Atmosphere-controlled dual reactivity of triarylphosphine in the photoexcited state: PC bond cleavage vs. electron transfer

Full text of DOI:10.1246/cl.130748

doi:10.1246/cl.130748
1478
Published on the web September 14, 2013
Atmosphere-controlled Dual Reactivity of Triarylphosphine in the Photoexcited State:
P-C Bond Cleavage vs. Electron Transfer
Shinro Yasui,*1 Yuya Ogawa,2 Kosei Shioji,2 and Shoko Yamazaki3
1Faculty of Contemporary Human Life Science, Tezukayama University, Gakuen-Minami, Nara 631-8585
2Department of Chemistry, Faculty of Science, Fukuoka University, Jonan-ku, Fukuoka 814-0180
3Department of Chemistry, Nara University of Education, Takabatake-cho, Nara 630-8528
(Received August 12, 2013; CL-130748; E-mail: yasui@tezukayama-u.ac.jp)
Steady-state photolysis of an acetonitrile solution of
triarylphosphine (Ar3P) was carried out using a xenon lamp.
The major products resulting from photolysis under deoxy-
genated conditions and under air were diaryl(cyanomethyl)-
phosphine Ar2PCH2CN and triarylphosphine oxide Ar3P=O,
respectively. Intrinsically, Ar3P in the excited state undergoes
homolytic cleavage of a P-C bond; however, electron transfer to
O2 is predominant under air. Reactivity of Ar3P in the excited
state is readily controlled by the choice of atmosphere used in
the reaction.
in the absence of O2
Ar
Ar2PCH2CN
2P-PAr2 + others
6
+
3
Xe lamp
(major)
(minor)
Ar3P
in CH3CN
1
in the presence of O2
1a; Ar = 2-methylphenyl
1b; Ar = 4-methylphenyl
1c; Ar = 2,4,6-trimethylphenyl
+
Ar2P-PAr2
Ar3P=O
2
6
(major)
(small amount)
Scheme 1. Photolysis of triarylphosphines under deoxygen-
ated conditons and under air.
Triarylphosphines (Ar3P) are a class of conventional organic
reagents, which exhibit absorption bands around 250-350 nm
due to an n ¼ π* transition. This suggests that the activation of
Ar3P by photoexcitation may be a useful strategy in the field of
organic synthesis. In order to be able to use Ar3P as an organic
reagent under irradiation, it is essential to understand its
photochemical behavior. In our previous work on laser flash
photolysis time-resolved infrared spectroscopy (LFP-TRIR) of
Ar3P, we found that Ar3P, when photoexcited under air,
undergoes electron transfer (ET) to oxygen O2, generating the
radical cation Ar2P•+.1 On the other hand, it has been reported
that UV laser flash of Ar3P under completely oxygen-free
conditions results in homolytic cleavage of a P-C bond to form
the diarylphosphinyl radical Ar2P• and aryl radical Ar•.2 We now
know through these studies that, depending on the atmosphere,
different transient intermediates are formed initially upon
irradiation. In the present study, we carried out steady-state
photolysis on tris(2-methylphenyl)phosphine (1a), tris(4-meth-
ylphenyl)phosphine (1b), and tris(2,4,6-trimethylphenyl)phos-
phine (1c) with varying wavelengths of irradiating light and
under different atmospheres, and the products were analyzed
by gas-chromatography (GC) and 31P NMR spectroscopy
(Scheme 1). A unique compound, diaryl(cyanomethyl)phos-
phine, was obtained in substantial yield under “deoxygenated”
conditions, without a rigorous degassing process.
An Ushio xenon short arc lamp (UXL-500D-0) was used as
the source of light. Wavelength of the irradiating light was
controlled by a Toshiba sharp-cut glass filter that was put
between the lamp and reaction vessel. The progress of the
reaction was followed by GC with a Shimadzu GC14A. In some
cases, 31P NMR spectra were recorded on a Varian INOVA400
operating at 161.9 MHz (85% phosphoric acid as the reference
standard (0.0 ppm)) to follow the reaction.
When an acetonitrile solution of 1 (1.0 mM) was photolyzed
under air, the corresponding phosphine oxide 2 was detected
along with 1 in GC experiments. The yields of 2 as well as
recovered 1 with varying wavelengths of irradiating light are
Table 1. Photoreaction of triarylphosphines Ar3P 1a
Yield/%b
Irradiated at
/nm
Time
/min
Run
1
Atmosphere
1c
2
3
1
2
3
4
5
6
7
8
9
10
11
12h
13
1a >310 (3.20)d
>350 (1.48)d
>310 (3.20)d
>350 (1.48)d
1b >310 (2.80)d
>350 (1.69)d
>310 (2.80)d
1c >310 (4.16)d
>350 (3.37)d
>390 (1.26)d
>420 (®)g
air
air
10
4
68
0
0
60 76 15
45
180 81 15
20 69
60 80 20
180 10 16
argone
argone
air
0
7
>29f
0
0
0
6
air
argone
air
>26f
2
10
0
2
46
58
0
0
0
air
air
air
60 19 67
90 89
7
6
5
5
0
53 (50)i
0
>310 (4.16)d
>420 (®)g
argone
argone
0
60 88
aIn acetonitrile. At room temperature. [1] = 1.0 © 10¹3 M.
bDetermined on GC based on the initial amount of 1.
d
cRecovered. The values in parentheses denote the logarithm
of the extinction coefficient (M¹1 cm¹1) of the shortest
e
wavelength of the irradiating light. The system may contain
f
a trace amount of O2; see the text. Based on an assumption
that the factor of GC analysis for 3 is larger than that for 1. The
g
h
assumption holds for 3c. No absorption observed. Mesity-
lene 5c (92%) was detected. iIsolate yield after 20 min
irradiation.
summarized in Table 1. To eliminate the effects of O2, the
photolysis of 1 (1.0 mM in acetonitrile) was carried out under
deoxygenated conditions, where air was replaced with as much
argon as possible. GC analysis of the reaction mixture showed a
major peak at a retention time slightly less than that of 1, along
with a small peak corresponding to 2.3 To identify this product,
GCMS analysis was performed with a JEOL JMS-GC mate II
for each photolyzed solution. The fragments obtained are listed
Chem. Lett. 2013, 42, 1478-1480
© 2013 The Chemical Society of Japan
www.csj.jp/journals/chem-lett/

1479
Table 2. Fragments resulting from the major product formed in
the photolysis of 1a-1c in acetonitrile under deoxygenated
conditionsa
m/z
1a
1b
1c
Fragment 1
Fragment 2
Fragment 3
253 [255]
238 [240]
213b [213]
253 [255]
309 [311]
294b [296]
269 [269]
213b [213]
aValues in brackets denote m/z values of the fragments
obtained in the photolysis in acetonitrile-d3. Base peak.
1000
1500
2000
2500
3000
3500
b
wavenumber / cm-1
Figure 1. IR spectra of Mes2PCH2CN (3c). Lower (in black);
Observed on FT-IR spectrometer measured in carbon
tetrachloride (0.04 mM). Upper (in gray); Simulated based on
DFT B3LYP/6-31G(d) with a scaling factor 0.9892. See ref 7.
a
in Table 2, where fragments 1, 2, and 3 are assignable to
Ar2PCH2CN, Ar2PCH2CN-15, and Ar2P, respectively, for each
derivative. For photolysis carried out in acetonitrile-d3 under
otherwise identical conditions, GCMS gave fragments that are
expected to result from Ar2PCD2CN (shown in brackets in
Table 2). With these results from GCMS analyses, the product
was assigned to diaryl(cyanomethyl)phosphine Ar2PCH2CN 3.
This assignment is further supported by spectroscopic character-
ization of the material isolated from the photolysis of 1c.4 Thus,
the 1H and 13C NMR spectra of the isolated material are
compatible with the structure of 3c,5 and the 31P NMR chemical
shift (¹23.1 ppm) also supports the structure of 3c, a compound
containing a trivalent phosphorus.6 The IR spectra (recorded on
a JASCO FT/IR-4200) of a carbon tetrachloride solution of the
material (0.06 mM) exhibit a characteristic absorption band of
C-N stretching at 2243 cm¹1. The IR spectrum over the entire
measured range is in excellent agreement with the spectrum
theoretically simulated for 3c based on DFT B3LYP/6-31G(d)
(Figure 1).7 To summarize these results, the photolysis of 1 in
acetonitrile under deoxygenated conditions affords 3 as a major
product. The yields of 3 as well as 2 are given in Table 1.
Besides 3 and 2, GCMS analysis of 1c, photolysed under
deoxygenated conditions, showed a tiny peak with a shorter
retention time. Its mass spectrum (m/z 105, 135, 150 (base),
255, and 270) was in accordance with the structure of bis(2,4,6-
trimethylphenyl)phosphine Mes2PH (4c). Mesitylene MesH (5c)
was also detected in GC experiments in 92% yield. The
photolysis of 1c in hexane under deoxygenated conditions did
not form 3c; however, a tiny peak corresponding to 4c was
observed in GC experiments.
The sum of the yields of 2 and 3 is smaller than the
conversion of 1 either under air or under deoxygenated
conditions (Table 1). To detect product(s) that were undetectable
in GC experiments, 31P NMR spectroscopy was used to analyze
the photolysis of 1c in acetonitrile-d3. The spectrum obtained
after a 2-min irradiation with light of wavelengths over 310 nm
under deoxygenated conditions, displayed signals resonating at
¹24.7 and ¹29.5 ppm, along with a small signal from 1c at
¹35.6 ppm. The signal at ¹24.7 ppm, which resonates at slightly
upper field than that from the isolated 3c, i.e., a hydrogen
analogue (¹23.1 ppm), most likely results from 3c-d2
(Mes2PCD2CN).8 The signal at ¹29.5 ppm is assignable to
1,1,2,2-tetrakis(2,4,6-trimethylphenyl)diphosphine Mes2PPMes2
(6c) based on the value of the chemical shift reported in
literature.9-11 The spectrum obtained on photolysis for 1 min
under air displayed a weak signal corresponding to 6c besides an
intense signal from 2c. Product 6c, which is undetectable on our
GC apparatus, accounts for the discrepancy between the sum of
the yields of 2 and 3 and the conversion yields of 1.
Table 1 lists the extinction coefficient ¾ at the shortest
wavelength of the irradiating light. Clearly, the stronger the
absorption of 1, the higher the conversion of 1, which clearly
indicates that reactions leading to the observed products are
initiated by the photoexcitation of 1 to its singlet excited state
11*.12 Intersystem crossing to the triplet state could take place
1
under some circumstances. It is more likely that 1* directly
undergoes subsequent reactions, because the system under
consideration in the present study does not have a triplet
sensitizer. Under deoxygenated conditions, 11* can undergo
homolytic P-C bond cleavage, as has been reported,2 to form the
radical pair, Ar2P• and Ar•. Nearly quantitative formation of
mesitylene 5c in the reaction of 1c (Run 12 in Table 1) strongly
suggests generation of Ar•, which would readily abstract a
hydrogen atom from the solvent CH3CN to give 5c as well as the
•
cyanomethyl radical CH2CN (eq 1).
Ar•
CH3CN
+
•CH2CN
ð1Þ
+
ArH
An almost theoretical amount of MesD (mesitylene-d1) was
obtained when photolysis was carried out in CD3CN. Radicals
•CH2CN and Ar2P•, if formed, explain the observed products.
Thus, they couple with each other to afford 3, while dimerization
of Ar2P• forms 6. A small amount of 4 results from hydrogen
abstraction from the solvent molecule by Ar2P• (Scheme 2). An
•
alternate possibility is that CH2CN and Ar2P• attack the parent
Ar3P 1, which exists in large amounts, resulting in the formation
of phosphoranyl radicals 7 and 8, respectively. These radicals
would afford 3 (eq 2) and 6 (eq 3), respectively, by ejecting Ar•.
However, theoretical calculations based on DFT B3LYP/6-
31G(d) predict extremely congested and unstable structures of
7 and 8, especially when Ar = mesityl. In addition, while the
reactions in eqs 2 and 3 predict a catalytic character in
producing 3 and 6, this is not the case. When the irradiation
was switched off, the progress of the reaction was found to be
halted.
•
•CH2CN + Ar3P
Ar3PCH2CN
ð2Þ
ð3Þ
3
6
- Ar•
- Ar•
1
7
•
Ar2P•
+
Ar3P
Ar2P-PAr3
1
8
Chem. Lett. 2013, 42, 1478-1480
© 2013 The Chemical Society of Japan
www.csj.jp/journals/chem-lett/

1480
of phosphinites on ClCH2CN,17 or through the reaction of
chlorophosphines with LiCH2CN.18 Compared with these
methods, the method of photolysis is quite simple. It is a one-
pot and apparently one-step reaction without any additives. No
rigorous degassing process is required to obtain 3 in practical
yields even though some phosphine is wasted by its oxidation to
phosphine oxide.
Ar2PH
4
1Ar3P*
Ar2P•
Ar2P-PAr2
11*
6
Ar•
•CH2CN
Ar2PCH2CN
hν
3
CH3CN
Further studies on using laser flash photolysis to reveal a
whole scheme of photoreactions of Ar3P, with an attempt to
specify multiplicity of the active species, is underway.
Ar3P
1
ArH
5
1
Scheme 2. Reaction of 1* under deoxygenated conditions.
This study was financially supported by a Tezukayama
research grant 2012.
The singlet excited state, 11*, if formed under air, may
undergo ET to molecular oxygen O2, generating the radical
References and Notes
•¹
cation Ar3P•+ 1•+ and superoxide radical anion O2 . Taking
1
a) S. Yasui, M. Mishima, Phosphorus, Sulfur Silicon Relat. Elem.
2011, 186, 838. b) S. Yasui, M. M. R. Badal, S. Kobayashi, M.
Mishima, Chem. Lett. 2013, 42, 866.
oxidation potentials of 1 in the ground state13 together with the
excitation energy of 1 into consideration, oxidation potentials
2
3
a) Y. Sakaguchi, H. Hayashi, Chem. Phys. Lett. 1995, 245, 591. b) Y.
Sakaguchi, H. Hayashi, J. Phys. Chem. A 2004, 108, 3421.
For photolysis under deoxygenated conditions, solvent was bubbled
with argon for several hours to degas oxygen from the solvent, and
air in each vessel was carefully replaced by argon. Nevertheless,
formation of a small amount of 2 in the photolysis under these
conditions indicates that the system was more or less contaminated by
O2. Argon gas used in this work may contain a trace amount of O2. Or,
we may have failed to get rid of O2 from the system completely.
Trimesitylphosphine (1c) was photolyzed for 20 min under deoxy-
genated conditions in several repeated experiments (54mg in total),
and the combined reaction mixture was subjected to column
chromatography using a mixture of dichloromethane-hexane (1:1
(v/v)) as eluent, which gave 20 mg of white powder (50% as 3c).
1H NMR (CDCl3): δ 2.26 (m, 18H, CH3), 3.26 (d, 2JPCH = 4.7 Hz, 2H,
CH2), 6.81 (s, 4H, CH), 13C NMR (CDCl3): δ 16.6, 21.1, 23.2, 118.3,
130.6, 139.1, 141.7.
1
of 1* are predicted to be between ¹4 and ¹3 V vs. Ag/Ag+.
These values are much more negative than the reduction
potential of O2 (¹1.05 V vs. Ag/Ag+),14 indicating that ET
from 11* to O2 is exothermic. The resulting radical cation 1•+ is
trapped by O2 to give the peroxidic radical cation Ar3P+OO• as
a transient intermediate, which eventually affords phosphine
oxide 2 (eq 4).1 That is, the formation of 2 is explained by the
radical cation intermediate 1•+. This mechanism predicts the
4
•¹
production of O2 , which was indeed suggested by an experi-
ment using the spin trap agent, 2-diphenylphosphinoyl-2-
methyl-3,4-dihydro-2H-pyrrole N-oxide (DPhPMPO). When a
solution of 1c in dimethyl sulfoxide (DMSO) was photolyzed
at 365 nm with bubbling O2 in the presence of DPhPMPO
5
6
7
•¹
for 3 min, the DPhPMPO-O2 adduct was detected by ESR
J. C. Tebby, in Phosphorus-31 NMR Spectroscopy in Stereochemical
Analysis, ed. by J. G. Verkade, L. D. Quin, VCH, Deerfield Beach, FL,
1987, Chap. 1.
spectroscopy on a JEOL RE ESR spectrometer.15 31P NMR
spectroscopy has confirmed the formation of a small amount of
6c from the photolysis of 1c under air as well, which suggests
that 11* undergoes P-C bond cleavage in competition with ET to
O2 under air. Interestingly, it has been reported that nitrogen
counterparts, aromatic amines such as N,N-dimethylanilines,
The wavenumbers of the major absorption bands in the simulated IR
spectrum are linearly correlated with those from the observed
spectrum, giving a slope of 0.9892. This value was taken as “a
scaling factor” to adjust the simulated spectrum to the observed one.
This value is close to the scaling factor 0.9613 valid for gas phase at
this level of theory. J. B. Foresman, Æ. Frisch, Exploring Chemistry
with Electronic Structure Methods, 2nd ed., Gaussian, Inc., Pittsburgh,
PA, 1996, Chap. 4.
also undergo ET to O2 under photochemical conditions to give
•¹ 16
the radical pair, aniline radical cation and O2
.
8
9
A. A. Borisenko, N. M. Sergeyev, Y. A. Ustynyuk, Mol. Phys. 1971,
22, 715.
H. H. Karsch, G. Hanika, B. Huber, J. Riede, G. Müller, J. Organomet.
Chem. 1989, 361, C25.
•−
O2
O2
+ O2
+
11*
1•+
Ar2POO•
ð4Þ
2
10 W. McFarlane, N. H. Rees, L. Constanza, M. Patel, I. J. Colquhoun,
J. Chem. Soc., Dalton Trans. 2000, 4453.
11 S. Sasaki, F. Murakami, M. Yoshifuji, Tetrahedron Lett. 1997, 38,
7095.
In summary, this study disclosed the dual reactivity of Ar3P
1
1 in the excited state, most likely the singlet state 1*, exerted
depending on the atmosphere. Intrinsically, 11* cleaves at a P-C
bond homolytically into Ar2P• and Ar•, which afford products
12 Our previous work based on chemical actinometry has approximated
intensity of the incident light from our apparatus on the surface of the
1
3-6. Meanwhile, 1* undergoes ET to O2 in the ground state
reaction vessel to be 7 © 10¹6 einstein s¹1, based on which we roughly
¹2
estimate quantum yield of the conversion of 1 in this work 10¹3-10
.
when O2 is present. The ET results in formation of the radical
cation 1•+, which eventually leads to phosphine oxide 2. This
study also showed that the number and the position of methyl
substituents on the aryl ligands exert only little effect on the
reactivity of 11*. It is worth pointing out that the photolysis of 1
in acetonitrile under deoxygenated conditions affords the unique
compound 3. This class of phosphines may be useful in
coordination chemistry, where there is a demand for electroni-
cally and sterically tuned ligands. Cyanomethylphosphines have
been previously prepared through the reduction of cyanometh-
ylphosphine oxides, which were obtained by the Arbuzov attack
S. Yasui, M. Tsujimoto, K. Shioji, A. Ohno, Chem. Ber. 1997, 130,
1699.
13 S. Yasui, M. Tsujimoto, J. Phys. Org. Chem. 2008, 21, 541.
14 a) D. S. Steichen, C. S. Foote, J. Am. Chem. Soc. 1981, 103, 1855. b)
K. Mizuno, T. Tamai, I. Hashida, Y. Otsuji, J. Org. Chem. 1995, 60,
2935. c) N. Somasundaram, C. Srinivasan, J. Org. Chem. 1996, 61,
2895.
15 M. Nishizawa, K. Shioji, Y. Kurauchi, K. Okuma, M. Kohno, Bull.
Chem. Soc. Jpn. 2007, 80, 495.
16 M. Hori, H. Itoi, H. Tsubomura, Bull. Chem. Soc. Jpn. 1970, 43, 3765.
17 O. Dahl, F. K. Jensen, Acta Chem. Scand., Ser. B 1975, 29, 863.
18 R. Zhang, Y. Zhou, Synthesis 1987, 938.
Chem. Lett. 2013, 42, 1478-1480
© 2013 The Chemical Society of Japan
www.csj.jp/journals/chem-lett/