Single-molecule-magnet carbon-nanotube hybrids

Article abstract of DOI:10.1002/anie.200804967

(Figure Presented) Devices and desires: The self-assembly of single-molecule-magnet (SMM) carbon-nanotube (CNT) hybrids (see picture) in conditions compatible to the creation of electronic devices is described. The process is controlled at the single-mole

Full text of DOI:10.1002/anie.200804967

Communications
DOI: 10.1002/anie.200804967
Single-Molecule Studies
Single-Molecule-Magnet Carbon-Nanotube Hybrids**
Lapo Bogani,* Chiara Danieli, Elisa Biavardi, Nedjma Bendiab, Anne-Laure Barra,
Enrico Dalcanale, Wolfgang Wernsdorfer, and Andrea Cornia
Dedicated to an anonymous bone-marrow donor in Cyprus
Carbon nanotubes (CNTs) hold great promise for sensing[1,2]
and nanoelectronics,[3] as core components of chemical and
biological[1,2] ultra-sensitive probes and of field-effect tran-
sistors (FETs).[3] CNT–SQUID devices[4] in particular could
constitute magnetic detectors with single-molecule sensitivity,
thus offering a viable route to the long-sought readout of
magnetic information stored in individual single-molecule
magnets (SMMs).[5] SMMs are metal-ion clusters with a large
easy-axis magnetic anisotropy,[6] exhibiting a magnetic hyste-
resis loop at low temperature and suggested as components
for quantum computing[7] and molecular spintronics.[5] To
date, the chemistry needed to bridge the domains of CNTs
and SMMs has remained unexplored. CNT hybrids with gold
or magnetic nanoparticles, proteins, enzymes, or luminescent
molecules are currently under intense investigation.[1,2,8] The
resulting materials usually entail a large number of nano-
particles or molecules per CNT, whereas CNT–SMM detec-
tors and spintronic devices require the sequential addition of
a small but very controlled number of nanomagnets. Grafting
through covalent bonds might introduce electron scattering
centers that may limit the performance of CNT devices. By
contrast, noncovalent p-stacking interactions with pristine
CNTs should largely preserve the CNT conductance, while
guaranteeing SMM–CNT interaction.
Herein we report the assembly of CNT–SMM hybrids
using a tailor-made tetrairon(III) SMM, [Fe4(L)2(dpm)6] (1;
Hdpm = dipivaloylmethane), designed to graft onto the walls
of CNTs. The ligand L3ꢀ (H3L = 2-hydroxymethyl-2-(4-
(pyren-1-yl)butoxy)methylpropane-1,3-diol), features an
alkyl chain with a terminal pyrenyl group and was synthesized
as in Figure 1a. Reduction of 4-pyren-1-yl-butyric acid gives
4-(1-pyrenil)butanol, which is then coupled with 4-bromo-
methyl-1-methyl-2,6,7-trioxa-bicyclo[2.2.2]octane.[9] A two-
steps deprotection of the trimethylol function affords H3L,
which is finally treated with the preformed[6] complex [Fe4-
(OMe)6(dpm)6] (2) to give 1 in excellent yield (95%).
The molecular structure of 1 (Figure 1b,c), determined by
single-crystal X-ray diffraction,[10] shows a tetrairon(III)
propeller-like core with idealized D3 symmetry held together
by two triply deprotonated H3L ligands lying at opposite sides
of the molecular plane (see Supporting Information). The
molecular size of 1 is 1.6–2.3 nm (av.: 1.9 nm). Low-temper-
ature high-frequency (HF)-EPR spectra at 190 and 230 GHz
(Figure 2a) and variable-temperature magnetic-susceptibility
measurements show the presence of an S = 5 high-spin ground
state with an easy-axis magnetic anisotropy (D =
ꢀ0.409 cmꢀ1; Supporting Information). Indeed, single-crystal
magnetic measurements reveal a hysteresis loop below 1 K
with characteristic quantum-tunneling[6] steps (Figure 2b),
confirming the SMM behavior.
[*] Dr. L. Bogani, Dr. N. Bendiab, Dr. W. Wernsdorfer
Institut Nꢀel, CNRS
25 Av. des Martyrs, 38042 Grenoble, Cedex 9 (France)
Fax: (+33)4-7688-1191
E-mail: lbogani@hotmail.com
Dr. A.-L. Barra
LCMI-CNRS
25 Av. des Martyrs, 38042 Grenoble, Cedex 9 (France)
CNT–FETs were obtained by electron-beam lithography
on degenerately n-doped silicon wafers covered with a 300 nm
thick SiO2 layer. Single CNTs were located by atomic force
microscopy (AFM) and connected by palladium leads sepa-
rated by 300 nm gaps. The hybrids were then produced by
immersion of the CNT–FETs in a 3.1 ꢀ 10ꢀ5 m solution of 1 in
1,2-dichloroethane (DCE) for 30 min, followed by extensive
washing with pure DCE. 1H NMR, ESI-MS, and fluorescence
techniques demonstrate that the complex is completely stable
in solution in the conditions used for the deposition (Sup-
porting Information). The grafting was reiterated to follow
the progressive addition of SMMs. After each treatment a few
SMMs were found to stick onto the CNT (Figure 3a), while
some others were also located on the surrounding surface.
The isostructural complex containing H3L’ = 2-hydroxy-
methyl-2-phenylpropane-1,3-diol[11] did not graft onto CNTs
in the same experimental conditions. This result is a strong
indication that 1 has been grafted as a result of the pyrenyl
functionalities.
Dr. C. Danieli, Prof. A. Cornia
Dipartimento di Chimica & INSTM
Universitꢁ di Modena e Reggio Emilia
Via G. Campi 183, 41100 Modena (Italy)
Dr. E. Biavardi, Prof. E. Dalcanale
Dipartimento di Chimica Organica e Industriale & INSTM
Universitꢁ di Parma
Viale G. P. Usberti 17/a, 43100 Parma (Italy)
[**] We thank Dr. F. Bianchi (Universitꢁ di Parma) for ESI-MS experi-
ments, Prof. G. Ponterini (Universitꢁ di Modena e Reggio Emilia) for
fluorescence spectra and Prof. R. Sessoli (Universitꢁ di Firenze) for
magnetic measurements. The work was financed by ANR-PNANO,
ANR-06-NANO-27 MolSpintronics, NE MAGMANET (FP6-NMP3-
CT-2005-515767), EC-RTN QuEMolNa (FP6-CT-2003-504880),
ERANET project “NanoSci-ERA: NanoScience in the European
Research Area” (SMMTRANS), PRIN and FIRB grants and the Marie
Curie action EIF-041565 MoST.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.200804967.
746
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 746 –750

Angewandte
Chemie
Micro-Raman investigation of the CNT–
SMM hybrids clearly revealed, superimposed
onto CNT and SiO2 spectra, several peaks
associated with intact molecules of 1. They are
not observed when the solutions used in the
deposition are thermally treated so as to inten-
tionally break the SMM magnetic core (Sup-
porting Information). Furthermore, the fluores-
cence of pyrene units is quenched in the hybrids,
as in solutions of intact 1 (Supporting Informa-
tion).
If 1 is randomly distributed along the CNT
the probability of having two consecutive mol-
ecules at a distance L is P(C,L) = CeꢀL C, where
C is the linear concentration of 1 on CNTs.[12]
Statistical analysis of the distribution of L,
performed on 40 CNTs, revealed good agree-
ment with the predicted law and allowed C to be
extracted for each repetition (Figure 4). C varies
linearly (R = 0.998) with the number of repeti-
tions, with 1.5 SMMmmꢀ1 deposited at each
iteration. Thus, in CNT–FETs with 300 nm
gaps, we can graft, on average, one SMM
every two treatments, allowing the sequential
addition of single SMMs to CNT devices.
CNTs are either metallic or semiconducting.
Herein we concentrate on the room-temper-
ature transport properties of devices based on
semiconducting CNTs, which are better suited
for ultra-sensitive detection.[2,3] Current (Id)
versus gate voltage (Vg) curves show p-type
FET behavior[4] (ON state at Vg < 0) with typical
ION/IOFF ratios of three decades (Figure 5a).
Hysteresis is observed, as a result of charge
injection from the CNT to the nearby region.[13]
The subthreshold swing Sw = [dVg/dLog(I)]
remains unchanged upon grafting, indicating
that the treatment leaves the gate unaffected.
While no appreciable variation is observed after
repeated treatment of the FETs with DCE
alone (Supporting Information), the ION current
decreases linearly upon SMM grafting, indicat-
Figure 1. a) Synthesis of 1: 1) BH3·THF, THF, 08C, 20 h, 2) NaOH, DMF, 908C, 12 h,
3) HCl, MeOH, 6 h, 4) Na2CO3, MeOH, 12 h, 5) Et2O, [Fe4(OMe)6(dpm)6] (2). b) Crystal
structure of 1, showing the Fe4 core and the peripheral pyrene-terminated chains
(hydrogen atoms omitted). c) View of the tetrairon(III) core along the idealized three-
fold axis, showing the preserved cluster geometry (ligands H3L and tert-butyl carbon
atoms of dpm omitted).
The AFM height profile of the same CNT reveals the
sequential addition, after each treatment, of objects with size
of 1–2 nm (Figure 3b). Statistical analysis of small CNTs with
diameter of about 1 nm (Figure 3c) shows a two-peak
distribution that can be fitted (R2 = 0.98) with two Gaussians
centered at (1.9 ꢁ 0.3) nm and (1.0 ꢁ 0.3) nm (errors are full-
widths-at-half-maxima), with an area ratio of 1:3.
ing that 1 interacts with the electron flow, producing scatter-
ing centers.[3] Linear fit of the trend (R = 0.989) indicates a
contribution of 560 ꢁ 40 pA per SMM (Figure 5b), which is
higher than our instrumental sensitivity of 10 pA and
demonstrates single-molecule detection capability. In the
Landauer formulation of incoherent transport regime,[14] such
ION decrease corresponds to an electron reflection coefficient
of about 7% per SMM. For comparison, covalent binding of
protein molecules on CNTs leads to an electron reflection
coefficient of 40% per molecule.[3]
The grafting also produces a gradual decrease in the ON/
OFF threshold voltage Vth1, along with an increase of the
hysteresis width (Supporting Information). This change is to
be attributed to charge-trapping by 1, as well as a charge
transfer between 1 and the CNT.
The former peak closely matches the molecular size
evaluated from X-ray data and can be assigned to SMMs lying
on top of CNTs. The feature of height 1 nm most likely arises
from SMMs lying beside a CNT, for which the CNT–SMM
height difference is detected (Figure 3d), because of the
limited lateral resolution of the AFM tip. Thus about 1= of the
4
grafted SMMs rest on top of the CNTs, with no direct
interaction with the surface. Inspection of CNTs or bundles
with diameter larger than 2 nm shows only one Gaussian peak
(R2 = 0.993) at 2.0 ꢁ 0.2 nm, confirming this interpretation.
Interestingly
1 can be burnt by laser irradiation
(30 mWmmꢀ2 at 576.5 nm), causing the disappearance of the
Angew. Chem. Int. Ed. 2009, 48, 746 –750
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
747

Communications
Figure 2. Magnetic properties of 1: a) Variable-temperature HF-EPR
spectra recorded at 230 GHz (black) on a powder sample, along with
the best-fit simulations (red); b) Hysteresis cycles recorded on a single
crystal of 1 at 11 different temperatures between 1 K (black) and
40 mK (red). The magnetic field is swept at a rate of 170 Oesꢀ1 along
the magnetic easy axis.
Figure 3. AFM topographic analysis of the hybrids. a) Height images
of the same CNT acquired on repeating the grafting process: left one
time, center four times, right ten times. b) Section profile along the
same CNT before (black line) and after (red line) multiple grafting of
1. c) Heights of the grafted objects. The statistics acquired on single
CNTs or small bundles with diameter <1 nm (yellow; 1 & 2 see (d))
differs from that acquired on CNTs or bundles with larger diameter
(blue; 3 & 4 see (d)). Lines are fittings with Gaussian distributions.
d) Corresponding different dispositions of 1 with respect to CNTs (1 &
2) or bundles (3 & 4).
Raman fingerprints and the recovery the original transport
curve (Figure 5a). The grafting–burning procedure can be
repeated up to a dozen times, but after this point further
iterations fail to recover the original curve, probably because
of accumulation of depleted material.
In conclusion we bridged the domains of molecular
magnetism and CNTs by fabricating the first CNT–SMMs
hybrids containing intact pyrene-functionalized SMMs in
conditions compatible to the creation of electronic devices.
We controlled the grafting of SMMs down to a single-
molecule level and we demonstrated the single-SMM sensi-
tivity of CNT–FETs. These results pave the way to the
construction of “double-dot” molecular spintronic devices,[5]
where a controlled number of nanomagnets is coupled to an
electronic nanodevice, and to the observation of the magneto-
Coulomb effect. The future use of CNT–SQUIDs may also
benefit from the chemical methodology described herein.
Finally, the approach may be extended to produce different
functional hybrids incorporating charge-transfer complexes,
valence tautomers, or photomagnetic materials.[15]
Figure 4. Probability of finding a distance L between SMMs grafted on
CNTs for sequential repetitions of the process. Symbols are experimen-
tal data for 1, 2, 4, and 6 repetition and lines are the corresponding
fits (see text). Inset: extracted linear concentration (C) as a function of
the number of repetitions of the process and linear fit.
748
www.angewandte.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 746 –750

Angewandte
Chemie
[2] Y. B. Zhang, M. Kanungo, A. J. Ho, P. Freimuth, D. van der Le-
lie, M. Chen, S. M. Khamis, S. S. Datta, A. T. C. Johnson, J. A.
Misewich, S. S. Wong, Nano Lett. 2007, 7, 3086 – 3091.
[3] a) M. Ouyang, J.-L. Huang, C. L. Cheung, C. M. Lieber, Science
2001, 292, 702 – 705; b) P. Avouris, Z. Chen, V. Perebeinos, Nat.
Nanotechnol. 2007, 2, 605 – 615; c) S. J. Tans, A. R. M. Verscheu-
ren, C. Dekker, Nature 1998, 393, 49 – 52; d) D. N. Futaba, K.
Hata, T. Yamada, T. Hiraoka, Y. Hayamizu, Y. Kakudate, O.
Tanaike, H. Hatori, M. Yumura, S. Iijima, Nat. Mater. 2006, 5,
987 – 994.
[4] SQUID = superconducting quantum interference device; a) J.-P.
Cleuziou, W. Wernsdorfer, V. Bouchiat, T. Ondarꢁuhu, M.
Monthioux, Nat. Nanotechnol. 2006, 1, 53 – 59; b) P. Jarillo-
Herrero, J. A. van Dam, L. P. Kouwenhoven, Nature 2006, 439,
953 – 956.
[5] a) L. Bogani, W. Wernsdorfer, Nat. Mater. 2008, 7, 179 – 186;
b) L. Bogani, W. Wernsdorfer, Inorg. Chim. Acta 2008, 361,
3807 – 3819.
[6] a) D. Gatteschi, R. Sessoli, J. Villain, Molecular Nanomagnets,
Oxford University Press, Oxford, UK, 2006; b) D. Gatteschi, R.
Sessoli, Angew. Chem. 2003, 115, 278 – 309; Angew. Chem. Int.
Ed. 2003, 42, 268 – 297; c) A. L. Barra, A. Caneschi, A. Cornia,
F. F. De Biani, D. Gatteschi, C. Sangregorio, R. Sessoli, L.
Sorace, J. Am. Chem. Soc. 1999, 121, 5302 – 5310.
[7] a) M. N. Leuenberger, D. Loss, Nature 2001, 410, 789; b) F.
Troiani, A. Ghirri, M. Affronte, S. Carretta, P. Santini, G.
Amoretti, S. Piligkos, G. Timco, R. E. P. Winpenny, Phys. Rev.
Lett. 2005, 94, 207208; c) F. Troiani, M. Affronte, S. Carretta, P.
Santini, G. Amoretti, Phys. Rev. Lett. 2005, 94, 190501; d) J.
Lehmann, A. Gaita-Ariꢂo, E. Coronado, D. Loss, Nat. Nano-
technol. 2007, 2, 312 – 317.
Figure 5. Transport properties of the SMM–CNT hybrids. a) Id versus
Vg curve of a typical CNT–FET shown for the pristine device (black line)
and for different numbers (n) of grafted SMMs (color scale). The
original curve is regained after destruction of 1 by laser burning (gray
dashed line) Vth1 and Vth2 indicate the ON!OFF and OFF!ON
threshold voltages, respectively. b) Variation of the ION as a function of
the number of grafted molecules (n; circles). The triangle represents
the value after laser burning and the line is a fit to the data.
[8] a) S. Banerjee, T. Hemraj-Benny, S. S. Wong, Adv. Mater. 2005,
17, 17 – 29; b) D. M. Guldi, G. M. A. Rahman, F. Zerbetto, M.
Prato, Acc. Chem. Res. 2005, 38, 871 – 878; c) V. Georgakilas, D.
Gournis, V. Tzitzios, L. Pasquato, D. Guldi, M. Prato, J. Mater.
Chem. 2007, 17, 2679 – 2694; d) K. Kurppa, H. Jiang, G. R.
Szilvay, A. G. Nasibulin, E. I. Kauppinen, M. B. Linder, Angew.
Chem. 2007, 119, 6566 – 6569; Angew. Chem. Int. Ed. 2007, 46,
6446 – 6449; e) K. Jiang, A. Eitan, L. S. Schadler, P. M. Ajayan,
R. W. Siegel, N. Grobert, M. Mayne, M. Reyes-Reyes, H.
Terrones, M. Terrones, Nano Lett. 2003, 3, 275 – 277.
Experimental Section
Detailed description of the synthetic procedures is reported in the
Supporting Information. CNT–FETs were fabricated on degenerately
n-doped Si wafers covered with 300 nm thick SiO2 (Siltronix sas).
Gold electrodes were deposited on the surface by deep-UV lithog-
raphy, providing a reference system to locate CNTs. Single-walled
CNTs (Rice Univ.) in DCE were spin-coated on the wafers, located by
AFM and connected with Pd leads by e-beam lithography (Support-
ing Information). 65% of contacts displayed a resistance lower than
100 kW. AFM images were acquired using a Veeco D3100 instrument
in tapping mode; hysteresis loops and HF-EPR spectra were recorded
on home-built instrumentation;[11] Raman spectra at 632.8 nm were
obtained on a LabRam Infinity spectrometer focusing on an area of
1 mm2.
[9] A. B. Padias, H. K. Hall, Macromolecules 1982, 15, 217 – 223.
[10] Crystal structure data for 1: C116H164Fe4O20, Mr = 2101.87, crystal
¯
dimensions 0.64 ꢀ 0.42 ꢀ 0.30 mm, triclinic, space group P1, a =
16.3398(4), b = 19.1784(5), c = 21.9168(5) ꢃ, a = 87.025(1)8, b =
68.784(1)8, g = 65.182(1)8, V= 5770.6(2) ꢃ3, Z = 2, 1calcd
=
1.210 gcmꢀ3, m = 0.556 mmꢀ1, MoKa radiation, l = 0.71073 ꢃ,
T= 120(2) K, 2qmax = 50.168, 42131/19160 reflections collected/
unique (Rint = 0.0264), R1 = 0.0733, wR2 = 0.2323, largest diff.
peak and hole 0.985/ꢀ0.558 eꢃꢀ3. CCDC-704776 (1) contains
the supplementary crystallographic data for this paper. These
data can be obtained free of charge from The Cambridge
Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_
request/cif.
Received: October 10, 2008
Published online: December 15, 2008
Keywords: iron · nanotubes · sensors · single-molecule magnets ·
.
single-molecule studies
[11] S. Accorsi, A.-L. Barra, A. Caneschi, G. Chastanet, A. Cornia,
A. C. Fabretti, D. Gatteschi, C. Mortalꢄ, E. Olivieri, F. Parenti, P.
Rosa, R. Sessoli, L. Sorace, W. Wernsdorfer, L. Zobbi, J. Am.
Chem. Soc. 2006, 128, 4742 – 4755.
[12] a) N. L. Johnson, S. Kotz, Distributions in Statistics, Wiley, New
York, 1972; b) For the conceptually identical problem of
randomly placed diamagnetic sites in a magnetic chain, see: L.
Bogani , A. Caneschi, M. Fedi, D. Gatteschi, M. Massi, M. A.
Novak, M. G. Pini, A. Rettori, R. Sessoli, A. Vindigni, Phys. Rev.
Lett. 2004, 92, 207204; c) L. Bogani , R. Sessoli, M. G. Pini, A.
Rettori, M. A. Novak, M. Massi, M. Fedi, L. Giuntini, A.
Caneschi, D. Gatteschi, Phys. Rev. B 2005, 72, 064405.
[1] a) J. Kong, N. R. Franklin, C. Zhou, M. G. Chapline, S. Peng, K.
Cho, H. Dai, Science 2000, 287, 622 – 625; b) P. G. Collins, K.
Bradley, M. Ishigami, A. Zettl, Science 2000, 287, 1801 – 1804;
c) K. Besteman, J.-O. Lee, F. G. M. Wiertz, H. A. Heering, C.
Dekker, Nano Lett. 2003, 3, 727 – 730; d) A. Star, J. C. P. Gabriel,
K. Bradley, G. Gruner, Nano Lett. 2003, 3, 459 – 462; e) R. J.
Chen, S. Bangsaruntip, K. A. Drouvalakis, N. Wong Shi Kam, M.
Shim, Y. Li, W. Kim, P. J. Utz, H. Dai, Proc. Natl. Acad. Sci. USA
2003, 100, 4984 – 4989.
Angew. Chem. Int. Ed. 2009, 48, 746 –750
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
749

Communications
[13] a) K. Bradley, J. Cumings, A. Star, J.-C. P. Gabriel, G. Gruner,
[15] a) J. S. Miller, J. C. Calabrese, R. S. McLean, A. J. Epstein, Adv.
Mater. 1992, 4, 490 – 493; b) A. Dei, Angew. Chem. 2005, 117,
1184 – 1187; Angew. Chem. Int. Ed. 2005, 44, 1160 – 1163; c) C.
Carbonera, A. Dei, J.-F. Lꢅtard, C. Sangregorio, L. Sorace,
Angew. Chem. 2004, 116, 3198 – 3200; Angew. Chem. Int. Ed.
2004, 43, 3136 – 3138; d) K. Matsuda, M. Irie, J. Am. Chem. Soc.
2000, 122, 8309 – 8310.
Nano Lett. 2003, 3, 639 – 641; b) W. Kim, A. Javey, O. Vermesh,
Q. Wang, Y. Li, H. Dai, Nano Lett. 2003, 3, 193 – 198.
[14] a) M. Bockrath, W. Liang, D. Bozovic, J. H. Hafner, C. M.
Lieber, M. Tinkham, H. Park, Science 2001, 291, 283 – 285; b) A.
Bachtold, M. S. Fuhrer, S. Plyasunov, M. Forero, E. H. Ander-
son, E. Zettl, P. L. McEuen, Phys. Rev. Lett. 2000, 84, 6082 –
6085.
750
www.angewandte.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 746 –750