Infrared spectroscopic observation of the radical XeF3 generated in solid argon

Article abstract of DOI:10.1021/ic9007312

Xenon trifluoride radicals were generated by the solid-state chemical reaction of mobile fluorine atoms with XeF₂ molecules isolated in a solid argon matrix. On the basis of spectroscopic and kinetic FTIR measurements and performed quantum chem

Full text of DOI:10.1021/ic9007312

Inorg. Chem. 2009, 48, 8723–8728 8723
DOI: 10.1021/ic9007312
Infrared spectroscopic observation of the radical •XeF3 generated in solid argon
Eugenii Ya. Misochko,*,† Alexander V. Akimov,† Vasilii A. Belov,† and Daniil A. Tyurin‡
†Institute of Problems of Chemical Physics of the Russian Academy of Sciences, 142432 Chernogolovka,
Moscow Region, Russia, and ‡Department of Chemistry, Lomonosov Moscow State University, 119899,
Moscow, Russia
Received April 15, 2009
Xenon trifluoride radicals were generated by the solid-state chemical reaction of mobile fluorine atoms with XeF2
molecules isolated in a solid argon matrix. On the basis of spectroscopic and kinetic FTIR measurements and
performed quantum chemical calculations, two infrared absorption bands at 568 (strong) and 523 (very weak) cm-1
have been assigned to asymmetric and symmetric Xe-F stretching vibrational modes of radical •XeF3, respectively.
Chemical reaction of fluorine atom with XeF2 in a solid argon cage obeys specific kinetic behavior indicating the
formation of a long-lived intermediate complex under the condition that the diffusing fluorine atom is attached to
isolated XeF2 at temperatures 20 K < T < 27 K. Subsequent thermally activated conversion in the complex is the main
source of novel xenon-containing radical species •XeF3. The rate constant and energy barrier are estimated for the
[XeF3], as Kr ∼ 7ꢀ10-5 c-1 at 27 K and E ≈ 1.2 kcal/mol, respectively.
Kr
reaction in an argon cage, [XeF -F]
s
f
2
•
Quantum chemistry calculations reveal that radical XeF3 has a planar C2v structure. DFT calculations show that
formation of the third Xe-F bond in the •XeF3 radical is exothermic, and the binding energy of the third Xe-F bond is
8-20 kcal/mol.
1. Introduction
structures in this series. Presently, only three neutral noble-
gas-containing open-shell species, namely, •XeF,7-9
HXeO•,10 and HXeCC•,11,12 were experimentally observed
and characterized at experimental and theoretical levels. In
the present study, we detected by FTIR spectroscopy the
stabilized radical •XeF3 in dilute mixtures of F2 and XeF2 in
solid argon using F2 as a photolytic precursor for fluorine
atoms. We applied the obtained infrared data as a starting
point for quantum-chemistry calculations to thus support
our assignment for IR spectra and to rationalize the stability
of chemical bonding in this radical.
The recent discovery of a new type of noble-gas-containing
hydrides, H-Rg-X (Rg = Ar, Kr, and Xe; X = electro-
negative element or group) renewed the field of rare gas
chemistry and stimulated an experimental and theoretical
search for novel noble-gas-containing molecules.1-4 Most of
these studies have been focused on closed-shell species, where
a noble gas atom displays divalent features. Meanwhile, a
series of well-known stable xenon fluorides, XeF4 and XeF6
(as well as xenon oxyfluorides5,6), demonstrates formal
oxidation states of xenon of +4 and +6 that considerably
broadens the range of possible xenon-containing structures
(Chart 1).
2. Experimental Details
The open-shell noble-gas-containing species, such as •XeF3
and •XeF5, attract particular interest as transient molecular
Dilute gas mixtures of F2/Ar and XeF2/Ar were deposited
through separate stainless steel vacuum manifolds onto the
golden surface of a flat cuprum finger held at 12 K in a high-
vacuum chamber. The composition of typical samples was
Ar/F2/XeF2=3000:2:1, though the relative concentrations of
reactants were varied from 10-3 to 2ꢀ10-4. The thickness of
the sampleswas approximately 100μm. Infrared spectra were
*To whom correspondence should be addressed. E-mail: misochko@
icp.ac.ru.
(1) Khriachtchev, L.; Pettersson, M.; Runeberg, N.; Lundell, J.; Rasanen,
:: ::
M. Nature (London) 2000, 406, 874.
(2) Gerber, R. B. Annu. Rev. Phys. Chem. 2004, 55, 55.
:: ::
(3) Pettersson, M.; Khriachtchev, L.; Lundell, J.; Rasanen, M. In
(8) Hoffman, G. J.; Colletto, M. J. Chem. Phys. 2001, 114, 2219.
(9) Misochko, E. Ya.; Akimov, A. V.; Goldshleger, I. U.; Tyurin, D. A.;
Laikov, D. N. J. Chem. Phys. 2005, 125, 4503.
Inorganic Chemistry in Focus II; Meyer, G., Naumann, D., Wesemann, L.,
Eds.; Wiley-VCH: New York, 2005; pp 15-34.
(4) Khriachtchev, L.; Rasanen, M.; Gerber, R. Acc. Chem. Res. 2009, 42,
:: ::
(10) Khriachtchev, L.; Pettersson, M.; Lundell, J.; Tanskanen, H.;
:: ::
Kiviniemi, T.; Runeb’erg, N.; Rasanen, M. J. Am. Chem. Soc. 2003, 125,
183.
(5) Schumacher, G. A.; Schrobilgen, G. J. Inorg. Chem. 1984, 23, 2923.
(6) Brock, D. S.; Bilir, V.; Mercier, H. P. A.; Schrobilgen, G. J. J. Am.
Chem. Soc. 2007, 129, 3598.
1454.
(11) Khriachtchev, L.; Tanskanen, H.; Lundell, J.; Pettersson, M.;
:: ::
Kiljunen, H.; Rasanen, M. J. Am. Chem. Soc. 2003, 125, 4696.
(12) Feldman, V. I.; Sukhov, F. F.; Orlov, A. Yu.; Tulpina, I. V.;
(7) Aquilanti, V.; Luzzatti, E.; Pirani, F.; Volpi, G. G. J. Chem. Phys.
1988, 89, 6165.
Logacheva, E. A.; Tyurin, D. A. Russ. Chem. Bull. 2005, 54, 1458.
r
2009 American Chemical Society
Published on Web 08/20/2009
pubs.acs.org/IC

8724 Inorganic Chemistry, Vol. 48, No. 18, 2009
Misochko et al.
Chart 1
recorded with a Bruker IFS-113 V FTIR spectrometer
(spectral region from 500 to 4000 cm-1 and spectral resolu-
tion 0.5 cm-1). The closed-cycle helium refrigerator (CTI
Cryogenics) was used to maintain the sample temperature in
these experiments. Fluorine atoms were generated by F2
photolysis at 337 nm with the pulse nitrogen laser. The laser
light entered the sample chamber through a fused silica
window and impinged on the sample at an incident angle of
45ꢀ. The beam was expanded to an area of about 4 cm2 at the
sample to ensure uniform irradiation. The average laser
power was varied from 1 to 5 mW/cm2. In some experiments,
weusedthesecondharmonicofaNd:YAGlaser(Continuum
model Surelite) at 532 nm at an average laser power of
20 mW/cm2 for photolysis of the reaction products.
To distinguish the chemical reactions involving photogen-
erated F atoms from those of diffusing thermal atoms,
photolysis of F2 molecules was performed at 12 K. Fluorine
atoms are able to diffuse in solid argon at temperatures above
20 K.13-15 Earlier, we used this peculiar feature of fluorine
atoms to generate various novel radical species in solid argon
matrices.16 To initiate reactions of thermally diffusing
F atoms, we performed annealing of photolyzed samples at
temperatures above 20 K. Annealing was carried out using a
step-by-step procedure. After completion of the photolysis at
12 K, the sample was annealed 2-3 min at T > 12 K. Then,
the temperature was lowered back to 12 K, and the spectrum
was recorded. This cycle was repeated 10-12 times until the
reactionswerecomplete. Inaseparateexperiment, weverified
that the 337 and 532 nm light did not induce any changes in
the infrared spectra of the Ar/XeF2 (= 3000:1) samples.
Quantumchemical calculationshavebeenperformed using
three density functionals: the Perdew-Burke-Ernzerhof
(PBE) nonempirical generalized gradient approximation,17
its hybrid extension (PBE0),18 and the popular semiempirical
B3LYP model.19 The scalar-relativistic Hamiltonian,20 the
high-quality relativistic correlation-consistent basis sets
L3(F) and L33(Xe) augmented by core polarization func-
tions,21 and a density-fitting technique22 as implemented in a
recentversionofthePrirodacode23 were used. In addition, we
have tested the scalar-relativistic20 ab initio riMP2(full)
Figure 1. Infrared spectra of the Ar/XeF2/F2 (= 3000:1:2) sample at
12 K. Trace a shows IR bands of XeF2 prior to photolysis. Trace b is the
difference spectrum showing changes in the spectrum following 337 nm
photolysis at 12 K. Traces c and d are the difference spectra showing
additionalchanges occurringduringshort-termannealingofthesample at
27 K (c) and after subsequent prolonged annealing at 27 K (d). The IR
bands labeled as B were assigned to the intermediate complex [XeF2-F],
and the IR bands labeled as C were assigned to complex [XeF2-F2], see
text below.
method24 to compare with the DFT results. Molecular geo-
metries have been fully optimized (tolerance on gradient:
10-5 au), and a very fine integration grid is used for the
DFT exchange-correlation terms (accuracy: 10-8 au per
atom). Unrestrictedandrestrictedapproacheshavebeenused
for calculations of the open- and closed-shell systems, respec-
tively.
3. Results and Discussion
(13) Feld, J.; Kunttu, H.; Apkarian, V. A. J. Chem. Phys. 1990, 93, 1009.
(14) Alimi, K.; Gerber, R. B.; Apkarian, V. A. J. Chem. Phys. 1990, 92,
3551.
(15) Misochko, E. Ya.; Akimov, A. V.; Wight, C. A. Chem. Phys. Lett.
1998, 293, 547.
(16) (a) Misochko, E. Ya.; Benderskii, V. A.; Goldschleger, A. U.;
Akimov, A. V. J. Am. Chem. Soc. 1995, 117, 11997. (b) Misochko, E. Ya.;
Akimov, A. V.; Goldschleger, A. U.; Boldyrev, A. I.; Wight, C. A. J. Am. Chem.
Soc. 1999, 121, 405. (c) Misochko, E. Ya.; Akimov, A. V.; Wight, C. A. J. Phys.
Chem. A 1999, 103, 7972. (d) Misochko, E. Ya.; Akimov, A. V.; Goldschleger, A.
U. Uspekhi Khim. 2003, 72, 262. Misochko, E. Ya.; Akimov, A. V.; Goldschleger,
A. U. Russ. Chem. Rev. 2003, 72, 233. (e) Misochko, E. Ya.; Akimov, A. V.;
Belov, V. A.; Tyurin, D. A.; Laikov, D. N. J. Chem. Phys. 2007, 127, 084301.
(17) (a) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. Rev. Lett. 1966, 77,
3865. (b) Perdew, J. P.; Ernzerhof, M.; Burke, K. J. Chem. Phys. 1996, 105, 9982.
(18) Adamo, C. J. Chem. Phys. 1999, 110, 6158.
A. Photolysis of the Samples Ar/F2/XeF2 at 12 K. The
IR spectrum of the Ar/XeF2 (= 3000:1) samples at 12 K
consists of a broad band, attributed to the isolated
molecules XeF2, with a maximum at 547 cm-1, which
corresponds to the asymmetric stretching vibrational
mode ν3 (in an argon matrix,25 ν3 =547 cm-1, and in a
gas phase,26 ν3=560.1 cm-1). Additionally, a very weak
band at 512 cm-1 is clearly seen in the spectrum that can
be ascribed to the symmetric stretching vibration of XeF2
in a matrix27 (a corresponding IR-inactive vibration
(24) (a) Feyereisen, M.; Fitzgerald, G.; Komornicki, A. Chem. Phys. Lett.
1993, 208, 359. (b) Vahtras, O; Almlof, J. E.; Feyereisen, M. W. Chem. Phys.
Lett. 1993, 213, 514.
(19) (a) Becke, A. D. Phys. Rev. A 1988, 38, 3098. (b) Lee, C.; Yang, W.;
Parr, R. G. Phys. Rev. B 1988, 37, 785. (c) Becke, A. D. J. Chem. Phys. 1993, 98,
5648.
(25) Turner, J. J.; Pimentel, G. C. In Noble-Gas Compounds; Hyman, H. H.,
Ed.; University Chicago Press: Chicago, IL, 1963; pp 101-105.
(26) (a) Agron, P. A.; Begun, G. M.; Levy, N. A.; Mason, A. A.; Jones,
S. F.; Smith, D. F. Science 1963, 139, 842. (b) Burger, H.; Ma, S. J. Mol.
Spectrosc. 1994, 164, 84.
(20) Dyall, G. J. Chem. Phys. 1994, 100, 2118.
(21) Laikov, D. N. Chem. Phys. Lett. 2005, 416, 116. Electronic basis sets
are designated by the symbol Λ. In the current literature, these basis sets are
designated by the symbol L.
(22) Laikov, D. N. Chem. Phys. Lett. 1997, 281, 151.
(23) Laikov, D. N.; Ustinjuk, Yu. A. Russ. Chem. Bull. 2005, 54, 820.
(27) Howard, W. H., Jr.; Andrews, L. J. Am. Chem. Soc. 1974, 96, 7862.

Article
Inorganic Chemistry, Vol. 48, No. 18, 2009 8725
Figure 3. IR spectrum of UV-photolyzed sample Ar/XeF2/F2
(= 3000:1:2) after exhaustive annealing at 27 K (a). (b) The difference
spectrum after irradiation with 532 nm light at 12 K. (c) The difference
spectrum after subsequent reactions in the dark at 12 K. All spectra were
recorded at 12 K.
Figure 2. Consumption of the XeF2 bandat547 cm-1 (O) and growthof
the product bandsduringannealingofthe photolyzed sample Ar/XeF2/F2
(= 3000:1:2): (0) FO2 at 1489 cm-1, (1) product B at 554 cm-1, (9)
from ν3(XeF2) and ν1(XeF2) for 20 and 11 cm-1, respec-
tively.
product C at 543 cm-1, (b) XeF4 at 576 cm-1, (2) product A at 568 cm-1
.
All spectra were recorded at 12 K. The last four points show changes
during annealing at 27 K.
The product bands demonstrate different kinetic beha-
vior. Growth of IR bands B, C, and that attributed to
XeF4 occurs at the initial stage of annealing treatment,
whereas IR band B disappears and growth of IR band A
dominates upon a prolonged annealing at 27 K. Addi-
tionally, a sharp IR band at 1490 cm-1 of the radical FO2
is observed in the infrared spectra of the reaction
products. This radical is formed in the reaction of F
atoms with O2 molecules, as revealed in our previous
reports.15,16 Molecular oxygen is a common impurity in
fluorine gas. Hence, O2 is always present in small con-
centrations (∼10-2 %) in samples. Since diffusing fluor-
ine atoms react with O2 molecules to form FO2 radicals,
we used this reaction as an internal standard for char-
acterizing the reaction rate of diffusing F atoms
(i.e., primary reactions of F atoms) upon annealing
cycles. Figure 2 shows changes of IR bands at annealing
cycles. Growth of the IR band at 1490 cm-1 reveals
reactions of diffusing F atoms occurring at the tempera-
ture range from 20 to 26-27 K. Consumption of XeF2
molecules agrees with the primary reaction of F atoms.
Besides this, growth of IR band B corresponds to the
primary reaction product at the initial stage. In contrast,
growth of the band, corresponding to XeF4 molecules,
which requires two F atoms, exhibits a second-order
nonlinear dependence on the concentration of primary
products. Infrared bands C demonstrate the same kinetic
behavior, justifying that product C requires two F atoms
as well. A prolonged annealing at 27 K leads to the
disappearance of IR bands B and the growth of IR bands
A at 568 and 523 cm-1, whereas the other product bands
remain the same. This kinetic observation allows us to
assume that primary product B corresponds to intermedi-
ate complex XeF2-F, formed at the diffusion of fluorine
atoms. Then, this complex converts to stable radical XeF3
with a low rate at 27 K. The rate constant of the reaction
mode of XeF2 in a gas phase:28 ν1=516.5 cm-1). The IR
spectrum of the Ar/F2/XeF2 (= 3000:2:1) samples at 12 K
is shown in Figure 1a. It contains the same IR bands of
isolated XeF2. In addition, a red-shifted, weak broad
band at 541 cm-1 appears in the spectrum. The intensity
of this band increases with the growth of the F2 concen-
tration in the samples. Respectively, we assigned this
band to the binary complexes XeF2-F2 formed upon
the sample preparation.
UV-photolysis of the Ar/F2/XeF2 sample at 12 K leads
to the disappearance of the IR band, attributed to the
complexes XeF2-F2. Simultaneously, IR bands at
554 cm-1 (labeled with symbol B) and 576 cm-1 appear
in the spectrum (see the difference spectrum in Figure 1b).
The latter may be assigned to the IR-active asymmetric
fundamental vibration ν6(eu) of the XeF4 molecule (in a
gas phase:29 ν6 =586 cm-1). Molecules of XeF4 can be
generated in the course of a cage reaction by photolysis of
the reactant complexes [XeF2-F2]:
λ ¼337nm
½XeF2-F2ꢁ
½XeF ꢁ
ð1Þ
f
4
B. Annealing of the Photolyzed Samples. Annealing of
the photolyzed samples leads to a decrease of the IR band
at 547 cm-1 of isolated XeF2 molecules. Up to 20-25% of
the reactant molecules XeF2 disappear in dark reactions
upon annealing; see the difference spectra in Figure 1c
and d. Simultaneously, new IR bands appear in the
spectrum at 568 (labeled as A), 554 (B), 543 (C), 523 (A,
very weak), 517 (B, very weak), 506 (C, weak), and
576 cm-1 (XeF4). Note that IR bands labeled as B and
C occur very close to fundamental absorptions of XeF2,
whereas IR bands A at 568 and 523 cm-1 are blue-shifted
(28) Brassington, N. J.; Edwards, H. G. M. J. Mol. Struct. 1987, 162, 69.
(29) Claasen, H. H.; Chernick, C. L.; Malm, J. G. J. Am. Chem. Soc. 1963,
85, 1927.
Kr
s
f
½XeF -Fꢁ
½XeF ꢁ
ð2Þ
2
3

8726 Inorganic Chemistry, Vol. 48, No. 18, 2009
Misochko et al.
Table 1. Molecular Parameters of the Series of Xenon Fluoride Compounds
•XeF (C¥v
)
XeF2 (D¥h
)
•XeF3 (C2v
)
XeF4 (D4h)
exptla
0.231
calcdb
exptl
calcdb
calcd,b r1 = r2, r3
exptl
calcdb
R(Xe-F), nm
0.234
0.231
0.237
0.213
0.200c
0.198e
0.201
0.197
0.200
0.197
0.195, 0.226
0.199, 0.223
0.197, 0.229
0.194, 0.203
178.6/176.6
178.0/178.6
7.5/20.0/9.0/-3.9
0.192d
0.195f
0.196
0.193
0.196
0.193
angle F-Xe-F, deg
Xe-F bond dissociation energy, kcal/mol
3.4
a Data from ref 7. b This work: PBE0/PBE/B3LYP/MP2. c Data from ref 26. d Data from ref 31. e Data from ref 30. f Data from ref 32.
Table 2. Frequencies (cm-1) of Xe-F Stretch Vibration Modes in the Series of Xenon Fluoride Compounds
XeF2
XeF3
XeF4
ν3 asym. stretch
ν1 sym. stretch
ν5 asym. stretch
ν1 sym. stretch
ν6 asym. stretch
calculationsa PBE0/PBE
B3LYP/MP2
gas phase
577.9/544.2
555.0/590.0
560.1b
541.1/489.3
515.3/541.4
516.5c
599.3/560.9
577.4/617.6
579/580d
553.0/508.2
530.2/562.8
534/526d
602.3/570.2
582.7/618.3
586e
581/587
568a
-11/-12/-13/-19
533/535
523a
-11/-3/-10/-12
argon matrix
matrix shift
547a, f
-13.1
512a,g
-4.5
578
-8
a This work. b Datafrom ref 26. c Data from ref 28. d Predicted from calculatedvalues and scaling factors, see text. e Datafrom ref 29. f Data from ref 25.
g Data from ref 27.
is estimated as Kr=1/τ ∼ 7ꢀ10-5 c-1 at 27 K, were τ is
characteristic time of the dark reaction. Annealing of the
photolyzed samples at different temperatures in the re-
gion of 26-30 K provides an estimate for the energy
barrier of reaction 2 in an argon cage: E ≈ 1.2 kcal/mol.
On the basis of the kinetic data, the reaction product (C)
can be assigned to complex [XeF2-F2] resulting from the
reaction of diffusing fluorine atoms with the primary
product B:
green-light photolysis, whereas a formed intermediate
product with IR bands at 545 and 504 cm-1 is highly
unstable, even at 12 K. If irradiation by green light is
performed at 18 K, then the intermediate complex
decays in the dark with a characteristic time of about
3-5 min.
In the next experiment, the same sample was exposed to
the irradiation with light at 337 nm. Photolysis leads to
the consumption of bands (C) at 543 and 506 cm-1 and
predominant growth of the IR band at 578 cm-1 assigned
to the XeF4 molecule, as well as to growth of IR bands (A)
at 568 and 523 cm-1, which we have assigned to radical
XeF3. In the frame of our assignments of IR product
bands, UV photolysis of the secondary product
(complexes C) creates two reaction channels:
½XeF -Fꢁ þ F K
s
½XeF ꢂF ꢁ
ð3Þ
F
f
2
2
2
Hence, the IR bands (C) occur very close to the IR bands
of reactant complex [XeF2-F2] formed during sample
preparation. The secondary reaction of diffusing F atoms
is the main source of XeF4 molecules, the formation of
which being clearly observed in spectra of the products.
According to the proposed scheme, XeF4 is formed in the
course of the reaction of mobile fluorine atoms with
XeF3:
½XeF ꢁ þ F K
s
½XeF ꢁ
ð4Þ
F
f
3
4
where [...] means a solid argon cage and Fstab stands for a
stabilized fluorine atom outside of a parent cage.
C. Photolysis of the Reaction Products with Laser Light
at 532 and 337 nm. To test the photostability of the
reaction products in visible and UV light, a sample
containing the reaction products was subjected to irra-
diation by 532 and 377 nm laser light. Figure 3 shows
changes in the spectrum after photolysis with 532 nm
green light at 12 K. Photolysis results in the disappear-
D. Quantum-Chemical Calculations of Radical XeF3.
We carried out quantum-chemical calculations to sup-
port our assignments for the infrared spectrum of radical
XeF3. We have calculated four molecular structures of
•XeF, XeF2, •XeF3, and XeF4 to compare our calculation
results with available experimental data. Table 1 provides
the obtained molecular parameters for these species. It
shows that the calculated distances Xe-F in XeF2 and
XeF4 agree well with the experimental data. Practically,
the same distance is calculated for symmetric fragment
F-Xe-F in the •XeF3 radical. Calculations predict that
radical •XeF3 corresponds to the C2v point group with a
F-Xe-F angle of 176-178ꢀ. Calculations show a much
ance of two product bands (A) at 568 and 523 cm-1
.
Simultaneously, growth of the bands at 545 and
504 cm-1 takes place. These bands disappear in the
subsequent dark reaction with a characteristic time of
about 100 min at 12 K, and IR bands at 568 and 523
cm-1 recover up to 80-90% of the starting intensities
(i.e., before the green light irradiation). So, this experi-
ment demonstrates decomposition of XeF3 by means of

Article
Inorganic Chemistry, Vol. 48, No. 18, 2009 8727
Table 3. Calculated vibrational frequencies of the radical XeF3
calculated harmonic frequencies, cm-1
vib. sym.
no
approximate type of mode
PBE0
PBE
B3LYP
MP2
a1
ν1
ν2
ν3
ν4
ν5
ν6
sym XeF2 stretch
XeF(3) stretch
XeF2 in-plane bend
out-of-plane bend
asym. XeF2 stretch
FXeF(3) bend (in-plane)
555.3
254.9
195.8
220.2
599.3
79.3
506.2
308.3
169.2
202.5
560.9
68.4
530.2
250.3
189.7
212.3
577.4
80.13
562.8
424.6
195.6
224.9
617.6
101.8
b1
b2
attributed to the symmetric stretch vibration ν1 in a solid
matrix. The other calculated fundamental vibrations of
XeF3 are presented in Table 3. The calculated harmonic
vibrational frequencies differ by only 1-5% from those
experimentally observed in the gas phase for ν1(XeF2),
ν3(XeF2), and ν6(XeF4). Better agreement between theore-
tical and experimental IR data for closed-shell species is
obtained by using the B3LYP functional. Calculations by
this method overestimate the corresponding frequencies in
the gas phase by factors of 1.009, 1.002, and 1.006, respec-
tively (the average scaling factor k = 1.006). The other
methods yield the average scaling factors k(PBE0) =
0.966, k(PBE)=1.034, and k(MP2)=0.95. We have used
these scaling factors to estimate the gas-phase values and
matrix shifts for vibrational frequencies ν1 and ν5 in the
radical •XeF3, see Table 2. IR-active modes are red-shifted
in a matrix: Δν3=-12 cm-1 in XeF2, Δν1=(-11 to -19)
cm-1 in •XeF3, and Δν6=-8 cm-1 in XeF4. The vibration
mode ν1 has a small matrix shift: Δν1=-4.5 cm-1 in XeF2
and Δν1=(-3 to +10) cm-1 in •XeF3.
In our measurements, we used naturally occurred
xenon, containing various xenon isotopes. The next series
of calculations was to obtain the vibrational frequencies
•
of XeF2 and XeF3 for different isotopes: 128Xe, 129Xe,
130Xe, 131Xe, 132Xe, 134Xe, and 136Xe. Using the above
estimated scaling factors and matrix shifts for the IR-
active mode, we calculated frequencies for each isotopo-
mer in a matrix:
νðmatrixÞ ¼ νðcalcdÞ ꢀ 0:966þΔν
ð5Þ
Figure 4 shows the simulated IR bands ν3 for XeF2 and
•
ν5 for XeF3 in comparison with those experimentally
Figure 4. IR absorption of IR-active mode in solid argon matrices
recorded at 12 K with spectral resolution of 0.5 cm-1, dark lines. Dashed
lines are the simulated spectra based on PBE0-calculated vibrational
frequencies of the ν3 mode for different xenon isotopomers XeF2 and the
ν5 mode for XeF3 (vertical sticks), scaled with eq 5. The Lorenzian shape
of the lines with a line width of δ = 0.85 cm-1 was applied. The height of
vertical sticks shows the abundance of xenon isotopes in naturally
occurring xenon (in percent).
registered at 12 K. This demonstrates that the calculated
spectra reproduce broad experimental bands in accor-
dance with the natural abundance of xenon isotopes. To
the contrary, the calculated frequencies of symmetric
stretch mode ν1 are not sensitive to xenon isotopes, as
was expected for such a type of vibration. Thus, both IR
bands ν1 at 512 (XeF2) and 523 cm-1 (XeF3) are narrow in
experimental spectra (line width e2 cm-1).
•
shorter distance for Xe-F in the XeF3 radical rather
than that in •XeF.
We also calculated the strength of the third Xe-F(3)
bond in radical XeF3. The DFT calculations indicate
Calculated harmonic vibrational frequencies of XeF2,
XeF3, and XeF4 are shown in Table 2 in comparison with
the experimental data. These molecules have one IR-active
vibration only in our experimental region, where ν > 500
cm-1. Thus, we assign a strong IR band at 568 cm-1 to the
asymmetric stretching vibrational mode ν5 of •XeF3. A very
weak band at 523 cm-1 (see Figure 1d and Figure 3b) can be
•
that •XeF3 in the gas phase is stable toward dissociation of
the Xe-F(3) bond, see Table 1. Our estimation predicts
thermal stability of the radical XeF3 at cryogenic tem-
•
peratures and exothermicity of the addition of the fluor-
ine atom to the XeF2 molecule. Also, it reveals that the
strength of the Xe-F(3) bond in the •XeF3 radical should
•
be higher than that in the XeF radical. In contrast, the
(30) Reichman, S.; Schreiner, F. J. Chem. Phys. 1969, 51, 2355.
(31) Ibers, J A.; Hamilton, W. C. Science 1963, 139, 106.
(32) Burns, J. H.; Agron, P. A.; Levy, N. A. In Noble-Gas Compounds;
Hyman, H. H., Ed.; University Chicago Press: Chicago, IL, 1963; p 211.
MP2 calculation predicts endothermicity for the addition
of the F atom to the XeF2 molecule and, subsequently, a
very high reaction barrier, E > 3.9 kcal/mol.

8728 Inorganic Chemistry, Vol. 48, No. 18, 2009
Misochko et al.
4. Conclusions
provides an effective reaction barrier, which depends on the
experimental conditions. On the basis of this fact, we assume
that the energy barrier for this reaction in the gas phase should
be lower by 1 kcal/mol. We believe that these data could serve
as a starting point for more accurate and detailed quantum-
chemical calculations aimed at the understanding of the
electronic structure and chemical properties of this intermedi-
ate. Ab initio calculations and experimental studies, employ-
ing the radical •XeF3 as a precursor in the synthesis of novel
closed-shell noble-gas compounds, are now being carried out.
The main experimental result of the present study is that the
reaction F þ XeF2 in solid argon produces new radical species
•XeF3. Kinetic and spectroscopic observations in combina-
tion with calculated vibrational frequencies of a series of
xenon fluoride compounds justify this conclusion. The for-
mation of radical •XeF3 obeys specific kinetic behavior
indicating the formation of a long-lived intermediate complex
under the condition that the diffusing fluorine atom is
attached to isolated XeF2 at temperatures 20 K < T < 27
K. Subsequent thermally activated conversion (eq 2) of this
complex is the main source of new radical species. At this
stage, we cannot make a definite conclusion about the origin
of the energy barrier for this reaction. Nevertheless, the
disappearance of radical •XeF3 upon the green-light irradia-
tion and the fast recoverability in the dark at 12 K (Figure 3)
suggest that a solid-argon cage assists in the formation of
an intermediate prereaction complex and, subsequently,
Acknowledgment. This research is supported by INTAS
(Grant No. 05-1000008-8017) and the Russian Founda-
tion for Basic Research (Grant No. 07-03-00768). We
thank Dr. D. Laikov for use of his software “Priroda” and
useful advice about calculations. We thank the Russian
Academy of Sciences Joint Supercomputer Center for
computational resources.