An experimental and theoretical study of NSCl decomposition in the presence of trace amounts of water

Article abstract of DOI:10.1021/jp802445r

An experimental and density functional theory study of the decomposition reaction of NSCl with water molecule(s) into NH₃ and SO₂ is described. Experiments involving the decomposition of NSCl in the presence of trace amounts of water

Full text of DOI:10.1021/jp802445r

J. Phys. Chem. A 2008, 112, 8561–8568
8561
An Experimental and Theoretical Study of NSCl Decomposition in the Presence of Trace
Amounts of Water
†
‡
,†
,§
Zhen Guo, Cunyuan Zhao, David Lee Phillips,* Evan G. Robertson,* and
,
§
Don McNaughton*
Department of Chemistry, The UniVersity of Hong Kong, Pokfulam Road, Hong Kong, P. R. China, School of
Chemistry and Chemical Engineering, Sun Yat-Sen UniVersity, Guangzhou 510275, P.R. China, and School of
Chemistry, Monash UniVersity, Victoria, 3800, Australia
ReceiVed: March 20, 2008; ReVised Manuscript ReceiVed: June 12, 2008
An experimental and density functional theory study of the decomposition reaction of NSCl with water
molecule(s) into NH
of trace amounts of water showed that the formation of both HNSO and SO
cell surface are fresh and more likely to be coated in adsorbed water. The density functional theory computational
results suggest that the decomposition involves two main steps: reactions 1 NSCl + nH O f HNSO + (n - 1)
O + HCl and 2 HNSO + nH O f NH + SO + (n - 1)H O. The barrier for reaction 1 in which NSCl
3
and SO
2
is described. Experiments involving the decomposition of NSCl in the presence
2
is faster when the cell and the
2
H
2
2
3
2
2
decomposes to form HNSO becomes substantially lower when the explicit hydrogen bonding of water molecules
is considered. The water assisted decomposition reaction mechanism can help account for the experimentally
observed conversion of NSCl to produce HNSO in the presence of water. The calculated barrier of reaction
2
(the decomposition of HNSO into NH
3 2 2
and SO ) can also explain the experimental observation of SO
bands from the decomposition of NSCl in the presence of water.
Introduction
co-workers failed to obtain the production of NSCN from NSCl
vapor and these bands at 2224 cm and 1337 cm were
-
1
-1
The properties, chemical reactions, and spectroscopy of
sulfur-nitrogen compounds have received much experimental
and theoretical attention due to their involvement in reactions
of interest in synthetic chemistry and in the environment. These
thionitroso compounds (especially those with nitrogen-sulfur
double bonds) are known to be very unstable and highly
assigned to N2O and SO2 using high resolution FTIR spectro-
12,13
scopy.
They also observed bands associated with HNSO and
SO2 from the hydrolysis of NSCl when trace quantities of water
were present and observed the half-life for conversion of NSCl
to HNSO to be a few minutes when the pyrolysis products were
isolated in a cell and monitored by rapid survey scan experi-
1
–3
reactive. A number of infrared spectroscopic investigations
have been carried out on two thionitroso containing molecules
12
ments. Here, additional experimental results are given for the
decomposition of NSCl in the presence of water and these results
indicate that the formation of both HNSO and SO2 is more rapid
when the cell and the cell surface are fresh and more likely to
be coated in adsorbed water. We also report density functional
theory (DFT) and second-order Moller-Plesset perturbation
theory (MP2) calculations to investigate the decomposition of
NSCl in the presence of trace amounts of water molecules. We
aim to elucidate reaction mechanism(s) that can account for a
relatively easy conversion of NSCl in the presence of water to
HNSO observed experimentally to take a few minutes of time
and to elucidate the reaction mechanisms involved in the
decomposition of HNSO by trace amounts of water to form a
SO2 byproduct.
4
–9
using matrix isolation and low resolution gas phase techniques.
An ab initio molecular orbital study performed at the QCISD
(
(
T):6-311G (3df, 2p) level of theory with correction of ZPE
zero point energy) by Nguyen and Flammang showed that
1
0
neutral NSCl is 77 kJ/mol more stable than ClsNdS. In
addition, the product of NSCl in the presence of trace amounts
of water, HNSO, has a nonlinear functional group -NdSdO
that has gained interest due to its geometry. A recent theoretical
study predicted that on the potential energy surface of the HNSO
system, cis-HNSO will be at the global minimum and fourteen
isomers are kinetically stable with five of these isomers being
1
1
experimentally suggested in the gas phase. More recently,
Robertson, McNaughton, and co-workers
12,13
measured the high
Computational and Experimental Details
resolution FTIR spectra of NSCl and HNSO in the attempted
generation of NSCN and NSSCN using NSCl as a precursor.
Previously in a low resolution spectroscopy study the bands at
Thiazyl chloride (NSCl), was prepared by heating a sample
of thiodiazyl dichloride (S3N2Cl2) to a temperature of 70 °C.
The NSCl(g) evolved was passed through a White cell (8 passes
-1
-1
14
2
224 cm and 1374 cm were mistakenly assigned to NSCN.
Recent experimental studies by Robertson, McNaughton and
×
0.53 m ) optical path 4.24 m), for high FTIR analysis. For
high resolution spectroscopy the cell was slowly pumped to
prevent buildup of hydrolysis products and to ensure a steady
flow of sample, with the pressure maintained at ca. 30 Pa. To
ascertain the half-life of the NSCl in the cell and to test the
conditions for optimum production, spectra were recorded at
time intervals with the products isolated in the cell and under
flowthrough conditions over 5 h of flow. Spectra were recorded
*
Authors to whom correspondence should be addressed. Telephone: 852-
2
8
859-2160(DLP); +61 3 99054566(EGR); +61 3 99054525(DM). FAX:
52-2857-1586 (DLP); +61 3 9905 4597 (EGR & DM). E-mail: phillips@
hkucc.hku.hk; evan.robertson@sci.monash.edu.au; Don.McNaughton@
sci.monash.edu.au.
†
The University of Hong Kong.
Sun Yat-Sen University.
Monash University.
‡
§
1
0.1021/jp802445r CCC: $40.75  2008 American Chemical Society
Published on Web 08/20/2008

8
562 J. Phys. Chem. A, Vol. 112, No. 37, 2008
Guo et al.
HNSO + nH O f NH +SO +(n - 1)H O (2)
SCHEME 1
2
3
2
2
We have considered the effect of water molecules on the
reactions by explicitly adding water molecules one by one into
the reactant complexes (or water-solvated clusters). The clusters
consist of reactants (NSCl) or HNSO molecules and one, two,
and three water molecules for reactions 1-1 and 2-1. The
reactions 1-2 and 2-2 were only considered with zero, one, or
two water molecules.
All of the reactions have been investigated by optimizing the
reactant complexes (RC), transition states (TS), and product
complexes (PC). These are denoted as (RC)ijn, (TS)ijn, and (PC)ijn
in the following text, tables, and figures, where i represents the
number of the reaction, j represents the step of the reaction,
and n is the number of water molecules in the reaction. For
example, (RC)113 represents the reactant complex of the first
step of the first reaction with three water molecules, i.e., the
reaction complex of NSCl + 3H2O f HNSCl(OH) + 2H2O.
Stationary structures for all of the reactant complexes and
transition states and product complexes were fully optimized
without a symmetry constraint (e.g., C1 symmetry) at the
B3LYP/aug-cc-pvdz level of theory. Analytical frequency
calculations were performed in order to confirm the optimized
structure to be either a minimum or a first order saddle point.
IRC calculations were performed for all of the reaction systems
to confirm the optimized transition state correctly connects the
relevant reactants and products. MP2 calculations were also used
for reaction 1-1 using the aug-cc-pvdz basis set. Previous work
indicates that it is probably necessary to use a basis set including
both polarization and diffuse functions for light and heavy atoms
in order to obtain a better description of the charge distributions
for the charge separated species found in these HCl elimination
on a Bruker HR120 spectrometer equipped with a liquid nitrogen
cooled MCT detector and KBr beamsplitter.
15
Density functional theory (DFT) and second-order Moller-
Plesset perturbation theory (MP2) computations were employed
to investigate the decomposition of thiazyl chloride (NSCl) in
the presence of water molecules. We have considered two major
types of reactions (see Scheme 1). The first reaction type
(
reaction 1) involves two possible reaction pathways. One is
the addition of water to NSCl to produce HNSCl(OH) in water
reaction 1-1-A) followed by the elimination of hydrogen
chloride (HCl) in water to produce the simplest sulfinylimide
HNSO) (reaction (1-2-A) as shown below:
(
(
NSCl + nH O f HNSCl(OH) + (n - 1)H O (n ) 1, 2, 3)
2
2
(1-1-A)
HNSCl(OH) + (n - 1)H O f HNSO + (n - 1)H O + HCl
2
2
17b
reactions.
(1-2-A)
Results and Discussion
These may be added together to obtain the first overall
reaction 1 given below:
A. Low Resolution and High Resolution FT-IR Spectra
of the Decomposition of NSCl. Figure 1 shows the spectra of
NSCl in the multipass cell first under flow through conditions
where NSCl predominates in the presence of HNSO and second
upon isolation of the cell and standing for 3 min. After 3 min,
SO2 is now apparent and HNSO signals have increased markedly
at the expense of NSCl. From these experiments it is apparent
that the lifetime of NSCl under static conditions in the glass
cell is on the order of a few minutes. Figure 2 shows high
resolution spectra at the start and finish of a 5 h flowthrough
experiment. After 5 h, the intensities of signals from HNSO
and SO2 have decreased markedly with SO2 showing the greatest
decrease. These spectra indicate that the production of both
NSCl + nH O f HNSO + (n - 1)H O + HCl (1-A)
2
2
The other is the decomposition of NSCl in trace moisture
e.g., in the presence of one or a few water molecules) to
hydrogen chloride (HCl) to produce the isomer (NSOH) of
HNSO (reaction 1-1-B), and then NSOH transformation to
HNSO via an H transfer assisted by water molecule(s) (reaction
(
1
-2-B as shown below:
NSCl + nH O f NSOH + (n - 1)H O + HCl
2
2
(
1-1-B)
(1-2-B)
If we combine the previous two equations, we have
NSCl + nH O f HNSO + (n - 1)H O + HCl (1-B)
NSOH + nH O f HNSO + nH O
HNSO and SO is more rapid when the cell and the cell surface
2
2
2
are fresh and more likely to be coated in adsorbed water. Hence,
we surmise that the decomposition of NSCl is assisted by the
presence of water through interaction with surface adsorbed
2
2
2
water on the ca. 0.25 m borosilicate glass surface and gaseous
The second type of reaction is the addition reaction of HNSO
water desorbed from the surface.
with a water molecule to produce H2NSO(OH) (reaction 2-1)
followed by decomposition of H2NSO(OH) into SO2 and NH3
in water (reaction 2-2) as shown below:
B. Reaction 1: Reactions of NSCl with Water to Produce
the Simplest Sulfinylimide HNSO. It has been experimentally
shown that HNSO was generated by the reaction of thiazyl
chloride (NSCl) with trace moisture in a cell. The conversion
of NSCl with water present to HNSO has a half-life of a
few minutes. Two possible pathways were examined for the
conversion reaction of NSCl to HNSO with water molecules
present. The first pathway is the one where NSCl in a trace
amount of moisture first undergoes water addition to produce
the intermediate HNSCl(OH). Then, the simplest sulfinylim-
ide, HNSO, is formed via the decomposition reaction of
HNSO + nH O f H NSO(OH) + (n - 1)H O
2
2
2
(
n ) 1, 2, 3) (2-1)
H NSO(OH) + (n - 1)H O f NH +SO +(n - 1)H O
2
2
3
2
2
(2-2)
These two reactions may be added together to obtain the
second overall reaction 2 shown below:

Decomposition Reaction of NSCl
J. Phys. Chem. A, Vol. 112, No. 37, 2008 8563
Figure 1. Low resolution (0.4 cm-1) FTIR spectra of NSCl and related
byproduct (HNSO and SO
length 4.2 m. Spectra were measured under flowthrough conditions
upper trace), and then 3 min later after the cell had been isolated.
2
), with total cell pressure 18 Pa, and path
(
Figure 3. Optimized geometries (bond lengths are in Å) for all the
reactants, reactant complexes, transition states, and product complexes
obtained from the B3LYP/aug-cc-pvdz calculations are shown for the
NSCl + nH
associated with structures (RC)11n, (TS)11n, and (PC)11n where n ) 1,
, 3).
2
O f HNSCl(OH) + (n - 1)H O (n ) 1, 2, 3) reactions
2
(
2
TABLE 1: NBO Charges on the Selected Atoms for the
Reaction NSCl + nH
2
2
O f HNSOCl(OH) + (n - 1)H O
Figure 2. High resolution (0.003 cm-1) FTIR spectra of NSCl and
related byproduct measured under flowthrough conditions, with total
cell pressure 30 Pa and path length 4.2 m. The upper spectrum was
measured first, while the lower spectrum was measured ca. 5 h later
when the amount of water in the vacuum system was depleted.
(
where n ) 1, 2, 3)
Species O1
H1
N
S
(
(
(
(
(
(
RC)111
TS)111
PC)111
RC)112
TS)112
PC)112
-0.972
-0.882
-0.882
-0.994
-0.948
-0.906
-0.970
-0.972
-0.928
0.493
0.483
0.406
0.504
0.499
0.434
0.513
0.517
0.443
-0.567
-0.872
-0.947
-0.652
-0.909
-0.938
-0.697
-0.909
-0.932
0.970
1.170
1.315
0.993
1.138
1.301
1.030
1.170
1.294
HNSCl(OH) + (n - 1)H2O f HNSO + (n - 1)(H2O) +
HCl (where n ) 1, 2, 3). The second type of pathway is the
decomposition of NSCl in trace moisture (e.g., in the presence
of one or a few water molecules) to hydrogen chloride (HCl)
and the isomer (NSOH) of HNSO, and then conversion of
NSOH to HNSO via an H transfer assisted by water
molecule(s).
(RC)113
(TS)113
(
PC)113
large reaction barrier, the corresponding small rate constant, and
the long lifetime predicted for the addition reaction of NSCl +
H2O f HNSCl(OH) suggests the reaction is unlikely to proceed.
However, incorporation of an additional water molecule into
the reaction dramatically catalyzes the reaction. As seen from
Table 2 and Figure 4, the reaction barrier of (RC)112 f (TS)112
1
-A: Addition of Water Molecule to NSCl and then
Decomposition of HNSCl(OH) into HNSO. 1-1-A: NSCl +
nH2O f HNSCl(OH) + (n - 1)H2O (where n ) 1, 2, 3). A
DFT investigation of the water assisted addition of NSCl species
as a function of the number of water molecules (up to 3) is
presented here. The effect of water molecules on the reaction
was considered by explicitly adding water molecules one by
one into the reactant complex (or water-solvated clusters). Figure
3
depicts simple schematic diagrams of the optimized structures
obtained for the (RC)11n, (TS)11n, and (PC)11n stationary points
for the addition reaction that involves one to three water
molecules. We also carried out a natural bond order analysis
(
NBO) to estimate the charges on the atoms of stationary
structures reported here, and selected results are given in Table
. Figure 4 displays the relative energy profiles (in kcal/mol)
1
obtained from the B3LYP/aug-cc-pvdz DFT calculations for the
NSCl + nH2O f HNSCl(OH) + (n - 1)H2O (where n ) 1, 2,
3
) addition reactions.
Inspection of Figure 4 shows that the reaction barriers for
the NSCl + nH2O f HNSCl(OH) + (n - 1)H2O (where n )
, 2, 3) addition reaction are about 48.6 kcal/mol for n ) 1,
5.7 kcal/mol for n ) 2, and 18.5 kcal/mol for n ) 3. The very
Figure 4. Schematic diagram (not to scale) of the activation free energy
profiles (in kcal/mol) obtained from the B3LYP/aug-cc-pvdz DFT
calculations for the (RC)11n f (TS)11n f (PC)11n (where n ) 1, 2, 3)
conversion reactions. See text for more details.
1
2

8
564 J. Phys. Chem. A, Vol. 112, No. 37, 2008
Guo et al.
TABLE 2: B3LYP/aug-cc-pvdz Calculated Energies (∆E)
the N, S, and O1 atoms whose NBO charges go from -0.567,
(
kcal/mol), Enthalpy (∆H) (kcal/mol), Free Energies (∆G)
kcal/mol), Entropy (∆S) (kcal/mol), Rate Constant (k, s 1),
0
.970, and -0.972 in (RC)111 to -0.872, 1.170, and -0.882 in
-
(
(
TS)111.
and Half-Life (t1/2, s) for Reaction 1 and Reaction 2
Reaction ∆E a ∆H ∆G kb
∆S
Inspection of structures and charges for the (RC)11n f (TS)11n
c
t
1/2
f (PC)11n (where n ) 1, 2, 3) addition reaction reveals some
trends that appear to correlate with a decrease in the reaction
barrier height. For example, the N-H1 bond lengths in the
reactants go from 4.33 Å in (RC)111 to 2.33 Å in (RC)112 and
1
1
1
1
11-113
-
-
-
23
7
22
5
11
46.4 44.9 48.6 -0.0124 1.61 × 10
23.2 21.4 25.7 -0.0144 9.76 × 10
15.8 13.8 18.5 -0.0158 1.75 × 10
4.3 × 10
7.1 × 10
3.96
12
13
1
2
.04 Å in (RC)113, and the O1-S bond lengths go from 2.87 Å
1
1
1
1
20-122
in (RC)111 to 2.84 Å in (RC)112 and 2.57 Å in (RC)113. It is
interesting to note that the trends in N-H1 and O-S bond
lengths suggest that as the charges of the reaction system in
6
-7
20
9.1 8.8 8.9 -0.0003 1.90 × 10
3.6 × 10
5.7 × 10
3.4 × 10
1
1
0
-11
-11
21
22
3
2.2 3.7 -0.005 1.21 × 10
0
2.9 1.9 3.4 -0.005 2.02 × 10
(RC)11n become more shared the barrier for the addition reaction
2
2
2
2
11-213
becomes lower. This also correlates with the changes in the
N-H1 bond lengths as one goes from (RC)11n to (TS)11n where
the differences are about 3.0 Å from (RC)111 to (TS)111, 0.94 Å
from (RC)112 to (TS)112, and 0.53 Å from (RC)113 to (TS)113.
Similar trends are also seen in the charges of the H1 atom. The
preceding trends in the structures of N-H1 and O1-S bonds
and the charges of H1 and O1 atoms associated with going from
the (RC)11n to their respective (TS)11n suggest less energy is
needed to change the structure and charge distribution of the
main addition coordinate to approach the transition state as the
number of water molecules increases from one to three in
the reaction system. This is consistent with the barrier height
decreasing from 48.6 kcal/mol for n ) 1 to 25.7 kcal/mol for
n ) 2 and to 18.5 kcal/mol for n ) 3 for the (RC)11n f (TS)11n
f (PC)11n (where n ) 1, 2, 3) addition reactions. We note
similar trends for a number of water assisted dehalogenation
reactions involving isopolyhalomethanes, halogenated metha-
-
-
-
14
4
13
3
11
31.7 30.3 35.8 -0.0184 3.67 × 10
19.3 17.3 22.6 -0.0178 1.75 × 10
17.2 15.1 20.6 -0.0184 5.13 × 10
1.9 × 10
4.0 × 10
1.4 × 10
12
13
3
2
2
2
2
20-221
-
4
3
20
21.6 21.2 22
-0.0027 4.81 × 10
1.5 × 10
5
-4
21
8.9
8
9.8 -0.006 4.11 × 10
1.7 × 10
a
∆
E, ∆H, ∆G, ∆S: Energies of activation with zero point
energy, enthalpies, free energies, and entropies of activation for the
reaction. k: the rate constant is calculated from the activated free
energy by
calculated by t1/2 ) ln 2/k.
b
c
k
)
(k
B
T/h) exp(-∆G/RT).
1/2
t : the half-life is
f (PC)112 decreases to about half that for the reaction involving
only one water molecule. The rate constant is increased by 16
orders of magnitude while the half-life is shortened by 16 orders
of magnitude. The significant catalytic effect of the second water
molecule can be accounted for by formation of a hydrogen bond
between the NSCl molecule and the water hydrogen (H) atom
16–18
nols, and halogenated formaldehyde molecules
(i.e., smaller
changes in the key structural features and atom charges as one
goes from the reactant complexes to their corresponding
transition states when the number of water molecules increases
in the water solvated reaction system and a correlation with
lower reaction barrier as the number of water molecules
increases).
(
such as the 2.33 Å distance for the N-H bond in (RC)111.
Incorporation of three water molecules reduces the reaction
barrier of the Gibbs free energy by another 7.2 kcal/mol. Both
the rate constant and half-life are changed by another 6 orders
of magnitude. This suggests that the hydrogen bonding of the
additional water molecules to the NSCl ·(H2O) reactant complex
greatly assists the addition process.
1
-2-A: HNSCl(OH) + nH2O ) HNSO + HCl + nH2O (where
n ) 0, 1, 2). Figure 5 displays the optimized geometries found
In order to better understand this effect, it is useful to compare
some structural and charge changes that occur upon going from
the reactions that have one water molecule to that where three
water molecules are explicitly added to the reaction system.
Compared to the reaction system with one water molecule, when
the second water molecule is explicitly added to the reaction
system there are substantial changes in the structures of the
complexes (RC)11n, (TS)11n, and (PC)11n (where n ) 1, 2). For
instance, the N-H1 bond becomes dramatically stronger and
changes its length from 4.34 Å in (RC)111 to 2.33 Å in (RC)112.
As the addition reaction proceeds from the reactants to reach
the transition states, (TS)111, larger changes in both the structures
and atom charges can be observed for the transition state
compared to its corresponding reactant complex. For example,
the N-H1 bond becomes much stronger and goes from 4.34 Å
in (RC)111 to 1.33 Å (TS)111. These large changes are also
accompanied by moderate changes in the O-S bond that
becomes stronger and goes from 2.87 Å in (RC)111 to 1.98 Å
in (TS)111. Meanwhile, the N-S bond becomes somewhat
weaker and goes from a bond length of 1.47 Å in (RC)111 to
for the reactions of HNSCl(OH) + (n - 1)H O f HNSO +
2
HCl + (n - 1)H O, i.e., (RC) f (TS)12n f (PC)12n (where
2
12n
n ) 0, 1, 2). Examination of Figure 5 shows that the gas
homogeneous decomposition reaction of HNSCl(OH) f HNSO
+ HCl proceeds through the four centered cyclic transition state
(TS)121 where the Cl atom is almost detached from the sulfur
atom (the S-Cl bond elongates from 2.32 Å to 2.85 Å) and
approaches the hydrogen atom in the HNSOCl(OH) H-O bond
(the H-Cl distance goes from 2.51 Å in (RC)120 to 1.64 Å in
(TS)120 and the HCl molecule is partially formed in the (TS)120).
The four-centered transition state is similar to that proposed by
1
9
Wallington and co-workers in the gas-homogeneous decom-
position reaction of CCl (OH) f HCl + Cl CO and also similar
3
2
to our previous study of the reaction of CH2Cl(OH) f HCl +
17b
HClO. In the one and two water-catalytic reactions, the water
molecules are incorporated into the leaving group H· · ·Cl
through cyclic hydrogen bonds for (RC)12n, (TS)12n, and (PC)12n
(n ) 1, 2). It is worth noting that there exist systematic changes
in going from the reactant complex to the corresponding
transition states and product complexes for (RC)12n f (TS)12n
f (PC)12n where n ) 0, 1, 2. For instance, the H1-O and the
S-Cl bonds become systematically elongated for the reactant
complexes (RC)12n as the number of water molecules (n)
increases from 0 to 2. More linear H-bonds are possible, and
they are further strengthened by cooperative effects. Meanwhile,
1.51 Å in (TS)111. These structural changes are also accompanied
by noticeable changes in the atom charges. For instance, the
charge on the H1 atom decreases significantly from 0.493 in
(
RC)111 to 0.483 in (TS)111 as the N-H1 bond begins to form.
This is accompanied by noticeable changes in the charges on

Decomposition Reaction of NSCl
J. Phys. Chem. A, Vol. 112, No. 37, 2008 8565
Figure 6. Schematic diagram (not to scale) of the activation free energy
profiles (in kcal/mol) obtained from the B3LYP/aug-cc-pvdz calcula-
tions for the (RC)12n f (TS)12n f (PC)11n (where n ) 0, 1, 2) conversion
reactions. See text for more details.
(
associated with structures (RC)11n, (TS)11n, and (PC)11n. Due
to the similarities of reaction 1-B with reaction 1-A under
explicit addition of water molecules, a brief discussion for
reaction 1-B is presented here. The calculated results indicate
that the reaction barriers (reaction 1-1-B) for the NSCl +
nH2O f NSOH + (n - 1)H2O + HCl (where n ) 1, 2, 3)
conversion reaction are about 19.1 kcal/mol for n ) 1, 16.7
kcal/mol for n ) 2, and 13.6 kcal/mol. This suggests water
molecules also play an important role in decreasing the
reaction barriers of conversion except that the water cluster
has a more moderate effect on reaction 1-1-B compared to
that of addition (reaction 1-1-A). As for the reaction 1-2-B,
NSOH + n(H2O) f HNSO + nH2O (where n ) 0, 1, 2), the
tautomerization barriers of NSOH and HNSO are calculated
to be 34.9 kcal/mol, 10.3 kcal/mol, and 7.7 kcal/mol upon
going from the reactions that have zero water molecules to
those where two water molecules are explicitly added to the
reaction system. Examination of the energy diagram (shown
in Figure 2S of the Supporting Information) of reaction 1-B
shows that the reaction rate is determined by two steps
Figure 5. The optimized geometry (bond lengths are in Å) for all of
the reactants, reactant complexes, transition states, and product
complexes obtained from the B3LYP/aug-cc-pvdz calculations are
shown for the HNSCl(OH) + nH
, 2) reactions (associated with structures (RC)12n, (TS)12n, and (PC)12n
where n ) 0, 1, 2).
2
O f HNSO + HCl + nH O (n ) 0,
2
1
the distances of the leaving group H · · ·Cl become shorter. The
systematic elongation of the S-Cl bonds and the shortening of
weak H-Cl interactions for the reactant complexes suggest there
is a systematic trend for activating the S-Cl bonds and partial
formation of the H-Cl leaving group as more water molecules
are explicitly incorporated into the reaction systems.
The H1-O bonds become elongated to 1.16 Å for (TS)120,
1
.31 Å for (TS)121, and 1.42 Å for (TS)122, and this indicates
(
1
conversion and tautomerization). The total barrier of reaction
-B (calculated from reactions when three water molecules
+
-
+
-
that the H3O Cl or H5O2 Cl ions become almost formed in
the transition states (TS)121 and (TS)122. The transition states
were added into the reaction system) is also 18.5 kcal/mol
which is about that of reaction 1-A.
+
-
(
TS)121 and (TS)122 are greatly stabilized by the H3O Cl or
+
-
H5O2 Cl pairs, in which the S · · ·Cl interactions are stronger
than those without the water molecule assistance (e.g., smaller
S· · ·Cl distances than that without a water molecule present).
However, in the product complexes (PC)120, (PC)121, and (PC)122,
In view of the energetics, reactions 1-A and 1-B are competing
in the gas phase and may all be operative in the conversion of
NSCl to HNSO. Furthermore, these moderate reaction barriers (18.5
kcal/mol) are also in reasonable agreement with our experimental
observation that the half-life of NSCl to HNSO is several minutes
in the presence of trace amounts of moisture.
C. Decomposition Reactions of Thionylimide (HNSO) to
Produce NH3 and SO2 in the Presence of Water. The
experimental studies outlined in a previous section show bands
of the byproduct HNSO and SO2. Here we employed density
functional theory calculations to investigate the reaction mech-
anism of HNSO with trace quantities of water to hopefully
provide a reasonable explanation for the experimental observa-
tions shown in Figures 1 and 2.
+
-
+
-
the zwitterions H3O Cl or H5O2 Cl will collapse and
associate to produce HCl and H2O molecules.
Examination of the reaction barriers of reaction 1-1 (RC)113
f (TS)113 f (PC)113 in Figure 4 and reaction 1-2 (RC)122 f
(
TS)122 f (PC)122 in Figure 6 shows that the activation free
energy of reaction 1-1 is about 15.1 kcal/mol higher than that
of reaction 1-2. This indicates that reaction 1-1 is the rate-
determining step in the water assisted reactions of NSCl to
produce the simplest sulfinylimide HNSO. The energies of the
stationary structures for reaction 1-1 were also optimized at the
MP2 level of theory using the aug-cc-pvdz basis set and gave
similar results to those from the DFT calculations.
HNSO + nH O f H NSO(OH) + (n - 1)H O
2
2
2
1
-B: ConWersion of NSCl into NSOH and Subsequent
Tautomerizations of NSOH and HNSO. Optimized geometries
bond lengths are in Å) for all the reactants, reactant
(where n ) 1, 2, 3) (2-1)
(
First, we present the addition reaction of water molecules
with HNSO to produce the H2NSO(OH) species. Figures 7
and 8 display the optimized structures and energy diagrams
for the reactions (RC)21n f (TC)21n f (PC)21n where n ) 1,
2, 3. Analogous to the addition reactions of (RC)11n f (TS)11n
f (PC)11n where n ) 1, 2, 3, the structures for (RC)21n,
complexes, transition states, and product complexes obtained
from the B3LYP/aug-cc-pvdz calculations are shown in
Figure 1S of the Supporting Information for the NSCl +
nH2O f NSOH + (n - 1)H2O + HCl (n ) 1, 2, 3) reactions
and NSOH + nH2O f HNSO + nH2O (n ) 0, 1, 2)

8
566 J. Phys. Chem. A, Vol. 112, No. 37, 2008
Guo et al.
H NSO(OH) + nH O f NH + SO + nH O
2
2
3
2
2
(where n ) 0, 1) (2-2)
The decomposition reaction of H2NSO(OH) to produce NH3
and SO2 in the presence of the moisture in air was demonstrated
1
2,13
experimentally here and in previous work.
Decomposition
of H2NSO(OH) into NH3 and SO2 in trace quantities of water
can be modeled as shown in Figures 9 and 10, namely,
H2NSO(OH) + nH2O f NH3 + SO2 + nH2O or (RC)22n f
(
(
TS)22n f (PC)22n where n ) 0, 1. The reaction (RC)220 f
TS)220 f (PC)220 undergoes a similar four-centered cyclic
transition state (TS)220 to produce NH3 and SO2 product. In our
water incorporated reactions, the (RC)221, (TS)221, and (PC)221
are structurally similar to the corresponding ones found in the
decomposition of HNSCl(OH). It is shown in Table 2 that the
reaction barrier without water molecules for the decomposition
reaction of HNSO into NH3 and SO2 is predicted to be 22.6
kcal/mol at the B3LYP/aug-cc-pvdz level of theory. Analogous
to the cases of reactions 1-1, 1-2, and 2-1, the activation energy
(
E), the activation enthalpies (H), and the activation Gibbs free
Figure 7. The optimized geometry (bond lengths are in Å) for all of
the reactants, reactant complexes, transition states, and product
complexes obtained from the B3LYP/aug-cc-pvdz calculations are
energy for reaction 2-2 decreased to be 8.9, 8.0, and 9.8 kcal/
mol, respectively, when the second water molecule is added to
the reaction system.
Considering that reaction 2-2 with two water molecules has
a lower reaction barrier (9.8 kcal/mol) than that of reaction 2-1,
we believe that for the decomposition of HSNO into NH3 and
SO2, reaction 2-1 is the rate-determining step. The reaction
barriers for reactions 2-2 and 2-1 appear to have reasonable
shown for the H
, 3) reactions (associated with structures (RC)21n, (TS)21n, and (PC)21n
where n ) 1, 2, 3).
2
NSO + nH
2
O f H
2
NSO(OH) + (n - 1)H O (n ) 1,
2
2
Figure 8. Schematic diagram (not to scale) of the activation free energy
profiles (in kcal/mol) obtained from the B3LYP/aug-cc-pvdz calcula-
tions for the (RC)21n f (TS)21n f (PC)21n (where n ) 1, 2, 3) conversion
reactions. See text for more details.
Figure 9. The optimized geometry (bond lengths are in Å) for all of
the reactants, reactant complexes, transition states, and product
complexes obtained from the B3LYP/aug-cc-pvdz calculations are
(
TC)21n, and (PC)21n are quite similar to those of the NSCl
addition reactions with water molecules. Examination of
Table 2 and Figure 8 shows that the addition reaction has
activation free energy barriers of 35.8 kcal/mol for n ) 1,
shown for the H
) reactions (associated with structures (RC)22n, (TS)22n, and (PC)22n
where n ) 0, 1). See text for more details.
2
NSOl(OH) + nH
2
O f NH
3
+ SO
2
+ nH O (n ) 0,
2
1
2
2.6 kcal/mol for n ) 2, and 20.6 kcal/mol for n ) 3.
Although the main qualitative difference is that reaction 2-1
with the same number of water molecules is faster than reaction
-1, it is still hard to compare the predicted activation data with
1
the experimental data. As mentioned above, water molecules
can noticeably reduce the reaction barrier and accelerate the
reaction. As shown in Table 2, the one water molecule reaction
has a high reaction barrier, while the second water molecule
addition has a significant effect on the reaction of HNSO leading
to a decrease in the reaction barrier by 13.2 kcal/mol. However,
the addition reaction barrier of reaction 2-1 is only moderately
affected (by 2.0 kcal/mol) when the third water molecule is
added to the reaction system. The calculated reaction barriers
are 22.6 and 20.6 kcal/mol at the B3LYP/aug-cc-pvdz level of
theory for the two and three water reaction systems, respectively.
Therefore, the gas-phase reaction of 2-1 can probably be
reasonably modeled by reactions involving two or three water
molecules.
Figure 10. Schematic diagram (not to scale) of the activation free
energy profiles (in kcal/mol) obtained from the B3LYP/aug-cc-pvdz
calculations for the (RC)22n f (TS)22n f (PC)22n (where n ) 0, 1)
conversion reactions. See text for more details.

Decomposition Reaction of NSCl
J. Phys. Chem. A, Vol. 112, No. 37, 2008 8567
agreement with the experimental observations of the band at
by hydrogen bonding effects which are enhanced as the number
of water molecules increases. The conversion reaction of NSCl
includes two competing reaction pathways: reaction 1-A and
reaction 1-B. The calculated results show that 1-1-A is rate-
determining in the pathway 1-A. However, the total reaction
rate in reaction 1-B is determined by two steps (reaction 1-1-B
and reaction 1-2-B). It is worth noting that the reaction barriers
of reactions 1-A and 1-B are quite similar and both can possibly
be used to explain the experimental observation of a few minutes
half-time of conversion of NSCl to HNSO in the presence of
trace amounts of moisture. The decomposition reaction of HNSO
involves two parts: reaction 2-1 and reaction 2-2. The overall
decomposition reaction barrier (20.6 kcal/mol obtained from the
B3LYP/aug-cc-pvdz calculations) of the rate-determining step
reaction 2-1 can help explain the experimental observation of
-
1
12
1
374 cm , belonging to SO2.
D. Discussion of Results. It is necessary to assemble several
water molecules for the water assisted reactions to occur and
in some cases water is also a necessary reactant molecule. Since
reaction 1-1 is the rate-determining step in the water reactions
of NSCl to produce the simplest sulfinylimine HNSO, we have
not considered the contribution of reaction 1-2 to the reaction
rate in this system and focus on reaction 1-1 to make a rough
comparison between the predicted kinetics and experimental
observations of the NSCl lifetime. Examination of Table 2
shows the rate constants of the NSCl(H2O)n reactant complexes
to form the product complexes (e.g., (RC)11n f (TS)11n f
-
23 -1
(
×
PC)11n) were estimated to be 1.61 × 10
s
for n ) 1, 9.76
for n ) 3 and
-
7
-1
-1 -1
10
s
for n ) 2, and 1.75 × 10
s
predicted lifetimes for the NSCl(H2O)n reactant complexes of
SO formation of from hydrolysis due to the presence of trace
2
2
2
5
4
.3 × 10 s for n ) 1, 7.1 × 10 s for n ) 2, and 3.96 s for
quantities of water.
n ) 3. These results indicate that a simple gas phase reaction
of NSCl with one H2O molecule cannot explain the observed
experimental lifetime of several minutes for NSCl in the
presence of trace amounts of water in the reaction chamber and
that two or more water molecules are needed to convert the
NSCl to HNSO on a time scale comparable to experimental
observations.
Acknowledgment. This research has been supported by
grants from the Research Grants Council of Hong Kong (HKU
7036/04P), the award of a Croucher Foundation Senior Research
Fellowship (2006-07) from the Croucher Foundation, and an
Outstanding Researcher Award (2006) from the University of
Hong Kong to D.L.P., E.G.R., and D.McN. thank the Australian
Research Council for support and Finlay Shanks for essential
instrument support.
At low vapor pressure and room temperature conditions, water
clusters will have very low concentrations in the gas phase and
purely gas phase reactions of NSCl with water clusters may
not fully account for the relatively fast conversion of NSCl to
HSNO or the further decomposition of HSNO to a SO2
byproduct as observed in the FT-IR experiments (see Figures 1
Supporting Information Available: The optimized geom-
etries for all of the reactants, reactant complexes, transition
states, and product complexes obtained from the B3LYP/aug-
2
1
2
13
and 2 and references and ). However, it is worth noting that
water has a strong preference for attachment to the walls of
containers like glass, Pyrex, and metals, and wall effect reactions
may be implicated in the production of the observed byproduct.
Indeed, experimental observations in the Monash laboratories
over a number of years have shown that the pressure buildup
and drop in an isolated glass cell is at least partly associated
with water molecules and highly dependent on the pretreatment
of the cell. It is conceivable that a few dangling OH bonds of
bound water molecules at the borosilicate surface of the
container would be able to effectively mimic a water cluster
and efficiently react with NSCl to convert it to HSNO in a few
minutes and also to react with HSNO to make a SO2 byproduct
as seen in the experiments. These types of water assisted
reactions are consistent with our present theoretical results and
appear reasonably consistent with the experimental observations
and conditions. It would be interesting to better characterize
the decomposition reactions of NSCl and HSNO as a function
of container size and cell pretreatment through baking or water
saturation in order to explore how much these wall effect
reactions may contribute to the overall decomposition reactions
observed experimentally.
cc-pvdz calculations are shown for the reactions: NSCl + nH O
f HNSCl(OH) + (n - 1)H O (n ) 1, 2, 3); HNSCl(OH) + (n
2
- 1)H O f HNSO + (n - 1)H O + HCl (n ) 1, 2, 3); HNSO
2
2
+ n(H O) f H NSO(OH) + (n - 1)H O (n ) 1, 2, 3); NSCl
2
2
2
+ nH O f NSOH + (n - 1)H O + HCl; NSOH + nH O f
2
2
2
HNSO + nH O; H NSO(OH) + (n - 1)(H O) f NH + SO
2
2
2
3
2
+ (n - 1)(H O) (n ) 1, 2). The energy diagram of conversion
2
NSCl to HNSO under three water molecules is also collected.
This material is available free of charge via the Internet at http://
pubs.acs.org.
References and Notes
(1) Middleton, W. J. J. Am. Chem. Soc. 1966, 88, 3842.
(2) Bryce, M. R.; Taylor, P. C. J. Chem. Soc., Perkin Trans I 1990,
225.
3
(
3) Takahashi, M.; Okasaki, R.; Inamoto, N.; Sugarawa, T.; Inamura,
H. J. Am. Chem. Soc. 1992, 114, 1830.
(
4) Miller, K. V.; Emken, W. C.; Duncan, L. C. J. Fluorine Chem.
1
984, 26, 125.
(5) Jacox, M. E. J. Phys. Chem. Ref. Data 1984, 13, 945.
(6) Peake, S. C.; Downs, A. J. J. Chem. Soc., Dalton Trans. 1974, 859.
(
(
7) Glemser, O.; Richert, H. Z. Anorg. Allg. Chem 1961, 307, 313.
8) Mirri, A. M.; Guarnieri, A. Spectrochim. Acta, Part A 1967, 23, 2159.
(
9) Muller, A.; Nagarajan, G.; Glemser, O.; Cyvin, S. F.; Wegener, J.
Spectrochim. Acta, Part A 1967, 23, 2683.
Conclusions
(
10) Nguyen, M. T.; Flammang, R. Chem. Ber. 1996, 129, 1379.
In our FTIR spectra, we observe bands associated with HNSO
and SO2 from hydrolysis with trace quantities of water. The
half-life for conversion of NSCl to HNSO was observed to be
a few minutes when the pyrolysis products were isolated in the
cell and monitored by rapid survey scans. A density functional
theory study of the reactions of NSCl and HSNO with water
was presented in which hydrogen bonded water molecules (up
to 3) were explicitly included in the reaction systems. The
calculations show that the conversion reactions of NSCl and
decomposition reactions of HNSO are significantly influenced
(11) Chi, Y. J.; Yu, H. T. Chin. J. Chem. 2003, 21, 30.
(12) Robertson, E. G.; Thompson, C. D.; Lucie, S.; McNaughton, D.
Phys. Chem. Chem. Phys. 2005, 7, 483.
(
13) (a) Puskar, L. P.; Robertson, E. G.; McNaughton, D. J. Mol.
Spectrosc. 2006, 240, 244. (b) Robertson, E. G.; McNaughton, D. J. Mol.
Spectrosc. 2006, 238, 56–63.
(14) (a) Allaf, A. W.; Suffolk, R. J. J. Chem. Res. 1994, 1, 186. (b)
Robertson, E. G.; McNaughton, D. J. Phys. Chem. A 2003, 107, 642–650.
(
15) (a) Beck, A. D. J. Chem. Phys. 1993, 98, 5648. (b) Becke, A. D.
Phys. ReV. A 1988, 38, 3098. (c) Vosko, S. H.; Wilk, L.; Nusair, M Can.
J. Phys. 1980, 58, 1200. (d) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B
1988, 37, 785.

8
568 J. Phys. Chem. A, Vol. 112, No. 37, 2008
Guo et al.
(
16) (a) Kwok, W. M.; Zhao, C.; Guan, X.; Li, Y. L.; Du, Y.; Phillips,
(18) (a) Kwok, W. M.; Zhao, C.; Li, Y. L.; Guan, X.; Wang, D.; Phillips,
D. L. J. Am. Chem. Soc. 2004, 126, 3119. (b) Zhao, C.; Lin, X.; Kwok,
W. M.; Guan, X.; Du, Y.; Wang, D.; Hung, K. F.; Phillips, D. L. Chem.
Eur. J. 2005, 11, 1093.
D. L. J. Chem. Phys. 2004, 120, 9017. (b) Lin, X.; Zhao, C.; Phillips, D. L
Chem. Phys. Lett. 2004, 397, 488. (c) Lin, X.; Guan, X.; Kwok, W. M.; Zhao,
C.; Du, Y.; Li, Y. L.; Phillips, D. L. J. Phys. Chem. A 2005, 109, 981.
(
17) (a) Li, Y. L.; Zhao, C.; Kwok, W. M.; Guan, X.; Zuo, P.; Phillips,
(
19) Wallington, T. J.; Schneider, W. F.; Barnes, I.; Becker, K. H.;
D. L. J. Chem. Phys. 2003, 119, 4671. (b) Kwok, W. M.; Zhao, C.; Li,
Y. L.; Guan, X.; Phillips, D. L. J. Chem. Phys. 2004, 120, 3323. (c) Phillips,
D. L.; Zhao, C.; Wang, D. J. Phys. Chem. A 2005, 109, 9653.
Sehested, J.; Nielsen, O. J. Chem. Phys. Lett. 2000, 322, 97–102.
JP802445R