Structure and stability of small nitrile sulfides and their attempted generation from 1,2,5-thiadiazoles

Article abstract of DOI:10.1021/jp010830e

The gas-phase generation and spectroscopic identification of nitrile sulfides by thermolysis of 1,2,5-thiadiazole precursors was attempted, but in all cases the thiadiazoles were found to produce sulfur and the corresponding nitrile. This prompted an investigation by ab initio and density functional calculations for the equilibrium geometries, stabilities, and decomposition mechanisms of several nitrile sulfides (XCNS, where X = H, F, Cl, CN, CH₃). Equilibrium geometries obtained from calculations at the B3LYP, MPn(n = 2-4), QCISD, QCISD(T), CCSD, and CCSD(T) levels with moderate to large basis sets indicate that the molecules have linear heavy atom geometries. The exception is the fluoro derivative, which is bent with a calculated barrier to linearity of 889 cm^-1 (B3LYP/cc-pVTZ). The nitrile sulfides are predicted by the B3LYP method to be stable in the dilute gas phase, whereas in the condensed phase they are suggested to be very unstable due to bimolecular decomposition. The mechanism of this loss process is complicated by various sulfur transfer and cyclization reactions between decomposition intermediates, with the predicted stable products being sulfur, nitriles, and thiadiazoles. The first step of the bimolecular decomposition is either a cycloaddition to thiofuroxan or a sulfur transfer with simultaneous S₂ loss to nitriles.

Full text of DOI:10.1021/jp010830e

6258
J. Phys. Chem. A 2001, 105, 6258-6265
Structure and Stability of Small Nitrile Sulfides and Their Attempted Generation from
1,2,5-Thiadiazoles
Tibor Pasinszki,*,† Tama´s Ka´rpa´ti,† and Nicholas P. C. Westwood*,‡
Department of Inorganic Chemistry, Budapest UniVersity of Technology and Economics, H-1521 Budapest,
Gelle´rt te´r 4, Hungary, and Guelph-Waterloo Centre for Graduate Work in Chemistry, Department of
Chemistry and Biochemistry, UniVersity of Guelph, Guelph, Ontario, Canada N1G 2W1
ReceiVed: March 5, 2001; In Final Form: April 23, 2001
The gas-phase generation and spectroscopic identification of nitrile sulfides by thermolysis of 1,2,5-thiadiazole
precursors was attempted, but in all cases the thiadiazoles were found to produce sulfur and the corresponding
nitrile. This prompted an investigation by ab initio and density functional calculations for the equilibrium
geometries, stabilities, and decomposition mechanisms of several nitrile sulfides (XCNS, where X ) H, F,
Cl, CN, CH3). Equilibrium geometries obtained from calculations at the B3LYP, MPn(n ) 2-4), QCISD,
QCISD(T), CCSD, and CCSD(T) levels with moderate to large basis sets indicate that the molecules have
linear heavy atom geometries. The exception is the fluoro derivative, which is bent with a calculated barrier
to linearity of 889 cm-1 (B3LYP/cc-pVTZ). The nitrile sulfides are predicted by the B3LYP method to be
stable in the dilute gas phase, whereas in the condensed phase they are suggested to be very unstable due to
bimolecular decomposition. The mechanism of this loss process is complicated by various sulfur transfer and
cyclization reactions between decomposition intermediates, with the predicted stable products being sulfur,
nitriles, and thiadiazoles. The first step of the bimolecular decomposition is either a cycloaddition to thiofuroxan
or a sulfur transfer with simultaneous S2 loss to nitriles.
Introduction
of this process has been the subject of some debate and is not
yet understood, although S2 has been identified as being a
primary product.2,3
Since the first indirect evidence for the existence of the first
nitrile sulfide, XCNS (benzonitrile sulfide) in 1970,1 these
compounds have become important transient species and
reactive intermediates in organic chemistry. Although much less
stable than the corresponding nitrile oxides, to the extent that
their isolation in the pure state is not possible, they have been
used in 1,3-dipolar cycloaddition reactions in solution, providing
routes to several classes of heterocycles accessible only with
difficulty by other means.2 They are usually generated in situ,
in the presence of a dipolarophile, by thermal decomposition
of a five-membered ring containing the CNS moiety. The
method of choice generally involves thermolysis of substituted
1,3,4-oxathiazol-2-ones. The chemical trapping reactions provide
the strongest evidence for the existence of nitrile sulfides,
suggesting that the nitrile sulfide forms and then instantly reacts.
Two reviews,2,3 considering work up to about 1991, summarize
their chemistry, mechanistic aspects, and some spectroscopy.
Spectroscopic studies on nitrile sulfides are limited due to
their instability at room temperature. PhCNS and CH3CNS have
been detected by infrared and ultraviolet spectroscopies when
generated photolytically in a cold matrix or when the gas-phase
pyrolysis products were condensed onto a cold window at
cryogenic temperatures using matrix isolation techniques (see
refs 2-4 and references therein). Nitrile sulfides are believed
to be thermally unstable, readily losing sulfur; photolysis of
nitrile sulfides at cryogenic temperatures or allowing the sample
to warm above 100 K resulted in desulfuration.2 The mechanism
The parent thiofulminic acid (HCNS) and other simple nitrile
sulfide derivatives, e.g., PhCNS, NCCNS, ClCNS, NH2CNS,
and CH3CNS have recently been generated and identified as
neutrals in the gas phase by neutralization-reionization mass
spectroscopy (NRMS),3-7 as have some of the corresponding
seleno compounds (RCNSe).8,9 In the case of the nitrile sulfides
a beam of mass-selected nitrile sulfide ions, produced either
by electron impact ionization and subsequent fragmentation of
suitable five-membered heterocycles containing at least one CNS
•+
linkage, or by sulfur ion transfer from CS3 to the nitrogen
atom of nitriles, is neutralized in a collision cell with Xe or
NH3 and then reionized in a second collision cell with O2. This
method, which establishes the existence of the neutral species
with lifetimes of at least microseconds in the dilute gas phase
in the mass spectrometer, has no preparative value for nitrile
sulfides. Because of the lack of successful methods to generate
these molecules into the dilute gas phase in sufficient yield, no
other spectroscopic data exist. The experimental structures of
nitrile sulfides are thus unknown and methods to generate them
into the gas phase would provide an important step toward their
full electronic and geometric characterization.
Given the paucity of experimental investigations, quantum-
chemical calculations can provide information on the structures
and stabilities. Early ab initio Hartree-Fock (HF) calculations
on HCNS predicted a linear structure,10 which seemed to be in
good agreement with the most frequently used mesomeric
structure of nitrile sulfides (XsCtNfS) which, with an sp
carbon atom, suggests a linear formulation. This was challenged
recently by MP2 calculations using various basis sets,7 predicting
a slightly bent equilibrium configuration. The structures of some
* Authors for correspondence. E-mail: pasinszki.inc@chem.bme.hu;
westwood@chembio.uoguelph.ca.
† Budapest University of Technology and Economics.
‡ University of Guelph.
10.1021/jp010830e CCC: $20.00 © 2001 American Chemical Society
Published on Web 06/05/2001

Structure and Stability of Small Nitrile Sulfides
J. Phys. Chem. A, Vol. 105, No. 25, 2001 6259
simple derivatives were also calculated at the MP2 level; the
frame of NCCNS11 and CH3CNS7 was calculated to be linear,
whereas the frame of CNCNS,11 BrCNS,7 ClCNS,6 and NH2-
CNS6 was predicted to be bent. Barriers to linearity and potential
quasi-linear behavior has not been investigated except for
CNCNS;11 this latter species being found to be quasi-linear with
a calculated barrier to linearity of 139 cm-1. There is still no
clear consensus on the structures of nitrile sulfides. They are
anticipated to be linear or quasi-linear by analogy with nitrile
oxides, but with calculated bond lengths sensitive to the bent-
linear question; e.g., the calculated CN bond length varies
between 1.140 and 1.208 Å in the various derivatives,6,7,10,11
which cannot be explained by substituent effects. We note, and
will show below, that neither MP2, which exaggerates the CN
bond length and tends to bend the molecule, nor HF, which
strongly underestimates the CN bond length, is the method of
choice for calculating the structures of nitrile sulfides, and
conclusions drawn on the basis of MP2 and HF calculations
must be taken cum grano salis. Calculations on nitrile sulfides,
as with the nitrile oxides,12,13 are sensitive to electron correlation
effects and the description of these latter effects is of crucial
importance.
In this paper we report a quantum-chemical study on the
equilibrium structures and stabilities of small nitrile sulfides
(XCNS, where X ) H (1), F (2), Cl (3), CN (4), CH3 (5)), and
their attempted generation from 1,2,5-thiadiazoles. These het-
erocycles were considered as potential precursors based on a
recent report6 of the formation of cyanogen N-sulfide (NCCNS)
in the flash vacuum thermolysis (FVP, 750 °C) of 3,4-dicyano-
1,2,5-thiadiazole. In addition, analogous heterocycles14 were
successfully used for the generation of unstable gas-phase nitrile
oxides, particularly CH3CNO13 and NCCNO15 for which various
spectroscopies (photoelectron and infrared in particular) provide
unambiguous evidence for their detection and electronic and
geometric structures. Of special computational interest are the
structures and potential mechanisms for the decomposition of
nitrile sulfides.
IR spectra were collected on a Nicolet 20SXC interferometer
equipped with a 20 cm single pass cell. The cell, with KBr
windows, gave a spectral range from 4000 to 400 cm-1. The
effluent from the pyrolysis tube was pumped continuously
through the cell using a rotary pump while maintaining the
pressure constant between 400 and 500 mTorr.
Computational Methods
The equilibrium structure of the parent nitrile sulfide, HCNS
(1) was calculated at the HF, MP2, MP3, MP4, QCISD, QCISD-
(T), CCSD, and CCSD(T) levels using standard 6-31G**,
6-311G(2d,2p), 6-311+G(3df,3pd), or cc-pVTZ basis sets, and
also using density functional theory in the form of Becke’s three-
parameter exchange functional in combination with the Lee,
Yang, and Parr correlation functional (B3LYP). The equilibrium
geometries of the computationally larger derivatives, FCNS (2),
ClCNS (3), NCCNS (4), and CH3CNS (5), were calculated at
the B3LYP/cc-pVTZ level. Calculations for the stabilities
(monomolecular and bimolecular sulfur loss) and decomposition
mechanisms were performed at the B3LYP/6-31G** level. The
monomolecular sulfur loss was also calculated at the CCSD-
(T)(full)/cc-pVTZ//B3LYP/cc-pVTZ level. Equilibrium molec-
ular geometries were fully optimized and harmonic vibrational
frequencies were then calculated at the minimum energy
geometries to confirm they were real minima on the potential
energy surface (zero imaginary frequencies). Transitional states
(TSs, first-order saddle points) were characterized with one
imaginary frequency. All calculations were performed with the
Gaussian-98 quantum chemistry package21 implemented on
Silicon Graphics Inc. Challenge/XL and Origin200 workstations.
Results and Discussion
Thermolysis of Thiadiazoles. All of the following observa-
tions were confirmed by PE, PIMS, and IR measurements.
Thermolyses of thiadiazoles 6-10 were carried out in a quartz
tube (8 mm i.d.) heated along 15 cm and attached directly to
the spectrometer (PE/PIMS) or gas cell (IR); for a more efficient
pyrolysis, the tube was loosely packed with quartz chips. We
find that 1,2,5-thiadiazoles are thermally very stable, with the
stability strongly influenced by the substituent. These thermoly-
ses, however, did not produce identifiable nitrile sulfides, not
even using 8, which was noted earlier6 to produce 4 by FVP,
one of the only nitrile sulfides, other than PhCNS,5 to be
generated by FVP, rather than in a mass spectrometer. The
detection,6 however, was by mass spectrometry, identification
being based upon the observation of a decrease of the m/z
intensity ratio of 136/84 by increasing the temperature of
pyrolysis (0.8 at 200 °C and 0.6 at 750 °C). It is possible that
this intensity change is due to secondary ion-molecule reactions
and does not arise from the formation of 4. However, it is more
likely that our present inability to observe 4 arises from a longer
contact time in our furnace and, more particularly, from the
higher pressures used in the present experiments, typically
around 0.1-1 mTorr in the ionization region of the UPS
experiment and ca. 450 mTorr in the FTIR cell. We note that
the νas/s(CNS) IR bands of nitrile sulfides have a predicted large
oscillator strength, expected to be distinctive for nitrile sulfides,
much as the corresponding vibrations are important fingerprints
for the nitrile oxide analogues.14 However, we could not identify
such bands in the IR spectra of the pyrolysates even after
prolonged accumulation of the spectra.
Experimental Section
3,4-Dichloro-1,2,5-thiadiazole (6) was a commercial product
(Aldrich). All other thiadiazole derivatives were synthesized
according to known literature methods: 3,4-difluoro-1,2,5-
thiadiazole (7) was synthesized from the dichloro-derivative with
potassium fluoride,16 3,4-dicyano-1,2,5-thiadiazole (8) from
diaminomaleonitrile with thionyl chloride,17 1,2,5-thiadiazole
(9), and 3,4-dimethyl-1,2,5-thiadiazole (10) from ethylenedi-
amine dihydrochloride and dimethylglyoxime, respectively, with
sulfur monochloride.18 5-Methyl-1,3,4-oxathiazole-2-one (11)
was synthesized from acetamide and ClC(O)SCl.19
For monitoring the gas-phase thermolysis and for identifying
the pyrolysis products a combination of three different spec-
troscopic methods were used: He I photoelectron (PE), pho-
toionization mass (PIMS), and infrared (IR) spectroscopies. He
I (21.2 eV) PE spectra were obtained on a home-built fast
pumping spectrometer20 used to monitor in situ the products of
fast flow thermolysis reactions. The capability exists for mass
analyzing ions produced in the photoionization process using a
quadrupole mass analyzer (Hiden Analytical, 320 amu) mounted
directly above the photoionization point. The conventional EI
source of the mass analyzer is removed, ionization being
provided by He I or unfiltered HL (10.2-12.7 eV) radiation.
Although not done in coincidence, PE and PIMS spectra can
be recorded within seconds of each other; thus it is assumed
that for a given PE spectrum the subsequent PIMS is of the
same compound.
The decomposition of thiadiazoles 6-10 rapidly and quan-
titatively produces nitriles and sulfur (Scheme 1) with the
decomposition rate depending strongly on the substituent. The

6260 J. Phys. Chem. A, Vol. 105, No. 25, 2001
Pasinszki et al.
TABLE 1: Calculated Equilibrium Structure, Energies, Barrier to Linearity, Dipole Moment, and Rotational Constants of
HCNSa
B3-LYP
MP2
MP3
MP4SDQ MP4SDTQ
Barrier to Linearity
QCISD
QCISD(T)
CCSD
CCSD(T)
basis set
6-31G**
6-311G(2d,2p)
cc-pVTZ
6-311+G(3df,3pd) 0
0
0
0
14
0
0
0
0
0
0
0
99
0
11
14
0
0
0
0
0
0
0
b
0
0
0
b
0
0
0
0
0
b
0
0.3
Structure for the 6-311+G(3df,3pd) Basis Set
H-C
C-N
N-S
HCN
CNS
tot. energy:
µc
1.062
1.159
1.596
180.0
180.0
1.062
1.180
1.574
174.1
179.0
1.061
1.150
1.600
180.0
180.0
1.063
1.158
1.607
180.0
180.0
1.066
1.185
1.584
165.0
177.3
1.063
1.157
1.609
180.0
180.0
1.065
1.167
1.601
180.0
180.0
1.063
1.155
1.609
180.0
180.0
1.064
1.167
1.601
180.0
180.0
-491.65082 -491.13688 -491.13812 -491.14550 -491.18636 -491.14530 -491.17829 -491.14281 -491.17701
3.57
6.170
4.88
6.205
5.20
6.184
5.21
6.120
4.81
6.150
5.23
6.116
5.14
6.114
5.24
6.125
5.14
6.120
Bd
a Bond angles in degrees, bond lengths in angstroms, total energies in au, and barrier to linearity in cm-1; all of the electrons were included in
the correlation energy calculations (“full”). b Not calculated. c Dipole moment in Debye. d Rotational constants in GHz; Isotopes: H-1, C-12, N-14,
S-32.
SCHEME 1
of the nitrile sulfide. Decomposition of 11 occurred at 350 °C
and was complete at 500 °C, much lower temperatures than
the substituted 1,2,5-thiadiazoles, 6-10. No nitrile sulfide could
be identified, 11 decomposing quantitatively to acetonitrile,
sulfur, and carbon dioxide (Scheme 1), in agreement with the
earlier investigation.
Equilibrium Structure of Nitrile Sulfides. Due to the lack
of experimental data on the structures of nitrile sulfides, the
reliability of the applied theoretical method cannot be tested
by the usual procedure of comparing experiment and calculation.
However, electron correlation and basis set effects are expected
to be crucial much as they are in the case of the nitrile oxides12,13
and so the smallest and computationally most tractable deriva-
tive, HCNS, 1, was selected as a test case with its geometry
calculated at various levels of theory and basis set size. Results
are shown in Table 1.
1 has a linear structure at all levels of theory and with all
basis sets, except for MP2 and MP4 using 6-31G**, cc-pVTZ,
and 6-311+G(3df,3pd) basis sets. The barrier to linearity at MP2
and MP4 is, however, small and decreases with increasing basis
set size. There is a linear-bent oscillation in the MPn expansion
series; MP1 ) HF(linear) f MP2(bent) f MP3(linear) f MP4-
(bent), which is also coupled to the bond lengths, especially
the CtN triple bond, for which oscillations can be seen in the
MPn series. The CtN bond lengths are 1.123, 1.180, 1.150,
and 1.185 Å at the HF (not shown in Table 1), MP2, MP3, and
MP4/6-311+G(3df,3pd) levels, respectively. The inclusion of
triple effects in the calculations strongly lengthens CN; compare
MP4SDQ vs MP4SDTQ, QCISD vs QCISD(T), and CCSD vs
CCSD(T). Such observations are a strong indication of large
and sensitive electron correlation effects and thus their descrip-
tion is of crucial importance. To provide an insight into
the nature of the electron correlation problem, single point
CISD(full) calculations were also done at the CCSD(T) geom-
etry using the 6-311+G(3dp,3pd) basis set. The CISD wave
function indicates that apart from the main HF configuration
there is no other important configuration (weight of configura-
tions larger than 0.05%: 93.6 (HF), 0.06, 0.06, 0.05, and 0.05%),
and thus single reference post-HF methods should be sophis-
ticated enough to describe the structure of 1. The description
of dynamic electron correlation, however, requires the use of
high-level ab initio methods.
dimethyl-substituted thiadiazole 10 is thermally the less stable.
Decomposition of 10 commenced at 650 °C with increasing
temperature, decreasing the amount of unreacted precursor, and
at 900 °C the decomposition was complete. Other than the major
products, acetonitrile and sulfur, a small amount of HCN was
detected, its relative amount gradually increasing with temper-
ature. Decomposition of 8 started above 700 °C, producing
dicyanogen and sulfur. In each case, sulfur precipitated on the
cold part of the glassware after the hot zone of the furnace. We
found that unreacted precursor 8 could be trapped with a simple
ice-water cooled U-trap inserted between the furnace and
spectrometer. Thus, thermolysis of 8 could provide a simple
and clean route to the on-line and small scale production of
NCCN for spectroscopic investigations, thereby avoiding trans-
port and storage of the highly poisonous NCCN gas. The
decomposition of 9 commenced at 800 °C, but even at 900 °C
the precursor composed the largest component, leaving the
furnace. The formation of HCN and sulfur could be clearly
observed. The halogen derivatives 6 and 7 were thermally even
more stable; they did not decompose at 800 °C. Only a small
amount of 6 decomposed at 900 °C, and the best indication of
this was the precipitation of a small amount of sulfur on the
glassware. We could not observe any sign of the thermal
decomposition of 7 up to 900 °C.
Although mentioned earlier4 that the FVP of 5-methyl-1,3,4-
oxathiazole-2-one, 11, does not produce acetonitrile sulfide, but
acetonitrile, sulfur, and carbon dioxide, we repeated this work
hoping that at a lower temperature we might identify formation
It is important to note that we have found previously similar
strong electron correlation effects, demonstrated by an oscilla-

Structure and Stability of Small Nitrile Sulfides
J. Phys. Chem. A, Vol. 105, No. 25, 2001 6261
TABLE 2: Calculateda Equilibrium Structure of Nitrile Sulfides, XCNS (X ) H, F, Cl, NC, and CH3)
HCNS, 1
1.061
1.160
1.603
FCNS,b 2
1.296
1.186
1.589
ClCNS, 3
1.625
1.163
1.609
NCCNS, 4
H3CCNS, 5
X-C
C-N
N-S
N-C
C-H
XCN
CNS
NCC
HCC
tot. energy
µc
1.349
1.171
1.577
1.161
1.449
1.159
1.619
1.092
180.0
180.0
180.0
180.0
139.5
169.2
180.0
180.0
180.0
180.0
180.0
110.5
-530.99779
5.28
159.982
2.406
-491.65340
3.57
-590.88996
1.32
206.578
2.562
-951.27123
3.89
-583.92544
1.42
Ad
B
6.135
1.522
1.498
C
2.530
2.406
a Calculated at the B3-LYP/cc-pVTZ level. Bond angles in degrees, bond lengths in angstroms, total energies in au. b Barrier to linearity is 888.6
cm-1 c Dipole moment in Debye. d Rotational constants in GHz. Isotopes: H-1, C-12, N-14, F-19, S-32, Cl-35.
.
TABLE 3: Calculateda Vibrational Frequencies (cm-1) and Infrared Intensities (km/mol) of Nitrile Sulfides, XCNS (X ) H, F,
Cl, NC, and CH3)
HCNS, 1 (C∞V
)
FCNS, 2 (Cs)
ClCNS, 3 (C∞V
)
NCCNS, 4 (C∞V
)
H3CCNS, 5 (C3V)
3084 (1.8) ν6(CH3 as.str)
3478 (362.3) ν1(HC str)
2142 (327.5) ν2(CN str)
2117 (64.0) ν1(CN str)
1156 (480.4) ν2(CF str)
2286 (177.1) ν1(CN str)
957 (163.6) ν2(NS str)
2344 (575.0) ν1(str)
2201 (157.9) ν2(str)
1123 (93.4) ν3(str)
3022 (18.0) ν1(CH3 s str)
2316 (294.3) ν2(CN str)
759 (50.7)
429 (5.2)
365 (62.6)
ν3(NS str)
737 (86.0)
ν3(NS str)
495 (2.8)
ν3(ClC str)
ν4(CNS def) 428 (108.4) ν4(CNS def) 370 (0.4)
ν4(CNS def) 578 (22.8)
ν5(ClCN def) 476 (0.5)
410 (7.4)
ν4(str)
1468 (9.5)
1414 (8.6)
1043 (2.5)
ν7(CH3 as def)
ν3(CH3 s def)
ν8(CH3 rock)
ν5(HCN def) 323 (0.4)
ν6(CNS def) 79 (0.004)
ν5(FCN bend)
ν5(def)
ν6(def)
209 (54.6)
121 (4.9)
ν7(CCN def) 1039 (64.4) ν4(str)
591 (29.0)
426 (0.5)
148 (1.3)
ν5(str)
ν9(CNS def)
ν10(CCN def)
a Calculated at the B3-LYP/cc-pVTZ level; unscaled frequencies.
TABLE 4: Calculated Vertical Ionization Potentials (in eV) of Nitrile Sulfides, XCNSa
HCNS, 1
FCNS, 2
ClCNS, 3
NCCNS, 4
H3CNS, 5
MO character
8.98 (8.89) π
9.27 (9.44) a′′
9.52 (9.46) a′
13.92 (1617) a′
14.33 (16.56) a′′
8.67 (8.75) π
9.66 (9.64) π
8.37 (8.38) e
πnb(CNS)
πb(CNS)
14.86 (16.51) π
15.72 (17.12) σ
13.46 (14.85) π
13.20 (14.75) π
14.00 (15.44) e
13.97 (17.17) σ
15.74 (18.01) π
16.65 (18.18) σ
nN
π(NC)
nS (S lone pair)
nCl
CH3
CH3
16.30 (17.91) a′
15.57 (17.10) σ
15.69 (17.74) π
15.12 (16.60) a1
16.35 (18.01) e
17.35 (19.65) a1
18.75 (21.01) σ
C-Cl
a Calculated using the ROVGF/cc-pVTZ method at the B3LYP/cc-pVTZ geometries; Koopmans’ IPs are in parentheses.
tory behavior in the MPn expansion series, basis set, barrier
height, and total energy, on calculated properties of the nitrile
oxide analogues.12,13,15,22,23 The other effect, triple excitations
markedly lengthening the CtN bond length and tending to bend
the structure, were also observed for nitrile oxides. It seems
that the computational behavior of nitrile sulfides is very similar
to that of nitrile oxides, which, given the structural similarity,
is not surprising. The possibility of oscillatory behavior of the
MPn series is known; starting from MP1 in the MPn series,
MP2 exaggerates the effects of the new incoming terms, the
double excitations, MP3 underestimates these due to DD
coupling, MP4 exaggerates the new triple effects, and MP5
underestimates these due to TT coupling.24 It seems that
MP4SDTQ strongly overestimates the CtN bond lengths and
barrier height, in accord with the known fact that the method
exaggerates triples contributions to the correlation energy.25
computational challenge. Calculations on 1 in this work suggest
the same for nitrile sulfides. We note that if very high level
conventional ab initio calculations are not feasible, DFT (e.g.,
in the form of B3LYP) can provide a reliable alternative. We
have found DFT to give surprisingly good results for ground-
state geometries and vibrational frequencies of nitrile oxides.12,15
Calculations on 1 predict the same for nitrile sulfides; compare
the B3LYP and CCSD(T) results in Table 1. For this reason
we will use the B3LYP method for calculation of the geometries
of substituted nitrile sulfides and for evaluating the decomposi-
tion process. To assist in detection of 1 by high-resolution
techniques, Table 1 also includes the calculated dipole moments
and rotational constants for all levels of theory.
Table 2 shows the calculated B3LYP/cc-pVTZ geometry,
dipole moments, and rotational constants for molecules 1-5,
with the calculated IR frequencies and intensities and the
ionization potentials (IPs) listed in Tables 3 and 4, respectively.
The calculated unscaled vibrational frequency (2316 cm-1) for
ν(CN) in 5 is in very good agreement with that observed (2230
What has emerged from calculation of structural properties
of nitrile oxides is the requirement for extra large basis sets
and very high levels of theory, making nitrile oxides a serious

6262 J. Phys. Chem. A, Vol. 105, No. 25, 2001
Pasinszki et al.
cm-1) by matrix IR.3,4 1, 3, 4, and 5 have a linear XsCsNsS
frame, and only the fluoro derivative, 2, is bent with a barrier
to linearity of 889 cm-1. The similarity to the nitrile oxides
must be noted again; FCNO is predicted to have a bent structure
(barrier to linearity is 1092 cm-1 at an identical level of
theory),26 but all other substituted nitrile oxides can be described
with a linear or quasi-linear frame.14 The propensity for bending
in XCNO species increases with the increase in electronegativity
of the substituent attached to the -CNO group and the increase
in the CtN bond length shows the same trend.14,26 This can be
seen in Table 2 for nitrile sulfides as well. From a comparison
of the bond lengths of 1-5 with those of molecules having
typical single and double NsS bonds (H2NSH and HNS) and
double and triple CN bonds (H2CNH and HCN),27 the CN bond
in nitrile sulfides is found to lie close to a triple bond with the
NS bond being close to a double bond. Calculated bond orders
for CN and NS bonds are 2.6-2.9 and 1.7-2.0, respectively,
using Gordy’s rule.28
TABLE 5: Energetics (in kcal/mol) of Reactions between
HCNS and Decomposition Intermediatesa
Stability and Decomposition Mechanism of Nitrile Sul-
fides. The thermal and photochemical instability of nitrile
sulfides is well-known; photolysis at cryogenic temperatures
or allowing the sample to warm above 100 K results in
desulfuration.2,3 The decomposition, however, seems not to be
a simple unimolecular reaction. In solution, the decay rate is
strongly concentration dependent, and the yields of 1,3-dipolar
cycloaddition products are dilution dependent.29 Isolated samples
of nitrile sulfides are stable only in an inert solid matrix at low
temperature. For example, upon heating, in EPA glass PhCNS
is consumed as soon the glass melts (around 140 K), but in a
rigid PVC matrix the decomposition is gradual and traces of
the compound are detectable at room temperature for a short
period of time.30,31 Since the experimental observations are
inconsistent with a unimolecular decomposition process, it has
been suggested29 that the decomposition is accelerated by
reaction with short sulfur chains (e.g., S1-S7), i.e., by a sulfur
transfer reaction, where the sulfur is expected to be extruded
initially as an S atom. The formation of S2 is often observed in
the pyrolysis mass spectrometry studies, and mechanisms
involving dimerization or sulfur atom transfer have been
postulated.3
Although clear from the accumulated experimental data3 that
the decomposition of nitrile sulfides follows higher order
kinetics, the mechanism of this process is still not known. To
clarify these questions, calculations have been performed using
the B3LYP method. The first question considered was that of
single sulfur atom formation, i.e., the direct rupture of the N-S
bond: XCNS f XCN + S. The formation of ground-state XCN
and ground-state sulfur S(3P) in this process is spin-forbidden.
According to our CCSD(T)(full)/cc-pVTZ//B3LYP/cc-pVTZ
calculations, the spin-allowed extrusion of a singlet sulfur atom,
S(1D), is endothermic, requiring quite high energies of 58.3,
51.4, 55.8, 63.7, and 58.6 kcal/mol (∆E) for 1-5, respectively.
Even formation of S(3P) would be endothermic with energies
of 27.0, 20.2, 24.6, 32.5, and 27.5 kcal/mol for 1-5, respec-
tively. The calculations demonstrate that direct sulfur atom
formation from nitrile sulfides cannot happen at low or ambient
temperatures. As noted from experiment,3 for example, the
activation energy for unimolecular decomposition cannot be
small; otherwise, it would not be possible to matrix isolate nitrile
sulfides formed by FVP.
a Calculated at the B3LYP/6-31G** level; relative energies (∆E),
relative Gibbs free energies (∆G), and energy barriers (∆Eq, ∆Gq), are
relative to the sum of the energies (corrected with ZPE) or Gibbs free
energies of starting materials. Gibbs free energies are calculated at
298.15 K and 1 atm. Hydrogen atoms are omitted on rings for
simplicity.
between 1 and various reaction products considered including
characterization of the minima and transition states along the
reaction paths. The most important reactions and their energetics
are listed in Table 5, and the essence of the decomposition
process of 1 is summarized in Figure 1.
Potential bimolecular reactions were thus considered to
account for the thermal instability of nitrile sulfides. B3LYP/
6-31G** calculations were performed first for 1, the compu-
tationally most tractable species, with more than forty reactions
As the initiating step, two bimolecular decomposition reac-
tions are found for 1, one via S-S and the other via C-C. In
the first, which is kinetically less favored, two molecules of 1
produce two molecules of nitrile (HCN) and singlet S2 directly

Structure and Stability of Small Nitrile Sulfides
J. Phys. Chem. A, Vol. 105, No. 25, 2001 6263
in the desulfuration process of 1. Thiofuroxan, like 1, does not
lose sulfur in a monomolecular process (R5) and should be stable
under isolated conditions, i.e., in the dilute gas phase or in an
inert solid matrix. In this context, a NCCNS dimer was seen in
the mass spectroscopy work.6 Thiofuroxan loss can, however,
occur in bimolecular processes by reaction with another
molecule of thiofuroxan (R6, R7) producing singlet S2 and
thiadiazole or with 1 (R8, R7) producing singlet S2, thiadiazole,
and HCN with barriers of 3.1 and 7.0 kcal/mol, respectively.
These reactions involve the intermediacy of a kinetically very
unstable sulfurated thiofuroxan (R7). S2 is a key intermediate
in the decomposition of 1. Either It reacts with other short sulfur
chains without any kinetic energy barrier to form S4, S6, S8,
etc., or more importantly, it reacts with 1 without any barrier,
forming a five-membered ring cycloadduct (R9). Other short
sulfur chains, like S4, react with 1 similarly and form cycload-
ducts; in the case of S4 a seven-membered ring containing five
sulfur atoms forms (R10). Sulfur chains Sn, longer than S2, can
also react with 1 with the formation of HCN and Sn+1 (e.g.
R11), but this is kinetically less favored than the cycloaddition
(R10). The sulfur-containing rings formed, e.g., in reactions R9
and R10 are good sulfur scavengers, remove sulfur from 1 (R12)
or from thiofuroxan (R13) in two steps, and grow into a larger
sulfur-containing ring (see R12 f R14) or react with S2 to form
the same larger ring (see R15 f R16). We have also investigated
whether the stable reaction products thiadiazole and HCN could
react with any component of the complicated reaction mixture,
particularly since 1 is a good 1,3-dipolarophile and HCN is a
1,2-dipolarophile. The reaction between 1 and HCN is exother-
mic but has a kinetic barrier of 16.5 kcal/mol (R17); therefore
the reaction is not expected to proceed at low temperatures,
but at room or higher temperatures this could contribute to an
increase in the yield of thiadiazole product. The reaction of HCN
with sulfur chains is endothermic, thus thermodynamically not
favored (e.g., R18). The reaction between thiadiazole and 1, a
possible sulfur transfer, is also endothermic and not favored
(R19, R20).
Figure 1. Decomposition of HCNS (1) (kinetic energy barriers for
reactions, ∆E‡, are shown in kcal/mol).
Summarizing the decomposition mechanism of 1, calculations
show that the initiating steps are, consecutively, the cycload-
dition of 1 and the decomposition of thiofuroxan with the
formation of S2. Short sulfur chains effectively remove sulfur
from 1 and grow larger and larger in the cycloaddition and sulfur
transfer processes discussed above and produce the correspond-
ing nitrile (HCN). Large rings containing one HCN unit could
possibly eliminate HCN at the end, but this could not be
investigated due to computational limitations. It is interesting
to note that the calculations suggest that besides sulfur and nitrile
(HCN), thiadiazole also forms in the decomposition of nitrile
sulfide. This latter product has never been considered in previous
work.
Figure 2. Bimolecular decomposition of nitrile sulfides, XCNS (X )
H (1), F (2), Cl (3), CN (4), CH3 (5)), in a “tail to tail” reaction. Energies
(∆E, in kcal/mol) are relative to that of the sum of total energies of
two XCNS molecules (ZPE included).
in a “tail to tail” (HCNS...SNCH) reaction after passing over a
kinetic energy barrier of 17.9 kcal/mol (∆E‡; ∆G‡ ) 25.2 kcal/
mol at 273.15 K and 1 atm; reaction number R2, Table 5; see
the TS structure in Figure 2). There is no intermediate in this
reaction, and the two N-S bonds simultaneously break,
releasing S2 and the nitrile. Thus, this is not a simple sulfur
atom transfer, as was suggested earlier,29 and calculations show
that a suggested intermediate such as HCNS2 does not exist on
the singlet potential energy surface. In the second route, two
molecules of 1 dimerize into 2-sulfo-1,2,5-thiadiazole (thio-
furoxan) in a typical Firestone-type [2+3] cycloaddition reaction
(R3, R4). The kinetic energy barrier of the first step (TS1)
leading to the Firestone intermediate (IM) is a mere 2.0 kcal/
mol, which clearly explains the instability of 1 if molecules
interact with each other, e.g., in the condensed phase, and may
explain our inability to observe nitrile sulfides under our present
experimental conditions. This dimerization is very similar to
the known dimerization of nitrile oxides leading to stable
furoxans,12,13,26 with the major difference, as we show below,
that thiofuroxan can readily react further and can also take part
The decomposition of 1 is a complicated process and follows
higher order kinetics. It is an interesting question how substit-
uents on the nitrile sulfide group influence this. Because of the
large number of potential reaction routes, it was not possible to
calculate decomposition mechanisms for 2-5 in detail, similar
to those described for 1, but the key first steps, cycloaddition
or “tail to tail” reactions between nitrile sulfides, is discussed
here. Results are shown in Figures 2 and 3 with calculated
reaction heats and kinetic energy barriers listed in Table 6. As
the results show, the cycloaddition (thiofuroxan route, Figure
3) for nitrile sulfides is both kinetically and thermodynamically
much more favored than the direct singlet S2
loss in the “tail to
tail” reaction (Figure 2). There is no energy barrier for the
cycloaddition of 2 and 3, the barrier is a mere 2.0 kcal/mol for

6264 J. Phys. Chem. A, Vol. 105, No. 25, 2001
Pasinszki et al.
TABLE 6: Calculated Reaction Heat (∆E, ∆G) and
Reaction Barriers (∆E‡, ∆G‡) for the Decomposition of
Nitrile Sulfides (XCNS, X ) H, F, Cl, CN, and CH3)a
spectroscopic investigations by thermolyzing 1,2,5-thiadiazoles,
but in all cases the thiadiazoles were found to produce sulfur
and the corresponding nitrile. This prompted a theoretical
investigation for the structures, stabilities, and decomposition
mechanisms of nitrile sulfides. Calculations indicate that the
molecules have linear heavy atom geometries, the exception is
the fluoro derivative, which is bent with a calculated barrier to
linearity of 889 cm-1. Calculations on the nitrile sulfide
structures are sensitive to electron correlation effects, and the
description of these latter effects is of crucial importance.
Nonetheless, DFT can be a cost-effective approach for such
molecules, and this approach was used to investigate decom-
position processes. Nitrile sulfides are unstable in the condensed
phase due to bimolecular reactions involving various sulfur
transfer and cyclization processes between decomposition
intermediates. Activation energy barriers of such reactions are
small, below 10 kcal/mol, depending on the substituent, predict-
ing a short lifetime for nitrile sulfides at ambient temperature.
Dimerization to the unstable thiofuroxan is one of the major
loss processes, and the instability and reactivity of thiofuroxan
is responsible for the complicated decomposition mechanism.
We note that in the case of nitrile oxides the dimerization to
the stable furoxan derivatives are the dominant loss pro-
cesses.12,13,22 Nitrile sulfides have long been known from
experiment to be stable to unimolecular decomposition,3 and
matrix IR and NRMS experiments have supported this. The
present calculations are in agreement, suggesting that the
generation of nitrile sulfides into the dilute gas phase for detailed
spectroscopic and structural observation is still feasible, although
remaining an experimental challenge.
molecule
∆E
∆G
∆E‡
∆G‡
2XCNS f 2XCN + S2
1
2
3
4
5
-8.7
-17.7
-30.4
-22.3
-5.4
17.9
9.6
14.3
24.3
15.3
25.2
18.7
23.4
30.9
21.3
-22.9
-14.7
+3.4
-9.5
-18.8
1
2
3
4
5
-41.9
-79.5
-54.2
-30.1
-34.9
-31.5
-66.5
-41.0
-19.1
-23.0
2.0
9.6
12.6
9.0
21.1
18.4
a Calculated at the B3LYP/6-31G** level; in kcal/mol.
Acknowledgment. We thank the Hungarian Scientific
Research Fund (OTKA Grant F022031) and the Natural
Sciences and Engineering Research Council of Canada (NSERC)
for research and equipment grants in support of this work. T.P.
thanks the Hungarian Soros Foundation for a travel grant and
the Hungarian Academy of Sciences for the award of a Ja´nos
Bolyai Research Scholarship.
References and Notes
Figure 3. Cycloaddition of nitrile sulfides, XCNS (X ) H (1), F (2),
Cl (3), CN (4), CH3 (5)). Energies (∆E, in kcal/mol) are relative to
that of the sum of total energies of two XCNS molecules (ZPE
included). In the case of the energetically close CN and CH3 substituted
molecules, the lower barrier for TS1 corresponds to -CH3.
(1) Franz, J. E.; Black, L. L. Tetrahedron Lett. 1970, 1381.
(2) Paton, R. M. Chem. Soc. ReV. 1989, 18, 33.
(3) Wentrup, C.; Kambouris, P. Chem. ReV. 1991, 91, 363.
(4) Kambouris, P.; Plisnier, M.; Flammang, R.; Terlouw, J. K.;
Wentrup, C. Tetrahedron Lett. 1991, 32, 1487.
(5) Maquestiau, A.; Flammang, R.; Plisnier, M.; Wentrup, C.; Kam-
bouris, P.; Paton, R. M.; Terlouw, J. K. Int. J. Mass Spectrom. Ion. Processes
1990, 100, 477.
(6) Flammang, R.; Gerbaux, P.; Morkved, E. H.; Wong, M. W.;
Wentrup. C. J. Phys. Chem. 1996, 100, 17452.
(7) Gerbaux, P.; Haverbeke, Y. V.; Flammang, R.; Wong, M. W.;
Wentrup. C. J. Phys. Chem. A 1997, 101, 6970.
(8) Gerbaux, P.; Flammang, R.; Morkved, E. H.; Wong, M. W.;
Wentrup. C. Tetrahedron Lett. 1998, 39, 533.
(9) Gerbaux, P.; Flammang, R.; Morkved, E. H.; Wong, M. W.;
Wentrup. C. J. Phys. Chem. A 1998, 102, 9021.
(10) Bak, B.; Christiansen, J. J.; Nielsen, O. J.; Svanholt, H. Acta Chem.
Scand. A 1977, 31, 666.
(11) Fehe´r, M.; Pasinszki, T.; Veszpre´mi, T. Inorg. Chem. 1995, 34,
945.
(12) Pasinszki, T.; Westwood, N. P. C. J. Phys. Chem. A 1998, 102,
4939.
(13) Pasinszki, T.; Westwood, N. P. C. J. Phys. Chem. A 2001, 105,
1244.
(14) Pasinszki, T.; Westwood, N. P. C. J. Mol. Struct. 1997, 408-409,
161.
(15) Pasinszki, T.; Westwood, N. P. C. J. Phys. Chem. 1996, 100, 16856.
(16) Geisel, M.; Mews, R. Chem. Ber. 1982, 115, 2135.
(17) Warren, J. D.; Lee, V. J.; Angier, R. B. J. Heterocycl. Chem. 1979,
16, 1617.
1, and it is 12.6 and 9.0 kcal/mol for 4 and 5, respectively,
which suggests possible dimerization at less than room tem-
perature, especially for 1-3, and would explain the instability
of nitrile sulfides. It is interesting to note that the kinetic barrier
to dimerization decreases with an increasing electronegativity
of the substituents, i.e., in the order CH3 > H > Cl > F. The
cyano-substituted derivative, 4, seems to be an exception in this
series and is predicted to be the less unstable. The completely
delocalized electronic structure of 4 might be an explanation
for this, but this has not been investigated in this work. Such
relative enhanced stability has been noted for NCCNO14,32 when
compared to the other nitrile oxides. Calculations suggest that
attaching a strong electron donating, sterically large, possibly
aromatic substituent to the nitrile sulfide group increase the
kinetic stability.
Conclusion
This work has sought to generate nitrile sulfides, XCNS
(where X ) H, F, Cl, CN, CH3), in the gas phase for

Structure and Stability of Small Nitrile Sulfides
J. Phys. Chem. A, Vol. 105, No. 25, 2001 6265
(18) Weinstock, L. M.; Davis, P.; Handelsman, B.; Tull, R. J. Org. Chem.
1967, 32, 2823.
(19) Muhlbauer, E.; Weiss, W. (Farbenfabriken Bayer, A.-G.) Brit. Pat.
1,079,348, 1967; Chem. Abstr. 1968, 68, P69000w.
(20) Pasinszki, T.; Westwood, N. P. C. J. Electron Spectrosc. Relat.
Phenom. 2000, 108, 63.
(21) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,
M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.;
Stratmann, R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A.
D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi,
M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.;
Ochterski, J.; Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick,
D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.;
Ortiz, J. V.; Baboul, A. G.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz,
P.; Komaromi, I.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-
Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe,
M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Andres, J. L.;
Gonzalez, C.; Head-Gordon, M.; Replogle, E. S.; Pople, J. A. Gaussian
98, Revision A.7; Gaussian, Inc.: Pittsburgh, PA, 1998.
(22) Pasinszki, T.; Westwood, N. P. C. J. Phys. Chem. 1995, 99, 6401.
(23) Pasinszki, T.; Westwood, N. P. C. J. Am. Chem. Soc. 1995, 117,
8425.
(24) He, Z.; Cremer, D. Int. J. Quantum Chem., Quantum Chem. Symp.
1991, 25, 43. He, Z.; Cremer, D. Theor. Chim. Acta 1993, 85, 305.
(25) Pople, J. A.; Head-Gordon, M.; Raghavachari, K. Int. J. Quantum
Chem., Quantum Chem. Symp. 1988, 22, 377. Kraka, E.; Gauss, J.; Cremer,
D. J. Mol. Struct. (THEOCHEM) 1991, 80, 95.
(26) Pasinszki, T.; Westwood, N. P. C. Manuscript in preparation.
(27) Calculated CN and NS bond lengths (B3LYP/cc-pVTZ): HCtN,
1.146 Å; H2CdNH, 1.263 Å; HNdS, 1.572 Å; H2N-SH, 1.734 Å.
(28) Gordy, W. J. Chem. Phys. 1947, 15, 305.
(29) Howe, R. K.; Shelton, B. R. J. Org. Chem. 1981, 46, 771.
(30) Harrit, N.; Holm, A.; Dunkin, I. R.; Poliakoff, M.; Turner, J. J. J.
Chem. Soc., Perkin Trans II 1987, 1227.
(31) Holm, A.; Harrit, N.; Toubro, N. H. J. Am. Chem. Soc. 1975, 97,
6197.
(32) Brupbacher, Th.; Bohn, R. K.; Jager, W.; Gerry, M. C. L.; Pasinszki,
T.; Westwood, N. P. C. J. Mol. Spectrosc. 1997, 181, 316.