Investigation of electrical transport in hydrogenated multiwalled carbon nanotubes

Abstract

Highly disordered multiwalled carbon nanotubes of large outer diameter (∼60 nm) fabricated by means of chemical vapor deposition process inside porous alumina templates exhibit ferromagnetism when annealed in a H2/Ar atmosphere. In the presence of an applied magnetic field, there is a transition from positive to negative magnetoresistance. The transition may be explained in terms of the Bright model for ordered and disordered carbon structures. Additionally, temperature dependent electrical transport experiments exhibit a zero-bias anomaly at low temperature.

Full text

Page 1

Investigation of electrical transport in hydrogenated multiwalled
carbon nanotubes
Adam L. Friedmana,b,n, Hyunkyung Chunc, Don Heimana, Yung Joon Jungc, Latika Menona
aDepartment of Physics, Northeastern University, Boston, MA 02115, USA
bCode 6876, US Naval Research Laboratory, Washington, DC 20375, USA
cDepartment of Mechanical and Industrial Engineering, Northeastern University, Boston, MA 02115, USA
a r t i c l e i n f o
Article history:
Received 30 November 2010
Accepted 2 December 2010
Keywords:
Carbon nanotubes
Electrical transport
Ferromagnetism
Carbon
a b s t r a c t
Highly disordered multiwalled carbon nanotubes of large outer diameter (?60 nm) fabricated by means
of chemical vapor deposition process inside porous alumina templates exhibit ferromagnetism when
annealed in a H2/Ar atmosphere. In the presence of an applied magnetic field, there is a transition from
positive to negative magnetoresistance. The transition may be explained in terms of the Bright model for
ordered and disordered carbon structures. Additionally, temperature dependent electrical transport
experiments exhibit a zero-bias anomaly at low temperature.
& 2010 Elsevier B.V. All rights reserved.
There are many reports on the electrical transport properties of
carbon nanotubes (CNTs). Most of them exhibit a zero-bias
anomaly (ZBA) associated with Luttinger liquid (LL) behavior in
highly ordered CNTs [1,2]. However, some have suggested that
apart from a LL explanation, for quasi-ballistic single electron
junctions in disordered conductors, a ZBA can arise due to either
electron or plasmon scattering near the tunnel barrier and finite
size effects or due to extremely high resistance contacts or
transmission lines [3–6]. However, there are very few studies
reported to date on the transport properties of disordered multi-
wall CNTs (MWCNTs). Initial measurements indicate that the
electrical transport properties of disordered MWCNTs can be
explained in terms of weak localization theory [7,8]. Weak
localization or Anderson localization also predicts that in the
presence of an applied magnetic field a sufficiently disordered
nanotube will exhibit negative magnetoresistance (MR) [9]. How-
ever, the magnetotransport behavior, even for well-ordered nano-
tubes, appears to be dominated by weak localization effects rather
than the LL [9]. This indicates that an applied magnetic field
effectively destroysthe LL. However, in order to observe such weak
localization effects, the temperature must be sufficiently low. In
most previous studies, negative MR was only observed for tem-
peratures below 5 K for ordered CNTs [9].
In this work, we report results of electrical transport measure-
ments on highly disordered MWCNTs fabricated by means of
chemical vapor deposition (CVD) inside nanoporous alumina
templates. Earlier we reported that ferromagnetism in such
nanotubes can be induced by annealing in hydrogen [12]. The
nature of this ferromagnetism has been discussed in the literature
[13–16]. In this work, we study temperature dependent transport
properties of ferromagnetic nanotubes. We show that at low
temperatures and in zero magnetic field, the nanotubes exhibit a
ZBA. MR measurements for the ferromagnetic MWCNTs show
positive MR at low temperatures, in some cases up to ?40 K,
beyond which they exhibit negative MR. Such a transition in MR is
not seen in non-ferromagnetic MWCNTs annealed without hydro-
gen. This transition may be attributed to increased disorder and
may possibly be explained in terms of the Bright model [10].
The MWCNTs are synthesized inside nanoporous alumina
templates fabricated by anodization of aluminum foil [17]. The
template acts as a deposition substrate for the MWCNTs yielding
nanotubes with outer diameter (?60 nm for our study) corre-
sponding with the pore diameter of the template. MWCNTs are
synthesized without catalyst in the templates in a CVD process at
660 1C with acetylene acting as the precursor gas. By adjusting the
CVD time, we control the inner diameter of the MWCNTs. Here, we
make MWCNTs with four different inner diameters using CVD
times of 60, 75, 90, and 100 min. The right inset of Fig. 1 shows the
inner diameter of MWCNTs as a function of CVD time.
The nanotubes are then removed from the templates by dissol-
ving the alumina in 15% sulfuric acid heated to 100 1C. The acid is
removed from the solution and replaced with 99.5% ethanol. The
nanotubes are then dispersed in the solution by sonicating for 3 s in
an ultrasonic bath. TEM studies show that the tubes are polycrystal-
line and extremely disordered (left inset of Fig. 1). However, the
nanotubes retain the SP2 bonding characteristic of a carbon nano-
tube,asevidencedbyXPSstudies[12].Adropofnanotubesolutionis
Contents lists available at ScienceDirect
journal homepage: www.elsevier.com/locate/physb
Physica B
0921-4526/$-see front matter & 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.physb.2010.12.009
nCorresponding author at: Code 6876, US Naval Research Laboratory, Electronic
Materials and Devices, 4555 Overlook Avenue SW, Washington, DC 20375, USA.
Tel.: +1 202 404 4573.
E-mail address: adam.friedman.ctr@nrl.navy.mil (A.L. Friedman).
Please cite this article as: A.L. Friedman, et al., Physica B (2010), doi:10.1016/j.physb.2010.12.009
Physica B ] (]]]]) ]]]–]]]

Page 2

deposited onto a Si/SiO2substrate and then the substrate is heated
slightly until the ethanol is evaporated out resulting in a well-
dispersed,sparsefilmofnanotubeson the Si/SiO2.The chipsare then
annealedinatubefurnacecontainingequalpartsArandH2at800 1C,
which induces ferromagnetism [12]. For comparison, some chips
wereannealedinAronly.Itwasfoundthatannealingalsolowersthe
contact resistances. Un-annealed devices resulted in 2-probe resis-
tancesof1–100 MO,whilethoseannealedhad2-proberesistancesof
4–60 kO. Lower annealing temperatures were tried, however, this
did not successfully lower the contact resistance. Moreover, the
temperaturenecessarytomake andbreakC–Hbondsisestimated to
be 600 1C [18]. We are well above this limit. Furthermore, TEM
studies on the annealed tubes show that the polycrystalline and
disorderednatureofthetubesisnegligiblyaltered.Wefoundthatthe
structure of the un-annealed tubes, H2/Ar annealed, and Ar annealed
tubes to be identical. This is further supported by studies that
consider 800 1C to be too low for significant graphitization [11].
Therefore, improved crystal ordering and enhanced transport effects
as a result of annealing are unlikely.
For electrical transport measurements, 2-probe devices are
made by photolithography. We also tested 4-probe devices and
found that the contact resistances are low and those 2-probe
devices, being easier to fabricate, will suffice for our experiments.
The gap between electrodes is approximately 3 mm. The nanotubes
are disordered to the extent that they do not contain continuous
walls over this measuring distance. Cr/Au contacts of 5/60 nm are
deposited in an electron beam evaporator. The photolithography
method makes many devices simultaneously on a single chip.
Suitable devices were chosen where only a few nanotubes bridge
the contacts. We found almost no difference in conductivity
between devices where 1, 2, or 3 nanotubes bridged the contacts.
Fig. 1 displays a scanning electron microscopy (SEM) image of one
such device. For MR measurements, we use a 2 terminal DC set up.
Measurements are made using a 14 T cryo-free superconducting
magnetic system and a variable temperature cryostat insert. The
magnetic field is applied normal to the long axis of the nanotube.
Fig.2showstheresistanceversustemperature,R(T),ofsometypical
hydrogen annealed and argon-only annealed nanotubes for four
different tube diameters. We see that for all samples, the resistance
versus temperature curve is punctuated by a ‘‘knee’’ at approximately
14 K.Beyondthisknee,theresistanceincreasesdramatically.Wemade
several batches of devices, and the behavior was consistent across all
devices made at different times. The position of the knee (which is
consistentlyatabout13or14 Kforallsamplesstudied)isindependent
of all other observed behavior in the nanotubes.
The conductance as a function of bias voltage, G(V), for a typical
hydrogenannealedsample and anargon-only annealedsample are
alsoshowninFig.2.Ondecreasingtemperature,theconductivityat
V¼0 decreases toward G¼0. One possible explanation for the ZBA
istheLuttingermodel.LLtheorypredictsnear-ohmicbehaviornear
room temperature with the development of a singularity in the
conductance curve as the temperature decreases, which at low
enough temperatures eventually become a conductance gap. We
can see that, indeed, our samples show the formation of the
predicted conductance gap, while approaching ohmic behavior
at higher temperatures; all the samples tested showed this
behavior. The depth of the conductance singularity and the
temperature at which the sample becomes ohmic appears to be
independent of inner diameter of the CNT. Also, this behavior
persisted for both Ar and H2annealed devices.
The strongly correlated electron system of a LL results in a power
law depends on the density of states as a function of energy which in
turn leads to a differential conductance, GLL¼dI/dV, given by [1,2]
GLLðV,TÞ ¼ ATaG zþ1
22
??
???? ????
2
sinh
x
? ?1
2coth
x
2
? ?
?1
pImC zþ1
2
????
ð1Þ
where
z¼ 1þa=2þix=2p
x¼ZeV=kBT
Thus,themodelhasthreeparameters:a,Z,andA.Aisaconstant
andZ accounts for voltage division of the nanotube at the electrical
contacts, i.e. tunnel barriers. It should be 0.5 for two contacts. The
value of the exponential scaling factor a strongly depends on the
type of electron tunneling in the device. There are two cases:
?
abulk¼
8N
aend¼
1
4N
1
1
g?1
1
gþg?2
?
??
ð2Þ
Here,Nreferstothenumberoflayers,orconductingchannelsandg
is the so-called Luttinger interaction parameter that measures the
strength of the interaction between electrons. The resistance (and
the LL) is contained at the points of contact between the nanotube
and the electrical contact. So, the first case is that an electron
tunnels from the contact into the LL from the end (aend), in the
direction along the length of the nanotube. The second case is that
an electron tunnels from the contact into the LL from the side
(abulk), at some angle from the length of the nanotube. In any case,
for a strongly correlated system such as this, we expect go1. For
g¼1, we would have the case of a non-interacting Fermi liquid. For
the case of g51, we would have very strong repulsive electron–
electron interactions and the material would become localized.
Based on this, one can make a theoretical prediction for a.
Graugnard et al. [2] predicted a range of 0.2–0.6. Bockrath et al.
[1] predicted that aend¼0.65 and abulk¼0.24. Indeed, the conduc-
tance in our samples decreases like Taand is consistent with LL
model and with previous studies [1,2]. A detailed analysis of
various samples and the calculated a are provided in Table 1.
AnotherpredictionoftheLuttingermodel,fromEq.(1),isthataplot
of G(V,T)/(G0/Ta) versus eV/kBT should show scaling where the data
collapsesontoauniversalcurve.OurresultsforanArannealedsample
andaH2annealedsamplearedisplayedinFig.3.Theconductancedata
was fit to Eq. (1) by varying the parameters A, a, and Z to the lowest
temperature data set. We see that the Luttinger model provides a
reasonably good fit to the data. However, there are significant
deviations from the predicted universal curve as also reported in
Fig.1. Scanningelectronmicroscope(SEM)imageofadisordered,multiwallcarbon
nanotubes device. Transmission electron microscope (TEM) image (left inset) of the
disordered nanotube, and (right inset) graph of inner diameter versus chemical
vapor deposition (CVD) time.
A.L. Friedman et al. / Physica B ] (]]]]) ]]]–]]]
2
Please cite this article as: A.L. Friedman, et al., Physica B (2010), doi:10.1016/j.physb.2010.12.009

Page 3

literature by other groups [1,2], in particular, the voltage drop in the
nanotube between the contacts [2]. Table 1 summarizes the results of
fitting the data for a variety of devices. Meas. a refers to the value
obtained directly from a least squares fit of the G(T) curve. Scaled a
refers to the value taken from the Luttinger scaling fit, as in Fig. 2. We
seethat,thevaluesofa areindeedclosetothepredictedvaluesforthe
two cases ofaendandabulk. However, we see a significant deviation for
the measured and scaled values. This is most likely due to the
appearance of the knee feature in the R(T) data. Therefore, the scaled
valueismostlikelythebestmeasuredvalue. Although we expectboth
casesofendandbulktunnelingtooccur,fromthescaledvalue,wecan
determine that the majority of tunneling is end tunneling. However,
our values of scaled a are slightly higher than the predicted value for
end tunneling. This can be attributed to a finite voltage drop in
nanotubebetweenthecontactsanddisorderintheMWCNTs.Basedon
our scaleda, the Luttinger parameter, g is estimated to be approaching
the limit of what can be considered ‘‘less than 1’’ before it becomes
‘‘much, much less than 1’’. As g decreases, it corresponds to increase in
strength of electron–electron interactions. Therefore, our material can
be judged to be just barely a LL and almost in the localized electron
region of mesoscopic systems. We also note that the value ofZ widely
varies with some values above the predicted value and some well
belowthepredictedvalue.Thiscanbeattributedtothecouplingofthe
electrical contacts to the nanotube. In the case of very strong coupling
(or very low contact resistance), the value of Z becomes significantly
less than the predicted value.
ThereareotherexplanationsforZBAsindisorderedconductors.For
instance, it has been shown theoretically that in the case of high
impedance contacts or transmission lines ZBA can occur [4–6]. This is
obviously not the case for our devices, as we have very low contact
resistances and the resistance of the nanotube is very small. However,
other studies have also shown that a ZBA can occur in a disordered
conductor due to effects associated with enhanced plasmon scattering
[5,19,20]. It has been proposed that here GðTÞ ? e?w=ffiffi
measureofdisorder[20].Wefound,surprisingly,thatthesemodelsdid
not fit our data as well as the Luttinger model. Although, based on the
disorderinournanotubeswecannotdefinitelyconcludethatthereisLL
behavior, it appears to be the most likely candidate.
Fig. 4 shows the temperature dependent MR for a hydrogen
annealed and an argon-only annealed device. For the hydrogen
annealed sample, the MR is positive at 6 K. At higher temperatures,
?15 K (for this sample) there is a transition to negative MR, which
continues to remain negative for higher temperatures. This beha-
viorwas consistentfor allsamplesstudied.However,thetransition
temperature seems to vary from sample to sample, ranging from
T
p
, withw being a
Fig. 2. Resistance versus temperature, R(T), and conductance versus voltage, G(V), curves as a function of temperature for hydrogen annealed and argon-only annealed
MWCNTs.TheG(V)plotsshowtheformationofasingularityindicativeofLLbehaviorfor(upperfigure)a100 min.CVDhydrogenannealeddeviceand(lowerimage)a75 min.
argon annealed device.
Table 1
Luttinger fitting parameters for hydrogen annealed MWCNTs device.
Sample AnnealingCVD
time
(min)
Meas.
a
Scaled
a
Scaled
g
A
0A1
1A1
2B1
3A3
AR0A2
AR1A1
AR2A1
AR3A1
H2/Ar
H2/Ar
H2/Ar
H2/Ar
Ar
Ar
Ar
Ar
60
75
90
100
60
75
90
100
0.282
0.420
0.171
0.221
0.119
0.277
0.158
0.168
0.833
0.683
0.768
0.784
0.843
0.821
0.846
0.847
0.00718
0.00871
1.694
0.00454
3.439
0.0223
1.287
0.0192
0.286
0.227
0.0546
0.407
0.229
0.311
0.142
0.278
A.L. Friedman et al. / Physica B ] (]]]]) ]]]–]]]
3
Please cite this article as: A.L. Friedman, et al., Physica B (2010), doi:10.1016/j.physb.2010.12.009

Page 4

?13 to 40 K. This confirms that the transition is not related to the
observed knee in the R(T) curves. Furthermore, the transition
temperature was also found to be independent of inner diameter
of the CNTs. In the case of the non-magnetic nanotubes, the MR is
always positive for all tested samples.
The MR transition in the ferromagnetic nanotubes is most
probably a result of an order-to-disorder transition as predicted
by Bright [10]. Bright formulated a model for transport in dis-
ordered carbon structures when contemplating negative MR in
different types of carbon fibers and filaments from the amorphous
to the highly graphitized [10]. It was suggested that negative MR is
the result of formation of Landau levels at the Fermi energy, which
increases the density of states and the current path arising from
disorder in the structure of the conductor. Taking into account
several considerations for disordered carbon structures, it was
shownthattheMRofadisorderedcarbonstructureinthepresence
of an applied field B follows:
rðBÞ ¼
1þm2B2
ðnþpÞem 1þm2B2ðp?nÞ2=ðpþnÞ2
hi
ð3Þ
Here, p and n are the particle and hole carrier densities, respec-
tively, while m is the carrier mobility. It is assumed that the
application of the electric field does not affect the effective mass,
and the scattering process or mobility. From Eq. (3) the MR ratio
Dr/r¼(r(B)?r(0))/r(0) can be calculated for two different limits.
First, if n5p, then the hole concentration is proportional to the
density of states and
Dr
r
? ?157m4B2
ð4Þ
Here, values for the parameters were taken from Ref. [11]. This is
the physical case of a very disordered nanotube. Second, when
p?n
Dr
r
nþp
This is the case of a well-ordered nanotube. Here the m2B2term
dominates in high fields. Thus, the MR approaches that of a regular
metallic conductor, for which Dr=r ¼m2B2.
InourdisorderedMWCNTs,atlowtemperaturesandintheabsence
of a magnetic field, the device behaves like a LL, but only barely, as the
disorder in the sample produces many electron–electron interactions
thatnearlymakeita2-dimensionalsystem.However,uponapplication
of a magnetic field the number of interactions increase significantly,
especially at fields higher than 1 T. This is evident in Fig. 4 that all the
interesting behavior happens well above 1 T, while for low fields it is
effectively a straight line. However, despite increased interactions, at
low temperatures the device behaves like a metal. As the temperature
increases, thermal fluctuations introduce even more disorder into the
system. Eventually, the MR transitions from the more ordered case to
the disordered case become negative.
In the case of non-magnetic argon-only annealed devices,
switching on a magnetic field will not significantly increase the
number of electron–electron interactions because the necessary
ferromagnetism is absent. Therefore, the device remains metallic
and as the temperature increases, thermal fluctuations cause the
mobility to decrease gradually.
On the other hand, there are at least two additional sources of
scattering available in the hydrogen annealed nanotubes that are not
present in the argon-only annealed nanotubes. We can rule out LL
behaviorbecausethefieldwilldestroythequantumcriticalstateofthe
LL.Wecanalsoruleoutweaklocalizationbecauseweaklocalizationis
expected to be unobservable if ferromagnetism is present. In order to
clarify this, we performed electrical measurements on samples
fabricated for the application of a gate voltage and found no
dependence on gate voltage. If the behavior was due to an Anderson
transitionwe wouldexpect the behaviorofa gatevoltagesweeptobe
different before and after the transition [21]. For a metal, we expect
Coulombblockadebehaviorleadingtolowtemperatureoscillationsof
a well-defined period [1]. For a semiconductor, we would expect the
appearance of a minimum corresponding to the carrier minimum.
However, we found that the behavior was roughly the same every-
where, showing irreproducible fluctuations. Thus, the device is most
probably a metal and our measurement temperatures were not low
enough to observe well-defined oscillations. This indicates that this is
notametal–insulatorormetal–semiconductortypeoftransition.Fig.4
shows the mobilities calculated as fitting parameters for the Bright
model using Eqs. (4) and (5). However, we see the mobility of the
hydrogendevicesfalloffmorequicklyatlowertemperatures,whichis
a graphical representation of the Bright transition from order to
disorder.
In conclusion, it is shown that at low temperatures, disordered
multiwalled carbon nanotubes annealed in argon-only and in
hydrogen exhibit a ZBA, which having explored several models,
canbebestfitbytheLuttingermodel.However,ontheapplication
of a magnetic field, negative magnetoresistance is observed only
for the hydrogen annealed nanotubes that are ferromagnetic.
This behavior is compatible with a Bright order-to-disorder
transition arising from a combination of increased disorder due
to electron–electron interactions, electron–magnon interactions,
and phonon scattering. The argon-only annealed nanotubes do
?? ??{pþn, then
?
n0þp0
?1
??
þn0þp0
nþpm2B2
ð5Þ
Fig. 3. Scaled conductivity versus eV/kBT for MWCNTs at various temperatures,
representing Luttinger log–log scaling plots. The insets are not logarithmic. The
upper graph shows data for an argon-only annealed device and the lower graph
shows data for a hydrogen annealed device. For both devices, the MWCNTs were
deposited for 75 min.
A.L. Friedman et al. / Physica B ] (]]]]) ]]]–]]]
4
Please cite this article as: A.L. Friedman, et al., Physica B (2010), doi:10.1016/j.physb.2010.12.009

Page 5

not exhibit this transition because they lack several scattering
pathways found in the hydrogen annealed nanotubes.
The authors acknowledge support from NSF CAREER Grant
ECCS-0551468.
References
[1] Marc Bockrath, David H. Cobden, Jia Lu, Andrew G. Rinzler, Richard E. Smalley,
Leon Balents, Paul McEuen, Nature 397 (1999) 598.
[2] E.Graugnard,P.J.dePablo,B.Walsh,A.W.Ghosh,S.Datta,R.Reifenberger,Phys.
Rev. B 64 (2001) 125407.
[3] Vikram V. Deshpande, Marc Bockrath, Leonid I. Glazman, Amir Yacoby, Nature
464 (2010) 209.
[4] S.M. Girvin, L.I. Glazman, M. Jonson, D.R. Penn, M.D. Stiles, Phys. Rev. Lett. 64
(1990) 3183.
[5] A.M. Rudin, I.L. Aleiner, L.I. Glazman, Phys. Rev. B 55 (1997) 9322.
[6] M.H.Devoret,D.Esteve,H.Grabert,G.-L.Ingold,H.Pothier,C.Urbina,Phys.Rev.
Lett. 64 (1824) 1990.
[7] Reeta Tarkiainen, Marcus Ahlskog, Pertti Hakonen, Mikko Paalanen, Physica E
18 (2003) 206.
[8] Reeta Tarkiainen, Marcus Ahlskog, A. Zyuzin, Pertti Hakonen, Mikko Paalanen,
Phy. Rev. B 69 (2004) 033402.
[9] R. Saito, G. Dresselhaus, M.S. Dresselhaus, Physical Properties of Carbon
Nanotubes, Imperial College Press, Singapore, 1998.
[10] A.A. Bright, Phys. Rev. B 20 (1979) 5412.
[11] M.S. Dresselhaus, G. Dresselhaus, K. Sugihara, I.L. Spain, H.A. Goldberg,
Graphite Fibers and Filaments, Springer, New York, 1988.
[12] A.L. Friedman, Hyunkyung Chun, Yung Joon Jung, Don Heiman, E.R. Glaser,
Latika Menon, Phys. Rev. B 81 (2010) 115461.
[13] P.Esquinazi,D.Spemann,R.Hohne,A.Setzer,K.-H.Han,T.Butz,Phys.Rev.Lett.
91 (2003) 227201.
[14] Y. Ma,P.O. Lehtinen, A.S.Foster, R.M. Nieminen,Phys.Rev. B 72(2005) 085451.
[15] S. Talapatra, P.G. Ganesan, T. Kim, R. Vajtai, M. Huang, M. Shima, G. Ramanath,
D. Srivastava, S.C. Deevi, P.M. Ajayan, Phys. Rev. Lett. 95 (2005) 097201.
[16] Y. Zhang, S. Talapatra, S. Kar, R. Vajtai, S.K. Nayak, P.M. Ajayan, Phys. Rev. Lett.
99 (2007) 107201.
[17] A.L. Friedman, Derrick Brittain, Latika Menon, J. Chem. Phys. 127 (2007)
154717 references contained therein..
[18] Hongwei Xuesong Li, Lijie Zhu, Cailu Ci, Zongqiang Xu, Bingquig Mao, Ji Wei,
Liang, Dehai Wu, Carbon 39 (2001) 2077.
[19] W. Yi, L. Hu, H. Hu, Z.W. Pan, S.S. Xie, Phys. Rev. Lett. 91 (2003) 076801.
[20] V. Krstic, S. Blumentritt, J. Muster, S. Roth, A. Rubio, Phys. Rev. B 67 (R) (2003)
041401.
[21] P.P. Edwards, R.L. Johnston, C.N.R. Rao, D.P. Tunstall, F. Hensel, Phil. Trans. R.
Soc. Lond. A 356 (1998) 5.
Fig. 4. Temperature dependent magnetoresistance (MR) for 60 min. CVD hydrogen annealed (upper) and argon-only annealed (lower) MWCNT devices on the left. Mobility
versus temperature for hydrogen annealed (upper) and argon-only annealed (lower) MWCNT devices calculated as parameters of the Bright model on the right.
A.L. Friedman et al. / Physica B ] (]]]]) ]]]–]]]
5
Please cite this article as: A.L. Friedman, et al., Physica B (2010), doi:10.1016/j.physb.2010.12.009