Anchoring of rare-earth-based single-molecule magnets on single-walled carbon nanotubes

Article abstract of DOI:10.1021/ja906165e

A new heteroleptic bis(phthalocyaninato) terbium(III) complex 1, bearing a pyrenyl group, exhibits temperature and frequency dependence of ac magnetic susceptibility, typical of single-molecule magnets. The complex was successfully attached to single-wall

Full text of DOI:10.1021/ja906165e

Published on Web 10/02/2009
Anchoring of Rare-Earth-Based Single-Molecule Magnets on
Single-Walled Carbon Nanotubes
Svetlana Kyatskaya,† Jose´ Ramo´n Gala´n Mascaro´s,‡ Lapo Bogani,*,§,|
Frank Hennrich,† Manfred Kappes,† Wolfgang Wernsdorfer,§ and Mario Ruben*,†
Institut fu¨r Nanotechnologie, Karlsruhe Institute of Technology (KIT), D-76021 Karlsruhe,
Germany, Institute of Chemical Research of Catalonia (ICIQ), AV. Països, Catalans 16,
E-43007, Tarragona, Spain, 1. Physikalisches Institut, UniVersita¨t Stuttgart, Pfaffenwaldring 57,
D-70550, Stuttgart, Germany, and Institut Ne´el & UniVersite´ J. Fourier, CNRS, Grenoble 25,
AV. des Martyrs, F-38042, Grenoble, France
Received January 7, 2009; E-mail: lapo.bogani@pi1.physik.uni-stuttgart.de; mario.ruben@int.fzk.de
Abstract: A new heteroleptic bis(phthalocyaninato) terbium(III) complex 1, bearing a pyrenyl group, exhibits
temperature and frequency dependence of ac magnetic susceptibility, typical of single-molecule magnets.
The complex was successfully attached to single-walled carbon nanotubes (SWNTs) using π-π interactions,
yielding a 1-SWNT conjugate. The supramolecular grafting of 1 to SWNTs was proven qualitatively and
quantitatively by high-resolution transmission electron microscopy, emission spectroscopy, and atomic force
spectroscopy. Giving a clear magnetic fingerprint, the anisotropy energy barrier and the magnetic relaxation
time of the 1-SWNT conjugate are both increased in comparison with the pure crystalline compound 1,
likely due to the suppression of intermolecular interactions. The obtained results propose the 1-SWNT
conjugate as a promising constituent unit in magnetic single-molecule measurements using molecular
spintronics devices.
Introduction
temperatures. The lack of hysteresis might be due to the fact
that the electrons passing through the molecule perturb the
The development of new molecules and hybrid materials
suited to single-molecule studies is a topic of considerable
current interest. This is particularly true for the field of molecular
spintronics, which holds great promise1 but needs tailored
magnetic molecules. One family of magnetic molecules that
seems particularly promising is that of single-molecule magnets
(SMMs), molecular complexes and clusters that display slow
dynamics of the magnetization at low temperatures and an
impressive array of quantum features. Combining SMMs with
the domain of spintronics would not only allow exploiting these
new properties but would also mean having unprecedented tools
to study the fundamental processes governing magnetization
dynamics.1b The field of molecular spintronics has thus received
growing attention from the scientific community, and several
approaches to its fundamental issues can be envisaged. One
consists of immobilizing single magnetic molecules between
two leads (or, respectively, between tip and surface) and then
passing a current through the magnet. Although this approach
has allowed the observation of several key issues, it has been
impossible, up to now, to establish the magnetic bistability of
individual SMMs, i.e., the hysteresis cycle opening at low
magnetization state of the molecules. In order to avoid the
difficulties linked to the passage of electrons through the mol-
ecule, a different scheme has been envisaged, that of a so-called
“spintronics double quantum dot”.1b In this scheme, the SMM,
which can be considered as a magnetic quantum dot, is weakly
coupled to a molecular single-electron transistor, being the other
molecular quantum dot. The magnetization of the molecule can
then be read by its effect on the electron flow through the
molecular quantum dot, with only weak perturbation of the
SMM itself. Extensions of this concept include the recently
developed nano-SQUID (superconducting quantum interference
device),2 whose high-sensitivity is due to the presence of a
single-walled carbon nanotube (SWNT) in the loop. It is
expected that such devices, as the SWNT cross-section is
comparable to the dimensions of SMMs, will allow the detection
of the single-molecule magnetic moments of a single molecule,
an ambitious goal since Ne´el’s pioneering work.3 The switching
of the magnetization of single molecules could also be detected
using the nanomechanical properties of a suspended nanotube,
which could act as a nanometer-sized vibrating sample mag-
netometer. The motion of the nanotube could be studied by the
current through the nanotube itself. However, several challenges
still remain to be solved before such spintronic schemes can be
experimentally implemented. One is the fabrication of the double
quantum dot device. While synthetic methods to graft nano-
† Karlsruhe Institute of Technology.
‡ Institute of Chemical Research of Catalonia.
§ Institut Ne´el & Universite´ J. Fourier, CNRS.
| Universita¨t Stuttgart.
(1) (a) Wolf, S. A.; Awschalom, D. D.; Buhrman, R. A.; Daughton, J. M.;
von Molna´r, S.; Roukes, M. L.; Chtchelkanova, A. Y.; Treger, D. M.
Science. 2001, 294, 1488–1495. (b) Bogani, L.; Wernsdorfer, W. Nat.
Mater. 2008, 7, 179–186. (c) Camarero, J.; Coronado, E. J. Mater.
Chem. 2009, 19, 1678–1684. (d) Ferrer, J.; Garc´ıa-Sua´rez, V. M. J.
Mater. Chem. 2009, 19, 1696–1717.
(2) Cleuziou, J.-P.; Wernsdorfer, W.; Bouchiat, V.; Ondarc¸uhu, T.;
Monthioux, M. Nat. Nanotechnol. 2006, 1, 53–59.
(3) (a) Ne´el, L. Ann. Geophys. (C.N.R.S.) 1949, 5, 99–136. (b) Ne´el,
L. C. R. Proc. Natl. Acad. Sci. U.S.A. 1949, 228, 664.
9
10.1021/ja906165e CCC: $40.75  2009 American Chemical Society
J. AM. CHEM. SOC. 2009, 131, 15143–15151 15143

A R T I C L E S
Kyatskaya et al.
particles or photoactive molecules onto SWNTs have been
developed in the last few years, only two reports have very
recently appeared on grafting SMMs onto SWNTs,4 and many
questions remain open: What groups allow producing a sizable
magnetic interaction without destroying the conductivity of the
SWNTs? Are the magnetic properties retained when the grafting
is performed? This last question is particularly important, as it
has been shown that environmental effects are fundamental in
SMMs and can lead to the progressive disappearance of the
hysteresis when the steric or electronic influence of a substrate,
e.g. of conducting surfaces, becomes considerable.5 Only
recently the persistence of SMM behavior in a monolayer of
tetranuclear Fe(III) complexes bonded to a substrate6 as well
as Fe6 wolframate molecules grafted on the side walls of
SWNTs4b could be demonstrated. The second great challenge
in the double quantum dot implementation is the characterization
of the SMM behavior with the device. One of the great scientific
challenges using these devices would be the possibility of
probing a single, isolated molecule. However, as mentioned
before,5 environmental interaction and hybridization effects can
exercise dramatic consequences on the magnetic properties,
possibly yielding different results for SWNT-SMM or
substrate-SMM conjugates than those observed for isolated
SMMs in the highly homogeneous environment of a bulk 3D
crystal. While tackling the challenge of single-molecule mag-
netism is certainly fascinating, the requirement for SMM
conjugates eliminates an important reference point, posing a
big experimental problem: in the conjugate device situation, how
can one be sure that he/she is really probing the right molecule,
i.e., respectively, the (magnetic) property of the molecule? A
robust magnetic fingerprint for the molecule becomes an
indispensable requirement for such cutting-edge experiments.
significantly higher than that of all other known SMMs.10-14
Moreover, Kramer’s theorem of spin parity provides a finger-
print of the magnetic behavior of these systems. The theorem
states that no matter how asymmetric the crystal field is, a
system possessing an odd number of electrons must have a
ground state that is at least doubly degenerate, even in presence
of crystal field or spin-orbit interactions. This means that no
avoided level crossing, and thus no zero field quantum tunnel-
ling, can be present for half-integer spins. In the case of rare-
earths, due to strong hyperfine interactions, the electronic spin
number J is not a good quantum number to calculate parity
effects and the sum of nuclear and electronic spin numbers is
more appropriate. The quantum tunnelling of the magnetization,
in particular, shows zero-field tunnelling only for rare-earths
whose summed electronic and nuclear moments give an integer
spin. Such effects are extremely robust, as they depend on the
intrinsic symmetry of the environment, and not on possible
ligand field distortions or electronic perturbations. They con-
stitute a perfect example of the desired fingerprint of the
magnetic behavior. The case of Tb(III) is of particular interest,
as hyperfine and nuclear quadrupole interactions need to
be considered to explain the magnetic behavior of [Pc2Tb]-/0/+
compounds. Also for the terbium complexes, the out-of-phase
component of the dynamic magnetic susceptibility appears well
above 40 K, due to an unusually high-energy barrier for reversal
of the magnetization, ∆E. ∆E can be increased even further by
oxidizing the ligand from the anion7,9,14 [Pc2Tb]- to the neutral
form8 [Pc2Tb]0 and even further on to the cation15 [Pc2Tb]+.
Such striking magnetic properties are coupled to a very robust
single metal ion structure, which is expected to retain its
magnetic properties in different environmental conditions, unlike
that of many of the structurally more complex d-metal
SMMs.10-13 Thus, the electronic structure of isolated bis(ph-
thalocyaninato) terbium(III) molecules, supported on the Cu(111)
surface, has been characterized by density functional theory and
scanning tunneling spectroscopy. These studies reveal that the
interaction with the metal surface preserves both the molecular
structure and the large spin magnetic moment of the metal center
on metallic surfaces.16 Furthermore, Kondo-physics behavior
was reported for molecules deposited on a Au(111) surface,
giving evidence for retained magnetic behavior.17
In this paper, we describe the creation and magnetic char-
acterization of a hybrid conjugate composed of SWNTs and a
Tb(III)-based single metal ion SMM, which constitutes a
solution to some of these problems. The SMM behavior of
mononuclear rare-earth (M) complexes in a low symmetry
environment was first reported by Ishikawa and co-workers7-9
on bis(phthalocyaninato) complexes [Pc2M]-/0/+ (Pc ) anion
of phthalocyanine) of Tb(III), Dy(III), and Ho(III).
This class of complexes shows large magnetic anisotropies
and slow relaxation of the magnetization in a temperature range
So far the unsubstituted complexes [Pc2Tb]0 afford no
molecular ordering on pyrolitic graphite or metallic surfaces
(4) (a) Bogani, L.; Danieli, C.; Biavardi, E.; Bendiab, N.; Barra, A.-L.;
Dalcanale, E.; Wernsdorfer, W.; Cornia, A. Angew. Chem., Int. Ed.
2009, 48, 746–750. (b) Giusti, A.; Charron, G.; Mazerat, S.; Compain,
J.-D.; Mialane, P.; Dolbecq, A.; Rivie`re, E.; Wernsdorfer, W.; Biboum,
R. N.; Kieta, B.; Nadjo, L.; Filoramo, A.; Bourgoin, J.-P.; Mallah, T.
Angew. Chem., Int. Ed. 2009, 48, 4949–4952.
(10) Gatteschi, D.; Sessoli, R.; Villain, J. Molecular Nanomagnets; Oxford
University Press: New York, 2006.
(11) Sessoli, R.; Tsai, H.-L.; Schake, A. R.; Wang, S.; Vincent, J. B.;
Folting, K.; Gatteschi, D.; Christou, G.; Hendrickson, D. N. J. Am.
Chem. Soc. 1993, 115, 1804–1816.
(5) (a) Bogani, L.; Cavigli, L.; Gurioli, M.; Novak, R. L.; Mannini, M.;
Caneschi, A.; Pineider, F.; Sessoli, R.; Clemente-Leo´n, M.; Coronado,
E.; Cornia, A.; Gatteschi, D. AdV. Mater. 2007, 19, 3906–3911. (b)
Mannini, M.; Sainctavit, P.; Sessoli, R.; Moulin, C. C. D.; Pineider,
F.; Arrio, M. A.; Cornia, A.; Matteschi, D. Chem.sEur. J. 2008, 14,
7530–7435.
(12) Zheng, Y.-Z.; Lan, Y.; Anson, C. E.; Powell, A. K. Inorg. Chem. 2008,
47, 10813–10815.
(13) Cangussu, D.; Pardo, E.; Dul, M.-C.; Lescouzec, R.; Herson, P.;
Journaux, Y.; Pedroso, E. F.; Pereira, C. L. M.; Stumpf, H. O.; Muo`oz,
M. C.; Ruiz-Garc´ıa, R.; Cano, J.; Julve, M.; Lloret, F. Inorg. Chim.
Acta 2008, 361, 3394–3402.
(6) Mannini, M.; Pineider, F.; Sainctavit, P.; Danieli, C.; Otero, E.;
Sciancalepore, C.; Talarico, A. M.; Arrio, M.-A.; Cornia, A.; Gatteschi,
D.; Sessoli, R. Nat. Mater. 2009, 8, 194–197.
(14) (a) Ishikawa, N.; Sugita, M.; Okubo, T.; Tanaka, N.; Lino, T.; Kaizu,
Y. Inorg. Chem. 2003, 42, 2440–2446. (b) Branzoli, F.; Carretta, P.;
Filibian, M.; Zoppellaro, G.; Graf, M. J.; Gala´n-Mascaro´s, J. R.; Fuhr,
O.; Brink, S.; Ruben, M. J. Am. Chem. Soc. 2009, 131, 4387–4396.
(15) Takamatsu, S.; Ishikawa, T.; Koshihara, S.; Ishikawa, N. Inorg. Chem.
2007, 46, 7250–7252.
(7) (a) Ishikawa, N.; Sugita, M.; Ishikawa, T.; Koshihara, S.; Kaizu, Y.
J. Am. Chem. Soc. 2003, 125, 8694–8695. (b) Ishikawa, N.; Sugita,
M.; Wernsdorfer, W. J. Am. Chem. Soc. 2005, 127, 3650–3651. (c)
Ishikawa, N.; Sugita, M.; Ishikawa, T.; Koshihara, S.; Kaizu, Y. J.
Phys. Chem. B 2004, 108, 11265–11271.
(16) Vitali, L.; Fabris, S.; Mosca Conte, A.; Brink, S.; Ruben, M.; Baroni,
S.; Kern, K. Nano Lett. 2008, 8, 3364–3368.
(8) Ishikawa, N.; Sugita, M.; Tanaka, N.; Ishikawa, T.; Koshihara, S.;
Kaizu, Y. Inorg. Chem. 2004, 43, 5498–5500.
(17) Katoh, K.; Yoshida, Y.; Yamashita, M.; Miyasaka, H.; Breedlove,
B. K.; Kajiwara, T.; Takaishi, S.; Ishikawa, N.; Isshiki, H.; Zhang,
Y. F.; Komeda, T.; Yamagishi, M.; Takeya, J. J. Am. Chem. Soc. 2009,
131, 9967–9976.
(9) Ishikawa, N.; Sugita, M.; Wernsdorfer, W. Angew. Chem., Int. Ed.
2005, 44, 2931–2935.
9
15144 J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009

Anchoring of Single-Molecule Magnets on SWNTs
A R T I C L E S
but only the instantaneous formation of molecular aggregates.18
Thus, a chemical engineering of the complex is necessary to
obtain hybrid SWNT-SMM materials. The rich and well-
studied chemistry of phthalocyanines allows adding tailored
functionalities to these SMMs, so as to organize them in hybrid
structures18 while the SMM behavior is maintained. Thus,
[Pc2Tb]-/0/+ family constitutes a very interesting class of
compounds that possess optimal characteristics for grafting onto
SWNTs.
Here we report the synthesis of a new pyrenyl-substituted
heteroleptical bis(phthalocyaninato)Tb(III) SMM (1), which
constitutes an appealing candidate for SMM-SWNT hybrids.
We demonstrate that this molecule can be subsequently confined
on to sidewalls of SWNTs under the persistence of pivotal
magnetic properties, some characteristics even being improved.
Among the chemical approaches for the grafting of organic
molecules to SWNTs, we have chosen the functionalization by
a noncovalent interaction via pyrene linkers to avoid possible
insertion of local defects by covalent bonding that may disrupt
the SWNT conducting properties.19-21 Moreover, recently,
successful grafting of a compound of the Fe4 SMM family
containing a pyrene moiety on to CNTs has been reported.4a
125.02, 125.12, 125.87, 126.71, 127.23, 127.38,127.51, 128.62,
129.92, 130.90, 131.43, 136.08. MALDI-ToF calcd for C20H17Br:
336.0508. Found: 335.7416.
Synthesis of the Asymmetrically Substituted Phthalocyanine
A3B (9). Lithium metal (35 mg, 5.04 mmol) was dissolved in
1-pentanol (25 mL) at 80 °C under an atmosphere of dry N2. To
this lithium pentanolate solution were added 7 (1.15 g, 3.88 mmol)
and 8 (0.13 g, 0.32 mmol), and the reaction mixture was heated at
135 °C for 3 h. On cooling, the dark blue-green solution was treated
with glacial acetic acid (25 mL). The resultant precipitate was
collected by filtration; washed thoroughly with water, methanol,
and Et2O; and then dried on air to yield a mixture (by MALDI-
ToF, main components, A3B, A4; minor components, A2B2, AB3,
B4) as a dark blue-green powder (630 mg). The product mixture
was dissolved in a minimum amount of chloroform, and the solution
was subjected to column chromatography (SiO2), eluting with
CHCl3. The first fraction was collected. After removal of the solvent
under reduced pressure, the blue residue was dried at 50 °C to afford
0.34 g (47%) of 10 (Rf ) 0.78; chloroform/hexane, 5/3 vv). The
second fraction (representing a mixture of two compounds with Rf
) 0.78 and 0.25) was initially applied to a radial chromatograph
eluting with chloroform/hexane (5/3 vv) to remove a trace amount
of A4 and then separated with chloroform to afford after removal
of the solvent under reduced pressure and desiccation at 50 °C a
green powder, phthalocyanine 9 (98 mg, 24%, Rf ) 0.25;
chloroform/hexane, 5/3 vv).
Experimental Section
A3B (9). 1H NMR (CDCl3; δ, ppm): 0.89-1.12 (m, 20 H), 1.32
(s, 12 H,), 1.53-2.34 (m, 54 H), 3.61 (m, 2 H), 7.57-8.39 (m, 17
H); 8.45 (d, J ) 9 Hz, 1 H). MALDI-ToF calcd for C88H106N8O:
1291.8516. Found: 1290.0271.
General Synthetic Remarks. Reactions requiring an inert gas
atmosphere were conducted under argon, and the glassware was
oven-dried (140 °C). All reagents were purchased from commercial
sources and used as received. Radial chromatography was per-
formed on Chromatotron 7924T, with plates prepared from Silica
gel 60PF254 containing gypsum. PcLi2 (4),22 1,2-dicyano-4,5-
dihexylbenzene (7),23 and 1-pyrenebutanol (12)24 were prepared
according to the literature procedures.
Synthesis of 1-Bromobutylpyrene (13). 1-Bromobutylpyrene
was obtained by a variation of the reported procedure.25 To a
solution of 1-pyrenebutanol 12 (2.32 g, 8.5 mmol) in dry benzene
(50 mL) phosphorus tribromide (1.16 g, 4.3 mmol) was slowly
added at 5 °C (ice bath). After the addition was complete, the
reaction was kept at room temperature until the starting compound
12 disappeared (ca. 1.5 h, TLC control). The mixture was poured
into ice-water (200 mL) and then extracted with diethyl ether (3
× 100 mL). The combined organic layers were washed twice with
water and dried over magnesium sulfate. The solvent was removed
under reduced pressure, and the residue was chromatographed on
SiO2 to give 1-bromobutylpyrene 13 (1.3 g, 45% yield), which was
recrystallized from light petroleum. Mp: 85-89 °C. 1H NMR
(CD2Cl2; δ, ppm): 1.97-2.05 (m, 4 H), 3.34-3.39 (m, 2 H),
3.45-3.50 (m, 2 H), 7.87 (d, J ) 9 Hz, 1 H), 7.96-8.03 (m, 3 H);
8.06-8.11 (m, 4H), 8.14 (d, J ) 9 Hz, 1 H). 13C NMR (CDCl3; δ,
ppm): 30.23, 32.62, 32.65, 33.66, 123.26, 124.78, 124.84, 124.96,
A4 (10). 1H NMR (CDCl3; δ, ppm): 0.87-1.12 (m, 24 H), 1.32
(s, 16 H,), 1.53-2.34 (m, 64 H), 3.31 (br. s, 2 H-N), 8.65 (br s,
8 H). MALDI-ToF calcd for C80H112N8: 1184.9004. Found:
1183.9338.
Synthesis of 4-(4-Pyren-1-ylbutoxy)phthalonitrile (8). 8 was
obtained according to the reported procedure26 with some modifica-
tion. A mixture of 1-bromobutylpyrene (13) (1.07 g, 3.2 mmol),
4-hydroxyphthalonitrile (14) (0.57 g, 3.9 mmol), and K2CO3 (2.09
g) in the mixture of DMSO and acetone (15 mL, 15 mL) in two-
necked round-bottom flask was heated at 60 °C overnight until the
starting compound 13 disappeared (TLC control). The reaction
mixture was poured into 200 mL of ice-water, and the precipitate
formed was collected by filtration and washed with water until pH
7 and than with methanol and ether. Crude product was recrystal-
lized from acetone/1,4-dioxane mixture to afford 1.15 g (77%) of
1
8 as yellowish powder. Mp: 154-157 °C. H NMR (CDCl3; δ,
ppm): 1.95-2.105 (m, 4 H), 3.42 (t, J ) 8 Hz, 2 H), 3.92 (t, J )
1
2
6 Hz, 2 H), 6.93 (dd, J ) 9 Hz, J ) 2 Hz, 1 H), 7.05 (d, J ) 2
Hz, 1 H); 7.47 (d, J ) 9 Hz, 1 H), 7.85 (d, J ) 8 Hz, 1H), 7.97 (d,
J ) 7 Hz, 1 H), 8.01 (br s, 2 H), 8.06-8.10 (m, 2 H), 8.14-8.17
(m, 2 H), 8.22 (d, J ) 8 Hz). 13C NMR (CDCl3; δ, ppm): 27.83,
29.00, 33.50, 69.19, 107.33, 117.57, 119.45, 119.64, 123.54, 125.17,
125.23, 125.54, 125.29, 126.41, 127.24, 127.69, 127.77, 127.84,
128.97, 130.37, 131.17, 131.78, 135.30, 136.24, 162.22. MALDI-
ToF calcd for C28H20N2O: 400.1570. Found: 399.8239.
(18) (a) Go´mez-Segura, J.; D´ıez-Pe´rez, I.; Ishikawa, N.; Nakano, M.;
Veciana, J.; Ruiz-Molina, D. Chem. Commun. 2006, 2866–2868. (b)
Ye, T.; Takami, T.; Wang, R.; Jiang, J.; Weiss, P. S. J. Am. Chem.
Soc. 2006, 128, 10984–10985. (c) Su, W.; Jiang, J.; Xiao, K.; Chen,
Y.; Zhao, Q.; Yu, G.; Liu, Y. Langmuir 2005, 21, 6527–6531.
(19) Bogani, L.; Wernsdorfer, W. Nat. Mater. 2008, 7, 179–186.
(20) Bogani, L.; Wernsdorfer, W. Inorg. Chim. Acta 2008, 361, 3807–
3819.
Synthesis of Lithium Phthalocyanine [A3B-Li2] (2). Under
an argon atmosphere, 85.7 mg (0.066 mmol) of phthalocyanine A3B
(9) was added to a solution of Li (0.12 mg, 0.016 mmol) in ethyl
alcohol (10 mL). A green color appeared, and when the mixture
was warmed, the color changing to deep blue. The mixture was
refluxed for 2 h and cooled, the solvent was removed under reduced
pressure, and the residue of dull blue lithium phthalocyanine was
extracted with acetone, previously dried over sodium sulfate, to
remove basic lithium compounds. Evaporation under reduced
pressure and desiccation under an argon atmosphere left dilithium
phthalocyanine 2 as a crystalline purple powder (yield, 67 mg;
(21) Ga˜mez-Navarro, C.; De Pablo, P. J.; Ga˜mez-Herrero, J.; Biel, B.;
Garcia-Vidal, F. J.; Rubio, A.; Flores, F. Nat. Mater. 2005, 4, 534–
539.
(22) Barrett, P. A.; Frye, D. A.; Linstead, R. P. J. Chem. Soc. 1938, 1157–
1163.
(23) Hanack, M.; Haisch, P.; Lehmann, H.; Subramanian, L. R. Synthesis
1993, 4, 387–390.
(24) Yamana, K.; Letsinger, R. L. Nucleic Acids Symp. Ser. 1985, 169,
72.
(25) Leardini, R.; Nanni, D.; Peduli, G.; Tundo, A.; Zanardi, G.; Foresti,
E.; Palmieri, P. J. Am. Chem. Soc. 1989, 111, 7723–7739.
(26) Mutsumi, K.; Takahisa, K.; Kazuchika, O.; Kenji, H.; Hirofusa, S.;
Nagao, K. Langmuir 2003, 19, 4825–4830.
9
J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009 15145

A R T I C L E S
Kyatskaya et al.
78%). The compound is unstable and has to be used for the next
step during the next few days. H NMR (acetone-d6; δ, ppm):
0.89-1.12 (m, 20 H), 1.32 (s, 12 H), 1.53-1.74 (m, 54 H), 3.21
(br s, 8 H), 4.63 (br s, 2 H), 7.58 (br s, 1 H), 8.0-8.31 (m, 17 H),
8.59 (d, J ) 9 Hz, 1 H), 9,14 (br s, 8 H). MALDI-ToF calcd for
C88H104N8OLi2: 1301.7491. Found: 1302.8651.
even after the first purification cycle a small quantity of catalyst
could still be detected, the gel filtration was performed twice. No
out-of-phase-signal in magnetic susceptibility measurements could
be detected after the second purification cycle.
1
Preparation of the 1-SWNT Hybrids. The SWNT-SCholate
suspension was diluted with acetone, and the SWNTs were collected
by filtration and washed extensively with water to remove the
cholate and then dried. The obtained SWNTs (2.5 mg) were
redispersed in a solution of 3.3 mg of 1 in 5 mL of dry and argon-
saturated CH2Cl , and the reaction mixture was stirred for 48 h;
whereupon, the2 green color of supernatant disappeared. The
resulting mixture was sonicated for 6 h at 20 °C followed by
centrifugation for 2 h. The precipitate was collected by centrifuga-
tion, washed several times with CH2Cl2, and desiccated under a
slow stream of argon. A 4.9 mg portion of the hybrid material was
used for ac susceptibility measurements.
Spectroscopic Characterization. UV/visible/NIR spectra were
measured on a Varian Cary 500 Scan UV/vis/NIR spectrophotom-
eter by using a 1 cm optical path length quartz cell. Fluorescence
spectra were measured on a Varian Cary Eclipse Fluorescence
spectrophotometer by using a 1 cm optical path length quartz cell.
MALDI-ToF spectra were recorded on a Perseptive Biosystems
Voyager-DePro. NMR spectra were recorded at room temperature
on a Bruker 300 Ultrashield series instrument. Melting points were
measured with Bu¨chi B-545 using open-ended glass capillaries.
Electrochemistry. The cyclic voltammogram was recorded on
an EG&G Model 263A potentiostat using a three-electrode system
in solvents containing 0.1 M n-Bu4NClO4 as the supporting
electrolyte. A platinum button electrode was used as the working
electrode. A platinum wire served as the counter electrode, and
Ag/AgCl was used as the reference electrode. The ferrocene/
ferrocenium (Fc/Fc+) redox couple was used as an internal
standard.
Synthesis of the Heteroleptic Bis(phthalocyaninato)terbium(III)
Complex (1). Under a slow stream of Ar, a mixture of PcLi2 (4)
(68 mg, 0.13 mmol) and Tb(acac)3 ·2H2O (5) (59 mg, 0.13 mmol)
in 1-chloronaphtalene (5 mL), percolated through a basic alumina
column just before being used, was heated at 185-195 °C for 1 h
until free 4 was no longer detected. The resulting dark blue solution
was cooled down and 2 (160 mg, 0.11 mmol) was added. The
mixture was heated up to 200-210 °C for 1 h until free 2 was no
longer detected (MALDI-ToF control). The mixture was subjected
to column chromatography (basic alumina oxide, 60 g), eluting with
CH2Cl2/hexane (7/4 vv). 1-Chloronaphtalene was eluted first and
then a greenish-blue band was collected and concentrated to yield
a product mixture (by MALDI-TOF, main components are [A3B-
TbPc]0/-, (A3B)Pc2Tb2] and minor components are [A3B-Tb-A3B]0/
-, [PcTbPc]0/-) as a dark blue-green powder (150 mg). The complex
1 was separated from the crude mixture by column chromatography
(basic alumina oxide), eluting with CH2Cl2/MeOH (10:1) followed
by reprecipitation from the hexane/CH2Cl mixture to afford a deep
green solid of neutral 1 (87 mg, 24% yield f2rom 2; Rf ) 0.38; CH2Cl2).
MALDI-ToF calculated for C120H120N16OTb (M+): 1960.9110. Found
1959.3799 (M+, 100%).
The symmetrical complex 6 with two pyrene anchors was
achieved by eluting with CH2Cl2/MeOH (2:1) mixture as a green
compound (9 mg, Rf ) 0.56; CH2Cl2/MeOH (10:3)). MALDI-ToF
calcd for C176H208N16O2Tb (M+): 2735.1264. Found 2736.9281 (M+,
100%).
Purification of Carbon Nanotubes. Typically, 10 mg of SWNT
material (HiPco, Unidym) was suspended in 25 mL of H2O with 1
wt % of SCholate using a tip sonicator (Bandelin, 200 W maximum
power, 20 kHz, in pulsed mode with 100 ms pulses) applied for
2 h at ∼20% power. A density gradient centrifugation (DGC)27
step was then used to remove larger agglomerates, amorphous
carbon, and catalyst particles while retaining SWNTs. However,
rather than preformed gradients, we used self-generated gradients
of Iodixanol plus 1 wt % of SCholate. Ultracentrifugation (Optima
Max-E, Beckman-Coulter) was carried out in a fixed-angle rotor
(ML-80, Beckman-Coulter) at 15 °C and at 45 000 rpm for 18-20
h using a polyallomer (8 mL Bell top Quick-Seal, Beckmann-
Coulter) centrifuge tubes. This rotation speed results in centripetal
accelerations of 103 650g and 140 400g, at the average/maximal
radii of 45.7/61.9 mm, respectively. After centrifugation, different
colored regionssapproximately 5 mL of the whole 8 mL from
bottom to topsare visible. Of these, only the last (i.e., topmost)
∼0.5 mL contains pure SWNTs. This last fraction was then isolated
and used as the starting suspension for gel filtration.
Gel filtration was performed using Sephacryl S-500 gel filtration
medium (Amersham Biosciences) in a glass column of 20 cm length
and 2 cm inner diameter. After filling the glass column with the
filtration medium, the gel was slightly compressed to yield a final
height of 14 cm. For the separation, 10 mL of SWNT starting
suspension was applied to the top of the column and a solution of
1 wt % SCholate in H2O as eluant was pushed through the column
by applying sufficient pressure with compressed air to ensure a
flow of ∼1 mL/min. Fractions were collected in 3-4 mL portions.
After applying ∼50 mL, most of the SWNTs (consisting of mainly
individualized SWNTs in a ca. 2:1 mixture of metallic and
semiconducting ones) were eluted, whereas smaller particles
(residue amorphous carbon and metal particles) remain on the
column. In order to prove the magnetic purity of the purified
SWNTs, we checked them before the grafting of the SMMs. Since
Microscopies. TEM specimens were prepared by dropping a
suspension of the 1-SWNT conjugate from dry dichloromethane
onto carbon-coated copper grids (Plano GMbH, 200 mesh). The
TEM investigations were performed on a FEI Tecnai F20 ST
microscope operated at an accelerating voltage of 200 kV. AFM
images were acquired using a Veeco Nanoman D3100 instrument
in tapping mode.
Magnetic Measurements. The ac and dc magnetic susceptibility
measurements were carried out on a SQUID magnetometer by
Quantum Design, model MPMS-XL-5. Hysteresis cycles were
acquired with a homemade micro-SQUID setup.28
Results and Discussion
Synthesis. The heteroleptic bis(phthalocyaninato)terbium(III)
complex 1 was synthesized by a multistep convergent method
(Scheme 1).29 Two different phthalocyaninato lithium salts (2
and 4) and [Tb(acac)3 ·2H2O] (acac ) acetylacetonato) (5) were
reacted in a 1:1:1 ratio under reflux in 1-chloronaphhalene
leading to a mixture of mono- and binuclear complexes, whose
formation was monitored during the course of the reaction by
thin-layer chromatography and matrix-assisted laser desorption/
ionization-time-of-flight mass spectrometry (MALDI-ToF MS).
The isolation of the neutral target complex [A3B-TbIIIPc]0 (1)
was achieved via column chromatography on basic alumina
oxide followed by reprecipitation from a CH2Cl2-hexane
mixture.
The unsymmetrical monophtalocyanine ligand 9 was syn-
thesized by statistical condensation of a 10:1 molar ratio of 7
(A moiety) and 8 (B moiety) in lithium pentanolate solution
(28) Wernsdorfer, W.; Hasselbach, K.; Mailly, D.; Barbara, B.; Thomas,
L.; Suran, S. J. Magn. Magn. Mater. 1995, 145, 33–39. (b) Werns-
dorfer, W. Supercond. Sci. Technol. 2009, 22, 064013/1–064013/13.
(29) Pondaven, A.; Cozien, Y.; L’Her, M. New J. Chem. 1992, 16, 711–
718.
(27) Arnold, M. S.; Green, A. A.; Hulvat, J. F.; Stupp, S. I.; Hersam, M. C.
Nat. Nanotechnol. 2006, 1, 60–65.
9
15146 J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009

Anchoring of Single-Molecule Magnets on SWNTs
A R T I C L E S
Scheme 1. Schematic Representation of the Synthesis of 1
Scheme 2. Schematic Representation of the Synthesis of 2
under reflux for 3 h, as shown in Scheme 2. The resulting
products contain a mixture of several possible topoisomeres,
i.e. a total of five species with different masses. MALDI-
ToF MS investigation of the mixture reveals that A3B and
A4 constitute the main components (24% and 47%, respec-
tively), while A2B2, AB3, and B4 are minor components.
Monophtalocyanines with identical (A4) or different substit-
uents (A3B) can then be isolated from this mixture via radial
chromatography on silica gel, thus leading to the desired
phtalocyanine to be used to create the heteroleptic bis-
phthalocyaninato compound.
The B moiety, 4-(4-pyren-1-ylbutoxy)phthalonitrile (8), was
synthesized as shown in Scheme 3. Reduction of 4-pyren-1-
ylbutyric acid (11) gives 4-(1-pyrenyl)butanol (12), which is
then converted into 1-bromobutylpyrene (13) in presence of
phosphorus tribromide. Subsequent coupling with 4-hydroxy-
phthalonitrile (14) in a mixture of DMSO and acetone in the
presence of K2CO3 gives 8 in a good yield (77%).
respective anionic forms, obtained in CDCl3 after the reduction
byhydrazinehydrate.The1HNMRspectrumofthe[A3B-TbIIIPc]-
anion shows manifold, broad, superimposed multiplets of the
ring protons (Pc and A3B) together with signals assignable to
the protons in the side chains and in the pyrene group. All bands
show strong paramagnetic upfield shifts, with the farthest band
at δ) -63 ppm, while the homoleptic parent complex [Pc2Tb]-
exhibits in solution only two signals with dipolar shifts for
R-protons (δ ) -82.38 ppm) and ꢀ-protons (δ ) -39.10 ppm)
of the phthalocyanine ligands.14
High-resolution MALDI-ToF MS spectra of compounds 1,
2, 4, 6, and 9 (see Figures 3S, 7S, and 8S in the Supporting
Information) prove the heteroleptic nature of 1 bearing two
different types of macrocycles (Pc and A3B). For all compounds,
intense ionic signals, corresponding to the intact singly charged
molecular ions, were observed. In all cases, the molecular ion
was the most abundant high mass ion with distinct isotopic
distributions. The relative abundances of the isotopic ions are
in good agreement with the simulated spectra, as reported in
Table 1S (Supporting Information), where the mass spectral
results and calculated values are summarized.
Structural Characterization of 1. Complexes 1 and 6, due to
the presence of delocalized unpaired electrons, are NMR silent.
1
Conclusive H NMR investigations could be acquired for the
9
J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009 15147

A R T I C L E S
Kyatskaya et al.
Scheme 3. Synthesis of 4-(4-Pyren-1ylbutoxy)phthalonitrile (8)
The electrochemical behavior of a [Pc2Lu]0 complex has been
reported exhibiting two reversible redox waves at 0.0 and -0.4
V.30 It was shown that the incorporation of electron-donating
groups into the Pc-macrocycle of [Pc′2Ln]0 complexes (Ln )
Eu, Gd, Yb, Lu) renders the respective Pc′-ligands more difficult
to reduce in comparison with the unsubstituted complexes.30
The electrochemical behavior of complex 1 was studied by
cyclic voltammetry (CV) obtained at Pt electrode from a solution
of 1 in methylene chloride containing 0.1 M [n-Bu4N][ClO4]
(Figure 10S, Supporting Information). Two quasireversible
monoelectronic redox systems appear near -0.19 and -0.61
V, due to the reduction of the pthtalocyanine ligands and in
accordance with the electron-donating nature of the substituents.
In addition, the cyclic voltammogram of 1 exhibits an irrevers-
ible oxidation at 0.64 V attributed to the oxidative coupling of
the pyrene moiety as the reported literature value.31
Assembly of the Hybrid 1-SWNT Conjugate. The pyrenyl
group is known to interact strongly with SWNTs via π-stacking
interactions. This has been used, for example, in the production
of SWNT-nanoparticle hybrids,4a the grafting of proteins and
other biological molecules to SWNTs,32 and to immobilize light-
harvesting groups on the SWNT, as well as in the design of
new photoelectric devices.33 Following this strategy, the SWNTs
were strongly purified by ultracentrifugation and gel filtration
to remove all magnetic iron catalyst particles. The so gained
SWNTs were proven to be diamagnetic in magnetic susceptibil-
ity measurements and did not show any out-of-phase-signal in
ac measurements. A sample of highly purified SWNTs (1 mg)
was dispersed by overnight sonication in a solution of 1 (5 mg)
in dry dichloromethane (25 mL) saturated with nitrogen.
Figure 1. HR-TEM image of (a) a bundle of two pristine SWNTs, (b) a
bundle of two SWNTs decorated with complex 1, and (c) the EDX spectrum
of the 1-SWNT conjugate taken from the area depicted in part b. The
blue curve corresponds to the experimental spectrum; the red lines show
the energy of the bands.
Transmission electron microscopy (TEM) inspection of
pristine SWNTs did not reveal any externally grafted objects,
as shown in Figure 1a. In contrast, the TEM images after
treatment with 1 show isolated SWNTs where fixed molecules
of 1 can be clearly distinguished as darker marks. Inspection
of the TEM images indicates that the grafted objects possess a
diameter of about 1 nm, in accordance with the size of complex
1. The energy-dispersive X-ray (EDX) spectroscopy study
proves the presence of terbium in the area of the grafted objects
(Figure 1c).
Atomic force microscopy (AFM) experiments were carried
out to characterize the 1-SWNT conjugate further. AFM height
images performed in tapping mode show the appearance of
protrusions on the SWNTs (Figure 2a), whereby only few
molecules are grafted on the SWNT surface. Statistical inspec-
tion of the heights of the spots reveals a rather narrow
distribution of sizes (Figure 2b), indicating a mean height of 1
nm, compatible with the dimensions of 1. The lateral AFM
resolution does not allow a quantitative estimation of the number
of grafted SMMs, as the size of the complex 1 is comparable
to that of SWNTs, which precludes exact lateral differentiation.
(30) (a) L’Her, M.; Cozien, Y.; Courtot-Coupez, J. J. Electroanal. Chem.
1983, 157, 183–187. (b) Markovitsi, D.; Tran-Thi, T.-H.; Even, R.;
Simon, J. Chem. Phys. Lett. 1987, 137, 107–112. (c) Pondaven, A.;
Cozien, Y.; L’Her, M. New J. Chem. 1991, 15, 515–516. (d) Jiang,
J.; Liu, R. C. W.; Mak, T. C. W.; Chan, T. W.D.; Ng, D. K. P.
Polyhedron 1997, 16, 515–520.
(31) (a) Coleman, A.; Pryce, M. T. Inorg. Chem. 2008, 47, 10980–10990.
(b) Lu, G.; Shi, G. J. Electroanal. Chem. 2006, 586, 154–160.
(32) Chen, R. J.; Zhang, Y.; Wang, D.; Dai, H. J. Am. Chem. Soc. 2001,
123, 3838–3839.
(33) D’Souza, F.; Chitta, R.; Sandanayaka, A. S. D.; Subbaiyan, N. K.;
D’Souza, L.; Araki, Y.; Ito, O. J. Am. Chem. Soc. 2007, 129, 15865–
15871.
9
15148 J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009

Anchoring of Single-Molecule Magnets on SWNTs
A R T I C L E S
Figure 2. (a) AFM topography of the 1-SWNT conjugate on a SiO2
surface. (b) Analysis of the AFM heights of the grafted molecules, showing
the highly monodisperse size around 1 nm. The width of the spots
corresponds to the width of the AFM tip used. (c) Statistical distribution of
distances between spots on a SWNT and fit to the data (see the text).
The distribution of distances between randomly placed objects
along a line has a characteristic statistical shape,34 given by the
form: P(L) ) L exp(-CL), where P(L) is the probability of
finding a segment of length L and C is the linear concentration
of objects on the line. Similar expressions have been used to
calculate the statistical distance between two defects in a
genomic sequence35 or to determine the ensemble of magnetic
segments in doped magnetic chains.36 As P(L) depends on C
the statistical analysis of the distances between the detectable
molecules can give an indication of the number of SMMs on
the SWNTs (Figure 2c). This method is more reliable than just
counting the number of protrusions on the SWNT, because it
allows extrapolating, on the assumption that the objects are
randomly placed, those complexes of 1 that lie too close to one
another to be detectable by AFM and thus appear as a single
spot in the image. The curve obtained is reported in Figure 2c,
together with a fit to the expected expression (R2 ) 0.998),
which affords C ) 12.5 molecules per micrometer. This means
that the number of detectable molecules on the SWNTs is about
1 for every 80 nm of SWNT length. Supposing that about three-
fourths of the complexes of 1 stick in fact besides the SWNT,
as observed before for a different SMMs,4 a linear concentration
of one molecule for every 11 nm of SWNT can be obtained, in
overall agreement with the value extracted with atomic emission
spectroscopy (vide infra).
Quantitative investigation of the Tb content in the 1-SWNT
conjugate was carried out with inductively coupled plasma
atomic emission spectroscopy, revealing a Tb mass percentage
of 4.57%; corresponding to one Tb atom every 300 C atoms.
Considering that SWNTs of about 1 nm in diameter are the
most abundant ones for the batches used, this leads to the
presence, on average, of one molecule of 1 for every 7 nm of
SWNT length. This is in overall agreement with the value
extrapolated from AFM analysis. The slightly higher value here
obtained could arise from the fact that here all the SWNT surface
is available for grafting while in the previous case the SWNTs
Figure 3. (a) UV/vis/NIR spectra in CH2Cl2 for the heteroleptic complex
1 (blue) and the respective homoleptic complex 6 (purple). The spectra of
hydrazine hydrate reduced complexes of 1 (red) and 6 (cyan) are also
displayed, proving quenching of the intervalence band. (b) Comparison of
the emission spectra of the pyrene-containing precursor 8 (black, reduced
60 times), the heteroleptic complex 1 (red), the homoleptic complex 6
(green), and the 1-SWNT hybrids (blue) (all in CH2Cl2).
are lying on the SiO2 surface, which prevents stacking from
one direction.
Spectroscopic Characterization. The characterization by
UV-visible/NIR of 1 confirms its electronic structure as well
as the neutral charge of the molecule. The UV-visible/NIR
spectrum exhibits two B-bands at 328 and 350 nm, a very
intense Q-band at 672 nm, and two bands at 480 and 913 nm
attributed to the π-levels of the substituted Pc′ ligands (Figure
3a).30 In the near-IR region, two main bands are observed at
906 nm (with a satellite at 800 nm) and 1300-1800 nm.
The high-energy band is related to the radical part and
l
attributed to the eg(π)fa1u(π) transition; the lower-energy
band is assigned to an intramolecular charge transfer, whereby
these near-IR bands disappear upon reduction of 1 by
hydrazine hydrate (1% vv).
The characterization by emission spectroscopy of 1-SWNT
samples (in situ dispersion in CH2Cl2, sonicated for 2 h)
additionally confirms the supramolecular fixation of 1 to the
SWNTs. The emission spectrum of compound 1, acquired with
excitation at λex ) 340 nm, shows a series of well-defined peaks
in the 350-450 nm region (Figure 3b). These bands correspond
to three of the five vibronic bands of pyrene fluorescence and
appear at 377, 398, and 412 nm. All these band values are red-
shifted by 60 nm with respect to the peak position of pristine
pyrene or the pyrene-containing fluorescent moiety 8 (black
line), indicating that the electron states of the ligand affect the
electronic properties of the pyrene moiety. This is an interesting
observation by itself, as it implies that there is an electronic
interaction between the pyrene and the Pc moieties, which could
be used, in principle, to transmit a magnetic interaction. In any
case, this peak is useful to confirm the formation of the
1-SWNT conjugate, as it is known that π-π-stacking interac-
tions shift and quench the fluorescence of pyrene groups.
Comparison of the emission spectra of 1 (Figure 3b, in red)
with that of the 1-SWNT conjugate (Figure 3b, in blue) clearly
shows the quenching of the fluorescence of 1, in agreement with
the general tendency of photochemical deactivation of pyrene
(34) Johnson, N. L.; Kotz, S. Distributions in Statistics. Continuous
MultiVariate Distributions; Wiley: New York, 1972.
(35) Brooker, R. J. Genetics: Analysis and Principles; McGraw-Hill: New
York, 2008.
(36) (a) Bogani, L.; Caneschi, A.; Fedi, M.; Gatteschi, D.; Massi, M.;
Novak, M. A.; Pini, M. G.; Rettori, A.; Sessoli, R.; Vindigni, A. Phys.
ReV. Lett. 2004, 92, 207204–8. (b) Bogani, L.; Sessoli, R.; Pini, M. G.;
Rettori, A.; Novak, M. A.; Massi, M.; Fedi, M.; Giuntini, L.; Caneschi,
A.; Gatteschi, D. Phys. ReV. B. 2005, 72, 064406/1–064406/10.
9
J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009 15149

A R T I C L E S
Kyatskaya et al.
derivatives by aggregation with SWNTs. Moreover, we can
observe a variation of the relative intensity of the vibronic peaks,
which is known to be related to the polarity of the surrounding
medium. Therefore, this can be used as an in situ probe of the
grafting process onto the SWNT wall, which should change the
polarity of the pyrene environment. The vibronic peaks become
more pronounced after grafting onto the SWNTs, which is what
is expected when passing from a higher-polarity environment
to a low-polarity one. The relative intensity of the vibronic bands
also changes in accordance with the passage from a more polar
to a less polar, aromatic environment.
The emission spectrum of homoleptic terbium(III) complex
6 bearing two Pc ligands with pyrene moieties exhibits
quenching of the fluorescence, even without the presence of
SWNT. This is consistent with intramolecular π-π-interaction
of pyrene chromophores taking place in complex 6, likely due
to syn-directed pyrene groups. The spectrum of 6 strongly
resembles that of 1-SWNT, showing the same vibronic fine
structure as indication for a low-polarity environment. Such a
strong resemblance is attributed to π-π-stacking interactions,
which, while absent in 1 alone, are present in both the 1-SWNT
and 6. This result highlights the importance of using a
heteroleptic pyrene-containing complex 1 for the grafting, so
as to monitor the grafting process via optical means. In addition,
quenching of the emission of the semiconducting SWNTs was
observed in the region of 875-1400 nm; however, this indicator
was considered as not being unambiguous, since aggregation
of the nanotubes might cause similar quenching.
Figure 4. (a) Temperature dependence of the out-of-phase (ꢁ′′) components
of the ac molar magnetic susceptibility of the 1-SWNT conjugate and (b)
the Arrhenius plot extracted from data in part a proving the SMM behavior.
pre-exponential factor of 6.2 × 10-8 s, in very good agreement
with the reported behavior of [Pc2Tb]0.8 This confirms the
persistence of superparamagnetic behavior in the 1-SWNT
conjugates, showing that the anchored molecules still behave
as SMMs. This behavior seems to indicate that the dynamic
properties are enhanced, as already reported also for diluted
samples of [Pc2Tb]0 SMMs.8,14 However, the rather broad peaks
in the ac susceptibility, the low signal, and the random
orientation of the SMMs in the conjugates do not allow a very
accurate comparison. Nevertheless, the energy barrier and the
relaxation time of the 1-SWNT conjugate seems to be increased
and lowered, respectively, in comparison to the pure crystalline
compound, likely due to the suppression of intermolecular
interactions, as similar results have been reported for diluted
Magnetic Characterization
The magnetic relaxation has been studied to prove the SMM
behavior of the 1-SWNT conjugates and of complex 1.
The T dependences of the ac susceptibilities of bulk samples
of complex 1 show (Figure 12S, Supporting Information), in
agreement with the behavior of homologous nonheteroleptic bis-
phtalocyanine [Pc2Tb]0, the presence of slow relaxation of the
magnetization at relatively high temperatures: an out-of-phase
signal ꢁ′′ appears at 46 K at 1 kHz, and the maximum is shifted
when varying the frequency.8 Below 30 K both ꢁ′′ and ꢁ′ start
increasing again, indicating the presence of non-negligible
magnetic interactions among SMMs, in agreement with that
observed for the dc susceptibility of undiluted [Pc2Tb] SMMs.14
This prevents a reliable extraction of the magnetic energy barrier
∆E from an Arrhenius plot. However, the ꢁ′′ plot clearly reveals
the presence of a shift in the peak frequency, comparable to
that previously reported. The large temperature range over which
the ꢁ′′ signal is spread is also attributable to intermolecular
interactions, which should at least partially disappear in
1-SWNT conjugates, due to increased inter-SMM distances.
The temperature dependence of the ꢁT product of 1, as
obtained with a SQUID magnetometer, is reported in Figure 12S
(Supporting Information). The ꢁT product steadily decreases on
lowering the temperature, as expected from the depopulation
of the spin levels of the rare-earth. A minimum value of 10.4
emu K/mol is reached at 7 K, and a steady increase is observed
below this temperature, indicating the onset of intramolecular
magnetic interactions.14
[Pc2Tb]0 and [Pc2Tb]-1 8,14b Note that this is in strong contrast
.
to recently reported Fe6-wolframate-based SMMs,4b where the
relaxation time becomes faster when the respective SMM-SWNT
conjugates are produced. This is probably due to the fact that
polynuclear transition-metal SMMs are more sensitive to small
structural deformations than monocenter rare-earth SMMs.
The persistence of SMM behavior in the conjugates is also
evidenced by the opening of a hysteresis cycle at low temper-
atures (Figure 5), as acquired with a micro-SQUID system.28
The hysteretic behavior was investigated as a function of
temperature and field sweeping rate (see also Figure 13S,
Supporting Information). The hysteresis loops of the 1-SWNT
conjugate exhibit increasing hysteresis with increasing field
sweep rate at a constant temperature (Figure 5a), increasing
hysteresis with decreasing temperature at a constant sweep rate,
as expected for the superparamagnet-like properties of a SMM.
Close to zero field the cycles present fast relaxation leading to
a step at about zero field, which is also the behavior observed
for such measurements performed on a powder sample of 1
(Figure 5b). This corresponds to the desired fingerprint of the
Tb-based SMM, which shows an increased magnetic relaxation
rate close to zero field. It is to be noted that powder measure-
ments of pure SMM complex 1 display a faster relaxation time,
close to zero field, than measurements acquired on diluted
[TbPc2] SMMs inserted in a crystalline matrix of diamagnetic
[YPc2] complexes.9 In the case here presented, 1-SWNT
conjugate, the hysteresis loops are more open at zero field, which
would hint to strongly reduced intermolecular interactions. In
isolated molecules, tunnelling is forbidden at zero-field because
of the Kramer’s theorem. As stated before, a system of such
kind must have a ground state that is at least doubly degenerate,
The ac susceptibility of the 1-SWNT conjugates reveals
magnetic properties of the conjugate: A ꢁ′′ signal appears at
even higher T than for bulk 1, with maxima between 44 K (997.3
Hz) and 27.5 K (1.00 Hz), as shown in Figure 4a. The Arrhenius
plot, extracted from the temperatures of the maxima for the
frequencies used (Figure 4b) affords ∆E ) 351 cm-1 with a
9
15150 J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009

Anchoring of Single-Molecule Magnets on SWNTs
A R T I C L E S
load of one molecule of 1 per 7-8 nm of SWNT length.
Emission spectra before and after the creation of the conjugates
indicate that the pyrene groups are actively involved in the
grafting, having a different environment in the hybrids than in
the crystalline compound or in dichloromethane solutions. This
nicely confirms the importance of π-interactions in creating
SMM/SWNT hybrids and constitutes an important experimental
finding for the future rationalization of interactions between
SMM and SWNT, both magnetic and nonmagnetic.
An investigation of the magnetic properties of the SMM-
SWNT conjugate was carried out by demonstrating that the
SMM behavior is retained and even improved in the hybride
systems, in contrast to many d-metal clusters on surfaces.37 This
demonstration is of primary importance for future experiments
and also already hints at the possibility of observing different
dynamics with respect to crystalline systems.
These results open a chemical route to single rare-earth
centers as constituent units for SWNT-based spintronic devices,
thus allowing the exploitation of their fingerprint parity effects7-14
in the quantum tunneling of the magnetization. Moreover, such
systems display coherence times longer than those of comparable
SMM that have been used to produce SMM-SWNT hybrids,38
and this could be used for quantum information single-molecule
experiments. Eventually, these compounds are very promising
candidates for single-molecule magnetic39 measurements, both
when grafted on SWNTs connected to superconducting elec-
trodes, as in the nano-SQUID, and when attached to metallic
SWNTs connected to Pd electrodes, which would allow the
observation of magneto-Coulomb effects.18,19 In this context,
it is interesting to note that facile bulk scale separation methods
have recently become available that yield both pure semicon-
ducting and pure metallic SWNT samples. In the future, we
will attempt analogous experiments with correspondingly puri-
fied SWNT materials.
Figure 5. Comparison of the micro-SQUID hysteresis cycles of (a) the
1-SWNT conjugate and (b) the powder sample of 1 recorded for different
scan rates. The hysteresis loops are more open at zero field for the 1-SWNT
conjugate (bottom), a result attributed to strongly reduced intermolecular
interactions in the conjugate.
even in the presence of crystal field or spin-orbit interactions
or in the presence of strong structural distortions. In the case of
terbium, due to hyperfine and nuclear quadrupole interactions,
the electronic spin number J is not a good quantum number to
calculate parity effects, and we shall consider the sum of nuclear
and electronic spin numbers. Terbium practically has only one
isotope with 100% abundance and nuclear quantum number I
) 3/2. The total quantum number is thus a half-integer and no
tunnel level splitting is allowed by Kramer’s parity theorem.
This is reflected in the shape of the cycles of the system, where
no neat tunnelling at zero field can be observed. The 1-SWNT
conjugate thus constitutes a perfect example of the desired
fingerprint in the magnetic behavior, showing retained and even
improved key characteristic feature of the SMMs relaxation in
close proximity of SWNTs.
Acknowledgment. We thank D. Wang, H. Ro¨ssner, and C.
Ku¨bel for HR-TEM measurements and A. Burlak for the help in
the manuscript preparation. Financial support was provided by
Deutsche Forschungsgemeinschaft (DFG), the ERA-NET-Chemistry
project “MultiFUN”, ANR-PNANO, ANR-06-NANO-27 MolSpin-
tronics and ANR-08-P110_MolNanoSpin, NE-MAGMANET (FP6-
NMP3-CT-2005-515767), EC-RTN QuEMolNa (FP6-CT-2003-
504880), ERC-08-226558- MolNanoSpin, and the Marie Curie
action EIF-041565 MoST.
1
Supporting Information Available: H NMR, UV/vis data,
MALDI-TOF MS of compounds 1, 2, and 9 and additional
microscopic and magnetic characteristics. This material is
available free of charge via the Internet at http://pubs.acs.org.
Conclusions
JA906165E
In conclusion, we have synthesized and structurally character-
ized a new heteroleptic pyrenyl-containing bis-phthalocyanine
SMM 1. The SMM behavior of the bulk complex 1 was proved
by ac susceptibility and micro-SQUID investigations. A
SMM-SWNT magnetic conjugate was obtained by supramo-
lecular assembly of complex 1 via π-interactions of the pyrene-
functionality with highly purified SWNTs (whereby the grafting
was qualitatively and quantitatively analyzed using HR-TEM
microscopy, emission spectroscopy, AFM imaging, energy-
dispersive X-ray, and elemental analysis), yielding an average
(37) (a) Ehli, C.; Aminur Rahman, G. M.; Jux, N.; Balbinot, D.; Guldi,
D. M.; Paolucci, F.; Marcaccio, M.; Paolucci, D.; Melle-Franco, M.;
Zerbetto, F.; Campidelli, S.; Prato, M. J. Am. Chem. Soc. 2006, 128,
11222–11231. (b) Fernando, K. A. S.; Lin, Y.; Wang, W.; Kumar, S.;
Zhou, B.; Xie, S.-Y.; Cureton, L. T.; Sun, Y.-P. J. Am. Chem. Soc.
2004, 126, 10234–10235.
(38) Bertaina, S.; Gambarelli, S.; Tkachuk, A.; Kurkin, I. N.; Malkin, B.;
Stepanov, A.; Barbara, B. Nature Nano. 2007, 2, 39–42.
(39) (a) Ruben, M. Angew. Chem., Int. Ed. 2005, 44, 1594–1596. (b) Ruben,
M.; Lehn, J.-M.; Mu¨ller, P. Chem. Soc. ReV. 2006, 35, 1056–1067.
(c) Ruben, M.; Landa, A.; Lo¨rtscher, E.; Riel, H.; Mayor, M.; Go¨rls,
H.; Weber, H. B.; Arnold, A.; Evers, F. Small 2008, 4, 2229–2235.
9
J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009 15151