The Vibrational Circular Dichroism Pattern of the ν(C=O) Bands in Isoindolinones

Article abstract of DOI:10.1002/cphc.201800244

The IR and vibrational circular dichroism (VCD) spectra of both enantiomers of Me-, iPr-, nBu-, Ph-, and CH₂Ph-substituted isoindolinones in solution and KBr pellets were measured and interpreted by DFT calculations. The spectra in solution revealed no important differences in the C=O stretching vibration region while the interpretation of very distinct spectra taken in pellets required determining the crystal structures. The studied compounds crystallized in the P2₁2₁2₁ (Me, iPr, CH₂Ph), P3₁ (nBu), and P2₁ (Ph) space groups. We found that the quality of simulated spectra strongly depends on the substituent, the structure of the molecular cluster assumed, basis set, and use of the dispersion correction. The IR spectra can be reproduced well based on the simplest linear arrangement of hydrogen-bonded chains mimicking the molecular arrangement in the crystals. We found no common approach to reproduce all the registered VCD spectra in the crystal phase. For the Me and nBu isoindolinones, the VCD pattern was the best reproduced by full optimization of the selected large molecular clusters. For iPr, Ph and CH₂Ph derivatives optimizing only the position of H-atoms in a fragment frozen as in the crystal provides the best results. Such an approach can reduce the computation time from months to one week.

Full text of DOI:10.1002/cphc.201800244

Accepted Article
Title: The VCD pattern of the ν(C=O) bands in isoindolinones
Authors: Joanna Ewa Rode, Krzysztof Lyczko, Magdalena Jawiczuk,
Robert Kawęcki, Wojciech Stańczyk, Agnieszka Jaglińska,
and Jan Cz. Dobrowolski
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemPhysChem 10.1002/cphc.201800244
Link to VoR: http://dx.doi.org/10.1002/cphc.201800244
A Journal of
www.chemphyschem.org

10.1002/cphc.201800244
ChemPhysChem
FULL PAPER
The VCD pattern of the ν(C=O) bands in isoindolinones
Joanna E. Rode,*[a] Krzysztof Lyczko,[a] Magdalena Jawiczuk,[b] Robert Kawęcki,[c] Wojciech Stańczyk,[c]
Agnieszka Jaglińska,[c] Jan Cz. Dobrowolski[a]
Abstract: The IR and VCD spectra of both enantiomers of Me, iPr,
nBu, Ph and CH2Ph substituted isoindolinones in solution and KBr
pellets were measured and interpreted with DFT calculations. The
spectra in solution revealed no important differences in the C=O
stretching vibration region while the interpretation of very distinct
spectra taken in pellets required determining the crystal structures.
The studied compounds crystallized in the P212121 (Me, iPr, CH2Ph),
P31 (nBu), and P21 (Ph) space groups. We found that the quality of
simulated spectra strongly depends on the substituent, the structure
of the molecular cluster assumed, basis set, and use of the dispersion
correction. The IR spectra can be reproduced well based on the
simplest linear arrangement of hydrogen-bonded chains mimicking
the molecular arrangement in the crystals. We found no common
approach to reproduce all the registered VCD spectra in the crystal
phase. For the Me and nBu isoindolinones, the VCD pattern was the
best reproduced by full optimization of the selected large molecular
clusters. For iPr, Ph and CH2Ph derivatives optimizing only the
position of H-atoms in a fragment frozen as in the crystal provides the
best results. Such an approach can reduce the computation time from
months to one week.
The vibrational bands of amides have been studied in IR and
Raman spectra for many decades.e.g.,[5] They have been studied
with vibrational circular dichroism spectroscopy (VCD) since the
late 1980s.[6] However, isoindolinones have only recently been
studied with the VCD method, and even then, only the amide’s
C=O stretching vibrations were used to determine their absolute
configuration.[7] Using this region of VCD spectra can be
challenging. First, unlike in IR, the VCD intensity of the C=O
stretching vibration band is weak or very weak.[8] Secondly, the
solvent can invert the sign of the C=O VCD band.[9] Thirdly, the
C=O bands may be non-robust, i.e. even a very small structural
change, a small external perturbation or use of the wrong
computational level can produce a sign change of these
bands.[6d,8a,10] Moreover, unlike the case of IR, empirical rules
governing the intensity and sign of the VCD bands are still
unknown and interpretation of the VCD spectra must rely on
quantum chemical computations.
(S)-Me
(S)-iPr
(S)-nBu
(S)-Ph
(S)-CH2Ph
Introduction
Scheme 1. The (S)-enantiomers of the studied isoindolinones.
The isoindolinone molecule consists of fused benzene and γ-
lactam rings (Scheme 1). The five membered γ-lactam ring
contains the amide HN-C=O moiety, which is present in many
important molecules.[1] Isoindolinones also occur in nature; they
are often bioactive and are widely used as building blocks in drug
synthesis.[2] In particular, 3-substituted chiral isoindolinones are
promising leads in pharmaceutical research.[3] A robust, reliable,
and straightforward synthesis of enantioenriched 3-substituted
isoindolinones has recently been developed by us.[4] The chiral
derivatives studied in this paper were obtained using that same
method.
We have recently observed that VCD spectra measured in
KBr pellets are much more intense than those measured in
solutions.[11] However, VCD spectra measured in solution were
completely different from those taken in KBr. This dichotomy
results from differences between conformational freedom in
solution and in the crystal phase. Surprisingly, although
reproducing the VCD spectrum in solution was very challenging
and required a detailed consideration of the conformational
equilibria and careful selection of theoretical approach, prediction
of the VCD spectra taken from a KBr pellet was much easier, even
when based on tetramers or dimers ordered as in the crystal.
VCD measurements in pellets may seem more difficult than in
solutions because of possible artefacts that may arise from the
light scattering on crystallites whose size is close to the
wavelength of the incident beam, linear birefringence from the
solid state medium and light polarization on crystal surfaces.[12]
Nevertheless, careful grinding makes dispersion of the solid in
KBr/KCl pellets much better, greatly enhancing their homogeneity.
Additionally, dual polarization modulation and rotation of the
sample holder help also to overcome these problems, what is the
best practice of the present stage of VCD instrumental
development.[13] Moreover, the use of a pellet medium for
insoluble compounds facilitates measurements which often
require a small amount of the sample. On the other hand,
[a]
[b]
Dr. J. E. Rode, Dr. K. Łyczko, Prof. Jan. Cz. Dobrowolski
Institute of Nuclear Chemistry and Technology
16 Dorodna Street, 03-195 Warsaw, Poland
E-mail: j.rode@ichtj.waw.pl
Dr. M. Jawiczuk
Institute of Organic Chemistry, Polish Academy of Sciences
Kasprzaka 44/52, 01-224 Warsaw, Poland
Current address: Centre of New Technologies, University of
Warsaw, Banacha 2C, 02-097 Warszawa, Poland
[c]
Prof. R. Kawęcki, MSc W Stańczyk, MSc A. Jaglińska
Siedlce University, Faculty of Science
3 Maja Street No 54, 08-110 Siedlce, Poland
Supporting information for this article is given via a link at the end of
the document.
This article is protected by copyright. All rights reserved.

10.1002/cphc.201800244
ChemPhysChem
FULL PAPER
Figure 1 The IR and VCD spectra of polycrystalline powders of the (S)-isoindolinones pressed in a KBr pellet.
reproduction of experimental solid-state VCD spectra can be a
challenge even if the X-ray structure is known. Indeed, large
molecular clusters may have to be calculated at sufficiently high
levels of theory. However, for large molecular systems,
optimizations with frozen heavy atoms and relaxed hydrogens
can be very efficient.[14] For even larger systems, fragment-base
quantum chemical calculations can be successfully applied.[15]
In this paper we present the IR and VCD spectra of both
enantiomers of Me, iPr, nBu, Ph and CH2Ph substituted
isoindolinones synthesized using the newly-developed method
(Scheme 1). Next, we demonstrate that the spectra in solution do
not display significant differences. Then, we show that the solid-
state IR and VCD spectra exhibit very distinct patterns in the
ν(C=O) range. To interpret these patterns, it was necessary to
solve the crystal structures and to calculate the spectra based on
molecular clusters cut out from the crystals. We found that there
is no uniform way to simulate all the VCD spectra, although the
simulation of IR spectra is relatively easy. In the paper we discuss
ways to obtain a satisfactory reproduction of the experimental
spectra for each of the derivatives.
On the other hand, for the polycrystalline powder of
isoindolinones pressed in KBr pellets, we registered quite intense
ν(C=O) VCD bands (Figs. 1 and S4). However, two or more C=O
IR bands in this region suggested some couplings of the C=O
modes. Moreover, in both IR and VCD spectra, the isoindolinone
substituents strongly changed the spectral pattern. It appeared
that for five substituents, we observe three types of IR and VCD
spectra. In the IR spectra the ν(C=O) spectral pattern basically
consists of two bands at ca. 1710 and 1680 cm-1. However, for
Me the two bands exhibit similar intensity but slightly different
width, for nBu the former is so dominant that the latter appears
rather as a shoulder, whereas for iPr, Ph, and CH2Ph the former
is slightly less intense and narrower than the latter (Fig. 1). The
two band separation, splitting, seems to be independent from the
substituent size or structure. The VCD patterns can be grouped
in the same way. Both Me and nBu substituted isoindolinones
form separate types. However, in the first approximation they both
can be described as if they were composed of two pairs of +/- and
-/+ signals separated differently: for nBu the signals seem to be
located closer than for Me (Fig. S5). For iPr, Ph, and CH2Ph, two
VCD bands are observed: a positive one at higher and a negative
one at lower wavenumbers (Fig. 1).
Results and Discussion
Verification of the 1:1 cyclic dimers hypothesis
Experimental IR and VCD spectra
The hypothetical ν(C=O) coupling seen in the IR spectra can be
explained as a result of the formation of cyclic dimers. In this case
one ν(C=O) band would correspond to the antisymmetric mode
and the other would correspond to the symmetric one. The former
band would absorb at higher wavenumbers and have a higher IR
intensity, and the latter would exhibit the opposite behavior.
Moreover, for the racemic Ph substituted isoindolinone, a type I
dimer (Fig. 2) was observed in the C2/c centrosymmetric
crystal,[16] and this kind of dimer has also been observed in similar
systems.[17] Thus, we first compared the DFT-calculated IR
spectra of 1:1 cyclic isoindolinone dimers with the experimental
ones. However, for all considered molecules no concordance was
observed between the IR and VCD spectra (Fig. 2 (Me), Figs. S6-
It is known that, in molecules with a single C=O vibrator, the
intensity of the ν(C=O) VCD bands is weak.[8] This is also the case
for the studied isoindolinones dissolved in CDCl3 (Fig. S1). Two
IR bands in the C=O stretching vibration range, located at ca.
1700 and 1620 cm-1, can be assigned to the ν(C=O) and β(NH)
vibrations, respectively. The IR spectra of the Me, iPr, nBu, and
CH2Ph derivatives measured in the solution are not particularly
distinguishable. Even so, the VCD bands are very weak and often
have the magnitude of measurement noise and for Ph the
solubility was too low to measure its spectrum (Fig. S1). Yet, this
is in line with computational predictions for both monomers and
dimers (Figs. S2 and S3). Moreover, the VCD spectra below
1600 cm-1 were hardly measurable because of the low solubility
of the studied compounds.
This article is protected by copyright. All rights reserved.

10.1002/cphc.201800244
ChemPhysChem
FULL PAPER
(S)-Me-I-dim
(S)-Me-II-dim
(S)-Me-I-stack
(S)-Me-II-stack
(S)-Me-I-stack2
(S)-Me-II-stack2
Figure 2 The schemes of possible isoindolinone dimers of type I and II. Comparison of experimental IR and VCD spectra of (S)–Me substituted isoindolinone with the
calculated ones obtained for dimers (B3LYP/TZVP) and stacking structures (B3LYP-D3/TZVP ). The VCD intensities of the (S)-Me-I-dim and (S)-Me-II-dim structures
were multiplied by factors of 20 and 40, respectively.
S9 (the other substituents)). Therefore, we checked whether a
stacking arrangement of dimers can improve the agreement. And
indeed, for some conformers of iPr and CH2Ph derivatives, the
concordance with the experimental IR and VCD spectra was
improved, while for others (Me, nBu, Ph) it was not (Fig. 2 (Me),
Figs. S10-S13 (the other substituents)). Thus, we considered
much weaker dimers of the II type and their stacking structures
(Fig. 2 (Me), Figs. S14-S21 (the other substituents)). In such sole
dimers, instead of two well-resolved IR bands, a band and its
shoulder were predicted. However, for the type II dimers of iPr,
Ph and CH2Ph isoindolinones, the +/- pattern observed in
experimental VCD spectra was obtained (Figs. S14, S16, S17).
Nevertheless, for larger structures built from the type II dimers,
dispersion forces. The non-planarity of dimers significantly
influences the calculated VCD intensities in the ν(C=O) range (Fig.
S22). On the other hand, the stacking interactions require good
reproduction of the dispersion forces and in this case the
Grimme's correction guarantee that the layers interact. Ultimately,
the calculated spectra based on dimer or stacking dimer
structures disagreed with the solid-state VCD spectra of
substituted isoindolinones and determination of the crystal
structures and spectra simulations based closely on the crystal
structures was necessary.
Structures of crystals
simulated VCD spectra disagree with the experimental ones (Figs. The X-ray structures of all studied chiral isoindolinones revealed
S18-S21). It is important to note that some stacking structures did
not converge if Grimme's D3 correction for the dispersion forces
was not applied. However, using the D3 correction for sole dimers
is not recommended because it can lead to strong non-planarity.
This is because the main interaction in dimers is hydrogen
bonding which is mostly electrostatic and in this particular case
the Grimme's correction probably overestimates role of the
the presence of infinite hydrogen bond chains (Fig. 3). The
molecular arrangements of these crystals are completely different
than those established so far for the racemic mixture.[16]
Furthermore, different substituents gave rise to different
symmetries: Me, iPr and CH2Ph crystallized in the P212121, nBu
in P31 while Ph crystallized in P21 space groups (Table 1 and S1).
This article is protected by copyright. All rights reserved.

10.1002/cphc.201800244
ChemPhysChem
FULL PAPER
view 1
view 2
view 3
(S)-Me
view 4
(S)-nBu
(S)-iPr
(S)-Ph
(R)-CH2Ph
Figure 3 Different views of the crystal arrangement in the studied isoindolinones.
This article is protected by copyright. All rights reserved.

10.1002/cphc.201800244
ChemPhysChem
FULL PAPER
In the net of hydrogen bonds, formed by so-called catemers,[18]
each molecule is simultaneously a proton and electron donor and
acceptor for the NH and C=O groups of the neighboring
molecules (Fig. 3). Generally, the neighboring molecules are
placed in two different planes; however, in the case of the nBu
derivative crystallizing in the P31 space group, binary interactions
between three planes occur (Fig. 3). On the other hand, the other
crystals belong to the P21 symmetry group and among them the
Me substituted isoindolinone has a different VCD spectrum. This
may be due to a different NH∙∙∙OC torsion angle, which is ca. 180°
for Me and ca. 90° for iPr, Ph and CH2Ph (Table 1).
Table 1 Comparison of crystal symmetry and selected intermolecular hydrogen
bond geometries (d-distance, a-angle, t-torsion angle) determined in the X-ray
measurements of the substituted isoindolinones.
Space
group
Z
d(Å)
(H∙∙∙O)
1.918
2.124
1.991
1.901
2.033
a
a
t
(CO∙∙∙H)
134.7
121.1
135.5
134.3
128.3
(NH∙∙∙C)
163.5
172.3
174.6
174.3
162.3
(NH∙∙∙OC)
-175.9
135.8
89.6
(S)-Me1
(S)-nBu2
(S)-iPr1
(S)-Ph3
(R)-CH2Ph1
P 212121
P 31
4
3
4
2
4
P 212121
P 21
89.4
P 212121
-102.9
Crystal system: 1orthorhombic, 2trigonal, 3monoclinic
LINEAR (L10)
STACKING (S12)
PAIRS OF COUPLETS (P16)
input
input
input
output
Output
output
Figure 4 Two views of the input and output structures for the (S)-Me substituted isoindolinone fragments extracted from the X-ray crystal structure. The B3LYP-D3/TZVP
level was used for L and B3LYP-D3/6-31G for S and P structures. L10, S12 and P16 denote the number of units considered.
In the Me crystal the herringbone stacking pattern is
observed and the distinct layers can be distinguished (Fig. 3). The
herringbone motifs are also observed in iPr and CH2Ph crystals,
but they are interlocked by substituent interactions. In the Ph
crystal, the herringbones are parallel but still are interlocked by
substituents. Moreover, the layers are visibly separated and the
border between them forms the infinite H-bond chains (Fig. 3). In
the Me crystal, the O1∙∙∙O1 distance equals 4.953(1) Å, while in
nBu, due to the substituent steric hindrance, it is as large as
8.414(1) Å. The distance between the stacking layers in the Me
This article is protected by copyright. All rights reserved.

10.1002/cphc.201800244
ChemPhysChem
FULL PAPER
Figure 5 Comparison of the experimental solid state IR (left) and VCD (right) spectra of the (S)-Me isoindolinone with the calculated ones based on the linear,
stacking and packing multimer models. For clarity, the wavenumbers of B3LYP-D3/TZVP (L) and B3LYP-D3/6-31G (S,P) calculated spectra were scaled.
crystal determined on the basis of isoindolinone ring planes
instead of O1 atoms is much shorter and equals 3.3 Å.
Furthermore, the distance between the herringbone layers
estimated from the distance between H3C planes is only 1.6 Å
(Fig. 3). The distances between the stacking layers in iPr, Ph and
CH2Ph are equal to 2.4, 3.0 and 2.3 Å, respectively. In conclusion,
the difference between Me and the other P21 crystals is probably
due to a larger distance between crystal layers and weaker
"Velcro fastening" between them.
31G* basis sets while the TZVP one was used only for smaller
fragments. Hereafter, we show only the details for the (S)-Me
system and the final results for the other systems while the other
details can be found in the Supplementary Information file.
First, we observed that in the course of optimization the initial
linear structures are strongly deformed and converge to a kind of
helix (Fig. 4). Such deformations introduce undesirable
contributions to the calculated VCD spectra (L-structures, Fig. 5).
In particular, the VCD pattern in the 1675-1750 cm-1 range does
not match the experimental one. However, it is worth noting that
the larger the number of units, the better the agreement between
the calculated and experimental IR spectra (Fig. 5).
Simulations of solid-state IR and VCD spectra
In this section we focus on the influence of several technical
parameters on the quality of the simulated solid-state IR and VCD
spectra of isoindolinones. These parameters are as follows: the
fragment’s crystal size and type, the basis set size and type, the
presence or absence of Grimme's D3 correction for dispersion
forces, and simplifications of the multimer optimization, such as
freezing the heavy atoms.
We considered the following fragments cut from crystals: (1)
the single linear multimers (L) of increasing lengths mimicking the
infinite H-bond chains; (2) a stacking (S) of two linear multimers;
(3) pairs of stacking couplets (P) extracted from different layers.
When selecting the fragments, it is important to conserve the
correspondence between the partners and to preserve the
hydrogen bond saturation. Examples of the structures prepared
for the (S)-Me isoindolinone are presented in Fig. 4. The B3LYP
calculations of all systems were combined with the 6-31G and 6-
Use of the larger TZVP basis set was hardly possible for the
S- and P-fragments. The TZVP basis set reproduces the band
positions, intensities and the band splitting much better than the
6-31G one. However, we found that for the L10 system the TZVP
and 6-31G basis sets yield similar overall patterns but visibly
different band separations (Fig. S23). On the other hand, the 6-
31G(d) basis set failed to reproduce both structure and spectra
regardless of whether Grimme's correction was used or not (Figs.
S23 and S24).
Optimization of the S- and P-fragments leads to smaller
deformations than for the linear systems (S12 and P16 vs. L10,
Fig. 4). Also, the IR spectra are well-reproduced regardless of the
S- or P-arrangement and the number of units considered (Fig. 5).
However, none of these models properly reproduce the IR and
VCD band intensities at ca. 1625 cm-1. Nevertheless, the P16
system with interaction between pairs of stacking couplets is a
This article is protected by copyright. All rights reserved.

10.1002/cphc.201800244
ChemPhysChem
FULL PAPER
Figure 6 Comparison of the experimental solid state IR and VCD spectra with the spectra computed based on the H-opt approach. The calculations were carried
out for different fragments of the (S)-Me isoindolinone which mimicked the arrangement of molecules in the crystal. The calculations were performed using the
B3LYP-D3 functional combined with the 6-31G and TZVP basis sets.
(S)-nBu-L9
(S)-nBu-P6
(S)-nBu-P9
(S)-iPr-L8
(S)-Ph-L8
(S)-iPr-P8
(S)-Ph-P8
(S)-iPr-P12
(S)-Ph-P12
(R)-CH2Ph-L8
(R)-CH2Ph-P8
(R)-CH2Ph-P12a
(R)-CH2Ph-P12b
Figure 7 X-ray structures of the substituted isoindolinones considered as inputs for the DFT calculations.
This article is protected by copyright. All rights reserved.

10.1002/cphc.201800244
ChemPhysChem
FULL PAPER
good match for the experimental VCD spectrum (Fig. 5). In
particular, the experimental -/+/- VCD band pattern above 1700
cm-1 is correctly predicted. Finally, it is worth noting that the
calculated IR and VCD spectra are only strongly structure-
dependent in the ν(C=O) range, whereas below 1500 cm-1 they
remain resistant to changes of adopted model (Fig. S25).
To find a more efficient way of modelling the spectra, we
checked whether freezing the heavy atoms while simultaneously
relaxing the positions of the hydrogen atoms would improve the
optimization time to spectrum quality ratio. The optimization of H-
atoms is necessary because their positions as established in X-
ray measurements are not fully accurate.[19] Calculations within
such an approximation (hereafter referred as H-opt) are much
faster than full optimization: a matter of a week vs. months. In the
case of ECD, they lead to quite a good agreement with the
experimental spectra.[20] For the VCD spectra, the H-opt approach
was used only once and yielded a good fit even though the
optimized structures were not in minima.[14]
Basically, the IR spectra obtained through the H-opt approach
do not depend on the fragment of the (S)-Me crystal considered
(Fig. 6). On the other hand, the H-opt simulated VCD spectra
strongly depend on the considered fragment of the crystal, but
only weakly on the number of the isoindolinone units taken into
account. Unfortunately, even for the largest P16 system, this
approach does not reflect the complexity of the VCD ν(C=O)
bands pattern. Finally, it is worth to add that below 1600 cm-1, the
VCD spectra simulated using different fragment types look similar
(Figs. S26).
B3LYP-D3
B3LYP
The selection of different fragments of crystals was done for
all studied isoindolinones calculated at different levels (Table S2).
The largest of the studied systems are presented in (Fig. 7). For
all of them, the structure and spectrum analyses were identical to
those of the Me derivative, except for the number of molecules
which had to be related to Z - the number of molecules in the
elementary cell (Table 1). The differences in crystal symmetry
prevent the selected molecular clusters from always having the S-
type of arrangement.
(S)-nBu-P6-tzvp-D3
(S)-nBu-P6-tzvp
The best fit of the experimental IR and VCD spectra of (S)-
nBu was achieved in full optimization of the L6 and P6 structures
at the B3LYP-D3/TZVP level (Fig. 8 and Fig. S27 for the other
levels of theory). The 1700 cm-1 and 1660 cm-1 IR bands
correspond to the +/- and -/+ VCD patterns, respectively (Fig.
S28). Oddly enough, the B3LYP-D3/TZVP calculations reproduce
the VCD spectrum well but destroy the geometry mimicking the
symmetry in the crystal, whereas the B3LYP/TZVP calculations
without Grimme's correction conserve such an arrangement but
improperly reproduce the IR and VCD spectra (Fig. 8). Although
the B3LYP-D3/TZVP calculations confirm the intuitive assignment
of the experimental bands to the coupled ν(C=O) and β(N-H)
vibrators, the modes cannot be resolved into in-phase and in-
counter-phase ones. The weak IR band at ca. 1625 cm-1
corresponds to ν(C=C) stretching vibrations of the six membered
ring. The H-opt approach, by definition, conserves the geometries
and, for the nBu derivative, it does not yield a proper reproduction
of the VCD spectrum (Fig. S29).
(S)-nBu-L6-tzvp-D3
(S)-nBu-L3-tzvp-D3
(S)-nBu-L6-tzvp
(S)-nBu-L3-tzvp
Figure 8 The (S)-nBu isoindolinone structures mimicking the crystal phase
arrangement and the comparison between the experimental and scaled
theoretical IR and VCD solid-state spectra. The B3LYP-D3 (solid) or B3LYP
(dashed) functionals are combined with the TZVP basis set.
sometimes also observed in other crystals of carbonyl
derivatives.[6l,17] Unlike for Me, for the above derivatives full
optimization significantly deforms the initial crystal-derived
geometry, especially for the P-fragments (Figs. S30-S35). On the
The experimental solid-state VCD spectra of the iPr, Ph and
CH2Ph derivatives in the (S)-configuration exhibit the +/- pattern
This article is protected by copyright. All rights reserved.

10.1002/cphc.201800244
ChemPhysChem
FULL PAPER
(a) Full-opt approach
(b) H-opt approach
(S)-iPr-L8-TZVP
(S)-Ph-P8-6-31G
(R)-CH2Ph-P12b-6-31G
Figure 9 The molecular clusters and IR and VCD spectra of the (S)-iPr, (S)-Ph and (R)-CH2Ph isoindolinones for which the best agreement with experiment
spectra was achieved: (a) full optimization and (b) H-opt approach. The B3LYP-D3 functional and the TZVP or 6-31G basis sets were applied. The
wavenumbers of the calculated spectra were scaled.
other hand, the linear multimers are less distorted and, generally,
the longer the linear multimer, the smaller the distortion. In fact, in
the L-multimers only the terminal molecules are disordered to
some extent.
In general, it is quite difficult to reproduce the solid state IR
and VCD spectra of the iPr, Ph and CH2Ph isoindolinone
derivatives in full optimization (Fig. S30-S35). This is mainly due
to limitations of the basis set size for the molecular clusters to fit
the molecular interactions in the crystal phase well. Thus, it is
difficult to recommend a common approach for all the derivatives
studied. Hereafter, we discuss only the best obtained results (Fig.
9).
For iPr, the best reproduction of the IR and VCD spectra was
achieved for the L8 structure calculated at the B3LYP-D3/TZVP
level (Fig. 9) and the same holds true for the 6-31G basis set (Fig.
S30). As for nBu, for iPr the use of Grimme's correction results in
a good reproduction of the VCD spectrum while it leads to
perturbed cluster symmetry. On the other hand, the lack of this
correction conserves the system symmetry, but the spectra do not
match the experimental ones well (Fig. S30 and S31). It is worth
emphasizing that the cost-efficient H-opt approach combined with
both the TZVP and 6-31G basis sets also reproduces the IR and
VCD spectra well, even for quite small molecular fragments (Fig.
S36).
reproduction of the IR and VCD spectra was achieved based on
the P8 structure calculated at the B3LYP-D3/6-31G level (Figs. 9
and S33). It is worth noticing that the H-opt approach also
provides quite a good description of the experimental spectra
(Figs. 9 and S37). However, it is necessary to add that, in spectra
calculated using the TZVP basis set, a relatively weak IR band at
ca. 1660 cm-1 can be interpreted as the one responsible for the
right-hand asymmetry of the experimental band located at ca.
1650 cm-1.
Again, for CH2Ph, the TZVP basis set could be used only for
the L4 and L6 structures, for which the agreement with the
experimental spectra is unsatisfactory (Fig. S34). The IR and VCD
spectra of the (R)-CH2Ph substituted isoindolinone was obtained
for the largest P12b structure calculated at the B3LYP-D3/6-31G
level (Figs. 9 and S35). It is worth noting that in this case the H-
opt approach also provides quite a good description of the
experimental spectra (Figs. 9 and S38). Analogously to the Ph
derivative, the apparent disagreement in calculations between the
experimental and theoretical IR spectra produced by the weak IR
band at ca. 1660 cm-1 can be assigned to the right-hand
asymmetry of the experimental band located at ca. 1650 cm-1.
Ending up, after going through such tedious calculations as in
this study, one desires to refer his/her results to simple
experimental rules. The appearance of bisignate VCD couplets as
in iPr, Ph and CH2Ph substituted isoindolonones brings to mind
“vibrational excitons’’ (VE) concept which links the VCD band's
sign sequence in the C=O stretching region with the mutual
For Ph, the TZVP basis set could be used only for the L4
and L6 structures, for which agreement with the experimental
spectra is unsatisfactory (Fig. S32). For this substituent, the best
This article is protected by copyright. All rights reserved.

10.1002/cphc.201800244
ChemPhysChem
FULL PAPER
arrangement of the C=O vibrators.[21] For the straightforward
application of the VE method a few criticalities were found.[22]
Nevertheless, we tested this method in our systems bearing in
mind that in isoindolinones the C=O vibrators are in two molecules
interacting by a H-bond, rather than in the same molecule as in
the systems for which the exciton method was developed for the
VCD spectra.[21] However, if we apply the concept directly by
measuring the appropriate dihedral angles between the C=O
moieties we end up with conclusions which are contradictory to
the original Taniguchi and Monde concept (Table S3).
Experimental Section
Synthesis All optically active 3-substituted isoindolinones were prepared
in a one-step synthesis as described earlier.[4] First, methyl 2-[(tert-butyl-
N-sulfinylimine)-methyl]benzoate was obtained from commercially
available (R) or (S) enantiomers of tert-butylsulfinamide and methyl 2-
formylbenzoate (Scheme 2). Then, the reaction of the Grignard reagent
with sulfinimine in CH2Cl2 produced the appropriate isoindolinones. The
products were purified by chromatography or trituration with a hexane-
diethyl ether mixture. The enantiomeric excess of the isoindolinones was
in the range of 58-91%.
Conclusions
In this paper we interpret the IR and VCD spectra in CDCl3
solution and in solid-state (in KBr pellets) for both enantiomers of
newly synthesized Me, iPr, nBu, Ph and CH2Ph substituted
isoindolinones. To this aim it was necessary to solve the crystal
structures and to calculate the spectra based on molecular
clusters cut from these crystals. The obtained results can be
summarized as follows:
i. The IR and VCD spectra of the studied isoindolinones in
CDCl3 solution do not reveal significant differences. There are
two IR bands at 1700 and 1620 cm-1, and the VCD spectra in
this region are almost too weak to measure. This is in line with
the calculations which for monomers and dimers predict very
weak VCD intensities of the bands in this region. The range
below 1600 cm-1 could not be measured because of the low
solubility of the studied compounds.
ii. For all studied chiral isoindolinones, the X-ray measurements
revealed the presence of infinite hydrogen bond chains. The
crystal symmetries found for the chiral crystals were the
following: Me, iPr and CH2Ph crystallized in the P212121, nBu
in the P31 and Ph in the P21 space group.
iii. The IR and VCD spectra of different isoindolinone derivatives
in crystals measured in the form of KBr pellets are very distinct.
They can be reproduced well by DFT calculations but the
agreement between the experimental and simulated spectra
strongly depends on the substituent, the structure of the
molecular cluster assumed, basis set, and use of the
dispersion correction.
Scheme 2 Synthesis of the (S)-isoindolinone derivatives.
IR and VCD measurements The substituted isoindolinones are hardly
soluble, thus registering the whole range of the VCD spectra in solution is
not always possible. However, all isoindolinones but the Ph-substituted
one are sufficiently soluble in CDCl3 to properly measure the ν(C=O)
stretching vibration IR bands. The solution VCD spectra were measured
for 45 minutes and corrected for the baseline given by the pure solvent.
The solid-phase IR and VCD spectra of substituted isoindolinones were
collected in the 2000–1000 cm−1 range. The sample weights were adjusted
to achieve IR absorbance between 0.4-0.6 in the most interesting ν(C=O)
region. The KBr pellets were prepared by mixing isoindolinone and KBr in
mortar instead of grinder, which destroys the isoindolinone crystals (ca.
2.0-2.5 vs. 100 mg, respectively) and pressed with ca. 10T for 3 minutes.
To average polarizations at crystallites, the pellets were assembled in a
rotating holder. To improve the signal-to-noise ratio, 4000 scans were
recorded and the measurements lasted approximately 45 minutes in total.
The baseline was corrected with a reference KBr pellet measured in the
same conditions. The spectra of both enantiomers were measured at a 4
cm−1 resolution using the single PEM FVS-6000 VCD spectrometer from
Jasco. The whole range spectra of both enantiomers of studied
isoindolinones are presented in the Supplementary Materials.
iv. For the IR spectra taken in KBr pellets, the simplest linear
arrangement of the hydrogen-bonded chains mimicking the
molecular arrangement in the crystals is sufficient for
reproducing the experimental spectra; however, the longer
the chain, the better the agreement.
v. We found no common approach to reproduce all the
registered VCD spectra in the crystal phase. For the Me and
nBu isoindolinones, the VCD pattern was best reproduced by
full optimization of the selected molecular clusters. The larger
the cluster, the better the agreement with the experimental
data.
vi. For iPr, Ph and CH2Ph substituted isoindolinones, restricted
optimization of only H-atom positions and having the
remaining atoms frozen in the crystal geometry provided the
best results. Such an approach can reduce the computations
from months to one week.
X-ray measurements Good crystals were obtained for all (S)-enantiomers
recrystallized from CH2Cl2 except for the CH2Ph substitution, for which a
good crystal was measured for the (R)-configuration. X-ray diffractions
were measured at 100 K using a Rigaku SuperNova (dual source) four
circle diffractometer along with an Eos CCD detector. The crystals were
mounted within a nylon loop by means of paratone-N cryoprotectant oil.
For data collection, mirror-monochromated Cu Kα radiation (= 1.54184 Å)
from a micro-focused Nova X-ray source was applied and CrysAlis PRO
software was used.[23] A multi-scan absorption correction was applied. The
structures were solved by direct methods and refined with the full-matrix
least squares technique on F2 data using SHELXTL and OLEX2
software.[24] All non-hydrogen atoms were refined anisotropically while H-
atoms were inserted in calculated positions and refined isotropically using
the riding model. The crystal structures were visualized by the MERCURY
program.[25] The structural data for the discussed five crystals: (S)-Me, (S)-
This article is protected by copyright. All rights reserved.

10.1002/cphc.201800244
ChemPhysChem
FULL PAPER
Heterocyclic Systems: Chemistry and Properties, Vol. 18 (Eds. O. A.
Attanasi, R. Noto, D. Spinelli) Italian Society of Chemistry, Rome, 2014,
pp. 113–140; c) D. L. Comins, S. Schilling, Y. Zhang, Org. Lett. 2005, 7,
95–98; d) S. M. Allin, C. J. Northfield, M. I. Page, A. M. Z. Slawin, J.
Chem. Soc., Perkin Trans. 1, 2000, 1715–1721.
iPr, (S)-nBu, (S)-Ph, and (R)-CH2Ph-isoindolinone, were deposited at the
Cambridge Crystallographic Data Center: CCDC-1819989, CCDC-
1819990, CCDC-1819991, CCDC-1819992, and CCDC-1819993,
respectively.
[4]
[5]
R. Kawęcki, W. Stańczyk. A. Jaglińska, Tetrahedron 2017, 74, 578-584.
a) T. Miyazawa, T. Shimanouchi, S. Mizushima, J. Chem. Phys. 1956,
24, 408; b) C. Y. S. Chen, C. A. Swenson, J. Phys. Chem. 1969, 73,
2999-3008; c) S. T. King, Spectr. Acta A: Mol. Spectr. 1972, 28, 165-
175; d) T. M. Watson, J. D. Hirst, J. Phys. Chem. A 2002, 106, 7858-
7867; e) P. Bouř, T. A. Keiderling, J. Chem. Phys. 2003, 119, 11253-
11262.
Calculations The conformational analysis of Me, nBut, iPr, Ph and CH2Ph
substituted isoindolinones was performed using the Conflex program
based on the MMFF94s force field.[26] For the nBut isoindolinone, the
aliphatic chain dihedrals were frozen in a zig-zag form. For Me and Ph
substituted isoindolinones, only one conformer was found, while for the
remaining isoindolinones three conformers were predicted. Next, dimers
and stacking dimer structures were computationally considered in the gas
phase. Two computational strategies were applied to reproduce the
spectra of substances in pellets. The first was based on fully optimizing
different extracted crystal fragments and confirming that the harmonic
frequencies were all real. The second strategy, a much faster H-opt
approach, only optimized the positions of H-atoms; here the calculated
spectra exhibited some imaginary frequencies which were not important
for interpreting the ν(C=O) stretching vibration region.
[6]
a) M. Diem. J. Am. Chem. Soc. 1988, 110, 6967–6970; b) T. B.
Freedman, A. C. Chernovitz, W. M. Zuk, M. G. Paterlini, L. A. Nafie, J.
Am. Chem. Soc. 1988, 110, 6970–6974; c) S. Góbi, G. Magyarfalvi, G.
Tarczay, Chirality 2015, 27, 625-634; d) S. Góbi, G. Magyarfalvi, Phys.
Chem. Chem. Phys. 2011, 13, 16130-16133; e) Z.-W. Qu, H. Zhu, V.
May, J. Comp. Chem. 2011, 32, 1500-1518; f) J. Kubelka, R. Huang, T.
A. Keiderling, J. Phys. Chem. B 2005, 109, 8231–8243; g) J. A. Grahnen,
K. E. Amunson, J. Kubelka, J. Phys. Chem. B 2010, 114, 13011–13020;
h) W. R. W. Welch, T. A. Keiderling, J. Kubelka, J. Phys. Chem. B 2013,
117, 10359−10369; i) M. Hollósi, E. Vass, G. Szilvágyi, A. Jakas, I.
Laczkó, ARKIVOC, 2012, 5, 291-300; j) A. M. Polyanichko, H. Wieser,
Biopolymers 2005, 78, 329–339; k) A. M. Polyanichko, V. V.
Andrushchenko, P. Bouř, H. Wieser in Circular Dichroism: Theory and
Spectroscopy, (Ed. D. S. Rodgers), Nova Science Publishers, Inc., 2012,
pp. 67-126; l) K. Knapp, M. Górecki, J. Frelek, R. Luboradzki, M. Hollósi,
Z. Majer, E. Vass, Chirality 2014, 26, 228-242.
All calculations were performed using the hybrid Becke three-
parameter Lee−Yang−Parr density functional theory (DFT) B3LYP
functional[27] with and without the Grimme’s D3 dispersion correction.[28]
The larger the system size, the smaller the appropriate basis set:
def2TZVP,[29] TZVP,[30] 6-31G(d),[31] and 6-31G,[32] respectively. The IR
and VCD spectra were presented using the Lorentzian functions with a 4
cm-1 peak width at half the maximum height. All calculations were
performed using the Gaussian 09 package of programs.[33]
[7]
[8]
a) A. Massa, P. Rizzo, F. Scorzelli, G. Monaco, R. Zanasi, J. Pharm.
Biomed. Anal. 2017, 144, 52-58; b) A. Massa, P. Rizzo, G. Monaco, R.
Zanasi, Tetrahedron Lett. 2013, 54, 6242-6246.
a) V. P. Nicu, E. Debie, W. Herrebout, B. Van der Veken, P. Bultinck, E.
J. Baerends, Chirality 2009, 21, E287-97; b) J. Kong, L. A. Joyce, J. Liu,
T. M. Jarrell, J. C. Culberson, E. C. Sherer, Chirality 2017, 29, 854–864;
c) R. F. Sprenger, S. S. Thomasi, A. G. Ferreira, Q. B. Cass, J. M. Batista
Junior, Org. Biomol. Chem. 2016, 14, 3369–3375; d) G. Mazzeo, S.
Abbate, G. Longhi, E. Castiglioni, C. Villani, Chirality 2015, 27, 907-913.
C. Merten, F. Li, K. Bravo-Rodriguez, E. Sanchez-Garcia, Y. Xu, W.
Sanderb, Phys. Chem. Chem. Phys. 2014, 16, 5627-5633.
Acknowledgements
This work was supported by the National Science Centre in
Poland Grant No. 2013/09/B/ST5/03664. Świerk Computing
Centre (CIS) is acknowledged for generous allotment of the
computing time. The authors thank Mr. Mateusz Nawara for his
help in English language corrections.
[9]
[10] a) R. Schweitzer-Stenner, T. J. Measey, Spectroscopy 2010, 24, 25–36;
b) L. Bednárová, P. Maloň, P. Bouř, Chirality 2007, 19, 775–786; c) J. H.
Choi, J. S. Kim, M. Cho, J. Chem. Phys. 2005, 122, 174903; d) A. Massa,
P. Rizzo, G. Monaco, R. Zanasi, Tetrahedron Lett. 2013, 54, 6242–6246.
[11] J. E. Rode, J. Cz. Dobrowolski, K. Lyczko, A. Wasiewicz, D. Kaczorek,
R. Kawęcki, G. Zając, M. Baranska, J. Phys. Chem. A 2017, 121, 6713-
6726.
Conflicts of interest
There are no conflicts to declare.
[12] a) L. A. Nafie, Vibrational optical activity: principles and Applications,
John Wiley & Sons, Ltd, 2011; b) G. Pescitelli, T. Kurtán, K. Krohn in
Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in
Stereochemical Analysis of Synthetic Compounds, Natural Products,
and Biomolecules (Eds. N. Berova, P. L. Polavarapu, K. Nakanishi, R. W.
Woody), John Wiley & Sons, 2012, pp. 217-249.
Keywords: VCD •X-ray • DFT • H-bonding • catemers
[13] a) L. A. Nafie in Comprehensive Chiroptical Spectroscopy, Volume 1:
Methodologies, and Theoretical Simulations (Eds. N. Berova, P. L.
Polavarapu, K. Nakanishi, R. W. Woody), John Wiley & Sons, 2012, pp.
115-146; c) E. Castiglioni, P. Biscarini, S. Abbate, Chirality 2009, 21,
E28–E36.
[1]
[2]
J. Caruano, G. G. Muccioli, R. Robiette, Org. Biomol. Chem. 2016, 14,
10134-10156.
a) R. K. Bhatia, Curr. Top Med. Chem. 2017, 17, 189 – 207; b) J. Liu, L.
Wang, N. Guo, Y. Teng, P. Yu, J. Chem. Pharm. Res. 2014, 256-260; c)
Y. Li, D. Liu, S. Cen, P. Proksch, W. Lin, Tetrahedron 2014, 70, 7010-
7015; d) K. Spec, T. Magauer, Beilstein J. Org. Chem. 2013, 9, 2048-
2078; e) A. Di Mola, M. Tiffner, F. Scorzelli, L. Palombi, R. Filosa, P. De
Caprariis, M. Waser, A. Massa, Beilstein J. Org. Chem. 2015, 11, 2591–
2599.
[14] M. M. Quesada-Moreno, J. R. Avilés-Moreno, J. J. López-González, K.
Jacob, L. Vendier, M. Etienne, I. Alkorta, J. Elguero, R. M. Claramunt,
Phys Chem Chem Phys. 2017, 19, 1632-1643.
[15] N. Jiang, R. X. Tan, J. Ma, J. Phys. Chem. B 2011, 115, 2801–2813.
[16] H. McNab, S. Parsons, D. A. Shannon, Acta Crystallogr., Sect. C: Cryst.
Struct. Commun. 1997, 53, 1098-1099.
[3]
a) C. Petronzi, S. Collarile, G. Croce, R. Filosa, P. De Caprariis, A.
Peduto, L. Palombi, V. Intintoli, A. Di Mola, A. Massa, Eur. J. Org. Chem.
2012, 5357–5365; b) A. Di Mola, L. Palombi, A. Massa, in Targets in
[17] A. Pérez-Mellor, A. Zehnacker, Chirality 2017, 29, 89–96.
[18] C. V. Berney, J. Am. Chem. Soc. 1973, 95, 708- 716.
This article is protected by copyright. All rights reserved.

10.1002/cphc.201800244
ChemPhysChem
FULL PAPER
[19] G. Pescitelli, T. Kurtán, U. Flörke, K. Krohn, Chirality 2009, 21, E181–
E201.
[20] a) T. Kurtán, R. Jia, Y. Li, G. Pescitelli, Y.-W. Guo, Eur. J. Org. Chem.
2012, 6722–6728; b) J. Frelek, M. Górecki, M. Łaszcz, A. Suszczyńska,
E. Vass, W. J. Szczepek, Chem. Commun. 2012, 5295–5297; c) N. U.
Rehman, H. Hussain, S. Al-Shidhani, S. K. Avula, G. Abbas, M. U. Anwar,
M. Górecki, G. Pescitelli, A. Al-Harrasi, RSC Adv., 2017, 7, 42357–42362.
[21] T. Taniguchi and K. Monde, J. Am. Chem. Soc. 2012, 134, 3695−3698.
[22] a) S. Abbate, G. Mazzeo, S. Meneghini, G. Longhi, S. E. Boiadjiev and D.
A. Lightner, J. Phys. Chem. A 2015, 119, 4261−4267; b) C. L. Covington,
V. P. Nicu and P. L. Polavarapu, J. Phys. Chem. A 2015, 119, 10589–
10601; c) P. L. Polavarapu, Chirality 2016, 28, 445–452; d) V.P. Nicu,
Phys.Chem.Chem.Phys. 2016, 18, 21202-21212.
[23] CrysAlis PRO. Agilent Technologies Ltd, Yarnton, Oxfordshire, England,
2014.
[24] a) G. M. Sheldrick, Acta Cryst. C 2015, 71, 3-8; b) O. V. Dolomanov, L .J.
Bourhis, R. J. Gildea, J. A. K. Howard, H. Puschmann, J. Appl. Cryst.
2009, 42, 339-341.
[25] C. F. Macrae, I. J. Bruno, J. A. Chisholm, P. R. Edgington, P. McCabe,
E. Pidcock, L. Rodriguez-Monge, R. Taylor, J. van de Streek, P. A. Wood,
J. Appl. Crystallogr. 2008, 41, 466-470.
[26] CONFLEX 7, Conflex Corp., Japan, 2012; b) H. Goto, E. Osawa, J. Am.
Chem. Soc. 1989, 111, 8950–8951; c) H. Goto, E. Osawa, J. Chem. Soc.,
Perkin Trans. 2 1993, 187–198.
[27] a) A. D. Becke, J. Chem. Phys. 1993, 98, 5648−5652; b) K. Burke, J. P.
Perdew,
Y Wang, Electronic Density Functional Theory: Recent
Progress and New Directions (Eds. J. F. Dobson, G. Vignale), M. P. Das,
Plenum, 1998.
[28] S. Grimme, J. Antony, S. Ehrlich, H. A. Krieg, J. Chem. Phys. 2010, 132,
154104.
[29] F. Weigend, R. Ahlrichs, Phys. Chem. Chem. Phys. 2005, 7, 3297-3305.
[30] A. Schaefer, C. Huber, R. Ahlrichs, J. Chem. Phys. 1994, 100, 5829-
5835.
[31] V. A. Rassolov, M. A. Ratner, J. A. Pople, P. C. Redfern, L. A. Curtiss, J.
Comp. Chem. 2001, 22, 976-84.
[32] W. J. Hehre, R. Ditchfield, J. A. Pople, J. Chem. Phys. 1972, 56, 2257.
[33] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb,
J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A.
Petersson, et al., GAUSSIAN 09 (Revision D.01), Gaussian, Inc.,
Wallingford, CT, 2013.
This article is protected by copyright. All rights reserved.

10.1002/cphc.201800244
ChemPhysChem
FULL PAPER
Entry for the Table of Contents (Please choose one layout)
FULL PAPER
J. E. Rode,* K. Lyczko,
M. Jawiczuk, R. Kawęcki,
W. Stańczyk, A. Jaglińska,
J. Cz. Dobrowolski
The
stretching
VCD
ν(C=O)
vibrations
bands in cyclic amides,
namely Me, iPr, nBu, Ph
and CH2Ph 3−substituted
isoindolinones, are un-
expectedly intense and
distinctive when measured
in KBr pellets. We show
how to reproduce these
specific patterns by X-ray
supported DFT calcula-
tions.
Page No. – Page No.
The VCD pattern of the
ν(C=O) bands in
isoindolinones
(S)-Ph
(S)-iPr
(S)-Me
(S)-CH2Ph
(S)-nBu
This article is protected by copyright. All rights reserved.