Regulating copper-binding affinity with photoisomerizable azobenzene ligand by construction of a self-assembled monolayer

Full text of DOI:10.1002/anie.200902048

Angewandte
Chemie
DOI: 10.1002/anie.200902048
Electrochemistry
Regulating Copper-Binding Affinity with Photoisomerizable
Azobenzene Ligand by Construction of a Self-Assembled Monolayer**
Isao Takahashi, Yuichiro Honda, and Shun Hirota*
Photoactive molecules have attracted much interest because
they allow molecular properties to be changed and switched
dramatically.[1–5] For example, functional molecules and
systems that both bind and release metal ions have been
constructed by reversible structural regulation using photo-
chromic molecules, such as spiropyran and azobenzene.[3,4] In
the case of a spiropyran derivative, one of the isomers bound a
transition-metal ion through its negatively charged group,
whereas the other isomer did not. This photoregulated metal-
binding system was useful both in solution and in a polymer
matrix.[3] However, the metal-binding ability was relatively
weak since the binding site was singly negatively charged, and
binding sites with two or more charged groups are required
for a more effectively regulated metal-binding system.
It has been reported that an azobenzene derivative
containing two iminodiacetic acid moieties (azo-IDA) could
strongly bind to a metal ion in aqueous solution, but the
difference in the binding constants of the trans and cis isomers
with the CuII ion were relatively small: KCu = 6.49 ꢀ 1010
(trans) and 8.99 ꢀ 1010 mÀ1 (cis).[4b] The cis ligand coordinates
to a CuII ion by intramolecular interaction in solution, whereas
the trans ligand coordinates to a CuII ion presumably through
intermolecular interaction, although it was not discussed in
detail. On a solid surface, the metal-binding behavior of a
ligand is not always identical to that in solution. For example,
although both isomers of an azopyridine derivative bind to
transition-metal ions in solution, only one of the isomers could
bind to them on an Au surface.[5] This difference in binding
ability could be useful in metal-binding regulation, but it is still
relatively weak (KNi = 1.72 ꢀ 104 mÀ1).[6]
To construct a photoregulated system, we synthesized an
asymmetrical azobenzene ligand (L) with an azo-IDA back-
bone and a disulfide group, and prepared its self-assembled
monolayer (SAM) on an Au surface (Scheme 1). The azo-
Scheme 1. a) Azobenzene ligand, trans-H4L. b) Preparation of the
SAMs of L: 1) modification of H4L to an Au surface (H4L–Au),
2) modification of hexanethiol (C6) to H4L–Au (H4L/C6–Au), 3) depro-
tonation of H4L/C6–Au (L/C6–Au). DMF=dimethylformamide.
IDA backbone was designed as the strong CuII-binding site
and the disulfide group as the surface adsorption site. SAMs
have been widely studied, especially in the electrochemical
evaluation of Cu ions bound to adsorbed ligands and in the
photoconversion of azobenzene-modified surfaces.[7–10] In this
study, the SAM of trans-L easily released the Cu ions, whereas
that of cis-L did not. The binding affinity of cis-L, therefore, is
apparently stronger than that of trans-L and irradiation of cis-
L with visible light (cis-to-trans photoisomerization) led to a
decrease in the number of adsorbed Cu ions. These results
clearly show photoregulation of the Cu-binding affinities
affixing the azobenzene ligand to the Au surface.
[*] Dr. I. Takahashi,[+] Y. Honda, Prof. S. Hirota
Graduate School of Materials Science
Nara Institute of Science and Technology
8916-5 Takayama, Ikoma, Nara 630-0192 (Japan)
Fax: (+81)743-72-6119
E-mail: hirota@ms.naist.jp
Homepage: http://mswebs.naist.jp/LABs/hirota/index_e.html
[+] Present address: College of Science and Engineering,
The protonated azobenzene ligand in the trans form
(trans-H4L, Scheme 1a) was synthesized according to
Scheme S1 (see the Supporting Information). Its deprotonat-
ed form (trans-L) was prepared by adding four equivalents of
NaOH to the protonated form. The absorption spectrum of
trans-L in aqueous solution showed three bands at 229 (1.3 ꢀ
104), 334 (2.1 ꢀ 104), and 420 nm (1.6 ꢀ 103 mÀ1 cmÀ1, Fig-
ure S2a in the Supporting Information), which corresponded
to typical p–p* and n–p* bands for a trans-azobenzene.[4b,9b,11]
Irradiation of trans-L with UV light (320 < l < 400 nm)
caused trans-to-cis photoisomerization together with an
absorption change (Supporting Information, Figure S2). The
Aoyama-Gakuin University
5-10-1 Fuchinobe, Sagamihara, Kanagawa 229-8558 (Japan)
[**] We acknowledge Prof. Hideki Masuda and Dr. Tomohiko Inomata
(Nagoya Institute of Technology) for the preparation of Au
substrates, and Dr. Akiharu Satake (Nara Institute of Science and
Technology) for help with the NMR measurements. This work was
partially supported by Grants-in-Aid for Scientific Research (No.
21350095) to S.H. and on Priority Areas (No. 20051016) from MEXT
and TRPCO Research Foundation. I.T. is financially supported as an
Inoue Fellow from the Inoue Foundation for Science.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.200902048.
Angew. Chem. Int. Ed. 2009, 48, 6065 –6068
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
6065

Communications
trans/cis ratio of L at the photostationary state (PSS) obtained
1 U) on the surface were estimated from the desorption waves
as 1.6 ꢀ 10À10 (trans) and 1.3 ꢀ 10À10 molcmÀ2 (cis). By com-
paring the obtained values with the fully covered monolayers
of other azobenzene derivatives ((1–5) ꢀ 10À10 molcmÀ2),[9a,10]
H4L–Au is estimated to be densely packed on the Au surface.
Cyclic voltammograms of (trans-L)/C6–Au before and
after dipping it into CuII(ClO4)2 solution were recorded in
aqueous 0.1m NaClO4 solution (Cu-free, Figure 1a). The
by UV irradiation was trans/cis = 38:62,[12] which was in
agreement with the value reported for azo-IDA.[4b] The
calculated spectrum of cis-L showed three bands at 240 (1.2 ꢀ
104), 295 (6.3 ꢀ 103), and 430 nm (2.1 ꢀ 103 mÀ1 cmÀ1, Fig-
ure S2a in the Supporting Information). Subsequent cis-to-
trans photoisomerization was performed by irradiation with
visible light (420 nm < l), and the ratio at the new PSS was
trans/cis = 80:20.
Adding one equivalent of CuII ions to the trans-L solution
resulted in a blue shift of the p–p* band (327 nm, Figure S3 in
the Supporting Information). The addition of CuII also
affected the efficiency of photoisomerization. The ratios of
CuII complexes for the PSSs obtained by UV and visible
irradiation were trans/cis = 54:46 and 90:10, respectively.
These values show that the trans-to-cis photoconversion is
unfavorable in the presence of CuII ions, whereas the cis-to-
trans conversion is favorable. It has been reported that the
thermal cis-to-trans isomerization of cis-azo-IDA was much
slower for its CuII complex than its unligated form, because of
stabilization of the CuII complex.[4b] The structural similarity
of L to azo-IDA suggests that the thermal stability of
[CuII(cis-L)] is increased by the CuII ion. Notably, the
stabilized [CuII(cis-L)] complex could be easily photocon-
verted to the trans form by irradiation with visible light, so
that the bound CuII ions could be removed.
Figure 1. a) Cyclic voltammograms of L/C6–Au (dotted line) and
[CuII(L)]/C6–Au (gray solid line, 1st scan; black solid line, 70th scan) in
aqueous 0.1m NaClO4 solution: trans (upper) and cis forms (lower).
Scan rate, 50 mVsÀ1. b) Plots of the relative anodic peak current (Ipa/
The absorption spectra of [CuII(trans-L)] and [CuII(cis-L)]
showed similar d–d transition bands around 700 nm (Support-
ing Information, Figure S4).[13] This similarity implies that the
coordination structures of [CuII(trans-L)] and [CuII(cis-L)] are
similar to each other in aqueous solution: two trans-L or one
cis-L ligand may bind to the CuII ion by intermolecular or
intramolecular interaction, respectively. Cyclic voltammo-
grams of both [CuII(trans-L)] and [CuII(cis-L)] gave similar
irreversible reduction waves at about À800 mV in aqueous
0.1m NaClO4 solution (Supporting Information, Figure S5).[13]
Since similar reduction waves were observed in the absence of
CuII ions for trans- and cis-L, the observed waves are
attributed to the reduction from the azo to the hydrazo
species (2eÀ, 2H+).[9a] No redox signals of the CuII centers
were detected in the voltammograms of [CuII(trans-L)] and
[CuII(cis-L)], and furthermore no free CuII ion was observed
(CuII(ClO4)2: cathodic peak potential Epc = À60 mV). These
results clearly demonstrate that the CuII ion binds tightly to
the trans and cis ligands in aqueous 0.1m NaClO4 solution.
SAMs of trans- and cis-L were prepared step by step
(Scheme 1b): 1) a densely packed monolayer of H4L was first
constructed on an Au surface (H4L–Au), 2) hexanethiol (C6)
was then filled into the unmodified space on the Au surface
(H4L/C6–Au),[14] and finally 3) H4L units were deprotonated
by NaOH (L/C6–Au). Since cis-H4L was immobilized by
dipping it into the UV-irradiated H4L solution which con-
tained some amount of its trans form (cis/trans = 80:20,
Figure S6 in the Supporting Information), the SAMs of cis-
L presumably contained both the cis and trans forms.[15] The
desorption waves of the cis and trans forms of H4L–Au were
both observed at about À900 mV (Supporting Information,
Figure S7), which was similar to the value obtained for the
SAM of a lipoic acid.[16] The coverages of the H4L units (2eÀ/
I
pa,0) versus the total cycle number for [CuII(L)]/C6–Au (upper, trans
form; lower, cis form). The anodic peak current (Ipa) is normalized to
that of the first cycle (Ipa,0). The arrows in (b) represent the first scan
after an interval (10–20 s) in the consecutive scans.
SAM after incubation of CuII ions ([CuII(trans-L)]/C6–Au)
gave a pair of redox waves at E1/2 = + 222 mV, whereas that
before incubation did not. These results indicate that the
observed waves are caused by the redox reaction of CuI/II
(1eÀ). From the quantity of the electricity of the cathodic
wave in the first cycle (Qpc,0), the surface coverage by
adsorbed Cu ions was 1.9 ꢀ 10À10 molcmÀ2. Taking into
account the coverage of trans-H4L (1.6 ꢀ 10À10 molcmÀ2),
about 1.2 equivalents of Cu ion was bound to trans-L on the
surface.
The intensities of the observed redox peaks of [CuII(trans-
L)]/C6–Au were gradually decreased by performing redox
cycles (Figure 1b). Isopotential points were observed at + 164
and + 286 mV during the process (Supporting Information,
Figure S8), which suggests that the initial Cu-bound state
changes.[17] The intensities of the redox peaks increased again
after an interval in consecutive scans (the arrows in
Figure 1b). These two observations indicated that the
decrease in the intensities of the redox peaks was not a
result of the desorption of trans-L but of the Cu ions: the Cu-
bound state ([CuII(trans-L)]/C6–Au) was converted to the Cu-
free one ((trans-L)/C6–Au). In addition, the redox waves of
CuI/II were almost completely diminished in the 70th cycle
(Figure 1). Decreases in the intensities of the redox peaks of
6066 www.angewandte.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 6065 –6068

Angewandte
Chemie
the Cu ions have been observed in relatively weak binding of
Cu ions to adsorbed molecules.[8f] Therefore, the Cu-binding
affinity to the immobilized trans-L is relatively weak and the
bound Cu ions may be easily desorbed from [CuII(trans-L)]/
C6–Au.
In conclusion, we have demonstrated that the Cu-binding
affinity of immobilized trans-L on an Au surface was weaker
than that of cis-L. Moreover, irradiation of the stronger CuII-
binding state, [CuII(cis-L)]/C6–Au, with UV light led to
removal of the bound Cu ions. These results indicate that
the Cu-binding state can be photoregulated, although the
binding affinity of the ligand is potentially high. Since the
coordination sites, the target metal ions, and the substrate
materials of the ligand could be changed systematically, this
type of azobenzene ligand could be useful to construct not
only a functional molecular unit, but also a novel reversible
metal-ion-binding system.
The redox wave of CuI/II was observed at E1/2 = + 280 mV
in the cyclic voltammograms of [CuII(cis-L)]/C6–Au (Fig-
ure 1a), and about 0.9 equivalents of Cu ion was bound to
each ligand L, judging by the coverage of the Cu ions (1.2 ꢀ
10À10 molcmÀ2). The intensities of the peak currents
decreased as in [CuII(trans-L)]/C6–Au, but the decrease
stopped around Ipa/Ipa,0 ꢀ 0.5 (Figure 1b). The (cis-L)/C6–Au
was presumably a mixed monolayer composed of cis- and
trans-L (and C6),[15] and the Cu ions bound to trans-L were
easily desorbed as stated above. The desorption of the Cu ions
in [CuII(cis-L)]/C6–Au is attributed to the coexisting trans-L
adsorbed on the Au surface, and desorption of the Cu ions
from immobilized cis-L was somewhat slower (or did not
occur) compared to that from trans-L. These results suggest
that the Cu-binding affinity of cis-L is stronger than that of
trans-L on the Au surface. The difference is likely to be
induced by the surface modification,[5] which makes it difficult
for the adsorbed trans-L to bind to a CuII ion by intermo-
lecular interaction.
To investigate the photoconversion from a stronger
binding state to a weaker one, [CuII(cis-L)]/C6–Au was
irradiated with visible light after 70 cycles of scans. Cyclic
voltammograms were recorded after the electrolyte solution
was renewed to avoid rebinding of the Cu ions. Interestingly,
the intensities of the redox peaks of CuI/II were gradually
decreased by increasing the irradiation time and finally
diminished after 270 min of irradiation (Figure 2). Two
isopotential points (+ 78 and + 333 mV) were detected
during the decrease in intensity, and the intensities of the
redox peaks of the bound Cu ions increased again after
incubation in the Cu solution. The change in the cyclic
voltammograms indicates that [CuII(cis-L)] (the stronger
binding state) was converted to trans-L (Cu-free state)
probably through [CuII(trans-L)] (the weaker binding state)
and the Cu-binding affinity was regulated by irradiation of
light.
Received: April 16, 2009
Published online: July 11, 2009
Keywords: copper · electrochemistry · ligand effects ·
monolayers · photoisomerization
.
[1] a) G. C. R. Ellis-Davies, J. H. Kaplan, Proc. Natl. Acad. Sci. USA
1994, 91, 187 – 191; b) G. A. Woolley, Acc. Chem. Res. 2005, 38,
486 – 493; c) C. Renner, L. Moroder, ChemBioChem 2006, 7,
868 – 878; d) K. L. Ciesienski, K. L. Haas, M. G. Dickens, Y. T.
Tesema, K. J. Franz, J. Am. Chem. Soc. 2008, 130, 12246 – 12247.
[2] a) S. Hirota, Y. Fujimoto, J. Choi, N. Baden, N. Katagiri, M.
Akiyama, R. Hulsker, M. Ubbink, T. Okajima, T. Takabe, N.
Funasaki, Y. Watanabe, M. Terazima, J. Am. Chem. Soc. 2006,
128, 7551 – 7558; b) H. Prakash, A. Shodai, H. Yasui, H. Sakurai,
S. Hirota, Inorg. Chem. 2008, 47, 5045 – 5047; c) I. Takahashi S.
Kuroiwa, H. E. Lindofors, L. A. Ndamba, Y. Hiruma, T. Yajima,
N. Okishio, M. Ubbink, S. Hirota, J. Pept. Sci. 2009, 15, 411 – 416.
[3] a) G. E. Collins, L.-S. Choi, K. J. Ewing, V. Michelet, C. M.
Bowen, J. D. Winkler, Chem. Commun. 1999, 321 – 322; b) T.
Suzuki, Y. Kawata, S. Kahata, T. Kato, Chem. Commun. 2003,
2004 – 2005.
[4] a) M. Blank, L. M. Soo, N. H. Wassermann, B. F. Erlanger,
Science 1981, 214, 70 – 72; b) S. Shinkai, S. Nakamura, M.
Nakashima, O. Manabe, M. Iwamoto, Bull. Chem. Soc. Jpn.
1985, 58, 2340 – 2347; c) A. Momotake, T. Arai, Tetrahedron Lett.
2003, 44, 7277 – 7280; d) M. Nanasawa, Y. Horikago, M. Hirai,
Polym. Bull. 1995, 35, 103 – 108.
[5] Z. Wang, M. J. Cook, A.-M. Nygꢁrd, D. A. Russell, Langmuir
2003, 19, 3779 – 3784.
[6] I. M. Klotz, W.-C. Loh Ming, J. Am. Chem. Soc. 1953, 75, 4159 –
4162.
[7] For examples, see: a) K. Ataka, J. Heberle, J. Am. Chem. Soc.
2003, 125, 4986 – 4987; b) M. Abe, T. Michi, A. Sato, T. Kondo,
W. Zhou, S. Ye, K. Uosaki, Y. Sasaki, Angew. Chem. 2003, 115,
3018 – 3021; Angew. Chem. Int. Ed. 2003, 42, 2912 – 2915; c) J. C.
Love, L. A. Estroff, J. K. Kriebel, R. G. Nuzzo, G. M. Whitesides,
Chem. Rev. 2005, 105, 1103 – 1169; d) I. Takahashi, T. Inomata,
Y. Funahashi, T. Ozawa, K. Jitsukawa, H. Masuda, Chem.
Commun. 2005, 471 – 473; e) S. Zhang, C. M. Cardona, L.
Echegoyen, Chem. Commun. 2006, 4461 – 4473; f) I. Takahashi,
T. Inomata, Y. Funahashi, T. Ozawa, H. Masuda, Chem. Eur. J.
2007, 13, 8007 – 8017; g) T. Inomata, K. Shinozaki, Y. Hayashi, H.
Arii, Y. Funahashi, T. Ozawa, H. Masuda, Chem. Commun.
2008, 392 – 394; h) C. Amatore, D. Genovese, E. Maisonhaute, N.
Raouafi, B. Schꢂllhorn, Angew. Chem. 2008, 120, 5289 – 5292;
Angew. Chem. Int. Ed. 2008, 47, 5211 – 5214; i) S. H. Choi, B.-S.
Kim, C. D. Frisbie, Science 2008, 320, 1482 – 1486.
Figure 2. Cyclic voltammograms of [CuII(cis-L)]/C6–Au before (black
solid line) and after irradiation with visible light for 90 min (dotted
line), 180 min (hashed solid line), and 270 min (gray solid line). The
SAM before light irradiation corresponds to that after 70 scans in
Figure 1. The modified Au surface was irradiated with visible light in
aqueous 0.1m NaClO4 solution. After irradiation, the electrolyte
solutions were renewed before the cyclic voltammograms were mea-
sured.
[8] a) I. Rubinstein, S. Steinberg, Y. Tor, A. Shanzer, J. Sagiv, Nature
1988, 332, 426 – 429; b) D. W. M. Arrigan, L. L. Bihan, Analyst
1999, 124, 1645 – 1649; c) W. Yang, D. Jaramillo, J. J. Gooding,
Angew. Chem. Int. Ed. 2009, 48, 6065 –6068
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 6067

Communications
D. B. Hibbert, R. Zhang, G. D. Willett, K. J. Fisher, Chem.
PSSs were trans/cis = 32:68 (UV irradiation) and 80:20 (visible-
light irradiation) by estimation from the H NMR spectra of L
(Supporting Information, Figure S2b).
1
Commun. 2001, 1982 – 1983; d) W. Yang, J. J. Gooding, D. B.
Hibbert, J. Electroanal. Chem. 2001, 516, 10 – 16; e) R. K.
Shervedani, A. Farahbakhsh, M. Bagherzadeh, Anal. Chim.
Acta 2007, 587, 254 – 262; f) A. M. Etorki, A. R. Hillman, K. S.
Ryder, A. Glidle, J. Electroanal. Chem. 2007, 599, 275 – 287; g) V.
Ganesh, M. P. C. Sanz, J. C. Mareque-Rivas, Chem. Commun.
[13] As the trans-to-cis photoconversion was unfavorable in the
presence of CuII ions, [CuII(cis-L)] was prepared by addition of
one equivalent of CuII ion after irradiation of trans-L with UV
light.
´
2007, 804 – 806; h) I. Szymanska, M. Stobiecka, C. Orlewska, T.
[14] We used hexanethiol to decrease the background currents in the
cyclic voltammograms (Refs. [7d] and [7f]) and to prevent
reactions of the CuII ions with the adsorbed sulfur atoms
(Ref. [8d]).
[15] When an Au surface is dipped into a solution with two different
molecules, the ratio of the molecules adsorbed on the surface is
not generally the same as that in the solution. For examples, see:
a) S. Imabayashi, T. Mita, T. Kakiuchi, Langmuir 2005, 21, 1470 –
1474; b) F. von Wrochem, F. Scholz, A. Schreiber, H.-G.
Nothofer, W. E. Ford, P. Morf, T. Jung, A. Yasuda, J. M. Wessels,
Langmuir 2008, 24, 6910 – 6917.
[16] a) J. Madoz, B. A. Kuznetzov, F. J. Medrano, J. L. Garcia, V. M.
Fernandez, J. Am. Chem. Soc. 1997, 119, 1043 – 1051; b) Y. Dong,
S. Abaci, C. Shannon, M. J. Bozack, Langmuir 2003, 19, 8922 –
8926.
[17] J. G. Gaudiello, T. C. Wright, R. A. Jones, A. J. Bard, J. Am.
Chem. Soc. 1985, 107, 888 – 897.
Rohand, D. Janssen, W. Dehaen, H. Radecka, Langmuir 2008,
24, 11239 – 11245.
[9] a) B. R. Herr, C. A. Mirkin, J. Am. Chem. Soc. 1994, 116, 1157 –
1158; b) T. Weidner, F. Bretthauer, N. Ballav, H. Motschmann,
H. Orendi, C. Bruhn, U. Siemeling, M. Zharnikov, Langmuir
2008, 24, 11691 – 11700.
[10] a) Y.-Q. Wang, H.-Z. Yu, J.-Z. Cheng, J.-W. Zhao, S.-M. Cai, Z.-
F. Liu, Langmuir 1996, 12, 5466 – 5471; b) H. Z. Yu, H. B. Shao,
Y. Luo, H. L. Zhang, Z. F. Liu, Langmuir 1997, 13, 5774 – 5778;
c) W.-W. Zhang, H.-F. Li, L. Liu, J.-L. Xie, C.-S. Lu, Y. Zhou, X.-
M. Ren, Q.-J. Meng, J. Colloid Interface Sci. 2003, 261, 82 – 87.
[11] N. Tamai, H. Miyasaka, Chem. Rev. 2000, 100, 1875 – 1890.
[12] The trans/cis ratios were calculated from the intensities of the p–
p* transition bands (334 (L) and 327 nm ([CuII(L)])): trans (%)
= APSS/Atrans ꢀ 100; APSS, absorbance of the PSS; Atrans, absorb-
ance of the 100% trans form (Ref. [4b]). The actual ratios of the
6068 www.angewandte.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 6065 –6068