Anion receptors containing - NH binding sites: Hydrogen-bond formation or neat proton transfer?

Article abstract of DOI:10.1002/chem.200400592

When the amide-containing receptor 1⁺ is in a solution of dimethyl sulfoxide (DMSO) in the presence of basic anions (CH₃COO, F, H₂PO₄), it undergoes deprotonation of the -NH fragment to give the corresponding zw

Full text of DOI:10.1002/chem.200400592

Anion Receptors Containing -NH Binding Sites: Hydrogen-Bond Formation
or Neat Proton Transfer?
[
a]
[b]
[a]
[a]
Valeria Amendola, Massimo Boiocchi, Luigi Fabbrizzi,* and Arianna Palchetti
ꢀ
ꢀ
ꢀ
+
+
Abstract: When the amide-containing
receptor 1 is in a solution of dimethyl
basic anions (Cl , Br , NO ), 1 es-
acidic receptor
2
undergoes neat
3
+
tablishes true hydrogen-bond interac-
tions of decreasing intensity. The less
proton transfer with only the more
basic anions CH COO and F , and es-
tablishes hydrogen-bond interactions
with H PO . An empirical criterion
for discerning neutralisation and hydro-
gen bonding, based on UV/Vis and
H NMR spectra, is proposed.
ꢀ
ꢀ
sulfoxide (DMSO) in the presence of
3
ꢀ
ꢀ
ꢀ
basic anions (CH COO , F , H PO ),
3
2
4
ꢀ
it undergoes deprotonation of the -NH
fragment to give the corresponding
zwitterion, which can be isolated as a
crystalline solid. In the presence of less
2
4
Keywords: anions
bonds molecular recognition
proton transfer · receptors
·
hydrogen
·
·
1
Introduction
most electronegative atoms, fluoride and oxygen (in the
form of inorganic and organic oxo anions), while recognition
studies are preferably carried out in aprotic media (e.g., di-
Anions play an important role in medicine, biology, environ-
[1]
mental and food sciences, accounting for the increasing in-
terest in the design of synthetic anion receptors. Methodolo-
gies for the prompt detection and quantitative determina-
tion of anions are being developed and, within the last
decade, a variety of molecular sensors for anions have been
methyl sulfoxide (DMSO), MeCN, CHCl ), to avoid compe-
3
tition of the solvent (e.g., water or alcohols) as a hydrogen-
bond donor.
From this perspective, we considered the possibility of de-
signing anion receptors that contained both a positively
charged group and a hydrogen-bond-donor group. The rea-
sons for this approach were two-fold: firstly, the presence of
a proximate, positive charge should enhance the tendency of
the -NH group to act as a hydrogen-bond donor, and sec-
ondly, the positively charged group itself should provide ad-
ditional electrostatic interaction with the negatively charged
[
2]
synthesised. Most of these sensors contain chromogenic or
fluorogenic groups that are covalently or non-covalently
linked to the receptor moiety, thus enabling the colorimetric
and fluorimetric sensing of anions with both temporal and
[3]
spatial resolution. The nature of the anion–receptor inter-
action is a matter of choice for the designer; it can be elec-
trostatic, in which case the receptor must be equipped with
positively charged groups (e.g., ammonium, alkylammoni-
+
analyte. We prepared the molecular ions 1 (1-benzyl-3-(tol-
uene-4-sulfonylamino)pyridinium; -NH binding site from
[4]
um, pyridinium, guanidinium), or it can be based on hy-
drogen bonding. In the latter case, the receptor must pro-
vide hydrogen-bond-donor groups, in most cases the -NH
fragment of carboxyamides, sulfonamides, ureas, thioureas
the p-toluenesulfonamide moiety, positive charge from the
+
benzylpyridinium fragment), and 2 (2-benzyl-9H-b-carbo-
lin-2-ium; -NH from the pyrrole subunit, positive charge
again from a benzylpyridinium fragment).
[
5]
and pyrroles. Receptors based on hydrogen bonding are
expected to interact principally with anions containing the
[
a] Dr. V. Amendola, Prof. L. Fabbrizzi, A. Palchetti
Dipartimento di Chimica Generale, Università di Pavia
viale Taramelli 12, 27100 Pavia (Italy)
Fax : (+39)0382-528-544
E-mail: luigi.fabbrizzi@unipv.it
+
+
The interaction of 1 and 2 with a variety of anions was
investigated by performing titration experiments in DMSO,
and analysing the products by UV/Vis spectrophotometry
[
b] Dr. M. Boiocchi
Centro Grandi Strumenti, Università di Pavia
via Bassi 21, 27100 Pavia (Italy)
120
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/chem.200400592
Chem. Eur. J. 2005, 11, 120 – 127

FULL PAPER
1
and H NMR spectroscopy. We demonstrated that anion–re-
ceptor interactions may involve either a neat proton transfer
or hydrogen-bond interactions, depending on the intrinsic
acidity of the -NH group and the basic tendencies of the
anion.
Results and Discussion
+
+
Both 1 and 2 were obtained as hexafluorophosphate salts,
following reaction of the parent molecules with benzylbro-
mide and subsequent recrystallisation in the presence of
[
NH ]PF . The crystalline form of 1-PF was isolated and its
4
6
6
molecular structure was determined by X-ray diffraction
studies (Figure 1). The structure reveals the reciprocal inter-
+
ꢀ5
Figure 2. Spectrophotometric titration of a solution of 1 (4”10 m) in
DMSO with a standard solution of tetrabutylammonium acetate in
DMSO. Inset: molar absorbance at 360 nm versus number of equivalents
ꢀ
6
+
of acetate (n) for a 5”10 m solution of 1 .
example, the family of spectra obtained during the titration
+
of 1 with acetate. The intensity of the bands at 290 and
3
60 nm (the latter being responsible for the yellow colour)
progressively increased, reaching a limiting value of the
ꢀ
1
ꢀ1
molar absorbance (15700 and 3100m cm , respectively)
after the addition of one equivalent of acetate.
The titration profile shows a steep curvature, which
should correspond to a particularly high equilibrium con-
stant; however, the p parameter, (p=[concentration of com-
plex]/[maximum possible concentration of complex]) was
greater than 0.8, which prevents the reliable determination
[6]
of an equilibrium constant. Therefore, we can only state
that logK is greater than 6. Notably, even in the case of a
ꢀ
6
more dilute solution (e.g., 5”10 m, as shown in the inset of
Figure 2), a steep curvature was obtained and p was greater
than 0.8.
+
Similar results were obtained following the titration of 1
6
Figure 1. The ORTEP view of the 1-PF molecule shows that two sym-
ꢀ
ꢀ
+
with F and H PO ; the bands at 290 and 360 nm increased
2 4
metrically equivalent 1 ions (related by an inversion centre) are linked
by hydrogen bonds (dashed lines) to form a molecular dimer (thermal el-
lipsoids are drawn at the 30% probability level, hydrogens bonded to
carbon atoms were omitted for clarity). Features of the NꢀH···O interac-
tion are N(1)···O(1)’=2.938(3) Š, H(1N)···O(1)’=2.074(24) Š, N(1)ꢀ
in intensity and the same limiting values of the molar absor-
ꢀ1
ꢀ1
bances were reached (15700 and 3100m cm , respectively,
see also Table 1). In all cases, a 1:1 reaction stoichiometry
was apparent; however, the steep curvature of the titration
H(1N)···O(1)’=172.8(23)8 (symmetry code ’=ꢀx, ꢀy+1, ꢀz).
+
Table 1. The equilibrium constants (logarithmic values) determined by
action of two 1 ions, by means of hydrogen bonds involving
+
+
ꢀ
the spectrophotometric titration of 1 (=LH ) with anions X , and the
the -NH group from one ion and a sulfonamide oxygen
atom from the other. Each 1 ion exhibits a conformation
ꢀ1
ꢀ1
molar absorbances (m cm ) at 290 and 360 nm of the solutions contain-
+
+
ing 1 and excess of the anion in DMSO.
that may minimise the intermolecular steric repulsions
within the hydrogen-bonded dimer. In any case, interactions
do not involve the PF6 ion, which was deliberately chosen
+
ꢀ
equilibrium: LH +X ÐL+HX
Anion
logK
e
2
90
e360
ꢀ
ꢀ
CH
F
H PO
2
3
COO
>6
>6
>6
15700
15700
15700
3100
3100
3100
ꢀ
to act as a very poor hydrogen-bond acceptor.
4ꢀ
Next, a solution of 1-PF in DMSO was titrated with a so-
6
ꢀ
+
ꢀ
ꢀ
lution of a [Bu N]X salt in DMSO, in which X represented
equilibrium: [LH] +X Ð[LH···X]
4
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
Anion
logK
e
290
e
360
the F , CH COO , H PO , NO3 , Cl or Br ion. In all
3
2
4
ꢀ
cases, the resulting solution had a yellow colour of greater
or lesser intensity; however, the UV/Vis absorption spectra
differed significantly from salt to salt. Figure 2 shows, as an
Cl
4.52Æ0.02
3.55Æ0.02
2.97Æ0.04
4500
3100
3500
450
310
350
ꢀ
Br
3ꢀ
NO
Chem. Eur. J. 2005, 11, 120 – 127
www.chemeurj.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
121

L. Fabbrizzi et al.
profile prevented determination of the equilibrium constant
(
logK>6).
+
Different results were obtained following titration of 1
ꢀ
ꢀ
ꢀ
with NO3 , Cl and Br . Although the intensity of absorp-
tion at 290 and 360 nm increased, the limiting values ob-
tained after addition of excess anion were distinctly lower
than those in the above examples, and varied according to
the nature of the anion (see Table 1). As an example,
Figure 3 displays the family of spectra obtained from the ti-
+
ꢀ4
Figure 3. Spectrophotometric titration of a solution of 1 (9”10 m) in
DMSO with a standard solution of benzyltributylammonium chloride in
DMSO. Inset: molar absorbance at 360 nm versus number of equivalents
of chloride (n).
ꢀ
tration with Cl . In all cases, nonlinear least-squares analysis
of the titration data indicated a 1:1 stoichiometry. In addi-
tion, the titration profiles displayed gentle curvature and the
p parameter was less than 0.8.
1
+
ꢀ2
Figure 4. H NMR titration of 1 in [D
6
]DMSO (10 m) with a standard
]DMSO. The spectra were
solution of tetrabutylammonium acetate in [D
recorded after the addition of a) 0, b) 0.25, c) 0.5 and d) 1.0 equivalents
of acetate.
6
The equilibrium constants, calculated from the nonlinear
fitting of titration profiles, are reported in Table 1. The logK
+
ꢀ
values for the 1 /X interaction decrease along the series:
ꢀ
ꢀ
ꢀ
4
ꢀ
ꢀ
ꢀ
CH COO , F , H PO > Cl > Br > NO3
.
changed. Similar results, although with less pronounced
shifts, were observed upon titration with bromide.
The slow evaporation of a water/acetonitrile (2:1 v/v) so-
lution containing equimolar amounts of 1-PF6 and
[Bu N]CH COO yielded pale yellow crystals. X-ray diffrac-
tion revealed that these crystals contained no anions; rather,
the positive charge of the benzyl pyridinium group was bal-
anced by the negative charge resulting from deprotonation
of the -NH fragment of the amide subunit. The zwitterion is
3
2
Further insights into the nature of receptor–anion interac-
1
tions were provided by the analysis of H NMR spectra.
ꢀ
2
Figure 4 displays the spectra for titration of a 10
m
[
[
D ]DMSO solution of 1-PF with a [D ]DMSO solution of
6
6
6
4
3
Bu N]CH COO. Upon addition of acetate, the -NH signal
4
3
at d=11.6 ppm disappears and the majority of signals perti-
nent to the -CH hydrogens shift distinctly upfield, with the
greatest shift observed for -C(3)-H (Dppm=ꢀ0.64). Similar
ꢀ
ꢀ
4
+
results were obtained for F and H PO , in particular, the
shown in Figure 6. Comparison with 1 revealed significant
2
same limiting d values (obtained after the addition of one or
more equivalents of anion) were recorded.
changes in the geometry of the sulfonamide group (Table 2).
Following deprotonation, the N(1)ꢀS(1) bond length de-
A different behaviour was observed with the remaining
anions. Figure 5 shows the spectra obtained in the course of
creases, whereas the S=O distance increases. This suggests
that the lone pair on the deprotonated nitrogen atom is
+
titration of 1 with benzyltributylammonium chloride in
partly delocalised by a p mechanism over the >SO frag-
2
ꢀ
ꢀ
[
D ]DMSO. In contrast to the results for CH COO , F and
ment. This is described in part by the resonance representa-
tion in Scheme 1 (formula b). On the other hand, some elec-
tronic charge must also delocalise through a p mechanism
over the pyridinium ring, as suggested by the distinct reduc-
tion of the N(1)ꢀC(1) distance and accounted for by the res-
6
3
ꢀ
H PO4 , the peakrelative to the amide group (peak1), did
2
not disappear, but instead was shifted significantly down-
field. The protons in positions 2, 3 and 5 were shifted down-
field, whereas those in positions 4 and 6 remained un-
122
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2005, 11, 120 – 127

Anion–Receptor Interactions
FULL PAPER
Table 2. Selected bond lengths [Š] and angles [8] for the sulfonamide
group in 1-PF and zwitterion 1.
6
1
-PF
6
zwitterion 1
N(1)ꢀC(1)
N(1)ꢀS(1)
S(1)ꢀO(1)
S(1)ꢀO(2)
S(1)ꢀC(13)
1.411(3)
1.644(3)
1.439(2)
1.419(2)
1.748(3)
124.36(20)
103.14(13)
107.57(14)
106.24(14)
120.51(14)
109.46(14)
108.92(14)
1.374(5)
1.570(3)
1.451(3)
1.451(3)
1.779(3)
120.51(26)
113.78(16)
106.39(16)
109.38(18)
114.84(16)
105.79(15)
106.36(16)
C(1)-N(1)-S(1)
N(1)-S(1)-O(1)
N(1)-S(1)-O(2)
N(1)-S(1)-C(13)
O(1)-S(1)-O(2)
O(1)-S(1)-C(13)
O(2)-S(1)-C(13)
Scheme 1. Resonance representation of the deprotonated form of recep-
+
tor 1 . Formula b accounts for the shortening of the NꢀS bond and for
1
+
ꢀ2
the lengthening of the S=O bonds. Formula c accounts for the shortening
of the NꢀC bond.
Figure 5. H NMR titration of 1 in [D
6
]DMSO (10 m) with a standard
]DMSO. The spectra
solution of benzyltributylammonium chloride in [D
6
were recorded after the addition of a) 0 and b) 10 equivalents of the
anion solution.
gous zwitterion, in which a methylpyridinium fragment is
linked to a methylsulfonamide group (NꢀC: 1.372 Š; NꢀS:
[
7]
1
.581 Š; S=O: 1.455 and 1.448 Š; SꢀC: 1.765 Š). Most in-
terestingly, a DMSO solution of the zwitterionic compound
showed the same spectroscopic features (UV/Vis and NMR
spectroscopy) as a solution containing the 1-PF salt and a
6
ꢀ
ꢀ
ꢀ
slight excess of [Bu N]X (in which X =F , CH COO or
H PO ). This unequivocally demonstrates that, on addition
of F , CH COO or H PO to a solution of 1-PF , a proton
4
3
ꢀ
2
4
ꢀ
ꢀ
ꢀ
3
2
4
6
+
ꢀ
is neatly transferred from 1 to X . In other words, a neu-
+
+
tralisation process takes place between the acid 1 (or BH )
ꢀ
ꢀ
and the base X (or A ), according to the classical Brønsted
proton transfer equilibrium (1).
þ
ꢀ
BH þ A Ð B þ HA
ð1Þ
The zwitterionic nature of 1 accounts for the spectroscop-
ic properties observed: The deprotonation of -NH causes
the formation of an electrical dipole with an especially large
moment, which explains the hyperchromic effect observed
in the absorption spectrum. Partial p delocalisation of the
negative electrical charge over the entire molecular frame-
workof the zwitterion accounts for an increase in the shield-
ing effect and the consequential general upfield shift of -CH
signals.
+
Figure 6. An ORTEP plot of the zwitterion 1 (thermal ellipsoids are
drawn at the 30% probability level, hydrogens bonded to carbon atoms
were omitted for clarity). The deprotonation of the N(1) atom ensures
the electroneutrality of the molecule.
ꢀ
ꢀ
onance formula c in Scheme 1 (which is one of the several
possible). The geometrical features of the sulfonamide
group of the zwitterion are very similar to those of an analo-
On the other hand, it is suggested that for Cl , Br and
ꢀ
NO3 ions, a genuine hydrogen-bond interaction is establish-
ed with the receptor 1 . The occurrence of such an interac-
+
Chem. Eur. J. 2005, 11, 120 – 127
www.chemeurj.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
123

L. Fabbrizzi et al.
tion removes charge density from the receptor, which gener-
ates a general deshielding of protons. This effect is especial-
ly detectable for the amide proton, which is directly in-
volved in the interaction, and for C(6)ꢀH, C(2)ꢀH and
C(3)ꢀH, which are adjacent to the positively charged pyridi-
nium nitrogen atom. The downfield shift is less pronounced
ꢀ
for Br , which, in view of the smaller electronegativity of
bromine, competes to a lesser extent for the -NH proton,
thus inducing a smaller deshielding effect.
ꢀ
ꢀ
ꢀ
The sequence of logK values (Cl >Br >NO ) reflects
3
the decreasing intensity of the anion–receptor hydrogen-
bond interaction. Notably, the different hydrogen-bond in-
teraction energies correspond to the different molar absor-
bances of the UV/Vis bands, even if e values do not corre-
late closely to logK values.
On the basis of this evidence, the following empirical cri-
terion for discerning neat proton transfer and true hydro-
gen-bond formation, following the interaction between an
+
Figure 7. Matching of the acidity scale for the receptor [BH] and the ba-
ꢀ
ꢀ
sicity scale for A (pK
a
values for anions are tentative; CH
have been arbitrarily assigned the same pK value). Receptor 1
a
3
COO and
-NH-containing receptor and an anion, is tentatively pro-
ꢀ
+
F
posed: a) A general upfield shift of proton signals within the
NMR spectrum, together with high and constant limiting ab-
sorbance of UV/Vis bands, indicate deprotonation; b) a se-
lective downfield shift and lower and variable e values indi-
cate authentic hydrogen-bond formation.
transfers a proton to anions that lie above the solid horizontal line and
establishes a true hydrogen-bond interaction with the anions that lie
below the line. Receptor 2 transfers a proton to anions that lie above
+
the dashed horizontal line and establishes a true hydrogen-bond interac-
ꢀ
tion with the H
2
PO
4
ion, which remains below the dashed horizontal
ꢀ
line. Cl and anions that lie below this line exhibit a basicity that is too
low and a hydrogen-bond-acceptor tendency that is too poor to enable
them to interact with 2 .
+
Next, we attempted to explain why receptor 1 transfers
ꢀ
+
a proton to some anions (e.g., CH COO ), but establishes
3
ꢀ
hydrogen-bond interactions with others (e.g., Cl ). The
equilibrium constant (K) for the acid–base neutralisation
equilibrium (1) results from the ratio of the two acidity con-
We tested the reliability of this approach by considering a
further -NH-containing receptor 2 . This possesses a pyr-
+
stants (KBH
+
/KHA), and the pKHA values for some organic
ꢀ
and inorganic anions are known (e.g., pK for CH COO
role-type -NH fragment, whose intrinsic acidity is lower
than that of sulfonamide, and should be enhanced by the
presence of a proximate pyridinium group. In this receptor,
HA
3
[8]
is 12.3 ). Therefore, having established that logK for
ꢀ
+
CH COO is >6, the pK
BH+
value for 1 is calculated to be
3
ꢀ
È
<
6.3. In Figure 7, pK (HA/A ) values for the anions investi-
the -NH fragment is separated from the >N group by two
a
+
gated are tentatively positioned on the vertical scale on the
benzene carbon atoms, as in 1 .
ꢀ
ꢀ
right (CH COO and F have been set at the same level,
Figure 8 displays the family of spectra obtained when a
3
ꢀ
4
even though acetate is slightly more basic). On the vertical
solution of 2-PF in DMSO (8.4”10 m) was titrated with a
6
+
scale on the left, the pK value of 1 is indicated. The two
a
vertical lines are juxtaposed in such a way that the pK of
a
+
ꢀ
1
lies somewhere between the pK
of H PO and the
HA 2 4
pKHA of Cl . For anions whose pK values are higher than
ꢀ
HA
+
the pKBH
+
of 1 , the equilibrium constant for the acid–base
neutralisation equilibrium (1) is distinctly higher than 1 and
proton transfer occurs. On the other hand, for anions whose
+
pK values are lower than the pK of 1 , the neutralisation
a
a
equilibrium is not favoured and the hydrogen-bond interac-
tion is not followed by definitive proton transfer. Therefore,
the occurrence of a Brønsted acid–base equilibrium or the
establishment of a hydrogen-bond interaction in a given
medium depends simply on the acidity of the receptorꢁs
-NH group with respect to the basicity of the anion. Anions
whose pKHA value is lower than the pK of the -NH acid
a
(
i.e., below the solid horizontal line in Figure 7) will estab-
lish a true hydrogen-bond interaction, whereas anions with
+
Figure 8. Spectrophotometric titration of a solution of 2 in DMSO (8.4”
higher pK values (above the dashed line) will induce defini-
ꢀ4
a
1
0
m) with
a
standard solution of tetrabutylammonium acetate in
DMSO. Inset: molar absorbance at 454 nm versus the number of equiva-
tive -NH deprotonation and the formation of HA.
ꢀ
4
+
lents of acetate (n) for a 1”10 m solution of 2 .
124
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2005, 11, 120 – 127

Anion–Receptor Interactions
FULL PAPER
standard solution of [Bu N]CH COO. Upon acetate addi-
4
3
tion, the absorption band at 380 nm decreases, whereas a
new band at 454 nm develops and increases in intensity to a
plateau after the addition of one equivalent of anion (the ti-
tration profile in the inset of Figure 8 refers to a solution of
ꢀ
4
1
”10 m). The smooth curvature of the titration profile ena-
bles the accurate determination of the equilibrium constant
see Table 3).
(
Table 3. The equilibrium constants (logarithmic values) determined by
+
+
ꢀ
the spectrophotometric titration of 2 (=LH ) with anions X , and the
+
ꢀ
1
ꢀ1
molar absorbances (m cm ) at 454 nm of the solutions containing 2
and excess of the anion in DMSO.
+
ꢀ
equilibrium: LH +X ÐL+HX
logK
Anion
e
454
ꢀ
CH
F
3
COO
5.10Æ0.05
4.77Æ0.02
2300
2300
ꢀ
+
ꢀ
ꢀ
equilibrium: [LH] +X Ð[LH···X]
Anion
logK
e
454
4ꢀ
H
2
PO
3.11Æ0.01
1800
1
Figure 9 shows the H NMR spectra obtained during the
+
titration of a [D ]DMSO solution of 2 with [Bu N]CH -
6
4
3
COO. All of the -CH signals are shifted upfield; the most
significant shifts are for the carbolinium unit, due to the de-
creasing positive charge on the pyridinium fragment (maxi-
mum Dppm=ꢀ0.4 for -C(5)ꢀH and -C(6)ꢀH). Similar re-
sults for limiting absorbance and d of -CH signals (upfield
shift) were obtained upon titration with fluoride. The logK
value, calculated from the titration profile, is 4.77Æ0.02.
These results indicate neutralisation and proton transfer, fol-
1
+
ꢀ2
+
6
Figure 9. H NMR spectra for the titration of 2 in [D ]DMSO (10 m)
lowing the interaction of 2 with acetate and fluoride. In ad-
6
with a standard solution of tetrabutylammonium acetate in [D ]DMSO.
The spectra were recorded after the addition of a) 0, b) 0.25, c) 0.75, d)
1.0 and e) 2.0 equivalents of acetate.
dition, using the relationship K=KBH+/KHA and the reported
ꢀ
+
pKHA value for CH COO in DMSO, the pKBH+ for 2 is
3
calculated to be 7.2. Therefore, the higher acidity of sulfona-
mide with respect to pyrrole is maintained in the presence
of a proximate, positively charged group. Deprotonation of
ꢀ
a pyrrole subunit in the presence of F in a CH Cl solution
2
2
[9]
has been recently observed.
Figure 10 shows the family of UV/Vis spectra obtained
+
upon titration of
a
solution of
2
in DMSO with
[
Bu N]H PO (titration profile in the inset, calculated
4
2
4
logK=3.11Æ0.01). The limiting value of the molar absorb-
ance at 454 nm is notably lower than that observed upon ti-
ꢀ
ꢀ
tration with CH COO and F (see Table 3).
3
1
Figure 11 displays the H NMR spectra obtained during
+
the titration of
a
solution of
2
in DMSO with
[
Bu N]H PO . There is a significant downfield shift of the
4 2 4
proton in position 5 (Dppm=+0.27); in addition, the up-
field shifts of the other carbolinium protons are less intense
than those observed in the titration with acetate. No shift is
observed for the aliphatic group -CH -, which is bound di-
2
rectly to the positively charged nitrogen atom. On the basis
of this, we suggest that a true hydrogen-bond interaction is
+
Figure 10. Spectrophotometric titration of a solution of 2 in DMSO
(
ꢀ4
8.2”10 m) with a standard solution of tetrabutylammonium monobasic
phosphate in the same solvent. Inset: the molar absorbance at 454 nm
+
ꢀ
established between 2 and H PO , and that -C(5)ꢀH is di-
2
4
rectly involved in this interaction. Notably, addition of the
versus the number of equivalents of phosphate (n).
Chem. Eur. J. 2005, 11, 120 – 127
www.chemeurj.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
125

L. Fabbrizzi et al.
would be expected for an analogous solvent, for example,
MeCN. Although most of the significant anions are repre-
sented on the right-hand vertical line, a number of -NH re-
ceptors of varying acidity have yet to be placed and correct-
ly juxtaposed on the left-hand vertical line. This diagram
should be helpful in the design of anion receptors. For ex-
ample, to increase the hydrogen-bond-donor tendency of
the receptor, particularly acidic -NH groups could be
chosen, or acidity could be enhanced by using electron-with-
drawing substituents (e.g., -NO , -CF , -CN). However, high
2
3
receptor acidity may induce neat proton transfer, which is
beyond the realm of supramolecular chemistry and within
the domain of classical Brønsted acid–base reactions. Ideal-
ly, the pK of the -NH-containing receptor should be coinci-
a
dent with (or slightly more positive than) the pK of the
HA
ꢀ
A
anion, corresponding to the premise that the strongest
hydrogen-bond interactions are established between an
ꢀ
anion A and its conjugate acid HA (which indeed refer to
[10]
the same pK value).
a
Experimental Section
4
-Methyl-N-pyridin-3-yl-benzensulfonamide: A solution containing 3-
aminopyridine (1.5 g, 15.9 mmol) in pyridine (70 mL) and 4-methylben-
zenesulfonyl chloride (3.8 g, 19.9 mmol) was refluxed for 2 h, then the re-
action mixture was poured onto iced water (150 mL). The white precipi-
tate was collected by filtration under vacuum and washed with pure
water (yield 3.35 g; 85%). The product purity was controlled by perform-
ꢀ
1
ing TLC (SiO
2
, 100% AcOEt:
H NMR (400 MHz, [D ]DMSO): d=10.50 (s, 1H; NH), 8.28 (d, 1H;
CH py), 8.25 (d, 1H; CH py), 7.63 (d, 2H; CH bz), 7.50 (dd, 1H; CH
py), 7.36 (d, 2H; CH bz), 7.28 (dd, 1H; CH py), 2.4 ppm (s, 3H; CH ).
-Benzyl-3-(toluene-4-sulfonylamino)pyridinium hexafluorophosphate (1-
f 12 12 2 2
R =0.5). C H N SO (248 gmol ).
1
6
3
1
+
ꢀ2
Figure 11. H NMR spectra for the titration of 2 in [D
with a standard solution of tetrabutylammonium monobasic phosphate in
]DMSO. The spectra were recorded after the addition of a) 0, b) 0.25,
c) 0.75, d) 1.0 and e) 2.0 equivalents of phosphate.
6
]DMSO (10 m)
1
PF
6
): An amount of 4-methyl-N-pyridin-3-yl-benzensulfonamide (0.4 g,
(70 mL). An excess of benzyl bromide
0.41 g, 2.4 mmol) was added and the resulting solution was refluxed for
4 h. The solvent was then removed by using a rotary evaporator to give
an oily residue, which was dissolved in hot water and treated with a satu-
rated aqueous solution of NH PF . The white precipitate of 1-PF was re-
covered by filtration under vacuum (yield 0.45 g; 83%). C19 SO PF
]DMSO): d=11.65 (s, 1H; NH),
.86 (d, 1H; CH(7)), 8.68 (s, 1H; CH(4)), 8.11 (d, 1H; CH(5)), 8.01 (dd,
1H; CH(6)), 7.65 (d, 2H; CH(2)), 7.46 (m, 5H; bz), 7.34 (d, 2H; CH(1)),
(d, 2H; CH(2)), 5.84 (s, 2H; CH (8)), 2.4 ppm (s, 3H; CH ).
2-Benzyl-9H-b-carbolin-2-ium (2-PF ): An amount of b-carboline (0.11 g,
[
D
6
1
(
2
.6 mmol) was dissolved in CHCl
3
ꢀ
butylammonium salts of other anions (NO3 and halides)
caused no significant modification of the UV/Vis and
4
6
6
H
17
N
2
2
6
ꢀ
1
1
(
484 gmol ). H NMR (400 MHz, [D
6
1
H NMR spectra, indicating a lackof interaction. Therefore,
the less acidic receptor 2 can still transfer a proton to the
moderately strong bases CH COO and F , establish hydro-
gen-bond interactions with H PO , and does not interact at
all with the weakly basic anions Cl , Br and NO3 . This
scenario is represented in Figure 7. The pKBH
now be positioned between the pK values of CH COO ,
F and H PO , thus distinguishing between acid–base neu-
tralisation (above the dashed horizontal line) and hydrogen-
bond formation (below). Cl and other anions exhibit a ba-
sicity that is too low and a hydrogen-bond-acceptor tenden-
cy that is too poor to enable them to interact with 2 .
8
+
ꢀ
ꢀ
3
2
3
ꢀ
2
4
6
ꢀ
ꢀ
ꢀ
0.65 mmol) was dissolved in CHCl3 (70 mL). An excess of benzyl bro-
mide (0.16 g, 0.98 mmol) was added and the resulting solution was re-
fluxed for 24 h. The solvent was then removed by using a rotary evapora-
tor to give a yellow solid residue, which was dissolved in hot water and
+
+
of 2 must
ꢀ
HA
3
ꢀ
ꢀ
treated with a saturated, aqueous solution of NH
tate of 1-PF was recovered by filtration under vacuum (yield 0.21 g;
80%). C18
s, 1H; CH(5)), 8.55 (d, 1H; CH(7)), 8.45 (s, 1H; CH(6)), 8.40 (d, 1H;
4 6
PF . The white precipi-
2
4
6
ꢀ
1
1
ꢀ
15 2 6 3
H N PF (404 gmol ). H NMR (400 MHz, CD CN): d=9.1
(
CH(4)), 7.80 (dd, 2H; CH(1)-CH(2)), 7.50 (m, 6H; CH(3), bz), 5.85 ppm
+
(s, 2H; CH (8)).
2
General procedures and materials: All reagents for syntheses were pur-
chased from Aldrich/Fluka and used without further purification. UV/Vis
spectra were recorded on a Varian CARY 100 spectrophotometer with a
quartz cuvette (path length: 1 cm). NMR spectra were recorded on a
Bruker Avance 400 spectrometer, operating at 9.37 T. Spectrophotomet-
Conclusion
Figure 7 is generally applicable for the solvent used in this
study (DMSO); however, a qualitatively similar diagram
ꢀ4
ꢀ5
ric titrations were performed at 258C on 10 and 10 m solutions of 1-
PF and 2-PF in DMSO (polarographic grade). Aliquots of a fresh
6
6
126
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2005, 11, 120 – 127

Anion–Receptor Interactions
FULL PAPER
[
15]
Bu
4
NX standard solution were added and the UV/Vis spectra of the sam-
(SHELXL 97). Anisotropic displacement parameters were refined for
all non-hydrogen atoms. Hydrogens bonded to carbon atoms were placed
at calculated positions with the appropriate AFIX instructions and re-
fined by using a riding model; hydrogens bonded to the N(1) atom of 1-
ples were recorded. Spectrophotometric titration curves were fitted by
using the HYPERQUAD program.
out in [D ]DMSO at moderately high concentrations of 1-PF
[
11] 1
H NMR titrations were carried
and 2-PF
6
6
6
ꢀ
2
(
10 m).
PF
6
were located in the DF map and refined, constraining the NꢀH dis-
tance to be 0.89Æ0.02 Š.
X-ray crystallographic studies: Diffraction data were collected at ambient
temperature by using an Enraf–Nonius CAD4 four-circle diffractometer,
working with graphite-monochromatised MoKa radiation (l=0.71073 Š).
Crystal data for the 1-PF
CCDC-237092 and CCDC-237093 contain the supplementary crystallo-
graphic data for this paper. These data can be obtained free of charge via
www.ccdc.cam.ac.uk/conts/retrieving.html (or from the Cambridge Crys-
tallographic Data Centre, 12 Union Road, Cambridge CB21EZ, UK; fax:
+
6
complex and for the 1 zwitterion are reported
in Table 4. Data reductions (including intensity integration, background,
(
+44)1223-336-033; or deposit@ccdc.cam.ac.uk).
Table 4. Crystallographic data for 1-PF
-PF
6
and zwitterion 1.
1
6
zwitterion 1
formula
M
colour
C
19
H
19
F
6
N
2
O
2
PS
19 18 2 2
C H N O S
338.42
pale yellow
484.40
colourless
Acknowledgement
dimension [mm]
crystal system
space group
a [Š]
b [Š]
c [Š]
0.80”0.43”0.21
triclinic
P1 (no. 2)
0.80”0.28”0.11
orthorhombic
P2 (no. 4)
1
10.3645(16)
8.5655(12)
10.7356(14)
–
115.083(12)
–
863.2(2)
2
1.302
0.201
w-2q scans
2–26
2181
2084
0.0213
1685
0.0393, 0.0849
0.0594, 0.0941
1.029
217
0.22/ꢀ0.17
We thankDr. Laura Linati (Centro Grandi Strumenti, Università di
Pavia) for the NMR characterisation of the samples and Dr. Enrico Mon-
zani (Dipartimento di Chimica Generale, Università di Pavia) for assis-
tance with the NMR titrations and helpful discussions. The financial sup-
port of the European Union (RTN Contract HPRN-CT-2000–00029) and
of the Italian Ministry of University and Research (PRIN, Dispositivi Su-
pramolecolari; FIRB, Project RBNE019H9K) is gratefully acknowl-
edged.
¯
6.1779 (8)
12.8573(17)
13.3607(18)
83.866(12)
87.840(16)
82.756(22)
1046.5(2)
2
1.537
0.303
w-2q scans
2–26
5213
4104
0.0116
2673
0.0492, 0.1136
0.0879, 0.1323
1.040
320
0.22/ꢀ0.17
a [8]
b [8]
g [8]
V [Š ]
3
Z
[
[
[
[
1] P. A. Gale, Coord. Chem. Rev. 2003, 240, 1.
ꢀ
3
1
calcd [gcm
]
2] C. Suksai, T. Tuntulani, Chem. Soc. Rev. 2003, 32, 192.
3] R. Martinez-MaÇez, F. Sancenón, Chem. Rev. 2003, 103, 4419.
4] R. J. Fitzmaurice, G. M. Kyne, D. Douheret, J. D. Kilburn, J. Chem.
Soc. Perkin Trans. 1 2002, 841.
ꢀ
1
mMoKa [mm
scan type
]
q range [8]
measured reflections
unique reflections
[
[
5] K. Choi and A. D. Hamilton, Coord. Chem. Rev. 2003, 240, 101.
6] C. S. Wilcox in Frontiers in Supramolecular Chemistry and Photo-
chemistry, Wiley-VCH, Weinheim, 1991, pp. 123–143.
7] N. Dennis, A. R. Katritzky, H. Wilde, E. Gavuzzo, A. Vaciago, J.
Chem. Soc. Perkin Trans. 2 1977, 1304.
[
a]
R
int
strong data [I
o
>2s(I
o
)]
[
[
b]
R1, wR2 (strong data)
[
b]
R1, wR2 (all data)
[
[
8] F. Bordwell, Acc. Chem. Res. 1988, 21, 456.
9] S. Camiolo, P. A. Gale, M. B. Hursthouse, M. E. Light, A. I. Shi,
Chem. Commun. 2002, 758.
[
c]
GoF
refined parameters
max./min. residuals [eŠ
ꢀ
3
]
[
10] T. Steiner, Angew. Chem. 2002, 114, 50; Angew. Chem. Int. Ed. 2002,
41, 48.
2
o
2
2
o
2
o
[
{
[
a] Rint =SjF ꢀF (mean)j/SF . [b] R1=SjjF
o
2
jꢀjF
c
jj/SjF j, wR2=
o
2
o
2
c
2
2
1/2
2
o
2
S[w(F ꢀF ) ]/S[w(F ) ]} , in which w=1/[s F +(aP) +bP] and P=
[11] P. Gans, A. Sabatini, A. Vacca, Talanta 1996, 43, 1739.
[12] L. J. Farrugia, J. Appl. Crystallogr. 1999, 32, 837.
[13] A. C. T. North, D. C. Phillips, F. S. Mathews, Acta. Crystallogr. Sect.
A 1968, 24, 351.
o
2
2
c
2
o
2
c
2
1/2
max(F ,0)+2F ]/3. [c] GoF={S[w(F ꢀF ) ]/(nꢀp)} , in which n is the
o
number of reflections and p is the total number of refined parameters.
[
14] A. Altomare, M. C. Burla, M. Camalli, G. L. Cascarano, C. Giaco-
vazzo, A. Guagliardi, A. G. G. Moliterni, G. Polidori, R. Spagna, J.
Appl. Crystallogr. 1999, 32, 115.
15] G. M. Sheldrick, SHELX97 Programs for Crystal Structure Analysis,
University of Gçttingen, Germany, 1997.
Lorentz and polarisation corrections) were performed with the WinGX
[
12]
package.
Absorption effects were evaluated with the psi-scan
method, and absorption correction was applied to the data (min./max.
transmission factors were 0.865/0.938 for 1-PF and 0.931/0.970 for 1).
Crystal structures were solved by direct methods (SIR 97), and refined
[
13]
[
6
[
14]
Received: June 10, 2004
Published online: November 10, 2004
2
by full-matrix least-square procedures on
F
using all reflections
Chem. Eur. J. 2005, 11, 120 – 127
www.chemeurj.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
127