Enhancing the reactivity of an electrophilic barbiturate dye by cooperative hydrogen bonding

Article abstract of DOI:10.1002/ejoc.200901066

The supramolecular complex formation by cooperative hydrogen bonding of a highly dipolar barbiturate dye and 2,6-diacetamidopyridine is shown to have a marked effect on the reactivity of the barbiturate, The increase of both Lewis acidity and electrophilicity by hydrogen bonding is demonstrated by thermodynamic and kinetic studies using the electrophile-nucleophile recombination reaction with cyanide.

Full text of DOI:10.1002/ejoc.200901066

FULL PAPER
DOI: 10.1002/ejoc.200901066
Enhancing the Reactivity of an Electrophilic Barbiturate Dye by Cooperative
Hydrogen Bonding
Mirko Bauer[a] and Stefan Spange*[a]
Keywords: Donor-acceptor systems / Supramolecular chemistry / Hydrogen bonds / Lewis acids / Acidity / Electrophilicity
The supramolecular complex formation by cooperative
hydrogen bonding of a highly dipolar barbiturate dye and
2,6-diacetamidopyridine is shown to have a marked effect on
the reactivity of the barbiturate. The increase of both Lewis
acidity and electrophilicity by hydrogen bonding is demon-
strated by thermodynamic and kinetic studies using the elec-
trophile–nucleophile recombination reaction with cyanide.
towards a single preferred isomer by adding a suitable re-
ceptor has recently been described by our group[9] and
others.[10] These findings strongly suggest an impact of hy-
drogen bonding onto a molecule’s electron distribution. But
is cooperative hydrogen bonding based on the model of
DNA base pairing also capable of changing the reactivity
of an electronically coupled electrophile? To answer this
question, a simple electrophile–nucleophile recombination
may serve as a suitable model reaction. Such recombina-
tions are generally well-defined second-order reactions
whose kinetics have already been studied in detail.[1] In this
work we will demonstrate for the first time that it actually
is possible to influence the reactivity of a neutral electro-
phile by using an external complementary hydrogen bond-
ing receptor.
Introduction
For polar reactions the reactivity of a substrate can be
described using thermodynamic and kinetic values such as
equilibrium or rate constants. This reactivity depends on
various chemical parameters in a complex way. For in-
stance, solvent effects play a crucial role regarding the nu-
cleophilicity of anions.[1a] The electron-donating ability of
a substituent has a great impact on the electrophilicity of
carbenium ions and related compounds.[1b,1c] Now the
question arises whether a substituent containing a hydrogen
bonding motif suitable for supramolecular recognition can
be externally modified by complex formation so that an ef-
fect on the chemical reactivity of the substrate results.
Indeed, the alteration of a substrate’s reactivity by supra-
molecular interactions, especially hydrogen bonding, is a
common principle in nature, for instance in enzyme cataly-
sis. Numerous model systems have been developed for mim-
icking this behaviour, yet most of them use hydrogen bonds
for preorganising template effects only.[2,3] On the other
hand, the impact on the reactivity by the direct modifica-
tion of the electronic structure of a molecule by hydrogen
bonds has been studied less intensely and is mainly restric-
ted to solvation effects of protic solvents.[4,5] The few exam-
ples comprising definite hydrogen bonded complex forma-
tion include the increase of the acidity of 1,3-β-dicarbonyl
compounds by one pKa unit by addition of complementary
receptors.[6] Also, the redox properties of flavine model
compounds are changed upon hydrogen bond formation as
could be shown by Rotello[7] and Goldenberg.[8] The dis-
placement of a tautomeric or configurational equilibrium
A suitable system for our studies has been found in the
barbituric acid-containing merocyanine dyes BA1 and BA2
(Figure 1). The zwitterionic resonance structure illustrates
[a] Department of Polymer Chemistry, Chemnitz University of
Technology,
Strasse der Nationen 62, 09111 Chemnitz, Germany
Fax: +49-371-531-21239
E-mail: stefan.spange@chemie.tu-chemnitz.de
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejoc.200901066.
Figure 1. Electrophilic barbiturate dyes used in this study and the
complementary receptor DAC (A = hydrogen-bond acceptor, D =
hydrogen-bond donor).
Eur. J. Org. Chem. 2010, 259–264
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
259

M. Bauer, S. Spange
FULL PAPER
that the barbiturate moiety serves as an electron-with-
drawing substituent. Thus, the β-C-atom at the polarised
double bond may act as a Lewis acid and electrophile as is
already known from similar benzylidene barbiturates.[1c,2,11]
As a complementary hydrogen bonding receptor the well
established 2,6-diacetamidopyridine (DAC) was used.[7–9,12]
When BA1 or BA2 are treated with DAC and a nucleo-
phile at the same time, multiple equilibria and their mutual
influence have to be considered. In general, the involved
equilibrium and rate constants can be depicted in a thermo-
dynamic circle as is exemplified in Figure 2.
Scheme 1. Synthesis of compounds BA1 and BA2.
For both dyes the UV/Vis absorption maximum is shifted
towards lower energy with increasing dipolarity and polar-
izability of the solvent. This is a typical finding for dyes
with a charge-transfer transition where the ground state is
less polar than the excited state which proves that in the
ground state the neutral resonance form of BA1 and BA2
(Figure 1, left) outweighs the zwitterionic one (Figure 1,
right). Furthermore, a bathochromic shift of the UV/Vis
absorption maxima with an increasing hydrogen bond do-
nating (HBD) ability of the solvent is observed. These HBD
solvents interact with the barbiturate’s carbonyl moieties by
which its electron density is reduced and the push-pull sys-
tem is enhanced. In contrast, the solvent’s hydrogen bond
accepting capacity, which has the opposite effect and leads
to a hypsochromic shift, is of minor importance. Thus,
upon complexation of BA1 or BA2 with a complementary
hydrogen bond receptor a net decrease of electron density
and therefore an increase of reactivity can be expected.
Figure 2. Thermodynamic circle comprising the equilibrium and
rate constants of a mixture of BA1, DAC and a nucleophile Nu.
Results and Discussion
General Considerations
The barbiturate BA1 was obtained by a simple one-step
synthesis starting from triflic anhydride activated Michler’s
ketone (Scheme 1) which provides an easier access to this
type of compounds compared to the previously published
two-step procedure for the synthesis of BA2.[13]
Binding Studies
The addition of the hydrogen bond donating receptor
DAC to BA1 or BA2 indeed results in a bathochromic shift
A first impression of the interactions of these dyes with of the UV/Vis absorption band which is in line with the
their surrounding can be derived from solvatochromic stud- results of solvatochromic studies. By UV/Vis titration in
ies whereby specific and nonspecific interactions can be sep- dichloromethane the binding constants KDAC,1 were found
arated.[13,14] Hereto, the UV/Vis absorption maxima of BA1 to be 1778 –1 (DAC·BA1) and 1093 –1 (DAC·BA2),
and BA2 were measured in a set of 22–24 solvents and respectively. These values indicate moderately strong hydro-
evaluated by multiple regression analysis using the solvent gen bonds and are comparable to other threefold hydrogen
parameter set of Catalán[15] (see Supporting Information).
bonded complexes.[16] In the more polar solvent acetonitrile
260
www.eurjoc.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2010, 259–264

Enhancing the Reactivity of an Electrophilic Barbiturate Dye
this binding constant drops dramatically to 57 –1 for SI) it is conspicuous that the signal of C-7 is shifted from δ
DAC·BA1 due to competitive hydrogen bonding to solvent = 178 ppm (BA1) to δ = 50 ppm (BA1–CN) indicating a
molecules. As in less polar solvents like toluene the solubil- change from sp2 to sp3 hybridization for C-7. The signal of
ity is too low, dichloromethane was used for all other in- C-5 shows an upfield shift of ∆δ = 19 ppm to δ = 86 ppm
vestigations.
for BA1–CN which is a typical value for barbiturate anions.
To determine the reactivity of BA1 and BA2, they were At δ = 126 ppm a signal for covalently bound cyanide is
treated with a variety of potential nucleophiles, such as ali- observed. From these results it can be concluded that the
phatic or aromatic amines, halide and pseudohalide ions, expected nucleophilic attack actually takes place according
tetramethylthiourea or mercaptoacetic acid. With added to Scheme 2 whereby the chromophoric system is interrup-
amines a characteristic UV/Vis shift is observed which can ted. A quantitative evaluation of the UV/Vis experiments
be attributed to general solvent effects. Among the other according to a 1:1 stoichiometry yields the equilibrium con-
nucleophiles only fluoride and cyanide produced a signifi- stants KNu,1 and second-order rate constants k2,1 given in
cant, but distinct, change in the UV/Vis absorption.
Table 1. From Equation (1) the monomolecular dissociation
With fluoride a hypso- and hypochromic effect is ob- rate constant k–1,1 is accessible, too. While k–1 is hardly af-
served which is completely reversed in the presence of traces fected, KNu and k2 are lower for BA1 which can be readily
of water. Identical spectra are also obtained when hydroxide explained with the electron-releasing effect of its N-(n-
is added. This indicates an increase of electron density at butyl) group.
the barbiturate moiety and a weaker push-pull system. The
K = k2/k–1
(1)
probable explanation of this effect is a deprotonation of the
NH group (Scheme 2) which is also described for urea de-
1
rivatives.[17] This assumption is further supported by an H
Table 1. Equilibrium and rate constants of the reaction of BA1 and
BA2 with cyanide in dichloromethane.
NMR titration where the disappearance of the NH signal
and an upfield shift of all other signals is observed. It
should be mentioned that the effects in the UV/Vis and
NMR spectra are less pronounced for BA1 than for BA2 as
the latter compound may be attacked at two NH groups.
BA1
BA2
KNu,1 [M–1]
k2,1 [M–1 s–1]
k–1,1 [s–1]
319
1654
3.1ϫ10–2
9.7ϫ10–5
10.5ϫ10–2
6.4ϫ10–5
In the presence of both, DAC and cyanide, a similar
evaluation is only possible for BA1. For BA2 an appropriate
fitting of the binding isotherm could not be performed, so
for the anionic BA2-CN the formation of complexes with a
higher stoichiometry cannot be excluded. Yet, the recombi-
nation of BA1 with cyanide proved to be a suitable model
reaction for studying thermodynamic and kinetic effects of
hydrogen bonding.
Hydrogen Bonding and Reactivity
While KDAC,1 and KNu,1 might be determined directly
(see above), this is not possible for the remaining KDAC,2
and KNu,2. However, when solutions of BA1 containing in-
creasing amounts of DAC are titrated with cyanide the nor-
malised binding isotherms get steeper and the observed val-
ues of KNu,obs are rising (Figure 3). As the equilibration of
BA1 with DAC is much faster than with cyanide, these ob-
served values can be interpreted as a weighed average of
KNu,1 and KNu,2 [Equation (2)]. Notably, the ratio of the
absorbance at different wavelengths remains almost con-
stant throughout any titration. Hence the concentration ef-
fect of a higher KDAC,2 is negligible and the mol fraction x
of DAC complexed BA1 is assumed to be constant.
Scheme 2. Reaction of BA1 with tetra-n-butylammonium fluoride
or cyanide and the respective UV/Vis spectra series.
On the other hand, when cyanide is added the UV/Vis
absorption band decreases within several hours without de-
velopment of a new band in the visible range. In a 1H NMR
experiment a second set of signals (including the NH signal)
KNu,obs = (1 – x) KNu,1 + x KNu,2 with x = cDAC·BA1/cBA1,0
(2)
Accordingly, a plot of KNu,obs against x shows a linear
appears upfield-shifted which slowly gains intensity while relationship as can be seen in Figure 4. Extrapolation to x
the original signals diminish. In the 13C NMR spectra (see = 0 and x = 1 then yields the values for BA1 (KNu,1
=
Eur. J. Org. Chem. 2010, 259–264
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
261

M. Bauer, S. Spange
FULL PAPER
= 5.2ϫ10–2

–1 s–1 are obtained. These results show that
indeed the moderately strong hydrogen bonding between
BA1 and DAC is capable of reducing the electron density at
the distant β-C-atom of BA1 and of increasing measurably
its electrophilicity as well as its Lewis acidity.
Finally, the remaining binding constant can be derived
from the thermodynamic circle, KDAC,2 = 16035 –1 (see
Supporting Information). This increase is expected due to
the anionic nature of BA1–CN which is known to enhance
the strength of hydrogen bonds.[7,18] Using this binding con-
stant the mol fraction of DAC complexed BA1–CN
(xBA1–CN) of the above data points can be calculated. A plot
of the monomolecular electrophile-nucleophile dissociation
rate constant k–1,obs against xBA1–CN then yields again a
fairly linear correlation from which k–1,1 = 9.7ϫ10–5 s–1
and k–1,2 = 1.3ϫ10–5 s–1 is obtained (Figure 6).
Figure 3. Normalised binding isothermes of the reaction of BA1
with cyanide in the presence of increasing amounts of DAC.
381 –1) and DAC·BA1 (KNu,2 = 3436 –1), respectively. A
likewise treatment is also possible for the kinetic data (see
Figure 5) whereby values of k2,1 = 3.0ϫ10–2 –1 s–1 and k2,2

Figure 6. Plot of the observed first-order dissociation rate constant
k–1,obs against the mol fraction xBA1–CN of DAC·BA1–CN.
In Figure 7 the aforementioned results are summarised.
It can be seen that the hydrogen bond formation of BA1
with DAC influences the dissociation rate constant of BA1–
CN (k–1) more than the recombination rate constant (k2).
This can be attributed to the formation of stronger hydro-
gen bonds of the anionic BA1–CN which results in a larger
decrease in electron density at the electrophilic centre com-
pared to the neutral BA1. The equilibrium constant KNu
experiences the largest influence by hydrogen bonding as it
comprises the opposing effects of both k2 and k–1.
Figure 4. Plot of the observed binding constant KNu,obs against the
mol fraction x of DAC·BA1.
Figure 5. Plot of the observed second-order rate constant k2,obs
against the mol fraction x of DAC·BA1.
Figure 7. Compilation of the determined equilibrium and rate con-
stants within the mixture of BA1, DAC and cyanide.
262
www.eurjoc.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2010, 259–264

Enhancing the Reactivity of an Electrophilic Barbiturate Dye
3.12 (s, 12 H, 12-H), 3.78 (m, 2 H, 13-H), 6.66 (d, J = 8.9 Hz, 4
H, 10-H), 7.22 (d, J = 8.9 Hz, 4 H, 9-H), 7.86 (s, 1 H, 3-H) ppm.
13C NMR (63 MHz, CD2Cl2): δ = 13.7 (C-16), 20.3 (C-15), 30.4
(C-14), 40.0 (C-12), 40.5 (C-13), 105.0 (C-5), 110.5 (C-10), 128.9
(C-8), 136.8 (C-9), 151.2 (C-2), 154.4 (C-11), 162.4 (C-6), 163.5 (C-
4), 178.3 (C-7) ppm. C25H30N4O3 (434.54): calcd. C 69.10, H 6.96,
N 12.89; found C 68.70, H 7.10, N 12.45.
When comparing the effects of hydrogen bond complex
formation (BA1 vs. DAC·BA1, Figure 7) and N-alkyl substi-
tution (BA1 vs. BA2, Table 1) substantial differences can be
recognised. As stated above, hydrogen bond formation has
a greater impact on k–1 than on k2 because hydrogen bonds
in the anionic product are much stronger than in the neutral
substrate. On the other hand, the N-(n-butyl) group in BA1
affects mainly the recombination rate constant k2 while k–1
remains almost constant compared to BA2. This is due to
the weak electron-releasing effect of the alkyl group which
is effective in the unreacted substrate but is almost cancelled
by the negative charge in the cyanide adduct. As a result,
the increase in KNu by hydrogen bonding complex forma-
tion exceeds the decrease by N-alkyl substitution which is
remarkable regarding the noncovalent nature of hydrogen
bonds.
NMR Investigation of BA1–CN: A solution of BA1 in CD2Cl2
(0.04 molL–1) was treated with 1.9 equiv. of tetra-n-butylammo-
nium cyanide. After 6 h the solution was only slightly colored and
1
the H NMR spectrum showed only signals of the product BA1–
1
CN. H NMR (250 MHz, CD2Cl2): δ = 0.9–1.1 (m, 25.5 H, alkyl),
1.3–1.7 (m, 34 H, alkyl), 2.91 (s, 12 H, 12-H), 3.13 (m, 15 H, alkyl),
3.69 (m, 2 H, 13-H), 6.61 (d, J = 9.0 Hz, 4 H, 10-H), 6.86 (s, 1 H,
3-H), 7.18 (d, J = 9.0 Hz, 4 H, 9-H) ppm. 13C NMR (63 MHz,
CD2Cl2): δ = 13.9, 14.4, 19.7, 20.5, 30.9, 39.4 (alkyl), 40.6 (C-12),
50.3 (C-7), 58.8 (alkyl), 86.2 (C-5), 111.8 (C-10), 125.6 (cyano),
128.7 (C-8), 132.2 (C-9), 149.2 (C-11), 152.4 (C-2), 161.2 (C-6),
162.9 (C-4) ppm.
Supporting Information (see also the footnote on the first page of
this article): Solvatochromic studies of BA1 and BA2, 13C NMR
spectra of BA1 and BA1–CN, and detailed evaluation of the
Conclusions
The recombination reaction of the electrophilic barbitu-
rate BA1 with cyanide was studied in the presence of dif- UV/Vis experiments.
ferent amounts of the complementary hydrogen bond re-
ceptor DAC. Our investigations revealed that the hydrogen
bonding leads to a decrease of the electron density in BA1
which is not limited to the hydrogen bonding site but is also
Acknowledgments
transmitted to a distant electrophilic centre via a conju-
Financial support by the Deutsche Forschungsgemeinschaft
gated π-system. Photometric measurements showed that
both the rate constant and the equilibrium constant of this
electrophile-nucleophile recombination increase upon hy-
drogen bond formation and allowed the calculation of those
values for pure BA1 and the complex DAC·BA1.
(DFG) is gratefully acknowledged.
[1] a) S. Minegishi, R. Loos, S. Kobayashi, H. Mayr, J. Am. Chem.
Soc. 2005, 127, 2641–2649; b) H. Mayr, T. Bug, M. F. Gotta,
N. Hering, B. Irrgang, B. Janker, B. Kempf, R. Loos, A. R.
Ofial, G. Y. Remennikov, H. Schimmel, J. Am. Chem. Soc.
2001, 123, 9500–9512; c) F. Seeliger, S. T. A. Berger, G. Y. Re-
mennikov, K. Polborn, H. Mayr, J. Org. Chem. 2007, 72, 9170–
9180.
[2] a) R. Ahuja, P.-L. Caruso, D. Möbius, W. Paulus, H. Rings-
dorf, G. Wildburg, Angew. Chem. 1993, 105, 1082–1085; An-
gew. Chem. Int. Ed. Engl. 1993, 32, 1033–1036; b) T. M.
Bohanon, P.-L. Caruso, S. Denzinger, R. Fink, D. Möbius, W.
Paulus, J. A. Preece, H. Ringsdorf, D. Schollmeyer, Langmuir
1999, 15, 174–184.
Experimental Section
General: Dichloromethane was freshly distilled from calcium hy-
dride. Michler’s ketone was crystallised several times from ethanol.
Tetra-n-butylammonium cyanide was dried in vacuo at 50 °C.
Tetra-n-butylammonium fluoride (1  solution in THF) and tetra-
n-butylammonium hydroxide (40% w/w solution in methanol) were
used as received. The syntheses of BA2[13] and DAC[19] were pub-
lished previously.
[3] For a recent review see: L. J. Prins, P. Scrimin, Angew. Chem.
2009, 121, 2324–2343; Angew. Chem. Int. Ed. 2009, 48, 2288–
2306.
Details of the evaluation of the UV/Vis experiments are given in
the Supporting Information.
[4] N. J. Pienta, R. J. Kessler, J. Am. Chem. Soc. 1993, 115, 8330–
8339.
Synthesis of 1-(n-Butyl)-5-{bis[4-(dimethylamino)phenyl]methylene}-
barbituric Acid (BA1): A solution of Michler’s ketone (794 mg,
2.96 mmol) in dichloromethane (10 mL) was treated with triflic an-
hydride (0.50 mL, 2.96 mmol) and stirred for 10 min. After ad-
dition of 1-n-butylbarbituric acid (818 mg, 4.41 mmol) stirring con-
tinued for 1 h. Afterwards 1,8-diazabicyclo[5.4.0]undec-7-ene
(1.00 mL, 6.71 mmol) was added slowly, whereby a highly exother-
mic reaction occurred, and the mixture was refluxed for 3 h. The
solution was washed 10 times with water and the solvent evapo-
rated. The residue was subjected to column chromatography (silica
gel, eluent: dichloromethane). The obtained solid was further puri-
fied by washing with toluene and dissolving in dichloromethane
followed by extraction with water. Greenish shiny crystals (544 mg,
42%); m.p. 228–231 °C. 1H NMR (250 MHz, CD2Cl2): δ = 0.92 (t,
J = 7.3 Hz, 3 H, 16-H), 1.33 (m, 2 H, 15-H), 1.57 (m, 2 H, 14-H),
[5] X. Chen, J. I. Brauman, J. Am. Chem. Soc. 2008, 130, 15038.
[6] A. M. Kelly-Rowley, V. M. Lynch, E. V. Anslyn, J. Am. Chem.
Soc. 1995, 117, 3438–3447.
[7] a) E. Breinlinger, E. Niemz, V. M. Rotello, J. Am. Chem. Soc.
1995, 117, 5379–5380; b) M. Gray, A. O. Cuello, G. Cooke,
V. M. Rotello, J. Am. Chem. Soc. 2003, 125, 7882–7888.
[8] L. M. Goldenberg, O. Neilands, J. Electroanal. Chem. 1999,
463, 212.
[9] I. Bolz, S. Spange, Chem. Eur. J. 2008, 14, 9338–9346.
[10] V. Berl, I. Huc, J.-M. Lehn, A. DeCian, J. Fischer, Eur. J. Org.
Chem. 1999, 3089–3094.
[11] R. Bednar, O. E. Polansky, P. Wolschann, Z. Naturforsch., Teil
B 1975, 30, 582–586.
[12] L. Yu, H.-J. Schneider, Eur. J. Org. Chem. 1999, 1619–1625.
[13] M. Bauer, A. Rollberg, A. Barth, S. Spange, Eur. J. Org. Chem.
2008, 4475–4481.
Eur. J. Org. Chem. 2010, 259–264
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
263

M. Bauer, S. Spange
FULL PAPER
[14] C. Reichardt, Solvents and Solvent Effects in Organic Chemis-
try, 3rd ed., Wiley-VCH, Weinheim, 2003.
[15] J. Catalán, J. Phys. Chem. B 2009, 113, 5951–5960.
[16] J. R. Quinn, S. C. Zimmerman, J. E. Del Bene, I. Shavitt, J.
Am. Chem. Soc. 2007, 129, 934–941.
[17] D. Esteban-Gómez, L. Fabrizzi, M. Licchelli, J. Org. Chem.
2005, 70, 5717–5720.
[18] T. Chin, Z. Gao, I. Lelouche, Y. K. Shin, A. Purandare, S.
Knapp, S. S. Isied, J. Am. Chem. Soc. 1997, 119, 12849–12858.
[19] K. Kurita, R. L. Williams, J. Heterocycl. Chem. 1975, 12, 789–
791.
Received: September 18, 2009
Published Online: December 3, 2009
264
www.eurjoc.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2010, 259–264