Optical properties of ultrashort semiconducting single-walled carbon nanotube capsules down to sub-10 nm.
ABSTRACT Single-walled carbon nanotubes (SWNTs) are typically long (greater than or approximately equal 100 nm) and have been well established as novel quasi one-dimensional systems with interesting electrical, mechanical, and optical properties. Here, quasi zero-dimensional SWNTs with finite lengths down to the molecular scale (7.5 nm in average) were obtained by length separation using a density gradient ultracentrifugation method. Different sedimentation rates of nanotubes with different lengths in a density gradient were taken advantage of to sort SWNTs according to length. Optical experiments on the SWNT fractions revealed that the UV-vis-NIR absorption and photoluminescence peaks of the ultrashort SWNTs blue-shift up to approximately 30 meV compared to long nanotubes, owing to quantum confinement effects along the length of ultrashort SWNTs. These nanotube capsules essentially correspond to SWNT quantum dots.
- [show abstract] [hide abstract]
ABSTRACT: The band gaps of self-assembled single-walled carbon nanotube (SWNT) films have been determined through curve fitting using the semi-empirical Tauc and Davis–Mott model, based on the measurement of optical absorption at the visible and near infrared range. This study provides a practicable option for the determination of band gaps for ultra-thin SWNT films or multi-walled carbon nanotube films whose vHs peaks cannot be well resolved in absorption spectra.Applied Physics A 04/2012; 97(2):341-344. · 1.55 Impact Factor
- [show abstract] [hide abstract]
ABSTRACT: CdS nanorods have been sorted by length using a density gradient ultracentrifuge rate separation method. The fractions containing longer rods showed relatively stronger oxygen-related surface trap emission, while the shorter ones had dominant band-edge emission. These results suggest that the final length distribution of CdS nanorods is not a result of random nucleation and growth, but is related to the local synthesis conditions. Inspired by these findings, different synthesis environments (N2, air, and O2) have been employed in order to tailor the length distribution. In addition to the rod length, the photoluminescence properties of CdS nanorods can also be manipulated. Increasing the oxygen partial pressure significantly changed the growth behavior of CdS nanorods by improving the anisotropic growth. KeywordsNanoseparation–quantum rods–CdS–photoluminescence–controlled synthesisNano Research 04/2012; 4(2):226-232. · 7.39 Impact Factor
- [show abstract] [hide abstract]
ABSTRACT: High quality gold nanorods (NRs) with a monodisperse size and aspect ratio are essential for many applications. Here, we describe how nearly monodisperse gold NRs can be separated from polydisperse samples using density gradient ultracentrifugation. Size and dimension analysis by transmission electron microscopy (TEM) and absorption spectroscopy revealed that the Au NRs were separated mainly as a function of their aspect ratio. The surface-enhanced Raman scattering (SERS) activity of Au NRs with lower aspect ratio is notably stronger than that of NRs with higher aspect ratio under 633 nm laser excitation, due to the size-dependent absorption of the longitudinal plasmon band. The separation approach provides a method to improve the quality of NRs produced by large scale synthetic methods. KeywordsAu nanorods–aspect ratio–density gradient–separation–surface-enhanced Raman scattering (SERS)Nano Research 04/2012; 4(8):723-728. · 7.39 Impact Factor
Optical Properties of Ultrashort Semiconducting
Single-Walled Carbon Nanotube Capsules Down to Sub-10 nm
Xiaoming Sun,†Sasa Zaric,‡Dan Daranciang,‡Kevin Welsher,‡Yuerui Lu,‡
Xiaolin Li,‡and Hongjie Dai*,‡
Department of Chemistry and Laboratory for AdVanced Materials, Stanford UniVersity,
Stanford, California 94305, and Chemistry Department, Beijing UniVersity of Chemical
Technology, Beijing 100029, P. R. China
Received January 28, 2008; E-mail: firstname.lastname@example.org
Abstract: Single-walled carbon nanotubes (SWNTs) are typically long (J100 nm) and have been well
established as novel quasi one-dimensional systems with interesting electrical, mechanical, and optical
properties. Here, quasi zero-dimensional SWNTs with finite lengths down to the molecular scale (7.5 nm
in average) were obtained by length separation using a density gradient ultracentrifugation method. Different
sedimentation rates of nanotubes with different lengths in a density gradient were taken advantage of to
sort SWNTs according to length. Optical experiments on the SWNT fractions revealed that the UV-vis-NIR
absorption and photoluminescence peaks of the ultrashort SWNTs blue-shift up to ∼30 meV compared to
long nanotubes, owing to quantum confinement effects along the length of ultrashort SWNTs. These
nanotube capsules essentially correspond to SWNT quantum dots.
Single-walled carbon nanotubes (SWNTs) are rolled up from
graphene sheets and have shown unique electronic, optical, and
mechanical properties.1–3Although the dependence of SWNT
physical properties on chirality has been extensively probed both
theoretically and experimentally,1,4less has been done for
length-dependent properties, including optical absorption and
photoluminescence. The ultimate miniaturization of SWNTs is
to produce nanotube capsules with nanometer-scale lengths.
Theory predicts that the bandgap of a nanotube increases with
shorter length in an oscillatory manner. The oscillation amplitude
increases as the nanotube length decreases. The underlying
physics is quantum confinement along the tube length as the
nanotube approaches zero-dimensional sizes. This is similar to
the quantum size effects observed in other semiconductor
systems.5–9These changes should be experimentally observable
for nanotube lengths below several tens of nanometers.5–7,10
Scanning tunneling microscope and spectroscopy methods
probed metallic nanotubes cut down to 30 nm length and indeed
showed the increase in the spacing between energy levels of
quantum states confined along the nanotube axis for shorter
UV-vis-NIR interband absorption and photoluminescence
(PL) spectroscopy methods have proven to be powerful tools
for characterizing SWNTs. Results have shown spectral features
related to the first and second subband gaps of semiconducting
nanotubes present in the sample,4,13and for the precise spectral
peak positions, exciton effects had to be taken into account.14
Optical probing of the length regime where strong quantum
confinement is expected has not been possible so far because
of SWNT sample preparation and separation limitations,
although a number of analytical methods have been used for
length, diameter, or chirality separation, such as size exclusion
chromatography,15,16ion exchange chromatography,17,18elec-
trophoresis19(or dielectrophoresis),20isopycnic density gradient
separation,21,22or combination of these methods.19,23The
†Beijing University of Chemical Technology.
(1) Dresselhaus, M. S.; Dresselhaus, G.; Eklund, P. C. Science of
Fullerenes and Carbon Nanotubes; Elsevier: San Diego, 1996.
(2) Baughman, R. H.; Zakhidov, A. A.; de Heer, W. A. Science 2002,
(3) Dai, H. J. Acc. Chem. Res. 2002, 35, 1035–1044.
(4) Bachilo, S. M.; Strano, M. S.; Kittrell, C.; Hauge, R. H.; Smalley,
R. E.; Weisman, R. B. Science 2002, 298, 2361–2366.
(5) Zhu, H. Y.; Klein, D. J.; Schmalz, T. G.; Rubio, A.; March, N. H. J.
Phys. Chem. Solids 1998, 59, 417–423.
(6) Liu, L.; Jayanthi, C. S.; Guo, H.; Wu, S. Y. Phys. ReV. B 2001, 6403,
(7) Rochefort, A.; Salahub, D. R.; Avouris, P. J. Phys. Chem. B 1999,
(8) Rubio, A.; Sanchez-Portal, D.; Artacho, E.; Ordejon, P.; Soler, J. M.
Phys. ReV. Lett. 1999, 82, 3520–3523.
(9) Lu, D. Y.; Li, Y.; Rotkin, S. V.; Ravaioli, U.; Schulten, K. Nano Lett.
2004, 4, 2383–2387.
(10) Wu, J.; Duan, W. H.; Gu, B. L.; Yu, J. Z.; Kawazoe, Y. Appl. Phys.
Lett. 2000, 77, 2554–2556.
(11) Venema, L. C. Appl. Phys. Lett. 1997, 71, 2629.
(12) Venema, L. C.; Wildoer, J. W. G.; Janssen, J. W.; Tans, S. J.; Tuinstra,
H. L. J. T.; Kouwenhoven, L. P.; Dekker, C. Science 1999, 283, 52–55.
(13) O’Connell, M. J.; Bachilo, S. M.; Huffman, C. B.; Moore, V. C.;
Strano, M. S.; Haroz, E. H.; Rialon, K. L.; Boul, P. J.; Noon, W. H.;
Kittrell, C.; Ma, J. P.; Hauge, R. H.; Weisman, R. B.; Smalley, R. E.
Science 2002, 297, 593–596.
(14) Spataru, C. D.; Ismail-Beigi, S.; Benedict, L. X.; Louie, S. G. Phys.
ReV. Lett. 2004, 92, 077402.
(15) Huang, X. Y.; McLean, R. S.; Zheng, M. Anal. Chem. 2005, 77, 6225–
(16) Hennrich, F.; Krupke, R.; Arnold, K.; Stutz, J. A. R.; Lebedkin, S.;
Koch, T.; Schimmel, T.; Kappes, M. M. J. Phys. Chem. B 2007, 111,
(17) Zheng, M.; Jagota, A.; Strano, M. S.; Santos, A. P.; Barone, P.; Chou,
S. G.; Diner, B. A.; Dresselhaus, M. S.; McLean, R. S.; Onoa, G. B.;
Samsonidze, G. G.; Semke, E. D.; Usrey, M.; Walls, D. J. Science
2003, 302, 1545–1548.
10.1021/ja8006929 CCC: $40.75 XXXX American Chemical Society
Published on Web 04/22/2008
J. AM. CHEM. SOC. XXXX, xxx, 000 9 A
shortest length obtained is typically ∼50 nm. Recently, Fagan
et al. obtained nanotube fractions down to 10 nm in length24
but observed no shifts in SWNT optical spectra as length was
Here, we report length separation of SWNTs by centrifugal
rate separation in a density gradient. We obtain nanotube
samples with average lengths down to ∼7.5 nm and reasonable
length distribution. The samples were characterized by atomic
force microscopy (AFM) and optical absorption and PL-
excitation spectroscopy methods. As the length of SWNTs is
reduced below 50 nm, the optical spectra show clear blue shifts,
demonstrating for the first time that optical peak positions of
ultrashort SWNTs are dependent on length because of quantum
Results and Discussions
The details of sample preparation and length separation of
HiPco SWNTs are given in the Supporting Information (part
I). In brief, a layer of 0.2 mL of sonication-cut SWNTs wrapped
with surfactant phospholipid (PL) based distearoyl-sn-glycero-
solution density gradient and centrifuged at ultrahigh speed
(∼300kg) for 3 h (see photograph of the ultracentrifuge tube
after separation in Supporting Information, Figure S1A). The
gradient containing separated SWNTs was manually sampled
and fractioned (100 µL each) from the centrifuge tube for
Tapping mode AFM images of typical fractions (see Figure
1) indicated that fraction 6 (labeled as f6 in Figure 1) contained
the shortest SWNTs. After AFM tip size correction (see
Supporting Information), the SWNTs showed an average length
of ∼7.5 nm (longest, 16 nm and shortest, ∼2 nm). The average
length of subsequent fractions (f8, f12, and f18) gradually
increased to 11, 27, and 58 nm, respectively, with a standard
deviation of 45-55%. Length histograms of the various fractions
obtained are shown next to the corresponding AFM images in
The method used in our work was different from isopycnic
density gradient separation used to separate SWNTs according
to diameter and bandgap.21,22Multilayer gradient (5, 7.5, 10%
layers) was used here, but all layers had a density smaller than
that of SWNT for separation; thus, all SWNTs had a chance to
go through the gradient, and only the time required to go through
the density gradient depended on the length of the nanotubes.
By making use of the different sedimentation rates of different
length SWNTs in density gradient and by terminating the
sedimentation by removing centrifugal force, different length
SWNTs are captured along the centrifuge tube at different spatial
locations. Longer SWNTs are located nearer to the bottom of
the centrifuge tube, and shorter SWNTs are nearer to the top.
The density gradient needs to be well designed. Isopycnic
separation experiment indicated that individual SWNTs wrapped
with PL-PEG have a density of ∼1.12 g/cm3, equal to ∼22%
iodixanol (Supporting Information, Figure S2). Therefore, the
gradient layers used should have a density <22%. As control
experiments indicated, iodixanol solution without density gradi-
ent could also separate according to length to a certain degree
but was poor in performance. Too-sharp an increase in density
gradient (e.g., 10%) would make tubes accumulate at boundary
areas and impair the separation (Supporting Information, Figure
S3A). We found that when the gradient was shallow (e.g.,
interlayer difference e5%) and the centrifugation time was long
enough (e.g., g2 h), the interlayer boundary became invisible
after ultracentrifugation. In that case, the step gradient gave
(18) Zheng, M.; Jagota, A.; Semke, E. D.; Diner, B. A.; Mclean, R. S.;
Lustig, S. R.; Richardson, R. E.; Tassi, N. G. Nat. Mater. 2003, 2,
(19) Heller, D. A.; Mayrhofer, R. M.; Baik, S.; Grinkova, Y. V.; Usrey,
M. L.; Strano, M. S. J. Am. Chem. Soc. 2004, 126, 14567–14573.
(20) Krupke, R.; Hennrich, F.; von Lohneysen, H.; Kappes, M. M. Science
2003, 301, 344–347.
(21) Arnold, M. S.; Stupp, S. I.; Hersam, M. C. Nano Lett. 2005, 5, 713–
(22) Arnold, M. S.; Green, A. A.; Hulvat, J. F.; Stupp, S. I.; Hersam, M. C.
Nat. Nanotechnol. 2006, 1, 60–65.
(23) Zheng, M.; Semke, E. D. J. Am. Chem. Soc. 2007, 129, 6084–6085.
(24) Fagan, J. A.; Simpson, J. R.; Bauer, B. J.; Lacerda, S. H. D.; Becker,
M. L.; Chun, J.; Migler, K. B.; Walker, A. R. H.; Hobbie, E. K. J. Am.
Chem. Soc. 2007, 129, 10607–10612.
(25) Liu, Z.; Sun, X. M.; Nakayama-Ratchford, N.; Dai, H. ACS Nano
2007, 1, 50–56.
Figure 1. AFM images of various SWNT fractions after centrifugal rate
separation (f6 means fraction 6 and so on) in gradient. The corresponding
length histograms of the fractions measured from 50 tubes in each fraction
are shown next to the corresponding AFM images.
B J. AM. CHEM. SOC. 9 VOL. xxx, NO. xx, XXXX
Sun et al.
separation results as good as results from the linear gradient.
To obtain primarily short tubes, we used a gradient with
relatively low density and long centrifugation times. We chose
5 + 7.5 + 10% three-layer step density gradient to effectively
control the sedimentation speed of individual SWNTs 2-50 nm
in length, thus providing higher resolution in rate separation of
short tubes. This design was similar to that used for biomac-
romolecule separation according to mass.26We should note that
the rational design on gradient should combine with good timing
for centrifugation. Too-long centrifugation time in the above
gradient will cause SWNTs with different lengths to settle down
near the bottom of the centrifuge tube (Supporting Information,
Figure S3B), whereas relatively long SWNTs can be obtained
by significantly shortened ultracentrifugation time (Supporting
Information, Figure S4A).
The UV-vis-NIR absorption spectra of several fractions are
shown in Figure 2A. The absorption peaks in 900-1500 nm
range are due to the lowest E11subband absorption, whereas
the peaks in 550-900 nm range correspond to their second E22
subband transitions4,13Multiple peaks corresponding to various
semiconducting SWNTs are present in the sample. Both first
and second subband absorption peaks show blue shifts as the
average nanotube length is reduced. The shift was continual
and monotonic: shorter fractions showed stronger blue shift.
By using a convenient excitation wavelength in the second
subband absorption range, a small number of chiralities are
selectively excited, and their PL in the first subband range can
be measured. Because of the reduced number of overlapping
PL peaks compared to the number of absorption peaks of various
(n,m) SWNTs in the first subband region, the measured NIR-PL
spectra show better defined spectral features. NIR-PL of various
fractions measured by using 720 nm excitation is shown in
Figure 2B. The three most prominent peaks correspond to (9,4),
(8,6), and (8,7) nanotubes. All three show blue shifts as the
fraction average length is reduced, in agreement with the
By measuring NIR-PL of various semiconducting (m,n)
SWNTs over a range of excitation wavelengths, length-
dependent spectral shifts for a number of semiconducting
chiralities present in the sample are clearly gleaned. Figure 3
shows resulting PL excitation (PLE) spectra for several fractions.
The PLE peaks were labeled by their chirality assignment.4All
PLE peaks blue-shifted as the fraction length was reduced both
in PL direction and excitation direction (in agreement with shifts
observed in first and second subband absorption data, respec-
(26) Price, C. A. Centrifugation in Density Gradients; Academic Press:
New York, 1982.
Figure 2. Optical characterization of length separated SWNT fractions.
(A) UV-vis-NIR absorption spectra after normalization to nanotube
concentration (on the basis of absorbance at 935 nm). (B) NIR-PL spectrum
under 720 nm excitation after normalization to concentration. Three SWNT
chiralities, (9,4), (8,6), and (8,7), were selectively excited. Blue shifts in
spectral peaks are seen in both absorption and PL spectra for shorter-length
Figure 3. PLE spectra of various SWNT fractions after length separation.
The vertical lines are guide to the eye. Chirality assignments were marked
for selected PLE peaks.
J. AM. CHEM. SOC. 9 VOL. xxx, NO. xx, XXXX
Optical Properties of Ultrashort SWNTs
Length-dependent PLE peak positions for several chiralities
are given in Figure 4A,B (the data for all measured chiralities
are given in the Supporting Information, Figure S5). The blue
shifts are highly pronounced for nanotube lengths below ∼20
nm. Dotted lines in Figure 4 represent PLE peak positions for
long nanotubes (∼140 nm. See Supporting Information, Figure
S4), which were obtained through a modified procedure by using
shorter sonication for cutting SWNTs and shorter centrifugation
time to fractionate long tubes. The peak shifts between PL peak
positions for the shortest fraction (f6) and the fraction containing
long nanotubes (denoted ∆E11,f6) for all measured chiralities
are shown in Figure 4C (excitation peak shifts ∆E22,f6are given
in the Supporting Information, Figure S5). Both ∆E11,f6 and
∆E22,f6ranged up to ∼30 meV, with the chirality averaged value
of 24 and 21 meV, respectively. Note that the ∆E11,f6and ∆E22,f6
peak shifts were generally smaller for smaller diameter nanotubes.
Before we can assign the observed spectral blue shifts to finite
length effects, we need to exclude other reported effects that
can lead to spectral peak shifts in SWNTs, such as bundling of
nanotubes13and the influence of the surrounding medium.27,28
The bundling of SWNTs has been shown to cause spectral red
shifts13and the decrease of PL yield.29We measured height
histograms of our SWNT fractions by using AFM and identified
some bundling in fraction f18 (see Supporting Information,
Figure S6), but no bundles were observed in shorter fractions
(f6, f8, and f10), where most of the blue shift happened. This
was further enforced by calculating relative PL quantum yield
from the measured absorption and PLE data, which showed
higher PL quantum yield for fractions f6-f10 relative to fraction
f18. Because all nanotubes had the same surfactant coating, the
only remaining difference in the surrounding medium of
different fractions was the percentage of iodixanol used to create
density gradient. The difference in iodixanol content between
fractions was less than 5%. To verify that iodixanol did not
influence the observed peak positions, PLE of HiPco/PL-PEG
nanotube samples was measured in the absence and presence
of 10% iodixanol, and no shifts were observed (see Supporting
Information, Figure S7).
We attribute the observed length-dependent spectra blue shifts
of SWNTs to finite length effects. Theoretical modeling (see
details of modeling in Supporting Information, Figures S9 and
S10) was carried out to obtain a good agreement between the
observed spectral shifts and theoretically predicted bandgap
changes due to quantum confinement along SWNT length.6,7,10
The black curves in Figure 4A,B are results of theoretical
modeling. Our calculated spectra shifts were based on SWNT
bandgap changes for each (m,n) as a function of length and
averaged over the length distribution of SWNTs in each fraction
(see Supporting Information). A monotonic bandgap change was
therefore seen for fractions with decreasing average tube length,
without oscillatory behavior between fractions because of
averaging over the length distribution. Bandgap change for a
(m,n) tube as a function of length was due to increased
quantization of states along the nanotube axis. The spectral shifts
were more pronounced for larger diameter SWNTs because of
the smaller bandgap and higher curvature of the E(k) dispersion
curve at the bandgap edge of larger tubes. This was consistent
with the observed spectral shifts being more pronounced for
larger diameter tubes than for smaller tubes (see Figure 4C).
Length distribution in each fraction, together with the increasing
bandgap oscillations as the nanotube length is decreased, might
have caused broadened PLE peaks, as experimentally observed
(see Figure 3 and the experimental length dependence of widths
in the Supporting Information, Figure S8). It is also possible
that decreased exciton lifetime as nanotube length approaches
the exciton size (several nanometers)14,30,31caused spectral peak
broadening. Interestingly, it is also observed that type-I tubes
[i.e., (m - n) mod 3 ) 1] showed spectral shifts larger than
those of type-II SWNTs [(m - n) mod 3 ) -1], especially in
the small diameter region (j1 nm, Figure 4C). This was
consistent with the fact that type-I SWNTs exhibited smaller
E11 bandgaps or higher band-edge curvatures than type-II
SWNTs of similar diameters because of trigonal warping effects,
which are more pronounced for small diameter tubes.32
The finite length effects described in this work were not
observed previously with separated CoMoCAT nanotubes.24We
(27) Moore, V. C.; Strano, M. S.; Haroz, E. H.; Hauge, R. H.; Smalley,
R. E.; Schmidt, J.; Talmon, Y. Nano Lett. 2003, 3, 1379–1382.
(28) Ohno, Y.; Iwasaki, S.; Murakami, Y.; Kishimoto, S.; Maruyama, S.;
Mizutani, T. Phys. ReV. B 2006, 73, 235427.
(29) Crochet, J.; Clemens, M.; Hertel, T. J. Am. Chem. Soc. 2007, 129,
(30) Chang, E.; Bussi, G.; Ruini, A.; Molinari, E. Phys. ReV. Lett. 2004,
(31) Perebeinos, V.; Tersoff, J.; Avouris, P. Phys. ReV. Lett. 2004, 92,
Figure 4. PLE peak positions of (8,4), (8,6), (8,7), and (9,7) tubes as
functions of tube length. (A) E11-PL peak positions vs length. (B) E22-
excitation peak position vs length. Black curves are modeling results (see
Supporting Information, Figure S10). (C) Dependence of the E11,f6shift on
tube diameter. For a (m,n) tube, type I means (m - n) mod 3 ) 1, and type
II means (m - n) mod 3 ) -1.
D J. AM. CHEM. SOC. 9 VOL. xxx, NO. xx, XXXX
Sun et al.
explain this by the smaller shifts observed for smaller diameter
nanotubes present in CoMoCAT samples (see Figure 4C) and
the shorter nanotube lengths of our samples. Also, the observa-
tion of systematic blue shifts requires that other factors, such
as bundling and too-wide length distributions, are excluded.
A new separation method, rate separation in density gradient,
was developed to sort 7-60 nm SWNTs acording to length. It
led to SWNTs less than 10 nm in length. Optical experiments
on the SWNT fractions revealed that the UV-vis-NIR absorp-
tion and PL peaks of the ultrashort SWNTs blue-shift up to
∼30 meV compared to long nanotubes, owing to quantum
confinement effects along the length of ultrashort SWNTs,
essentially corresponding to SWNT quantum dots.
Acknowledgment We thank Dr. Jie Deng for helpful discus-
sion. This work was supported by MARCO-MSD and Intel.
Supporting Information Available: Experimental details,
supplementary figures, and details of modeling of SWNT
bandgap vs length. This material is available free of charge via
the Internet at http://pubs.acs.org.
(32) Reich, S.; Thomsen, C.; Maultzsch, J. Chapter I. Carbon Nanotubes:
Basic Concepts and Physical Properties; Wiley-VCH: Weinheim,
J. AM. CHEM. SOC. 9 VOL. xxx, NO. xx, XXXX
Optical Properties of Ultrashort SWNTs