Methylammonium groups at the solid walls of nanometer-sized, water-filled monolayer gaps as binding sites for a tetraanionic porphyrin

Article abstract of DOI:10.1021/ja003882a

Long-chain hydrosulfides containing two secondary amide functions and either electron-poor or electron-rich carbon - carbon double bonds were self-assembled on gold surfaces around a flat-lying, octaanionic porphyrin. Rigid and reactive surface monolayers

Full text of DOI:10.1021/ja003882a

3454
J. Am. Chem. Soc. 2001, 123, 3454-3461
Methylammonium Groups at the Solid Walls of Nanometer-Sized,
Water-Filled Monolayer Gaps as Binding Sites for a Tetraanionic
Porphyrin
Marc Skupin,† Guangtao Li,† Werner Fudickar,† Jo1rg Zimmermann,†,‡ Beate Ro1der,‡ and
Ju1rgen-Hinrich Fuhrhop*,†
Contribution from the Freie UniVersita¨t Berlin, FB Biologie, Chemie, Pharmazie, Institut fu¨r Chemie/
Organische Chemie, Takustrasse 3, D-14195 Berlin, Germany, and Humboldt-UniVersita¨t zu Berlin,
Institut fu¨r Physik, InValidenstrasse 110, D-10115 Berlin, Germany
ReceiVed NoVember 6, 2000
Abstract: Long-chain hydrosulfides containing two secondary amide functions and either electron-poor or
electron-rich carbon-carbon double bonds were self-assembled on gold surfaces around a flat-lying, octaanionic
porphyrin. Rigid and reactive surface monolayers with 2 nm-wide, porphyrin-based gaps were thus obtained.
The gold electrodes were then immersed in water, and the double bonds on the gaps’ surfaces reacted with
methylamine. It was added to the double bonds either by Michael addition or by bromination with hypobromite
followed by methylamine substitution. Only the double bonds at the border of the gaps were accessible to
methylamine dissolved in the bulk water volume and could react. The walls of the rigid membrane gaps now
contained methylammonium groups at the sites of the double bonds in defined heights. A tetracationic copper(II)
porphyrinate could not diffuse any more into the gap and did not quench the fluorescence of the octaanionic
porphyrin on the bottom of the gap. A tetraanionic porphyrin, on the other hand, was fixated by the ring of
ammonium groups. The bound porphyrin then acted as molecular cover for the gap with respect to ferricyanide
transport from bulk water to the electrode. It was removed by raising the pH to a value of 12, where the
methylammonium groups were neutralized to amines. Lowering the pH to 7 again and addition of more of the
anionic porphyrin reclosed the gap.The porphyrin “cover” should be localized at distances of 8-10 and 20 Å
from the bottom porphyrin by multiple charge interactions. The 8-10 Å distance is ideal for studies of
photoinduced electron transfer between two porphyrin monomers of different redox potential. Furthermore it
was found, that redox-active tyrosine could be trapped in the water volume above the porphyrin on gold.
Introduction
diffused into these gaps and reached their bottom. They quan-
titatively quenched the fluorescence of free-base porphyrins
there, if they carried opposite charges and paramagnetic central
ions.5 Finally, it was repeatedly shown, that cyclic water-soluble
edge amphiphiles derived from cyclohexane or benzene, which
fitted into the structure of hexagonal ice,7 for example, trans-
1,2-cyclohexanediol, or corresponding phenols or amines8,9
stopped the flow of ferricyanide ions through the membrane
gaps.2-5 This was related to the “immobilization” of the water
molecules8,9 within the hydrophobic gaps.
We now explored the rigidity of the walls of the membrane
gaps after having tagged methylammonium groups in different
heights to them. trans-Configured C-C double bonds were
chosen as reactive sites. Since the walls are impermeable to
amines2, it was anticipated that reagents in bulk water would
only attack the edges of the gaps and otherwise would leave
the monolayer intact, which was rigidified by amide hydrogen-
bond chains. If this was indeed the case, one could use the wall-
substituents, for example, ammonium groups, to fixate a second
porphyrin or other fitting redox-active molecules within the
membrane gap and at a chosen distance from the bottom
porphyrin. Furthermore, the bottom porphyrin would be in a
more hydrophobic environment, the top porphyrin, in a more
We have described earlier self-assembly procedures for
bolaamphiphiles (“bolas”) containing two secondary amide
groups close to both headgroups on poly(acrylonitrile)1 and on
gold.2-4 Bolas are amphiphiles with headgroups on each end
of a hydrophobic core.5 Rigid, impermeable monolayers have
been obtained, which were impermeable to amines1 and formed
rigid walls around 1-3 nm-wide gaps around steroids3 or
porphyrins.4,5 When the gold electrodes were submersed in
aqueous ferricyanide solutions, the gaps were filled with water,
and cyclic voltammograms (CVs) of the solute in the bulk
solution were obtained. Furthermore, the 2 nm gaps were
surrounded by solid walls, in which the monomeric building
blocks were fixed by covalent gold-sulfur bonds6 as well as
by two hydrogen-bond chains.1-5 Fitting metalloporphyrins
* To whom correspondence should be addressed.
† Freie Universita¨t Berlin.
‡ Humboldt-Universita¨t zu Berlin.
(1) Bo¨hme, P.; Hicke, H.-G.; Bo¨ttcher, C.; Fuhrhop, J.-H. J. Am. Chem.
Soc. 1995, 117, 5824.
(2) Fuhrhop, J.-H.; Bedurke, T.; Gnade, M.; Schneider, J.; Doblhofer,
K. Langmuir 1997, 13, 455.
(3) Fuhrhop, J.-H.; Ruhlmann, L.; Messerschmidt, C.; Fudickar, W.;
Zimmermann, J.; Ro¨der, B. Pure Appl. Chem. 1998, 70, 2385.
(4) Fudickar, W.; Zimmermann, J.; Ruhlmann, L.; Roeder, B.; Siggel,
U.; Fuhrhop, J.-H. J. Am. Chem. Soc. 1999, 121, 9539.
(5) Fuhrhop, J.-H.; Fritsch, D. Acc. Chem. Res. 1986, 19, 130.
(6) Whitesides, G. M.; Bain, C. D. Angew. Chem. 1989, 101, 533.
(7) Franks, F. Pure Appl. Chem. 1987, 59, 1189.
(8) Vossen, M.; Forstmann, F. J. Chem. Phys. 1995, 101, 2379.
(9) Gompper, G.; Hauser, M.; Kornyshev, A. A. J. Chem. Phys. 1995,
101, 3378.
10.1021/ja003882a CCC: $20.00 © 2001 American Chemical Society
Published on Web 03/27/2001

Binding Sites for a Tetraanionic Porphyrin
J. Am. Chem. Soc., Vol. 123, No. 15, 2001 3455
Scheme 2
Figure 1. Model of non-covalent porphyrin heterodimers in rigid
membrane gaps containing water and redox-active tyrosine.
Scheme 1
bolas. Syntheses became long and tedious. We therefore settled
for an electron-rich double bond instead. A simple synthesis of
the asymmetric bola 2 was realized in high yield by applying a
Claisen rearrangement11 of a protected allylic vinyl ester
(Scheme 2). Attempts with Wittig-type reactions of bolas only
gave low yields.
Both bolas 1 and 2 were insoluble in water. Their reactivity
with amines and oxidants in bulk solutions was therefore studied
in methanol/water mixtures. It was found that acrylamide reacted
with ammonia at pH 12 within 60 min10 to give the 3-amino
compound 3 in quantitative yield. The nonactivated double bond
in a model bola was more difficult to oxidize in methanol/water.
Attempts with persulfates or peroxides resulted in complex
mixtures of products in low yields; peracids yielded the epoxides
in quantitative yields but destroyed the gold-sulfur bond in
experiments with the membrane gap systems. Oxidation with
basic osmium tetroxide12 was successful with the model bola
but again did not lead to functional membrane gaps. Reaction
with sodium hypobromite followed by methylamine substitution
gave, however, the expected mixtures of amines 4 in 80% yield
hydrophilic environment, which would differentiate the redox
potentials of otherwise identical porphyrins (Figure 1). Variation
of distance in the range of membrane thickness and of polarity
of two photoreactive molecules could thus be achieved by a
three-step self-assembly procedure plus a functionalization step.
Tyrosine could be trapped in the water-filled gaps.
Results
We envisaged bolaamphiphilic diamide 1 with a Michael-
type terminal double bond as a simple candidate for amination
in aqueous medium. Corresponding amination of acrylamides
has been demonstrated.10 The ammonium groups could be used
to fixate an anionic “top porphyrin” relatively far away from
the “bottom porphyrin” and very close to the bulk water volume.
Bola 1 has been synthesized from 12-hydroxydodecanoic acid
via the 12-aldehyde and Emmons-Horner reaction with the tert-
butyl ester of bromoacetic acid. The overall yield starting from
the aldehyde was 60% (Scheme 1).
1
as shown by H NMR spectroscopy and was also compatible
with the membrane gaps (Scheme 3).
The porphyrins 5a,b,c and 6 applied in this work have been
described earlier4 (Scheme 4).
The self-assembly procedure on gold followed a modified
protocol as was described for the analogous compounds without
a double bond:5 at first the gold electrodes were treated with
porphyrin solutions for 1-4 days. Afterward, the dried elec-
trodes were immersed in chloroform solutions of the amphiphiles
For shorter porphyrin-porphyrin distances and a more
hydrophobic environment of the “top” porphyrin a double bond
in the center was more appropriate. Michael-type double bonds
were, however, difficult to realize in the center of asymmetric
(11) Mulzer, J.; Bo¨hm, I.; Bats, J. W. Tetrahedron Lett. 1998, 39, 9643.
(12) . Kwong, H. L.; Soranto, C.; Ogino, Y.; Sharpless, K. B. Tetrahedron
Lett. 1990, 31, 2999.
(10) . Johnson, M. R. J. Org. Chem. 1986, 51, 833.

3456 J. Am. Chem. Soc., Vol. 123, No. 15, 2001
Skupin et al.
Scheme 3
Scheme 4
Figure 2. CV curves and pictograms of ferricyanide ions (1 mM) in
bulk water depending on the coating of the gold electrode: (a) bare
gold, (b) bola 1-coated, (c) bola 1 + porphyrin 5a, (d) after Michael-
type methylamination (see 3 and Figure 3b), (e) after addition of
porphyrin 5b cover (see Figure 3b), (f, g) removal and placement of
cover 5b.
2a), which completely disappeared when a closed monolayer
of 1 blocked the gold electrode (Figure 2b). When porphyrin
5a was at first bound to the subphase followed by bolas 1 or 2,
about 50% of the gold surface was covered by flat-lying
porphyrins, and about 60% of the current on bare gold was
measured as determined from peak heights of the CVs (Figure
2c). The area of the remaining CV curves was always roughly
proportional to the amount of porphyrin bound, as measured
by the intensity of the Soret band. Electroneutral, covalently
bound and octaanionic porphyrins behaved similarly; no charge
repulsion effects were observed. Treatment of this porous
system with methylamine at pH 12, and subsequent washing
and neutralization caused a retardation of the ferricyanide ion
transport (Figure 2d). The cationic methylammonium groups
presumably bind to the ferricyanide anions on their way through.
The tetraphenylcarboxylato porphyrin 5b was then added to the
system in absence of ferricyanide. UV/vis spectroscopy of the
electrode first indicated a large excess of 5b with respect to 5a.
1 or 2 in the form of their reduced hydrosulfides at room
temperature (see Experimental Section).
Cyclic Voltammetry (CV). In the first CV-experiments it
was shown that bolas 1 and 2 containing trans-configured
carbon-carbon double bonds behaved on gold in the same way
as their saturated analogues. Ferricyanide ions (10-4 M)
produced a strong cyclic voltammogram on bare gold (Figure

Binding Sites for a Tetraanionic Porphyrin
J. Am. Chem. Soc., Vol. 123, No. 15, 2001 3457
When the electrode was washed with water, the Soret band
intensities of 5a and 5b were similar. In FTIR-spectra of the
aminated monolayer, however, we could not detect any new
bands. Use of CD3NH2 did not change this situation. The
concentration was too low. The electrode was then plunged into
a KCl/ferricyanide solution, and it was found that the ion
transport was blocked by the porphyrin as indicated by a
complete loss of the CV current. The tetraanionic porphyrin
was thus also bound to the methylammonium groups within
the gap and acted as a cover (Figure 2e). Upon neutralization
of the methylammonium groups by sodium hydroxide, the
porphyrin 5b cover was released again, and the ion flow was
restored (Figure 2f). Lowering the pH and addition of more
porphyrin 5b again led to a complete interruption of the flow
of ferricyanide ions (Figure 2g). A change of pH alone did not
lead to porphyrin coverage, because the concentration of 5b in
the supernatant was too low (∼10-12 M). The process was
repeated six times and was fully reproducible. No significant
clogging of the pore was detected, and no extra current was
observed when the porphyrin cover had been replaced. The
cover effect only occurred when the gap had been treated with
methylamine. The original gaps containing only unreacted
double bonds were not blocked by porphyrin 5b.
The models in Figure 3 summarize the results: rigid, por-
phyrin 5a-based membrane gaps allowed the free transport of
ferricyanide ions to the gold electrode (Figure 3a). Upon Michael
addition of methylamine an unknown number of methylammo-
nium groups was covalently attached to the gaps. When ferri-
cyanide was added to the bulk water volume, the ammonium
groups formed salts with them, and their diffusion was retarded.
When porphyrin 5b was first added and ferricyanide afterward,
the porphyrin acted as an effective cover. It blocked the ion
transport completely (Figure 3b). The cover could be removed
by raising the pH to 12, thus neutralizing the amino groups.
However, it did not come back by simply lowering the pH again.
Addition of a large excess of porphyrin 5b was always needed
to close the gap again in a reversible manner. The large excess
is presumably needed, because 5b adsorbs to the membrane.
Washing with water removed this excess. This process was
repeated successfully five times.
Figure 3. (a) Model of the self-assembled mixed monolayer of bola
1 around porphyrin 5a on gold. Ferricyanide ions pass freely (compare
CV curve c) in Figure 2). (b) Same model after Michael addition of
methylamine to the double bonds and a tetraanionic porphyrin to the
resulting methylammonium groups. The ferricyanide current is blocked
(compare CV curves e and g in Figure 2).
We then performed a similar series of experiments with bola
2 containing an electron-rich double bond. Bola 2 was again
bound to gold electrodes in the same fashion as described for
1 and then functionalized with methylamine. For this purpose
the gold electrode covered with porphyrin 5a and bola 2 was
first immersed in an aqueous sodium hypobromite solution (0.1
M, pH 12) for 5 min, then thoroughly washed with water, and
treated overnight with methylamine in water (0.1 M, pH 10).
This sequence of reactions presumably yielded regioisomers of
1,2-amino alcohols similar to 4. The ferricyanide CV currents
were similar to those observed with bola 1 before and after
addition of methylamine (Figure 4a, compare with Figure 2d).
Then porphyrin 5b was added again, and the CV curve was
again strongly reduced (Figure 4b, compare with Figure 2e).
Raising the pH value to 12 released 5b from the gap again, but
this process now took about 30 min as compared to a few
seconds only in the experiments with the gaps in bola 1 walls,
where porphyrin 5b was situated at the top of the nanometer
well. Furthermore, the release of the porphyrin did not occur
quantitatively in the first cycle. Later cycles were however fully
reversible. The deeply hidden porphyrin is obviously much more
difficult to remove from the gap than the one close to the bulk
water volume.
The model in Figure 5 characterizes the new situation with
the top porphyrin in a more hydrophobic environment and much
closer to bottom porphyrin. The distance should now vary
between 8 and 10 Å, because two different carbon atoms of the
double bond can carry the ammonium group.
The same reversible closing and opening experiments with
the cationic membrane gaps originating from bola 1 were also
performed with Mn(III) TPPS 5c, which has its Soret band at
450 nm. In this case, it could be shown by visible absorption
spectroscopy of the surface monolayer that the 450 and 410
nm bands of both porphyrins 5a and 5c appeared in the probes
with aminated gaps.13 In this case the porphyrin 5c was 20 Å
apart from porphyrin 5a we also detected a weak CV curve for
the oxidation of Mn(III) to Mn(IV) at 450 mV.13
Finally, we also closed the porphyrin 5a gaps in the
membrane made of bola 1 by plunging the electrode for 2 h
into a 0.1 M solution of tyrosine and washing it with water.
After this treatment, ferricyanide ions could not pass the gap
(Figure 6). The blocking effect lasted at least a few days, when
the electrode was immersed in ferricyanide potassium chloride
(13) Spectra and CVs are given in the Supporting Information.

3458 J. Am. Chem. Soc., Vol. 123, No. 15, 2001
Skupin et al.
Figure 4. Cyclic voltammograms (CVs) of ferricyanide in water
observed with gold electrodes covered with (a) methylaminated bola 2
and porphyrin 5a, (b) after addition of tetraanionic porphyrin 5b at pH
7, (c) after rising the pH to 12, (d) after lowering the pH to a value of
7 and adding more of porphyrin 5b, and (e) after raising the pH to 12
again. The resulting CV changes are analogous to those in Figure 2a.
solutions, which did not contain any tyrosine. Addition of 30%
ethanol to the solution reopened the pore. The details of the
experiments with tyrosine and several other solutes in the
entrapped aqueous volume will be published elsewhere.14
In the CV curves of the figures 2, 4, and 6, the oxidation
and reduction peaks first became separated by very large voltage
differences and then disappeared completely. Currents were,
however, still observable. Such a behavior is indicative of a
change from linear to radial diffusion, corresponding to close
and far away pores.15-17 The gaps, into which ferricyanide ions
diffuse at a given time, seem to get further and further apart
after amination or plugging.15-17 The ferricyanide current was
not dependent on the kind of porphyrin which was on the bottom
of the gap. Electroneutral porphyrin thiols were also attached
there by self-assembly and gave the same CV curves as
octaanion 5a. A charge repulsion as observed with negatively
charged monolayers18 was not found with the thin porphyrin
layer.
Figure 5. (a) Model of the membrane gap made of 2, (b) containing
methylammonium groups which fixate porphyrin 5b at a distance of
about 8 ( 10 Å to the bottom porphyrin 5a.
Figure 6. CV of a 0.2 M ferricyanide solution with a gold electrode
covered with porphyrin 5a and bola 1 before (a) and after (b) treatment
with 0.1 M tyrosine.
Fluorescence Quenching. Octaanionic porphyrin 5a adsorbed
to gold electrodes of intermediate smoothness showed fluores-
cence.5 The emission has been totally quenched by adding
tetracationic porphyrins containing paramagnetic Mn(III) or
Cu(II) central ions.5
of bola 1 and 2 membranes. The two fluorescence bands at 670
and 740 nm of porphyrin 5a on gold in the presence and absence
of the enveloping bola 1 showed the usual inversion of
intensities. Addition of the tetracationic copper(II) porphyrin 6
led to quantitative quenching, both in the absence and in the
presence of the rigid membrane. The quenching process without
the membrane occurred immediately; in the membrane gap it
took about 100 s (Figure 7). This was somewhat faster than
with the corresponding gap made of the bola without a double
bond4 (∼600 s).
The heterodimerization of the gold-bound porphyrin and the
water-dissolved metalloporphyrin also took place in the gaps
(14) Li, G.; Fuhrhop, J.-H. Submitted for publication.
(15) Amatore, C.; Saveant, J. M.; Tessier, D. J. Electroanal. Chem. 1983,
147, 39.
(16) Amatore, C.; Saveant, J. M.; Tessier, D. J. Electroanal. Chem. 1983,
146, 37.
(17) Murray, R. W. In Molecular Design of Electrode Surfaces; Murray,
R. W., Ed.; Techniques of Chemistry Series; Wiley-Interscience: New York,
1992; Vol. 22, Chapter 1.
(18) Gadzekpo, V. P. Y.; Xiao, K. P.; Aoki, H.; Bu¨hlmann, P.; Umezawa,
Y. Anal. Chem. 1999, 72, 5109.
The gap was again totally closed by bathing the electrode
for 2 h in a 0.1 M solution of 1,2-trans-cyclohexanediol and

Binding Sites for a Tetraanionic Porphyrin
J. Am. Chem. Soc., Vol. 123, No. 15, 2001 3459
Figure 7. Fluorescence spectra of porphyrin 5a before (upper spectra) and after (lower spectra) addition of copper porphyrinate 6. Inserts: time
course of fluorescence decrease (a) bare gold, (b) bola 1 fence around porphyrin 5a, (c) bola 1 fence after methylamination.
subsequently washing it with pure water. Treatment with
ammonia and washing with water also led to inaccessibility of
the porphyrin at the gap’s bottom by the cationic copper
porphyrinate 6 in bulk solution. The fluorescence intensity
dropped at first quickly by 30% but then remained constant.
Upon the pH rising to 11, no change in fluorescence was
observed. Further addition of a large excess of copper porphy-
rinate 6 at pH 11 had no effect either. Size and charge of the
methylammonium groups prevented the passage of 6.
functionalization with charged groups in water with reagents
in the gap’s water volume, (iv) filling of the gap with redox-
active solutes, for example, tyrosine, (v) self-assembly of a
fitting second porphyrin or other electron acceptor or donor,
which binds to the charged gap components. Thus far, this
sequence has only been realized on a planar gold electrode.
Analytical tools, which allow following the five-step-assembly,
have been limited to cyclic voltammetry of ferricyanide and
fluorescence quenching experiments. Several experiments using
various methods of scanning microscopy (SFM, STM, SNOM)
failed thus far to detect and characterize the water-filled 2 nm
gaps. Nevertheless, we are confident that the model given in
Figure 8 is close to reality. The π-orbitals of the carbon-carbon
double bond will point into the gap if one assumes (i) that the
amide hydrogen-bond chains run around the gap and (ii) that
the chains occur in an all-trans conformation. Michael addition
of an amine or a bromination/substitution sequence could easily
be started with reagents in the water-filled gap.
3. Discussion
The procedures described above allow the construction of a
porphyrin heterodimer or a corresponding porphyrin-quinone,
porphyrin-flavin, etc. pair by application of a simple synkinesis
()synthesis of noncovalent molecular systems19,20) sequence:
(i) porphyrin self-assembly (ii) bola self-assembly, (iii) bola
(19) Fuhrhop, J.-H.; Ko¨ning, J. Molecular Assemblies and Membranes;
Stoddart, J. F., Ed.; Monographs in Supramolecular Chemistry; Royal
Society of Chemistry: London, 1994.
(20) Fuhrhop, J.-H.; Endisch, C. Molecular and Supramolecular Chem-
istry of Natural Products and Model Compounds; Marcel Dekker: New
York, 2000.
Nature arranges porphyrin and other redox-active systems in
the center of assemblies of protein helices, which envelop the
rigid gap and provide the large variety of amino acid side chains

3460 J. Am. Chem. Soc., Vol. 123, No. 15, 2001
Skupin et al.
separation or to apply solid-state NMR experiments, which could
characterize the immobilized solutes in the gaps. There is not
enough material in one square centimeter of a molecular
monolayer to allow such investigations with the equipment
available.
We therefore attempted to transfer the membrane system to
the surface of colloidal gold particles. Small, water-soluble
particles (diameter < 4 nm) were found to be useless as a basis
for synkinesis of well-organized membrane gaps. Neither the
four-step self-assembly process of the doubly occupied gaps
nor the blockade of the gaps with cyclohexanediol worked
reproducibly. We assume that the high curvature of these par-
ticles does not allow the preparation of wall-like domains around
the adsorbed dyes. First experiments with 30-50 nm citrate
gold particles have been more successful, and results will be
reported in due course.
Figure 8. Bird’s-eye view of the porphyrin 5a-based membrane gap
on a gold electrode. The π-orbitals of the carbon-carbon double bonds
(size is exaggerated) point into the gap and are replaced by methy-
lammonium groups after a Michael addition. The arrows indicate the
assumed direction of amide hydrogen-bond chains.
as binding sites. Photosynthetic and catalytic sites are thus
realized. The membrane gaps developed in this work are much
less organized. The site and type of a single side chain can be
determined, but differentiation of substituents within a gap is
not possible, and the gap cannot adjust to solutes. Except for
the distances of the top porphyrin from the bottom porphyrin
and from the bulk water volume, the stereochemistry of the
assembly cannot be controlled. Furthermore, neither the number
nor the relative position of the substituents, for example,
methylammonium groups, are known accurately.
Nevertheless, one may assume at least 50% occupancy of
the double bonds in the wall by methylammonium groups, which
would lead to a more or less continuous ring of positive charges
in the gap, and the system introduced here offers at least six
advantages, mostly of preparative nature:
(i) The olefinic amphiphiles can easily be prepared and
adjusted around a dye which is covalently bound to a gold or
other reactive subphase.
(ii) The membrane’s integrity is not disturbed by adding
charged or other highly water-soluble groups to the hydrophobic
core.
(iii) The distance between two reactive molecules can be
made much longer than in covalent assemblies. Low solubility
of rigid systems is not a problem.
(iv) Analysis by electrochemical and spectroscopic methods
is straightforward.
Experimental Section
Syntheses of compounds 1-6 are described in the Supporting
Information and refs 22-32.
Self-Assembled Monolayers (SAMs). The gold electrodes5 were
initially washed with spectroscopic grade dichloromethane. After drying,
they were cleaned with a solution of 30% H2O2 and concentrated
H2SO4 (3:1) for 30 s, rinsed with MilliQ water, and dried with a strong
stream of nitrogen. They were exposed to a 10-3 M aqueous solution
of porphyrin 5a at pH 12 for 3 days. The porphyrin-covered electrodes
were then rinsed with MilliQ water and dried again with nitrogen. The
modified electrodes were then exposed to a 10-2 M chloroform solution
of bola 1 or 2 overnight, washed with distilled chloroform, dried with
nitrogen and washed with MilliQ water.
Functionalization of Bolas (1) and (2) on the Electrodes. The
Michael addition to bola 1 was carried out by immersing the electrodes
in a solution of 0.5 M methylamine overnight. The electrodes were
then washed with MilliQ water. The electrode with bola 2 gaps were
exposed to an aqueous solution of 5 × 10-2 M HOBr for 5 min at 5
°C. After washing with milliQ water, the electrodes were plunged into
a 0.5 M methylamine solution overnight and washed again with MilliQ
water.
Addition and Removal of Porphyrin 5b. After the Michael addition
described above, the gold electrode was rinsed with waterm and a CV
curve was measured in ferricyanide/KCl, and the electrode rinsed again
with 5 mL of distilled water. The electrode was then plunged into a
1 × 10-4 M solution of the tetracarboxylate porphyrin 5b for 15 min,
removed, rinsed again with water, and then used for measurements of
CV curves in the ferricyanide/KCl solution at pH 7. The detectable
current had been diminished drastically (see Figure 5b and 5c). The
electrode was then removed again, rinsed with water, and plunged into
EtOH/KOH (pH 12) for 15 min. The use of ethanol was essential. Water
or water/methanol solutions did not remove significant amounts of
porphyrin 5b. After washing with water again, the CV curve in water
containing ferricanide/KCl was reestablished (Figure 5e). This cyclic
procedure was repeated five times with similar results.
(v) The water volume between the reactive molecules can
be doped with tyrosine or ascorbic acid, which may act as
electron-transfer agents.
(vi) The systems can be transferred to gold colloids and may
then be used in much larger quantities in bulk water volumes
(see Outlook section).
Reaction of Carbon-Carbon Double Bonds in Solution. Michael
Addition of Ammonia to Acrylamide in Water. Acrylamide (4.1 g,
Outlook
We assume that any central metal ion may be introduced into
both the bottom and top porphyrins. The first oxidation and
reduction potentials of the porphyrin ligands can thus be varied
in a range from 0.5 to 1.5 V, and from -0.8 to -2.2 V,
respectively.21,22 Long-distance, light-induced charge separation
thus become possible in simple noncovalent systems. Further-
more, tyrosine or ascorbic acid can be localized in the aqueous
volume between the electron donor and acceptor dyes.
(23) Corey, E. J.; Suggs, J. W. Tetrahedron Lett. 1975, 31, 2647.
(24) Bardili, B.; Marschner, H.; Weyerstahl, P. Liebigs Ann. Chem. 1985,
275.
(25) Anderson, G. W.; Callahan, F. M. J. Am. Chem. Soc. 1960, 82,
3359.
(26) Schneider, J.; Messerschmidt, C.; Schulz, A.; Gnade, M.; Schade,
B.; Luger, P.; Bombicz, P.; Hubert, V.; Fuhrhop, J.-H., Langmuir 2001. In
press.
(27) Corey, E. J.; Wei-Guo, S. Tetrahedron Lett. 1990, 31, 2089.
(28) Arnold, R. C.; Lien, A. P.; Alm, R. M. J. Am. Chem. Soc. 1950,
74, 731.
Thus far, however, it has not been possible either to perform
flash photolysis experiments looking for light-induced charge
(29) Chaudhary, S. K.; Hernandez, O. Tetrahedron Lett. 1979, 2, 99.
(30) Mulzer, J.; Dehmlow, H.; Buschmann, J.; Luger P. J. Org. Chem.
1992, 57, 3194.
(21) Fuhrhop, J.-H.; Kadish, K.; Davis, D. G. J. Am. Chem. Soc. 1973,
95, 5140.
(31) Evans, P. A.; Roseman, J. D.; Gaber, L. T. Synth. Commun. 1996,
4685.
(22) Holt, D. A.; Luengo, J. I.; Clardy, J. J. Am. Chem. Soc. 1993, 115,
9936.
(32) Kergomard, A. Bull. Soc. 1961, 2360.

Binding Sites for a Tetraanionic Porphyrin
J. Am. Chem. Soc., Vol. 123, No. 15, 2001 3461
58 mmol) was dissolved in 100 mL of water. Aqueous ammonia (25
mL, 32%) was added, and the solution stirred for 2 h. The solvent was
removed in vacuo, and a H NMR spectrum of the residue was taken
with an integral of 29.6 for eight methylene protons (δ ) 1.2 and 1.4
ppm). The yield of amines was estimated to be more than 90%. MS
(FAB,pos.,Xe) m/z 262 (M) 283 (M + Na+); 249 (M - diol).
FTIR Spectroscopy. FTIR spectroscopy was also used to follow
the conversion of the assembled Michael-type bola 1 on gold to the
amine. Deuterated (CD3NH2) methylamine was used to set off the
signals of the added group (2100-2200 cm-1). There was found no
signal in the region between 2100 and 2200 cm-1. The spectra only
showed the CH2 vibrations of the chains at 2919 and 2849 cm-1. Even
the terminal CH3 groups (2960 cm-1) of bola 1 material in the
monolayer were not intense enough to be detected.
UV/vis Spectrum on Gold Electrode. The porphyrins on the
covered gold electrodes were also characterized after each self-assembly
process with electronic spectra measured with a fiber optic (Perkin-
Elmer). One example of such a spectrum with approximately equimolar
amounts of 5a and 5b after amination is given in the Supporting
Information.
1
in d6-DMSO. Two triplets of equal intensity (δ ) 2.0 and 2.5) and the
lack of a signal at δ ) 6.0 (<5%) pointed to a quantitative amination.
Control experiments with longer chain R,â-unsaturated amides gave
similar results.
Michael Addition of Methylamine to Bola 1 in Methanol-
Water32. Bola 1 (300 mg, 0.9 mmol) was dissolved in 100 mL of
methanol and 50 mL of water. Aqueous methylamine (10 mL, 40%)
was added at room temperature, and the solution stirred overnight. The
solvents were removed in vacuo, and the residue was characterized by
TLC and mass spectrometry. TLC (silica gel, CHCl3:MeOH ) 10:1)
showed the educt (Rf ) 0.35) and one product of similar mobility(Rf
) 0.20). The major peaks in the mass spectrum (FAB,pos.,Xe) were
at m/z 360 (M),and 329 (educt).
Bromination and Methylamine Substitution. R-Carboxy-ω-car-
bomethoxy-4-decene (400 mg) (Scheme 4) was dissolved in a mixture
of 5 mL of acetic acid and 35 mL of water at 10 °C. A solution of
5.84 g of KBr in 12.5 mL of water oxidized with 25 mL of sodium
hypochlorite (12%) was then added to the decenoic acid solution. The
mixture was stirred for 2 h at 10 °C, 20 mL of dilute HCl was added,
and the solution was extracted with CHCl3 (3 × 10 mL) and dried
with MgSO4. The solvent was removed. A white powder (350 mg)
remained, corresponding to a mixture of R-bromo-alcohol (Scheme 3).
This mixture was then treated in methanol (100 mL) with 5 mL of
aqueous methylamine (40%) overnight. The solvent and excessive
methylamine were removed in vacuo. The residue containing amino
alcohols 4 was redissolved in chloroform. The major proof for the
introduction of an aminomethyl group was given by the appearance of
a doublet of δ ) 2.43 ppm with an integral of 13.5 which compares
Acknowledgment. Financial support by the Deutsche For-
schungsgemeinschaft (SFB 312 “Vectorial Membrane Pro-
cesses” and SFB 448 “Mesoscopic Systems”), the Fonds der
Deutschen Chemischen Industrie and by the FNK of the Free
University is gratefully acknowledged.
Supporting Information Available: Detailed descriptions
of syntheses and a sample of a UV/vis spectrum of porphyrins
5a and 5b on gold after self-assembly are given (PDF). This
material is available free of charge via the Internet at
http://pubs.acs.org.
JA003882A