S. Tanaka, N. Nomura, T. Nishioka et al.
Journal of Organometallic Chemistry 943 (2021) 121843
solution was then stirred at 70 °C for 3 h under an N2 atmo-
sphere to obtain an orange suspension. The precipitate was fil-
tered under reduced pressure, and the filtrate was evaporated to
dryness. The orange residue was suspended in ice-cold n-hexane
and collected by filtration. The collected solid was recrystallized in
CH2Cl2/cyclohexane = 1: 10 (v/v) at 0 °C to produce fine crystals,
which were collected by decantation, washed with THF, and dried
in vacuo (137 mg, 2.1 × 10−4 mol, 93%). The product was stored
under ambient conditions until use.
are shown in Figure S12. The selected bond distances and angles
for the complexes are shown in Table S1, and the molecular coor-
dinates are listed in Tables S2-S4. Harmonic vibrational frequen-
cies were calculated for the optimized geometries using the 6-
311G(d,p) level.
Time-dependent DFT calculations for [3]+ and [3H+]+ were per-
formed using the optimized structures at the uB3LYP/6-311+G(d,p)
level. Electronic excitations and the related molecular orbitals are
given in Tables S7 and S5 for [3]+ and Tables S8 and S6 for [3H+]+,
respectively. Selected Mulliken atomic spin densities for [3]+ and
[3H+]+ are shown in Table S9.
A single crystal suitable for X-ray crystallography was obtained
by recrystallization with CH2Cl2/toluene/n-hexane = 1: 1: 4 at 0
2
°C. 1H NMR (300 MHz, C6D6): δ 3.43 (d, 1H, JHH = 15.6 Hz, PTA),
2
2
3.56 (d, 1H, JHH = 15.1 Hz, PTA), 3.70 (d, 1H, JHH = 12.8 Hz, PTA),
2.8. X-ray crystallography
2
3
3.90 (d, 1H, JHH = 12.5 Hz, PTA), 6.04 (ddd, 1H, JHH = 5.7, 7.3
4
3
Hz, JHH = 1.4 Hz, 5-py), 6.74 (t, 1H, JHH = 7.5 Hz, 4-py), 6.89
Diffraction data for
3 was collected by using a Rigaku
3
4
(ddd, 1H, JHH = 7.2, 7.5 Hz, JHH = 1.6 Hz, 4-BPT), 6.97 (ddd, 1H,
AFC11/Saturn 724+ CCD diffractometer with monochromated Mo-
3JHH = 7.2, 7.8 Hz, JHH = 1.6 Hz, 5-BPT), 7.07 (d, 1H, JHH = 7.9
4
3
˚
Kα radiation (λ = 0.710747 A) (Rigaku Co.). The data were pro-
3
Hz, 3-py), 7.21 (tt, 1H, JHH = 7.7 Hz, JPH = 1.3 Hz, 5-BPT), 7.48 (dd,
cessed and corrected for Lorentz and polarization effects using the
CrystalClear software package. The analyses were carried out us-
ing the WinGX software [50]. Absorption corrections were applied
using the MultiScan method. The structures were solved using di-
rect methods (SIR97) [51] and refined by full-matrix least-squares
with F2 using SHELXL-2018/3 [52]. Crystallographic data are sum-
marized in Table S10. Non-hydrogen atoms were refined anisotrop-
ically. Hydrogen atoms were located in a difference Fourier map
and isotopically refined.
´
3
4
3
1H, JHH = 7.7 Hz, JHH = 1.0 Hz, 4-BPT), 7.60 (dd, 1H, JHH = 7.7
´
4
3
Hz, JHH = 1.4 Hz, 6-BPT), 7.87 (d, 1H, JHH = 7.6 Hz, 6-BPT),
´
3
4
8.02 (dd, 1H, JHH = 7.7 Hz, JHH = 1.6 Hz, 3-BPT) 8.51 (d, 1H,
3JHH = 5.6 Hz, 6-py). 31P{1H} NMR (121 MHz, C6D6): δ −32.4. IR
(KBr): νCO = 1908 cm−1. Anal. Calcd for 3 (C30H35FeN7OP2S): C,
54.63; H, 5.36; N, 14.87. Found: C, 54.23; H, 5.41; N, 14.59.
2.5. Synthesis of trans-[FeIII(PyBPT-κ3N,C,S)(CO)(PTA)2]PF6 ([3]PF6)
Tris(4-tolyl)aminium hexafluorophosphate (113 mg, 2.7 × 10−4
mol) was mixed with 3 (100 mg, 1.5 × 10−4 mol) in CH2Cl2 (20
mL) in the dark. The reaction mixture was stirred at room temper-
ature for 1 h, followed by the addition of toluene (10 mL). After
removing the solvent to a minimum volume under reduced pres-
sure, the suspension was filtered to collect a dark greenish-blue
powder, which was rinsed with toluene and dried in vacuo (83 mg,
1.0 × 10−4 mol, 68%). The mass spectrum for the product was con-
sistent with the composition formula of [3]+ (Figure S1). The prod-
uct was stored in the dark under ambient conditions until use.
3. Results and discussion
3.1. Synthesis of N,C,S-pincer iron(II/III) carbonyl complex with PTA
ligand
The synthesis of trans-[Fe(PyBPT-κ3N,C,S)(CO)(PTA)2] (3) was
achieved in analogy with the synthesis of 1 and 2 [41]. Simple re-
placement of PMe3 of 1 with PTA in synthetic procedures provided
3. Complexes 1 - 3 are the one-electron reduced forms and syn-
thetic sources of [1]PF6 - [3]PF6, respectively. The assignment of
the 1H NMR signals for 3 in C6D6 was carried out in comparison
with the spectra obtained for [{Fe(μ-PyBPT-κ3N,C,S)(CO)2}Fe(CO)3]
[48] and intact PTA (Figure S2). Signals at 6.05-8.52 ppm were as-
signed to the protons of PyBPT. The 31P{1H} NMR spectrum for 3
in C6D6 showed a single signal at -32.4 ppm, which was similar
to that observed for PMe3 of 1 [41] and confirmed that the PTA
ligands were bound to the central iron (Figure S3). The single sig-
nal suggested that the coordination spheres of the two PTA ligands
were equivalent in solution [41].
2.6. Photoreaction of [3]PF6 and [3H+]PF6
A 0.24 mM CH2Cl2 solution of the complex placed in a 1 cm2
square optical cell (1 mL) was irradiated with a monochrome
light (λ = 400 - 800 nm) generated by a fluorescent spectrome-
ter (HITACHI F-7000 fluorescence spectrometer, Hitachi Co.) in the
dark. A UV light cut filter was used to remove light diffracted at
a higher order of the incident light from the monochrome light
before irradiating the sample solution. The photon flux density
(/mol•s−1•m−2) at each wavelength adopted in this study was cal-
culated from the optical power density measured on the surface of
the sample cell using a pyranometer (ML-01, EKO Instruments Co.)
with an effective spectral range of 400-1100 nm (Figure S11). UV-
Vis spectra for the reaction solution were recorded every 15 min
to 30 min during irradiation. The photolytic kinetics was traced by
changes in absorbance at 600 and 612 nm for [3]PF6 and [3H+]PF6,
respectively. Photolysis plots were obtained as the average of three
independent experiments and analyzed by the first-order kinetic
equation.
A crystal structure of 3 obtained by X-ray analysis is depicted
Selected bond distances and angles for 3 are listed in Table 1.
Corresponding data for 1 and 2 are reproduced in the same table
for comparison [41]. The overall conformation of 3 was similar to
those of 1 and 2,[41] except for a larger bending of the equato-
rial tridentate ligand at the thiophenol ring. As with cases 1 and
2, the structural variance in the crystal and NMR data implied a
quick flip-flop motion of the equatorial ligand in solution [41]. The
˚
mean Fe-P distances followed the order of 2.243 A (1, PMe3) >
˚
˚
2.223 A (3, PTA) > 2.206 A (2, P(OEt)3), which was reversed be-
tween 3 and 2 in terms of the steric bulk of the ligands estimated
from the Tolman cone angles (118° (1, PMe3) > 109° (2, P(OEt)3)
> 103° (3, PTA)) [44,47]. The more bent shape of the equatorial
ligand observed in the crystal structure of 3 is a possible cause of
larger steric hindrance, which prevents the PTA ligands from ap-
proaching more closely the central iron despite the smaller cone
angle compared to P(OEt)3.
2.7. Computational details
The structures of complexes 3, [3]+, and [3H+]+ were optimized
by DFT calculations using the Gaussian 09 program package [49].
The crystal structure of 3 was used as the initial model of 3, [3]+,
and [3H+]+. The rB3LYP (3) and uB3LYP ([3]+ and [3H+]+) den-
sity functional methods and the 6-311+G(d,p) basis set were used
for the calculations. Optimized structures for 3, [3]+, and [3H+]+
Alternatively, the stronger π-electron acceptance of P(OEt)3
from the central iron in comparison to the other phosphine lig-
3