Molecular structure and infrared spectra of the monomeric 3-(methoxy)-1,2-benzisothiazole 1,1-dioxide (methyl pseudosaccharyl ether)

Article abstract of DOI:10.1016/j.molstruc.2007.06.004

The computational description of saccharin (1,2-benzisothiazol-3(2H)-one-1,1-dioxide) and its derivatives is difficult due to the presence of hypervalent S{double bond, long}O bonds in their structures. Therefore, in this investigation, the HF, DFT/B3LYP and MP2 methods were used to predict the geometry and the infrared spectrum of the saccharyl derivative 3-(methoxy)-1,2-benzisothiazole 1,1-dioxide (MBID). Their relative predictive capabilities were then evaluated by comparing the obtained results with experimentally available data, namely the newly obtained IR spectra of MBID isolated in low-temperature inert matrices. For each method, different basis sets [6-31++G(d,p), 6-31++G(3df,3pd), 6-311++G(d,p), 6-311++G(2df,2pd), 6-311++G(3df,3pd), aug-cc-pVDZ and aug-cc-pVTZ] were considered. The best overall agreement has been achieved at the B3LYP/6-311++G(3df,3pd) and B3LYP/6-31++G(3df,3pd) levels of theory, showing the adequacy of the B3LYP functional to describe the investigated properties in this type of compounds and stressing the relevance of including high-order polarization functions in the basis set. The chosen level of theory [B3LYP/6-311++G(3df,3pd)] was applied to analyze the vibrational spectra and the geometry of the title molecule. In agreement with the experiment, the C{single bond}O{single bond}C linkage in MBID is predicted by these calculations to exhibit considerably short (1.320 ?) and long (1.442 ?) (N{double bond, long})C{single bond}O and (H₃)C{single bond}O bonds, respectively, and a hybridization of the central oxygen atom close to sp² (the C{single bond}O{single bond}C angle is predicted to be ca. 117°). This C{single bond}O{single bond}C bonding pattern fits the well-known high reactivity of MBID upon thermal rearrangement, which has been shown to result in easy selective [1,3′]-isomerization of the compound.

Full text of DOI:10.1016/j.molstruc.2007.06.004

Available online at www.sciencedirect.com
Journal of Molecular Structure 876 (2008) 77–85
www.elsevier.com/locate/molstruc
Molecular structure and infrared spectra of the monomeric
3-(methoxy)-1,2-benzisothiazole 1,1-dioxide
(methyl pseudosaccharyl ether)
a,b,
a,c
a,d
*, Rui Almeida , Andrea Gomez-Zavaglia
,
´
Agnieszka Kaczor
b
a
Maria de Lurdes S. Cristiano , Rui Fausto
a
Department of Chemistry, University of Coimbra, Coimbra P-3004-535, Portugal
Faculty of Chemistry, Jagiellonian University, Ingardena 3, 30-060 Krakow, Poland
b
c
Department of Chemistry and Biochemistry, F.C.T., and CCMAR, University of Algarve, Campus de Gambelas, 8005-039 Faro, Portugal
d
Faculty of Pharmacy and Biochemistry, University of Buenos Aires, C.P. 1113 Buenos Aires, Argentina
Received 27 April 2007; received in revised form 29 May 2007; accepted 1 June 2007
Available online 20 June 2007
Abstract
The computational description of saccharin (1,2-benzisothiazol-3(2H)-one-1,1-dioxide) and its derivatives is difficult due to the pres-
ence of hypervalent S@O bonds in their structures. Therefore, in this investigation, the HF, DFT/B3LYP and MP2 methods were used to
predict the geometry and the infrared spectrum of the saccharyl derivative 3-(methoxy)-1,2-benzisothiazole 1,1-dioxide (MBID). Their
relative predictive capabilities were then evaluated by comparing the obtained results with experimentally available data, namely the
newly obtained IR spectra of MBID isolated in low-temperature inert matrices. For each method, different basis sets [6-31++G(d,p),
6-31++G(3df,3pd), 6-311++G(d,p), 6-311++G(2df,2pd), 6-311++G(3df,3pd), aug-cc-pVDZ and aug-cc-pVTZ] were considered. The
best overall agreement has been achieved at the B3LYP/6-311++G(3df,3pd) and B3LYP/6-31++G(3df,3pd) levels of theory, showing
the adequacy of the B3LYP functional to describe the investigated properties in this type of compounds and stressing the relevance of
including high-order polarization functions in the basis set.
The chosen level of theory [B3LYP/6-311++G(3df,3pd)] was applied to analyze the vibrational spectra and the geometry of the title
molecule. In agreement with the experiment, the CAOAC linkage in MBID is predicted by these calculations to exhibit considerably
˚
˚
short (1.320 A) and long (1.442 A) (N@)CAO and (H3)CAO bonds, respectively, and a hybridization of the central oxygen atom close
to sp2 (the CAOAC angle is predicted to be ca. 117ꢀ). This CAOAC bonding pattern fits the well-known high reactivity of MBID upon
thermal rearrangement, which has been shown to result in easy selective [1,30]-isomerization of the compound.
ꢁ 2007 Elsevier B.V. All rights reserved.
Keywords: Methyl pseudosaccharyl ether; Molecular structure; IR spectra; Matrix isolation; DFT and ab initio calculations
1. Introduction
attention as they show herbicidal [1,2], antimicrobial and
antifungal activity [3–5] or potential in enzymatic inhibi-
tion [6]. Additionally, substituted 1,2-benzisothiazole 1,1-
dioxides (pseudosaccharins) are important intermediates
in the organic synthesis, as their O-ethers provide efficient
intermediates for reductive cleavage of the CAO bond in
phenols [7,8], benzylic [9] and naphthylmethylic alcohols
[10], through heterogeneous catalytic transfer hydrogenoly-
sis, or through cross-coupling with organometalic reagents
[11–13]. Though the presence of p-type orbitals in the
Saccharin (1,2-benzisothiazol-3(2H)-one-1,1-dioxide) is
a commonly known substance, as it is the oldest artificial
sweetener. Apart from the wide commercial applications
of saccharin itself, its derivatives receive an increased
*
Corresponding author. Address: Faculty of Chemistry, Jagiellonian
University, Ingardena 3, 30-060 Krakow, Poland.
E-mail address: kaczor@chemia.uj.edu.pl (A. Kaczor).
0022-2860/$ - see front matter ꢁ 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.molstruc.2007.06.004

78
A. Kaczor et al. / Journal of Molecular Structure 876 (2008) 77–85
substituent is an additional requirement, ensuring a proper
alignment of the compound on catalyst surface
matrices. This method allowed obtaining, for the first time,
well-resolved pure vibrational spectra of the isolated
MBID, which can be directly compared with the theoretical
results obtained for the molecule in vacuum, due to the
very weak interactions between the medium (noble gas)
and the target molecule.
a
(alkylpseudosaccharyl ethers are unreactive) [8], the struc-
tural basis of the enhanced reactivity of this kind of
pseudosaccharyl ethers towards transition metal-catalyzed
ipso-replacement lies in the unusual bond lengths and
geometry of their CRAOACA linkage (where R = heteroa-
romatic ring and A = aliphatic or aryl group) [8–10,13–18].
The strong electron withdrawing effect of the pseudosaccha-
ryl group results in significant lengthening of the CAAO
bond at the expense of the CRAO bond, which then gains
partial double-bond character [8–10,13,19]. This character-
istic structural feature, which is directly related with the
potential of the pseudosaccharyl-oxygen system as nucleo-
fuge in TM-catalyzed reductive cleavage, makes pseudosac-
charins interesting compounds for structural analysis.
Nowadays, quantum-chemical methods provide power-
ful support for experimental structural studies. Therefore,
both geometry and vibrational properties of saccharin
and several metal saccharinates have been extensively stud-
ied by ab initio and DFT methods [20–25]. The key prob-
lem in these investigations is the difficulty to obtain an
accurate description of the S@O bond. Indeed, most of
the standard computational methods and basis sets have
been shown to be unable to properly treat this moiety, lead-
ing to considerable discrepancies between the experimental
and theoretically predicted values for the S@O bond
lengths and for the vibrational frequencies associated with
this bond [22–25]. The underestimation of the theoretically
predicted frequencies for the S@O stretching vibrations
was noticed previously for different types of compounds
containing the S@O bond, from simple molecules, as
dimethyl sulfate [26], dimethyl sulfite [27] or the SSOꢀ
anion [28], to more complex species, like 5-methyl-
2. Methods
2.1. Experimental
MBID was synthesized from 3-chloro-1,2-benzisothiaz-
ole-1,1-dioxide (pseudosaccharyl chloride), through solvol-
ysis. The saccharyl halide (0.105 g, 5.2 mmol) was
dissolved in anhydrous methanol (10 mL) and the mixture
was stirred under nitrogen, at room temperature, until
TLC analysis (DCM as eluent) indicated the absence of
pseudosaccharyl chloride (10 min). A white precipitate
was formed upon stirring. The reaction mixture was diluted
with water (10 mL) and the excess of methanol was
removed under reduced pressure. The organic product
was then extracted with ethyl acetate (3 · 20 mL) and the
organic extract washed with aqueous HCl (pH 2–3), later
with aqueous sodium bicarbonate (pH 11–12) and finally
with water. The organic solution was finally dried over
anhydrous sodium sulfate, filtered, and the filtrate evapo-
rated to dryness, under reduced pressure, to give a colour-
less solid. Recrystallization from toluene gave the required
product, 3-methoxy-1,2-benzisothiazol-1,1-dioxide as col-
ourless crystals (62% yield; m.p. 184–185 ꢀC).
1H NMR (300 MHz, CDCl3): d, 3.39 (3H, s), 7.40 (1H,
t), 7.89–8.02 (3H, m) MS (EI): m/z 181([M]+, 100).
The infrared spectra were recorded in the 400–
4000 cmꢀ1 range using a Mattson Infinity 60AR series
FT-IR spectrometer, with 0.5 cmꢀ1 resolution. The sample
was co-deposited with argon (N60, Air Liquide) or krypton
(N48, Air Liquide) isolant gas onto a cryogenically cooled
CsI window. The compound was sublimated at ca. 330–
351 K from a specially designed mini-furnace assembled
inside the cryostat. The selected temperature of the optical
substrate (10 K) was obtained using an APD Cryogenics
closed-cycle helium refrigeration system with a DE-202A
expander. The temperature was measured directly at the
sample holder by a silicon diode temperature sensor, con-
nected to a digital controller (Scientific Instruments, Model
9650-1) with the accuracy of 0.1 K.
1H,3H-pyrrolo[1,2c][1,3]thiazole-6,7-dicarboxylate
2,2-
dioxide [29]. Although a good agreement of the experimen-
tal and predicted frequencies was achieved for both
dimethyl sulfate and sulfite when the B3LYP method was
used together with the split valence quadruple-f basis set
(aug-cc-pVQZ) [26,27], calculations on pseudosaccharins
at this level of theory are currently unpractical due to its
excessive computational cost.
In the present investigation, a comprehensive evaluation
of commonly used theoretical methods (HF, B3LYP and
MP2) with various split valence double and triple-f basis
sets was carried out for 3-(methoxy)-1,2-benzisothiazole
1,1-dioxide (MBID), in order to find a suitable and cost-
effective level of theory to be used in the study of saccharin
derivatives. MBID was chosen due to its conformational
simplicity (it possesses a single conformer with practical
significance). The choice of the target compound was also
motivated by our interest to subsequently address the prob-
lem of [1,30]-isomerization in pseudosaccharin O-ethers,
MBID being one of the simplest compounds prone to
undergo this type of rearrangement. The experimental
method selected to evaluate the computational results
was infrared spectroscopy in low-temperature inert
2.2. Computational
Geometry optimizations were performed at the HF,
DFT (B3LYP [30,31]) and MP2 levels of theory, using
the 6-31++G(d,p) and 6-311++G(d,p) basis sets for
MP2, the 6-31++G(d,p), 6-311++G(d,p) and 6-
311++G(3df,3pd) basis sets for HF and all these basis sets
plus the 6-31++G(3df,3pd), 6-311++G(2df,2pd), aug-cc-
pVDZ and aug-cc-pVTZ bases for B3LYP. In some cases,
the Geometry Direct Inversion of the Invariant Subspace

A. Kaczor et al. / Journal of Molecular Structure 876 (2008) 77–85
79
(GDIIS) method of optimization [32,33] was applied.
50
40
30
20
10
0
Vibrational frequencies were calculated at each level of the-
ory and the nature of the stationary points on the potential
energy surface (PES) resulting from optimization was
determined by inspection of the corresponding calculated
Hessian matrix. The optimized structures described in this
study were confirmed to correspond to true minimum
energy conformations on the different potential energy sur-
faces investigated.
Potential energy profile for internal rotation was calcu-
lated performing a relaxed scan on the DFT (B3LYP)/6-
311++G(3df,3pd) PES along the relevant reaction
coordinate.
All the above-mentioned calculations were performed
using the Gaussian98 [34] or Gaussian03 [35] suites of pro-
grams. Potential energy distributions (PEDs) of the normal
modes were computed in terms of natural internal coordi-
nates [36] with the Gar2ped [37] program.
-200
-150
-100
-50
0
50
100
150
200
N8=C9-O15-C17 angle (deg)
Fig. 2. Potential energy profile for internal rotation around the
N8@C9AO15AC17 dihedral angle, calculated at the B3LYP/6-
311++G(3df,3pd) level of theory.
3. Results and discussion
3.2. Vibrational spectra of MBID and evaluation of
computational methods
3.1. Conformational distribution in MBID
The infrared spectra of MBID isolated in solid argon
and krypton at 10 K are given in Fig. 3 along with the the-
oretical spectra of the most stable monomer calculated at
different levels of theory. Computed frequencies have been
scaled down to account for anharmonicity and limitations
of the basis sets. The scaling factors were found by linear
fitting of a criteriously chosen set of well-defined experi-
mental (argon) and calculated frequencies (given in Table
S1 of the Supplementary Material), with the intercept of
the line set on zero. The scaling factors used in the present
analysis are given in Table 1, which also includes the fre-
quencies of the bands attributed to the S@O asymmetric
and symmetric stretching vibrations. Definition of the
internal coordinates used in the normal mode analysis of
MBID and computed B3LYP/6-311++G(3df,3pd) fre-
quencies and IR intensities are given in Tables S2 and S3.
The assignments for the matrix spectra are provided in
Table 2.
The comparison of the computed vibrational frequen-
cies with the matrix data enabled unequivocal evaluation
of the applied methods. As it will be shown below, the
B3LYP/6-311++G(3df,3pd) calculations give the best
overall agreement between theoretically predicted and
observed spectral features, and then, among the different
alternatives tested, appear as the most convenient theoret-
ical approach to the study of the type of compounds under
consideration.
The N8@C9AO15AC17 dihedral angle is the only rele-
vant coordinate for conformational flexibility within the
MBID molecule (see Fig. 1). The potential energy profile
for internal rotation around the C9AO15 bond, calculated
at the B3LYP/6-311++G(3df,3pd) level, is given in
Fig. 2. In the global minimum conformation, the
N8@C9AO15AC17 dihedral angle is 0ꢀ and the molecule
exhibits Cs symmetry. The energy is computed to increase
quite steeply until the N8@C9AO15AC17 angle is 120ꢀ;
from this point on, a nearly flat profile is predicted, exhib-
iting a very shallow minimum at 180ꢀ. The less stable
conformer is ca. 47 kJ molꢀ1 higher in energy than the con-
formational ground state.
Taking into account the large energy difference between
the two MBID conformers, it can confidently be stated
that, in the range of temperatures used to sublime the com-
pound for matrix preparation (ca. 330–351 K), the higher
energy form is of no practical importance due to its exceed-
ingly small population. This form will then be neglected
from now on.
O14
H13
O16
6
S7
H12
5
4
1
2
N8
The most striking differences between the experimental
and calculated spectra occur for the absorptions associated
with the S@O stretching vibrations. For these vibrations, a
considerable discrepancy can be noticed for all the levels of
theory tested, apart from HF/6-311++G(3df,3pd),
B3LYP/6-31++G(3df,3pd) and B3LYP/6-311++G(3df,
3pd). Note, however, that although surprisingly good
9
H11
H19
3
H20
O15
H10
C17
H18
Fig. 1. Adopted atom numbering scheme for the MBID molecule. Ring
carbon atoms are only represented by their numbers.

80
A. Kaczor et al. / Journal of Molecular Structure 876 (2008) 77–85
ν2
100
50
HF/6-31++G(d,p)
ν1
0
100
50
HF/6-311++G(d,p)
ν2
ν1
0
ν1
HF/6-311++G(3df,3pd)
B3LYP/aug-cc-pVDZ
B3LYP/aug-cc-pVTZ
ν2
100
0
ν2
ν1
50
0
ν2
ν1
50
0
100
B3LYP/6-31++G(d,p)
ν1
ν1
ν2
50
0
B3LYP/6-31++G(3df,3pd)
ν2
50
0
100
B3LYP/6-311++G(d,p)
ν1
ν2
ν2
ν2
50
0
B3LYP/6-311++G(2df,2pd)
B3LYP/6-311++G(3df,3pd)
ν1
ν1
50
0
50
0
100
MP2/6-31++G(d,p)
ν1
ν1
ν2
50
0
100
MP2/6-311++G(d,p)
ν2
ν2
ν2
50
0
Ar
Kr
ν1
ν1
0.3
0.0
0.3
0.0
400
600
800
1000
1200
1400
1600
1800
Wavenumber (cm-1)
Fig. 3. Comparison of the FT-IR spectra of monomeric MBID isolated in argon and krypton matrices with the theoretical spectra obtained at different
levels of theory. m1 and m2 denote the bands associated mainly with the asymmetric and symmetric S@O stretching vibrations, respectively.
estimations for the S@O stretching frequencies were
obtained at the HF/6-311++G(3df,3pd) level, considerable
disagreement with the experimental data was observed for
both frequencies and intensities of other absorptions, dis-
approving the use of this method. For example, the
description of the relative intensities of the bands in the
1650–1400 cmꢀ1 region is erroneous, with the relative
intensities of the two intense bands attributed mainly to
the C@N bond stretching coordinate being reversed at
the HF level (whatever the basis set used; see Fig. 3). The
frequency separation between these two bands is also over-
estimated by the HF method (63 cmꢀ1 for all basis sets) rel-
ative to experiment (54 cmꢀ1 for both Ar and Kr) and all
other methods applied (51–55 cmꢀ1, with the exception of
B3LYP/aug-cc-pVDZ, for which it is 58 cmꢀ1).
In the experiment, the absorption due to the asymmet-
ric S@O stretching mode is observed at 1370 cmꢀ1 in
argon and at 1367 cmꢀ1 in krypton, whereas that corre-
sponding to the symmetric S@O stretching is found at
1189 cmꢀ1 (Ar) and at 1188 cmꢀ1 (Kr). An error of

A. Kaczor et al. / Journal of Molecular Structure 876 (2008) 77–85
81
Table 1
set strongly influences the agreement between the
experimental and calculated values. Indeed, both the
general performance as well as the theoretical estimation
of the mS@O frequencies progressively lose quality
when B3LYP/6-311++G(3df,3pd) is restricted to
B3LYP/6-311++G(2df,2pd) and further to B3LYP/6-
311++G(d,p). This conclusion is also reinforced if
one compares the results obtained with the 6-
31++G(3df,3pd) and 6-311++G(2df,2pd) basis sets used
together with the B3LYP functional. The obtained agree-
ment between the experimental and computed mS@O fre-
quencies was definitely better for the double-f basis with
the full set of polarization functions (error: 1.3% for
both frequencies; calc.: 1351 cmꢀ1, 1173 cmꢀ1) than for
the triple-f basis set with lower number of polarization
functions.
The number of basis functions in the basis set, however,
does not significantly influences the results if the same set
of polarization functions is used. Vibrational frequencies
obtained with the ‘‘full’’ 6-311++G(3df,3pd) basis set are
only slightly better than those obtained with the 6-
31++G(3df,3pd) basis. Therefore, the latter is certainly
an adequate choice for the theoretical treatment of
pseudosaccharins of bigger size.
Comparison of frequencies of the S@O stretching vibrations calculated at
different levels of theory with the experimental values
Frequency (cmꢀ1 a
)
Scaling
factorb
m(S@O)
m(S@O)
asymmetric
symmetric
Experimental
Argon matrix
Krypton matrix
1370
1367
1189
1188
Calculated
HF/6-31++G(d,p)
1306 (4.6)
1311 (4.2)
1370 (0.1)
1248 (8.8)
1319 (3.6)
1290 (5.7)
1351 (1.3)
1296 (5.3)
1337 (2.3)
1355 (1.0)
1316 (3.8)
1339 (2.2)
1146 (3.6)
1148 (3.4)
1197 (0.7)
1093 (8.0)
1146 (3.6)
1116 (6.1)
1173 (1.3)
1120 (5.7)
1162 (2.2)
1177 (1.0)
1146 (3.6)
1161 (2.3)
0.897
0.901
0.901
0.995
0.990
0.986
0.988
0.991
0.988
0.987
0.984
0.991
HF/6-311++G(d,p)
HF/6-311++G(3df,3pd)
B3LYP/aug-cc-pVDZ
B3LYP/aug-cc-pVTZ
B3LYP/6-31++G(d,p)
B3LYP/6-31++G(3df,3pd)
B3LYP/6-311++G(d,p)
B3LYP/6-311++G(2df,2pd)
B3LYP/6-311++G(3df,3pd)
MP2/6-31++G(d,p)
MP2/6-311++G(d,p)
a
The percent error relative to the averaged experimental (argon and
krypton) results is given in the parentheses.
b
Scaling factors were found by linear fitting (assuming a zero intercept)
of the chosen set of experimental and calculated frequencies (for details
refer to the text).
On the other hand, the choice of the type of the basis set
significantly affects the vibrational frequencies. Compari-
son of B3LYP/6-311++G(2df,2pd) with B3LYP/aug-cc-
pVTZ (see Fig. 3), both having the same set of polarization
functions, clearly shows that the Dunning’s correlation
consistent basis set performs much worse than the Pople’s
basis set. This trend is even more evident upon reducing the
level of calculation, i.e., when the B3LYP/6-31++G(d,p)
vs. the B3LYP/aug-cc-pVDZ results are compared, with
the latter being completely unsuccessful to perform reason-
ably in the whole considered range of frequencies (see
Fig. 3).
A final note must be made regarding use of MP2 calcu-
lations. For a given basis set, the MP2 method produces
the best description of the vibrational spectra of MBID
when compared with both B3LYP and HF. However, at
the MP2/6-311++G(d,p) level, the mS@O frequencies are
overestimated by ca. 30 cmꢀ1. Much better results are
obtained at the much less time-consuming B3LYP/6-
311++G(3df,3pd) level, which yields the mS@O frequencies
overestimated only by ca. 15 cmꢀ1. Therefore, taking into
account the time factor, the MP2 was not tested at the
higher levels, as the method does not seem practical to
the computation of the vibrational spectra of pseudosac-
charyl ethers.
1.0% relative to the averaged (argon and krypton)
experimental frequencies was obtained for both S@O
stretching frequencies calculated at the B3LYP/6-
311++G(3df,3pd) level (see Table 1). Compared to these
results, the error doubled when the significantly more
time-consuming MP2/6-311++G(d,p) level of theory
was used (2.2/2.3%; calc.: 1339 cmꢀ1, 1161 cmꢀ1) and it
nearly quadrupled for the same method, but with the
double-f basis set (3.8/3.6%; calc.: 1316 cmꢀ1
,
1146 cmꢀ1).
On the other hand, the B3LYP functional used together
with the very common and usually efficient 6-31++G(d,p)
and 6-311++G(d,p) basis sets gave results that are signifi-
cantly worse (errors in the range of 5.3–6.1%; calc.:
1290 cmꢀ1, 1116 cmꢀ1 and 1296 cmꢀ1, 1120 cmꢀ1, for dou-
ble-f and triple-f basis set, respectively) not only than
MP2, but also HF. For the latter, errors were in the range
of 3.4–4.6% (calc.: 1306 cmꢀ1, 1146 cmꢀ1 and 1311 cmꢀ1
,
1148 cmꢀ1, for double-f and triple-f basis set, respectively).
Furthermore, application of this functional with the aug-
cc-pVDZ basis set brought an even higher discrepancy
between the experimental and calculated frequencies (8.8/
8.0%; calc.: 1248 cmꢀ1, 1093 cmꢀ1). Using the respective
triple-f basis set (aug-cc-pVTZ) instead of the double-f
one reduced the error to 3.6% for both modes (calc.:
1319 cmꢀ1, 1146 cmꢀ1). Nevertheless, this error is still
much larger than that obtained when the B3LYP/6-
311++G(3df,3pd) level of theory is used.
3.3. Geometry of MBID
As the best agreement between the experimental and
computed vibrational frequencies was achieved at the
B3LYP/6-311++G(3df,3pd) level of theory, the latter will
be here used as the reference method in the analysis of
geometry of MBID (see Fig. 4).
From the whole set of results, it became evident that
the number of polarization functions present in the basis

Table 2
Assignment of experimental IR spectra of MBID isolated in an argon and krypton matrices
Calculated valuesa B3LYP/6-311++G(3df,3pd)
Argon (10 K)
m/cmꢀ1
Krypton (10 K)
Approximate assignment
m/cmꢀ1
I/km molꢀ1
I
m/cmꢀ1
I
3167
3164
3152
3139
3127
3095
3019
1634
1610
1583
1478
1476
1466
1465
1457
5.4
2.2
3076
vw
mCAH6R
mCAH6R
mCAH6R
mCAH6R
mCAHM
mCAHM
mCAHM
mC@N, mCAC6R
mCAC6R
mCAC6R, mC@N
dCACAH6R
dM
4.6
3049
1.2
vw
10.3
8.9
3004
vw
24.8
150.4
3.6
2956
w
3013
w
1635b/1633/1629
1599
m/m/vs
vw
1632/1627
1597
1568/1565b
1475
w/vs
vw
164.7
11.1
10.8
2.7
1575/1568b
1479
vs/m
vw
vs/w
w
1473
m
1472
w
1459
w
1456
w
dM
dM
dM
11.3
33.6
1457
w
1454
w
1443
m
1442
m
1441
m
1440
m
dM
1371
1355
340.8
210.6
1383b/1378/1376b
1370/1368b/1366b
1355/1354
w/vs/vs
vs/vs/w
w/w
1380b/1376
1367/1365b/1361b
1354
m/vs
mCAO, mCAC5R
mS@O
vs/w/w
w
1349/1347
1340
w/w
1342
w
w
1336
w
1334
w
1322
w
1321
w
1328
1284
70.8
1.5
1315b/1313b/1312
1277b/1274
1242b/1239
1208
w/w/m
wv/wv
wv/wv
w
1313b/1310
1275/1273/1272
1243b/1240b/1237
1206
w/m
mCAC6R
dCACAH6R
wv/wv/wv
wv/wv/wv
1211
3.6
w
dM
1205
w
1204
w
dM
1177
1172
1160
1150
1129
1049
1025
1005
972
270.7
52.9
0.9
1199/1197/1194/1192/1190
1187/1185
1161
w/m/m/m/vs
s/m
1193
vs
mS@O
dCACAH6R
dM
dCACAH6R, mCAC6R
dCACAH6R, mCAC6R
dCACAC6R
mCAC6R
c6R
mC17AO
c6R, c5R
mNAS, dXACAO
1187b
m
vw
1160
wv
m/m
25.1
0.2
1148/1144
m/m
1146/1142
18.9
0.9
1061b/1057
1017
w/m
vw
1062b/1058/1056
1017
vw/m/m
vw
0.0
29.1
0.6
973/970/966
m/w/w
974/973/970/968
915/908/906
m/m/m/m
m/m/m
970
896
63.5
0.0
917b/916/913/911/910/907/903
m/m/m/m/m/m
891
c6R
787
13.6
41.3
32.3
10.2
20.8
39.4
90.0
15.5
20.1
7.5
780/779/771/770
w/w/w/m
783b/780/771/
768/764/762
752/751b/750
705b/703
w/w/m
m/w/w
w/m/m
m/w
c6R
dCACAC6R
c6R, s6R
dCACAC6R
762
768b/765
m/w
754
756/753/751b/750
w/w/m/m
700
704
673b/671
m
672
w/m
673b/972b/670
621
w/w/m
m
s6R
617
621
m
dCACAC6R, dXAS@O
cCAS (@O)-N, dCAXAX5R
cCAS (@O)AN, dCAXAX5R, s6R
dXAS@O
588
600/597/596/592
w/s/s/vw
599/596/590/589b
540
536b/543
w/s/vw/vw
m
538
542
535
507
m
w
w
529
m/m
501
507
w
dCAXAX5R
a
Frequencies scaled uniformly by the factor of 0.987.
Shoulder.
b

A. Kaczor et al. / Journal of Molecular Structure 876 (2008) 77–85
83
The computed values of bond angles and dihedral angles
are practically coincident for all methods used, the major
differences being 3.5ꢀ and 1.3ꢀ, respectively. Also the calcu-
lated values of the bond lengths are considerably conserved
In the thorough study of the C2S2H2 isomers, the
authors stated also that the HF method reproduces the
energy trend, but they ascribed it to incidental cancellation
of errors [40]. Therefore, like for computation of the vibra-
tional spectra, HF does not seem to be appropriate to
describe the geometry of pseudosaccharins. Indeed, the
bond lengths obtained at the HF level of theory (with all
basis set tested) are considerably shorter with respect to
bond lengths calculated using the MP2 and DFT methods.
As it could be expected, this is particularly noticeable for
bonds involving the sulfur atom. Inclusion of polarization
functions in the HF calculations resulted in further short-
ening of all bonds, markedly apparent for CAO and
SAX bonds (X = C, O or N). For example, the deviation
of the HF/6-311++G(3df,3dp) calculated values from
those obtained at the B3LYP/6-311++G(3df,3pd) level
˚
throughout the molecule (major difference 60.02 A) with
the exception of the bonds linking the substituent to the
saccharin moiety (both CAO bonds) and, specially, the
bonds involving the sulfur atom, for which considerable
deviations are noticed (see Table 3).
The bonding description of formally hypervalent mol-
ecules requires inclusion of both electron correlation and
polarization functions [38]. The crucial importance of
polarization functions, particularly of the f-type was also
previously emphasized in studies of the relative energies
of sulfur-containing molecules of the C2S2H2 type
[39,40]. The interesting conclusion presented by Vijay
et al. was the apparent inadequacy of DFT methods to
predict relative energies of the C2S2H2 isomers [40]. On
the other hand, the present study supports the adequacy
of the DFT approach to study sulfur compounds once
the proper basis set and functional has been chosen
(although it has to be underlined that we did not con-
sider specifically relative conformational energies, due
to the fact that MBID has only one experimentally
meaningful conformer, but the relative performances of
several theoretical methods to calculate accurate vibra-
tional frequencies.) This apparent discrepancy might be
rationalized if we take into consideration the different
basis sets applied by Vijay et al. (6-31G*, 6-311G*, 6-
311+G*, 6-311+G**, 6-311++G**, 6-31G(2df), 6-
31G(3df), 6-31G(3df,3pd) [40]) and in the present work
or the dissimilarity of the systems under study.
˚
are within the range 0.023–0.043 A, the latter value for
the SAN bond. The inability of HF to predict correctly
the geometry of MBID is also clear if a comparison is made
between the calculated results obtained at this level and
available crystallographic data for other substituted
arylpseudosaccharyl ethers. Although the overall value for
the bond lengths in the CAOAC linkage for a series of
arylpseudosaccharyl ethers is not as close as in the case of
˚
substituted aryltetrazolyl ethers (2.777 A) [8], it is found
˚
in the quite narrow range of 2.752–2.802 A [13,16–19].
The HF predicts this values to fall between 2.716 and
˚
2.728 A, depending on the basis set applied, which is a
clearly wrong prediction, even taking into account that
the gas-phase and crystal data may differ somewhat. On
the other hand, reasonable values for the CAOAC linkage,
˚
in the 2.762–2.783 A range, are obtained for all other levels
of theory used.
In contrast to HF results, bond lengths resulting both
from MP2 and B3LYP calculations with basis sets with
restricted number of polarization functions are overesti-
mated relatively to the B3LYP/6-311++G(3df,3pd) refer-
ence data (see Table 3). In particular, quite distinctly
different values for the S@O and SAN bond lengths are
predicted by MP2 and B3LYP/6-311++G(3df,3pd). In this
case, however, the reason of deviation might be ascribed to
the basis set used rather than to the method itself, as a
nearly coincidence is observed in the values obtained with
both MP2 and B3LYP methods equipped with the same
basis set [MP2/6-31++G(d,p) vs. B3LYP/6-31++G(d,p)
and MP2/6-311++G(d,p) vs. B3LYP/6-311++G(d,p)].
The overestimation of the bond lengths does not depend
only on the number of polarization functions present in the
basis set. For example, the B3LYP/aug-cc-pVDZ basis set,
which is comparable in this sense with the B3LYP/6-
31++G(d,p) basis, predicts exceptionally long bonds (par-
ticularly evident for the SAN and S@O bonds), resulting in
its deep inability to reproduce vibrational frequencies (refer
to previous section). On the contrary, the computed bond
lengths do not depend strongly on the additional splitting
of the valence shell upon moving from the 6-31 type to
the 6-311 type basis sets, which for the same method and
Fig. 4. The B3LYP/6-311++G(3df,3pd) optimized geometry of MBID.
Values of bond lengths (in italics, A) and valence angles (degrees) are
given.
˚

84
A. Kaczor et al. / Journal of Molecular Structure 876 (2008) 77–85
Table 3
The chosen bond lengths for MBID optimized at different levels of theory
˚
Method/basis set
Bond length/A
C1AS7
S7AN8
S7@O14/S7@O16
C9@N8
C9AO15
C17AO15
HF/6-31++G(d,p)
HF/6-311++G(d,p)
1.779
1.779
1.768
1.817
1.802
1.809
1.790
1.810
1.797
1.791
1.791
1.789
1.660
1.656
1.636
1.734
1.699
1.715
1.680
1.714
1.689
1.679
1.715
1.710
1.427
1.419
1.401
1.483
1.449
1.464
1.437
1.456
1.441
1.434
1.464
1.449
1.270
1.266
1.264
1.293
1.284
1.292
1.289
1.287
1.284
1.284
1.298
1.294
1.300
1.297
1.294
1.329
1.323
1.327
1.323
1.324
1.321
1.320
1.333
1.326
1.428
1.426
1.422
1.447
1.443
1.447
1.444
1.446
1.442
1.442
1.450
1.443
HF/6-311++G(3df,3pd)
B3LYP/aug-cc-pVDZ
B3LYP/aug-cc-pVTZ
B3LYP/6-31++G(d,p)
B3LYP/6-31++G(3df,3pd)
B3LYP/6-311++G(d,p)
B3LYP/6-311++G(2df,2pd)
B3LYP/6-311++G(3df,3pd)
MP2/6-31++G(d,p)
MP2/6-311++G(d,p)
set of polarization functions produced practically coinci-
dent results.
B3LYP can only be said to reproduce the experimental
spectra better than HF if the whole frequency range is
taken into consideration, because accidental cancellation
of errors lead to an excellent agreement between the exper-
imental mS@O stretching frequencies and the correspond-
ing HF calculated data. Very interestingly, it was found
that Pople’s type basis sets perform much better than Dun-
ning’s correlation consistent basis sets having the same
number of polarization functions to calculate the geometry
and infrared spectrum of the studied pseudosaccharin. On
the whole, the B3LYP/6-311++G(3df,3pd) calculations
give the best overall agreement between theoretically pre-
dicted and observed spectral and geometrical features,
and then, among the different alternatives tested, appear
as the most convenient theoretical approach to the study
of the type of compounds under consideration.
The typical geometry for substituted pseudosaccharyl
ethers is predicted by the calculations to be preserved in
MBID. As mentioned in the Section 1, the very character-
istic feature of this group of compounds is its uncommon
CRAOACA linkage (where R = heteroaromatic ring and
A = allyl, benzyl, naphtyl or aryl group), which contains
˚
a short CRAO bond (1.311–1.335 A) and a much longer
˚
(1.417–1.488 A) CAAO bond, together with a large
CAOAC angle (115.9–118.7ꢀ) that indicates a nearly sp2
hybridization of the oxygen atom [9,10,13,16–19]. The
observed features are a result of the strong electronegativ-
ity of the pseudosaccharyl group, quantitatively close to
that of the fluorine atom [8]. The B3LYP/6-
311++G(3df,3pd) predicted CRAO and CAAO bond
˚
lengths in MBID are equal to 1.320 and 1.442 A, respec-
Finally, the latter level of theory was applied to analyze
the vibrational spectra and the geometry of MBID. In
agreement with the experiment, the CAOAC linkage in
MBID is predicted by these calculations to exhibit consid-
tively, well within the expected value ranges, whereas the
calculated CRAOACA angle of 117.0ꢀ is in agreement with
the suggested close to sp2 hybridization of the oxygen ether
atom.
˚
˚
erably short (1.320 A) and long (1.442 A) (N@)CAO and
(H3)CAO bonds, respectively, and a hybridization of the
central oxygen atom close to sp2 (the CAOAC angle is pre-
dicted to be ca. 117ꢀ). This CAOAC bonding pattern fits
the well-known high reactivity of MBID upon thermal
rearrangement, which has been shown to result in easy
selective [1,30]-isomerization of the compound.
4. Conclusions
Quantum-chemical methods have been applied to model
the geometry and the infrared spectra of MBID, a simple
substituted pseudosaccharyl ether. The calculations were
done at different levels of theory, employing both ab initio
(HF, MP2) and DFT (B3LYP) methods combined with
different basis sets [6-31++G(d,p), 6-31++G(3df,3pd), 6-
311++G(d,p), 6-311++G(2df,2pd), 6-311++G(3df,3pd),
aug-cc-pVDZ and aug-cc-pVTZ]. The results of the calcu-
lations were evaluated based on their relative capability to
predict correctly the IR spectra of monomeric MBID
obtained in low-temperature argon and krypton matrices.
An extensive set of polarization functions is necessary to
correctly reproduce both the geometry and frequencies of
the hypervalent S@O bonds. However, the proper choice
of the method and basis set is also crucial in order to obtain
a good general agreement with the experimental data, i.e.,
Acknowledgements
Calculations were partially done at the Academic Com-
puter Center ‘‘Cyfronet’’, Krakow, Poland (Grant KBN/
SGI_ORIGIN_2000/UJ/044/1999), which is acknowledged
for computing time. The research was supported by the
Portuguese Fundac¸ao para a Cieˆncia e a Tecnologia (Grant
˜
FCT #SFRH/BPD/17081/2004 and projects POCI/QUI/
59019/2004 and POCI/QUI/58937/2004). AGZ is member
of the research career Conicet (National Research Council,
Argentina).

A. Kaczor et al. / Journal of Molecular Structure 876 (2008) 77–85
85
[25] G. Jovanovski, S. Tanceva, B. Soptrajanov, Spectrosc. Lett. 28 (1995)
1095.
Appendix A. Supplementary data
[26] A. Borba, A. Gomez-Zavaglia, P. Simoes, R. Fausto, Spectrochim.
Acta, Part A – Mol. Biomol. Spectrosc. 61 (2005) 1461.
[27] A. Borba, A. Gomes-Zavaglia, P. Simoes, R. Fausto, J. Phys. Chem.
A 109 (2005) 3578.
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/
j.molstruc.2007.06.004.
[28] A.H. Zeng, L. Yu, Y. Wang, Q.Y. Kong, Q. Xu, M.F. Zhou, J. Phys.
Chem. A 108 (2004) 6656.
[29] A. Kaczor, T. Melo, M.I.L. Soares, R. Fausto, J. Phys. Chem. A 110
(2006) 6531.
References
[1] E.B. Wilson, Phys. Rev. 45 (1934) 706.
[2] M. Otten, W. von Deyn, S. Engel, R.L. Hill, U. Kardorff, M. Vossen,
P. Plath, H. Walter, EP0961774, (1999).
[3] F. Zani, M.R. Mingiardi, C.A. Maggiali, P. Mazza, Farmaco 51
(1996) 707.
[4] M. Mor, F. Zani, P. Mazza, C. Silva, F. Bordi, G. Morini, P.V.
Plazzi, Farmaco 51 (1996) 493.
[5] F. Zani, P. Vicini, Arch. Pharm. 331 (1998) 219.
[6] P.I. Eacho, P.S. Foxworthy-Mason, H.-S. Lin, J.E. Lopez, M.K.
Mosior, M.E. Richett, EP1274708, (2006).
[7] A.F. Brigas, R.A.W. Johnstone, Tetrahedron Lett. 31 (1990) 5789.
[8] J.A.C. Alves, J.V. Barkley, A.F. Brigas, R.A.W. Johnstone, J. Chem.
Soc. Perkin Trans. 2 (1997) 669.
[9] N.C.P. Araujo, A.F. Brigas, M.L.S. Cristiano, L.M.T. Frija, E.M.O.
Guimaraes, R.M.S. Loureiro, J. Mol. Catal. A – Chem. 215 (2004)
113.
[10] L.M.T. Frija, M.L.S. Cristiano, E.M.O. Guimaraes, N.C. Martins,
R.M.S. Loureiro, J.F. Bickley, J. Mol. Catal. A – Chem. 242 (2005)
241.
[30] A.D. Becke, Phys. Rev. A 38 (1988) 3098.
[31] C.T. Lee, W.T. Yang, R.G. Parr, Phys. Rev. B 37 (1988) 785.
[32] P. Csaszar, P. Pulay, J. Mol. Struct. 114 (1984) 31.
[33] O. Farkas, H.B. Schlegel, J. Chem. Phys. 111 (1999) 10806.
[34] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb,
J.R. Cheeseman, V.G. Zakrzewski, J.A. Montgomery, Jr., R.E.
Stratmann, J.C. Burant, S. Dapprich, J.M. Millam, A.D. Daniels,
K.N. Kudin, M.C. Strain, O. Farkas, J. Tomasi, V. Barone, M. Cossi,
R. Cammi, B. Mennucci, C. Pomelli, C. Adamo, S. Clifford, J.W.
Ochterski, G.A. Petersson, P.Y. Ayala, Q. Cui, K. Morokuma, P.
Salvador, J.J. Dannenberg, D.K. Malick, A.D. Rabuck, K. Raghav-
achari, J.B. Foresman, J. Cioslowski, J.V. Ortiz, A.G. Baboul, B.B.
Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R.
Gomperts, R.L. Martin, D.J. Fox, T. Keith, M.A. Al-Laham, C.Y.
Peng, A. Nanayakkara, M. Challacombe, P.M.W. Gill, B. Johnson,
W. Chen, M.W. Wong, J.L. Andres, C. Gonzalez, M. Head-Gordon,
E.S. Replogle, J.A. Pople, GAUSSIAN 98, Revision B.01. In
Gaussian, Inc., Pittsburgh, PA, 2001.
[35] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb,
J.R. Cheeseman, J.A. Montgomery, Jr., T. Vreven, K.N. Kudin, J.C.
Burant, J.M. Millam, S.S. Iyengar, J. Tomasi, V. Barone, B.
Mennucci, M. Cossi, G. Scalmani, N. Rega, G.A. Petersson, H.
Nakatsuji, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa,
M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, M. Klene, X.
Li, J.E. Knox, H.P. Hratchian, J.B. Cross, V. Bakken, C. Adamo, J.
Jaramillo, R. Gomperts, R.E. Stratmann, O. Yazyev, A.J. Austin, R.
Cammi, C. Pomelli, J.W. Ochterski, P.Y. Ayala, K. Morokuma, G.A.
Voth, P. Salvador, J.J. Dannenberg, V.G. Zakrzewski, S. Dapprich,
A.D. Daniels, M.C. Strain, O. Farkas, D.K. Malick, A.D. Rabuck,
K. Raghavachari, J.B. Foresman, J.V. Ortiz, Q. Cui, A.G. Baboul, S.
Clifford, J. Cioslowski, B.B. Stefanov, G. Liu, A. Liashenko, P.
Piskorz, I. Komaromi, R.L. Martin, D.J. Fox, T. Keith, M.A. Al-
Laham, C.Y. Peng, A. Nanayakkara, M. Challacombe, P.M.W. Gill,
B. Johnson, W. Chen, M.W. Wong, C. Gonzalez, J.A. Pople,
GAUSSIAN 03, Revision C.02. In Gaussian, Inc., Wallingford, CT,
2004.
[36] P. Pulay, G. Fogarasi, F. Pang, J.E. Boggs, J. Am. Chem. Soc. 101
(1979) 2550.
[37] J.M.L. Martin, C. Van Alsenoy, GAR2PED, In University of
Antwerp, 1995.
[38] V. Typke, M. Dakkouri, J. Mol. Struct. 599 (2001) 177.
[39] V. Jonas, G. Frenking, Chem. Phys. Lett. 177 (1991) 175.
[40] D. Vijay, U.D. Priyakumar, G.N. Sastry, Chem. Phys. Lett. 383
(2004) 192.
[11] B.J. Hussey, R.A.W. Johnstone, I.D. Entwistle, Tetrahedron 38
(1982) 3775.
[12] A.F. Brigas, R.A.W. Johnstone, J. Chem. Soc. Chem. Commun.
(1994) 1923.
[13] J.V. Barkley, M.L.S. Cristiano, R.A.W. Johnstone, R.M.S. Loureiro,
Acta Crystallogr. C – Cryst. Struct. Commun. 53 (1997) 383.
[14] S. Brodersen, A. Langseth, Mat. Fys. Skr. Dan. Vid. Selsk 1 (1956).
[15] A.F. Brigas, R.A.W. Johnstone, J. Chem. Soc. Perkin Trans. 1 11
(2000) 1735.
[16] J.A.C. Alves, A.F. Brigas, R.A.W. Johnstone, Acta Crystallogr. C –
Cryst. Struct. Commun. 52 (1996) 1576.
[17] A.F. Brigas, P.M. Goncalves, R.A.W. Johnstone, Acta Crystallogr. C
– Cryst. Struct. Commun. 54 (1998) 251.
[18] A.F. Brigas, R.A.W. Johnstone, Acta Crystallogr. C – Cryst. Struct.
Commun. 52 (1996) 1293.
[19] N.C.P. Araujo, P.M.M. Barroca, J.F. Bickley, A.F. Brigas, M.L.S.
Cristiano, R.A.W. Johnstone, R.M.S. Loureiro, P.C.A. Pena, J.
Chem. Soc. Perkin Trans. 2 (2002) 1213.
[20] M.A.R. Matos, M.S. Miranda, V.M.F. Morais, J.F. Liebman, Mol.
Phys. 103 (2005) 221.
[21] M. Remko, J. Phys. Chem. A 107 (2003) 720.
[22] I.G. Binev, B.A. Stamboliyska, E.A. Velcheva, Spectrochim. Acta,
Part A – Mol. Biomol. Spectrosc. 52 (1996) 1135.
[23] Y.I. Binev, C.T. Petkov, L. Pejov, Spectrochim. Acta, Part A – Mol.
Biomol. Spectrosc. 56 (2000) 1949.
[24] P. Naumov, G. Jovanovski, Struct. Chem. 11 (2000) 19.