Vibronic and Cationic Features of 2-Fluorobenzonitrile and 3-Fluorobenzonitrile Studied by REMPI and MATI Spectroscopy and Franck–Condon Simulations

Abstract

Fluorinated organic compounds have superior physicochemical properties than general organic compounds due to the strong C-F single bond; they are widely used in medicine, biology, pesticides, and materials science. In order to gain a deeper understanding of the physicochemical properties of fluorinated organic compounds, fluorinated aromatic compounds have been investigated by various spectroscopic techniques. 2-fluorobenzonitrile and 3-fluorobenzonitrile are important fine chemical intermediates and their excited state S₁ and cationic ground state D₀ vibrational features remain unknown. In this paper, we used two-color resonance two photon ionization (2-color REMPI) and mass analyzed threshold ionization (MATI) spectroscopy to study S₁ and D₀ state vibrational features of 2-fluorobenzonitrile and 3-fluorobenzonitrile. The precise excitation energy (band origin) and adiabatic ionization energy were determined to be 36,028 ± 2 cm⁻¹ and 78,650 ± 5 cm⁻¹ for 2-fluorobenzonitrile and 35,989 ± 2 cm⁻¹ and 78,873 ± 5 cm⁻¹ for 3-fluorobenzonitrile, respectively. The density functional theory (DFT) at the levels of RB3LYP/aug-cc-pvtz, TD-B3LYP/aug-cc-pvtz, and UB3LYP/aug-cc-pvtz were used to calculate the stable structures and vibrational frequencies for the ground state S₀, excited state S₁, and cationic ground state D₀, respectively. Franck–Condon spectral simulations for transitions of S₁ ← S₀ and D₀ ← S₁ were performed based on the above DFT calculations. The theoretical and experimental results were in good agreement. The observed vibrational features in S₁ and D₀ states were assigned according to the simulated spectra and the comparison with structurally similar molecules. Several experimental findings and molecular features were discussed in detail.

1. Introduction

Due to the presence of the strong C-F single bond within the molecule, fluorinated organic compounds have superior physicochemical properties and are widely used in medicine, biology, pesticides, and materials science [1,2,3,4,5]. In recent years, a large number of fluorinated aromatic compounds have been investigated by various spectroscopic techniques. Ling et al. used femtosecond time-resolved photoelectron imaging to study the conformation of bi-fluorophenol and bi-fluoroaniline in the excited state S₁ after photoexcitation [6,7]. Wijngaarden’s group used high-resolution microwave spectroscopy to measure the rotation spectra of fluorine substituted benzaldehyde, benzonitrile, phenol, and pyridine derivatives to study the molecular structure changes caused by fluorination and intramolecular hydrogen bonding interactions [8,9,10]. Many experimental groups have also studied the vibrational spectra of excited state S₁ and cationic ground state D₀ of fluorine-substituted phenol, anisole, and aniline derivatives using laser-induced fluorescence (LIF), resonance-enhanced multiphoton ionization (REMPI), and mass-analyzed threshold ionization (MATI) spectroscopy [11,12,13,14]. Mono-fluorobenzonitrile is a very important class of intermediate for organic synthesis; its vibrational and rotational properties have been reported in many studies [15,16,17,18]. Kamaee et al. investigated the structural trends in mono-, di-, and pentafluorobenzonitriles using Fourier transform microwave spectroscopy [10]. Palmer et al. measured photoelectron spectroscopy of 2-fluorobenzonitrile (2FBN) and 3-fluorobenzonitrile (3FBN) and reported the ionization energies (IEs) of 9.78 eV and 9.79 eV, respectively [19]. Jiang and Levy used laser-induced fluorescence and dispersive fluorescence spectroscopy to study the vibrational relaxation of the excited state of 4-fluorobenzonitrile (4FBN) molecule [20]. In 2018, Zhao et al. [21] studied the vibrational features of the excited and cationic ground states of 4FBN by REMPI and MATI techniques. Silva et al. [16] measured the UV-Vis spectra of monofluorobenzonitriles in dichloromethane; from the curves they measured, the approximate origins of 2FBN and 3FBN could be estimated at 283 nm. To the best of our knowledge, the vibrational properties of the excited states and cationic ground states of 2FBN and 3FBN have not been reported in the literature.

MATI and zero kinetic energy (ZEKE) spectroscopy are currently the most popular high-resolution techniques for measuring vibrational features of cationic ground states. Kwon’s group built a vacuum ultraviolet single photon MATI system to study many cationic vibrational features [22,23,24,25,26,27,28,29]. Tzeng’s group and Ketkov’s group used two-color MATI to study the cationic spectra of many benzene derivatives and sandwich molecules [30,31,32,33,34,35,36]. Wright’s group used ZEKE technology to research cationic vibrational features of many halogenated benzene and their derivatives [37,38,39,40,41,42]. In this paper, we used two-color REMPI and MATI techniques to study the vibrational features of the excited states and cationic ground states of 2FBN and 3FBN. The precise excitation energies and adiabatic ionization energies were determined. The measured vibrational features were assigned and several experimental findings were analyzed and discussed in detail.

2. Results

The stable structures of 2- and 3-fluorobenzonitrile with atomic labels are shown in Figure 1. 2FBN and 3FBN molecules consist of 13 atoms with a total of 33 normal vibrational modes, 30 modes of which are at aromatic ring and 3 modes at CN group. The labeling convention of the vibrational modes followed the Varsanyi system [43]. Vibronic transitions were expressed in the Wilson notation based on the benzene modes, where the v′ ← v″ transition in the normal mode n was represented by ${\mathrm{n}}_{\mathrm{v}″}^{\mathrm{v}\prime }$ [44]; subscript V″ was omitted in the present research as it was a constant 0 (the low energy level of the transition is the vibrationless or zero point energy level of the low electronic state).

2.1. Vibronic Features of 2-Fluorobenzonitrile in the S₁ State

The vibronic spectrum of the S₁ ← S₀ transition of 2FBN was measured by a two-color REMPI experiment with the vibration frequency range of 0–1350 cm⁻¹. The experimental result is shown in Figure 2a and its Franck–Condon simulation calculated at TD-B3LYP/aug-cc-pvtz level is shown in Figure 2b. It can be seen that the experimental result was in good agreement with the calculated one. The obvious feature of both REMPI and its simulation in Figure 2 was that the rate of signal-to-noise in the low frequency region was greater than in the high frequency region. The simulation spectrum showed that the bands in the high frequency region were dense and consisted of many fundamentals, overtones, and combinations of various modes, many of which were very weak. Thus, dense and weak bands raised the spectral baseline and resulted in a bad rate of signal-to-noise in high frequency regions. The band at certain frequencies in the spectrum maybe came from several component (or vibration mode) contributions. For simplicity’s sake, we only list the largest contributor in

Table 1. Observed bands in the vibronic spectrum of 2FBN and their possible assignments ^a.

Transition Energy (cm−1)	Relative Intensity	Shift (cm−1)	Calc. (cm−1)	Assignment b
36,028	100	0	0	00
36,164	4	136	137	151
36,198	7	170	156	γCN2
36,290	3	262		10b1γCN1
36,356	12	328		10a1γCN1
36,369	42	341	335	6b1
36,452	9	424	421	9b1
36,528	9	500	495	6a1
36,638	3	610		16a110b1
36,663	31	635	646	16b2
36,696	30	668	668	11
36,721	30	693	698	17a1
36,738	7	710	711	6b110b2
36,792	6	764	755	9b16b1
36,843	34	815	813	121
36,874	2	846	841	9b2
36,945	4	917	910	βCN16b1
36,950	5	922	916	6a19b1
36,974	74	946	960	18b1
36,982	15	954	969	121γCN2
37,008	10	980	981	6b116b2
37,037	14	1009	1003	116b1
37,062	6	1034	1022	16b210b2
37,199	14	1171	1156	131
37,285	12	1257	1224	7a1
37,315	4	1287	1295	18b16b1

^a Experimental values are shifts from 36,028 cm⁻¹ and the calculated ones (scaled by 0.9649) are obtained from the TD-B3LYP/aug-cc-pvtz calculations. ^b β, in-plane bending; γ, out-of-plane bending.

.

Based on DFT calculation and spectral simulation, we analyzed and assigned the vibronic spectra of 2FBN. It is very clear in Figure 2a that the band at 36,028 cm⁻¹ was assigned to the band origin of the S₁ ← S₀ transition. Many in-plane vibrational modes of the ring were active and most of them were very strong in the REMPI spectrum. The bands at 136, 341, 424, 500, 668, 815, 946, 1171, and 1257 cm⁻¹ were assigned to fundamental modes 15, 6b, 9b, 6a, 1, 12, 18b, 13, and 7a, respectively. One out-of-plane fundamental mode at the ring was observed, which appeared at 693 cm⁻¹ and was assigned to mode 17a. Several overtone vibrations were observed, which appeared at 170, 635, and 846 cm⁻¹ and assigned to γCN², 16b², and 9b², respectively. Other bands observed in the REMPI spectrum were assigned to the combined vibrations of several modes. All the measured vibrational frequencies, calculated frequencies, and possible assignments are listed in Table 1.

From the measured REMPI spectra in Figure 2a, we found that the vibronic band 1¹ was much wider than other bands. From the simulation calculation, we knew that the band 1¹ consisted of three components: 16a¹10a¹ (665.2 cm⁻¹), 1¹ (667.5 cm⁻¹), and 6b² (669.7 cm⁻¹). The calculated dipole strengths at the level of TD-B3LYP/aug-cc-pvtz for these three components were 3.365 × 10⁻⁵, 1.254 × 10⁻², and 3.2 × 10⁻³, respectively. Due to the very close vibrational energy, the resonance interactions may have played a role in the broadening of the experimental spectral line.

2.2. Photoionization Efficiency (PIE) Spectra of 2FBN

In order to measure the cationic spectra, we first required to know the ionization energy (IE). With the present experimental setup, the IE could be measured by photoionization efficiency (PIE) or MATI experiments. The PIE approach detected the prompt ions involving the field-ionization of high Rydberg neutrals and yielded a strong signal that led to an abruptly rising step near the ionization limit. In contrast, the MATI spectrum detected the threshold ions and yielded a sharp peak at the ionization threshold and vibrational features of the cation. We recorded both the PIE and MATI spectra by scanning the frequency of the ionization laser over a large range to determine the IE of 2FBN. Figure 3a,b show the PIE and MATI spectra via the intermediate state S₁0⁰ (36,028 cm⁻¹). The adiabatic IE of 2FBN was determined to be 78,647 ± 10 cm⁻¹ by PIE and 78,650 ± 5 cm⁻¹ (9.7514 ± 0.0006 eV) by MATI, including the correction of the Stark effect, respectively. These results were in good agreement with the previous measured value of 9.78 eV (78,881 cm⁻¹) [19] by photoelectron spectroscopy with an He I UV-light source.

2.3. Cationic Spectra of 2FBN

To investigate the molecular geometry and vibrational features of the 2FBN cation, the MATI spectra were recorded by ionizing via the S₁0⁰, S₁6b¹ (0⁰ + 341 cm⁻¹), S₁1¹ (0⁰ + 668 cm⁻¹), S₁12¹ (0⁰ + 815 cm⁻¹), and S₁18b¹ (0⁰ + 946 cm⁻¹) intermediate states.

We first performed the theoretical calculation and spectral simulation. The Franck–Condon simulation is shown in Figure 4a; the corresponding MATI spectrum via S₁0⁰ is shown in Figure 4b. From Figure 4a,b we know that the theoretical and experimental spectra were in good agreement. The most intense peak corresponded to the origin of the D₀ ← S₁ transition of 2FBN. Spectral features were assigned, mainly based on DFT calculations, the Franck–Condon simulation, and comparisons with the available data on substituted benzonitriles. Spectral assignment is a very tedious and error prone matter. Accurate assignments can be obtained by high dimensional or even full dimensional vibrational calculations [45,46]. For the present work, we used the Franck–Condon simulation, which greatly facilitated the spectral identification work. The bands at 131, 333, 530, 571, 683, 826, 972, 1268, and 1555 cm⁻¹ were relatively intense and assigned to ring or CN group in-plane motion modes 15, 6b, 6a, βCN, 1, 12, 18b, 7a, and 8b, respectively. Several out-of-plane bending vibrations were also observed, such as γCN and 10b, appearing at 106 and 197 cm⁻¹, respectively. Other bands were weak and assigned to overtone or combination vibrations. The measured and calculated cationic vibrational frequencies and their possible assignments are listed in

Table 2. Assignment of the observed bands (cm⁻¹) in the MATI spectra of 2FBN ^a.

Intermediate Levels in the S1 State					Calc.	Assignment b
00	6b1	11	121	18b1
106		106			104	γCN1
131			130	132	139	151
197			193	199	205	10b1
		209				γCN2
		323	321	322	335	10a1
333	335				341	6b1
407	408		409	410	410	9b1
		430				6b1γCN1
	466					6b1151
474		473				152γCN2
		520				10a110b1
530			531	532	547	6a1
	532					6b110b1
		582				6b1151γCN1
		643				10a2
571			570	571	584	βCN1
	672	672				6b2
683		687		687	696	11
698						16a110b1
	746					6b19b1
		796				11γCN1
		818				11151
826			823		826	121
	865					6b16a1
		888				1110b1
	906					6b1βCN1
		950				11152
			955			121151
972				973	976	18b1
		1023				116b1
	1036					6b3
1059			1064			6a2
				1104		18b1151
	1164		1160			6b1121
		1217				116a1
1268					1271	7a1
				1292		18b110a1
	1316			1321		6b118b1
		1329				1110a2
			1349			1216a1
			1394			121βCN1
1396					1410	121βCN1
1466					1466	19b1
				1503		18b16a1
		1513				11121
				1545		18b1βCN1
1555					1551	8b1
	1607					6b17a1
			1646			122

^a The experimental values are shifts from 78,650 cm⁻¹, whereas the calculated ones are obtained from the B3LYP/aug-cc-pVDZ calculations, scaled by 0.9849. ^b β, in-plane bending; γ, out-of-plane bending.

.

In order to find more vibrational modes of 2FBN cation, the different intermediate states were used to record the MATI spectra. Figure 4c shows the MATI spectra via S₁6b¹ (0⁰ + 341 cm⁻¹). In comparison with Figure 4b, we found that, when S₁6b¹ was used as the intermediate state, most of the spectral features could be assigned to combinations of 6b and the modes found in MATI via S₁0⁰. This could be verified by shifting Figure 4c to the left to align its band 6b with the 0⁺ band in Figure 4b. No more fundamental modes than the MATI via S₁0⁰ were found.

Figure 5 shows the MATI spectra via the intermediate states of S₁1¹ (0⁰ + 668 cm⁻¹), S₁12¹ (0⁰ + 815 cm⁻¹), and S₁18b¹ (0⁰ + 946 cm⁻¹). Similarly, when S₁1¹ (0⁰ + 668 cm⁻¹) was used as the intermediate state, a lot of bands were assigned to the combination vibrations of the mode 1 and those found in MATI via S₁0⁰. In the lower frequency region, substituent CN out-of-plane bending γCN and its overtone γCN² were found. Aromatic ring out-of-plane bending 10a and its overtone 10a² were also observed. Other bands were weak and were assigned to combination vibrations of several modes.

Similarly, when S₁12¹ was used as the intermediate, as shown in Figure 5b, except for the bands at 1064 and 1646 cm⁻¹ being assigned to 6a² and 12², other bands greater than 823 cm⁻¹ (D₀12¹) were assigned to combinations of 12¹ and other modes. In lower frequency regions, some fundamental modes were active, which were found in the MATI spectrum via S₁0⁰ or S₁1¹. For the MATI via S₁18b¹ in Figure 5c, the spectral feature was similar to the MATI via S₁12¹; all the assignments, as well as the calculated and measured values, are listed in Table 2.

2.4. Vibronic Features of 3-Fluorobenznitrile in the S₁ State

The vibronic spectrum in the S₁ state of 3FBN is shown in Figure 6a, together with its Franck–Condon simulation shown in Figure 6b for comparison. The entire simulated spectra appeared comparable to the 2-color REMPI spectra in Figure 6a. The distinct band corresponding to the transition energy of 35,989 cm⁻¹ was identified as the origin of the S₁←S₀ electronic transition.

Table 3. Assignment of observed bands (cm⁻¹) in the 2-color REMPI spectrum of 3FBN ^a.

Transition Energy	Exp.	Relative Intensity	Calc. a	Assignment b
35,989	0	100		00, band origin
36,129	140	11	142	151
36,182	193	21	171	γ(CN)2
36,255	266	7	251	10b1γ(CN)1
36,324	335	16	342	10b2
36,374	385	35	381	9a1, β(C-F)
36,390	401	58	399	6b1, β(CCC)
36,403	414	31	417	10b1γ(CN)3
36,459	470	11	473	6a1, β(CCC)
36,469	480	6	484	10b2151
36,509	520	6	523	9a1151
36,549	560	9	567	βCN
36,584	595	9	593	10b3γ(CN)1
36,611	622	9	623	10a2γ(CN)2
36,649	660	36	663	11, breather
36,675	686	33	684	9a1151γ(CN)2
36,692	703	20	709	151β(CN)1
36,727	738	8	741	6b110b2
36,789	800	6	798	6b2
36,864	875	8	872	6a16b1
36,900	911	13	915	1110b1γ(CN)1
36,927	938	13	940	6b2151
36,947	958	24	957	121
36,963	974	27	975	1110a1γ(CN)1
36,985	996	11	996	6a19a1151
37,051	1062	17	1063	116b1
37,075	1086	9	1087	416a1γ(CN)1
37,136	1147	17	1142	9b1
37,241	1252	12	1253	6a16b19a1
37,260	1271	22	1266	131
37,304	1315	7	1314	1116b116b1
37,324	1335	14	1338	1219a1
37,347	1358	15	1356	1216b1
37,363	1374	19	1385	19a1

^a The experimental values are shifts from 35,989 cm⁻¹, whereas the calculated ones are obtained from the TD-B3LYP/aug-cc-pVDZ calculations, scaled by 0.9722. ^b β, in-plane bending; γ, out-of-plane bending.

lists the observed vibronic transition energies, along with the energy shifts with respect to the band origin, band relative intensities, and possible assignments. The spectral assignment of 3FBN was accomplished by comparing with those of 4-fluorobenzonitrile, 3-fluorophenol, TD-B3LYP/aug-cc-pvtz calculation, and the Franck–Condon simulation. The spectral features in Figure 6a mainly resulted from vibronic transitions related to the in-plane ring deformation and substituent sensitive bending vibrations. The bands appearing at 140, 385, 401, 470, 560, 660, 958, 1147, 1271, and 1374 cm⁻¹ were assigned to in-plane stretching or bending vibrations 15, 9a, 6b, 6a, βCN, 1, 12, 9b, 13, and 19a, respectively. The out-of-plane overtone vibrations γ(CN)² and 10b² were also observed in lower frequency regions. Other bands were assigned to combination vibrations of several modes.

2.5. PIE Spectra of 3FBN

Similar to the 2FBN, ionization energy was very important for the cationic spectral measurements. We first performed the PIE experiment to determine the IE of 3FBN to be 78,873 ± 10 cm⁻¹. Then, we measured the MATI spectra to give the precise IE of 3FBN to be 78,873 ± 5 cm⁻¹. The PIE and MATI spectra via S₁0⁰ are shown in Figure 7a,b for comparison. It was obvious that they were very consistent.

2.6. MATI Spectra of 3FBN

Figure 8a,b show the calculated Franck–Condon spectrum and measured MATI spectrum via S₁0⁰ state at 35,989 cm⁻¹, respectively. We can see that they were in good agreement. Many in-plane vibrations were active, such as modes 15, 6b, 6a, 1, 12, 18a, 9b, 18b, 13, and 8a appearing at 133, 371, 498, 668, 978, 1066, 1117, 1144, 1307, and 1566 cm⁻¹, respectively. Out-of-plane bending modes 10b and 10a were also observed, but they were weak. Other bands were assigned to combinations of several modes. All the experimental and calculated cationic vibrational frequencies of 3FBN and corresponding assignments are listed in

Table 4. Assignment of observed bands (in cm⁻¹) in the MATI spectra of 3FBN ^a.

Intermediate Levels in the S1 State				Calc.	Assignment b
00	γ(CN)2	6b1	11
	120			118	γ(CN)1
133				141	151
	238			237	γ(CN)2
185				188	10b1
322				331	10a1
371		370		374	6b1, β(CCC)
		399		394	9a1
498				495	6a1, β(CCC)
		506			6b1151
	609				γ(CN)26b1
615				605	16a1
668			668	679	11, breathing
		688			6b110b1151
	730				γ(CN)26a1
		744			6b2
			775		4110b1
			803		11151
863		860			6b16a1
893		894			6a19a1
978				965	121, β(CCC)
997				990	6a2
1066				1067	18a1, β(CH)
	1098				γ(CN)26a16b1
1117				1105	9b1, β(CH)
1144				1133	18b1, β(CH)
			1166		116b1
	1218				γ(CN)2121
		1235			6b26a1
		1258			6b16a19a1
	1299				γ(CN)218a1
1307				1302	131, β(CH)
		1350			6b1121
	1357				γ(CN)29b1
		1374			6b16a2
		1385		1382	19a1
		1489			6b19b1
1566				1537	8a1, ν(CC)
		1516			6b118b1
		1574			6b118a1151

^a The experimental values are shifts from 78,873 cm⁻¹, whereas the calculated ones are obtained from the B3LYP/aug-cc-pVDZ calculations, scaled by 0.9704. ^b β, in-plane bending; γ, out-of-plane bending.

.

When measuring the MATI via S₁γCN² (Figure 8c), the distinct feature at 238 cm⁻¹ was assigned to D₀γCN², which followed the propensity rule Δν = 0. The fundamental vibration γCN¹ was also observed at 120 cm⁻¹, with a weak intensity, which did not appear in the REMPI spectrum. Other bands were assigned to combination vibrations of γCN² and fundamental vibrations.

When measuring the MATI via S₁6b¹, as shown in Figure 9a, the distinct feature at 370 cm⁻¹ was assigned to D₀6b¹, which followed the propensity rule Δν = 0. The intense band at 399 cm⁻¹ was assigned to 9a¹. Other bands were assigned to combination vibrations of 6b¹ and fundamental vibrations. Figure 9b shows the MATI spectrum via S₁1¹, where the cationic vibration 1¹ (668 cm⁻¹) was most intense. The bands at 775, 803, and 1166 cm⁻¹ were assigned to combination vibrations 4¹10b¹, 1¹15¹, and 1¹6b¹, respectively.

3. Discussion

3.1. Breathing Vibrational Band of 2FBN

Whether the vibrational spectra of excited state S₁ or cationic ground state D₀, the frequencies of different vibrations in the high-frequency region may have been very close or even the same, which may have come from the fundamental, overtone, or combination vibrations. For example, the breathing vibration 1¹ of 2FBN in the REMPI spectrum (see Figure 2a) appeared at 668 cm⁻¹. The theoretical calculation showed that there were also two weaker vibrations 16a¹10a¹ and 6b², whose vibrational frequencies were close to that of mode 1¹. The calculated vibration frequencies of 16a¹10a¹, 1¹, and 6b² were 665.2 cm⁻¹, 667.5 cm⁻¹ and 669.7 cm⁻¹, respectively. They were so close that the spaces between them were less than the experimental resolution, which led to a wide spectral band in the REMPI spectrum. When using this band as the intermediate state to perform the MATI experiment, according to the propensity rule of Δν = 0, these three vibrational modes of cation may be observed with great intensity. Generally, the vibration frequency of cation is slightly different from that of the excited state for the same vibrational mode and the frequency change can be not consistent for various vibration modes. So, these three vibration modes of cation of 2FBN may be separated in the MATI spectrum. As shown in Figure 5a, the cationic mode 1¹ appeared at 687 cm⁻¹, 6b² appeared at 672 cm⁻¹, and 10a¹ and 10a² were also observed at 322.7 and 643.2 cm⁻¹, respectively. The strength of the MATI signal was not only related to the Franck–Condon factor but also to the population of the intermediate state S₁ and further related to the resonance degree of each vibration mode with the excitation (S₁ ← S₀) photon frequency. The experimental results demonstrated that the superposition band of several vibrations could be used as an intermediate state to perform the MATI experiments and more vibrational modes of cation could be observed.

3.2. Molecular Structure in S₀, S₁, and D₀ States and Vibrational Frequencies

Theoretical calculations showed that the stable configurations of the ground state S₀, excited state S₁, and cationic ground state D₀ of 2FBN and 3FBN molecules all had Cs symmetry and all the atoms were in the ring plane. This was consistent with their large Franck–Condon factors, intense REMPI and MATI signals, and MATI spectra, following the propensity rule of Δν = 0. However, in the transitions of S₁ ← S₀ and D₀ ← S₁, the bond length and bond angle of molecules changed slightly.

Table 5. Bond length and bond angle of electronic ground state S₀, first excited state S₁, and cationic ground state D₀ of 2-fluorobenzonitrile calculated at RB3LYP/aug-cc-pvtz, TD-B3LYP/aug-cc-pvtz, and UB3LYP/aug-cc-pvtz levels, respectively.

	S0	S1	D0	Δ(S1 − S0)	Δ(D0 − S1)	Δ(D0 − S0)
Bond length (Å)
C1−C2	1.396	1.436	1.455	0.040	0.019	0.059
C2−C3	1.381	1.406	1.393	0.025	−0.013	0.012
C3−C4	1.389	1.409	1.372	0.020	−0.037	−0.017
C4−C5	1.392	1.408	1.434	0.016	0.026	0.042
C5−C6	1.385	1.421	1.386	0.036	−0.035	0.001
C6−C1	1.401	1.424	1.391	0.023	−0.033	−0.010
C1−C11	1.427	1.395	1.407	−0.032	0.012	−0.020
C11−N12	1.152	1.165	1.158	0.013	−0.007	0.006
C2−F13	1.341	1.324	1.296	−0.017	−0.028	−0.045
Bond angle (°)
C1−C2−C3	122.0	124.7	122.5	2.7	−2.2	0.5
C2−C3−C4	118.8	119.0	117.5	0.2	−1.5	−1.3
C3−C4−C5	120.5	118.1	121.1	−2.4	3.0	0.6
C4−C5−C6	120.0	122.7	121.4	2.7	−1.3	1.4
C5−C6−C1	120.4	120.6	119.0	0.2	−1.6	−1.4
C6−C1−C2	118.3	114.9	118.4	−3.4	3.5	0.1

and

Table 6. Bond length and bond angle of the electronic ground state S₀, first excited state S₁, and cationic ground state D₀ of 3-fluorobenzonitrile calculated at RB3LYP/aug-cc-pvtz, TD-B3LYP/aug-cc-pvtz, and UB3LYP/aug-cc-pvtz levels, respectively.

	S0	S1	D0	Δ(S1 − S0)	Δ(D0 − S1)	Δ(D0 − S0)
Bond length (Å)
C1−C2	1.398	1.425	1.380	0.026	−0.045	−0.018
C2−C3	1.380	1.413	1.393	0.033	−0.019	0.013
C3−C4	1.384	1.406	1.433	0.022	0.027	0.049
C4−C5	1.390	1.404	1.374	0.014	−0.029	−0.016
C5−C6	1.387	1.411	1.395	0.024	−0.016	0.008
C6−C1	1.398	1.438	1.449	0.039	0.010	0.051
C1−C11	1.430	1.397	1.415	−0.033	0.017	−0.015
C11−N12	1.152	1.165	1.155	0.013	−0.009	0.003
C2−F13	1.346	1.330	1.300	−0.016	−0.030	−0.046
Bond angle (°)
C1−C2−C3	118.2	118.4	116.9	0.2	−1.4	−1.3
C2−C3−C4	122.5	125.2	123.4	2.7	−1.7	0.9
C3−C4−C5	118.5	115.9	118.8	−2.6	2.9	0.3
C4−C5−C6	120.6	121.1	119.2	0.4	−1.9	−1.4
C5−C6−C1	119.6	122.2	120.9	2.6	−1.3	1.3
C6−C1−C2	120.4	116.9	120.5	−3.4	3.5	0.1

show the bond lengths and bond angles of the S₀, S₁, and D₀ states of 2FBN and 3FBN calculated at levels of RB3LYP/ang-cc-pvtz, TD-B3LYP/ang-cc -pvtz, and UB3LYP/ang-cc-pvtz, respectively. It can be seen that the bond lengths between adjacent carbon atoms of the ring of 2FBN were very close to the corresponding bond lengths of 3FBN. After the transition of S₁ ← S₀, each C−C bond length increased, resulting in the perimeters of ring of 2FBN and 3FBN increasing by 0.160 Å and 0.158 Å, respectively. The transition of D₀ ← S₁ led to the shortening of four C−C bonds and the lengthening of two C−C bonds. The overall effect of D₀ ← S₁ transition was that the perimeters of ring of 2FBN and 3FBN decreased by 0.073 Å and 0.072 Å, respectively. Further, the ring C−C bond lengths of D₀ state was averagely larger than that of S₀ state. The perimeters of the aromatic ring of 2FBN and 3FBN at the cationic ground states were 0.087 Å and 0.086 Å larger than those of the neutral ground state S₀, respectively. That is, the perimeters or average bond lengths of the ring in the ground state S₀, excited state S₁, and cationic ground state D₀ met the relationship: S₀ < D₀ < S₁. The length of a chemical bond reflects, to some extent, the strength of that bond. The greater the bond length, the weaker the bond strength. The frequency of an ideal oscillator is proportional to the square root of the bond strength, so the larger the bond length, the lower the vibration frequency. On this basis, we could predict that, on average, the vibration frequencies of the ground state S₀, the excited state S_1, and the cationic ground state D₀ met the relationship: S₀ > D₀ > S₁. The 33 normal vibration frequencies calculated at the B3LYP/ang-cc-pvtz level of 3FBN were statistically analyzed. On average, the vibration mode frequency of the ground state S₀ was about 21 cm⁻¹ greater than that of the cationic ground state D₀ and the vibration frequency of D₀ was about 43 cm⁻¹ greater than that of S₁. For example, the frequencies of breathing vibration mode 1 for S₀, D₀, and S₁ of 2FBN measured in the experiment were 724 [18], 685, and 668 cm⁻¹, respectively; for mode 12, they were 835 [18], 823, and 815 cm⁻¹, respectively; for mode 18b, they were 1100 [18], 973, and 946 cm⁻¹, respectively. The reported experimental and theoretical data of mFBT and mDFB [47] also indicated that most of the vibrational modes of these two molecules followed this rule. Furthermore, from the above vibration data, we know that the frequency variation was larger for the out-of-plane mode (such as 18b of 2FBN) than for the in-plane mode (such as modes 1 and 12 of 2FBN). Our DFT theoretical results showed that this law held for most vibration modes of benzene derivative.

For 2FBN and 3FBN, the bond lengths of C−N in S₀ state were equal (1.152 Å), also equal in S₁ state (1.165 Å), and almost equal in D₀ state (1.158 and 1.155 Å). This means that the C−N bond was very strong and did not change with the substitution position (ortho- or meta-). The C−F bond length yielded a slight change, with different substitution positions.

The aromatic ring included six bond angles of C−C−C. In the electronic transition, the bond angle of the ring also underwent a certain degree of change. Four angles had a variation of approximately 2–3° and the other two had relatively small changes. In the transitions of S₁ ← S₀ and D₀ ← S₁, variations of the bond angle and bond length of rings led to the ring expansion and contraction, further activating a large number of in-plane vibration modes. Most of the observed vibronic features in the experiments were assigned to in-plane vibrations, only a few of out-of-plane modes were observed. Many benzene derivative molecules have exhibited such characteristics [48,49,50,51,52,53].

3.3. Substitution Effect on Ionization Energy

Molecular IE is an important parameter of molecular characteristics. In order to study the effect of fluorine and CN substitutions on ionization energy, we listed the IEs of benzene [54], fluorobenzene [55], benzonitrile [56], 2-fluorobenzonitrile, 3-fluorobenzonitrile, p-fluorobenzonitrile [21], phenol [57], o-fluorophenol [58], m-fluorophenol [59,60], and p-fluorophenol [11] in

Table 7. Ionization energy of benzene and phenol and their F- and CN-substituted molecules (cm⁻¹).

Molecule	IE	ΔIE	Molecule	IE	ΔIE
Benzene a	74,557	0	Benzonitrile c	78,490	0
Fluorobenzene b	74,227	−330	2-Fluorobenzonitrile d	78,650	160
Benzonitrile c	78,490	3933	3-Fluorobenzonitrile d	78,873	383
			4-Fluorobenzonitrile e	78,000	−490
Phenol f	68,625	0
2-Fluorophenol g	70,006	1381	Fluorobenzene b	74,227	0
3-Fluorophenol, cis h,i	70,188	1563	2-Fluorobenzonitrile d	78,650	4423
3-Fluorophenol, trans h,i	70,449	1824	3-Fluorobenzonitrile d	78,873	4646
4-Fluorophenol j	68,577	−48	4-Fluorobenzonitrile e	78,000	3773

^a Ref. [54]. ^b Ref. [55]. ^c Ref. [56]. ^d This work. ^e Ref. [21]. ^f Ref. [57]. ^g Ref. [58]. ^h Refs. [59,60]. ⁱ Ref. [60]. ^j Ref. [11].

and divided them into four groups for comparison. First, we found that three molecular IEs reduced with respect to their parent molecules, i.e., for the fluorine substitution, the ionization energy of fluorobenzene, 4-fluorophenol, and 4-fluorobenzonitrile was reduced by 330, 490, and 48 cm⁻¹ compared with their parent molecules, respectively, where fluorine played a role of electron donor. However, fluorine-substituted ortho and meta benzonitrile slightly increased the IEs by 160 and 383 cm⁻¹, respectively; fluorine-substituted ortho and meta (cis and trans) phenols increased the IEs by 1381, 1563, and 1824 cm⁻¹, respectively. For these substitutions, fluorine exhibited electron withdrawing properties. It could be seen that the role of fluorine changed with the characteristics of the parent molecule and different substitution positions. Unlike fluorine substitution, CN-substituted benzene and fluorobenzene at ortho, meta, and para positions increased the ionization energy by 3933, 4423, 4646, and 3773 cm⁻¹, respectively, playing a role of strong electron withdrawing.

In addition, we can see from Table 7 that the effects of ortho and meta substitution on ionization energy were very close, while the effect of para substitution was relatively weak. Moreover, the IEs of molecules formed by ortho, meta, and para substitutions met the relative relationship: para < ortho < meta. Most benzene derivative molecules followed this rule.

4. Materials and Methods

4.1. Experimental Methods

The 2-fluorobenzonitrile and 3-fluorobenzonitrile samples were purchased from J&K Chemical and Sigma-Aldrich company, respectively. They were used without further purification. They were a colorless or a light brown liquid with a purity of 99%. The sample was heated to about 130 °C for 2FBN and 60 °C for 3FBN to obtain sufficient vapor pressure. Then, 3 bar krypton for 2FBN and 2.5 bar argon for 3FBN were used as the carrier gases; they carried the sample molecules into the beam source chamber through a pulse valve of 0.5 mm diameter nozzle (0.8 mm for 3FBN). Then, the molecule beam entered the ionization chamber through a skimmer located 20 mm downstream from the nozzle orifice. The vacuum pressures of the beam source and ionization chambers were ~4 × 10⁻⁴ Pa and ~6 × 10⁻⁶ Pa, respectively.

The light source consisted of two sets of dye lasers pumped by YAG lasers. One dye laser (CBR-D-24, Sirah) pumped by a frequency-tripled Nd: YAG laser (Qsmart 850, Quantel) was used as the excitation laser. Another dye laser (Precision Scan-D, Sirah) pumped by another frequency-tripled Nd: YAG laser (Qsmart 850, Quantel) was used as the ionization laser for two-color REMPI or probe laser for MATI experiments. The dyes of coumarin 540A and coumarin 460 or coumarin 480 were used for the excitation and ionization lasers, respectively. The dye laser wavelengths were calibrated by a wavemeter (WS7-60 UV-I). The fundamental outputs of the dye lasers were further frequency-doubled by BBO crystals.

Due to the strong electron-withdrawing ability of the CN group, the transition energies of S₁ ← S₀ were lower than those of D₀ ← S₁ for 2FBN and 3FBN. Such an energy structure indicated that two sets of light sources were required for the measurement of excited state spectra. In the REMPI experiments, we fixed the ionization laser at 232 nm, then scanned the excitation laser from 265 to 279 nm to obtain vibronic spectra of the first electronically excited state S₁ for 2FBN and 3FBN.

In the MATI experiments, the molecules in neutral ground state S₀ were resonantly excited to specific vibronic levels in the S₁ state, further excited to the high Rydberg state by the probe laser, which had a scanning range of 224–240 nm. A −0.5 V/cm pulsed electric field was applied to remove the prompt ions. After a time delay of about 29 μs, the Rydberg molecules were ionized by a 143 V/cm pulsed electric field. Newly formed threshold ions passed through a 48 cm field-free region to be detected by a microchannel plate (MCP) detector. The signal was collected by a multichannel scaler (SRS: SR430) and recorded by a computer. Each mass spectrum was accumulated for 300 laser shots. The time sequence of the whole system was controlled by a pulse delay generator (SRS: DG645). More details on the experimental system have been described in our previous publications [61,62,63].

4.2. Theoretical Methods

All calculations were performed using the Gaussian 16 program package [64]. The geometry optimization and vibrational frequencies of S₀, S₁, and D₀ states were calculated at the levels of RB3LYP/aug-cc-pvtz, TD-B3LYP/aug-cc-pvtz, and UB3LYP/aug-cc-pvtz, respectively. Prior to the experiments, we also used the G4 and CBS-QB3 methods to predict IEs in order to select the appropriate dyes. The theoretical predicted adiabatic ionization energies (AIEs) by CBS-QB3 and G4 for 2FBN were 79,111 and 78,972 cm⁻¹, respectively, with relative errors of +0.59% and +0.41%. The predicted AIEs by CBS-QB3 and G4 for 3FBN were 79,480 and 79,043 cm⁻¹, respectively, with relative errors of +0.77% and +0.22%. The spectral simulations were performed based on the above B3LYP/aug-cc-pvtz calculations, which provided reliable accuracy. The broadening of the REMPI spectral lines was mainly caused by the Doppler effect, while the width of the MATI spectral lines was mainly caused by the ionization field applied to the Rydberg state. The Gaussian line shape, adiabatic Hessian, and time-independent model were used in constructing the spectra [65]. Combined with the theoretical calculations and simulated spectra, the vibrational features of 2FBN and 3FBN measured by the REMPI and MATI experiments were assigned.

5. Conclusions

The high-resolution vibrational spectra of the first electronically excited state S₁ and cationic ground state D₀ of 2-fluorobenzonitrile and 3-fluorobenzonitrile were measured by two-color resonance-enhanced multiphoton ionization and mass-analyzed threshold ionization spectroscopy. The precise band origins of S₁ ← S₀ transition and adiabatic ionization energies were determined to be 36,028 ± 2 cm⁻¹ and 78,650 ± 5 cm⁻¹ for 2-fluorobenzonitrile and 35,989 ± 2 cm⁻¹ and 78,873 ± 5 cm⁻¹ for 3-fluorobenzonitrile, respectively. DFT theory at the level of B3LYP/aug-cc-pvtz was used to calculate the molecular structure, vibrational frequency, and to further perform the Franck–Condon simulations. The theoretical results were in good agreement with the experimental measurements. The vibrational features of S₁ and D₀ states were analyzed in detail and assigned.

The MATI spectra followed well the propensity rule Δν = 0, indicating that the molecular structures of the cationic ground states were similar to that of the excited states. The molecular structures and vibration frequencies in S₀, S₁, and D₀ states were discussed in detail. The ring C-C bond lengths in S₀, S₁, and D₀ states averagely obeyed the rule of S₁ > D₀ > S₀. The bond length reflected the bond strength; further, the bond length was related to the vibration frequency. On average, or for most vibrational modes, the vibration frequencies of the ground state S₀, excited state S₁, and cationic ground state D₀ met the relative relationship: S₁ < D₀ < S₀. At the transition of S₁ ← S₀ and D₀ ← S₁, a lot of vibrational modes associated with ring in-plane distortion were active and only a few out-of-plane fundamental vibrations were observed. The substitution effects of F and CN were discussed. Whether the electron donating group or the electron withdrawing group, the ionization energies of molecules formed by ortho, meta, and para substitutions meet the relative relationship: para < ortho < meta.