A R T I C L E S
Tulevski et al.
and 1570 cm-1 are assigned to the G+ and G- bands,
respectively, and are associated with the C-C stretch modes
along the nanotube axis (G+) and the circumferential direction
(G-). The D-band, 1350 cm-1, is commonly referred to as the
defect band and is a phonon mode activated by the presence of
defects in the structure of the SWCNT. The peaks at 378, 417,
and 645 cm-1 are assigned to the Eg and A1g peaks of the
sapphire substrate and were used to calibrate the spectra. Since
the functionalization results in the hybridization of the bonded
carbon atoms (from sp2 to sp3) of the SWCNTs, defects are
introduced into the crystal structure upon functionalization,
increasing the intensity of the D-band mode. Monitoring the
intensity ratio between the D-band and the G-band, ID/IG, will
indicate whether the reaction took place, to what degree, and if
it was reversed. The ID/IG ratio is significantly higher in the
spectrum taken of the functionalized SWCNTs (Figure 2, red
trace). The G-band intensity is attenuated and broadened as well.
This ratio reverts to the initial value upon annealing the
functionalized SWCNTs, as is seen in the blue trace in Figure
2, suggesting that the SWCNTs are defunctionalized and that
the sp3 carbon atoms have rehybridized back to the sp2 state.
The RBM peak at 220 cm-1 also decays upon functionalization
as the symmetry of the radial mode is broken due to covalent
attachment of the organic moiety.24 This mode is also fully
recovered upon annealing. The data confirm that the SWCNTs
are chemically functionalized with 1 and that the functional-
ization is reversible.
Figure 2. Raman spectra of pristine SWCNT (black trace), after function-
alization with 1 (red trace), and after annealing to remove 1 (blue trace).
The data were acquired by depositing the samples onto sapphire windows
and using a 488 nm excitation.
nium salt (1) reacted readily with SWCNTs to yield the
functionalized species (f-SWCNT). The diazonium salt formed
a charge-transfer complex at the surface of the SWCNT with
electron transfer from the SWCNT to the organic compound
leading to the formation of a SWCNT-aryl carbon-carbon bond,
converting the sp2 hybridized SWCNT carbon atom to a sp3
hybridized carbon atom.13-17,22 Annealing the reacted material
to 600 °C, under argon, cleaved the SWCNT-aryl bond, restoring
the original sp2 hybridization of the SWCNT. Compound 1 was
end-functionalized with hydroxamic acid. Hydroxamic acids are
known to bind strongly and selectively to basic metal oxides
(i.e., Al2O3 and HfO2) as opposed to silicon oxide.12,18 We use
a short-chain (x ) 0) hydroxamic acid terminated compound
that renders the functionalized SWCNTs water-soluble. An
advantage of using aqueous solutions, as opposed to common
organic solvents as used in the previous reports,10,11 is that the
solution pH can be precisely controlled to mediate the subtle
acid-base interaction between the terminal hydroxamic acid
and the desired metal-oxide surface allowing for increased
SWCNT density and, subsequently, increased device yield. The
SWCNT functionalization is completely reversible upon an-
nealing to 600 °C. This reversibility is critical since the
attachment chemistry rehybridizes the carbon atoms at the point
of attachment, introducing scattering sites and severely degrad-
ing the SWCNTs device performance.
The hydroxoxamic acid functionality was chosen due to its
ability to discriminate between basic metal oxides and silicon
oxide surfaces.12,18 Metal oxides with large isoelectric points
(IEP) are able to deprotonate the weakly acidic hydroxamic acids
(pKa ) ∼9.3), resulting in a hydroxamate ion that can more
readily chelate to the oxide surface. Using the patterned
substrates, the weakly bonding hydroxamic acid species exists
at the acidic SiO2 surface (IEP ) ∼2), while the mostly
deprotonated, strongly bonding hydroxamate species exists on
the more basic HfO2 surface (IEP ) ∼8), leading to selective
binding to the basic metal oxide surface. The pH can be carefully
tuned to mediate this acid-base interaction to maximize the
binding density, while keeping the functionalized tubes stably
suspended. This, coupled with the high oxidation state of the
hafnium ion (4+), leads to strong and selective binding of the
f-SWCNTs to the HfO2 surface. It is also important to note that
using dielectrics deposited via atomic layer deposition (ALD)
yielded a higher density than those deposited via electron beam
deposition. ALD yields atomically smooth surfaces and therefore
a more homogeneous surface structure. It is presumably this
homogeneity that results in these superior results. To take
advantage of the binding selectivity, substrates were lithographi-
cally patterned to expose HfO2 channels for the placement of
the modified SWCNTs, as illustrated in Figure 3a. Trenches of
various line widths were fabricated by chemically etching the
SiO2 overlayer through a photoresist mask. The substrates were
then exposed to a solution of the f-SWCNTs for 3 h, removed
from the solution, and rinsed thoroughly with deionized water
before being imaged with scanning electron and atomic force
microscopes. The images (Figure 3b-d) reveal the functional-
ized carbon nanotubes assembled strictly in the region with
exposed HfO2. The rest of the surface, composed of SiO2, was
completely free of any f-SWCNTs.
Raman spectroscopy is a powerful tool for probing the
dynamics of SWCNT functionalization.19,20 Figure 2 shows the
Raman spectra, at 488 nm excitation, of pristine SWCNTs prior
to any chemical treatment (black), of SWCNTs functionalized
with 1 (red), and of f-SWCNTs annealed to 600 °C (blue). The
pristine SWCNTs (Figure 2, black trace) show all the requisite
peaks associated with typical Raman spectra of SWCNTs.19,20
The radial-breathing mode (RBM), associated with the ring
vibration perpendicular to the long axis of the nanotube, is
assigned to the peak at ∼220 cm-1, while the peaks at 1590
(13) Bahr, J. L.; Yang, J.; Kosynkin, D. V.; Bronikowski, M. J.; Smalley, R.
E.; Tour, J. M. J. Am. Chem. Soc. 2001, 123, 6536.
(14) Kooi, S. E.; Schlecht, U.; Burghard, M.; Kern, K. Angew. Chem., Int. Ed.
2002, 41, 1353.
(15) Niyogi, S.; Hamon, M. A.; Hu, H.; Zhao, B.; Bhowmik, P.; Sen, R.; Itkis,
M. E.; Haddon, R. C. Acc. Chem. Res. 2002, 35, 1105.
(16) Heald, C. G. R.; Wildgoose, G. G.; Jiang, L.; Jones, T. G. J.; Compton, R.
G. ChemPhysChem 2004, 5, 1794.
(17) Banerjee, S.; Hemraj-Benny, T.; Wong, S. S. AdV. Mater. 2005, 17, 17.
(18) Folkers, J. P.; Gorman, C. B.; Laibinis, P. E.; Buchholz, S.; Whitesides,
G. M.; Nuzzo, R. G. Langmuir 1995, 11, 813.
(19) Dresselhaus, M. S.; Dresselhaus, G.; Jorio, A.; Souza Filho, A. G.; Pimenta,
M. A.; Saito, R. Acc. Chem. Res. 2002, 35, 1070.
(20) Jorio, A.; Saito, R.; Hertel, T.; Weisman, R. B.; Dresselhaus, G.;
Dresselhaus, M. S. MRS Bull. 2004, 29, 276.
9
11966 J. AM. CHEM. SOC. VOL. 129, NO. 39, 2007