Electrochemically controlled hydrogen bonding. O-quinones as simple redox-dependent receptors for arylureas

Article abstract of DOI:10.1021/jo000520c

9,10-Phenanthrenequinone and acenaphthenequinone are shown to act as simple redox-dependent receptors toward aromatic ureas in CH₂Cl₂ and DMF. Reduction of the o-quinones to their radical anions greatly increases the strength of hydr

Full text of DOI:10.1021/jo000520c

VOLUME 65, NUMBER 26
DECEMBER 29, 2000
©
Copyright 2000 by the American Chemical Society
Articles
Electr och em ica lly Con tr olled Hyd r ogen Bon d in g. o-Qu in on es a s
Sim p le Red ox-Dep en d en t Recep tor s for Ar ylu r ea s
Yu Ge, Laura Miller, Teresa Ouimet, and Diane K. Smith*
Department of Chemistry San Diego State University San Diego, California 92118-1030
dsmith@sciences.sdsu.edu
Received April 3, 2000
9
,10-Phenanthrenequinone and acenaphthenequinone are shown to act as simple redox-dependent
2 2
receptors toward aromatic ureas in CH Cl and DMF. Reduction of the o-quinones to their radical
anions greatly increases the strength of hydrogen bonding between the quinone carbonyl oxygens
and the urea N-hydrogens. This is detected by large positive shifts in the redox potential of the
quinones with no change in electrochemical reversibility upon addition of urea guests. Cyclic
voltammetric studies with a variety of possible guests show that the effect is quite selective. Only
guests with two strong hydrogen donors, such as O-H bonds or amide N-H bonds, that are capable
of simultaneously interacting with both carbonyl oxygens give large shifts in the redox potential of
the quinones. The electronic character and conformational preference of the guest are also shown
to significantly affect the magnitude of the observed potential shift. In the presence of strong proton
donors the electrochemistry of the quinone becomes irreversible indicating that proton transfer
has taken place. Experiments with compounds of different acidity show that the pK
protonated quinone radical is about 15 on the DMSO scale, >4 pK units smaller than that of
,3-diphenylurea. This is further proof that hydrogen bonding and not proton transfer is responsible
for the large potential shifts observed with this and similar guests.
a
of the
a
1
Hydrogen bonds are one of the most important types
of intermolecular interactions. Due to their strength and
directionality, they are ubiquitous in biological systems,
providing essential recognition, structural and control
elements needed to coordinate and run the complex
molecular machinery required for life. In recent years
there has been tremendous interest in utilizing hydrogen
bonds for similar purposes in synthetic systems. They
elements in the creation of large molecular assemblies
in solution3 and “engineered” solid-state materials. In
contrast, hydrogen bonds have not been used extensively
as control elements in synthetic systems, even though
they are often used as such in biological ones. For
example, in flavin-containing redox proteins, hydrogen
bonds are used to modulate the redox potential of the
,4
5,6
have been used extensively as recognition elements in
(2) For some recent examples, see: (a) Inouye, M.; Takahashi, K.;
Nakazumi, H. J . Am. Chem. Soc. 1999, 121, 341-345. (b) Schmuck,
C. Chem. Commun. 1999, 843-844. (c) Bell, T. W.; Hou, Z. Angew.
Chem., Int. Ed. Engl. 1997, 36, 1536-1538.
(3) For reviews see: (a) Conn, M. M.; Rebek, J .Chem. Rev. 1997,
97, 1647-1668. (b) Whitesides, G. M.; Simanek, E. E.; Mathias, J . P.;
Seto, C. T.; Chin, D. N.; Mammen, M.; Gordon, D. M. Acc. Chem. Res.
1995, 28, 37-44.
the design of synthetic receptors1,2 and as structural
(
1) For reviews see: (a) Webb, T. H.; Wilcox, C. S. Chem. Soc. Rev.
1
2
993, 383-395. (b) Rebek, J . Angew. Chem., Intl. Ed. Engl. 1990, 29,
45-255. (c) Hamilton, A. D. J . Chem. Educ. 1990, 67, 821-828. (d)
Schmidtchen, F. P.; Berger, M. Chem. Rev. 1997, 97, 1609-1646.
1
0.1021/jo000520c CCC: $19.00 © 2000 American Chemical Society
Published on Web 12/06/2000

8
832 J . Org. Chem., Vol. 65, No. 26, 2000
Ge et al.
flavins by stabilizing one oxidation state relative to
another. This has been demonstrated by Rotello and co-
workers in several model systems.7
reason, substrates with appropriately positioned NH
groups hydrogen bond with the carbonyl oxygens much
stronger in the reduced states. Examples of good sub-
strates are urea derivatives with PQ11 and the diamido-
pyridine derivative 1 with NI. The increase in binding
strength is detected by large positive shifts in the redox
potential of PQ and NI in the presence of the substrates.
In effect, the substrates stabilize the radical anions
through hydrogen bonding, making it easier to reduce
the quinone or diimide in their presence.
Previously, we have showed how hydrogen bonds
coupled with electrochemistry can create a powerful
8
control element in synthetic receptor-substrate systems.
The concept, which is also being actively explored by
9
10
Rotello and others, is straightforward. Hydrogen bonds
are known to have substantial electrostatic character.
Therefore, a reduction or oxidation process that leads to
a change in partial charge on one of the components in a
hydrogen bond will have a significant effect on the
strength of that hydrogen bond. In particular, if the
negative charge on the hydrogen acceptor or the positive
charge on the hydrogen donor is increased, the strength
of the hydrogen bond will be increased. Alternatively, if
the negative charge on the hydrogen acceptor or the
positive charge on the hydrogen donor is decreased, the
strength of the hydrogen bond will be decreased.
Positive shifts in the redox potential of quinones in the
presence of hydrogen bond donors have been studied
previously.12 The difference here is the large magnitude
of the observed shifts at very low concentrations of
hydrogen bond donor. The maximum potential shift is
related to the ratio of binding constants in the oxidized
1
3
(Kox) and reduced states (Kred) as given in eq 3. On the
basis of this, the binding strength of PQ with the oxidized
8
In our earlier work, we showed that substantial effects
can be produced even in very simple systems. Specifically
we looked at the behavior of 9,10-phenanthrenequinone,
PQ, and 1,8-naphthalimide, NI. Both PQ and NI undergo
reversible one-electron reductions in aprotic media to
form radical anions, eqs 1 and 2. Although the radical is
diphenylurea increases by >2000 times upon reduction
2 2
of PQ to its radical anion in CH Cl , and the binding
strength of NI to 1 increases by >100 times upon
reduction.
In this paper, we report the results of our full study
on the redox-dependent binding properties of the ortho-
quinone/radical anion redox couple. In addition to PQ,
we studied acenaphthenequinone, ANQ, which also un-
dergoes a reversible, one-electron reduction in aprotic
media. A large variety of substrates were studied with
both quinones in order to better understand the nature
of the binding interaction. We also explored in detail the
question of hydrogen bonding vs proton transfer in these
systems.
delocalized, most of the negative charge resides on the
oxygens due to their greater electronegativity. For this
(4) For some recent examples, see: (a) Gong, B.; Yan, Y.; Zeng, E.;
Skrzypczak-J ankunn, E.; Kim, Y. W.; Zhu, J .; Ickes, H. J . Am. Chem.
Soc. 1999, 121, 5607-5608. (b) Corbin, P. S.; Zimmerman, S. C. J .
Am. Chem. Soc. 1998, 120, 9710-9711. (c) Martin, T.; Obst, U.; Rebek,
J . Science 1998, 281, 1842-1845. (d) Clark, T. D.; Buriak, J . M.;
Kobayashi, K.; Isler, M. P.; McRee, D. E.; Ghadiri, M. R. J . Am. Chem.
Soc. 1998, 120, 8949-8962. (e) Sakai, N.; Majumbar, N.; Matile, S. J .
Am. Chem. Soc. 1999, 121, 4294-4295.
Exp er im en ta l Section
Syn th etic P r oced u r es. Elementary analyses were per-
formed by Desert Analytics, Tucson, AZ. Et
freshly distilled from CaH . 1-Phenyl-3-(4-methoxyphenyl)-
urea,14 1,3-di-(4-methoxyphenyl)urea, 1-phenyl-3-(4-trifluoro-
2
O was dried and
2
15
(
5) For reviews see: (a) Aakeroy, C. B.; Seddon, K. B. Chem. Soc.
16
17
methylphenyl)urea, 1,3-di-(4-trifluoromethylphenyl)urea,
-phenyl-3-(2,5-di-tert-butylphenyl)urea, 1-phenyl-3-propyl-
Rev. 1993, 397-407. (b) MacDonald, J . C.; Whitesides, G. M. Chem.
1
Rev. 1994, 94, 2383-2420. (c) Zaworotko, M. J . Chem Soc. Rev. 1994,
2
83-288. (d) Bishop, R. Chem Soc. Rev. 1996, 311-319.
(
6) For some recent examples, see: (a) Holy, P.; Zavazka, J .;
(11) Ureas are known to strongly hydrogen bond with oxygens that
Cisarova, I.; Podlaha, J . Angew. Chem., Int. Ed. Engl. 1999, 38, 381-
carry partial negative charges, so they are often used as key compo-
nents in receptors for oxyanions. See, for example: (a) Hughes, M. P.;
Shang, M.; Smith, B. D. J . Org. Chem. 1996, 61, 4510-4511. (b)
Nishizawa, S.; Buhlmann, P.; Iwao, M.; Umezawa, Y. Tetrahedron Lett.
1995, 36, 6483-6486.
3
83. (b) Brunet, P.; Simara, M.; Wuest, J . D. J . Am. Chem. Soc. 1997,
19, 2737-2738. (c) Evans, C. C.; Sukarto, L.; Ward, M.. D. J . Am.
1
Chem. Soc. 1999, 121, 320-325. (d) Aoyama, Y.; Endu, K.; Anzai, T.;
Yamaguchi, Y.; Sawaki, T.; Kobayashi, K.; Kanehisa, N.; Hashimoto,
H. J . Am. Chem. Soc. 1996, 118, 5562-5571.
(12) See Gupta, N.; Linschitz, H. J . Am. Chem. Soc. 1997, 11, 6384-
(7) (a) Breinlinger, E.; Niemz, A.; Rotello, V. M. J . Am. Chem. Soc.
6391, and references therein.
1
995, 117, 5379-5380. (b) Breinlinger, E.; Rotello, V. M. J . Am. Chem.
Soc. 1997, 119, 1165-1166.
8) (a) Ge, Y.; Lilienthal, R. R.; Smith D. K. J . Am. Chem. Soc. 1996,
(13) Miller, S. R.; Gustowski, D. A.; Chen, Z.-H.; Gokel, G. W.;
Echegoyen, L.; Kaifer, A. E. Anal. Chem. 1988, 60, 2021-2024.
(14) Miyahara, M. Chem. Pharm. Bull. 1986, 34, 1950-1960.
(15) Miyahara, M.; Kamiya, S.; Nakadate, M., Chem. Pharm. Bull.
1983, 31, 41-44.
(
1
1
18, 3976-3977. (b) Ge, Y.; Smith D. K. Anal. Chem. 2000, 72, 1860-
865..
(9) (a) Deans, R.; Niemz, E.; Breinlinger, E.; Rotello, V. M. J . Am.
(16) Routaboul, J .-M.; Mougin, C.; Ravanel, P.; Tissut, M.; MrLina,
G.; Calmon, J .-P. Phytochemistry 1991, 30, 733-738.
(17) Lin, H. C.; Cotter, B. R.; Bieron, J . F.; Krishnamurti, R. J .
Fluorine Chem. 1991, 52, 107-116.
Chem. Soc. 1997, 119, 10863-10864. (b) Niemz, A.; Rotello, V. M. Acc.
Chem. Res. 1999, 32, 44-52.
(10) Carr, J . D.; Lambert, L.; Hursthouse, M. B.; Malik, K. M. A.;
Tucker, J . H. R. Chem. Commun. 1997, 1649-1650.
(18) Izdebski, J .; Pawlak, D. Synthesis 1989, 423.

Electrochemically Controlled Hydrogen Bonding
J . Org. Chem., Vol. 65, No. 26, 2000 8833
urea,18 and 1-phenyl-3-dibutylurea19 were prepared from the
appropriate isocyanates and amines using the general proce-
dure described below. Other solvents and chemicals were of
reagent grade and used as supplied from the manufacturers.
Gen er a l P r oced u r e for th e P r ep a r a tion of Ur ea s. A
stirred solution of the amine (10 mmol) in ethyl ether (10 mL)
was cooled to 0 °C. The isocyanate (10 mmol) in ethyl ether
(
10 mL) was added dropwise over 15 min. The solution was
then allowed to warm to room temperature and stirred for
-24 h. The solvent was removed in vacuo and the residue
1
recrystallized from ethanol to yield a crystalline solid. All
compounds, except 1-phenyl-3-(2,5-di-tert-butylphenyl)urea,
are known, and analytical data are in accordance with
1
4-19
literature values.
1
-P h en yl-3-(2,5-d i-ter t-bu tylp h en yl)u r ea . 2,5-Di-tert-bu-
tylaniline and phenyl isocyanate were used as starting materi-
als. The reaction was finished in 2 h with 90% yield: mp
1
F igu r e 1. Cyclic voltammograms of 1 mM ANQ in CH Cl in
the presence of increasing amounts of 1,3-diphenylurea. Scan
2
32.0-232.5 °C; H NMR 1.321 (s, 9 H), 1.326 (s, 9 H), 7.007
2
2
1
3
(tt, J ) 9.6, 1.6 Hz, 1 H), 7.21-7.46 (m, 8 H); C NMR 31.74,
rate ) 100 mV/s. [1,3-Diphenylurea] (mM) ) (a) 0, (b) 0.5, (c)
3
1
3
2.28, 35.85, 36.55, 121.30, 124.69, 125.61, 128.37, 128.57,
30.08, 135.75, 139.20, 144.75, 151.82, 155.65; IR (CHCl
1
, (d) 2, (e) 5, (f) 10.
3
+
)
320, 2965, 1644, 1600, 1557, 1505,1235; MS m/z 325 (M )
O: C, 77.73; H, 8.70; N, 8.63. Found:
Anal. Calcd for C21
H
28
N
2
and alpha ) 0.5. The E , D, and R values were then used to
Q
u
C, 77.80; H, 9.00; N, 8.60.
fit the CVs in the presence of urea by adjusting Kox and Kred
In str u m en ta l P r oced u r es. In str u m en ta tion a n d Re-
a gen ts. Cyclic voltammetry (CV) experiments were performed
with a PAR Model 263 digital potentiostat using the Model
to give the best fit to the experimental CVs at different values
-
6
of k
f
. All species were assumed to have equal D, ∼5 × 10
2
5
cm /s. Values of k
f
e 10 M/s resulted in much poorer fits to
2
70 electrochemistry software package. The acquisition mode
was set to “ramp” in order to simulate an analogue experiment.
CH Cl was freshly distilled from P and filtered through a
the experimental voltammograms. In the final simulations k
f
values of 106 were used for both the oxidized and reduced
quinones.
2
2
2 5
O
column of activated alumina immediately before use. Anhy-
drous grade N,N-dimethylformamide (DMF) was purchased
from Aldrich and also filtered through an activated alumina
column immediately before use. NBu
as the electrolyte. All measurements were conducted at 22 °C
under N in a jacketed, one-compartment cell with a Au disk
4 6
PF (0.1 M) was used
2
working electrode (2 mm diameter) and a Pt wire counter
electrode. A Ag wire was used as a pseudo-reference in most
experiments. Ferrocene, N,N,N′,N′-tetramethylphenylenedi-
amine and cobalticinium hexafluorophosphate were used as
internal references. For the CVs used for computer simulation,
a Ag wire in a fritted tube containing 10 mM AgNO
3
in 0.1 M
Resu lts a n d Discu ssion
Bu NPF /DMF was placed in a separate compartment and
4
6
used as a reference. The cell for these experiments was covered
with a black cloth to prevent photodegradation of the quinones.
Gen er a l P r oced u r e for Mea su r em en t of ∆E1/2 Va lu es.
Electrolyte and solvent were added to a cell with working,
Volta m m etr y of Qu in on es in th e P r esen ce of 1,3-
Dip h en ylu r ea . Figure 1 shows cyclic voltammograms
CVs) of acenaphthenequinone (ANQ) in the presence of
(
different amounts of 1,3-diphenylurea in CH Cl . The
2
2
reference and counter electrodes protected under N
2
. The
solvent was degassed by purging with N for a few minutes.
2
CVs of ANQ are similar to the previously reported CVs
of phenanthrenequinone (PQ) in the presence of 1,3-
After taking a background scan, the quinone and internal
reference were added and a CV taken. Then the guest was
added to the cell and another CV taken. The difference in half-
wave potentials of the quinone relative to the internal refer-
ence before and after adding the guest, ∆E1/2, was then
calculated.
CV Sim u la tion . DigiSim v2.1 software (BioAnalytical
Systems) was used to simulate the CVs and calculate the
binding constants of PQ and ANQ with 1,3-diphenylurea.
Background subtracted CVs of the quinones in the presence
of 0, 0.5, 1.0, 2.0, 5.0, 10, 20 and 50 mM diphenylurea were fit
to the following “square-scheme”, eq 4. The formal redox
8a
diphenylurea. CV (a) in Figure 1 (dashed line) is ANQ
by itself. Addition of 0.5 equiv of 1,3-diphenylurea, CV
b), causes noticeable broadening of the anodic peak, and,
(
in the presence of 1 equiv of 1,3-diphenylurea, CV (c),
the half-wave potential (E1/2) is shifted significantly
positive of the original. The wave continues to shift
positive as more 1,3-diphenylurea is added until at 10
mM, which is close to the solubility limit, the wave is
1
70 mV positive of the original, CV (f).
The behavior observed in Figure 1 is consistent with
potential of the quinone, E
quinone, D, and the uncompensated resistance in the cell, R
Q
, the diffusion coefficient of the
2
0
strong binding of the urea to the ANQ radical anion.
However, attempts to simulate the CVs of PQ and ANQ
in the presence of different amounts of 1,3-diphenylurea
u
,
were determined by fitting the CVs with no added urea at scan
rates between 200 and 1000 mV/s, assuming Butler-Volmer
kinetics with an electron-transfer rate constant, k , ) 10 cm/s
s
in CH
mental data. We believe this is due to self-aggregation
of the urea in CH Cl2. This hypothesis is supported by a
concentration dependence study of 1,3-diphenylurea in
2 2
Cl resulted in unsatisfactory fits to the experi-
(
19) Van Landeghem, H.; de Aguirre, I. Bull. Soc. Chim. Fr. 1967,
72-178.
20) Since quinones are known to have very fast electron-transfer
rate constant in aprotic solvents (see, for example, Howell, J . O.;
Wightman, R. M. Anal.Chem. 1984, 56, 524-529), ∆E ’s > 60 mV at
these relatively slow scan rates were assumed to be entirely due to
the R in the cell. R values of 800-900 Ω were typically determined
in this fashion.
2
1
(
1
2 2
CD Cl using HNMR. The peak for the urea N-hydro-
p
gens moved upfield by 0.19 ppm when the urea solution
was diluted from 1 mM to 20 µM. It then moved
downfield by 0.03 ppm upon dilution from 20 µM to 10
u
u

8
834 J . Org. Chem., Vol. 65, No. 26, 2000
Ge et al.
background subtraction.21 However the peak potentials
in the simulated CVs fit the experimental CVs nicely.
The simulated CV’s shown in Figures 2 and 3 cor-
-
1
-1
respond to Kox ) 1 M and Kred ) 905 M for PQ and
-
1
-1
22
K
ox ) 1 M and Kred ) 715 M for ANQ. It should be
noted that it is not possible to accurately determine Kox
from the simulations since values ranging from 0.01 to
-
1
1
M
produce very little change in the voltammograms.
Despite this, the values for Kred appear reliable since
varying Kox over this range only causes Kred to vary
-
1
-1
-1
between 893 M and 905 M for PQ and 700 M and
7
-
1
15 M for ANQ. Values of Kox > 1 lead to larger values
for Kred, but the quality of the fit deteriorates.
The lack of any significant interaction between the
neutral quinones and diphenylurea in DMF is confirmed
by H NMR. Adding up to 30 mM PQ to a 1mM solution
F igu r e 2. Experimental (solid lines) and simulated (dashed
lines) cyclic voltammograms of 1 mM PQ in DMF in the
presence of increasing amounts of 1,3-diphenylurea. Scan rate
1
7
of diphenylurea in DMF-d produced a less than 0.01 ppm
)
200 mV/s. [1,3-Diphenylurea] (mM) ) (a) 0, (b) 2, (c) 10, (d)
change in the chemical shift of the urea protons.
The above studies indicate that reduction acts es-
sentially as an on/off switch for o-quinone/urea binding
in DMF. Binding is “off” in the oxidized state and “on”
in the reduced state.
5
0. The simulated CVs correspond to Kox ) 1 and Kred ) 905
-
1
M
.
The simulations also suggest that the PQ radical anion
binds the urea slightly better than the ANQ radical
anion.23 This appears general since consistently larger
E
1/2 shifts are observed with PQ for a variety of ureas.
(See Tables 2 - 4.) The reason for this difference may be
structural, electrostatic, or both. Semiempirical molecular
orbital calculations (AM1) indicate that the O-O distance
in the PQ radical anion (2.754 D) is smaller than in the
ANQ radical anion (2.951 D). This makes the O-O
distance in PQ more comparable to the urea H-H
distance in diphenylurea (2.633 D), allowing at least the
possibility of forming two more linear hydrogen bonds
with PQ than ANQ. In addition, a slightly larger per-
centage of the excess negative charge appears localized
F igu r e 3. Experimental (solid lines) and simulated (dashed
lines) cyclic voltammograms of 1 mM ANQ in DMF in the
presence of increasing amounts of 1,3-diphenylurea. Scan rate
(21) This was a greater problem with ANQ because of its more
negative redox potential. Increases in the background current due to
increasing moisture and other changes during long experiments result
in an actual background that is greater than the original background
taken at the beginning of the experiment, particularly at the negative
limit. As a result the background subtraction is incomplete and causes
the experimental currents to be larger than the simulated ones on the
forward scan and smaller on the reverse.
)
200 mV/s. [1,3-Diphenylurea] (mM) ) (a) 0, (b) 1, (c) 5, (d)
2
0. The simulated CVs correspond to Kox ) 1 and Kred ) 715
-
1
M
.
µM. Further dilution resulted in the chemical shift
moving upfield again. This behavior indicates that the
(
22) In our earlier paper (ref 8a), we reported a value of Kred ) 660
M- for PQ under the same conditions. There are several reasons for
the different value reported here. The first and most significant
involves a mistake that was made with the diffusion coefficients in
the original simulations. Specifically, while the diffusion coefficients
1
urea does aggregate in CD
is complicated.
2 2
Cl and the aggregation mode
Since self-aggregation competes with binding to the
quinone radical anion, successful fitting of the experi-
-6
2
of PQ, PQ- and the urea were set at 5.45 × 10 cm /s, the value
determined for PQ in this solvent system, the diffusion coefficients of
-
the PQ-urea complexes, PQ-urea and PQ-urea , were mistakenly left
mental CVs in CH
gregation into the mechanism. In contrast, H NMR data
of 1,3-diphenylurea in DMF-d shows no concentration
2 2
Cl would require incorporating ag-
-5
2
as 1 × 10 cm /s (the program’s default value). Since it is usually the
solvent system that has the greatest effect on diffusion coefficients,
the standard assumption is that the diffusion coefficients of all the
species are the same and this more reasonable assumption was made
in the simulations reported here. In addition to the above, a number
of other changes were made which we believe make the simulation
more realistic, although they have less effect on the value of Kred.
First, in our original simulations we only used three equilibria, Q +
1
7
dependence over the same concentration range, which
indicates urea self-aggregation does not occur in DMF.
This simplifies the binding, and, as a result, we are able
to fit the experimental CVs reasonably well in DMF.
Figure 2 shows CVs of PQ in DMF in the presence of
different amounts of 1,3-diphenylurea (solid lines) over-
lapped with the corresponding simulated CVs (dotted
lines). Similar CVs for ANQ are shown in Figure 3. For
both quinones, the E1/2 shift in DMF is significantly
-
-
-
e- a Q , Q + urea a Q-urea, and Q + urea a Q-urea , to fit the
data rather than the complete square scheme (eq 4) used here. Second,
we assumed everything was at equilibrium and set all the rate
s
constants very large. The electron-transfer rate constant, k , was set
4
at 1 × 10 cm/s and the forward rate constants for complex formation,
9
k , were set at 1 × 10
f
s
M/s. In these simulations we set k to the more
f
reasonable value of 10 cm/s and we explored the effect of changing k .
6
This lead us to choosing a value of 1 × 10 M/s for both k
f
’s.
smaller than in CH
is expected since DMF should solvate the hydrogen
bonding sites much more effectively than CH Cl . With
2
Cl
2
, indicating weaker binding. This
-1
(
23) The value of Kred for ANQ reported here (715 M ) is actually
larger than the original value (ref 8a) reported for Kred for PQ (660
-1
M
). This is because of differences in the simulation model as
2
2
described in footnote 22. It should be noted that if the ANQ and PQ
data are fitted to the same model, whether it is the old or new one,
PQ the simulated CVs fit the observed ones very well.
The fits are not as good with ANQ due to problems with
the ANQ data always gives a smaller value for Kred
.

Electrochemically Controlled Hydrogen Bonding
J . Org. Chem., Vol. 65, No. 26, 2000 8835
Ta ble 1. Sh ift in Ha lf-Wa ve P oten tia l, ∆E1/2, for
Ta ble 2. Sh ift in Ha lf-Wa ve P oten tia ls, ∆E1/2, for
P h en a n th r en equ in on e, P Q, a n d Acen a p h th en equ in on e,
ANQ, in th e P r esen ce of Va r iou s NH- a n d OH-Con ta in in g
Su bstr a tesa
Differ en t Dica r bon yl Com p ou n d s in th e P r esen ce of 5
a
m M Dip h en ylu r ea
/1-
∆
E1/20
(mV)
DMF
∆
E1/2 P Q0
/1-
∆E1/2 ANQ0/1-
dicarbonyl (1.0 mM)
CH2Cl2
(mV)
(mV)
phenanthrenequinone, PQ
131 ( 13
141 ( 9
61 ( 2
58 ( 3
40 ( 6
8 ( 7
substrate (50 mM)
CH2Cl2
DMF
1
-2
2 ( 5
2 ( 2
CH2Cl2
DMF
1
,2-naphthoquinone, 2
acenaphthenequinone, ANQ
anthraquinone, 3
benzil, 4
117 ( 4
benzophenone
1,4-dicyanobenzene
water
ethanol
ethylene glycol
propylamine
,2-diaminobenzene
,3-diaminobenzene
propionamide
benzamide
formanilide
butylurea
,3-dipropylurea
-phenyl-3-propylurea
,3-diphenylurea
6
11
11
12
10 ( 9
11 ( 5
3
1
49 ( 5
irreversible
5 ( 17
15 ( 13
14 ( 7
111 ( 4
0 ( 1
18 ( 14
2 ( 18
20 ( 9
31 ( 11
60 ( 5
5 ( 4
2 ( 9
a
The reported ∆E1/2 values are the average of at least three
independent measurements. The ranges correspond to the 95%
confidence limits.
11 ( 2 109 ( 6 14 ( 6
1 ( 3
7 ( 2
9 ( 8
5 ( 7
7
2 ( 1
6 ( 5
3 ( 3
2 ( 1
3 ( 4
1 ( 8
5
1
1
on the O’s in the PQ radical anion than in the ANQ
radical anion (32.8% in PQ- vs 30.4% in ANQ ).
For comparison, the voltammetry of another o-quinone,
,2-naphthoquinone, 2, and two other related dicarbonyl
compounds, anthraquinone, 3, and benzil, 4, were stud-
ied. Like the o-quinones, anthraquinone and benzil can
20 ( 11
25 ( 2
-
4
22
61 ( 5 16
80 ( 12 20 ( 9
51 ( 4 28 ( 10 39 ( 11 10 ( 7
156 ( 11 66 ( 17 144 ( 4 54 ( 12
131 ( 13 107 ( 28 117 ( 4 90 ( 5
63 ( 10 14 ( 9
1
1
1
1
(5 mM in CH2Cl2)
1
1
-phenyl-3-(2,5-di-tert-
butylphenyl)urea
52 ( 6
21 ( 5
41 ( 3 11 ( 7
(5 mM in CH2Cl2)
,1-dibutyl-3-phenylurea
4 ( 5
5 ( 2
5 ( 9
5 ( 5
a
∆E1/2 values reported with ranges are the average of at least
three independent measurements. The ranges correspond to the
5% confidence limits.
9
and benzil. Anthraquinone does gives a significant, albeit
much smaller, positive shift in CH Cl2, presumably due
2
to the hydrogen bonding motif depicted in structure B.
However, no significant shift is observed in DMF. It is
not possible to evaluate the redox-dependent binding
be reversibly reduced to radical anions in aprotic media
resulting in increased negative charge on the oxygens.
With the o-quinones one urea can simultaneously hydro-
gen bond with two oxygens as shown in structure A
below. However, with anthraquinone one urea can only
hydrogen bond to one oxygen at a time as shown in
structure B. This is likely to be the case with benzil as
well, since rotation about the central C-C bond will be
hindered in the radical anion due to partial double bond
formation and the lower energy structure will have the
oxygens trans due both to electrostatic repulsion and
steric effects.
2 2
behavior of benzil in CH Cl since the CV becomes
irreversible upon the addition of 1,3-diphenylurea. How-
ever, it also gives no significant E1/2 shift in DMF.
The much larger E1/2 shifts observed for the o-quinones
compared to anthraquinone and benzil support the
hypothesis that both carbonyl oxygens in the o-quinones
are hydrogen-bonded to the urea. Anthraquinone and
benzil give smaller shifts because the urea is only able
to interact with one carbonyl at a time.
Volta m m etr y of o-Qu in on es in th e P r esen ce of
Oth er Su bstr a tes. Table 2 lists the shifts in E1/2
observed for PQ and ANQ upon addition of 50 mM of
2 2
various compounds in CH Cl and DMF. These data
shows that despite their simple structure PQ and ANQ
are rather selective redox-dependent receptors. Most
compounds, even those that contain hydrogen bonding
sites, do not produce significant shifts in the E1/2 of the
quinone/radical anion redox couple.
As shown in Table 2, addition of electron-poor aromatic
substrates that do not have N-H or O-H bonds (1,4-
dicyanobenzene and benzophenone) result in small, posi-
The results with benzil, anthraquinone, and o-naphtho-
quinone are summarized along with PQ and ANQ in
Table 1, which lists the change in E1/2 of the neutral/
radical anion redox couples upon addition of 5 mM
diphenylurea. The data show that 1,2-naphthoquinone
behaves almost identically to PQ. This is not surprising
given the structural similarities between the two and it
offers further evidence that the angles of the carbonyls
in this type of o-quinone provide a somewhat better fit
to urea than the angles in ANQ. However, despite these
slight differences, all the o-quinones give E1/2 shifts that
are considerably larger than those with anthraquinone
2 2
tive shifts in the E1/2’s of the quinones in CH Cl , but not
in DMF. However, these shifts are much, much smaller
than those observed with the ureas. This supports our
assumption that the interaction with diphenylurea is
primarily due to hydrogen bonding and not due to a
donor-acceptor or other type of π-π interaction between
the aromatic systems of the relatively electron-rich
radical anion and the somewhat electron-poor urea.
The data in Table 2 also show that the presence of a
N-H or O-H bond in the substrate is not sufficient to
produce a significant ∆E1/2. (For the purpose of this
discussion we define “significant” as a shift of at least

8
836 J . Org. Chem., Vol. 65, No. 26, 2000
Ge et al.
1
0 mV after the error limits are taken into consideration.)
Ta ble 3. Sh ift in Ha lf-Wa ve P oten tia ls, ∆E1/2, of
P h en a n th r en equ in on e, P Q, a n d Acen a p h th en equ in on e,
ANQ, in th e P r esen ce of Ur ea s w ith Differ en t
Con for m a tion sa
Of those substrates that contain an O-H bond, water,
ethanol, and ethylene glycol, only the latter produces a
significant shift. Unlike the others, ethylene glycol has
two O-hydrogens that could simultaneously hydrogen
bond to both quinone carbonyl oxygens.
A greater variety of compounds containing N-H bonds
were examined. With these it is obvious that the type of
bond also plays a role. None of the amino compounds
produced a significant shift, including 1,2-diaminoben-
zene which should be able to hydrogen bond with both
carbonyls. (Ethylenediamine was also tried as a substrate
but it could not be studied due to rapid reaction with the
quinones.)
∆
E1/2 P Q0
/1-
∆E1/2 ANQ0/1-
(mV)
(mV)
substrate (5.0 mM)
CH2Cl2 DMF CH2Cl2 DMF
1-phenyl-3-(2-pyridyl)urea, 5 42 ( 6 16 ( 8 18 ( 10 6 ( 1
1
1
,3-diphenylurea
131 ( 13 61 ( 2 117 ( 4 40 ( 6
-phenyl-3-(4-pyridyl)urea, 6 141 ( 6 85 ( 4 130 ( 5 60 ( 5
a
The reported ∆E1/2 values are the average of at least three
independent measurements. The ranges correspond to the 95%
confidence limits.
gives significantly smaller shifts than both 6 and 1,3-
diphenylurea. These results provide further evidence that
the structure of the o-quinone-urea complex is as depicted
in eq 1 with the urea in the trans-trans conformation and
both N-hydrogens simultaneously bonded to the carbonyl
O’s.
The data in Table 3 also show that the phenylpyridyl-
urea 6 binds more strongly than diphenylurea itself. This
can be attributed to an electronic effect. The greater
electronegativity of nitrogen compared to carbon makes
a pyridyl group more electron-withdrawing than a phenyl
group. This causes the urea N-hydrogen to be more
positive in 6 than diphenylurea. (A similar effect is seen
Amides are better hydrogen donors than amines due
to the electron-withdrawing nature of the carbonyl. As a
result, the simple amides propionamide and benzamide
2 2
do produce small but significant shifts in CH Cl . Amides
formed from arylamines should be even better hydrogen
donors, and indeed, formanilide, which has only one N-H
bond, gives much larger shifts than the other amides. In
fact the magnitude of the shift is equivalent to that
observed with the alkylureas, butylurea and dipropyl-
urea. The N-hydrogen in these compounds should be less
positively charged than the N-hydrogen in formanilide,
but there are two of them making it possible to form two
hydrogen bonds with the o-quinones. Not surprisingly,
even larger shifts are observed when one of the urea
nitrogens has an aryl substituent as in phenylpropylurea
and the largest shift is observed when both nitrogens
have aryl substituents as in diphenylurea. This com-
pound has two highly polarized N-H bonds capable of
forming two strong hydrogen bonds with the o-quinones.
The importance of forming two strong hydrogen bonds
is supported by the behavior of 1,1-dibutyl-3-phenylurea.
This arylurea only has one N-hydrogen and gives insig-
nificant potential shifts. However, part of the reason for
the small shifts with this urea is that there is likely to
be a considerable amount of steric hindrance at the
binding site.
in the pK
-aminopyridine and 4-aminopyridine have pK
DMSO of 27.6 and 26.5, respectively, considerably smaller
a
’s of related compoundssthe N-hydrogen of
2
a
’s in
2
5
a
than the pK of aniline (30.5). )
This electronic effect was explored further by testing
diphenylurea derivatives containing electron-donating
methoxy groups and electron-withdrawing trifluoro-
methyl groups. Table 4 lists the shifts in E1/2 observed
for PQ and ANQ upon addition of 5 mM of these
2 2
substrates in CH Cl and DMF. As expected, the mag-
nitude of the shift steadily increases going from the urea
with the most electron-donating substituents,1,3-di(4-
methoxyl)phenylurea, to the urea with the most electron-
withdrawing substituents,1,3-di(4-trifluoromethyl)phen-
ylurea. The former would be expected to have the least
positively charged N-hydrogens and the latter the most
positively charged.
P r oton Tr a n sfer vs Hyd r ogen Bon d in g. In an
attempt to find even stronger bonding substrates for the
o-quinones, diphenylthiourea was also tried as a guest
with PQ. As with the oxyureas, a large positive shift in
the cathodic peak potential of PQ was observed, but,
unlike the oxyureas, this was accompanied by substantial
changes in wave shape.26 Most notably the cathodic peak
current doubled in size, the original anodic peak disap-
peared and another anodic peak appeared at more
positive potentials. Such behavior is strongly indicative
of proton transfer. Indeed, very similar behavior is
observed when 1,3-diphenyl-1,3-propanedione is added
to PQ, Figure 4. This compound should not hydrogen
bond to the urea, but the methylene protons have an
acidity comparable to diphenylthiourea.
A better argument supporting the importance of both
urea N-hydrogens being able to hydrogen bond to the
o-quinones can be made by considering the behavior of
the two pyridylureas 5 and 6. The electronic character
of both compounds should be very similar. However, with
5
, a strong intramolecular hydrogen bond will be formed
between the 2-pyridyl nitrogen and one of the urea
N-hydrogens.24 This means that only one hydrogen will
be available for intermolecular hydrogen bonding. In
contrast, with 6, both N-hydrogens will be available for
intermolecular hydrogen bonding.
The behavior seen in Figure 4, has also been observed
upon addition of acids to other quinones in aprotic
solvents.27 It can be explained by an ECE mechanism,
Table 3 lists the observed ∆E1/2 values for these two
pyridyl ureas along with 1,3-diphenylurea. Compound 5
(
24) (a) Singha, N. C.; Sathyanarayana, D. N., J . Chem. Soc., Perkin
Trans. 2 1997, 157-162. (b) Sudha, L. V.; Sathyanarayana, D. N. J .
Mol. Struct. 1985, 131, 253-259. (c) Sudha, L. V.; Sathyanarayana,
D. N. J . Mol. Struct. 1984, 125, 89-96. (d) Camilleri, P.; Odell, B.;
Rzepa, H. S.; Sheppard, R. N. Chem. Commun. 1988, 1132-1133.
(25) Bordwell, F. G. Acc. Chem. Res. 1988, 21, 456-463.
(26) Similar electrochemical behavior was recently reported for the
addition of thioureas to ubiquinone in CH
2 2
Cl . Greaves, M. D.; Niemz,
A.; Rotello, V. M. J . Am. Chem. Soc. 1999, 121, 266-267.

Electrochemically Controlled Hydrogen Bonding
J . Org. Chem., Vol. 65, No. 26, 2000 8837
Ta ble 4. Sh ift in Ha lf-Wa ve P oten tia ls, ∆E1/2, of P h en a n th r en equ in on e, P Q, a n d Acen a p h th en equ in on e, ANQ, in th e
P r esen ce of Differ en t Su bstitu ted Ur ea sa
E1/2 P Q0
/1-
(mV)
DMF
∆E1/2 ANQ0/1- (mV)
∆
substrate (5.0 mM)
CH2Cl2
CH2Cl2
DMF
1
1
1
1
1
,3-di-(4-methoxy)phenylurea
-phenyl-3-(4-methoxy)phenylurea
,3-diphenylurea
-phenyl-3-(p-trifluoromethyl)phenylurea
,3-di-(4-trifluoromethyl)phenylurea
90 ( 7
121 ( 17
131 ( 13
171 ( 9
193 ( 4
50 ( 2
57 ( 6
68 ( 4
105 ( 2
116 ( 4
136 ( 13
169 ( 4
30 ( 7
35 ( 4
40 ( 6
56 ( 4
68 ( 9
61 ( 2
84 ( 8
100 ( 12
a
The reported ∆E1/2 values are the average of at least three independent measurements. The ranges correspond to the 95% confidence
limits.
Ta ble 5. Rever sibility of th e CV of P Q a n d ANQ in DMF a n d CH2Cl2 in th e P r esen ce of Com p ou n d s w ith Differ en t p Ka
Va lu esa
DMF
CH2Cl2
PQ0
/1-
ANQ0/1-
PQ0/1-
ANQ0/1-
pK
substrate
,3-diphenylurea
diethyl malonate
dimethyl methylmalonate
ethyl acetoacetate
(DMSO)
reversible?
reversible?
reversible?
reversible?
1
19.55
16.4
15.05
14.4
yes
yes
yes
no
yes
yes
yes
no
yes
yes
yes
no
yes
yes
yes
no
1
,3- diphenyl-1,3-propanedione
13.35
11.0
no
no
no
no
no
no
no
no
benzoic acid
a
CV conditions: 1 mM quinone + 10 mM acid in 0.1 M NBu4PF6 in either DMF or CH2Cl2; 100 mV/s scan rate. b pKa values in DMSO,
from ref 25.
radical PQH. PQH is easier to reduce than the original
quinone, so at the potential of PQ reduction, PQH is
-
immediately reduced to PQH . Thus the cathodic peak
current doubles and the original anodic peak disappears
-
since PQ is no longer present. The irreversible oxidation
peak at more positive potentials has been ascribed to the
-
27
oxidation of the final anion, PQH in this case.
2 2
F igu r e 4. Cyclic voltammograms of 1 mM PQ in CH Cl in
the presence of increasing amounts of 1,3-diphenyl-1,3-pro-
panedione. Scan rate ) 100 mV/s. [1,3-Diphenyl-1,3-pro-
panedione] (mM) ) (a) 0, (b) 0.4, (c) 1, (d) 2, (e) 5.
2 2
Addition of 1,3-diphenyl-1,3-dione to ANQ in CH Cl
also causes the quinone reduction to become irreversible,
but the voltammetry differs substantially from that
observed for PQ. As shown in Figure 5, instead of the
original cathodic peak doubling in size, a new, broad,
irreversible reduction peak is observed at more negative
potentials. This is accompanied by a decrease in the
original anodic peak, but no new anodic peak appears.
The behavior of ANQ also suggests proton transfer is
occurring, but it is accompanied by a different mech-
anism than in the case of PQ. This is not that sur-
prising given the structural differences. In the case of
-
F igu r e 5. Cyclic voltammograms of 1 mM ANQ in CH
2
Cl
2
in
PQ, addition of the second electron creates the 14 e
the presence of increasing amounts of 1,3-diphenyl-1,3-pro-
panedione. Scan rate ) 100 mV/s. [1,3-Diphenyl-1,3-pro-
panedione] (mM) ) (a) 0, (b) 0.4, (c) 1, (d) 2, (e) 5.
aromatic phenanthrene ring system, but adding a second
electron to ANQ only gives a 12 e system, so the new
double bond is not aromatic. The resulting bond length
changes may cause significant reorganizational energy
-
eq 5. In the absence of proton donors, PQ is reversibly
-
reduced to the radical anion, PQ . In the presence of a
(27) Eggins, B. R.; Chamber, J . Q., J . Electrochem. Soc. 1970, 117,
-
strong enough acid, PQ is protonated to form the neutral
186-191.

8
838 J . Org. Chem., Vol. 65, No. 26, 2000
Ge et al.
leading to slow electron transfer kinetics. It is also
possible that the protonated radical anion dimerizes and
the second electron addition is due to reduction of the
dimer product. Further investigation would be necessary
to clarify the mechanism. The main point at present is
that protonation clearly occurs and leads to an irrevers-
ible CV.
anions is observed even in this highly competitive solvent.
Even stronger binding is observed in CH Cl , a solvent
2 2
more comparable to the solvents normally used for the
study of hydrogen bonding receptors.
The present work leaves little doubt that the strong
bonding between the ureas and quinone radical anions
is due to the formation of two strong, partially ionic
hydrogen bonds between the urea NH’s and the quinone
carbonyl O’s. To see such strong binding it is necessary
to have the two carbonyls aligned in the same direction
as well as two strong hydrogen donors on the guest that
can also align in the same direction.
Beyond the specific example of the o-quinone-urea
complex, this study, along with the work of Rotello and
others, illustrates the powerful role electrochemistry can
play in creating and characterizing strongly hydrogen
bonded complexes. Reversibly reducible carbonyl com-
pounds such as quinones and imides are not the only
examples of redox-dependent hydrogen bonding receptors
and in future papers we will report examples of other
simple compounds that also can be converted electro-
chemically into strong hydrogen bonding receptors.
By taking CVs of the quinones in the presence of proton
sources with different pK
pK of the protonated radical anion. Table 5 gives the
results of this study. The pK values for these compounds
in DMSO25 are used since not all the values in DMF and
CH Cl are available. It is assumed that the relative
acidity scale will be the same in DMF and CH Cl even
though the actual pK ’s will be different. Based on these
results the protonated radical for both quinones appears
to have a pK of about 15 on the DMSO scale. It is
noteworthy that this is >4 pK units smaller than that
a
’s, it is possible to estimate the
a
a
2
2
2
2
a
a
a
of the ureas. Therefore, the positive redox potential shifts
we observe are indeed the result of hydrogen bonding and
not due to proton transfer.
Con clu sion s
We have shown that two simple o-quinones, phenan-
threnequinone and acenaphthenequinone, act as quite
selective redox-dependent receptors toward aromatic
ureas in aprotic solvents. Whereas essentially no interac-
tion is observed between diphenylurea and the neutral
quinones in DMF, strong binding to the quinone radical
Ack n ow led gm en t. We would like to thank the
donors of the Petroleum Research Fund administered
by the American Chemical Society and the National
Science Foundation for partial support of this work.
J O000520C