Infrared spectra and density functional theory calculations of group 8 transition metal sulfide molecules

Article abstract of DOI:10.1021/jp900994c

Small sulfur molecules were reacted with laser-ablated Fe, Ru, and Os atoms in excess argon and condensed at 7 K. Reaction products were identified from matrix infrared spectra through sulfur-34 isotopic shifts, spectra of sulfur isotopic mixtures, and fr

Full text of DOI:10.1021/jp900994c

J. Phys. Chem. A 2009, 113, 5375–5384
5375
Infrared Spectra and Density Functional Theory Calculations of Group 8 Transition Metal
Sulfide Molecules
Binyong Liang, Xuefeng Wang, and Lester Andrews*
Department of Chemistry, UniVersity of Virginia, P.O. Box 400319, CharlottesVille, Virginia 22904-4319
ReceiVed: February 3, 2009; ReVised Manuscript ReceiVed: February 26, 2009
Small sulfur molecules were reacted with laser-ablated Fe, Ru, and Os atoms in excess argon and condensed
at 7 K. Reaction products were identified from matrix infrared spectra through sulfur-34 isotopic shifts, spectra
of sulfur isotopic mixtures, and frequencies from density functional calculations. The strongest absorptions
of the MS2 disulfide molecules are observed at 540.2, 535.5, and 537.5 cm-1, respectively, for the group 8
-
metals, and a 523.2 cm-1 band is assigned to the FeS vibrational fundamental in solid argon. The FeS2
anion was detected in the spectrum at 542.1 cm-1. The RuS2 absorption exhibited resolved natural ruthenium
isotopic splittings. Evidence is also presented for side-bound M(S2) isomers and MS4 molecules with different
structures including Fe(S2)2, (S2)RuS2, and tetrahedral OsS4. Although OsO4 is a well-known molecule, this,
we believe, is the first experimental observation of OsS4.
Introduction
a band at 542 cm-1 to the FeS fundamental in the argon matrix,20
and a recent microwave spectrum of FeS provided a basis for
estimation of the fundamental frequency (506 cm-1) from
centrifugal distortion constants.21 Theoretical calculations have
determined the 5∆ ground state for FeS depending on the method
employed.22-25 Recent DFT calculations compared cyclic Fe(S2)
and FeS2 and found bent FeS2 to be 0.18 eV lower in energy.18
To the best of our knowledge, there has been no investigation
of ruthenium and osmium sulfide molecules so the present report
will make an initial contribution.
Comparison of metal oxide and sulfide molecules formed with
the different metals in a transition metal series or group is of
chemical interest, but there have been many fewer investigations
of molecular sulfides.21,22 Recent Fe atom reactions with O2
found FeO, the bent OFeO dioxide, and side-bound dioxygen
complexes Fe(O2) as major products.26 Later reactions with
laser-ablated Ru and Os atoms provided the MO2 dioxide
molecules as major products, and the MO3 and MO4 molecules
were formed on sample annealing.27
Previous reactions of group 4 and 5 transition metals with
small sulfur molecules gave mostly monosulfide and disulfide
molecular products.28,29 Thus, we anticipate that metal sulfides
will also be produced here. The present work reports the reaction
of laser-ablated Fe, Ru, and Os atoms with small sulfur
molecules S1,2,3,4 produced through microwave discharge of
elemental sulfur and sulfur-34 vapor30 and density functional
calculations of the anticipated metal sulfide product molecules.
Transition metal sulfides have attracted considerable attention
because of their widely recognized importance in many biologi-
cal and industrial processes.1 In particular, iron sulfides have
generated the most interest, and iron pyrite, FeS2, has a textbook
structure.2,3 Iron sulfides are relevant to the origin of life on
Earth,4,5 and they are also cited as evidence for ancient life on
extraterrestrial samples.6 In biological systems, iron-sulfur
clusters are not only involved in electron transfer and organic
radical generations but also serve as immediate sulfur donors
in various processes.7-9 In industrial applications, bulk transition
metal sulfides are commonly employed as catalysts for hy-
drodesulfurization. Ruthenium sulfide clusters have been ex-
tensively studied, in part, because bulk RuS2 is among the most
active catalysts for binary metal sulfides.10-12 In addition, reports
have also shown that osmium sulfides play important roles in
volcanic studies.13
On the molecular level, group 8 transition metal sulfides have
been investigated less, and most research has focused on the
ionic iron sulfides. A Fourier tranform mass spectrometry study
+
reported the formation of FeSn (n ) 1-6) in the sequential
reactions of Fe+ with ethylene sulfide, and photodissociation
and ion-molecule reactions provided various bond energies.14
Photoelectron spectroscopy of ion-sulfur cluster anions has
been reported. In the case of FeS-, the energy differences
between three excited electronic states and the ground state of
FeS were determined,15 and a higher resolution study found a
520 ( 30 cm-1 vibrational spacing for FeS and measured
electron affinities for FeSn (n ) 1-6) species.16 Using mass-
spectrometric techniques and complementary ab initio and
density functional theory calculations, structural and thermo-
dynamic aspects of FeS+ and FeS2+ in the gas phase have been
determined.17,18 By means of two successive collision-induced
electron detachments from FeS-, the ionization energy of FeS
has also been determined.17 A recent photodissociation spec-
troscopic study of FeS+ observed the vibrational frequency for
the excited 6Π state.19 An earlier matrix isolation work assigned
Experimental and Computational Methods
The technique for investigating reactions of laser-ablated
metal atoms has been described in detail previously.31 The Nd:
YAG laser fundamental (1064 nm, 10 Hz repetition rate with
10 ns pulse width) was focused onto a rotating high-purity iron,
ruthenium (Johnson Matthey), or osmium (E-Vac Products)
metal target. The laser energy was varied from 5 to 10 mJ/
pulse. Metal atoms were codeposited with a sulfur-doped argon
stream onto a 7 K CsI cryogenic window at 2-4 mmol/h for
1 h. Infrared spectra were recorded at 0.5 cm-1 resolution on a
Nicolet 550 spectrometer with 0.1 cm-1 accuracy using a
* To whom correspondence should be addressed. E-mail: lsa@
virginia.edu.
10.1021/jp900994c CCC: $40.75  2009 American Chemical Society
Published on Web 03/30/2009

5376 J. Phys. Chem. A, Vol. 113, No. 18, 2009
Liang et al.
mercury cadmium telluride detector. Matrix samples were
annealed at different temperatures, and selected samples were
subjected to irradiation using a medium-pressure mercury lamp
(>220 nm) with the globe removed.
A microwave discharge in argon seeded with sulfur vapor
was used as a source of sulfur atoms and small molecules as
reagents. The coaxial quartz discharge tubes used here evolved
from the one described in earlier experiments.30 Natural isotopic
sulfur (Electronic Space Products, Inc., recrystallized) and
enriched sulfur (98% 34S, Cambridge Isotope Laboratories) were
used as received: different mixtures of the two isotopic samples
were also employed. The vapor pressure of sulfur feeding the
discharge was controlled by resistance heating. The microwave
discharge maintained in the argon-sulfur mixture (Opthos
Instruments 120 W microwave discharge, 30-50% of the
maximum power level) with an Evenson-Broida cavity ex-
tended from 5 cm downstream of the sulfur reservoir to the
end of the discharge tube.
Figure 1. Infrared spectra in the 600-450 cm-1 region for products
formed in the laser-ablated Fe reaction with discharged sulfur vapor
during condensation in excess argon at 7 K: (a) sample deposited for
60 min, (b) after annealing to 25 K, (c) after annealing to 35 K, (d)
after >220 nm irradiation, and (e) after annealing to 40 K.
Following previous work,28,29,31,32 DFT calculations were
performed on anticipated metal sulfide products first using the
Gaussian 98 program,33 the B3LYP or BPW91 density
functional,34,35 the 6-311+G(d) basis set for sulfur and iron, and
the LanL2DZ effective core potential and basis for Ru and
Os.36,37 Additional calculations were performed using the
Gaussian 03 program system,38 the B3LYP density functional,34
the large Gaussian basis 6-311+G(3df) for S and Fe, and the
SDD pseudopotential and basis for Ru and Os atoms.36,39
A very weak band at 542.1 cm-1 becomes 5-fold stronger
relative to the 540.2 cm-1 band when the laser energy is reduced
using a 10% neutral-density filter, as shown in Figure 3. This
band is reduced substantially on full arc irradiation while the
540.2 cm-1 band increases by 10%. The sulfur-34 counterpart
at 534.4 cm-1 defines a very low 1.0144 isotopic frequency ratio
and decreases 50% on full arc irradiation while the 531.3 cm-1
band increases 10%. Mixed isotopic spectra from an experiment
using lower laser energy gave a 1/2/1 triplet absorption with
538.4 cm-1 intermediate component for this photosensitive
species.
Ru + Sx. Ruthenium atom reactions with the sulfur vapor
mix gave spectra shown in Figure 4. The strongest features are
a single band at 556.1 and the resolved ruthenium isotopic
multiplet27,41 terminated by the Ru-102 and 104 peaks at 535.5
and 533.9 cm-1. Ultraviolet irradiation had no effect, but
annealing slightly decreased the multiplet and increased the
single band. A weaker ruthenium isotopic multiplet in the 540
cm-1 region and a 518.7 cm-1 band are also observed. These
absorptions shifted with sulfur-34 substitution, and exhibited
different sulfur isotopic frequency ratios as given in Table 2.
Isotopic spectra illustrated in Figure 5 reveal triplet mixed sulfur
isotopic spectra for the major bands.
Os + Sx. Osmium atom reaction product absorption spectra
are shown in Figure 6 for natural sulfur-32, sulfur-34, and a
mixture of the two. The band patterns are different for the two
pure isotopic spectra owing to different isotopic shifts for the
major product species (Table 2). The most straightforward is
the lower region containing 503.7 and 491.6 cm-1 bands for
the pure isotopes and a slightly broadened lower band for the
mixture. The upper region contains several overlapping mixed
isotopic triplets, which will be discussed below. Figure 7 shows
the behavior on annealing and UV irradiation in the sulfur-34
experiment where there is less band overlap: the stronger bands
increase on annealing to 30 K, but UV irradiation decreases
the upper bands in favor of the OsS4 band.
Results and Discussion
Infrared Spectra. Elemental sulfur vapor in a flowing argon
discharge gives S2 as judged from its blue emission40 and S3
and S4 based on their known infrared absorption spectra.30 Laser-
ablated group 8 metal atoms were codeposited with such an
argon stream, and new metal-dependent sulfur reaction product
spectra will be presented and assigned. Relatively low laser
energy was employed, and the concentration of metal atoms is
sufficiently low as to minimize the contribution of dimetal
species to the observed spectra: resolved natural ruthenium
isotopic splittings support this point.
Fe + Sx. Five experiments were done with iron and sulfur
vapor changing the sulfur content by a factor of 4 as judged by
the strong S3 absorption intensities and the ablation laser energy
by a factor of 10 using a neutral-density filter. There were a
large number of variables to control in these experiments.
Representative spectra are illustrated in Figure 1, where the S3
absorptions in the 600 cm-1 region (not shown) are as intense
as the S4 absorptions, and weak S3 bands are detected at 584.9,
581.7 cm-1. The major absorptions appeared at 567.4, 540.2,
523.2, 467.2, and 461.8 cm-1 as listed in Table 1. Annealing
slightly increased the 540.2 and 467.2 cm-1 and decreased the
523.2 and 461.8 cm-1 bands and ultraviolet irradiation had little
effect. Higher annealing decreased the 567.4 cm-1 peak in favor
of its 565.0 cm-1 satellite. These new product absorptions
exhibited 8 -13 cm-1 shifts on sulfur-34 substitution and 32/
34 isotopic frequency ratios appropriate for Fe-S vibrational
modes. Figure 2 compares product spectra with mixed sulfur-
32,34 and with sulfur-34 substitution. The mixed sulfur isotopic
multiplets reveal product stoichiometric information. Notice that
the 567.4 cm-1 band reveals a sextet, the 540.2 cm-1 band a
triplet, and the 523.2 cm-1 band a doublet. On higher temper-
ature annealing, the sextet sharpens and gives way to a broad
triplet, which corresponds to the 565.0 cm-1 satellite in the
original spectrum.
Calculations. Density functional calculations were done for
anticipated product molecules in different electronic states using
different basis sets and pseudopotentials for the metals. The
results of these calculations are summarized in Table 2 and
discussed with each molecule below.

Group 8 Transition Metal Sulfide Molecules
J. Phys. Chem. A, Vol. 113, No. 18, 2009 5377
TABLE 1: Infrared Absorptions (cm-1) from Codeposition
of Laser-Ablated Fe, Ru, and Os Atoms with Discharged
Sulfur Vapor in Excess Argon
32S
34S
32S + 34S
R(32/34)
identity
Iron
567.4 554.6 567.4, 564.4, 562.1,
561.2, 558.3, 554.6
1.0231
Fe(S2)2
565.0 552.2 565.2, 559.0, 552.3
542.1 534.4 542.1, 538.4, 534.4
540.2 531.3 540.2, 536.1, 531.3
526.6 515.6
523.2 513.4
518.3 508.6 518.3, 508.6
1.0232
1.0144
1.0168
1.0213
1.0191
1.0191
1.0205
1.0190
1.0216
1.0190
Fe(S2)2 isomer
FeS2
FeS2, ν3
-
FeS
FeS site
FexSy
493.7
483.8
471.1 462.3
467.2 457.2 467.2, 461.9, 457.5
461.8 453.1 457.5, 453.0
FeS2, ν1
Ruthenium
572.0 556.3
569.7 553.5
556.1 540.7 556.1, 548.4, 540.8
546.1 535.6
545.2 534.6
544.2 533.6
543.4 532.8
541.5 531.0
542.6 531.8
Figure 2. Infrared spectra in the 600-440 cm-1 region for products
formed in the laser-ablated Fe reaction with discharged sulfur-34
enriched vapor during condensation in excess argon at 7 K: (a) normal
32S isotopic sample deposited for 60 min, (b) 50/50 mixed 32,34S isotopic
sample, (c) after annealing to 30 K, (d) after annealing to 40 K, (e)
after annealing to 48 K, (f) 34S after sample deposition, (g) after
annealing to 5 K, and (h) (c) after annealing to 43 K.
1.0282
1.0293
1.0285
1.0196
1.0198
1.0199
1.0199
1.0198
1.0203
1.0202
1.0205
1.0207
1.0207
1.0208
1.0208
1.0245
Rux(S2)
Rux(S2) site
Ru(S2)
99RuS3
100RuS3
101RuS3
102RuS3
104RuS3
(S2)RuS2, site
(S2)RuS2 b2
99RuS2, ν3
100RuS2, ν3
101RuS2, ν3
102RuS2, ν3
104RuS2, ν3
(S2)RuS2 a1
540.8 530.1
537.9 527.1 537.9, -, 527.1
537.1 526.2 -, -, 526.2
536.3 525.4 -, 532.7, 525.4
535.5 524.6 -, 531.9, 524.6
533.9 523.0 -, 530.3, 523.0
518.7 506.3 518.5, 511.5, 506.8
481.5
[
[
[
[
[
[
100RuS4]
102RuS4]
104RuS4-]
100RuS2-
102RuS2-
104RuS2
490
479.9
478.5
475.5
473.9
472.5
1.021
]
]
]
Osmium
546.9 531.2 546.9, 539.1, 531.2
545.1 532.6 545.1, 543.1, 532.6
537.5 524.9
536.8 522.6 536.9, 526.5, 522.6
520.2 505.4 520.2, 513.7, 505.4
503.7 491.6 503.7, 498.5, 495.9,
493.0, 491.6
1.0296
1.0235
1.0240
1.0271
1.0293
1.0246
Os(S2)
(S2)OsS2
OsS2
(S2)OsS2
(S2)OsS2
OsS4
Figure 3. Infrared spectra in the 550-530 cm-1 region for products
formed in the low-energy laser-ablated Fe reaction with discharged
sulfur vapor during condensation in excess argon at 7 K: (a) normal
32S isotopic sample deposited for 60 min, (b) after >220 nm irradiation,
(c) 50/50 mixed 32,34S isotopic sample, (d) after >220 nm irradiation,
(e) 34S after sample deposition, and (f) after >220 nm irradiation.
492.4 479.9
1.0260
[OsS]
FeS. The 523.2 cm-1 band in iron-sulfur experiments
decreases on annealing and shifts to 513.4 cm-1 with sulfur-
34, and this 32/34 isotopic frequency ratio, 1.0191, is in excellent
agreement with the harmonic FeS diatomic isotopic frequency
ratio (1.0192). The mixed sulfur isotopic spectrum (Figure 2)
reveals a pure isotopic doublet, which shows conclusively that
this absorption involves a single sulfur atom. The microwave
and the previous 542 cm-1 band could be due to FeS2. Without
benefit of isotopic mixtures, these workers also reversed the
identifications of VS and VS2 reaction products.28 In summary,
the present 523.2 cm-1 band is assigned to the FeS vibrational
fundamental in solid argon on the basis of the foregoing
evidence, particularly sulfur-34 substitution. It is difficult to
predict the gas-phase fundamental from this single observation,
but we certainly expect it to be within 10 cm-1 of our argon
matrix value. The FeS molecule is probably produced here from
the combination of Fe and S atoms during formation of the argon
matrix, which attests to the deposition of S atoms from the
discharge source.
5
spectrum finds a ∆ ground state for FeS and predicts a 506
cm-1 fundamental from centrifugal distortion constants,21 while
the high-resolution photoelectron spectrum also determined a
5∆ ground state with a 520 ( 30 cm-1 vibrational spacing.16
Theoretical calculations predict values 544, 532, 521, and 513
cm-1 depending on the methods employed,22-25 and our
calculation gives 514 cm-1. The early matrix isolation reaction
of thermal Fe with OCS reveled 542 and 522 cm-1 product
absorptions in solid argon without isotopic substitution, and
these workers assigned the 522 cm-1 band to SO2 impurity and
the 542 cm-1 band to FeS.20 The FeS2 molecule is identified
here at 540.2 cm-1 from mixed isotopic absorptions (see below),
FeS2. The 540.2 cm-1 absorption increases slightly on
annealing and shifts to 531.3 cm-1 on sulfur-34 substitution,
and the isotopic frequency ratio so defined (1.0168) reveals less
sulfur participation in this vibrational mode than for the above
FeS diatomic molecule. This can arise from more involvement

5378 J. Phys. Chem. A, Vol. 113, No. 18, 2009
Liang et al.
characterizes the vibration of two equivalent sulfur atoms, and
the low 32/34 isotopic frequency ratio 1.0144 indicates the
increased involvement of the Fe partner as would be found with
a larger obtuse angle. Anion photodetachment indicates that
-
FeS2 is a very stable anion (FeS2 electron affinity 74.3 kcal/
mol), and B3LYP calculations have predicted a nearly linear
6A1 iron disulfide anion structure.16,18 Using a larger basis set,
our DFT calculation predicts almost the same structure, an
electron affinity of 74.5 kcal/mol, and a very strong antisym-
metric mode at 529.6 cm-1 with isotopic 32/34 ratio 1.0142. In
addition, the calculated mixed 32,34 isotopic component is 0.1
cm-1 above the median of pure isotopic values, just as is the
observed value. Hence, the weak 542.1 cm-1 band can be
assigned to the iron disulfide molecular anion made here by
electron capture during codeposition of laser-ablated iron atoms
and electrons with sulfur molecules. It is perhaps fortuitous that
DFT/B3LYP calculations work so well for the nearly linear 6A1
iron disulfide anion in view of the fact that the isoelectronic
FeO2- anion is a linear doublet ground state and a multireference
computational problem.42c Our B3LYP calculation predicts a
sextet state with no spin contamination [〈S(S + 1)〉 is 8.78 before
and 8.75 after annihilation].We invite higher level calculations
to confirm this assignment and application of the single reference
DFT method.
Figure 4. Infrared spectra in the 580-500 cm-1 region for products
formed in the laser-ablated Ru reaction with discharged sulfur vapor
during condensation in excess argon at 7 K: (a) sample deposited for
60 min, (b) after annealing to 30 K, (c) after >220 nm irradiation, (d)
after annealing to 35 K, and (e) after annealing to 40 K.
of Fe in the antisymmetric motion between two S atoms. The
mixed isotopic spectrum reveals a triplet pattern at 540.2, 536.1,
531.3 cm-1, and the stronger single intermediate component
characterizes the motion of two equivalent sulfur atoms. It is
significant that the intermediate component is higher (0.3 cm-1)
than the average of pure isotopic bands (535.8 cm-1) as this
points to interaction with the lower frequency symmetric
stretching mode of the mixed isotopic molecule, which now
are of the same symmetry. Our best calculation (6-311+(3df)
5
Fe(S2). Our best calculation finds a A1 ground state for the
side-bound Fe(S2) complex (Table 2), which is 3 kcal/mol lower
5
energy than the above B2 ground state of the open disulfide
molecule, in disagreement with earlier workers who used a
smaller basis set.18 However, this energy is within the accuracy
of our calculation, and these two states have almost the same
energy. Most importantly, this side-bound Fe(S2) complex has
its strongest computed mode at 477.7 cm-1 and this mode has
more S-S stretching character and a higher 32/34 ratio (1.0234),
which is not compatible with the observed 476.2 cm-1 band.
In addition, the mixed isotopic band in the case of the Fe(S2)
complex is expected to be higher than the median isotopic band
because the interacting partner is computed to be lower (Table
2), and such is not the case. Hence, we have no evidence for
the Fe(S2) complex in our experiment. The calculated structure
for the Fe(S2) complex reveals an S-S distance close to that in
iron pyrite (2.171 Å), but our Fe-S distance is shorter than the
solid value (2.259 Å).2 The most stable form of iron trisulfide
is SFe(S2) (Table 2), but we do not have evidence for this
molecule either.
5
basis) for FeS2 finds the B2 ground state, in agreement with
previous workers,18 and the strong b2 mode at 546.8 cm-1 with
a weaker a1 mode at 476.1 cm-1 (Table 2). The computed 32/
34 isotopic frequency ratios for the b2 and a1 modes of this
molecule of angle 113.5° are 1.0167 and 1.0219, respectively.
The agreement of the 540.2 cm-1 observed band and isotopic
5
characteristics with the computed b2 mode values for the B2
ground state is sufficient to confirm the identification of the open
iron disulfide molecule; however, there is more information.
The 467.2 cm-1 band tracks with the 540.2 cm-1 band on
annealing, but it is slightly stronger, and our calculation
predicted a weaker symmetric stretching mode for FeS2. The
467.2 cm-1 band shifts to 457.2 cm-1 (ratio 1.0216) and forms
a mixed isotopic triplet although there is overlap with compo-
nents for the 461.8 cm-1 band. Significantly, the 461.9 cm-1
intermediate component is 0.3 cm-1 lower than the average of
the pure isotopic bands precisely matching the 0.3 cm-1 higher
intermediate component in the antisymmetric stretching mode
triplet absorption. This is due to coupling of the two stretching
modes of 32SFe34S, which now have the same symmetry. The
467.2 cm-1 band and isotopic counterparts are also in excellent
agreement with computed a1 mode values for the FeS2 molecule
and add further support for its identification. Our B3LYP
calculation describes this quintet ground state amazingly well:
there is little spin contamination [〈S(S + 1)〉 is 6.18 before and
6.00 after annihilation].
Fe(S2)2. The highest new band at 567.4 cm-1 shifts to 554.6
cm-1 with sulfur-34 and defines a 1.0231 ratio, which suggests
a mixed S-S and Fe-S stretching mode. The mixed sulfur
32,34 sample gave a partially resolved sextet, which sharpened
on annealing and then gave way to a broad triplet at 565.2,
559.0, 552.3 cm-1. This triplet characterizes the 565.0 cm-1
satellite absorption as being due to a species with two almost
equivalent S2 subunits where each gives a triplet mixed isotopic
band and the two almost overlap. The sextet is expected for
two equivalent S2 subunits each with equivalent S atoms, which
has been observed with textbook 1/4/4/2/4/1 relative intensities
for the group 10 metal atom and S2 reactions.43 The most stable
iron tetrasulfide stoichiometry molecules we converged after
many attempts are the Fe(S2)2 structures detailed in Table 2.
These have intense high-frequency mixed S-S and Fe-S
stretching modes computed at 579 or 547 cm-1. The former
5B1 species has a computed mixed isotopic sextet, which fits
the 567.4 cm-1 band very well. This quintet state is not spin
contaminated [〈S(S + 1)〉 is 6.11 before and 6.00 after
FeS2-. The weak 542.1 cm-1 band increases in intensity
relative to the strong FeS2 absorption at 540.2 cm-1 with reduced
laser energy, and full arc irradiation reduces the former and
increases the latter absorptions (Figure 3), which are charac-
teristic of a molecular anion.31a,42 The photosensitive 542.1 cm-1
band shifts to 534.4 cm-1 with sulfur-34 and forms a 1/2/1 triplet
with median 538.4 cm-1 band using mixed isotopes. This triplet
3
annihilation]. The latter B1 state exhibits a calculated mixed

Group 8 Transition Metal Sulfide Molecules
J. Phys. Chem. A, Vol. 113, No. 18, 2009 5379
TABLE 2: Calculated Structural Parameters and Vibrational Frequencies (cm-1) for Group 8 Sulfide Molecules and
Complexesa
lengths (Å)
angles (deg)
freq (cm-1
(symmetry, intensities, km/mol)
)
species
state
rel energy (kcal/mol)
-
S2
S3
3Σg
SS: 1.902
0
0
715.9 (0)
1A1 SS: 1.922
SSS: 118.5
686.5(b2,134), 594.5(a1,2), 262.3(a1,1)
iron
FeS
FeS
5
∆
FeS: 2.028
0
Fe32S: 514.5(30); Fe34S: 504.8(29)
Fe32S: 528.0(39); Fe34S: 518.1(38)
Fe32S: 451.6(105); Fe34S: 443.1(99)
477.3 (a1,29), 331.8 (b2,3), 297.9(a1,2)
5Σ+ FeS: 2.004
+5
-34
0
6
FeS-
∆
FeS: 2.137
Fe(S2)
5Α1 FeS: 2.166
SS: 2.180
FeS2
5Β2 FeS: 2.013
SFeS: 113.5
+3
Fe32S2: 546.8(b2,32), 476.1(a1,13), 126.7(a1,2)
Fe32S34S: 542.6(a,31), 470.7(a, 12), 125.3(a,2)
Fe34S2: 537.8(b2,31), 465.9(a1,12), 123.9(a1,2)
Fe32S2: 471.5(b2,41), 357.9(a1,6), 49.8(a1,3)
FeS2
3Β1 FeS: 2.056
SFeS: 146.3
+20
Fe(S2)
5Β1 FeS: 2.330
SS: 2.082
22
509 (a1,6), 241 (b2,0), 237(a1,8)
FeS2
[SDD for Fe]
SFe(S2)
5Β2 FeS: 2.013
SFeS:112.9
Fe32S2: 537.5(b2,25), 465.1(a1,13), 124.3(a1,2)
Fe32S2: 592 (a1,45), 419(a1,1), 242(a1,11)
3B1 FeS: 2.091, 2.339
SS: 2.003
0
FeS3-(D3h)
1A1′ FeS: 1.937
6A1 FeS: 2.117
SFeS: 169.7
45
-72
631 (e,47 × 2), 489 (a1,0), 187 (e,7 × 2), 112 (a2′′,0)
Fe32S2: 529.6(b2,169), 389.8(a1,1), 29.7(a1,33)
Fe32S34S: 526.0(a,166), 383.9(a,1), 29.5(a,32)
FeS2
Fe34S2: 522.2(b2,162), 378.3(a1,1), 29.3(a1,31)
34
Fe(S2)2 (D2)
5B1 FeS: 2.218
578.9 (b3,61), 569(a,0), [576.4, 575.5, 571.1, 568.5, 563.1 (b3) for
S
]
1,2,2,3,4
[ SDD for Fe]
348(b2,0), 330 (b3,52)
SS: 2.0230
Fe(S2)2 (C2v)
[ SDD for Fe]
3B1 FeS: 2.179, 2.285
600.0 (a1,24),
546.7(b2, 133)
[546.1, 540.0, 546.3, 539.9, 533.2, 539.7, 533.1, 532.9 (b2) for
34
S
in order]
1,2,2,3,4
SS: 2.049, 1.9940
1A1 FeS: 1.988
FeS4 (Td)
Ruthenium
RuS
RuS
RuS2
105
517(t2,62 × 3), 462(a1,0), 242(t2,5 × 3), 193(e,0 × 2)
5
∆
RuS: 2.134
100RuS: 479.5, 102RuS: 478.3(26), 104RuS; 477.2
5Σ+ RuS: 2.145
3B1 RuS: 2.063
SRuS: 119.9
14
0
102Ru32S: 474.1(17), 102Ru34S: 463.4(16)
102Ru32S2: 550.5(b2,78),526.3(a1,9), 127.9 (a1,1)
102Ru32S34S: 546.3(a,74), 518.3(a,12), 126.5(a,1)
102Ru34S2: 539.4(b2,75),513.0(a1,9), 125.0 (a1,1)
102Ru32S2: 500.4(a1,8),495.1(b2,31), 155.4 (a1, 1)
100Ru32S2: 531.8(b2,78),504.8(a1,11), 125.4 (a1, 1)
102Ru32S2: 530.2(b2,78), 504.0(a1,11), 125.1(a1,1)
104Ru32S2: 528.6(b2,78),503.3(a1,11), 124.8 (a1,1)
100Ru32S2: 483.5(b2,139),463.2(a1,18), 106.4 (a1,2)
102Ru32S2: 481.9(b2,139),462.5(a1,18), 106.1 (a1,2)
104Ru32S2: 480.4(b2,138),461.8(a1,18), 105.9 (a1,2)
102Ru32S2: 423.8(b2,98),409.1(a1,8), 104.1 (a1,3)
RuS2
RuS2
5B2 RuS: 2.108 SRuS: 111.4
3B1 RuS: 2.089
2
0
SRuS: 117.5
-
RuS2
4B2 RuS: 2.134
SRuS: 122.1
-29
-
RuS2
6A1 RuS: 2.215
SRuS: 135.7
-16
Ru(S2)
3A2 RuS: 2.271
SS: 2.037
14
102Ru32S2: 549.7 (a1,15), 362.1(a1,2), 238.4(b2,4)
102Ru32S34S: 541.8(a,14), 358.7(a,2), 235.2(a,4)
102Ru34S2: 533.7 (a1,14), 355.3(a1,2), 232.1(b2,4)
102Ru32S2: 471.3 (a1,14), 208.0(a1,0), 186.6(b2,1)
Ru(S2)
5A1 RuS: 2.252
SS: 2.206
27
8
Ru (S2)
3A2 RuS: 2.318
SS: 2.074
102Ru32S2: 522.7 (a1,17), 341.7(a1,3), 228.6(b2,7)
RuS3 (D3h)
SRu(S2)
1A1′ RuS: 2.084
3B1 RuS: 2.127, 2.450
SS: 2.003
0
13
549 (e,50 × 2), 464 (a1,0), 160 (e,4 × 2), 73 (a2′′,0)
102Ru32S2: 586 (a1,32), 488(a1,53), 243(a1,7)
(S2)RuS2 (C2V)
nonplanar
1A1 RuS: 2.083
0
102Ru32S2: 542.8(b2,66),530.3(a1,77), 514.6 (a1,8)
102Ru34S2: 532.1 (b2,65), 516.6(a175), 500.9(a1,7)
SRuS: 116.0
RuS: 2.311
SS: 2.0820
RuS4 (Td)
(S2)RuS2 (C2V)
nonplanar
1A1 RuS: 2.126
1A1 RuS: 2.067
19
483(t2,50 × 3), 464(a1,0), 201(t2,3 × 3), 180(e,0 × 2)
102Ru32S2: 559.0(b2,69),555.2(a1,64), 532.7 (a1, 24)
RuS: 116.1
RuS: 2.274
102Ru34S2: 548.0 (b2,66), 539.9(a166), 519.4(a1,17)
SS: 2.0510
RuS4 (Td)
osmium
OsS
1A1 RuS: 2.105
19
511(t2,57 × 3), 488(a1,0), 209(t2,2 × 3), 185(e,0 × 2)
5Σ+ OsS: 2.12
0
9
Os32S: 509.3(9); Os34S: 496.3(8)
5
OsS
∆
OsS: 2.13
Os32S: 494.2(16); Os34S: 481.7(8)

5380 J. Phys. Chem. A, Vol. 113, No. 18, 2009
Liang et al.
TABLE 2: Continued
lengths (Å)
freq (cm-1
)
species
OsS2
state
3B1
angles (deg)
rel energy (kcal/mol)
0
(symmetry, intensities, km/mol)
OsS: 2.078
SRuS: 122.2
Os32S2: 547.4(a1,6), 538.8(b2,67), 138.1 (a1,1)
Os32S34S: 544.0(a,22), 528.3(a,49), 136.4(a,0)
Os34S2: 532.5(a1,6),526.1(b2,64), 134.7 (a1,1)
Os32S2: 486 (a1,12), 476(b2,99), 124 (a1, 2)
-
OsS2
4B2
3A2
OsS: 2.142
SRuS: 121.3
OsS: 2.274
-56
Os(S2)
Os(S2)
44
Os32S2: 546.0 (a1,15), 360.8(a1,0), 226.9(b2,2) Os32S34S: 538.1(a,15),
356.8(a,0), 223.7(a,2) s34S2: 530.0 (a1,15), 352.6(a1,0), 220.6(b2,2)
SS: 2.047
OsS: 2.210
SS: 2.007
OsS: 2.128
OsS: 2.312, 2.097
1A1
48
Os32S2: 590.7 (a1,9), 389.6(a1,0), 223.9(b2,)
OsS4 (Td)
(S2)OsS2 (C2V)
nonplanar
1A1
1A1
0
3
512.5(a1,0), 500.5(t2,58 × 3), 194.0(t2,2 × 3), 186.2(e,0 × 2)
(S2)Os32S2: 545.3(b1,61), 537.3(a1,38), 523.0 (a1, 36), 337.3(a1,3),
281.7(b1,3), 162.8(a1,0)
SS: 2.108
SOsS: 129.8
(32S2)Os32S34S: 542.6(a,61), 526.4(a,38), 522.9(a,36)
(32S34S)Os32S2: 545.3(b1,61), 537.3(a1,38), 515.4 (a1,36)
a Calculations used B3LYP/6-311+G(3df) for Fe and B3LYP/6-311+G(3df)/SDD for Ru and Os except where the formulas are in italics and
B3LYP/6-311+G(d)/LANL was used.
Figure 5. Infrared spectra in the 590-470 cm-1 region for products
formed in the laser-ablated Ru reaction with discharged sulfur vapor
during condensation in excess argon at 7 K: (a) normal 32S sample
Figure 6. Infrared spectra in the 560-480 cm-1 region for products
formed in the laser-ablated Os reaction with discharged sulfur vapor
during condensation in excess argon at 7 K: (a) normal 32S sample
deposited for 60 min and (b) after annealing to 40 K, (c) 34S sample
deposited for 60 min and (b) after annealing to 30 K, (c) 34S sample
after deposition and (d) after annealing to 30 K, (e) 50/50 mixed 32,34
isotopic sample after deposition, and (f) after annealing to 30 K.
S
after deposition and (d) after annealing to 40 K, (e) 50/50 mixed 32,34
isotopic sample after deposition, and (f) after annealing to 40 K.
S
isotopic triplet, which is appropriate for the 565.0 cm-1 satellite
absorption, but this calculation is an approximation as spin
contamination is evident [〈S(S + 1)〉 is 3.07 before and 2.24
after annihilation].
of quintets with some overlap (Figure 5, most intense Ru-102
positions marked). The ruthenium-102 mixed sulfur 32, 34 band
at 531.9 cm-1 is 1.8 cm-1 higher than the median of pure sulfur
isotopic bands, which demonstrates interaction with a relatively
close lower frequency stretching mode now allowed in the lower
symmetry mixed isotopic molecule.
Other Absorptions. Several other absorptions in Table 1
cannot be identified from the information available. The weak
526.6 cm-1 absorption appears to be favored at lower laser
energy, and the mixed isotopic spectrum reveals an intermediate
component, so two sulfur atoms are likely. The 461.8 cm-1 band
is favored at higher laser energy, but it decreases on annealing.
Overlapping of bands in the mixed isotopic experiments make
the identification of intermediate components less than straight-
forward, but it appears that a mixed isotopic counterpart falls
under the 457.2 cm-1 Fe34S2 band. We considered cyclic Fe2S2
as a possibility because of the unique (diatomic FeS) sulfur 32/
34 isotopic frequency ratio, but our DFT calculation finds a 9A
Our B3LYP/6-311+G(3df)/SDD calculation for RuS2 predicts
a
3B1 ground state with intense 550.5 cm-1 antisymmetric
stretching frequency and 1.0206 sulfur 32/34 ratio while the
B3LYP/6-311+G(d)/LANL calculation finds a 530.2 cm-1
frequency with 1.0208 sulfur 32/34 ratio and 3.2 cm-1 Ru
100-104 separation (3.2 cm-1 observed). Although the 5B2 state
was observed for FeS2 and it is only 0-2 kcal/mol higher than
3
state with highest observable frequency at 426 cm-1
.
the B1 state for RuS2, depending on the basis set, the strong
RuS2. The resolved ruthenium natural isotopic quintet from
533.9 to 537.9 cm-1 (Figure 4, Table 1) with natural abundance
isotopic intensities is characteristic of the stretching mode of a
single Ru atom.27,41 The quintet shifts almost 11 cm-1 upon
sulfur-34 substitution, and the sulfur 32/34 isotopic ratios range
from 1.0208 to 1.0205 as the Ru mass decreases from 104 to
99 and the lighter metal atom contributes more to the vibrational
reduced mass. The mixed sulfur 32,34 reaction gave a triplet
computed frequency for the 5B2 state, 495 cm-1, is much lower
than the observed value. Hence, we assign the ruthenium natural
3
isotopic quintet from 533.9 to 537.9 cm-1 to the B1 ground
state of RuS2 in the argon matrix. The much weaker symmetric
stretching mode, predicted in the low 500 cm-1 region, is not
observed. The B3LYP calculation describes this triplet ground
state with no spin contamination [〈S(S + 1)〉 are 2.01 before
and 2.00 after annihilation].

Group 8 Transition Metal Sulfide Molecules
J. Phys. Chem. A, Vol. 113, No. 18, 2009 5381
is due to weak interaction with the symmetric Ru-S stretching
mode predicted at 362 cm-1. The antisymmetric Ru-S stretch-
ing mode is computed at 238 cm-1, which is too low to interact
with the higher symmetric modes in the mixed isotopic molecule
5
of lower symmetry. In addition, the A1 state is 14 kcal/mol
higher energy, and the leading mode at 471 cm-1 is too low for
assignment to the observed band. The very good agreement
between observed and computed vibrational characteristics for
3
Ru(S2) in the A2 ground state confirms its identification.
RuS3. A weaker resolved ruthenium natural isotopic quintet
from 541.5 to 546.1 cm-1 is 8 cm-1 higher than in the RuS2
quintet. The ruthenium isotopic splittings are the same within
experimental error, but the sulfur 32/34 frequency ratio is
slightly lower for the weaker quintet suggesting a slightly larger
obtuse valence angle for this product molecule with a similar
antisymmetric Ru-S stretching mode in isotopic character. Our
B3LYP/6-311+G(d)/LANL calculation for RuS3 predicts a 1A1′
ground state in D3h symmetry with intense antisymmetric
stretching frequency 19 cm-1 higher than RuS2 and 1.0197 sulfur
32/34 ratio. Unfortunately, the mixed sulfur 32,34 spectrum of
the disulfide covers this region. The weaker ruthenium isotopic
quintet is assigned to RuS3 on the basis of its observed and
calculated relationship with that for RuS2.
(S2)RuS2. Our B3LYP/6-311+G(d)/LANL calculation finds
the lowest energy structure for ruthenium tetrasulfide to have
the 1A1 ground state in the nonplanar C2V structure with strong
absorptions at 543 cm-1 (b2) and 530 cm-1 (a1). The tetrahedral
structure is 19 kcal/mol higher with strong stretching mode at
483 cm-1. The 540.8 cm-1 peak appears to be the Ru-102
absorption for a multiplet that is mostly covered by the multiplet
for RuS3. This band exhibits the 32/34 isotopic ratio 1.0202,
which is characteristic of an antisymmetric S-Ru-S stretching
mode. The stronger and broader 518.7 cm-1 band shows a
1.0245 isotopic 32/34 ratio and a broadened asymmetric triplet
mixed sulfur isotopic band with intermediate component 1.2
cm-1 below the pure isotopic median value, which indicates
interaction with two additional equivalent sulfur atoms and a
higher mode in the lower symmetry molecule. The 1.0245
isotopic frequency ratio characterizes a mixed Ru-S, S-S
stretching mode. Our calculation predicts 32/34 isotopic ratios
1.0201 and 1.0265 for the above b2 and a1 modes, which
considering the approximation involved, is very good agreement
with the observed frequencies and isotopic data.
Figure 7. Infrared spectra in the 550-480 cm-1 region for products
formed in the laser-ablated Os reaction with discharged sulfur-34 vapor
during condensation in excess argon at 7 K: (a) sample deposited for
60 min, (b) after annealing to 30 K, (c) after >220 nm irradiation, (d)
after annealing to 35 K, and (e) after annealing to 40 K.
The isotopic frequencies for central and terminal isotopic
substitution in the antisymmetric stretching mode for a C2V
molecule can be used to estimate the valence angle, as detailed
for the S3 molecule.30 Using the ruthenium isotopic data, a 112
( 2° lower limit and using the sulfur isotopic data, a 117 ( 1°
upper limit are predicted. These experimental values are in
excellent agreement with the 117.5° and 119.9° values computed
3
for the B1 state of RuS2, depending on the basis set.
Ru(S)2. The strong single band at 556.1 cm-1 shifts to 540.7
cm-1 with sulfur-34 and defines a larger 32/34 ratio, 1.0285,
which is characteristic of a S-S vibrational mode, hence no
ruthenium isotopic structure. In the mixed 32, 34 isotopic
experiment, a symmetrical triplet is observed with 548.4 cm-1
intermediate component at the median position. This means that
two equivalent sulfur atoms are involved and there is no nearby
stretching mode to interact in the lower symmetry mixed isotopic
molecule.
The B3LYP/6-311+G(3df)/SDD calculation for Ru(S)2 pre-
3
dicts a A2 ground state with medium intensity 549.7 cm-1
symmetric stretching frequency and 1.0300 harmonic sulfur 32/
34 isotopic frequency ratio. The slightly lower observed value
Figure 8. Structures calculated (B3LYP/6-311+G(3df)/SDD) for group 8 metal sulfur species.

5382 J. Phys. Chem. A, Vol. 113, No. 18, 2009
Liang et al.
-
Weak Low-Frequency Absorptions: RuS4. and RuS2 . We
observed the three strongest ruthenium isotopic components (Ru-
100, -102, -104) at 481.5, 479.9, and 478.5 cm-1 for a weak
absorption in the most productive experiment, which used sulfur-
34. This is where tetrahedral RuS4 is expected to absorb, and
our calculation predicts a 3.0 cm-1 Ru 100-104 shift, which is
in agreement with the observed band separation. The weak
absorptions increased 50% on UV irradiation and decreased on
higher temperature annealing and are tentatively assigned to the
higher energy RuS4 structural isomer. A weaker sulfur-32
counterpart was detected at 490 cm-1. Recall that RuO4 was
observed in the corresponding oxygen investigation at 96.4%
of the OsO4 frequency, and 490 cm-1 is 97.4% of the OsS4
frequency to be assigned below.27 Calculations with SDD tend
to overestimate and with LANL tend to underestimate the
observed RuS2 frequency, and this molecule is a case in point
(Table 2). Our calculation using SDD gave 511 cm-1, in
agreement with previous work,44 and using LANL found 483
cm-1: the computed values bracket the 490 cm-1 observed
frequency in the same way found for RuS2.
ground electronic state molecule indicates substantial interaction
between both stretching modes in the 32-Os-34 isotopic
molecule, but these bands are not resolved from other product
absorptions in the mixed isotopic spectrum. Thus, the identifica-
tion of the OsS2 molecule rests on the unique sulfur 32/34
isotopic ratio, which is calculated as 1.0241 and observed as
1.0240.
Although the cyclic Os(S)2 molecule or complex appears to
increase on annealing, apparently the more stable OsS2 molecule
does not, which suggests that its formation may require
activation energy. It is also suggested that OsS2 combines with
S2 on annealing to form both (S2)OsS2 and OsS4. Ultraviolet
irradiation clearly favors the final, most stable OsS4 product.
Os + S2 f Os(S2)
Os + S2 f OsS2
(1)
(2)
(3)
OsS2 + S2 f (S)2OsS2 f OsS4
Three additional weak absorptions with similar profile at
475.5, 473.9, 472.5 cm-1 were almost destroyed by UV
photolysis, and these are within a few cm-1 of the position
computed for the very strong b2 mode of 4B2 ground-state RuS2-,
and the 3.1 cm-1 predicted Ru 100-104 separation is in
agreement with the observed band splitting. These bands are
appropriate for the anion assignment, but without more observa-
tions, this must be considered tentative. The photosensitive
(S2)OsS2. Three bands at 545.1, 536.8, and 520.2 cm-1 track
together on annealing and exhibit different sulfur 32/34 isotopic
ratios. All three of these bands exhibit asymmetric mixed
isotopic triplets with the higher two showing the most coupling
between the two modes of the OsS2 functional group when the
symmetry is reduced and the lower mode revealing characteristic
behavior for an Os(S2) group. The three bands are calculated at
545.3, 537.4, and 523.0 cm-1 for the nonplanar C2V (S2)OsS2
-
RuO2 anion was also observed in the earlier experiments.27
Our DFT calculation finds a 5∆ ground state for RuS with a
478.3 cm-1 harmonic frequency and 2.3 cm-1 Ru 100-104 shift,
which is not in agreement with the observed band spacings. In
addition, the infrared intensity for RuS is much less than for
the above antisymmetric stretching modes. Hence, we believe
that RuS is not observed in these experiments.
1
complex in the A1 ground state with sulfur 32/34 isotopic
frequency ratios 1.0242, 1.0275, and 1.0301, which are in
extremely good agreement with the observed values (Table 1).
Even more impressive is the calculation of the mixed isotopic
triplets: one S-34 in the OsS2 functional group is predicted to
give intermediate components 2.7 and 10.9 cm-1 below the all-
S-32 values and we observed these bands 2.0 and 10.4 cm-1
below, and one S-34 in the Os(S2) group is predicted to give a
mixed component 7.6 cm-1 below the all-S-32 value and this
band appears 6.5 cm-1 lower. In summary, the observation of
the three strongest stretching modes and their detailed sulfur-
34 substitution character in excellent agreement with B3LYP
calculated values confirms the identification of the (S2)OsS2
5
5
OsS. Our DFT calculation finds low Σ+ and ∆ states for
OsS with 509 and 494 cm-1 harmonic frequencies and 1.0261
sulfur 32/34 isotopic frequency ratios. These are very ap-
proximate predictions of the heavy metal OsS vibrational
fundamental. A weak 492.4 cm-1 band in these experiments
sharpens on UV photolysis and shifts to 479.9 cm-1 with sulfur-
34 (ratio 1.0260). Accordingly, this weak band is appropriate
for OsS, which merits a tentative assignment.
1
complex in the A1 ground state.
OsS2 and Os(S2). The osmium and sulfur product spectrum
is complicated as four absorptions with different sulfur isotopic
shifts fall in the upper region. The global minimum energy
species is open OsS2 in the 3B1 ground electronic state, as found
for RuS2, but in the OsS2 case the 5B2 state is now 11 kcal/mol
higher in energy. Our best calculation finds the 3A2 ground state
of cyclic Os(S2) to be 44 kcal/mol higher energy than the global
minimum energy open species. The first band to notice is the
sharp 546.9 cm-1 peak, which shifts to 531.2 cm-1 with sulfur-
34 and defines the large (1.0296) sulfur 32/34 isotopic frequency
ratio characteristic of an almost pure S-S stretching mode. The
single sharp mixed isotopic counterpart at 539.1 cm-1 is
precisely at the median of pure isotopic counterparts, as expected
for the Os(S2) molecule without nearby interacting vibrational
modes.B3LYP calculation predicts the S-S stretching mode for
this molecule at 546.0 cm-1, very close to the observed band.
The OsS2 molecule unfortunately shares the 537 cm-1 band
profile with natural isotopic sulfur, but the two bands are
resolved with sulfur-34 at 524.9 and 522.6 cm-1, and the
shoulder at 537.5 cm-1 defines the sulfur-32 counterpart and
the 1.0240 isotopic frequency ratio. Our calculation for the 3B1
OsS4. The anticipated tetrahedral OsS4 molecule is 3 kcal/
mol lower in energy than the above (S2)OsS2 complex. The same
calculation that predicted the above complex fundamentals
within 3 cm-1 gave the intense triply degenerate mode of OsS4
at 499.4 cm-1 with sulfur isotopic frequency ratio 1.0246, and
the large Gaussian basis increased this to 500.5 cm-1. Our
computation of tetrahedral OsS4 is in excellent agreement with
a previous theoretical report.44 This calculation was repeated
for all possible mixed sulfur isotopes, and most of the intensity
remains in the pure isotopic bands as there is little mode
coupling thought the heavy Os metal center, just as observed
for OsO4.27 Weak mixed isotopic bands are predicted 1.2, 2.6,
and 4.6 cm-1 above the all-34 peak.
The strong 503.7 cm-1 band increases on annealing and UV
irradiation and shifts to 491.6 cm-1 using sulfur-34 with the
same frequency ratio as calculated for OsS4. Broadening of the
sulfur-34 component in the mixed isotopic experiment is
ascribed to the first of the mixed isotopic bands computed for
Os32S334S. The chemical behavior of these absorptions is
appropriate for their assignment to the stable OsS4 molecule,
which is confirmed by the frequency and isotopic shift calcula-

Group 8 Transition Metal Sulfide Molecules
J. Phys. Chem. A, Vol. 113, No. 18, 2009 5383
(3) Wells, A. F. Structural Inorganic Chemistry, 4th ed.; Clarendon:
Oxford, UK, 1975. (b) Sourisseau, C.; Cavagnat, R.; Fouassier, M. J. Phys.
Chem. Solids 1991, 52, 537.
(4) (a) Williams, R. J. P. Nature (London) 1990, 343, 213. (b) Drobner,
E.; Huber, H.; Wa¨chtersha¨user, G.; Rose, D.; Stetter, K. O. Nature (London)
1990, 346, 742.
(5) Posfai, M.; Buseck, P. R.; Bazylinski, D. A.; Frankel, R. B. Science
1998, 280, 880.
(6) McKay, D. S.; Gibson, E. K.; ThomasKeprta, K. L.; Vali, H.;
Romanek, C. S.; Clemett, S. J.; Chilier, X. D. F.; Maechling, C. R.; Zare,
R. N. Science 1996, 273, 924.
(7) Cosper, N. J.; Booker, S. J.; Ruzicka, F.; Frey, P. A.; Scott, R. A.
Biochemistry, 2000, 39, 15668.
(8) Ugulava, N. B.; Carlos, J. S.; Jarrett, J. T. Biochemistry 2001, 40,
8352.
(9) Anxolabehere-Mallart, E.; Glaser, T.; Frank, P.; Aliverti, A.; Zanetti,
G.; Hedman, B.; Hodgson, K. O.; Solomon, E. I. J. Am. Chem. Soc. 2001,
123, 5444.
tions. Although OsO4 is a well-known molecule of tetrahedral
symmetry,45,46 we believe that this is the first experimental
observation of OsS4.
Structure and Bonding Trends within the Group 8 Family.
First, let us compare the group 8 disulfide and dioxide molecules,
which have all been observed in solid argon matrices. Our DFT
3
5
calculations showed that the B1 and B2 states for FeO2 are
close in energy, but the observed 945.8 and 797.1 cm-1
stretching frequencies and oxygen 16/18 isotopic ratios clearly
3
5
fit predictions for the B1 and not the B2 state.26 The reverse
situation holds for FeS2 where the 540.2 and 467.2 cm-1
absorptions and sulfur 32/34 isotopic data match the calculations
5
3
for the B2 state and not the higher energy B1 state (Table 2).
-
-
Both FeO2 and FeS2 were observed in low laser energy
experiments where electron capture products are favored.42c
The bent RuO2 molecule was well described by isotopic
frequencies and calculations for the 1A1 ground electronic state,27
and the case for the 3B1 ground state of RuS2 has been made in
this work. Both OsO2 and OsS2 were adequately described by
isotopic frequencies and calculations for the 3B1 ground
electronic state. In the case of Fe and Ru, the disulfide has higher
spin multiplicity than the dioxide.
(10) Frenchard, F.; Sautet, P. J. Catal. 1997, 170, 402.
(11) Jeevanandam, P.; Koltypin, Y.; Gofer, Y.; Diamant, Y.; Gedanken,
A. J. Mater. Chem. 2000, 10, 2769.
(12) Eckermann, A.; Fenske, D.; Rauchfuss, T. B. Inorg. Chem. 2001,
40, 1459.
(13) Lassiter, J. C.; Luhr, J. F. Geochem. Geophys. Geosys. 2001, 2,
U1.
(14) MacMahon, T. J.; Jackson, T. C.; Freiser, B. S. J. Am. Chem. Soc.
1989, 111, 421.
The open MS2 disulfide is more stable than the side-bonded
M(S2) complex, and this difference increases with group 8 metal
size.
(15) Zhang, N.; Hayase, T.; Kawamata, H.; Nakao, K.; Nakajima, A.;
Kaya, K. J. Chem. Phys. 1996, 104, 3413.
(16) Zhai, H. J.; Kiran, B.; Wang, L.-S. J. Phys. Chem. 2003, 107, 2821.
(17) Harvey, J. N.; Heinemann, C.; Fiedler, A.; Schro¨der, D.; Schwartz,
The side-bonded (η2-O2)FeO2 complex has been character-
ized,26 but the sulfur analog Fe(S2)2 has both disulfur molecules
side-bonded. There is no evidence for the higher energy FeO4
or FeS4 molecules. Both RuO4 and the higher energy side-
bonded (η2-O2)RuO2 complex were observed,27 but the lower
energy (η2-S2)RuS2 complex clearly dominated the RuS4
molecule in the present experiments. In the osmium case, the
much lower energy OsO4 molecule exceeded the (η2-O2)OsO2
complex band absorbance by more than an order of magnitude
whereas OsS4 is only slightly lower in energy than the (η2-
S2)OsS2 complex, and both species are clearly observed here.
These observations are in line with the increasing stability of
the VIII oxidation state on going down the group 8 metal family.
H. Chem.sEur. J. 1996, 2, 1230.
(18) Schro¨der, D.; Kretzschmar, I.; Schwartz, H.; Rue, C.; Armentrout,
P. B. Inorg. Chem. 1999, 38, 3474.
(19) Husband, J.; Aguirre, F.; Thompson, C. J.; Metz, R. B. Chem. Phys.
Lett. 2001, 342, 75.
(20) DeVore, T. C.; Franzen, H. F. High Temp. Sci. 1975, 7, 220. The
542 cm-1 Fe and OCS reaction product first assigned to FeS is due to another
species.
(21) Takano, S.; Yamamoto, S.; Saito, S. J. Mol. Spectrosc. 2004, 224,
137.
(22) Bauschlicher, C. W. Jr.; Maitre, P. Theor. Chim. Acta 1995, 90,
189.
(23) (a) Hu¨bner, O.; Termath, V.; Berning, A.; Sauer, J. Chem. Phys.
Lett. 1998, 294, 37. (b) Hu¨bner, O.; Sauer, J. Phys. Chem. Chem. Phys.
Lett. 2002, 4, 5234.
(24) (a) Bridgeman,A. J.; Rothery, J. J. Chem. Soc., Dalton Trans.
2000, 211. (b) Mebel, A. M.; Hwang, D.-Y. J. Phys. Chem. A 2001,
105, 7460.
Conclusions
(25) Clima, S.; Hendrickx, M. F. A. Chem. Phys. Lett. 2007, 436, 341.
(26) (a) Andrews, L.; Chertihin, G V.; Ricca, A.; Bauschlicher, C. W.,
Jr. J. Am. Chem. Soc. 1996, 118, 467. (b) Chertihin, G. V.; Saffel, W.;
Yustein, J. T.; Andrews, L.; Neurock, M.; Ricca, A.; Bauschlicher, C. W.
Jr. J. Phys. Chem. 1996, 100, 5261. (c) Gong, Y.; Zhou, M.; Andrews, L.,
J. Phys. Chem. A 2007, 111, 12001 (Fe + O2).
Laser-ablated Fe, Ru, and Os atoms were reacted with small
sulfur molecules from a microwave discharge in argon and
condensed at 7 K. Numerous reaction products were identified
from matrix infrared spectra, sulfur-34 isotopic shifts, spectra
of sulfur isotopic mixtures, and frequencies from density
functional calculations. The strongest absorptions of the MS2
disulfide molecules are observed at 540.2, 535.5, and 537.5
cm-1, respectively, for the group 8 metals. The photosensitive
FeS2- anion was detected at 542.1 cm-1. The RuS2 absorption
exhibited five resolved natural ruthenium isotopic splittings.
Evidence is also presented for side-bound M(S2) isomers and
MS4 molecules with the structures Fe(S2)2, (S2)RuS2, and
tetrahedral OsS4, which were computed to be the lowest energy
species for this stoichiometry.
(27) Zhou, M.; Citra, A.; Liang, B.; Andrews, L. J. Phys. Chem. A 2000,
104, 3457.
(28) Liang, B.; Andrews, L. J. Phys. Chem. A 2002, 106, 3738.
(29) Liang, B.; Andrews, L. J. Phys. Chem. A 2002, 106, 6295.
(30) Brabson, G. D.; Mielke, Z.; Andrews, L. J. Phys. Chem. 1991, 95,
79.
(31) (a) Andrews, L.; Citra, A. Chem. ReV. 2002, 102, 885. (b) Andrews,
L.; Cho, H.-G. Organometallics 2006, 25, 4040.
(32) Cho, H.-G.; Lyon, J. T.; Andrews, L. Organometallics 2008, 27,
5241.
(33) Frisch, M. J.; et al. Gaussian 98, reVision A.1; Gaussian, Inc.:
Pittsburgh, PA, 1998.
(34) (a) Becke, A. D. J. Chem. Phys. 1993, 98, 5648. (b) Lee, C.; Yang,
Y.; Parr, R. G. Phys. ReV. B 1988, 37, 785.
(35) Perdew, J. P.; Wang, Y. Phys. ReV. B 1992, 45, 13244.
(36) Raghavachari, K.; Trucks, G. W. J. Chem. Phys. 1989, 91, 1062.
(37) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 270. (a) Wadt,
W. R.; Hay, P. J. J. Chem. Phys. 1985, 82, 284. (b) Hay, P. J.; Wadt, W. R.
J. Chem. Phys. 1985, 82, 299.
Acknowledgment. We gratefully acknowledge financial
support from NSF Grant CHE 00-78836 and NCSA computing
Grant No. CHE07-0004N to L.A.
(38) Frisch, M. J.; et al. Gaussian 03, ReVision C.02; Gaussian, Inc.:
Pittsburgh, PA, 2003.
References and Notes
(1) Stiefel, E. I., Matsumoto, K., Eds.; Transition Metal Sulfur
Chemistry, Biological and Industrial Significance; American Chemical
Society: Washington, DC, 1997.
(39) Andrae, D.; Haeussermann, U.; Dolg, M.; Stoll, H.; Preuss, H.
Theor. Chim. Acta 1990, 77, 123.
(40) (a) Long, S. R.; Pimentel, G. C. J. Chem. Phys. 1977, 66, 2219.
(b) Smardzewski, R. R. J. Chem. Phys. 1978, 68, 2878.
(2) Rickard, D.; Luther, G. W., III Chem. ReV 2007, 107, 514.

5384 J. Phys. Chem. A, Vol. 113, No. 18, 2009
Liang et al.
(41) The natural abundances of 102Ru and 104Ru are 32% and 19%. CRC
Handbook of Chemistry and Physics; CRC Press: Boca Raton, FL, 1985.
(42) (a) Zhou, M.; Andrews, L. J. Am. Chem. Soc. 1998, 120, 11499.
(b) Zhou, M.; Andrews, L.; Bauschlicher, C. W., Jr. Chem. ReV 2001, 101,
1931. (c) Li, Z. H.; Gong, Y.; Fan, K.; Zhou, M. J. Phys. Chem. A 2008,
112, 13641.
(44) Zli, S.; Stoll, H.; Baerends, E. J.; Kaim, W. Inorg. Chem. 1999,
38, 6101.
(45) McDowell, R. S.; Goldblatt, M. Inorg. Chem. 1971, 10, 625.
(46) Cotton, F. A.; Wilkinson, G.; Murillo, C. A.; Bochmann, M.
AdVanced Inorganic Chemistry, 6th ed.; Wiley: New York, 1999.
(43) Liang, B.; Wang, X.; Andrews, L. J. Phys. Chem. A 2009, 113, in
press (Ni, Pd, Pt + S2).
JP900994C