DFT Computations to Simulate the IR Spectrum of a Transient Intermediate Generated upon Laser Flash Photolysisof Triarylphosphines Shinro Yasui,* 1 Md. Mizanur Rahman Badal, 2 Shinjiro Kobayashi, 2 and Masaaki Mishima 2 1 Faculty of Contemporary Human Life Science, Tezukayama University, Gakuen-Minami, Nara 631-8585 2 Institute for Materials Chemistry and Engineering, Kyushu University, Hakozaki,Higashi-ku, Fukuoka 812-8581 (Received March 25, 2013; CL-130258; E-mail: yasui@tezukayama-u.ac.jp) The mechanism of photooxidation of triarylphosphines Ar 3 P under air was explored using a combination of a theoretical technique and an experimental technique. Computations based on density functional theory (DFT) were performed to simulate an IR spectrum of a possible intermediate Ar 3 P + OO • , the absorption bands of which are in good agreement with the bands on the transient spectrum observed by laser flash photolysis- time-resolved infrared spectroscopy (LFP-TRIR). The radical cation Ar 3 P •+ generated upon LFP is likely to be trapped by O 2 . Density functional theory (DFT) is an economical and useful tool to simulate many aspects of a molecule as well as dynamics of the molecule. An example is the simulation of IR spectra, about which several reports have been published. 1 Meanwhile, laser flash photolysiswith time-resolved infrared spectroscopy (LFP-TRIR) has been developed to detect a transient intermedi- ate during photoprocesses. 2 It appears that the combination of a theoretical technique such as DFT computation and an exper- imental technique such as LFP-TRIR might provide a powerful tool to explore the mechanistic aspects of a photoreaction. 3 Recently, we found that steady-state photolysisof Ar 3 P 1 under air results in the oxidation of 1 to the corresponding phosphine oxide Ar 3 P=O 2. Dependency of the reaction rate on the wavelengths ofirradiating light suggested strongly that the singlet excited state of 1, 1 Ar 3 P* 3, 4 leads to 2. The excited state 3,either from this state or after intersystem crossing to its triplet state, likely undergoes electron transfer (ET) to oxygen O 2 to generate a radicalion pair, namely, the radical cation Ar 3 P •+ 4 and superoxide radical anion O 2 •¹ . In fact, our previous experi- ments using LFP with time-resolved UV-visible spectroscopy (LFP-TRUV) 5 have shown that LFP on 1 results in the formation of a transient intermediate assignable to 4 within tens of nano- seconds after the flash. 6 The resulting radical cation 4 would undergo reaction with O 2 under the reaction conditions, giving peroxidic radical cation Ar 3 P + OO • 5 (path A in Scheme 1). 7 Alternatively, if the radicalion pair, 4 and O 2 •¹ , couplewith each other before each component diffuses into the solvent bulk, then the transient intermediate would be phosphadioxirane 6 (path B in Scheme 1). Such a phosphadioxirane intermediate is some- times postulated in the photooxidation of trivalent phosphorus compounds by singlet oxygen ( 1 O 2 ). 8 That is, 6 could be formed here as well if 1 O 2 is generated under our photochemical conditions. An objective of this study is to disclose which intermediate, 5 or 6, is formed in the course of the photooxidation of 1 to 2. To accomplish this objective, we attempted to detect a transient intermediate on the LFP-TRIR and identify the structure of the intermediate by comparing its IR spectrum with DFT- simulated spectra for candidate intermediates. Triarylphosphines 1a-1j used in this study are given in Chart 1. LFP-TRIR experiments were performed with a continuum cw Q-sw Nd:YAG laser equipped with a JASCO TRIR-1000 dispersive-type IR spectrometer. When the LFP experiment was conducted in an acetonitrile solution of 1 (1-6 mM) under air at 266 nm, a transient spectrum consisting of several absorption bands appeared at the region of 1050-1300 cm ¹1 on the TRIR with a microsecond time-scale. Representative examples are shown inFigure 1. Resolution with respect to wavenumber was 4-12 cm ¹1 . The LFP was carried out also using dichloromethane as solvent, resulting in the formation of a spectrum with almost identical absorption bands for each derivative. To identify the transient intermediate observed here, simulations of IR spectra by theoretical computation were performed with a Gaussian 09 program package. 9 The IR spectra of 5a-5j were simulated by computation with the DFT/B3LYP level using 6-31G(d) as a basis set. The IR spectra of 1a-1j were likewise simulated. When any band in the simulated spectra for 1 was sufficiently intense, it was subtracted from the corresponding band from 5 because, in the present LFP-TRIR experiments, absorption bands from 1 in the whole observable region are offset to zero at t = 0. 10 The wavenumbers of the absorption bands in the observed spectra (listed in the second column in Table 1) were then plotted against those of the corresponding absorption bands of the simulated spectrum for 5 (in the third column in Table 1). Table1also lists the vibration 1 Ar 3 P * [ Ar 3 P •+ O 2 •− ] O 2 3 + Ar 3 POO • • Ar 3 P O O 6' 6 Ar 3 P •+ O 2 •− 4 + + O 2 Ar 3 POO • + 5 path A path B ET Ar 3 P 1 LFP at 266 nm Scheme 1. A supposed mechanism offormation of a transient intermediate 5 or 6. X = 1a; H 1b; 2-Me 1c; 3-Me 1d; 4-Me 3 P X 2 P Y Y = 1i ; 2-Me 1j ; 4-Me X = 1e; 4-MeO 1f; 4-Cl 1g; 4-F 1h; 2,4,6-Me 3 Chart 1. Triarylphosphines used in this work. Published on the web May 30, 2013 866 doi:10.1246/cl.130258 © 2013 The Chemical Society of Japan Chem. Lett. 2013, 42, 866-868 www.csj.jp/journals/chem-lett/