Benzoylguanidines as Anion-Responsive Systems

Article abstract of DOI:10.1002/cplu.201800247

A series of benzoylguanidinium salts was prepared and the changes in UV/Vis spectra, triggered by the presence of anions, were investigated. All compounds undergo deprotonation with basic anions like dihydrogenphosphate and acetate in acetonitrile. The most pronounced spectral changes were obtained by deprotonation of N¹-benzoyl-N³-(p-nitrophenyl) guanidinium chloride which shows the naked-eye visible color change from colorless to yellow. Measured pK_a(BH⁺) in acetonitrile ranges from 12–16, which is comparable to the pyridinium cations. The proton transfer equilibria were also tested in acetonitrile/water mixture where all but the most acidic derivatives showed pK_a(BH⁺) of 4–6 units which corresponds to apparent association constants of 10⁴–10⁶ dm³ mol^?1. UV/Vis spectra of neutral and protonated forms were modelled by the TD-DFT approach using CAM-B3LYP and PBE0 functionals and compared to CC2 results. In the case of CAM-B3LYP, a parameter ω, defining amount of long-range exchange correction, was varied to achieve the best agreement with the experimental spectra. The optimized ω parameters are 0.10 a₀^?1 for neutral benzoylguanidines and 0.20 a₀^?1 for neutral nitrobenzoyl and protonated systems. The larger ω parameter in the latter is ascribed to more pronounced charge transfer character of the HOMO-LUMO transition – the one responsible for the lowest energy absorption band.

Full text of DOI:10.1002/cplu.201800247

Accepted Article
Title: Benzoylguanidines as the Anion Responsive Systems
Authors: Zoran Glasovac, Luka Barešić, Ivana Antol, and Davor
Margetić
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: ChemPlusChem 10.1002/cplu.201800247
Link to VoR: http://dx.doi.org/10.1002/cplu.201800247
A Journal of
www.chempluschem.org

10.1002/cplu.201800247
ChemPlusChem
FULL PAPER
Benzoylguanidines as the Anion Responsive Systems
Zoran Glasovac,*[a] Luka Barešić,[a] Ivana Antol[a] and Davor Margetić[a]
Abstract: A series of benzoylguanidinium salts was prepared and
the changes in UV/Vis spectra, triggered by the presence of anions,
were investigated. All compounds undergo deprotonation with basic
anions like dihydrogenphosphate and acetate in acetonitrile. The
most pronounced spectral changes were obtained by deprotonation
of N1-benzoyl-N3-(p-nitrophenyl) guanidinium chloride which shows
the naked-eye visible color change from colorless to yellow.
Measured pKa(BH+) in acetonitrile ranges from 12-16 being
comparable to the pyridinium cations. The proton transfer equilibria
were also tested in acetonitrile/water mixture where all but the most
acidic derivatives showed pKa(BH+) of 4 – 6 units what corresponds
to the apparent association constants of 104-106 dm3mol-1. UV/Vis
spectra of neutral and protonated forms were modelled by TD-DFT
approach using CAM-B3LYP and PBE0 functionals and compared to
CC2 results. In the case of CAM-B3LYP, a parameter  defining
amount of long-range exchange correction, was varied to achieve
the best agreement with the experimental spectra. The optimized 
anion sensors[4-6,15] as well as active organocatalysts.[16-18]
Protonated guanidine subunit will form charge-assisted
hydrogen bonds with anions which are significantly stronger with
respect to the ordinary neutral-neutral or neutral-anion
interactions.[19]
Benzoylguanidines, as typical representatives of acylguanidines,
are expected to act as anion receptors better than simple alkyl
and aryl guanidines due to improved hydrogen bond donating
properties. Indeed, their anion binding affinity is known for a long
time.[20] It has also been proven that benzoylguanidine-type
drugs inhibit urokinase and NHE with multiple charge assisted
hydrogen bonds being
a
dominant interaction motif.[21-23]
Schmuck[24] emphasized the following advantages of
acylguanidines over non-acylated ones in context of anion
sensing: (i) increased acidity with pKa of 7-9, (ii) conversion of
guanidine N-H bond to amidic one, (iii) increased planarity and
rigidity of the receptor and (iv) more possibilities for the desired
additional structural modifications targeting increased selectivity
of the receptor. First two points (quantified through pKa and
hydrogen bond donating ability) are only loosely mutually
correlated,[25] however, higher anion binding affinities of
thioureas with respect to ureas have been often ascribed to their
increased acidity.[26]
-1
parameters are 0.10 a0-1 for neutral benzoylguanidines and 0.20 a0
for neutral nitrobenzoyl and protonated systems. The larger 
parameter in the latter is ascribed to more pronounced charge
transfer character of HOMO-LUMO transition - the one responsible
for the lowest energy absorption band.
Increasing acidity of the receptor is therefore desired physico-
chemical property in context of anion sensing and binding.
However, sufficiently basic anions, like acetate and fluoride
could cause deprotonation resulting in acid-base equilibrium
forming two separated entities.[27-32] While deprotonation is not
desirable in context of supramolecular chemistry and
geometry-based selectivity,[27,28,31] it may be useful in indicators
when a large UV bathochromic shift is preferred over the actual
binding. Deprotonation generally leads to larger shift in
absorption maximum with respect to the anion binding.[31,33]
Drawback related to the incorporation of the benzoyl (acyl)
moiety is the formation of intramolecular hydrogen bonds (IHB)
involving protons usually aimed to bind anions which reduces
usefulness of the receptor. For example, Jiang and coworkers
observed no acetate binding of N1-[4-(dimethylaminobenzoyl-)]-
N3-phenyl thiourea in acetonitrile while the corresponding
diarylthiourea binds acetate with the constant of 104 dm3mol-1
order of magnitude.[34,35] Even analogous N3- unsubstituted
benzoylthiourea showed a weak acetate binding effect, in spite
to the availability of two protons for the interaction with anion.[35]
In comparison with corresponding benzoylthioureas, N1,N3-
disubstituted benzoylguanidines are expected to work better as
sensors due to larger number of protons available for hydrogen
bonding especially if protonated. Like benzoylthioureas,
aroylguanidines possess co-planar substructure[36] due to
intramolecular hydrogen bonding and larger orbital overlap
between aroyl chromophore and guanidine moiety.
Introduction
Anions are omnipresent in the environment and in living
organisms. Among many other sources, they could originate
from well-known pollutants like the herbicide – glyphosate[1] or
chemical warfare agents[2] being indicators of their usage. To
control their presence and potentially negative impact on human
health and environment, efficient sensors are needed. Therefore,
development of anion sensors has become one of the most
active areas of supramolecular chemistry.[3]
Guanidine moiety has been used in a vast number of the
artificial anion sensors and receptors as the anion recognition
site.[4-6] Due to its Y-shaped structure[7-9] it possesses high
intrinsic and solution basicity[10,11] what ensures its protonation
over wide range of pH. It can form strong multiple hydrogen
bonds with anions or polar groups either inter- or
intramolecularly[12-14] what makes it suitable active subunit in
[a]
Z. Glasovac*, L. Barešić, I. Antol, D. Margetić
Division of Organic Chemistry and Biochemistry
Ruđer Bošković Institute
Bijenička cesta 54, 10000 Zagreb, Croatia
E-mail: glasovac@irb.hr; ORCID: 0000-0002-3441-6595
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/cplu.201800247
ChemPlusChem
FULL PAPER
Our previous results showed significant change in aryl-guanidine Results and Discussion
overlapping induced by protonation while the arylguanidinium
cations show the spectroscopic properties similar to the isolated
chromophores.[33] In this sense, N1-aryl-N3-aroylguanidines are
suitable for construction of protonation sensitive system in which
the electronic properties of two subunits of one chromophore
can be tuned separately. This is particularly valuable for the
design of charge transfer (CT) systems.
To achieve desired goals, we conducted an anion exchange
titrations in which several equilibria are possible: dissociation of
the starting salt (eq. 1), formation of the ion pair (eq. 2), proton
exchange within the ion pair (eq. 3) and finally dissociation of the
neutral–neutral pair to the solvated neutral entities (eq. 4).
Under submillimolar concentration typically used for UV/Vis
measurements, the first step could take place before
measurements during the solution preparation. To test this
hypothesis we prepared 4xTfOH salt by in situ protonation of 4
with trifluormethanesulfonic acid (TfOH). The most intense
absorption bands in the spectra of 4xTfOH and independently
prepared 4xHCl are located at 260 and 263 nm, respectively.
This anion induced shift of 3 nm is verified by titration of 4xTfOH
salt with tetrabutylammonium chloride in acetonitrile (Figure S6
in Supporting Information) confirming the presence of chloride
anion in the close proximity of benzoylguanidinium cation.
For successful design of new anion sensors,
a good
computational model is needed which is able to simulate
electronic spectra of both neutral and protonated structures with
similar accuracy. Recently we have investigated changes in UV
spectra of arylguanidines and their salts employing modern TD-
DFT approaches.[33] Comparison with CASPT2 and experimental
data revealed two well-established DFT functionals (PBE0[37,38]
and CAM-B3LYP[39]) as the best choice from the viewpoint of
computational cost, efficiency and accuracy. In continuation of
this research and our long term interest in acid/base and
hydrogen bonding properties of guanidine derivatives,[40] we
focused here on substituted N1-phenyl-N3-benzoylguanidine
systems. This simple model system is suitable for investigating
electronic interactions between three fragments (phenyl,
guanidinyl and benzoyl) that affect the spectroscopic properties
and their changes upon protonation. Such structure will also
allow independent change of the substituents for the fine tuning
of the acid/base and electronic properties of the receptor.
k1
0
+ Cl-
BH+
BH+.....Cl-
(1)
k-1
k2
0
0
BH+......A-
B..........HA
+ A-
BH+
(2)
(3)
(4)
k-2
k3
0
0
BH+......A-
B..........HA
k-3
k4
The aim of this work is to investigate
a
series of
benzoylguanidines (Scheme 1) in a form of hydrochloride or
trifluormethanesulfonate salts as potential anion detection
subunits in acetonitrile, water and their 1:1 mixture. From the
pool of anions we have selected representatives covering highly
and weakly basic anions. Acetate (model for carboxylate) and
dihydrogenphosphate were of our primary interest as they are
very common in living organisms and also present in glyphosate
and formed by its hydrolysis. The second target of this work is to
test the modern TD-DFT approaches for simulation of their
UV/Vis spectra. This would allow future computational design of
the efficient anion sensor based on benzoylguanidine subunit
showing a naked-eye visible color changes. Computed TD-DFT
UV spectra are benchmarked against CC2 method and the
experimental data.
B
+ HA
k-4
The second process is the binding and it is usually target of the
measurements. The last two stages (3 and 4) are not always
present and they occur only if the difference in acidity between
BH+ and HA is sufficiently large[25] (pKa > 1). If present, these
two processes occur as the one step because of low stability of
neutral-neutral complexes in polar media. They are hardly
separable from the binding and very difficult to measure
individually. The processes (2)-(4) represent a typical acid-base
proton exchange and could be used to estimate pKa of one of
the partners. In our experiments with highly and moderately
basic anions, formation of the ionic pairs could be excluded due
to at least two reasons: (i) there is no indication of isosbestic
points other than those due to deprotonation and (ii) calculated
Gibbs energy of complex formation (Gcplx([1H+Ac-]) from ionic
species (eq. 5) is ca. -31 kJ mol-1 and from neutrals (eq. 6) is
positive and more than 19 kJ mol-1 (Table 1) above the summed
energies of neutrals indicating that proton transfer followed by
dissociation of neutrals is thermodynamically favorable in
acetonitrile.

G°cplx([1H+Ac-]) = G°([1H+Ac-]) – [G°(1H+) + G°(Ac-)]
(5)
(6)

G°cplx([1H+Ac-]) = G°([1H+Ac-]) – [G°(1) + G°(HAc)]
Similar
thermodynamics
was
obtained
for
other
benzoylguanidines as well. This is not surprising having in mind
the expected large difference in pKa between acetate,(Ac),
dihydrogenphosphate (DHP) and fluoride (F) anions and target
Scheme 1. Structures of the investigated benzoylguanidines.
This article is protected by copyright. All rights reserved.

10.1002/cplu.201800247
ChemPlusChem
FULL PAPER
benzoylguanidines of more than 2 pKa units. This result is in line
with our previous work[33] where ionic complexes (salt bridges) of
arylguanidines were observed only for phenylguanidine/acetate
and naphthylguanidine/formate; i.e. with the anions having the
basicity within 1-2 pKa units from the target guanidine derivative.
Regardless, we tried to fit our titration experiments both to the
formation of complex according to the eq. (2) and to the proton
exchange equilibrium (pKa, eqs. 2-4). Namely, Perez-Casas
and Yatsimirsky have shown that in some cases the same data
could be fitted to both "apparent association constant" and to the
overall constant involving multiple equilibria including proton
transfer with sufficient quality.[27] These, and also other
authors[41] pointed out the importance of self-association of
anions with their conjugate acids in non-aqueous media and this
equilibrium is also included in Kapp. The Kapp will serve for the
comparison with literature data and those obtained by titration of
selected benzoylthioureas (BTU1 and BTU4-6) and neutral
benzoylguanidines (4 and 6) prepared in this work.
Table 1. Calculated pKa and Gcplx as well as the experimental pKa, log Kapp and log Kassoc determined from the titrations of investigated
benzoylguanidines and benzoylthioureas in different media.
[b]
Mol.
BH+
1H+
pKa(BH+)[a]
(calc)
Gcplx
pKa(BH+) or log Kapp(AcO-)
Mol.
B
log Kassoc (ACN)
/kJ mol-1
-31.2 (19.1)
ACN (A=DHP)[c]
50-50 (A=Ac)[c,d]
14.6
14.2 ± 0.02
14.2 ± 0.01
6.1 ± 0.02
(4.6 ± 0.01)
4
n/b [d]
2H+
3H+
4H+
5H+
6H+
15.5
-28.2 (19.5)
-31.9 (21.4)
-33.0 (24.2)
-38.0 (23.8)
-30.6 (23.1)
15.7 ± 0.02
6.1 ± 0.02
(4.6 ± 0.01)
5
5.0 ± 0.01 (Ac)
n/b [d]
14.3
13.5 ± 0.01
13.2 ± 0.01
5.1 ± 0.02
(5.2 ± 0.06)
BTU1
BTU4
BTU5
BTU6
13.2
12.5 ± 0.01
(12.4 ± 0.01)[e]
4.5 ± 0.02
(6.0 ± 0.03)
4.9 ± 0.01 (Ac)
3.7 ± 0.01 (DHP)
13.6 (12.2)[f]
14.0
12.6 ± 0.01
(12.4 ± 0.01)[e]
mostly
deprotonated
4.7 ± 0.01 (Ac)
n/b [d]
13.4 ± 0.01
(13.2 ± 0.01) [e]
4.5 ± 0.02
(5.6 ± 0.02)
[a] pKa values of the conjugate acid (BH+) were calculated using the approach analogous to the one described in ref.[42] except that CPCM was used as
the solvation model instead of IPCM. Details are given in Supporting Information material. [b] Gcplx(BHA) is calculated according to eqs. (5) and (6); the
latter are given in parentheses. Gibbs energies were calculated at SMD(ACN)/B3LYP/6-311+G(d,p)//B3LYP/6-31+G(d,p) level of theory. [c]
experimental data from the titrations experiments; values in parentheses relate to the pKapp,50-50. pKa(B) was determined using the literature value of
pKa,ACN (H3PO4) = 17.5[44] and pKa,50-50(HAc) = 6.1[43]; The errors indicate quality of fitting and not the precision of the measured constants. The actual
error on pKa is ± 0.2 pKa units as determined from two independent titrations either form the hydrochlorides (1H+ and 3H+) or from different salts (4H+ -
6H+). [d] n/b – no binding. [e] titration of in situ generated trifluormethanesulfonic salt with solution of DHP anion. [f] measured by competitive titration of
5 against 4 with TfOH; pKa(4) = 12.5 was used as the reference value.
In acetonitrile, three most basic anions (Ac, F and DHP)
deprotonate benzoylguanidinium cations what is in accordance
with their high basicity. The best results were obtained in
titrations with DHP in which case we were able to obtain
satisfactory fit to pKa and to determine pKa of each target
benzoylguanidinium cation. Titration of the 5HCl with DHP anion
is shown in Figure 1 as an example. Obtained results are in very
good agreement with the calculated data (Table 1). Since the
DHP/H3PO4 self-association constant in acetonitrile (log K = 8)[44]
is quite large, it was necessary to include it into overall
equilibrium. Attempts to fit the acetate titration data to pKa
resulted in good overall fit but with unrealistically high pKa values
(above 19 pKa units). No improvement was obtained if self-
association of the acetate/acetic acid was taken into account.
The reason for this behavior was ascribed to the difference in
pKa of the titration partners larger than 5 pKa units. Fitting the
DHP or Ac titration data to Kapp resulted in significant deviations
from the experimental titration curve. In a view of these findings,
fitting the data for titration with fluoride was not attempted. As
expected, the last two anions (N and P) induced no observable
spectral change and they were not used for the titrations in
water and 50-50 mixture.
Titrations in acetonitrile were also conducted for the selected
benzoylthioureas (BTU1 and BTU4-BTU6) and neutral
benzoylguanidines 4 and 5 prepared in this work (Table 1). For
the derivatives BTU4, BTU5 and guanidine 5 obtained log Kassoc
(in ACN) are similar and close to 5. No binding was obtained for
other derivatives. It is interesting to note that nitro- substituted
benzoylthioureas appear to be acidified enough to become
acceptable anion binding system. On the other hand, modifying
BTU4 to BTU5 by introducing dimethylamino group retards the
binding ability of benzoylthiourea group leading to no response
This article is protected by copyright. All rights reserved.

10.1002/cplu.201800247
ChemPlusChem
FULL PAPER
of BTU6 to the presence of acetate. This effect could be
ascribed to the lowering acidity similar to that calculated for
benzoylguanidines 4 and 6 (Table 1). In the case of guanidines,
position of nitro group plays important role giving reasonable
response to acetate anion only for p-nitrophenyl derivative 5.
CAM density functionals assuming acetonitrile as a solvent
amounts to 79 %, 79 % and 81 % respectively. The very good
agreement between experiment and theory is achieved with all
density functionals in spite of the difference in solvent. The
isomer ratios were calculated for all investigated guanidines and
their relative contributions are given in Table S1. The Ba isomer
is preferred for derivative 5, its conjugate acid 5H+ and for 6H+
while in all other cases, Bb structure is the dominant one. For
each isomer, UV spectra were calculated with long-range
corrected CAM-B3LYP and hybrid PBE0 functionals. For the
neutral derivatives 1-4, CC2 method was also employed. We
tested influence of basis sets using 6-31G(d) and 6-311+G(d,p),
termed hereafter as BS1 and BS2, respectively. Calculated UV
spectra for each pair of isomers (Ba and Bb) were averaged
according to their contribution in the mixture.
The experimental spectrum of the neutral base 1 shows two
clearly visible bands at 273 and 235 nm (I and II, Figure 2) with
the third one appearing as a shoulder at approximately 286 nm
(Ib, Figure 2). CC2 approach predicts both absorption bands at
lower wavelengths with the band II being almost invisible. The
PBE predict the spectra in a very good agreement to the
experimental one especially in combination with smaller basis
set (BS1) where the only band I is red-shifted by 6 nm. Using the
larger basis set (BS2) shifts the complete spectrum for another 7
nm toward longer wavelengths keeping the same intensity ratio
for the bands I and II. The distances between two observable
absorption maxima (max) of 43 nm for both PBE and CC2
Figure 1. Titration of 5xHCl with TBA DHP in acetonitrile. Additional aliquots of
an anion up to 2 eq. did not changed spectra.
We were able to determine both log Kapp and pKa(BH+) in 50-50
ACN-W mixture with sufficient reliability for all target derivatives.
No formation of [Ac-H-Ac]- complex was assumed in this solvent
system. The measured pKa,50-50 values range from 4.5 to 6.1
what is comparable to methylpyridinium and 2-aminopyridinium
cations[43] and they follow the same trend as observed in
acetonitrile. The Kapp values show the expected inverse trend
with respect to pKa meaning that the benzoylguanidinium cations
with lower pKa give higher log Kapp. Obtained log Kapp spans the
values from 4.2 to 6, almost approaching the data for the
benzamido-benzoylthioureas obtained previously.[45] It is also of
the same order of magnitude as the binding constants measured
for benzoylthioureas in acetonitrile (see above).
Titration of benzoylguanidinium salts 1xHCl and 4xHCl with
TBAAc was also carried out in water. Weak, but observable
spectral changes were obtained. For the less acidic derivative
1xHCl (Table 1), adding 5.8 equivalents of acetate anion led to
the roughly estimated 38 % of 1. Analogous titration with more
acidic compound 4xHCl resulted with almost complete
deprotonation after addition of 6 equivalents of acetate anion. In
neither case, we were able to determine Kapp or pKa reliably.
Crude estimates of Kapp are around 101 and 103 dm3mol-1 for 1H+
and 4H+, respectively.
agree satisfactorily with the 38 nm of the experimental max
.
CAM/BS2 approach predicts only one strong band located at
263 nm what is very close to the CC2 approach. Similarly to
PBE, changing from CAM/BS2 to CAM/BS1 shifts the maximum
of the absorption band I toward longer wavelengths by 7 nm
while the band II is still missing. To achieve better results with
CAM approach, we decided to optimize the extent of long-range
correction by variation of the  parameter. Namely, for the
calculations of excitation energies in a donor-acceptor system,
the optimized long-range corrected functionals outperformed the
default ones.[48] In our previous work, we also obtained better
agreement of the calculated and experimental UV spectra in a
series of arylguanidines using optimized CAM-B3LYP approach
(= 0.20 a0−1 instead of default value 0.33 a0−1).[33] Compound 1
was taken as the model system for optimization of  parameter
and its value is varied from 0.10 to 0.20 in steps of 0.01 a0−1. UV
spectrum modeled using CAM(0.10)/BS1 approach resulted in
agreement with the experimental one in a sense that it correctly
predicts two maxima located at 282 and 235 nm (Figure 2a),
although with slightly overestimated mutual spacing of 47 nm.
Additionally, relative intensity of the second band (at 235 nm) is
underestimated with respect to the first one. Further decrease in
 value gives better intensity ratio and smaller spacing between
two bands. However, it also shifts both bands to the longer
wavelengths leading to the larger deviation from the
experimental data.
Structural dependence of UV spectra
Synthesized benzoylguanidines (B) exist as a mixture of two
isomers: "Ba" and "Bb" (B = 1-7, Scheme 1) and, in the case of
the neutral benzoylguanidine 4, their ratio was determined by
1
low–temperature H NMR spectroscopy in CD2Cl2: DMSO-d6 =
10:1, analogously to Gschwind et al. [46,47] By integration of the
baseline separated NH signals for both isomeric conformations,
which are located at 9.2 (4a) and 8.9 (4b) ppm, we obtained the
ratio of 4a : 4b = 23 % : 77 % (Figure S1a in Supp. Info.). The
contribution of 4b structure calculated by the B3LYP, PBE and
This article is protected by copyright. All rights reserved.

10.1002/cplu.201800247
ChemPlusChem
FULL PAPER
Table 2. Calculated contributions of the isomers in the mixture and comparison of the calculated and experimental UV spectra.[a]
Comput.
method
Mol.
B
B
B
BH+
Mol.
B
B
B
BH+
max/nm
max(I)/nm
max(II)/nm
max/nm
max(I)/nm
max(II)/nm
Exp.
1/1H+
2/2H+
3/3H+
273
279
263
282
275
265
234
235
-
235
-
245
246
239
251
248
n/c[d]
4/4H+
5/5H+
6/6H+
276
246 (sh)[c]
-
234
-
243
232
266
265
268
273
264
n/c [d]
PBE/BS1
CAM/BS2
CAM(0.10)/BS1
CAM(0.20)/BS1
CC2
270 (~358 sh)[c]
290
273 (~350 sh)[c]
293
270
222
Exp.
268
ca 278 (sh)[c]
257
276
220
234
240
-
239
262
220
240
241
237
245
238
n/c[d]
332
335
321
340
319
268; 234
244 (br)
253
261 (br)
252
246
293
284
297
281
PBE/BS1
CAM/BS2
CAM(0.10)/BS1
CAM(0.20)/BS1
CC2
258
Exp.
296
293
279
297
283
280
256
-
-
-
248
265
252
269
256
n/c[d]
270-280
-
280
-
225
247
-
265
260
259
256
PBE/BS1
CAM/BS2
CAM(0.10)/BS1
CAM(0.20)/BS1
CC2
247
-
281 (278)
243 (~240 sh)[c]
256 (259)
226
[a] B3LYP = SMD(ACN)/B3LYP/6-311+G(d,p)//OPT, CAM = SMD(ACN)/CAM-B3LYP/BS(1 or 2)//OPT; PBE = SMD(ACN)/PBE0 /BS(1 or 2)//OPT; CC2 = RI-
CC2/def2-TZVP//OPT where OPT = SMD(ACN)/B3LYP/6-31+G(d,p); CAM(0.10) and CAM(0.20) use non-default value for  parameter. [b] isomer ratio was
calculated by B3LYP approach. Similar values were obtained by CAM and PBE functionals. [c] sh – shoulder with no defined maximum. [d] n/c = not calculated.
Figure 2. Comparison of the experimental UV spectra of 1 and 1H+ with those simulated using PBE and CAM approaches (See footnote in Table 2). The
calculated excitation energies for both isomers were shown as the color-coded lines, cumulatively for both isomers.
UV spectrum of 1H+ is characterized by absorption band with the
maximum located at max = 245 nm (Figure 2b) which tails to ca
310 nm. Again, both DFT methods in combination with BS2
predict this absorption maximum blue- and red- shifted with
respect to the experimental value in the same way as for the
neutral compound 1 with |max(exp.-calc.)| being 6-7 nm. Again,
switching to BS1 reduces |max(exp.-calc.)| to 1 nm for PBE but
shifts the CAM result from 239 to 233 nm. Reducing the  value
leads to red shift of the absorption band as in the case of neutral
derivative what compensate the effect of the basis set change.
Thus, the satisfactory agreement between the experiment and
CAM predicted spectra is achieved with  = 0.20 and BS1 with
|max(exp.-calc.)| = 2 nm. The CAM(0.10)/BS1 approach does
not impair the results significantly since it leads to additional red
shift of the band by 4 nm giving overall deviation from the
experiment of 6 nm; but this deviation is still smaller than for the
neutral derivative 1.
Toward the visible region of the UV/Vis spectra.
As the next step, we have tested PBE and CAM models for
simulation of other benzoylguanidines. Successful model would
predict changes in the UV spectra due to protonation and/or
This article is protected by copyright. All rights reserved.

10.1002/cplu.201800247
ChemPlusChem
FULL PAPER
substitution effect consistently. Also, it would help us to
rationalize apparently unusual observed substitution effect on
the UV spectra. Namely, introduction of the methoxy group at
para position of the N1-phenyl moiety in 1 induces small
hypsochromic shift ( = 4 nm, Table 2) of the first band. On the
other hand, a large bathochromic shift of 26 nm and 60 nm was
obtained by para substitution at the same moiety with either
methylcarboxy (3) or nitro groups (5). Inspection of orbitals
(Figure S10 in Supp. Info.) indicates that substitution at N1-
phenyl moiety will affect the HOMO more strongly than LUMO. It
is general knowledge that electron withdrawing groups (EWG)
lower the orbital energies what, in our case, should lead to the
increased HOMO-LUMO gap and, consequently, to the
bathochromic shift – exactly opposite to what we observed.
predicting the band above 350 nm (Ph → *). The default
CAM/BS2 approach predicts band I red-shifted by > 10 nm with
overestimated distance between band I and II ( = 56 nm).
Change in basis sets from BS2 to BS1 improves the position of
the first band while the (II-I) is only marginally shortened (max
= 278 nm,  = 52 nm). Better agreement with the experiment
was obtained by optimization  value by its variation from 0.10
to 0.30 in steps of 0.01 |a0-1|. The value of  = 0.25 was taken as
the best estimate for the distance between bands I and II with
the first peak located at 287 nm (11 nm red-shifted from the
experiment) and with the second peak positioned at 235 nm (11
nm blue-shifted), which is similar to the CC2 value. Reducing the
 to 0.20 |a0-1| improves the peak tailing on account of the first
band shift to 293 nm. It also improves the position of the second
band, although its intensity is highly overestimated.
Figure 3. Lowest unoccupied orbitals for 1,
3 and 5 calculated at
CAM(0.10)/BS1 level of theory.
Detailed inspection of orbitals reveals systematic increase of
orbital coefficients at the phenyl moiety in LUMO on going from
1 to 5 (Figures 3 and 4) due to electron density shift induced by
EWG. As a consequence LUMO(5) becomes very similar to
LUMO+2(1) and HOMO-LUMO transition becomes the locally
excited state localized mostly within the N-phenyl moiety. The
HOMO is affected to the lesser extent (Figure 4) due to
increased antibonding interaction with guanidine fragment.
These two effects lead to the observed EWG induced
bathochromic shifts. Comparison of the experimental and
calculated spectra (Figure 5) proves the consistency of the
employed computational methods. Positions of the absorption
bands in the UV/VIS spectra of the derivatives 1, 2, 3 and 5 are
similarly well described by PBE/BS1 and CAM(0.10)/BS1 and
significantly better than with BS2. Both methods show
systematic deviation from the experiment by 3-6 nm in the case
of PBE/BS1 and 3-10 nm for CAM(0.10)/BS1. Intensities of the
band II in 3 and 5 (Figures 5c and 5e) were not predicted
correctly by the employed DFT methods (with BS1 or BS2).
However, all of them indicate presence of a number of states
with small oscillator strengths in the appropriate regions.
Conversion of benzoyl to nitrobenzoyl derivative (1 → 4) leads to
broadening of the first absorption band with the maximum
shifted by 4 nm toward longer wavelengths (Table 2) but tailing
to approximately 380 nm. CC2 approach simulates the
experimental spectrum acceptably well. The PBE/(BS1 or BS2)
approach completely fails to give the acceptable spectra
Figure 4. HOMO and LUMO energies in benzoylguanidines 1-6. Results for 1,
3 and 5 clearly indicate similar HOMO energies and strong stabilization of
LUMO. In the case of 6, the state with largest oscillator strength corresponds
to the excitation from HOMO-1. The HOMO-LUMO transition has pronounced
charge transfer character and, consequently, low oscillator strength. This
transition (Eexc = 3.47 eV, exc. = 358 nm, f = 0.08; data for the isomer 6a)
could be associated with pronounced peak tailing above 350 nm..
The system 6 shows again the broad first absorption band with
maximum positioned at 274 nm – practically the same as in the
parent compound 1. However, the maximum is wide and tails to
>400 nm. Similarly to 4, this system also proved to be the
problematic case for PBE method which predicts existence of
low-energy states located above 500 nm (Ph → *). However,
no absorption band was experimentally observed above 400 nm.
The energy of the same transition is 332 nm (6a) and 308 nm
(6b) as calculated by the CAM approach. Again, the optimized
method includes  value of 0.20 in combination with BS1.
This article is protected by copyright. All rights reserved.

10.1002/cplu.201800247
ChemPlusChem
FULL PAPER
Figure 5. Comparison of the calculated and experimental UV/Vis spectra for the neutral benzoylguanidines 1-6. The color-coded bars indicate the calculated
excitation energies and their oscillator strengths.
Besides quite good guess for the position of the absorption band
maximum, it also reproduces a peak tailing quite well. These
presence of additional overlapped bands. Both 4H+ and 6H+
have p-nitrobenzoyl group and the described spectral features
are most likely related to the influence of nitro group. Indeed, for
the 4H+ and 6H+, both DFT approaches predict the dark states
around 330 nm with dominant configuration corresponding to n-
(NO2) → *(LUMO) transition what is most likely responsible for
the weak absorption in this region.
All approaches successfully predict significant hypsochromic
shift on going from neutral to the protonated state. In general,
the most intense band is correctly described, but neither of the
methods predicts the experimentally observable shoulder for any
of benzoylguanidinum cations (Figure 6 and Figure S11a,b in
results indicate that the default
 parameter should be
decreased but to the lesser extent in comparison to 1-3 and 5.
The protonation effect on the spectra of neutral
benzoylguanidines is similar to the one described for 1 (see the
previous paragraph). Except for the 4H+ and 6H+, the spectra of
protonated derivatives contain one clearly visible band followed
by more or less pronounced shoulder at ca 275-280 nm and
tailing up to 340 nm (Figure S11b in the Supp. Info.). The
spectra of 4H+ and 6H+ show the maximum at 265-266 nm with
non-negligible absorbance up to 380 nm what indicates the
This article is protected by copyright. All rights reserved.

10.1002/cplu.201800247
ChemPlusChem
FULL PAPER
Supp. Info.). Thus, PBE/BS1 approach performs surprisingly
well for the derivatives 1H+, 2H+, and 4H+, with the deviation of
ca 1 nm with respect to experiment. Both CAM(0.10)/BS1 and
CAM(0.20)/BS1 produce correct spectra with the latter giving
slightly smaller deviation from experiment ((exp.-calc.) = -2 - -
3 nm for 1H+, 2H+, and 4H+ and 9 nm for 6H+). The more
problematic cases were 3H+ and 5H+ for which PBE method
predict the maximum at significantly longer wavelengths (Figure
6). Again, CAM(0.20)/BS1 improves the simulated spectra
(Table 2) but the deviation from the experiment for the 5H+is still
too large. Switching back to CAM/BS2 with default  parameter
provides the results with no significant improvement over
CAM(0.20)/BS1. In the spectra of 5H+ all employed methods
indicate presence of two bands corresponding to the
experimental ones but with opposite trend in intensities.
Inspection of the states for the 5H+ indicates presence of two
bright states in both conformers responsible for the mentioned
bands. The first one belongs to the Ph → * (HOMO-LUMO)
excitation while the second one encompasses at least two states
both of which could be best described as the GU → *Bz
excitations. Apparently, neither of the employed calculations
predicts their oscillator strengths correctly. Increase in  value to
0.5 does improve the relative ratio of the intensities of these two
bands from 3.7 to 1.7 in favor of the first band but not sufficiently
to correctly reproduce experimental spectrum.
Figure 6. Comparison of the experimental and simulated UV/Vis spectra of 3H+ and 5H+ with the employed computational models. CAM and CAM(0.10) models
use default and reduced  parameter, respectively.
chloride and trifluoromethanesulfonate salts confirmed that the
influence of counterion is no larger than the experimental error.
All investigated compounds show large changes of UV/Vis
spectra upon deprotonation. The largest bathochromic shift from
Conclusions
UV to visible region of spectra was observed for
benzoylguanidinium derivative 5H+. Therefore, it could be
suitable for the naked eye detection of the basic anions like
In this work we have prepared a series of benzoylguanidinium
salts and investigated their spectral features in the presence of
anions. All benzoylguanidinium salts undergo deprotonation by
acetate and dihydrogenphosphate in acetonitrile and in 50-50
H2O-ACN mixture. Their selectivity toward anions is related to
the difference in pKas between conjugated acids of the used
anion and benzoylguanidine. From the titration data, pKa for all
derivatives were determined. Also, in the case of H2O-ACN
mixture, apparent association constants between 104 and 106
dm3 mol-1 were measured. Comparison of pKas obtained using
dihydrogenphosphate
and
acetate.
However,
water
deprotonates 5H+ limiting its potential use to non-aqueous
medium. Obtained data also indicate pKa,ACN(BH+) of ca 13.0 as
the lower bound ensuring high sensitivity of benzoylguanidine
toward basic anions as well as their applicability in the presence
of water. Neutral derivative 5 and benzoylthioureas BTU4 and
BTU5 also show the acetate sensing capabilities in acetonitrile
This article is protected by copyright. All rights reserved.

10.1002/cplu.201800247
ChemPlusChem
FULL PAPER
FG2
with the binding constant logKassoc = 4.7-5.0. No binding was
observed for other neutral benzoylguanidines.
O
S
O
1) NH4SCN / dry acetone,
reflux, 30 min.
N
N
Cl
H
H
The spectra were modeled by TD-DFT and CC2 methods. The
results indicate CAM/BS1, with the long-range correction ()
FG2
FG1
FG1
2)
-1
H2N
parameter of 0.10 a0 for neutral benzoylguanidines and 0.20
-1
reflux, 30 min.
a0 for neutral nitrobenzoyl and protonated derivatives is the
Method A or B
best cost/accuracy computational option.
FG1 = H or NO2
Our results clearly show suitability of benzoylguanidinium motif
as the pKa dependent anion responsive subunit. We envision its
usage within the larger and substrate selective sensor having at
least one chromophore absorbing in the visible region of spectra.
The selection of suitable chromophores for this purpose could
be preferably done by the TD CAM-B3YLP approach due to its
applicability for large systems.
FG2 = H, CH3O, COOCH3 or NO2
FG2
O
NH2
Methods:
N
N
H
A) HDMS, EDCI, dry acetonitrile, r.t., 2 h
B) HgCl2, Et3N, NH3(aq), DMF, r.t., 72 h
FG1
Scheme 2. Synthetic procedures employed for the preparation of target
benzoylguanidines..
Computational details.
Experimental Section
All structures were optimized in two conformations resembling the
solution structures described by Gschwind and coworkers (Scheme
1).[46,47] Weight of each conformer was determined from the relative
Gibbs energies of the structures optimized in the acetonitrile assuming
solvent as the polarized continuum. The optimizations were conducted at
SMD/B3LYP/6-31+G(d,p) level of theory[61,62] and the minima were
confirmed by the vibrational analysis. UV spectra were also calculated in
acetonitrile by using either PBE0[37] or CAM-B3LYP[39] in conjunction with
either 6-31G(d) (BS1) or 6-311+G(d,p) (BS2) basis set. Simulation of
spectra from the calculated excitation energies and oscillator strengths
was conducted by in-house created Perl script written by dr. Mario
Barbatti during our recent collaboration.[33] For the peak broadening
Gaussian functions with a half-width of 0.3 eV were used. The CC2
calculations were performed using Def2-TZVP basis set and they were
conducted for the neutral compounds 1-4. The calculated spectra were
scaled to approximately fit the experimental peak heights. All DFT
calculations were performed by Gaussian 09[63] while CC2[64,65]
calculations were carried out by Turbomole 6.3.1.[66]
Synthesis.
Target benzoylguanidines were prepared from the corresponding
benzoylthioureas[49,50] employing following synthetic approaches:
Shinada-Ohfune HMDS[51] procedure (Method A, Scheme 2, guanidines
1-3,) or modification of HgCl2 procedure used by Cunha et al.[52] (Method
B, Scheme 2, guanidines 4 and 5,). In the Method B, a 25 % w/w
ammonia in water was used as the amine reagent in analogy to Estevez
et al.[53] All guanidines were prepared in satisfactory yields. Compounds
4 and 6 were also prepared by the Method A, but the reactions were low-
yielding with significantly larger amounts of the side-products. In our
hands, EtOAc/H2O extraction within the HMDS procedure (Method A) did
not provide complete removal of the side-products formed from EDCI.
Therefore, reaction was quenched by pouring the product mixture to the
1M aq. hydrochloric acid. From this solution, pure benzoylguanidines
were extracted with EtOAc. The products were obtained in a form of
hydrochloride salts as confirmed by 1H NMR and UV spectra.
Titration experiments.
Acknowledgements
Titration experiments were conducted starting from the benzoylguanidine
hydrochloride salts (BGU, c = 0.02-0.05 mM). Aliquots of anions (A) were
added in the form of tetrabutylammonium (TBA) salts until titration
endpoint was reached unless stated otherwise. Titration endpoint was
chosen when no change in spectra other than dilution effect was
observed and, in acetonitrile, it amounts 2-6 mol-equivalents. The
titrations were performed in acetonitrile (ACN), in water (W) and in 50-50
mixture of ACN and W (50-50). Concentrations of the anion solutions
were of 10-4 M in ACN or 10-3 M in W or 50-50. For the titrations we
choose highly basic acetate (Ac, pKa,ACN(HAc) = 22.3[54]), and fluoride (F,
pKa,ACN(HF) = 25.2[55]), moderately basic dihydrogenphosphate (DHP,
pKa,ACN(HDHP) = 17.5[44]) and weakly basic nitrate (N, pKa,ACN(HN) =
8.8[56]) and perchlorate (P) anions. In water, HDHP and HAc have pKas of
2.12 and 4.75, respectively[57] what is slightly lower than pKa2 and pKa3 of
glyphosate[58,59]. Fluoride, nitrate and perchlorate anions were used for
the titrations only in acetonitrile and only from the qualitative aspects. All
titrations were preceded by concentration dependence measurements.
No deviations from linearity up to the 1 x 10-4 mol dm-3 were observed
indicating no significant self-association in the concentration range 1-10 x
10-5 mol dm-3. All association constants were determined by use of
HypSpec software.[60] The spectras were recorded on PG Instruments
Ltd T80+ spectrophotometer
The work was supported by the Croatian Science Foundation
grant No. 9310. The calculations were performed on the Isabella
cluster (http://isabella.srce.hr) at the Zagreb University
Computing Center (SRCE). We also acknowledge Professor V.
Tomišić and dr. N. Bregović for the helpful discussion on the
determination of binding constants and pKa from the titration
data.
Keywords: benzoylguanidines • TD-DFT calculations •
guanylation • deprotonation • anion binding
[1]
[2]
B. C. Barja, M. S. Afonso, Environ. Sci. Techmol. 1998, 32, 3331−3335.
E. M. Jakubowski, J. M. McGuire, R. A. Evans, J. L. Edwards, S. W.
Hulet, B. J. Benton, J. S. Forster, D. C. Burnett, W. T. Muse, K. Matson,
C. L. Crouse, R. J. Mioduszewski, S. A. Thomson, J. Anal. Toxicol.
2004, 28, 357−363.
[3]
[4]
N. Busschaert, C. Caltagirone, W. Van Rossom, P. A. Gale, Chem. Rev.
2015, 115, 8038-8155.
K. A. Schug, W. Lindner, Chem. Rev. 2005, 105, 67-113.
This article is protected by copyright. All rights reserved.

10.1002/cplu.201800247
ChemPlusChem
FULL PAPER
[5]
A. E. Hargrove, S. Nieto, T. Zhang, J. L. Sessler, E. V. Anslyn, Chem.
[38] M. Ernzerhof, G. E. Scuseria, J. Chem. Phys. 1999, 110, 5029-5036.
[39] T. Yanai, D. Tew, N. Handy, Chem. Phys. Lett. 2004, 393, 51-57.
[40] Z. Glasovac, F. Pavošević, V. Štrukil, M. Eckert-Maksić, M. Schlangen,
R. Kretschmer, Int. J. Mass Spectrom. 2013, 354-355, 113-122.
[41] N. Bregović, N. Cindro, L. Frkanec, K. Užarević, V. Tomišić, Chem Eur.
J. 2014, 20, 15863-15871.
Rev. 2011, 111, 6603-6782.
[6]
[7]
[8]
[9]
F. P. Schmidtchen, M. Berger, Chem. Rev. 1997, 97, 1609-1646.
P.J. Gund, J. Chem. Educat. 1972, 49, 100–106.
A. Gobbi, G. Frenking, J. Am. Chem. Soc. 1993, 115, 2363–2372.
E. D. Raczynska, M. K. Cyranski, M. Gutowski, J. Rak, J.-F. Gal, P.-C.
Maria, M. Darowska, K. Duczmal, J. Phys. Org. Chem. 2003, 16, 91-
106.
[42] Z. Glasovac, M. Eckert-Maksić, Z. B. Maksić, New J. Chem. 2009, 33,
588–597.
[10] Z. B. Maksić, B. Kovačević, R. Vianello, Chem. Rev. 2012, 112, 5240-
5270.
[43] S. Espinosa, E. Bosch, M. Rosés, J. Chromatogr. A 2002, 964, 55–66.
[44] D. Barišić, N. Bregović, V. Tomišić, unpublished results.
[45] W.-X. Liu, Y.-B. Jiang, J. Org. Chem. 2008, 73, 1124-1127.
[46] R. Kleinmaier, R. M. Gschwind, Magn. Res. Chem. 2010, 48, 678-684.
[47] R. Kleinmaier, M. Keller, P. Igel, A. Buschauer, R. M. Gschwind, J. Am.
Chem. Soc. 2010, 132, 11223-11233.
[11] D. Margetić, Physico-chemical properties of organosuperbases, in T.
Ishikawa (Ed.), Superbases for Organic Synthesis: Guanidines,
Amidines, Phosphazenes and Related Organocatalysts, Wiley,
Chichester, 2009, pp. 9-48.
[12] B. Kovačević, Z. Glasovac, Z. B. Maksić, J. Phys. Org. Chem. 2002, 15,
765-774.
[48] C.-R. Zhang, J. S. Sears, B. Yang, S. G. Aziz, V. Coropceanu, J.-L.
Brꢀdas, J. Chem. Theory Comput. 2014, 10, 2379−2388.
[49] R. L. Frank P. V. Smith, Org. Synth. Coll. Vol. 3 (1955) p.735; 28 (1948)
p.89.
[13] Z. Glasovac, B. Kovačević, E. Meštrović, M. Eckert-Maksić,
Tetrahedron Lett. 2005, 46, 8733-8736.
[14] E. van Aken, H. Wynberg, F. van Bolhuis, J. Chem. Soc. Chem.
Commun. 1992, 629-630.
[50] C. R. Rasmussen, F. J. Villani Jr., L. E. Weaner, B. E. Reynolds, A. R.
Hood, L. R. Hecker, S. O. Nortey, A. Hanslin, M. J. Constanzo, E. T.
Powell, A. J. Molinari, Synthesis 1988, 456-459.
[15] P. A. Gale, C. Caltagirone, Coord. Chem. Rev. 2018, 354, 2-27.
[16] D. Leow. C.-H. Tan, Synlett, 2010, 1589-1605. See also other book
chapters in ref. 11.
[51] T. Shinada, T. Umezawa, T. Ando, H. Kozuma, Y. Ohfune, Tetrahedron
Lett. 2006, 47, 1945–1947.
[17] P. Selig, Synthesis, 2013, 45, 703-718.
[52] S. Cunha, M. B. Costa, H. B. Napolitano, C. Lariucci, I. Vencato,
Tetrahedron 2001, 57, 1671-1675.
[18] X. Fu, C.-H. Tan, Chem. Commun. 2011, 47, 8210-8222.
[19] P. I. Nagy, P. W. Erhardt, J. Phys. Chem. B, 2010, 114, 16436-16442.
[20] R. P. Dixon, S. J. Geib, A. D. Hamilton, J. Am. Chem. Soc. 1992, 114,
365-366.
[53] H. Esteves, Â. Fátima, R. de P. Castro, J. R. Sabino, F. Macedo Jr., T.
O. Brito, Tetrahedron Lett. 2015, 56, 6872–6874.
[54] K. Sarmini, E. Kenndler, J. Biochem. Biophys. Methods 1999, 38, 123–
137.
[21] E. Zeslawska, A. Schweinitz, A. Karcher, P. Sondermann, S. Sperl, J.
Stürzebecher, U. Jacob, J. Mol. Biol. 2000, 301, 465-475.
[22] E. Zeslawska, B. Oleksyn, K. Stadnicka, Struct. Chem. 2004, 15, 567-
572.
[55] C. R. Nicoleti, V. G. Marini, L. M. Zimmermann, V. G. Machado, J. Braz.
Chem. Soc. 2012, 23, 1488-1500. pKa,ACN(HF) value was estimated
from the data in DMSO (pKa,DMSO(HF) = 15.0) using linear correlation
function.
[23] S.-H. Loh, B. Sun, R. D. Vaughan-Jones, Brit. J. Pharm. 1996, 118,
1905-1912.
[56] J. T. Muckerman, J. H. Skone, M. Ning, Y. Wasada-Tsutsui, Biochim.
Biophys. Acta 2013, 1827, 882–891.
[24] C. Schmuck, Coord. Chem. Rev. 2006, 250, 3053-3067.
[25] K. Haav, S. A. Kadam, L. Toom, P. A. Gale, N. Busschaert, M. Wenzel,
J. R. Hiscock, I. L. Kirby, T. Haljasorg, M. Lõkov, I. Leito, J. Org. Chem.
2013, 78, 7796-7808.
[57] R. C. Weast (Editor-in-chief), "CRC Handbook of Chemistry and
Physics 70th Ed." CRC Press Inc. Boca Raton, FL, USA, 1989.
[58] Glyphosate pKa2, pKa3 and pKa4 = 2.09, 5.52 and 10.28, respectively;
see ref. 1.
[26] D. Esteban Gómez, L. Fabrizzi, M. Licchelli, E. Monzani, Org. Biomol.
Chem. 2005, 3, 1495-1500.
[59] Glyphosate pKa2, pKa3 and pKa4 = 2.27, 5.57 and 10.25; T. G. Appleton,
J. R. Hall, I. J. McMahon, Inorg. Chem. 1986, 25, 726−734.
[60] HypSpec software P. Gans, A. Sabatini, A. Vacca, Talanta 1996, 43,
1739-1753.
[27] C. Perez-Casas, A. Yatsimirsky, J. Org. Chem. 2008, 73, 2275-2284.
[28] P. A. Gale, Chem. Commun. 2011, 47, 82-86.
[29] L. S. Evans, P. A. Gale, M. E. Light, R. Quesada, Chem. Commun.
2006, 965-967.
[61] A. V. Marenich, C. J. Cramer, D. G. Truhlar, J. Phys. Chem. B, 2009,
113, 6378-6396.
[30] S. Camiolo, P. A. Gale, M. B. Hursthouse, M. E. Light, A. J. Shi, Chem.
Commun. 2002, 758–759.
[62] A. D. Becke, J. Chem. Phys. 1993, 98, 5648-5652.
[63] 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, 2009.
[31] V. Amendola, D. Esteban Gómez, L. Fabrizzi, M. Licchelli, Acc. Chem.
Res. 2006, 39, 343-353.
[32] V. Amendola, L. Fabrizzi, L. Mosca, Chem. Soc. Rev. 2010, 39, 3889-
3915.
[33] I. Antol, Z. Glasovac, D. Margetić, R. Crespo-Otero, M. Barbatti, J.
Phys. Chem. A 2016, 120, 7088-7100.
[64] O. Christiansen, H. Koch, P. Jørgensen, Chem. Phys. Lett. 1995, 243,
409–418.
[34] L. Nie, Z. Li, J. Han, X. Zhang, R. Yang, W.-X. Liu, F.-Y. Wu, J.-W. Xie,
Y.-F. Zhao, Y.-B. Jiang, J. Org. Chem. 2004, 69, 6449-6454.
[35] F.-Y. Wu, Z. Li, L. Guo, X. Wang, M.-H. Lin, Y.-F. Zhao, Y-B Jiang, Org.
Biomol. Chem. 2006, 4, 624–630.
[65] a) C. Hättig, F. Weigend, J. Chem. Phys. 2000, 113, 5154-5161; b) C.
Hättig, A. Köhn, J. Chem. Phys. 2002, 117, 6939-6951; c) C. Hättig, A.
Hellweg, A. Köhn, Phys. Chem. Chem. Phys. 2006, 8, 1159-1169.
[66] Turbomole 6.3. R. Ahlrichs, M. Bär, M. Häser, H. Horn, C. Kölmel,
Chem. Phys. Lett. 1989, 162, 165–169.
[36] C. Rether, W. Sicking, R. Boese, C. Schmuck, Beil. J. Org. Chem. 2010,
6, doi:10.3762/bjoc.6.3.
[37] C. Adamo, V. Barone, J. Chem. Phys. 1999, 110, 6158-6169.
This article is protected by copyright. All rights reserved.

10.1002/cplu.201800247
ChemPlusChem
FULL PAPER
Entry for the Table of Contents (Please choose one layout)
Layout 1:
FULL PAPER
In search of novel anion sensors,
benzoylguanidinium salts were
Z. Glasovac*, L. Barešić, I. Antol, D.
Margetić
investigated experimentally and
computationally. Electronic structure
properties and observable changes in
UV/Vis absorption spectra upon anion
binding were tuned by variation of the
functional groups attached. The most
responsive derivative is suitable for
the naked eye detection of the
Page No. – Page No.
Benzoylguanidines as the Anion
Responsive Systems
dihydrogenphosphate and acetate.
Layout 2:
FULL PAPER
Author(s), Corresponding Author(s)*
((Insert TOC Graphic here; max. width: 11.5 cm; max. height: 2.5 cm))
Page No. – Page No.
Title
Text for Table of Contents
This article is protected by copyright. All rights reserved.