−8 mbar), respectively. In the MATI experiments, both prompt ions and Rydberg neutrals were formed simultaneously in the laser-molecular beam interaction zone. A pulsed electric field of −1 V/cm (−2.5 V at TOF plate U1; 0 V at plate U2; distance of plate U1 and U2 is 2.5 cm) was applied to the region between TOF plates U1 and U2 about 180 ns after the occurrence of the laser pulses to guide the prompt ions toward the opposite direction of the flight tube.49 As a result, the prompt ions were not detected. Because the neutral Rydberg molecules were not affected by the electric field, they kept moving with the molecular beam velocity of ∼1500 m/s for 1.2 cm to reach TOF plate U2. During this period, the states with low quantum number n do not survive owing to the rapid predissociation.51 After a time delay of ∼11.8 μs, a second pulsed electric field of +200 V/cm (+2250 V at TOF plate U2; +2050 V at plate U3; 0 V at plate U4; distance between any two of these three plates is 1 cm) was applied to field-ionize the high n Rydberg neutrals in the region between TOF plate U2 and U3. The n value is estimated to be 129 by using the formula given by Chupka for calculating the lifetimes of very high Rydberg states of aromatic molecules.52 We usually slightly changed the experimental conditions to obtain the best MATI signal. Therefore, the n value may also change. However, the slight variation of the ion intensity does not affect experimental results. It is noted that TOF plates U2, U3, and U4 form a Wiley−McLaren arrangement to have a good TOF mass resolution.53 After plate U4, the threshold ions have a constant kinetic energy of ∼2150 eV and fly through a 1.0 m field-free region before being detected by a dual-stack microchannel plate (MCP) particle detector with an estimated amplification factor of 4 × 107. The ion signal from the MCP detector was discriminated at a level of −10 mV and accumulated. The typical count of these threshold ions for the present MATI experiments is 100−150/ s. All of these ion detection and analysis processes are achieved by using a multichannel scaler (MCS, Stanford Research System, SR430). The MCS and the transient digitizer were interfaced to a personal computer. Mass spectra were accumulated at a 0.020 nm (equivalent to 1.2 cm−1) spacing for 300 laser shots. The obtained optical spectra were normalized to the laser power to avoid spurious signals due to shot-to-shot laser fluctuation. The typical full width at halfmaximum (fwhm) of a typical MATI band is 7−10 cm−1. 2.2. Computational Method. All of the ab initio and density functional theory (DFT) calculations were performed by using the Gaussian 09 package.54 The computations provide information about the structure, total energy, vibrational frequency, and molecular properties of OFPA, MFPA and PFPA in their S0, S1, and D0 states. In the DFT calculations, both the Becke three-parameter functional with the Lee− Yang−Parr functional (B3LYP) and the PW91 functional (B3PW91) calculations with the 6-311++G(d,p) basis set can predict the electronic excitation and IEs with an uncertainty of no more than 4%. Because the latter gives a slightly better prediction on the vibrational frequencies in the S1 and D0 states, we only quote the values from the B3PW91 calculations for comparison with the experiments. The validity of the calculations for the cation are confirmed by the spin multiplicity
ambiguity resulting from presence of complexes, clusters, or impurities in the chemical sample. In the present study, we apply the REMPI and MATI techniques to record the vibronic and cation spectra of OFPA, MFPA, and PFPA. For these isomeric species, the band origin of the S1 ← S0 electronic transition (E1) is close to one-half of the IE. One can use the R2PI technique to record the vibronic spectrum. In the cases of OFPA and MFPA, the E1 is less than one-half of the IE. Thus, the two-color (2C) R2PI method is required to record their vibronic spectra.48 It is achieved by fixing the frequency of the ionization laser and scanning the excitation laser while collecting the prompt ions. For PFPA whose E1 is slightly greater than one-half of the IE, the vibronic spectrum can be recorded by using the one-color (1C) R2PI method. To obtain information about more active vibrations and to investigate the molecular geometry of the cation, we recorded the two-color resonant two-photon MATI spectra by ionizing through several intermediate vibronic levels. The detailed procedure will be described later. These results yield information about the active vibrations of OFPA, MFPA, and PFPA in the electronically excited S1 and cationic ground D0 states in addition to the E1 and adiabatic IEs. Because the vibration frequency of each mode of a multiple substituted benzene is related to substituent−ring and substituent− substituent interactions, each molecule has a unique set of spectral features. Therefore, a well-resolved spectrum can be used for molecular identification, although the spectral assignment is not trivial. Comparison of these new experimental data with those of PA4−8,11,13,14 and its derivatives21−24 allows us to gain knowledge about the fluoro substitution effect on the transition energy and molecular vibration.
2. EXPERIMENTAL AND COMPUTATIONAL DETAILS 2.1. Experimental Method. We performed all of the experiments reported in this paper by using a coaxial laserbased photoionization time-of-flight (TOF) mass spectrometer, described previously.49 OFPA, MFPA, and PFPA (99% purity) were purchased from Sigma-Aldrich Corporation and used without further purification. The vapors were seeded into 2 to 3 bar of helium and expanded into the vacuum through a pulsed valve with a 0.15 mm diameter orifice. The free jet passed through a skimmer (1.0 mm diameter) located ∼10 mm downstream from the nozzle orifice to form a molecular beam. The two-color resonant two-photon excitation process was achieved by utilizing two independent tunable UV lasers controlled by a delay/pulse generator (Stanford Research Systems DG535). The excitation (pump) source is a Nd:YAG pumped dye laser (Quanta-Ray PRO-190−10/Lambda-Physik, Scanmate UV with BBO-III crystal; Fluorocence-548, Rhodamine-575, Rhodamine-590 dyes). The output visible radiation is frequency-doubled to produce UV radiation. The ionization (probe) UV laser (Lambda-Physik, Scanmate UV with BBO-III crystal; Coumarin-540A, Rhodamine-575 dyes) is pumped by a frequency-doubled Nd:YAG laser (Quanta-Ray LAB-150). The wavelengths of both lasers were calibrated by using a laser wavelength meter (Coherent, WaveMaster). To correct the systematic instrument errors for obtaining the true wavelength, the displayed values of the excitation and ionization lasers need to be added by 0.20 and 0.28 nm, respectively. The results are in excellent agreement with those obtained from the highresolution electronic spectroscopic experiments.50 These two counter-propagating laser beams were focused and intersected 8278
dx.doi.org/10.1021/jp501059c | J. Phys. Chem. A 2014, 118, 8277−8286
The Journal of Physical Chemistry A
Article
Table 1. Observed Vibronic Bands (in cm−1) of OFPA in Figure 1a
of ∼0.75. The predicted IE is deduced from the difference in the zero-point energy (ZPE) levels of the cation in the D0 state and the neutral in the S0 state. The simulated MATI spectra were obtained by calculating Franck−Condon (FC) factors from the predicted harmonic vibrational frequencies and geometries of the molecule in S1 and D0 states. The calculated vibrational frequencies and scaling factors are listed in the Tables along with the measured values.
3. RESULTS 3.1. Vibronic and Cation Spectra of OFPA. A wise choice on the frequency of the ionization laser can minimize excess energy of the R2PI process to increase the signal-tonoise ratio for recording the spectrum.48,55,56 The results from our preliminary R2PI and MATI experiments show that the E1 is less than one-half of the IE of OFPA. Therefore, we performed the 2C-R2PI experiments to record the vibronic spectrum by scanning the excitation laser in the range of 270.00−284.00 nm while fixing the ionization laser at 274.55 nm (36 423 cm−1). As seen in Figure 1, the E1 of OFPA
energy
shift
calcd
assignment and approximate descriptionb
35586 35756 35926 36082 36259 36391 36535 36758 36806
0 168 340 496 673 805 949 1172 1220
183 321 502 691 814 963 1133 1231
0°0, origin 10b10, γ(ring-CCH) 9b10, β(CF), β(ring-CCH) 6a10, β(CCC) 110, breathing 1210, β(CCC) 18b10, β(ring-H) 9a10, β(ring-H) 1410, v(CC)
a Experimental values are shifts from 35 586 cm−1, whereas the predicted frequencies (scaled by 0.98) are obtained from the TDB3PW91/6-311++G(d,p) calculation. bβ, in-plane bending; γ, out-ofplane bending; v, stretching
The intense vibronic bands at 673, 949, 1172, and 1220 cm−1 result from vibronic transitions 110, 18b10, 9a10, and 1410 of OFPA. Mode 1 (breathing motion) mainly involves in-plane ring deformation, whereas mode 14 resembles the kekule-type in-plane CC stretching vibration of benzene. Figure 2 shows
Figure 1. 2C-R2PI spectrum of OFPA, recorded by scanning the excitation laser while fixing the ionization laser at 274.55 nm (36 423 cm−1).
appears at 35 586 ± 2 cm−1, which is in very good agreement with that reported on the basis of the excitation LIF experiments.22 Two weak bands on the low-energy side of the origin band result from hot band transition. This argument has been confirmed by observing the decrease in intensity of these bands as the stagnation pressure of the pulsed valve increases. The assignment of the vibronic bands has been accomplished by comparing the present data with those of PA,4−8 MFPA, PFPA, and the predicted values from the time-dependent (TD) B3PW91calculations with the 6-311++G(d,p) basis set. Table 1 lists the observed vibronic bands in Figure 1 along with the excitation photon energy, relative intensity, energy shift from the band origin, calculated vibrational frequency, and possible assignment. The numbering scheme of the normal vibrations of OFPA follows that used by Varsanyi and Szoke4,57 for o-Di“light” substituted benzene derivative and is based on Wilson’s notations.58 The vibrational motions can be viewed by following the Gaussview procedure of the Gaussian 09 program.54
Figure 2. Some observed active vibrations of OFPA, MFPA, and PFPA in S1 and D0 states. The open circles designate the original locations of the atoms, whereas the solid dots mark the displacements. The measured and calculated (in the parentheses) frequencies are included for each mode.
some active normal vibrations of OFPA, MFPA, and PFPA in S1 and D0 states observed in the present studies. Modes 18b and 9a result from the in-plane ring-H bending vibration. The moderately intense bands at 496 and 805 cm−1 correspond to transitions 6a10 and 1210, related to in-plane ring deformation vibrations. The weak band at 340 cm−1 results from substituentsensitive in-plane C−F and ring−CCH bending vibration 8279
dx.doi.org/10.1021/jp501059c | J. Phys. Chem. A 2014, 118, 8277−8286
The Journal of Physical Chemistry A
Article
9b1. The low-frequency band at 168 cm−1 is assigned to the out-of-plane bending vibration 10b1. Because there were no available data concerning the ionic properties of OFPA, we carried out the photoionization efficiency (PIE) experiment to locate the ionization limit. This experiment was also performed by applying the 2C-R2PI technique. However, the operational scheme is different from the one for recording the vibronic spectrum stated previously. The PIE curve of OFPA (not shown and available on request) was recorded by scanning the ionization laser while fixing the excitation laser at the S10° level (281.01 nm or 35 586 cm−1). Because this method collects the non-energy-selected prompt ions, it gives strong intensity and leads to an abruptly rising step at the ionization limit.28 The IE was found to be near 71 976 with an uncertainty of ∼10 cm−1. In contrast, the MATI technique detects only the threshold ions and leads to a sharp peak at the ionization limit,38−41 as seen in Figure 3. Analysis
Table 2. Observed MATI Bands (in cm−1) of OFPA in Figure 3a intermediate level in the S1 state S10° 128 358 402 538 715 823
S111
716
calcd
assignment and approximate descriptionb
135 359 412 537 709 820
β(ring-CCH) 9b1, β(CF), β(ring-CCH) 6b1, β(CCC) 6a1, β(CCC) 11, breathing 121, β(CCC)
Experimental values are shifts from 71 976 cm−1, whereas the predicted frequencies (scaled by 0.97) are obtained from the UB3PW91/6-311++G(d,p) calculation. bβ, in-plane bending a
frequencies of the observed MATI bands and their possible assignments. The bands at 402, 538, 715, and 823 cm−1 result from the in-plane ring deformations 6b1, 6a1, 11, and 121 vibrations, respectively. Low-frequency bands at 128 and 358 cm−1 result from the substituent-sensitive in-plane ring−C CH bending and 9b1 vibrations. We applied the FC principle to the B3PW91/6-311++G(d,p) calculations to simulate the MATI spectrum associated with the D0 ← S1 transition from the S10° state. As seen in Figure 3b, the simulated spectral features resemble those of Figure 3a. Figure 3c shows MATI spectrum recorded by ionization through S111 (0° + 673 cm−1) level. The strongest band results from the breathing motion 11 of the OFPA cation. 3.2. Vibronic and Cation Spectra of MFPA. Similar to the case of OFPA, the vibronic spectrum of MFPA was recorded by using the 2C-R2PI technique with the ionization laser fixed at 269.86 nm (37 056 cm−1). Figure 4 displays the vibronic spectrum of MFPA, in which the E1 appears at 35 573 ± 2 cm−1. Two weak hot bands also appear at the low-energy side of the 0°0 band. Table 3 lists the observed vibronic bands along with possible assignments. According to Varsanyi and Szoke4,57 MFPA is classified as a m-Di-“light”-substituted benzene derivative. Bands at 400, 464, 678, 965, and 991 cm−1 result from vibronic transitions 9a10, 6a10, 110, 1210 and 1810, respectively. Some of these normal modes are shown in Figure 2. It is noted
Figure 3. MATI spectra of OFPA, recorded by ionizing via the (a) 0° and (c) 11 levels in the S1 state. (b) Simulated spectrum by applying the Franck−Condon principle in the B3PW91/6-311++G(d,p) calculation.
on the 0+ bands with consideration of the uncertainty in the laser photon energy, the spectral width, and the Stark effect gives the adiabatic IE of OFPA to be 71 976 ± 5 cm−1 (8.9239 ± 0.0006 eV), which is in excellent agreement with that obtained by our PIE experiment. A major advantage of the MATI (or ZEKE) over the PIE and conventional photoelectron spectroscopic techniques is that it provides information about the internal motions of the cation. Because the rotational constants of OFPA, MFPA, and PFPF are very small (
