Giant dendritic molecular electrochrome batteries with ferrocenyl and pentamethylferrocenyl termini

Article abstract of DOI:10.1021/ja8062343

Giant redox dendrimers were synthesized with ferrocenyl and pentamethylferrocenyl termini up to a theoretical number of 3⁹ tethers (seventh generation). Lengthening of the tethers proved to be a reliable strategy to overcome the bulk constraint at the dendrimers periphery. These redox metallodendrimers were characterized by 1H, ¹³C, and ²⁹Si NMR; MALDI-TOF mass spectrometry (for the low generations); elemental analysis; UV-vis spectroscopy; dynamic light scattering (DLS); atomic force microscopy (AFM); electron-force microscopy (EFM) for half- or fully oxidized dendrimers; cyclic voltammetry; and coulometry. UV-vis spectroscopy, coulometry, and analytical data are consistent with an increasing amount of defects as the generation number increases, with this amount remaining relatively weak up to G₅. AFM shows that the dendrimers form aggregates of discrete size on the mica surface, recalling the agglomeration of metal atoms in monodisperse nanoparticles. Cyclic voltammetry reveals full chemical and electrochemical reversibility up to G₇, showing that electron transfer is fast among the flexible peripheral redox sites. Indeed, the redox stability of these new electrochromic dendrimers, i.e., a battery behavior, was established by complete chemical oxido-reduction cycles, and the blue 17-electron ferrocenium and deep-green mixed-valence Fe(III)/Fe(II) dendritic complexes were isolated and characterized. AFM studies also show the reversible dendrimer size changes from upon redox switching between Fe(II) and Fe(III), suggesting a breathing mechanism controlled by the redox potential. Considerable adsorption of highgeneration dendrimers on Pt electrodes such as G₇-Fc allows the easy formation of modified electrodes that sense the ATP anion only involving the electrostatic factor even in the absence of any other type of interaction with the redox tethers.

Full text of DOI:10.1021/ja8062343

Published on Web 12/29/2008
Giant Dendritic Molecular Electrochrome Batteries with
Ferrocenyl and Pentamethylferrocenyl Termini
Catia Ornelas, Jaime Ruiz, Colette Belin, and Didier Astruc*
Institut des Sciences Mole´culaires, UMR CNRS 5255, UniVersite´ Bordeaux 1, 351 Cours de la
Libe´ration, 33405 Talence Cedex, France
Received August 7, 2008; E-mail: d.astruc@ism.u-bordeaux1.fr
Abstract: Giant redox dendrimers were synthesized with ferrocenyl and pentamethylferrocenyl termini up to a
theoretical number of 39 tethers (seventh generation). Lengthening of the tethers proved to be a reliable strategy
to overcome the bulk constraint at the dendrimers periphery. These redox metallodendrimers were characterized
by 1H, 13C, and 29Si NMR; MALDI-TOF mass spectrometry (for the low generations); elemental analysis; UV-vis
spectroscopy; dynamic light scattering (DLS); atomic force microscopy (AFM); electron-force microscopy (EFM)
for half- or fully oxidized dendrimers; cyclic voltammetry; and coulometry. UV-vis spectroscopy, coulometry,
and analytical data are consistent with an increasing amount of defects as the generation number increases,
with this amount remaining relatively weak up to G5. AFM shows that the dendrimers form aggregates of discrete
size on the mica surface, recalling the agglomeration of metal atoms in monodisperse nanoparticles. Cyclic
voltammetry reveals full chemical and electrochemical reversibility up to G7, showing that electron transfer is
fast among the flexible peripheral redox sites. Indeed, the redox stability of these new electrochromic dendrimers,
i.e., a battery behavior, was established by complete chemical oxido-reduction cycles, and the blue 17-electron
ferrocenium and deep-green mixed-valence Fe(III)/Fe(II) dendritic complexes were isolated and characterized.
AFM studies also show the reversible dendrimer size changes from upon redox switching between Fe(II) and
Fe(III), suggesting a breathing mechanism controlled by the redox potential. Considerable adsorption of high-
generation dendrimers on Pt electrodes such as G7-Fc allows the easy formation of modified electrodes that
sense the ATP anion only involving the electrostatic factor even in the absence of any other type of interaction
with the redox tethers.
Introduction
physical properties found in organometallic (in particular
ferrocenyl) polymers5 encouraged chemists to introduce orga-
Dendrimers1 have numerous potential applications in biol-
nometallic complexes into the dendritic structure in order to
obtain well-defined macromolecules with interesting physical
properties.1-4 In particular, metallocenyl dendrimers6 have
already been applied in redox sensing7a-c and various types of
catalysis.7d-h
ogy,2 catalysis3 and materials science.4 The richness of the
(1) (a) Newkome, G. R.; Moorefield, C. N.; Vo¨gtle, F. Dendrons and
Dendrimers. Concepts, Synthesis and Applications; Wiley-VCH,
Weinheim, Germany, 2001. (b) Dendrimers and Other Dendritic
Polymers; Tomalia, D., Fre´chet, J. M. J., Eds.; Wiley-VCH: New York,
2002. (c) Dendrimers and Nanosciences; Astruc, D., Ed.; Elsevier:
New York, 2003; Vol. 6, pp 709-1208; (d) Reviews on metalloden-
drimers: Newkome, G. R.; He, E.; Moorefield, C. N. Chem. ReV. 1999,
99, 1689–1746. (e) Hwang, S. H.; Shreiner, C. D.; Moorefield, C. N.;
Newkome, G. R. New J. Chem. 2007, 31, 1192–1217.
Although dendrimer science is now well-spread,1 only a few
syntheses and characterizations of large dendrimers have been
reported.8 The seminal synthesis of PAMAM dendrimers by
Tomalia8b-d and his group in the early 1980s was a remarkable
landmark, however, in terms of very large dendrimers.8a-c Our
(2) (a) Astruc, D. C. R. Acad. Sci. 1996, 322, 757–766, Se´r. II b. (b)
Kojima, C.; Kono, K.; Maruyama, K.; Takagishi, T. Bioconjugate
Chem. 2000, 11, 910–917. (c) Hecht, S.; Fre´chet, J. M. J. An-
gew.Chem., Int. Ed. 2001, 40, 74–91. (d) Sadler, K.; Tam, J. P.
J. Biotechnol. 2002, 90, 195–205. (e) Cloninger, M. J. Curr. Opin.
Chem. Biol. 2002, 6, 742–748. (f) Stiriba, S. E.; Frey, H.; Haag, R.
Angew. Chem., Int. Ed. 2002, 41, 1329–1334. (g) Grinstaff, M. W.
Chem.sEur. J. 2002, 8, 2838–2846. (h) Boas, U.; Heegarard, P. M. H.
Chem. Soc. ReV. 2004, 33, 43–63. (i) Lee, C.; Mac Kay, J. A.; Fre´chet,
J. M. J.; Szoka, F. Nat. Biotechnol. 2005, 23, 1517–1526.
(4) (a) Zeng, F.; Zimmerman, S. C. Chem. ReV. 1997, 97, 1681–1712.
(b) Matthews, O. A.; Shipway, A. N.; Stoddart, J. F. Prog. Polym.
Sci. 1998, 23, 1–56. (c) Venturi, M.; Serroni, S.; Juris, A.; Campagna,
S.; Balzani, V. In Dendrimers; Vo¨gtle, F., Ed.; Topics in Current
Chemistry; Springer-Verlag: Berlin, 1998; Vol. 197, pp 193-228; (d)
Bosman, A. W.; Bruining, M. J.; Kooijman; Spek, A. L.; Janssen,
R. A. J.; Meijer, E. W. J. Am. Chem. Soc. 1998, 120, 5547–8548. (e)
Bosman, A. W.; Janssen, H. M.; Meijer, E. W. Chem. ReV. 1999, 99,
1665–1688. (f) Grayson, S. M.; Fre´chet, J. M. J. Chem. ReV. 2001,
101, 3819–3867. (g) Crooks, R. M.; Zhao, M.; Sun, L.; Chechik, V.;
Yeung, L. K. Acc. Chem. Res. 2001, 34, 181–190.
(3) (a) Oosterom, G. E.; Reek, J. N. H.; Kamer, P. C. J.; van Leeuwen,
P. W. N. M. Angew. Chem., Int. Ed. 2001, 40, 1828–1849. (b) Astruc,
D.; Chardac, F. Chem. ReV. 2001, 101, 2991–3024. (c) Kreiter, R.;
Kleij, A. W.; Klein Gebbink, R. J. M.; van Koten, G. In Dendrimers
IV: Metal Coordination, Self Assembly, Catalysis; Vo¨gtle, F., Schalley,
C. A., Ed.; Topics in Current Chemistry; Springer-Verlag: Berlin,
2001’ Vol. 217, p 163. (d) van Heerbeeck, R.; Kamer, P. C. J.; van
Leeuwen, P. W. N. M.; Reek, J. N. H. Chem. ReV. 2002, 102, 3717–
3756. (e) Liang, C.; Fre´chet, J. M. J. Prog. Polym. Sci. 2005, 30, 385.
(f) Me´ry, D.; Astruc, D. Coord. Chem. ReV. 2006, 250, 1965–1979.
(5) (a) Foucher, D. A.; Tang, B. Z.; Manners, I. J. Am. Chem. Soc. 1992,
114, 6246–6248. (b) Manners, I. AdVan. Organomet. Chem. 1995, 37,
131–168. (c) Nguyen, P.; Gomez-Elipe, P.; Manners, I. Chem. ReV.
1999, 99, 1515–1548. (d) Whittell, G. R.; Manners, I. AdV. Mater.
2007, 19, 3439–3468. (e) Eloi, J. C.; Chabanne, L; Whittell, G. R.;
Manners, I. Mater. Today 2008, 11, 28–36.
9
590 J. AM. CHEM. SOC. 2009, 131, 590–601
10.1021/ja8062343 CCC: $40.75
2009 American Chemical Society

Giant Dendritic Molecular Electrochrome Batteries
A R T I C L E S
Scheme 1. Schematic Representation of the Tether-Lengthening Strategy for the Introduction of Bulky Functional Groups at the Periphery of
Large Dendrimers
group has recently described the synthesis and characterizations
of giant organic dendrimers utilizing an organo-iron synthetic
route.9 Our efforts are now focusing on the functionalization
of such dendrimers with robust redox groups, in order to
investigate whether a limit to redox reversibility is attained for
large dendrimers and to forecast applications in the nanoelec-
tronics field.
The strategy of lengthening the dendrimer tethers was applied
by attaching a long alkyl chain to the metallocene followed by
its reaction with the dendrimers. The syntheses, characteriza-
tions, and remarkable physical properties of the resulting
ferrocenyl and pentamethylferrocenyl dendrimers are detailed
in this article.
The functionalization of dendrimers requires clean and
quantitative reactions to avoid the formation of mixtures among
the branch termini. Low-generation dendrimers can be func-
tionalized by classic reactions, but giant dendrimers require
special attention due to the bulk constraint at the periphery.
Indeed, it has been proposed that backfolding of dendritic
termini is all the less pronounced as these terminal groups are
bulkier.10c,d Although the synthesis of ferrocenyl dendrimers
has been reported by various research groups,11 the introduction
of bulky organometallic complexes at the termini of the branches
of giant dendrimers remains a challenge. We now show that
lengthening of the dendrimers tethers is a reliable strategy to
overcome these steric problems (Scheme 1).
Results and Discussion
1. Synthesis of the Organic Dendrimers. The dendritic
construction of the dendritic cores follows the known procedure
shown in Scheme 2 illustrating the multiplication by three of
the number of terminal allyl groups, generations being built up
by sequences involving Williamson reaction followed by
hydrosilylation.9
2. Functionalization of the Dendrimers with Redox-Active
Termini. In order to search applications in the nanoelectronics
field, such as molecular batteries12a-d or capacitors,12e-g robust
(8) (a) de Gennes, P.-G.; Hervet, H. J. Phys., Lett. 1983, 44, L351–360.
(b) Denkewalker, R. G.; Kolc, J. F.; Lukasavage, W. J. U.S. Patent
1983 4410688; Chem. Abstr. 1984, 100, 103907. (c) Tomalia, D. A.;
Baker, H.; Dewald, J. R.; Hall, M.; Kallos, G.; Martin, S.; Roeck, J.;
Ryder, J.; Smith, P. Macromolecules 1986, 19, 2466–2470. (d)
Tomalia, D. A.; Naylor, A. M.; Goddart, W. A. Angew. Chem., Int.
Ed. 1990, 29, 138–175. (e) Sherf, U.; Mu¨llen, K. Synthesis 1992, 1–
2. (f) Tomalia, D. A.; Durst, H. D. Top. Curr. Chem. 1993, 165, 193–
313. (g) van der Made, A. W.; van Leeuwen, P. W. N. M.; de Wilde,
J. C.; Brandes, R. A. C. AdV. Mater. 1993, 5, 466–468. (h) Zhou,
L. L.; Roovers, J. Macromolecules 1993, 26, 963–968. (i) Sournies,
F.; Crasnier, F.; Graffeuil, M.; Faucher, J.-P.; Lahana, R.; Labarre,
J.-F. Angew. Chem., Int. Ed. 1995, 34, 578–581. (j) Devadoss, C.;
Bharathi, P.; Moore, J. S. J. Am. Chem. Soc. 1996, 118, 9635–9644.
(k) Moore, J. S. Acc. Chem. Res. 1997, 30, 402–413. (l) Caminade,
A. M.; Majoral, J. P. In Dendrimers; Vo¨gtle, F., Ed.; Topics in Current
Chemistry; Springer: Berlin, 1998; Vol. 170, pp 79-124. (m)
Berresheim, A. J.; Mu¨ller, M.; Mu¨llen, K. Chem. ReV. 1999, 99, 1747–
1785. (n) Watson, M. D.; Fechtenkotter, A.; Mu¨llen, K. Chem. ReV.
2001, 101, 1267–1300. (o) Ihre, H.; Padilla de Jesu´s, O. L.; Fre´chet,
J. M. J. J. Am. Chem. Soc. 2001, 123, 5908–5917. (p) Dvornic, P. R.;
Li, J.; de Leuze-Jallouli, A. M.; Reeves, S. D.; Owen, M. J.
Macromolecules 2002, 35, 9323. (q) Dvornic, P. R. J. Polym. Sci.,
Part A 2006, 44, 2755–2773. (r) Furuta, P.; Brooks, J.; Thompson,
M. E.; Fre´chet, J. M. J. J. Am. Chem. Soc. 2006, 125, 13165–13172.
(6) (a) For a review on ferrocenyl dendrimers and their electrochemical
properties, see: Casado, C. M.; Cuadrado, I.; Moran, M.; Alonso, B.;
Garcia, B.; Gonzales, B.; Losada, J. Coord. Chem. ReV. 1999, 185-
186, 53–79. (b) For electrochemical studies of ferrocenes, see:
Holloday, J. D. L.; Bowden, W. L.; Geiger, W. E. J. Am. Chem. Soc.
1977, 99, 7089–7090. (c) Holloday, J. D. L.; Geiger, W. E. J. Am.
Chem. Soc. 1979, 101, 2038–2044. (d) For electrochemical studies of
multi-ferrocenyl oligomers and polymers, see Rulkens, R.; Lough, A.-
J.; Manners, I.; Lovelace, S. R.; Grant, C.; Geiger, W. E. J. Am. Chem.
Soc. 1996, 118, 12683–12695. (e) Camine, N.; Mueller-Westerhoff,
U. T.; Geiger, W. E. J. Organomet. Chem. 2001, 637, 823–826.
(7) (a) Vale´rio, C.; Fillaut, J.-L.; Ruiz, J; Guittard, J.; Blais, J.-C.; Astruc,
D. J. Am. Chem. Soc. 1997, 119, 2588–2589. (b) Casado, C. M.;
Cuadrado, I.; Alonso, B.; Mora´n, M.; Losada, J. J. Electroanal. Chem.
1999, 463, 87–92. (c) Astruc, D.; Daniel, M.-C.; Ruiz, J. Chem.
Commun. 2004, 2637–2649. (d) Rigaut, S.; Delville, M.-H.; Astruc,
D. J. Am. Chem. Soc. 1997, 119, 11132–11133. (e) Albrecht, M.;
Hovestad, N. J.; Boersma, J.; van Koten, G. Chem.sEur. J. 2001, 7,
1289–1294. (f) Daniel, M.-C.; Astruc, D. Chem. ReV. 2004, 104, 293–
346. (g) Ornelas, C.; Ruiz, J.; Cloutet, E.; Alves, S.; Astruc, D. Angew.
Chem., Int. Ed. 2007, 46, 872–877. (h) Astruc, D.; Ornelas, C.; Ruiz,
J. Acc. Chem. Res. 2008, 41, 841–856.
9
J. AM. CHEM. SOC. VOL. 131, NO. 2, 2009 591

A R T I C L E S
Ornelas et al.
Scheme 2. Synthesis of the G1-27-Allyl Dendrimer9a
a This dendritic construction was continued, using the phenoltriallyl dendron as a building block, until G7 to synthesize the precursors of the ferrocenyl
and pentamethylferrocenyl dendrimers (see eqs 1-7).
Scheme 3. Synthesis of the Ferrocenyl Derivative 14
redox active termini, i.e. ferrocenyl (Fc) and pentamethylfer-
rocenyl (Fc*) were used to functionalize the organic dendrimers.
Knowing that these metallocenes present fully reversible redox
properties in the monomer form, their introduction into the
dendrimers will extend their redox chemistry to the nanoscale.
The strategy of lengthening the branches was used in order
to achieve the functionalization of the giant dendrimers9 with
these metallocenes, because of the forecasted bulk problem.
Phenolic derivatives of ferrocene and pentamethylferrocene
possessing a “long tether” were synthesized, in order to attach
them to the dendrimers using the Williamson reaction (which
has already been shown to be an efficient reaction to construct
and functionalize dendrimers9). The polyferrocenyl dendrimers
were synthesized for eight generations, from G0 (with nine
termini) to G7 (with a theoretical total number of termini, TNT,
of 19 683) and pentamethylferrocenyl dendrimers were synthe-
sized for G0 (nine termini) and G7 (19 683 termini).
2.1. Ferrocenyl Dendrimers Series (Fc). The synthesis of a
phenolic derivative of ferrocene containing a long alkyl chain
was achieved by Friedel-Craft reaction between 11-bromoun-
decanoyl chloride and ferrocene in the presence of AlCl3,
yielding the 11-bromoundecanoyl ferrocene 12, followed by
Clemmensen reduction of the ketone group, yielding the 11-
bromoundecyl ferrocene 13.15 Williamson reaction between 13
and hydroquinone afforded the phenolic derivative 14 (Scheme
3).
(9) (a) Sartor, V.; Djakovitch, L.; Fillaut, J.-L.; Moulines, F.; Neveu, F.;
Marvaud, V; Guittard, J.; Blais, J.-C.; Astruc, D. J. Am. Chem. Soc.
1999, 121, 2929–2930. (b) Ruiz, J.; Lafuente, G.; Marcen, S.; Ornelas,
C.; Lazare, S.; Blais, J.-C.; Cloutet, E.; Astruc, D. J. Am. Chem. Soc.
2003, 125, 7250–7257.
(10) (a) Newkome, G. R.; Yao, Z.; Baker, G. R.; Gupta, V. K. J. Org.
Chem. 1985, 50, 2003. (b) Newkome, G. R.; Moorefield, C. N.; Baker,
G. R. Aldrichim. Acta 1992, 25 (N°2), 21. (c) Lescanec, R. L.;
Muthukumar, M. Macromolecules 1990, 23, 2280–2288. (d) Mansfield,
M. L.; Klushin, L. I. Macromolecules 1993, 26, 4262–4268. (e) Boris,
D.; Rubinstein, M. Macromolecules 1996, 29, 7251–7260. (f) Naidoo,
K. J.; Hughes, S. J.; Moss, J. R. Macromolecules 1999, 32, 331–341.
(11) See refs 1, 6, 7 and (a) Takada, K.; Diaz, D.; Abrun˜a, H.; Cuadrado,
I.; Casado, C. M.; Alonso, B.; Lobete, F.; Garcia, B.; Losada, J. J. Am.
Chem. Soc. 1997, 119, 10763–10770. (b) Shu, C.-F.; Shen, H. M. J.
Mater. Chem. 1997, 7, 47. (c) Tokuhisa, H.; Zhao, M.; Baker, L. A.;
Phan, V. T.; Dermody, D. L.; Garcia, M. E.; Peez, R. F.; Crooks,
R. M.; Mayer, T. M. J. Am. Chem. Soc. 1998, 120, 4492. (d) Schneider,
R.; Ko¨llner, C.; Weber, I.; Togni, A. Chem. Commun. 1999, 2415–
2416. (e) Casado, C. M.; Alonso, B.; Mora´n, M.; Cuadrado, I.; Losada,
J. J. Electroanal. Chem. 1999, 463, 87–93. (f) Turin, C. O.; Chiffre,
J.; de Montauzon, D.; Daran, J.-C.; Caminade, A.-M.; Manoury, E.;
Balavoine, G.; Majoral, J.-P. Macromolecules 2000, 33, 7328–7336.
(g) Casado, C. M.; Gonzales, B.; Cuadrado, I.; Alonso, B.; Mora´n,
M.; Losada, J. Angew. Chem., Int. Ed. 2000, 39, 2135–2138. (h)
Gonzales, B.; Cuadrado, I.; Casado, C. M.; Alonso, B.; Pastor, C.
Organometallics 2000, 120, 4023–4029. (i) Alonso, B.; Casado, C. M.;
Cuadrado, I.; Moran, M.; Kaifer, A. E. Chem. Commun. 2002, 16,
1778–1779. (j) Brettar, J.; Burgi, T.; Donnio, B.; Guillon, D.; Klappert,
R.; Sharf, T.; Descheneau, R. AdV. Funct. Mater. 2006, 16, 260–267.
(k) Kaifer, A. E. Eur. J. Inorg. Chem. 2007, 5015–5027.
The ferrocenyl dendrimers were synthesized by Williamson
reaction between the iodomethyl dendrimers and 14 using an
excess of K2CO3, in DMF, at 80 °C, for 48 h. Eight different
generations (from G0 with nine branches to G7 with a TNT of
19 683) were synthesized in order to study their physical
properties with the increasing number of generations (eqs 1-8).
1
All the ferrocenyl dendrimers were characterized by H, 13C,
and 29Si NMR, elemental analysis, UV-vis. spectroscopy,
(12) (a) Meyer, W. H. AdV. Mater. 1998, 6, 439–448. (b) Romero, P. G.;
Torres-Go´mez, G. AdV. Mater. 2000, 19, 1454–1456. (c) Stanish, I.;
Lowy, D. A.; Hung, C. W.; Singh, A. AdV. Mater. 2005, 9, 1194–
1198. (d) Song, H.-K.; Palmore, G. T. R. AdV. Mater. 2006, 13, 1764–
1768. (e) Jeon, N. L.; Clem, P.; Jung, D. Y.; Lin, W.; Girolami, G. S.;
Payne, D. A.; Nuzzo, R. G. AdV. Mater. 1997, 11, 891–895. (f)
Oldfield, G.; Ung, T.; Mulvaney, P. AdV. Mater. 2000, 20, 1519–
1522. (g) Lidgi-Guigui, N.; Dablemont, C.; Veautier, D.; Viau, G.;
Seneor, P.; van Dau, F. N.; Mangeney, C.; Vaure`s, A.; Deranlot, C.;
Friederich, A. AdV. Mater. 2007, 13, 1729–1733.
9
592 J. AM. CHEM. SOC. VOL. 131, NO. 2, 2009

Giant Dendritic Molecular Electrochrome Batteries
A R T I C L E S
Scheme 4. Synthesis of the Pentamethylferrocenyl Derivative Containing a Long Alkyl Chain and a Phenol Termini.
dynamic light scattering (DLS), atomic force microscopy
(AFM), cyclic voltammetry (CV) and coulometry. The two first
generations (G0-Fc and G1-Fc) show the corresponding molec-
ular peaks at 5 167.50 (calcd. for C306H444O18Si9Fe9, 5 166.27)
and 17 241.55 (calcd for C1017H1488O63Si36Fe27, 17 241.55) in
the MALDI-TOF mass spectra, confirming the structure of the
polyferrocenyl dendrimers (see the Supporting Information,
pages SI10 and SI13). The analytical data confirm the structures
up to G5, but above this generation (G6 and especially G7) the
carbon analysis that is found is low because of structural defects
and increased ability to encapsulate external impurities (solvents,
water, inorganic salts).
presence of substituted cyclopentadienyl derivatives have been
shown to provide a method of choice to synthesize ferrocenes
containing two different rings.13 Hydrosilylation of 25 with
chloromethyl(dimethyl)silane in the presence of the Kartstedt
catalyst yielded compound 26. Subsequent to the substitution
of chloride by iodide in compound 26, Williamson reaction with
hydroquinone afforded the phenolic derivative 28 (Scheme 4).
The pentamethylferrocenyl dendrimers were synthesized by
Williamson reaction between the iodomethyl-terminated den-
drimers and 28 using excess of K2CO3 in DMF at 80 °C for
48 h. The reaction was carried out for G0 and G7 (eqs 10 and
1
10), and the products were characterized by H, 13C and 29Si
NMR, elemental analysis, UV-vis spectroscopy, dynamic light
scattering, and cyclic voltammetry. G0-Fc* shows in the
MALDI-TOF mass spectra the corresponding molecular peak
at 6 450.30 (calcd for C378H606O18Si18Fe9, 6647.12) confirming
the expected structure (see the Supporting Information, page
SI31).
14, K2CO3
G1-27-CH2I
8 G1-27-Fc
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(3)
DMF, 80oC, 48h
(16)
14, K2CO3
G2-81-CH2I
8 G2-81-Fc
(4)
DMF, 80oC, 48h
(17)
14, K2CO3
G3-243-CH2I
8 G3-243-Fc
G7-19683-CH2I K CO , DMF8 G7-19683-Fc*
(10)
DMF, 80oC, 48h
(18)
2
3
(5)
(9)
28, 80oC, 48h
(30)
14, K2CO3
G4-729-CH2I
8 G4-729-Fc
DMF, 80oC, 48h
(19)
3. Characterization of the Metallodendrimers. 3.1.
UV-Vis Spectroscopy. In order to estimate the number of
defects in the dendritic structures due to the divergent synthesis,
we investigated the UV-vis spectroscopy of the dendrimers.
The Lambert-Beer law (A ) εlc) was used to determine the
actual total number of metallocene groups in the dendrimers.18
The UV-vis spectra of the metallocenes present an absorption
band at 440 nm in the Fc series, and at 423 nm in the Fc* series.
The number of metallocene termini in each dendrimer can be
estimated by comparing the molar extinction coefficient ε of
(6)
14, K2CO3
G5-2187-CH2I
8 G5-2187-Fc
(7)
DMF, 80oC, 48h
(20)
14, K2CO3
G6-6561-CH2I
8 G5-6561-Fc
(8)
DMF, 80oC, 48h
(21)
14, K2CO3
G7-19683-CH2I
8 G7-19683-Fc
(9)
DMF, 80oC, 48h
(22)
2.2. Pentamethylferrocenyl Dendrimers Series (Fc*). To
synthesize a phenolic derivative of pentamethylferrocene, we
first functionalized cyclopentadiene with a long alkyl chain and
an olefin termini (23) by reaction of sodium cyclopentadienyl
with 11-bromo-1-undecene. This cyclopentadienyl derivative 23
was then used in a ligand substitution reaction in the complex
[Cp*Fe(CO)2(NCMe)][PF6] (24), induced by UV light at -40
°C, affording the functionalized pentamethylferrocene 25.
Indeed, previous photocleavage studies of cationic [CpFeL3]+
and [Cp*FeL3]+ (L3 ) arene or 3 CO’s) derivatives in the
(13) (a) Gill, T. P.; Mann, K. R. Inorg. Chem. 1983, 22, 1986–1988. (b)
Catheline, D.; Astruc, D. J. Organomet. Chem. 1983, 248, C9–C12.
(c) Catheline, D.; Astruc, D. J. Organomet. Chem. 1984, 266, C11–
C14. (d) Catheline, D.; Astruc, D. Organometallics 1984, 3, 1094–
1100. (e) McNair, A. M.; Schrenck, J. L.; Mann, K. R. Inorg. Chem.
1984, 23, 633–636. (f) Schrenk, J. L.; McNair, A. M.; McCormick,
F. B.; Mann, K. R. Inorg. Chem. 1986, 25, 3504–3508. (g) Ruiz, J.;
Astruc, D. Inorg. Chim. Acta 2008, 361, 1–4.
(14) (a) Koppel, D. E. J. Chem. Phys. 1972, 57, 4814–4820. (b) Frisken,
B. J. Appl. Opt. 2001, 40, 4087–4091. (c) Hassan, P. A.; Kulshreshtha,
S. K. J. Colloid Interface Sci. 2006, 300, 744–748.
9
J. AM. CHEM. SOC. VOL. 131, NO. 2, 2009 593

A R T I C L E S
Ornelas et al.
Table 1. Number of Termini Calculated of the Metallocene-Terminated Dendrimers Using the Lambert-Beer Law and Coulometry Results
series
product
λ (nm)
theoretical no. of termini
ε
calcd no. of terminia
electrolysisb (%)
Fc
monomer 14
G0-Fc (15)
G1-Fc (16)
G2-Fc (17)
G3-Fc (18)
G4-Fc (19)
G5-Fc (20)
G6-Fc (21)
G7-Fc (22)
monomer 28
G0-Fc* (29)
G7-Fc* (30)
440
440
440
440
440
440
440
440
440
423
423
423
1
9
27
ε0 ) 116
1064
3096
9 ( 1
27 ( 1
98
99
95
94
90
92
88
83
81
9417
81 ( 4
243
729
2 187
6 561
19 683
1
24 808
81 709
231 678
724 728
1 629 242
ε0 ) 148
1364
210 ( 10
700 ( 40
2 000 ( 100
6 200 ( 300
14 000 ( 1000
Fc*
9
9 ( 1
14 000 ( 1000
19 683
2 100 122
a ε/ε0: represents the experimental branch number calculated from Lambert-Beer law. b The data are provided with an estimated error of (2%.
the dendrimers with that of the corresponding monomer (ε0).
The linearity of the absorption with the number of ferrocenyl
groups in the low-generation dendrimers confirms that there is
no interaction between the ferrocenyl groups that would perturb
the validity of the Lambert-Beer law. Thus, the number of
termini calculated using the Lambert-Beer law confirms the
existence of the defects in the dendritic structures that is marked
for the high generations (Table 1). More specifically, these
defects are relatively not very numerous until generation 5, but
their number becomes higher when the generations 6 and 7 are
reached. The structural defects follow the same tendency,
presenting similar results independently of the nature of the
peripheral redox group (for example, G7-Fc and G7-Fc* both
present an experimental number of termini of 14 000 ( 1000).
The comparison of the three last generations of ferrocenyl
dendrimers in Figure 1 at the same concentration of dendrimer
shows the progression of the number of ferrocenyl termini from
each generation to the next.
3.2. Cyclic Voltammetry. All the cyclic voltammograms of
the metallodendrimers present a single reversible redox wave
at the same potential as that of the corresponding metallocene
(Figure 2).6,7 The single wave can be explained by the weak
electrostatic factor in dendrimers, the redox centers being very
far from one another. The electrochemical reversibility observed
for all the generations shows that the electron-hopping mech-
Figure 1. UV-vis spectra of G5, G6, and G7 of ferrocenyl dendrimers at
5 × 10-7 M in dichloromethane.
anism among the redox centers of the periphery is extremely
of mechanism in metallodendrimers has been proposed by
efficient. Thus, electron transfer is fast between all the redox
Amatore et al. for dendrimers containg 64 peripheral branches
groups of the dendrimers and the electrode, subsequent to
electron transfer at the optimum distance among the flexible
peripheral redox centers. A precise analysis of this later type
terminated by [Ru(terpy)2]2+. It is based on the measured
electron hopping rate constant using ultramicroelectrodes and
on a Smoluchowski-type model developed to take into account
viscosity effects during the displacement of the Ru(III)/
Ru(II)(tpy)2 redox centers around their equilibrium positions.16a,b
This analysis involves very fast scan rates, thus the events
proposed by Amatore et al. have time to proceed within the
time scale of the present studies.
In addition to electron hopping through space, the dendrimers
rotate fast within the electrochemical time scale, so that all the
redox groups come close to the electrode within this time
scale.6,7 Fast rotation alone can explain the fast heterogeneous
electron transfer with the small dendrimers. For the high
generations, however, the hopping mechanism between nearby
centers appears to be much easier than the fast-rotation
mechanism. In favor of this explanation, it is noteworthy that
with derivatized electrodes with the high-generation dendrimers,
complete rotation is no longer possible, but electron transfer is
still fast.
(15) Saji, T.; Hoshino, K.; Ishii, Y.; Goto, M. J. Am. Chem. Soc. 1991,
113, 450–456.
(16) (a) Amatore, C.; Grun, F.; Maisonhaute, E. Angew. Chem., Int. Ed.
2003, 40, 4944–4947. (b) Amatore, C.; Bouret, Y.; Maisonhaute, E.;
Goldsmith, J. I.; Abrun˜a, H. D. Chem.sEur. J. 2001, 7, 2206–2226.
(c) For electrochemical redox recognition with ferrocene systems, see:
Beer, P. D.; Gale, P. A. Angew. Chem., Int. Ed. 2001, 40, 486–506.
(17) For the first observation of the flattening of dendrimers on a surface,
see: (a) Hierlemann, A.; Campbell, J. K.; Baker, L. A.; Crooks, R. M.;
Rico, A. J. J. Am. Chem. Soc. 1998, 120, 5323–5324, see also. (b) Li,
J.; Piehler, L. T.; Qin, D., Jr.; Tomalia, D. A. Langmuir 2000, 16,
5613–5617. (c) Zhang, H.; Grim, P. C. M.; Vosch, T.; Wiesler, U.-
M.; Berresheim, A. J.; Mullen, K.; De Schryver, F. C. Langmuir 2000,
16, 9294–9298. (d) Coen, M. C.; Lorenz, K.; Kressler, J.; Frey, H.;
Mulhaupt, R. Macromolecules 2000, 29, 8069–8079. (e) Mecke, A.;
Lee, I.; Baker, J. R., Jr.; Holl, M. M. B.; Orr, B. G. Eur. Phys. J. E
2004, 14, 7–10.
(18) (a) Cheon, K.-S.; Kazmaier, P. M.; Keum, S.-R.; Park, K.-T.; Buncel,
E. Can. J. Chem. 2004, 551–556. (b) Liu, D.; De Feyter, S.; Cotlet,
M.; Stefan, A.; Wiesler, U.-M.; Herrman, A.; Grebel-Koehler, D.; Qu,
J.; Mu¨llen, K.; De Schryver, F. C. Macromolecules 2003, 16, 5918–
5928.
The ferrocenyl dendrimers present a single redox wave at
490 mV vs FeCp*2 in dichloromethane (electrolyte,
9
594 J. AM. CHEM. SOC. VOL. 131, NO. 2, 2009

Giant Dendritic Molecular Electrochrome Batteries
A R T I C L E S
Figure 3. (a) Modified Pt electrode with G7-Fc; (b) linearity of the intensity/
sweep rate function; (c) ATP2- sensing, and regeneration of the modified
electrode for reuse.
though there is no interaction site near the ferrocenyl group on
the dendrimer tethers to perturb the ferrocenyl group and modify
its redox potential.16c Indeed, the electrostatic interaction
between the multiply charged ferrocenium dendrimer and ATP2-
is large and sufficient by itself to significantly perturb the redox
system and decrease the ferrocenyl potential as can be seen in
Figure 3. After the experiment, [ATP][n-Bu4N]2 can be washed,
and the modified electrode can be used for another experiment.
The cyclic voltammogramms of the pentamethylferrocenyl
dendrimers were also recorded in dichloromethane, and they
were shown to present a single redox wave at 210 mV vs FeCp*2
(electrolyte, [n-Bu4N][PF6] 0.1 M; scan rate, 200 mV s-1; 20
°C; working and counter electrodes, Pt; quasi-reference elec-
trode, Ag). Also in this case, the number of electrons could be
calculated for G0-Fc* (9) but for G7-Fc* the absorption
phenomenon induces erroneous values using the Bard-Anson
equation.
3.3. Coulometry. Whereas the number of electrons for
metallodendrimers could not be calculated for the higher
generations due to the absorption phenomenon, the ferrocenyl
metallodendrimers were also characterized by coulometry for
all the generations, and the data provided satisfactory results in
terms of numbers of metallocene redox units in the dendrimers.
The total current Q (Q ) nFN) used to oxidize the dendrimers
can be compared with the calculated one (one electron per
ferrocene). The results found for the electrolysis of the two first
generations fit to the theoretical redox group number (98 and
99%), and they are higher than 90% until G5, showing the rather
good correspondence between the actual redox group numbers
and the theoretical ones. For the last generation (G7), the
Figure 2. Cyclic voltammograms obtained for the eight generations from
G0-Fc to G7-Fc using decamethylferrocene (FeCp*2) as an internal reference
(solvent, dichloromethane; electrolyte, [n-Bu4N][PF6] 0.1 M; scan rate, 200
mV s-1; 20 °C; working and counter electrodes, Pt; quasi-reference
electrode, Ag).
[n-Bu4N][PF6] 0.1 M; scan rate, 200 mV s-1; 20 °C; working
and counter electrodes, Pt; quasi-reference electrode, Ag). The
number of electrons involved in each redox process was
calculated using the Bard-Anson equation:19a,b for G0 (9
ferrocenyl termini), G1 (27 ferrocenyl termini), and G2 (81
ferrocenyl termini), the numbers 9, 27, and 103 electrons were
found, respectively. After G2, this kind of calculation is marred
by the adsorption of the ferrocenyl dendrimers onto the electrode
surface (for example: for G3 with 243 ferrocenyl termini and
G6 with 6561 ferrocenyl termini, numbers obtained are 700 and
30000 electrons respectively using the Bard-Anson equation).
The adsorption phenomenon becomes all the more important
as the generation is higher, and this can be evidenced by the
difference between the anodic and cathodic potential that
becomes lower (<60 mV) and lower as the generation number
increases (Figure 2). This pronounced adsorption of large
ferrocenyl dendrimers leads to the facile formation of derivatized
electrodes19c-e with such redox dendrimers upon dipping the
electrode into the solution of dendrimer and scanning a few
times around the region of the ferrocenyl redox potential. These
modified electrodes are actually very robust, and washing with
dichloromethane does not remove the dendrimer. Moreover,
recognition of ATP2- is possible as its n-Bu4N+ salt, even
9
J. AM. CHEM. SOC. VOL. 131, NO. 2, 2009 595

A R T I C L E S
Ornelas et al.
Table 2. Hydrodynamic Diameters Obtained by DLS and Calculated Diffusion Coefficients and Volumes of the Metallodendrimers
theoretical MW
(g mol-1
hydrodynamic
diameterc (nm)
diffusion
coefficientd (m2 s-1
series
producta
)
estimated MWb (g mol-1
)
)
volumee (m3)
densityf(Kg/m3)
Fc
G2-Fc (17)
G3-Fc (18)
G4-Fc (19)
G5-Fc (20)
G6-Fc (21)
G7-Fc (22)
G7-Fc* (30)
53 468
162 146
488 181
1 466 316
4 400 608
13 203 566
16 005 053
(5.2 ( 0.14) × 104
(1.54 ( 0.08) × 105
(4.60 ( 0.3) × 105
(1.35 ( 0.1) × 106
(4.0 ( 0.4) × 106
(1.0 ( 0.2) × 107
(1.2 ( 0.3) × 107
11.1 ( 0.8
16.6 ( 0.5
20.5 ( 0.4
24.3 ( 0.5
27.6 ( 0.5
29.5 ( 1.1
30.0 ( 1.4
8.76 × 10-14
6.57 × 10-14
4.63 × 10-14
3.89 × 10-14
3.52 × 10-14
3.29 × 10-14
3.24 × 10-14
(7.2 ( 1.4) × 10-25
(1.70 ( 0.15) × 10-24
(4.5 ( 0.2) × 10-24
(7.5 ( 0.4) × 10-24
(1.10 ( 0.10) × 10-23
(1.34 ( 0.20) × 10-23
(1.41 ( 0.20) × 10-23
(1.2 ( 0.1) × 102
(1.5 ( 0.2) × 102
(1.7 ( 0.2) × 102
(3.0 ( 0.3) × 102
(6.0 ( 0.6) × 102
(1.2 ( 0.2) × 103
(1.4 ( 0.2) × 103
Fc*
a It was not possible to obtain the hydrodynamic diameter of dendrimers of G0 and G1 by DLS, because their size is below the limit for this
technique. b Molecular weight calculated taking into account the defects estimated by UV-vis. and coulometry (2.5% for G2, 5% for G3, 6% for G4, 8%
for G5, 10% for G6,and 23% for G7). c Measured at 25 °C in dichloromethane; the hydrodynamic diameter value was obtained from statistical CONTIN
analysis with 2% error for G4, G5, and G6; the error is larger for the low generations, because their size is at or near the low limit for this technique, and
for G7, which contains many defects. d Calculated using the Stokes-Einstein equation. e Considering the globular dendrimer shape as a perfect sphere (V
) (4/3)πr3). f The density was calculated taking into account the average MW calculated by UV-vis and coulometry.
Table 3. Average Height Obtained by AFM on a Mica Surface for
Each Generation Gn of Ferrocenyl Dendrimers Gn-Fc.
dendrimer
no. of branches
height (nm)
G0-Fc (15)
G1-Fc (16)
G2-Fc (17)
G3-Fc (18)
G4-Fc (19)
G5-Fc (20)
G6-Fc (21)
G7-Fc (22)
9
27
81
243
729
2187
6561
19 683
0.8 ( 0.1
1.5 ( 0.2
2.5 ( 0.2
3.0 ( 0.3
3.5 ( 0.3
4.5 ( 0.4
6.0 ( 0.6
7.5 ( 1.5
electrolysis shows a difference of 17% between the theoretical
number of electrons and the obtained one. The results obtained
for the coulometry of the ferrocenyl dendrimers series are
gathered in Table 1.
3.4. Dynamic Light Scattering (DLS). Dynamic light scat-
tering (DLS)14 is a useful technique to characterize the metal-
lodendrimers, allowing to determine, or to estimate, the size
(hydrodynamic diameter) of the dendrimers in solution and
consequentlytheirdiffusioncoefficient(D)usingtheStokes-Einstein
equation: D ) kT/6πηRh, where the Rh is the hydrodynamic
radius, η is the solvent viscosity, k is the Boltzmann constant,
and T is the temperature. All the values of the hydrodynamic
diameter reported in Table 2 were found to be constant for three
different concentrations. These hydrodynamic diameter values
in this way usually overestimate the true dendrimer diameter,
because (i) the solvation sphere is included in this hydrodynamic
value, (ii) for the small dendrimers, the DSL technique is not
applicable, and (iii) for average generations such as G2, the
dendrimers are not fully spherical, and the technique reaches
its limit. The progression of the observed values when the
generation increases clearly shows the size progression, however.
It is known that dendrimers present globular shapes,1 and
they are usually considered as perfect spheres in order to
calculate their physical properties. We also used this approxima-
tion to estimate, the volume and the density of the metalloden-
drimers (Table 2). The difference between theoretical and actual
values becomes significant for G6 and especially G7 that present
many defects and whose molecular masses are much reduced
compared to the theoretical ones. There is no significant
difference in the diameter for both neutral series, as expected,
i.e., G7-Fc has an observed diameter of 29.5 ( 1.1 nm and G7-
Fc* has an observed diameter of 30.0 ( 1.4 nm.
Figure 4. Relationship between the (a) number of termini vs the generation
of dendrimer: y ) 5.15 + 8.56e(x/0.903), (R ) 1.00); (b) number of termini
vs the hydrodynamic diameter; (c) the hydrodynamic diameter vs the density
of the dendrimer: y ) 136 + 0.00619e(x/2.44), (R ) 0.996).
In these series of dendrimers, the number of branches is
multiplied by three at each generation and, according to Figure
4a, this increase perfectly fits the exponential function (with R
9
596 J. AM. CHEM. SOC. VOL. 131, NO. 2, 2009

Giant Dendritic Molecular Electrochrome Batteries
A R T I C L E S
Figure 5. AFM images obtained for the ferrocenyl dendrimers (Gn-Fc).
9
J. AM. CHEM. SOC. VOL. 131, NO. 2, 2009 597

A R T I C L E S
Ornelas et al.
Figure 9. Statistical distribution of the FeII and FeIII species on the mixed-
valence dendrimers. For instance, with G0 (9 terminal branches) the statistical
distribution obtained with a simple binomial law gives 24.6% of 4FeII-
5FeIII, 16.4% 3FeII-6FeIII, 7% 2FeII-7FeIII, 1.7% 1FeII-8FeIII, and 0.2% 0FeII-
9FeIII, etc.
Figure 6. UV-vis spectra of the oxidized G0-Fc in dichloromethane.
are extensively deformed from their globular shape when they
are neat, in contact with a surface presenting a flat disk.17 In
this work, AFM was used to study the behavior of the giant
ferrocenyl dendrimers on a mica surface. An ether solution of
the dendrimer was deposited on a mica surface, and the solvent
was evaporated before AFM observation. In spite of the
flattening of the dendrimers on a mica surface, it is possible to
observe a progression of the dendrimer size upon increasing
the generation number (Table 3). The height of dendrimers in
each generation was quite uniform in the monolayers, as can
be observed in Figure 5. Interestingly, the width of each AFM
spot is relatively regular suggesting agglomeration of the
metallodendrimers among one another and also the discrete
limits of this agglomeration (as for metallic nanoparticles) that
is rather regular from a dendrimer particle to the next. In other
words, on mica surface, the dendrimers behave like atoms in
forming relatively monodisperse nanoparticles. This feature
looks quite general, and is best observed for the dendrimers
below G7-Fc.
Figure 7. Representation of the “breathing mechanism” of the giant
metallodendrimers upon redox switching between the neutral orange form
and the cationic blue form.
4. Chemical Oxidation of the Fc and Fc* Series. The
ferrocenyl and pentamethylferrocenyl dendrimers were chemi-
cally oxidized and isolated and the dendritic polycations were
reduced, in order to confirm by NMR their redox stability after
a complete oxidoreduction cycle.
The ferrocenyl dendrimers were oxidized using a stoichio-
metric amount of acetylferrocenium tetrafluoroborate [Fc-
COMe][BF4],20 in dry dichloromethane at room temperature for
5 min. After the oxidized dendrimers were washed with ether
in order to remove the acetylferrocene formed, they were
isolated as blue solids and characterized by UV-vis spectros-
copy, showing an absorption band at 624 nm (Figure 6). The
reduction of the ferrocenium dendrimers was carried out with
an aqueous solution of TiCl3. The ferrocenyl dendrimers series
was thus shown to be highly stable in both the oxidized
ferrocenium and reduced ferrocene forms by quantitative
Figure 8. UV-vis spectrum of the oxidized G0-Fc* in dichloromethane.
) 1). The DLS data show that the hydrodynamic diameter
increases almost linearly with the number of termini until G5
(2187 branches). For the next generations, the increase of the
diameter seems to be less significant (Figure 4b), but the
accuracy of the technique is not sufficient for a conclusion
concerning this point. Figure 4c shows that the density increases
exponentially with the increase of the number of termini.
3.5. Atomic Force Microscopy (AFM). For the design of
nanodevices with optimal properties, it is essential to study the
structure and the behavior of dendrimers in the condensed state
in contact with surfaces subsequent to evaporation of the solvent.
Flat substrates such as mica have already been shown to be
useful models for this kind of interaction.17 Mica possesses a
negative surface charge density and is readily accessible to high-
resolution scanning probe techniques.17 Previous studies, using
AFM experiments, have already demonstrated that dendrimers
1
interconversion, their structure being confirmed by H NMR
after a complete oxido-reduction cycle.
(19) (a) Flanagan, J. B.; Margel, S.; Bard, A. J.; Anson, F. C. J. Am. Chem.
Soc. 1978, 100, 4248–4253. (b) Bard, A. J.; Faulkner, R. L.
Electrochemical Methods, 2nd ed.; Wiley: New York, 2001. (c)
Abruna, H. D. In ElectroresponsiVe Molecular and Polymer Systems;
Stotheim, T. A. Ed.; Dekker: New York, 1988; Vol. 1, p 97. (d)
Murray, R. D. In Molecular Design of Electrode Surfaces; Techniques
of Chemistry XII, Wiley: New York, 1992; p 1. (e) Yamada, M.;
Quiros, I.; Mizutani, J.; Kubo, K.; Nishihara, I. Phys. Chem. Chem.
Phys. 2001, 3, 3377–3383.
9
598 J. AM. CHEM. SOC. VOL. 131, NO. 2, 2009

Giant Dendritic Molecular Electrochrome Batteries
A R T I C L E S
Figure 10. EFM images of (a) solution of the oxidized dendrimer G5-Fc+ (left); (b) solution of equimolar G5-Fc and G5-Fc+ (right).
The oxidized form of G5-Fc was also characterized by AFM
in order to compare the size of the cationic form with that of
the neutral form. It was found that the average height obtained
for the G5-Fc+ (6.5 ( 0.6 nm) is much larger than that of its
neutral form (4.5 ( 0.4 nm). This experimental difference must
be considered with caution, because the type of force that is
measured for the neutral ferrocenyl dendrimer is different of
the one that is measured by EFM for the ferrocenium dendrimer.
On the other hand, this difference can be expected from the
repulsive forces between the charged termini in the oxidized
dendrimer. Also, in the solid state, the ferrocenium cation forms,
5. Mixed Valence in Redox Dendrimers. Mixed valence
complexes have been extensively investigated since Henry
Taube’s work because of their particular properties in molecular
electronics.21 Extension to multinuclear ferrocenyl complexes
and dendrimer chemistry was emphasized by Casado et al. with
mixed-valence dendrimers by cyclic voltammetry. In this case,
the short distance between the two ferrocenyl groups induced
an electrostatic factor and eventually an electronic communica-
tion between them, the superposition of these two significant
factors causing a potential difference with two distinct waves
in cyclic voltammetry.22 Our interest now is to discuss the
mixed-valence problem in dendrimers in which the redox termini
are almost independent because of the large number of σ-bonds
between them. Under these conditions, there is no through-bond
electronic communication between the redox centers. Also, the
redox potentials of the various redox centers are extremely close
to one another, because the electrostatic factor is almost nil.
What remains, however, is the entropic factor that was
highlighted by Suttin and Taube23 in the expression of the
different factors that are involved in mixed valency. This factor
is a statistic one, i.e., the various potentials of the redox sites
of a dendrimer are provided by a binomial law around a mean
value. This distribution is represented in Figure 9 for the half-
oxidized dendrimer G0-Fc containing nine ferrocenyl branches.
The ideal cyclic voltammogram is identical to that of a single
electron transfer in a monoredox-site compound as discussed
in the cyclic voltammetry section, and the redox potentials of
all the redox sites of the dendrimers are statistically distributed
along this cyclic voltammetry curve.19
with the anion BF4 , contact ion pairs22b located at the periphery
-
of the dendrimers. The counteranion adds bulk that still increases
the dendrimer size, which contributes to the observed large size
difference. Thus, these giant redox metallodendrimers exhibit
a “breathing mechanism” controlled by the redox potential
(Figure 7). In CH2Cl2, a solvent that favors ion pairing, this
breathing is accompanied by a displacement of the anions from
the oxidant toward all the redox termini of the dendrimer, and
the opposite is occurring upon reduction of the ferrocenium
dendrimer. In an aqueous solvent, the formation of solvent-
separated ion pairs would also be accompanied by a large size
increase upon oxidation because of the solvation of each
ferrocenium termini by several solvent molecules in addition
to increased repulsion among the peripheral positive charges.
The pentamethylferrocenyl dendrimers were oxidized using
a stoichiometric amount of ferrocenium hexafluorophosphate
[FeCp2][PF6],20a in dry dichloromethane at room temperature
for 5 min. After washing the oxidized dendrimers with ether in
order to remove the ferrocene formed, they were isolated as
green solids and characterized by UV-vis. spectroscopy,
showing two absorption bands at 624 and 745 nm (Figure 8).
The reduction of the pentamethylferrocenium dendrimers was
The oxidized dendrimer G0-Fc+ (blue) was mixed with a
stoichiometric amount of its G0-Fc in neutral form (orange) in
dichloromethane and the resulting solution of half-oxidized
presents a deep green color (eq. 11).
20a
carried out with an aqueous TiCl3 solution. The pentameth-
ylferrocenyl dendrimers series was thus shown to be highly
stable in both the oxidized and reduced forms by quantitative
interconversion, their structure being confirmed by H NMR
(21) Creuz, C.; Taube, H. J. Am. Chem. Soc. 1969, 91, 3988–3989.
(22) Taube, H. Angew. Chem., Int. Ed. 1984, 23, 329–339 (Nobel lecture).
(a) Cuadrado, I.; Casado, C. M.; Alonso, B.; Mo´ran, M.; Losada, J.;
Belsky, V. J. Am. Chem. Soc. 1997, 32, 7613–7614. (b) Loupy, A.;
Tchoubar, B.; Astruc, D. Chem. ReV. 1992, 92, 1141–1165.
(23) (a) Taube, H. Electron-Transfer Reactions of Complex Ions in Solution;
Academic Press: New York, 1970. (b) Astruc, D. Electron-Transfer
and Radical Processes in Transition Metal Chemistry; VCH: Wein-
heim, Germany, 1995; Chapter 1.
1
after a complete oxidoreduction cycle.
(20) (a) Connelly, N. G.; Geiger, W. E. Chem. ReV. 1996, 2, 877–910. (b)
Geiger, W. E. Organometallics 2007, 26, 5738–5765 (a historical and
perspective article)
9
J. AM. CHEM. SOC. VOL. 131, NO. 2, 2009 599

A R T I C L E S
Ornelas et al.
As described by Henry Taube23 when the two species FeII
and FeIII are together in solution, there is a fast electron
transfer among all the redox centers. Thus, applying this
statement to dendrimers, we assume that once both forms
are mixed in solution, a statistical distribution of mixed-
valence dendrimers is obtained with FeII and FeIII (eq 11 and
Figure 9). The presence of mixed-valence dendrimers in such
tended the redox properties of the ferrocenyl and pentameth-
ylferrocenyl complexes to the nanoscale. Thus, these nanomol-
ecules have potential applications in nanoelectronics, such as
molecular batteries, capacitors, anion sensors with modified Pt
electrodes and electrochromes, because they are highly colored
and able to very rapidly transfer a large number of electrons at
once (14 000 electrons for G7) (eqs 12 and 13).
Experimental Section
a mixture could be confirmed using the EFM technique
(electron force microscopy). With this technique, the nano-
molecules that possess a charge density can be distinguished.
Thus, a mixture containing equal molar amounts of G5-Fc
and its corresponding oxidized form G5-Fc+ were character-
ized by EFM, and it could be observed that all the nano-
molecules present in the images have a charge density with
a value, on average, of half (0.44) of the charge density of
a pure solution of G5-Fc+ (0.82) (all the analyzed spots
present, in both cases, similar heights). This is clear evidence
for the presence of mixed-valence dendrimers (Figure 10).
General Data. For general data including solvents, compounds
and reactions, NMR spectroscopy, MALDI TOF mass spectrometry,
cyclic voltammetry including modified Pt electrodes and elemental
analysis, see the Supporting Information.
Dynamic Light Scattering Measurements (DLS). The DLS
measurements were made using a Malvern Zetasizer 3000 HSA
instrument at an angle of 90°, in methylene chloride solution at 25
°C. Measurements were carried out at different concentrations until
the hydrodynamic diameter was found to be constant at three
different concentrations (for high concentration, higher hydrody-
namic diameter values were found because of aggregation).
Atomic Force Microscopy (AFM) and Electron Force
Microscopy (EFM) Experiments. AFM samples were prepared
by spin-coating of a suitable solution (adjusted on trial-error basis)
of dendrimer in diethyl ether. Before spin coating, the mica surface
was cleaved with Scotch tape. The freshly cleaved highly oriented
mica was covered with the solution and spinned at 1000 rpm with
subsequent 10-15 s extra spinning at 3000 rpm for complete drying
in air. The AFM apparatus is a Thermomicroscope CP Research
capable of measurements in multiple modes and the sample imaging
is done in air immediately after spin-coating. The noncontact or
NC dynamic mode was used giving the weakest interaction with
the surface and therefore the less chance of alteration. The
cantilever/tip systems used were either Ultralever or Nanosensors,
with typical tip (silicon tip) radius of curvature of 10 nm. The
amplitude and phase imaging of this mode were used. A sample
was considered as good when a monolayer of dendrimers was
measured by AFM.
Conclusion
The metallodendrimers reported here have high redox stability
and could be isolated and characterized in oxidized, reduced
and mixed-valence forms. Their size progression was followed
from the zeroth to the seventh generation by DLS, AFM,
UV-vis, coulometry, and cyclic voltammetry. UV-vis spec-
troscopy and coulometry show that the defect amount increases
as the generation number increases (as expected in classic
divergent construction) and reaches 20-30% of the theoretical
number of ferrocenyl termini at the seventh generation, i.e., the
largest dendrimers somewhat resemble hyperbranched polymers.
Finally, however, it has been possible to reach a number of
about 14000 ferrocenyl termini whereby, in spite of the strong
adsorption, cyclic voltammetry still shows fast electron transfer
because of an electron-hopping mechanism. The large amount
of defects in G7 does not perturb this electrochemical revers-
ibility. This strong adsorption of giant ferrocenyl dendrimers
should prove very useful for the design of recoverable redox
sensors. AFM shows an atom-like quite monodisperse aggrega-
tion and the breathing upon redox switching, whereas EFM
shows the charged and half-oxidized, mixed-valent systems. All
these synthesis, characterizations and stability tests have ex-
General Synthesis of Ferrocenyl Dendrimers. The iodometh-
ylsilane-dendrimer, 14 (2 equiv. per branch) and K2CO3 (10 equiv.
per branch) were introduced into a Schlenk flask, then dry DMF
(30 mL) was added, and the reaction mixture was heated at 80 °C
for 48 h under magnetic stirring. After the solvent was removed in
vacuo, dichloromethane was added and the mixture was filtered
on celite to remove K2CO3. The solvent was removed in vacuo
and the residue was washed with methanol (2 × 10 mL and
9
600 J. AM. CHEM. SOC. VOL. 131, NO. 2, 2009

Giant Dendritic Molecular Electrochrome Batteries
A R T I C L E S
precipitated with CH2Cl2/MeOH. After drying in vacuo, ferrocenyl
dendrimers were obtained as yellow waxy products. For details
concerning each generation, see the Supporting Information and
the example of G5-Fc below.
1.69 (s, 36H, CH2CH2CH2Si), 1.30 (m, 126H, (CH2)7), 1.25 (s, 36H,
CH2CH2CH2Si), 0.69 (s, 36H, CH2CH2CH2Si), 0.15 and 0.085 (d,
108H, Si(CH3)2). 13C NMR (CH3COCH3, 62 MHz), δ: 155.6 (arom.
Cq), 114.2 (arom. CH), 88.5 (Cq of Cp), 79.9 (Cq of Cp*), 71.4
and 70.8 (CH of Cp), 60.9 (OCH2Si), 44.0 (benzylic Cq of the core),
33.6 (CH2CH2CH2Si), 32.0 (CpCH2), 29.7 ((CH2)7), 23.7
(CH2CH2CH2Si), 13.8 (CH2CH2CH2Si), 11.2 (CH3 of Cp*), -4.64
(Si(CH3)2). 29Si NMR (CDCl3, 59.62 MHz), δ: 0.45 and 0.32
(Si(CH3)2). MS (MALDI-TOF; m/z): Calcd for C378H606O18Si18Fe9,
6647.1; found, 6450.30. Anal. Calcd for C378H606O18Si18Fe9: C,
70.42; H, 9.47. Found: C, 69.78; H, 9.83.
Synthesis of G5-Fc, 20. The G5-Fc dendrimer 20 was synthesized
from 2187-iodide dendrimer 7 (0.107 g, 0.000140 mmol), 14 (0.275
g, 0.614 mmol), and K2CO3 (0.392 g, 2.80 mmol) using the general
synthesis for ferrocenyl dendrimers. The product 20 was obtained
as a yellow waxy product in 77% yield (0.158 g, 0.000108 mmol).
1H NMR (CDCl3, 300 MHz), δ: 7.13 (d, arom), 6.80 (m, arom),
4.08 (t, Cp), 3.86 (t, CH2CH2O), 3.51 (s, inner SiCH2O), 3.45 (s,
outer SiCH2O), 2.31 (t, CH2Cp), 1.72 (t, CH2CH2O), 1.61 (s,
CH2CH2CH2Si), 1.29 (s, Cp(CH2)8), 1.13 (s, CH2CH2CH2Si), 0.56
(s, CH2CH2CH2Si), 0.033 (s, Si(CH3)2). 13C NMR (CDCl3, 62
MHz), δ: 159.0 (inner SiCH2OCAr), 155.7 (arom. CO(CH2)11), 153.0
(arom. COCH2Si), 139.2 (arom. Cq), 127.2 and 113.4 (arene CH
of the dendron), 115.3 and 114.8 (arom. CH), 89.5 (Cq of Cp),
68.4 and 68.0 (Cp), 67.0 (O(CH2)11), 60.9 (OCH2Si), 43.0
(CH2CH2CH2Si), 42.0 (benzylic Cq), 31.1 (OCH2CH2), 29.6
((CH2)8), 26.1 (CpCH2), 17.7 (CH2CH2CH2Si), 14.6
(CH2CH2CH2Si), -4.6 (Si(CH3)2). 29Si NMR (CDCl3, 59.62 MHz),
Synthesis of G7-Fc*, 30. The G7-Fc* dendrimer 30 was
synthesized from 19683-iodide dendrimer 9 (0.050 g, 7.9 × 10-6
mmol) using the general synthesis for the pentamethylferrocenyl
dendrimers. The dendrimer 30 was obtained as a brown waxy
1
product in 51% yield (0.073 g). H NMR (CDCl3, 300 MHz), δ:
6.86 (m, arom), 3.57 and 3.48 (m, Cp and SiCH2O), 2.16 (t,
CH2Cp), 1.87 (s, CH3 of Cp* and CH2CH2CH2Si), 1.25 (m, (CH2)7
and CH2CH2CH2Si), 0.64 (s, CH2CH2CH2Si), 0.075 (d, Si(CH3)2).
13C NMR (CH3COCH3, 62 MHz), δ: 155.6 (arom. Cq), 114.6 (arom.
CH), 88.4 (Cq of Cp), 79.8 (Cq of Cp*), 71.3 and 70.7 (CH of
Cp), 60.9 (OCH2Si), 33.6 (CH2CH2CH2Si), 32.0 (CpCH2), 29.7
((CH2)7), 23.7 (CH2CH2CH2Si), 13.8 (CH2CH2CH2Si), 11.1 (CH3
of Cp*), -4.7 (Si(CH3)2). 29Si NMR (CDCl3, 59.62 MHz), δ: 0.43
(Si(CH3)2). Anal. Calcd for C934893H1495830O49203Si49203Fe19683: C,
70.16; H, 9.42. Found: C, 67.78; H, 9.23.
δ:
86337H126768O5463Si3276Fe2187: C, 70.72; H, 8.71. Found: C, 70.09;
H, 9.01.
General Synthesis of Pentamethylferrocenyl Dendrimers. The
0.46
and
0.24
(Si(CH3)2).
Anal.
Calcd
for
C
iodomethylsilane dendrimer, 28 (2 equiv. per branch) and K2CO3
(10 equiv. per branch) were introduced into a Schlenk flask, then
dry DMF (30 mL) was added and the reaction mixture was heated
at 80 °C for 48 h under magnetic stirring. After removing the solvent
in vacuo, CH2Cl2 (20 mL) was added and the mixture was filtered
on celite to remove K2CO3. The solvent was removed in vacuo
and the residue was washed with methanol (2 × 10 mL) and
precipitated with CH2Cl2/MeOH. After being dried in vacuo,
pentamethylferrocenyl dendrimers were obtained as yellow waxy
products. For all details concerning the synthesis of 28 and its
precursors, see the Supporting Information.
Synthesis of G0-Fc*, 29. The G0-Fc* dendrimer 29 was
synthesized from 9-iodide dendrimer 2 (0.050 g, 0.022 mmol), 28
(0.224 g, 0.307 mmol) and K2CO3 (0.276 g, 1.97 mmol) using the
general synthesis for the pentamethylferrocenyl dendrimers. The
dendrimer 29 was obtained as a brown waxy product in 79% yield
(0.110 g). 1H NMR (CDCl3, 300 MHz), δ: 7.05 (s, 3H, CH core),
6.87 (d, 36H, arom), 3.62 and 3.53 (d, 36H, Cp), 3.54 and 3.51 (d,
36H, SiCH2O), 2.21 (t, 18H, CH2Cp), 1.90 (s, 135H, CH3 of Cp*),
Acknowledgment. Financial support from the Fundac¸a˜o para
a Cieˆncia e a Tecnologia (FCT), Portugal (PhD grant to CO), the
Institut Universitaire de France (IUF, DA), the University Bordeaux
1, the Centre National de la Recherche Scientifique (CNRS), and
the Agence Nationale de la Recherche (Project ANR-06-NANO-
026) is gratefully acknowledged.
Supporting Information Available: General data, description
of the electrochemical techniques, detailed experimental pro-
cedures of the syntheses and characterizations of the metallocene
precursors and metallocene dendrimers, and MALDI-TOF mass
and NMR spectra (PDF). This material is available free of
charge via the Internet at http://pubs.acs.org.
JA8062343
9
J. AM. CHEM. SOC. VOL. 131, NO. 2, 2009 601