Electrochemically driven release of picomole amounts of calcium ions with temporal and spatial resolution

Article abstract of DOI:10.1002/anie.200705274

(Figure Presented) Controlled delivery: Redox commutation of a paraphenylenediamine and aza-crown ether assembly anchored as a self-assembled monolayer to an electrode surface allows the time and spatially controlled release of picomole amounts of calcium ions. The release occurs on a submillisecond time scale, and is irreversible provided that the two-electron-oxidized redox center is not returned electrochemically to its neutral initial state (see scheme).

Full text of DOI:10.1002/anie.200705274

Angewandte
Chemie
DOI: 10.1002/anie.200705274
Electrochemistry
Electrochemically Driven Release of Picomole Amounts of Calcium
Ions with Temporal and Spatial Resolution**
Christian Amatore,* Damiano Genovese, Emmanuel Maisonhaute,* Noureddine Raouafi, and
Bernd Schöllhorn*
In memory of Patrick Allenet
The ability to trigger the delivery or capture of a nanoscale
amount of active molecules or ions within a fast (namely,
millisecond) controlled time scale and with a micrometric or
sub-micrometric spatial resolution is a current challenge.
Several cellular functions are indeed controlled by such
nanoquantities of molecular and ionic effectors, and their
investigation requires the possibility of monitoring their
fluxes externally.[1,2] Similarly, the increasing awareness in
nanosciences of the necessity of delivering chemicals at the
nanoscale level will also demand that new ways are found to
perform such operations. Such ability will concern many
research areas as well as advanced industrial microfabrication
processes.
In this respect, electrochemically commanded delivery
ought to offer an interesting solution as a result of the
lightness of electrochemical equipment and its easy imple-
mentation. Recent studies by Mirkin and co-workers estab-
lished that electrical control of electrocapillary forces at a
polarized liquid/liquid interface allows attoliters of any
desired solution to be released following an electrochemical
command.[3,4] Despite its seminal value, this method requires
the simultaneous flow of a solvent which contains and carries
the species to be delivered. Another option consists of finding
methods for delivering only the species of concern.
could not otherwise be dissolved. The availability of such
specific ligands has led to the design of analytical sensors in
which a redox center is connected to the ligand. Whenever the
standard potential of the coupled redox moiety is modified
upon capture of the ion of choice by the ligand, the change in
electrochemical behavior demonstrates the presence of the
target ion and may even provide information about its
concentration.[6–9]
It occurred to us that this strategy could be reversed and
used for the fast and precise delivery of picomole amounts of
an ion with a spatial scale defined by the dimensions of an
electrode bearing a specific ligand attached to it and coupled
with an appropriate redox center. Indeed, the electrostatic
interaction, which is generally monitored through its action
on the potential of the redox reporter upon chelation of the
ionic target, could be directed so that a change in the charge
of the redox center provoked by its oxidation or reduction
may affect sufficiently the complexation ability of the ligand,
thus promoting release of a chelated ion.
The validity of the concept has already been demon-
strated by using photochemical activation.[10] In that case, the
excited state populated on absorption of light involves an
intramolecular charge transfer that enforces a drastic change
in binding ability. However, deactivation of the excited state
of the molecule results in recomplexation occurring at a near-
diffusion-limited rate. Since the life time of excited states is
generally a few nanoseconds, the ion released initially may
only move a few nanometers and be captured again when the
molecule returns to its ground state. In contrast, electro-
chemical activation may induce a permanent change (at least
while the electrochemical driving force is applied from the
electrode) provided that the redox center is chemically stable.
Hence, one may consider that the electrochemical commuta-
tion of the complexation ability of a ligand has a similar
efficiency as that of photochemical processes, but also has the
important advantage of maintaining the commutated status as
demanded by the experiment to be carried out. Furthermore,
the use of microelectrodes allows the space in which release
occurs to be perfectly defined through tailoring the diffusion
layer by careful choice of the dimensions of the electrode.[11]
Herein we validate this concept for the direct release of
picomole amounts of ions through an electrochemical com-
mand. In doing so, we used a system in which the ligand/redox
center assembly is anchored as a self-assembled monolayer
(SAM) to the electrode surface, since ultimately this is
certainly the option which offers the largest synthetic
versatility, the most precise control of released quantities,
Following the discovery of crown ethers, a great body of
research has been, and is currently, focused on the design of
ligands for the selective complexation of ions,[5] with the
purpose of either sequestrating them from a solution or
allowing their dissolution into apolar solvents, in which they
[*] Dr. C. Amatore, D. Genovese, Prof. E. Maisonhaute, Dr. N. Raouafi,
Dr. B. Schöllhorn
UMR CNRS 8640 “PASTEUR” and LIA CNRS XiamENS “Nano-
BioChem”, DØpartement de Chimie, Ecole Normale SupØrieure,
UniversitØ Pierre et Marie Curie-Paris 6, 24 rue Lhomond, 75231
Paris Cedex 05 (France)
Fax: (+33)1-4432-3863
E-mail: christian.amatore@ens.fr
emmanuel.maisonhaute@ens.fr
bernd.schollhorn@ens.fr
[**] This work was supported by the CNRS (UMR 8640, cooperation
CNRS-DGRST, and the LIA XiamENS), the Ecole Normale SupØr-
ieure, the UniversitØ Pierre et Marie Curie, and the French ministry
of research through ANR REEL. We gratefully thank Monique Martin
and Pascal Plaza for helpful discussions, and also Alain Fuxa and
Guillaume Bertrand for some preliminary experiments. D.G. thanks
the University of Bologna for a travel grant.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2008, 47, 5211 –5214
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5211

Communications
and the fastest commutation times (since no significant time
constant is involved). Our aim in this study was only to
validate this concept, so we selected conditions which allowed
the best physicochemical characterization of the system
properties. For this reason, we used a microelectrode with
sufficiently large dimensions, rather than an ultra-micro-
electrode, to enable its surface status to be controlled with
adequate precision and reproducibility. Similarly, since most
crown ethers and similar ligands have been designed to
perform in non-aqueous solvents, we carried out this vali-
dation study in acetonitrile with Ca2+ used as a model ion to
be released. We do not view these choices as potential
limitations for future applications, such as those envisioned
above. Indeed, ultra-microelectrodes with nanometric dimen-
sions may be readily manufactured[12] and their surfaces can
be modified by a considerable variety of SAMs or other
specific chemicals.[13]
Figure 1. Square scheme illustrating the mechanistic path (gray lines)
followed by ligands 1 and 2 in the presence of Ca2+ ions during the
oxidation (forward scan) and the reduction (reverse scan; see
Figures 2 and 3).
We thus focused the present study on molecules 1 and 2
(Scheme 1). These compounds were synthesized from 3
according to the procedures described in the Supporting
Information and initially published by Pearson et al.[14–18]
Figure 2. Cyclic voltammograms of 1 (2 mm) in acetonitrile/0.1m
tetrabutylammonium tetrafluoroborate solution in the absence
(dashed line) and presence (solid line) of 1.1 equivalents of Ca2+ ions.
Scan rate: 1 Vsꢀ1; glassy carbon electrode (diameter: 1 mm). The
potential scale refers to the saturated calomel electrode.
wave Red4 occurs nearly at the same potential as Red2, which
indicates that the dication remains almost uncoordinated.
Wave Red5 is more anodic than wave Red1, thus showing that
its electrochemical process is facilitated. This finding shows
that only when the cation radical is reduced to the neutral
ligand form, namely, at Red5, can a fast back-complexation of
Ca2+ ions take place, which drives the reduction process (EC
system,[19] Figure 1). This proposal was confirmed since
inverting the reverse scan potential after scanning over
wave Red2 (data not shown) resulted in a wave very close
to that of the Ox2 wave, which shows that the monocation was
formed and not significantly coordinated on the voltammetric
timescale. Since wave Ox3 evolves after Ox2, two electrons are
transferred at the same potential (ECE mechanism).[19,21]
This study of the parent compound 1 kinetics and
mechanism in solution clearly demonstrates that efficient
release of a Ca2+ ion occurs upon oxidation of the [Ca(1)]2+
complex. However, the use of a system in solution is certainly
not adequate for applications such as those envisioned above.
Anchoring the [Ca(1)]2+ complex onto the electrode surface
would meet such requirements, and the dimensions of the
electrode would define those of the space in which release
occurs through precise definition of its diffusion layer.[11,22]
A common strategy for such purpose consists of exploiting
the high affinity of sulfur for gold to produce self-assembled
monolayers (SAMs). We chose to adapt a recent technique
based on the formation of dithiocarbamate SAMs to anchor
our system (2),[23–26] since this should allow synthetic flexibility
Scheme 1. Phenylenediamine–aza-crown ether derivatives.
In compounds 1–3, the nitrogen atom of the aza-crown
ether ligand also belongs to the paraphenylenediamine redox-
active center. Upon oxidation, the coordinating ability of this
electron-donating nitrogen atom is expected to be strongly
reduced because of the partial positive charge delocalized on
it. This charge should induce strong coulombic repulsive
forces on any complexed cation, since it is created only a few
angstroms away. Our goal was then to optimize the expected
square-scheme mechanism (Figure 1), which ultimately con-
trols the efficiency of the release of the calcium ions.
We used cyclic voltammetry to fully characterize the
dynamics of the system.[19] The electrochemical behavior of a
solution of 1 in acetonitrile containing 0.1m tetrabutylammo-
nium tetrafluoroborate is very similar to that of the well-
characterized tetramethylparaphenylenediamine.[20] Two
reversible systems Ox1/Red1 and Ox2/Red2 are observed,
which define the two standard potentials E01 and E20 of the
cation-free molecule (Figure 2).
Upon addition of one equivalent of Ca2+ ions, both peaks
disappear, and an irreversible two-electron peak Ox3 is
observed at much higher potential. The shift from the first
oxidation peak is 875 mV, which illustrates that 1 has a very
high affinity constant for the complexation of Ca2+ ions. On
the reverse scan, provided that the Ox3 wave was scanned
over, two reduction waves Red4 and Red5 are observed. The
5212
www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 5211 –5214

Angewandte
Chemie
in the future. Dithiocarbamates have been used only recently
for the preparation of SAMs. Their affinity to gold is equal or
even higher than for structurally comparable thiols. Reported
procedures imply that prior to adsorption onto the gold
suface, a dithiocarbamate is formed in the presence of carbon
disulfide and a strong base, such as sodium hydroxide, in
water or ethanol, which is accompanied by a partial precip-
itation of the corresponding dithiocarbamate salt.
This experimental protocol was slightly modified and a
solution of 3 and carbon disulfide in chloroform was prepared
without additional base. We did not observe the formation of
any precipitate. In our case, since the aliphatic amine is a
weak base and the dithiocarbamic acid is thermodynamically
unstable,[27] it is likely that the chemisorption on the gold
electrode is sufficient to promote the conversion of the acid
into the dithiocarbamate, with the formation of the sulfur–
gold bond being the driving force of the reaction. Avoiding
the generation of a precipitate allows easier control of the
system during the formation of the SAMs and in particular for
diluting the active ligand on the gold surface with inert thiols
(see below).[28,29]
ing, and a broader and larger irreversible band (Ox3*)
developing at a much higher potential.
Two reduction waves (Red4* and Red5*) were observed
on the reverse voltammogram upon scanning over wave Ox3*.
Reduction Red4* was close to the former wave Red2*
observed for the free SAM, while Red5* developed before
Red1*. This observation is similar to that described and
rationalized above for 1 in solution, and shows that confine-
ment of the system within a SAM assembly does not alter the
overall process depicted in Scheme 1. This was again evident
while scanning over wave Ox3*, that is, in less than 1ms. The
quantity of released calcium ions is given by the half of the
charge corresponding to the oxidation at Ox3* (two-electron
process versus one calcium ion per center), namely, 2.3 ”
10ꢀ12 mol for the system in Figure 3. Since the electrode
surface area is 4.8 mm2, this value corresponds to the release
of 4.7 ” 10ꢀ11 molcmꢀ2 ions per unit of surface area. Such a
value is compatible with the expected SAM coverage of the
electrode surface and with the coverage of 6.5 ”
10ꢀ11 molcmꢀ2 deduced by integration of the voltammetric
signal in the absence of calcium ions (waves Ox1* or Ox2* in
Figure 3). The difference of 27% stems from the fact that the
anodic potential scan in the presence of calcium ions was
interrupted before the end of the wave Ox3* to avoid
damaging the SAM. This value shows also that the amount
of calcium ions to be released may be adjusted broadly by
using ultra-microelectrodes (which have much smaller surface
areas than the microelectrode used in the present study) and
finely by the potential excursion range and the dilution of the
active SAM component with inert ones.[31]
Monolayers of 2 formed spontaneously on gold electro-
des. After rinsing the electrodes with chloroform and
acetonitrile, the electrochemical properties were evaluated
in acetonitrile (Figure 3). They were very similar to those
reported above for 1 in solution, but the shape of both
voltammetric waves was characteristic of surface-confined
redox molecules. Hence, they strongly depend on the
procedure used to construct the monolayer. This behavior is
typical for SAMs because of lateral interactions between
electroactive molecules.[19,30] By competitive adsorption of n-
hexanethiol, which dilutes the amount of active molecule in
the SAM, such lateral interactions may be minimized, and the
coverage may be adjusted at will between zero and maximum
coverage.[28,29,31]
Furthermore, although this is not the property that we
aimed for in this study, this result establishes that our system
also behaves as an extremely good electroanalytical sensor by
drastically modifying its electrochemical properties upon
complexation.[6–9]
The gradual addition of Ca2+ ions to the solution results in
the two reversible oxidation waves of the free ligand (Ox1*/
Red1* and Ox2*/Red2* in Figure 3) progressively disappear-
In conclusion, the validity of the principle of electro-
chemically driven release of picomole amounts of a specific
ion stored in a complexing self-assembled monolayer has
been established. The results demonstrate that release
occurred on a sub-millisecond time scale, and was irreversible
provided that the two-electron-oxidized redox center was not
returned electrochemically to its neutral initial state. Further
work will involve a full kinetic characterization of the process
through the use of ultrafast cyclic voltammetry.[22] The system
may be incorporated on ultra-microelectrodes of various
sizes, and this will allow the nanometric shape and dimensions
of the space into which ions are delivered to be selected.[12]
Finally, the present syntheses have been developed for a
system in non-aqueous media. Our future goal concerns the
ability of provoking concentration jumps of cation effectors
near a living cell,[1,2] to extend our current studies on vesicular
release of neurotransmitters and of oxidative stress by
triggering the response of single cells with precise temporal
and spatial resolution.[1,2,32–34] This is the fundamental reason
for our choice of evaluating the performance of the system
with calcium ions. Experiments not reported here established
that using aqueous electrolytes is not a problem for the
stability of the redox center. However, the complexing ability
of the present ligand in its the neutral state is clearly not high
Figure 3. a) Schematic representation of the mixed self-assembled
monolayer formed by 2 upon dipping a gold electrode (area: 4.8 mm2)
in chloroform containing CS2 (0.124 mm), 3 (0.24 mm), and hexane-
thiol (0.71 mm) as a dilutant. b) Cyclic voltammograms of the mixed
monolayer in (a) in acetonitrile/0.1m tetraethylammonium tetrafluor-
oborate in the absence (dashed line) and in the presence (solid line)
of Ca2+ ions (0.92 mm). Scan rate: 50 Vsꢀ1. The surface coverage of 2
is 6.5”10ꢀ11 molcmꢀ2. Note, that in the presence of Ca2+ ions the
anodic scan was reversed before the electrolysis at wave Ox3* was
complete.
Angew. Chem. Int. Ed. 2008, 47, 5211 –5214
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5213

Communications
enough for impeding spontaneous leaking of Ca2+ ions into a
calcium-free aqueous medium. Therefore, fulfilling our
ultimate goal requires the synthesis of redox-center-coupled
ligands with better complexing ability.[6–9]
Figure 3, when the system is immobilized and diluted on the
electrode surface. Except for this peculiar observation, and the
consequences of confinement on the wave shapes, all voltam-
metric features of the two calcium-bound systems are in exact
correspondence (Ox3$Ox3*, Red4$Red4*, and Red5$Red5*)
in terms of relative magnitudes and potential locations. This
finding shows that the small magnitude of wave Ox3 (hence,
those of Red4 and Red5) reflects a drastic reduction (by a factor
ca. 2.6) of the diffusion coefficient D of the electroactive system
in solution when a calcium ion is bound. This result suggests a
change in diffusion coefficient, which is in full agreement with
the fact that this behavior is not observed in the adsorbed phase
where no molecular diffusion is involved. We have no concrete
rationalization for a reduction of the solution diffusion coef-
ficient by ca. 2.6, since this would correspond to a drastic change
in the size of individual molecules which cannot be accounted for
by any realistic change in the conformation upon calcium
binding. However, it is possible that calcium binding in free
solution involves some calcium-mediated molecular clustering
so as to produce a complex aggregate [Ca(1)]2+n. Such aggre-
gates, because of their size, would experience a drastic decrease
in their diffusion coefficient(s) Dn compared to that (D0) of the
unbound 1 or of an isolated [Ca(1)]2+ complex. However,
oxidation would lead to the correct electron stoichiometry (since
2nF/mol of aggregate is equivalent to 2F/mol of each [Ca(1)]2+
moiety) although the current peak height of wave Ox3* (hence,
those of Red4* and Red5*) would decrease as (Dn/D0)1/2.
[22] C. Amatore, E. Maisonhaute, Anal. Chem. 2005, 77, 303 A.
[23] Y. Zhao, W. Perez-Segarra, Q. Shi, A. Wei, J. Am. Chem. Soc.
2005, 127, 7328.
[24] J. M. Wessels, H. G. Nothofer, W. E. Ford, F. von Wrochen, F.
Scholz, T. Vessmeyer, T. Schroedter, H. Weller, A. Yasuda, J.
Am. Chem. Soc. 2004, 126, 3349.
[25] M. Lieder, Electrochim. Acta 2004, 49, 1813.
[26] E. Almirall, A. Fragoso, R. Cao, Electrochem. Commun. 1999, 1,
10.
[27] P. Morf, F. Raimondi, H.-G. Nothofer, B. Schnyder, A. Yasuda,
J. M. Wessels, T. A. Jung, Langmuir 2006, 22, 658.
[28] C. Amatore, E. Maisonhaute, B. Schꢀllhorn, J. Wadhawan,
ChemPhysChem 2007, 8, 1321.
[29] C. Amatore, S. Gazard, E. Maisonhaute, C. Pebay, B. Schꢀllhorn,
J.-L. Syssa-Magalꢁ, J. Wadhawan, Eur. J. Inorg. Chem. 2007,
4035.
[30] C. P. Smith, H. S. White, Anal. Chem. 1992, 64, 2398.
[31] H. O. Finklea in EncyclopØdia of Electrochemistry, Vol. 10 (Eds.:
M. Fujihira, I. Rubinstein, J. F. Rusling), Wiley-VCH, Weinheim,
2005.
Received: November 16, 2007
Revised: March 10, 2008
Published online: June 2, 2008
Keywords: calcium · crown compounds · electrochemistry ·
.
monolayers · self-assembly
[1] Y. M. Tang, E. R. Travis, R. M. Wightman, J. Neurochem. 2000,
74, 702.
[2] C. L. Haynes, L. A. Buhler, R. M. Wightman, Biophys. Chem.
2006, 123, 20.
[3] F. O. Laforge, J. Carpino, S. A. Rotenberg, M. V. Mirkin, Proc.
Natl. Acad. Sci. USA 2007, 104, 11895.
[4] K. T. Rodolfa, A. Bruckbauer, D. Zhou, A. I. Shevchuk, Y. E.
Korchev, D. Klenerman, Nano Lett. 2006, 6, 252.
[5] C. J. Pedersen, J. Am. Chem. Soc. 1967, 89, 7017.
[6] L. Zhang, C. M. Cardona, L. Echegoyen, Chem. Commun. 2006,
4461.
[7] A. E. Kaifer, Acc. Chem. Res. 1999, 32, 62.
[8] G. W. Gokel, W. M. Leevy, M. E. Weber, Chem. Rev. 2004, 104,
2723.
[9] P. D. Beer, P. A. Gale, G. Z. Chen, Coord. Chem. Rev. 1999, 185,
3.
[10] M. M. Martin, P. Plaza, Y. H. Meyer, F. Badaoui, J. Bourson, J.-P.
Lefevre, B. Valeur, J. Phys. Chem. 1996, 100, 6879.
[11] C. Amatore in Physical Electrochemistry (Ed.: I. Rubinstein),
Marcel Dekker, New York, 1995.
[12] N. Baltes, L. Thouin, C. Amatore, Angew. Chem. 2004, 116, 1455;
Angew. Chem. Int. Ed. 2004, 43, 1431.
[13] J. J. Watkins, J. Chen, H. S. White, H. D. Abruna, E. Maison-
haute, C. Amatore, Anal. Chem. 2003, 75, 3962.
[14] A. J. Pearson, W. J. Xiao, J. Org. Chem. 2003, 68, 5361.
[15] A. J. Pearson, W. J. Xiao, J. Org. Chem. 2003, 68, 2161.
[16] A. J. Pearson, J. J. Hwang, M. E. Ignatov, Tetrahedron Lett. 2001,
42, 3537.
[17] A. J. Pearson, J. J. Hwang, Tetrahedron Lett. 2001, 42, 3541.
[18] A. J. Pearson, J. J. Hwang, Tetrahedron Lett. 2001, 42, 3533.
[19] A. J. Bard, L. R. Faulkner, Electrochemical Methods, Wiley, New
York, 2001.
[20] R. G. Evans, O. V. Klymenko, C. Hardacre, K. R. Seddon, R. G.
Compton, J. Electroanal. Chem. 2003, 556, 179.
[21] In Figure 2 the current peak value for wave Ox3 is about the
same as those of waves Ox1 or Ox2, although one would expect a
twofold increase to account for the two-electron process which
occurs at this wave. Conversely, the correct result is obtained in
[32] Q. Xin, R. M. Wightman, Anal. Chem. 1998, 70, 1677.
[33] C. Amatore, S. Arbault, M. Guille, F. Lemaꢂtre, ChemPhysChem
2007, 8, 1597.
[34] C. Amatore, S. Arbault, C. Bouton, K. Coffi, J.-C. Drapier, H.
Ghandour, Y. Tong, ChemBioChem 2006, 7, 653.
5214
www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 5211 –5214