Photoinduced oxidation of triphenylphosphine isolated in a low-temperature oxygen matrix

Article abstract of DOI:10.1016/j.cplett.2008.11.034

Photooxidation reactions of triphenylphosphine (Ph₃P) monomers isolated in matrices of solid oxygen at 10 K were characterized by means of infrared spectroscopy. Upon UV (λ > 280 nm) irradiation of O₂ matrices, ca. 90% of Ph₃P was converted to triphenylphosphine oxide (Ph₃P{double bond, long}O), with concomitant formation of ozone. In the competing photoreaction, ca. 10% of Ph₃P was converted to diphenyl-phenyl-phosphinate, Ph₂(Ph-O-)P{double bond, long}O. The interpretation was assisted by theoretical [DFT(B3LYP)/6-31G(d, p)] calculations of vibrational spectra, as well as by comparison with the experimental vibrational data from separate experiments in which monomeric Ph₃P and Ph₃P{double bond, long}O were isolated in argon and oxygen matrices at 10 K.

Full text of DOI:10.1016/j.cplett.2008.11.034

Chemical Physics Letters 467 (2008) 97–100
Contents lists available at ScienceDirect
Chemical Physics Letters
journal homepage: www.elsevier.com/locate/cplett
Photoinduced oxidation of triphenylphosphine isolated in a low-temperature
oxygen matrix
a,b,*
, Leszek Lapinski , Maciej J. Nowak b
b
Igor Reva
a
Department of Chemistry, University of Coimbra, Rua Larga, 3004-535 Coimbra, Portugal
Institute of Physics, Polish Academy of Sciences, Al. Lotnikow 32/46, 02-668 Warsaw, Poland
b
a r t i c l e i n f o
a b s t r a c t
Article history:
Photooxidation reactions of triphenylphosphine (Ph P) monomers isolated in matrices of solid oxygen at
3
Received 8 October 2008
In final form 10 November 2008
Available online 14 November 2008
1
0 K were characterized by means of infrared spectroscopy. Upon UV (k > 280 nm) irradiation of O
ces, ca. 90% of Ph P was converted to triphenylphosphine oxide (Ph P@O), with concomitant formation of
ozone. In the competing photoreaction, ca. 10% of Ph P was converted to diphenyl-phenyl-phosphinate,
Ph
(PhꢀOꢀ)P@O. The interpretation was assisted by theoretical [DFT(B3LYP)/6-31G(d,p)] calculations of
vibrational spectra, as well as by comparison with the experimental vibrational data from separate exper-
2
matri-
3
3
3
2
iments in which monomeric Ph
3
P and Ph
3
P@O were isolated in argon and oxygen matrices at 10 K.
Ó 2008 Elsevier B.V. All rights reserved.
1
. Introduction
The rich chemistry of phosphines is dominated by strong nucle-
2. Experimental and computational methods
Commercial samples of Ph
3
P (99%) and Ph
3
P@O (98%) supplied
of
spectral purity were supplied by Linde AG. Matrices were pre-
pared by deposition of the studied compounds, together with a
ophilicity and reducing character of these compounds. Trivalent
phosphorus can be easily oxidized and converted into quinqueva-
lent phosphorus (see Scheme 1). Recently, it has been reported that
singlet molecular oxygen (generated photochemically with inter-
mediacy of a photosensitizer) reacts with phosphines (1) in solu-
tions at room temperature. In such reactions, the corresponding
phosphine oxides (2) are generated as dominating products,
whereas phosphinates (3) are produced only in very small amounts
by Aldrich were used in this study. Matrix gases Ar and O
2
2
large excess of the matrix gas (Ar or O ), onto a cold CsI window
(10 K) directly attached to the cold tip of a continuous flow liquid
helium cryostat. The compounds were sublimated from a minia-
ture glass oven by resistive heating. The FTIR spectra were re-
ꢀ1
ꢀ1
corded in the 4000–400 cm
range, with 0.5 cm
resolution,
[
1–4].
The aim of the present work is an investigation of photochem-
using a Thermo Nicolet Nexus 670 FTIR spectrometer, equipped
with a KBr beamsplitter.
ical behavior of triphenylphosphine (triphenylphosphane accord-
ing to the IUPAC nomenclature) embedded in low-temperature
matrices of solid oxygen at 10 K. Solid O is optically transparent
2
in a broad spectral region: from far IR to visible and UV. In spite
of this advantage, solid oxygen was only scarcely applied as a host
in matrix-isolation studies [5–12]. As far as matrix-isolation photo-
chemistry is concerned, low-temperature solid oxygen environ-
ment enables selective delivery of UV-excitation as well as
structural characterization of photogenerated products based on
The samples were irradiated through the outer quartz window
of the cryostat with the light from HBO200 high-pressure mercury
lamp. This lamp was fitted with a water filter to remove IR radia-
tion. The energy of the incident UV light was controlled by cut-
off filters UG5, UG11 or WG295 (Schott). In particular, photoreac-
3
tion of Ph P was induced using WG295 long pass filter transmitting
light with k > 280 nm. In addition, the 400–680 nm range of the
visible spectrum was suppressed with a band pass filter (UG5,
Schott). Other details of the experimental setup can be found in
our recent publications [13,14].
Theoretical calculations were performed using the GAUSSIAN 03
program package [15] at the DFT(B3LYP) level of theory [16–18],
with the standard 6-31G(d,p) basis set. This basis set was adopted
from our previous study on triphenylamine [19], where it pro-
duced vibrational data quantitatively comparable with the aug-
cc-pVDZ basis set. The scaling factor of 0.975, used in the present
work for the calculated frequencies, was obtained from the least
squares linear fit (R = 0.99985).
their vibrational spectra. The excess of O
able conditions for the photooxidation reactions. Moreover, in
the low-temperature matrices of solid O , the photoproducts can
2
creates the most favor-
2
be stabilized and subsequently identified using stationary spectro-
scopic techniques.
*
Corresponding author. Address: Department of Chemistry, University of Coim-
bra, Rua Larga, 3004-535 Coimbra, Portugal. Fax: +351 239 827 703.
E-mail address: reva@qui.uc.pt (I. Reva).
0
009-2614/$ - see front matter Ó 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.cplett.2008.11.034

9
8
I. Reva et al. / Chemical Physics Letters 467 (2008) 97–100
O
O
1.2
a
b
Ph P : O + h
ν
@ 10 K
X
O2
3
2
R P
R P
+
R P(OR)
3
3
2
0
.6
O
3
1
2
3
0
1
.0
.2
Scheme 1. Reaction products that may result from oxidation of trisubstituted
phosphine (1): phosphine oxide (2) and phosphinate (3).
Ph P=O : O @ 10 K
3
2
0
.6
3
. Results and discussion
0
1
.0
.2
The Raman [20–22] and infrared [22–24] spectra of triphenyl-
3
phosphine (Ph P) in condensed phases were the subject of several
c
d
Ph P=O : Ar @ 10 K
3
studies, however, no vibrational spectra of monomeric Ph
3
P have
0
0
.6
.0
been reported hitherto.
Within the current work, the experimental FTIR spectra were re-
corded for monomers of Ph P isolated in O and Ar low-temperature
matrices. These spectra are presented in Fig. 1, where they are com-
pared with the IR spectra of Ph P monomers calculated at the
3
2
4
0
0
0
Ph P=O Calc.
3
3
2
DFT(B3LYP) level of theory, with the standard 6-31G(d,p) basis set.
The theoretical DFT(B3LYP)/6-31G(d,p) calculations were per-
formed for the optimized geometry of Ph
previous gas-phase electron diffraction experimental data [25], this
optimized structure was predicted to be of C symmetry. The simi-
larity of the vibrational signatures of Ph P isolated in O and Ar
3
P. In agreement with the
1
300 1200 1100 1000 900
800
700
600
500
Wavenumber / cm–
1
3
3
2
Fig. 2. Experimental FTIR spectrum of the photoproduct(s) obtained upon 20 min of
matrices (Fig. 1a and b) with thetheoretically calculated IR spectrum
for (Fig. 1c) is in line with a non-planar structure at the central phos-
UV-irradiation (k > 280 nm) of Ph P isolated in O2 matrix at 10 K (a) compared with
3
experimental FTIR spectra of Ph
and with the theoretically simulated spectrum of Ph
3
P@O isolated in O
2
(b) and Ar (c) matrices at 10 K
3
P@O monomer (d). The ‘O ’
3
phorus atom of Ph
The complete optimized geometry of the Ph
anearlier study [25] and can be found in theSupplementary material
Table S1). The most prominent bands correspond to the modes
3
P and the overall C
3
symmetry of the molecule.
mark designates the band due to the matrix-isolated ozone photoproduct. The ‘X’
mark (in frame a) designates the absorption due to a photoproduct other than
Ph P@O or O .
3 3
3
P monomer agrees with
(
ꢀ
1
involving PꢀC stretching (1092.4 cm ); out-of-plane CH bending
ꢀ1
ꢀ1
(
746.5 cm ); out-of-plane ring bending (696.4 and 499.7 cm ).
In order to investigate the photochemical behavior of Ph P iso-
lated in solid O , the matrix was subjected to series of UV-irradia-
gests that the main photoproduct, generated upon UV
3
(k > 280 nm) irradiation of Ph P isolated in an O matrix, can be
3
2
2
3
reliably identified as Ph P@O.
tions, where the wavelength of the incident light was controlled by
cut-off filters. Upon exposure to UV light with k > 280 nm, the IR
spectrum of the starting compound disappeared completely within
several minutes. The new IR spectrum of the photogenerated prod-
ucts was recorded after the UV-irradiation. This spectrum is pre-
sented in Fig. 2a.
In separate experiments, IR spectra of monomeric triphenyl-
3 2
phosphine oxide Ph P@O isolated in O and Ar matrices were re-
corded. These spectra are presented in Fig. 2b and c, respectively.
A very good agreement between spectra 2a and 2b strongly sug-
The experimental IR spectra of Ph P@O are well reproduced by
the results of theoretical simulations performed at the DFT(B3LYP)/
6-31G(d,p) level (Fig. 2d). These calculations were carried out for
the optimized geometry of the compound (see Table S2 in the Sup-
plementary material). Using the calculated forms of normal modes,
3
the strong band found in the spectrum of Ph P@O at 1219/
3
ꢀ
1
1216 cm was assigned to the stretching vibration of the P@O
group. This band does not have its analogue in the spectrum of
Ph P. Other strong bands found in the spectrum of Ph P@O at posi-
3
3
tions clearly different from any IR absorptions of Ph P were ob-
3
ꢀ1
served at 1123.7/1121.2 and at 544.8/542.3 cm
Comparison of the spectra presented in Fig. 2 as traces (a) and
(b) shows also that Ph P@O is not the only photoproduct generated
upon exposure of Ph P isolated in an O matrix to UV (k > 280 nm)
light. Alongside the bands belonging to the spectrum of Ph P@O, a
new band around 1036 cm was found in the spectrum recorded
after UV-irradiation of Ph P monomers. This band is characteristic
of ozone (O ) isolated in an O matrix [26].
Hence, the main UV-induced process observed for Ph
in an O matrix can be summarized as follows:
.
0
0
.4
.2
3
a
b
c
Ph P : O @ 10 K
3
2
3
2
3
ꢀ1
0
0
.0
.4
3
3
2
Ph P : Ar @ 10 K
3
3
P isolated
0
0
.2
.0
2
h
v
Ph
3
P þ 2O
2
! Ph
3
P@O þ O
3
There are several reasons to believe that the driving force of the
Ph P Calc.
3
8
4
0
studied photoreaction is the promotion (by absorption of UV pho-
ton) of Ph P to its excited state and not a direct excitation of O ma-
3
2
trix host. It has been demonstrated within the present work, as
well as in the previous work of Schriver-Mazzuoli et al. [27], that
exposure of neat O matrices to UV radiation of a lamp fitted with
2
a cut-off filter transmitting light with k > 280 nm (or with a mono-
chromatic laser light k = 266 nm [27]) does not result in generation
of ozone. Direct excitation of solid oxygen and generation of ozone
requires irradiation with UV light of shorter wavelengths.
1
300 1200 1100 1000 900
800
700
600
500
Wavenumber / cm–
1
Fig. 1. Experimental FTIR spectra of Ph
3
P isolated in O
2
(a) and Ar (b) matrices at
1
3
0 K compared with the spectrum of Ph P monomer simulated at the DFT(B3LYP)/
6
-31G(d,p) level (c).

I. Reva et al. / Chemical Physics Letters 467 (2008) 97–100
99
However, the UV (k > 280 nm) light, used in the current exper-
iments, can be effectively absorbed by the molecules of Ph
trapped in O . The UV absorption spectrum of this species (as it
was shown for Ph P dissolved in cyclohexane) exhibits a broad
band that has maximum around 260 nm and extends approxi-
mately to 300 nm [28]. Hence, the energy of the UV (k > 280 nm)
light must be a very good match to promote direct excitation of
3 2
The three-membered ring peroxides with the form R PO ,
3
P
where the R group represents alkyl or aryl substituents, were re-
ported to act as intermediates in reactions of singlet oxygen with
organic phosphines in solutions and in the gaseous phase at room
temperature and higher temperatures [1–4,30–32]. There has been
much recent interest to such ‘exotic peroxides’ – to quote from the
title of a recent review by Sawwan and Greer [33]. Under the pres-
ent matrix-isolation conditions, at low temperatures, no absorp-
tion bands ascribable to the peroxide intermediate could be
observed experimentally.
2
3
Ph
be related with initial excitation of Ph
consumption of Ph P, the subsequent UV (k > 280 nm) irradiations
did not result in any increase of the O absorption band.
careful inspection of the spectrum recorded after UV
k > 280 nm) irradiation of Ph P isolated in an O matrix reveals
the presence of a new absorption around 900 cm (marked by
X’ in Fig. 2a). Absorption in this spectral region is characteristic
3
P. Generation of ozone in the current experiments must also
3
P. Indeed, upon the total
3
3
The present matrix-isolation study is the first experimental evi-
A
dence of phosphinate photoproduct formed from Ph
without addition of any photosensitizers. The experiments carried
out in the current work demonstrated that excitation of Ph P with
UV (k > 280 nm) light induces the oxygenation of this compound.
No direct excitation of O molecules of the matrix host (that would
3 2
P and O
(
3
2
ꢀ1
3
‘
of the PꢀOꢀC moiety. For matrix-isolated trimethyl phosphate, a
2
strong infrared absorption due to the PꢀOꢀC stretching vibration
require irradiation with the more energetic UV light k ꢁ 248 nm
ꢀ1
was observed at 860 cm [29]. In the present case, appearance
of a similar absorption indicates occurrence of an additional photo-
reaction, where an oxygen atom is inserted in a PꢀC bond to pro-
[27]) was necessary to promote this photooxidation.
3
The pattern of photochemical reactions of Ph P with molecular
oxygen, investigated in the current work, shows similarities with
the pattern of thermal oxidations of phosphines. These latter pro-
cesses were investigated already as early as in XIX century [34–37].
Elucidation of the complicated mechanisms of these reactions is
important for many practical applications. For example, substi-
tuted arylphosphines are being evaluated as potential additives
for enhancing the stability of future jet fuels toward thermal oxida-
duce diphenyl-phenyl-phosphinate, Ph
2
(PhꢀOꢀ)P@O (or simply
‘
phosphinate’). In order to assess the efficiency of such photochem-
ical reaction, several theoretical spectra were simulated, where the
phosphinate and Ph P@O photoproducts contributed with varying
3
percentage. These simulated spectra (shown in Fig. 3b) are com-
pared with the experimental spectrum (Fig. 3a). This comparison
leaves no doubt that (in the reported experiment) ca. 10% of ma-
3
tion [32,38]. Ph P was shown to improve the thermal oxidative sta-
3
trix-isolated Ph P was converted to the phosphinate photoproduct.
bility of jet fuel by at least 50% [39].
The new bands, which arise (upon UV k > 280 nm irradiation) at
1
5
258, 1205, 1168, 1132, 918, 900, 774, 740, 731, 588, 536,
23 cm (marked by crosses in Fig. 3a), have their counterparts
4
. Conclusions
ꢀ1
in the calculated spectrum of the phosphinate photoproduct (see
Fig. 3b). The complete optimized geometry of the Ph
monomer can be found in the Supplementary material (Table S3).
As an additional test for the reaction mechanism, monomers of
New vibrational data on monomeric Ph
in Ar and O matrices are reported. The O
2 2
3
P and Ph
matrices can be consid-
ered inert in absence of UV-irradiation. The photochemical changes
could be triggered by the excitation of Ph P with UV (k > 280 nm)
3
P@O isolated
2
(PhꢀOꢀ)P@O
3
3
Ph P@O deposited at 10 K with a large excess of oxygen (to form an
light. Under such conditions, the direct excitation of molecular
oxygen and formation of ozone is not possible. The main UV-in-
O
2
matrix) were exposed to prolonged UV (k > 280 nm) irradiation.
No bands ascribable to the phosphinate photoproduct appeared
upon such experimental conditions. This proves that phosphinate
duced process, observed upon irradiation of Ph
matrix, concerns generation of Ph P@O and O
the competing minor photoreaction, where both atoms of O
ecule are incorporated into the photoproduct structure [4], diphe-
nyl-phenylphosphinate Ph
(PhꢀOꢀ)P@O is formed. In the present
study, both Ph P@O and Ph
(PhꢀOꢀ)P@O products were photo-
3
P isolated in an
products. In
mol-
O
2
3
3
is not a product of a secondary, UV-induced reaction of Ph
with oxygen, but that it is generated as a result of the branched
character of the photochemistry of Ph P.
3
P@O
2
3
2
3
2
chemically generated and trapped in a low-temperature oxygen
matrix. The recorded IR spectral characteristics of these species al-
lowed their positive identification. The isolation of the photochem-
ical targets in low-temperature oxygen matrices proved to be very
instrumental for studies of UV-induced oxygenation reactions,
where the photochemical products can be preserved at low tem-
peratures and studied spectroscopically.
0
0
0
.2
.1
a
b
O3
x
x
x x
x
x
x
x
x
x
x
x
.0
6
3
0
Appendix A. Supplementary material
Supplementary data associated with this Letter can be found, in
the online version, at doi:10.1016/j.cplett.2008.11.034. These data
contain the complete optimized geometries of the Ph3P, Ph3P@O,
and Ph2(PhꢀOꢀ)P@O monomers (Tables S1, S2, S3, respectively).
1
300 1200 1100 1000 900
800
700
600
500
Wavenumber / cm–
1
Fig. 3. (a, foreground layer): FTIR spectrum of Ph
3
P@O isolated in an O
2
matrix (the
area under the spectrum trace is filled white); (a, background layer): the FTIR
spectrum of photoproducts generated upon UV(k > 280 nm) irradiation of Ph
isolated in an O matrix (the area under the spectrum trace is filled grey). Crosses
designate absorption bands due to the minor photoproduct; (b, foreground layer):
IR spectrum calculated for Ph P@O (the area under the spectrum trace is filled
white); (b, background layer): sum of the IR spectra calculated for Ph P@O and for
Ph (Ph–O)P@O; the spectra were added with weights 9 and 1, respectively (the area
3
P
References
2
[
[
1] D. Zhang, B. Ye, D.G. Ho, R.M. Gao, M. Selke, Tetrahedron 62 (2006) 10729.
2] D.G. Ho, R.M. Gao, J. Celaje, H.Y. Chung, M. Selke, Science 302 (2003) 259.
3
3
[3] R.M. Gao, D.G. Ho, T. Dong, D. Khuu, N. Franco, O. Sezer, M. Selke, Org. Lett. 3
(2001) 3719.
2
under the spectrum trace is filled grey). The theoretical spectra were calculated at
the DFT(B3LYP)/6-31G(d,p) level. The theoretical frequencies were scaled by a
uniform factor of 0.975.
[4] S. Tsuji, M. Kondo, K. Ishiguro, Y. Sawaki, J. Org. Chem. 58 (1993) 5055.
[5] T.L. Tso, E.K.C. Lee, J. Phys. Chem. 88 (1984) 5465.
[6] T.L. Tso, E.K.C. Lee, J. Phys. Chem. 88 (1984) 5475.

1
00
I. Reva et al. / Chemical Physics Letters 467 (2008) 97–100
[
[
[
7] T. Akasaka, A. Yabe, W. Ando, J. Am. Chem. Soc. 109 (1987) 8085.
8] S. Hashimoto, H. Akimoto, J. Phys. Chem. 91 (1987) 1347.
9] S. Hashimoto, H. Akimoto, J. Phys. Chem. 93 (1989) 571.
[25] V.A. Naumov, M.A. Tafipol’skii, A.V. Naumov, D.Y. Shorokhov, S. Samdal, Russ. J.
Gen. Chem. 71 (2001) 1225.
[26] L. Schriver-Mazzuoli, Phys. Chem. Earth Part C–Solar-Terrest. Planet. Sci. 26
(2001) 495.
[
[
10] R.J.H. Clark, J.R. Dann, J. Phys. Chem. 100 (1996) 9271.
11] M. Bahou, L. Schriver-Mazzuoli, C. Camy-Peyret, A. Schriver, Chem. Phys. Lett.
[27] L. Schriver-Mazzuoli, A. de Saxcé, C. Lugez, C. Camy-Peyret, A. Schriver, J.
Chem. Phys. 102 (1995) 690.
2
73 (1997) 31.
[
12] N. Miyauchi, H. Hidaka, T. Chigai, A. Nagaoka, N. Watanabe, A. Kouchi, Chem.
Phys. Lett. 456 (2008) 27.
[28] P. Changenet, P. Plaza, M.M. Martin, Y.H. Meyer, W. Rettig, Chem. Phys. 221
(1997) 311.
[
[
[
[
[
[
[
13] I. Reva, M.J. Nowak, L. Lapinski, R. Fausto, Chem. Phys. Lett. 452 (2008) 20.
14] I.D. Reva, M.J. Nowak, L. Lapinski, R. Fausto, Chem. Phys. Lett. 429 (2006) 382.
15] M.J. Frisch et al., GAUSSIAN 03. Gaussian, Inc., Wallingford, CT, 2004.
16] A.D. Becke, Phys. Rev. A 38 (1988) 3098.
17] C.T. Lee, W.T. Yang, R.G. Parr, Phys. Rev. B 37 (1988) 785.
18] S.H. Vosko, L. Wilk, M. Nusair, Can. J. Phys. 58 (1980) 1200.
19] I. Reva, L. Lapinski, N. Chattopadhyay, R. Fausto, Phys. Chem. Chem. Phys. 5
[29] I. Reva, A. Simão, R. Fausto, Chem. Phys. Lett. 406 (2005) 126.
[30] K. Nahm, Y. Li, J.D. Evanseck, K.N. Houk, C.S. Foote, J. Am. Chem. Soc. 115 (1993)
4879.
[31] B.D. Beaver, L. Gao, M.G. Fedak, M.M. Coleman, M. Sobkowiak, Energ. Fuel 16
(2002) 1134.
[32] B.D. Beaver, C. Burgess-Clifford, M.G. Fedak, L. Gao, P.S. Iyer, M. Sobkowiak,
Energ. Fuel 20 (2006) 1639.
(
2003) 3844.
[33] N. Sawwan, A. Greer, Chem. Rev. 107 (2007) 3247.
[34] A. Cahours, A.W. Hofmann, Ann. Chem. Pharm. 104 (1857) 1.
[35] J.H. van’t Hoff, Z. Phys. Chem. Stöch. Verwand. 16 (1895) 411.
[36] W.P. Jorissen, Z. Phys. Chem. 22 (1897) 34.
[
[
20] E. Steger, K. Stopperka, Chem. Ber.-Recl. 94 (1961) 3023.
21] G. Bouquet, M. Bigorgne, Spectrochim. Acta Part A–Mol. Spectros. A 23 (1967)
1
231.
[
[
[
22] J. Goubeau, G. Wenzel, Z. Phys. Chem. Neue Fol. 45 (1965) 31.
23] K.A. Jensen, P.H. Nielsen, Acta Chem. Scand. 17 (1963) 1875.
24] G.B. Deacon, J.H.S. Green, Spectrochim. Acta Part A–Mol. Spectrosc. A 24 (1968)
[37] C. Engler, J. Weissberg, Ber. Deut. Chem. Ges. 31 (1898) 3055.
[38] B. Beaver, D. Rawlings, P. Neta, Z.B. Alfassi, T.N. Das, Heteroatom Chem. 9
(1998) 133.
8
45.
[39] Z.B. Alfassi, P. Neta, B. Beaver, J. Phys. Chem. A 101 (1997) 2153.