NMR study of boron nitride nanotubes

Article abstract of DOI:10.1016/j.ssc.2004.01.012

¹¹B nuclear magnetic resonance (NMR) spectra and the spin lattice relaxation times (T₁) have been measured in boron nitride nanotubes (BNNTs), grown by chemical vapour deposition(CVD), as a function of temperature and applied magnetic field. From the analysis of the ¹¹B NMR spectra and fitting of the T₁ data, two components arising from the hexagonal and rhombohedral phases in the BNNTs were decomposed and identified. The fractions for each phase were also determined. The temperature dependence of 1/T₁ reveals that the dominant relaxation mechanism for the rhombohedral phase is a Raman relaxation mechanism driven by lattice vibration.

Full text of DOI:10.1016/j.ssc.2004.01.012

Solid State Communications 130 (2004) 45–48
www.elsevier.com/locate/ssc
NMR study of boron nitride nanotubes
a,
Jae Kap Jung , Kwon-Sang Ryu , Yong-Il Kim , Chengchun Tang
a
a
b
*
a
Electro magnetic Metrology, Korea Research Institute of Standards and Science, P. O. Box 102, Yuseong, Daejeon, 305-600, South Korea
b
Advanced Materials Laboratory, National Institute for Materials Science, 1-1 Namiki, Tsukuba 305-0044 Ibaraki, Japan
Received 15 November 2003; received in revised form 6 January 2004; accepted 12 January 2004 by A. Pinczuk
Abstract
1
1
B nuclear magnetic resonance (NMR) spectra and the spin lattice relaxation times ðT Þ have been measured in boron nitride
1
nanotubes (BNNTs), grown by chemical vapour deposition(CVD), as a function of temperature and applied magnetic field.
1
From the analysis of the B NMR spectra and fitting of the T1 data, two components arising from the hexagonal and
1
rhombohedral phases in the BNNTs were decomposed and identified. The fractions for each phase were also determined. The
temperature dependence of 1=T reveals that the dominant relaxation mechanism for the rhombohedral phase is a Raman
1
relaxation mechanism driven by lattice vibration.
q 2004 Elsevier Ltd. All rights reserved.
PACS: 76.60. 2 K
Keywords: A. Nanotubes; D. Spin lattice relaxation; E. Nuclear magnetic resonance
1
1
1
. Introduction
relaxation mechanism discovery of
B nucleus in
BNNTs. This NMR method is effective and sensitive to
the investigations mentioned earlier. To our knowledge,
Chopra et al. were the first to synthesize pure boron
1
1
nitride nanotubes (BNNTs) using an arc-discharge method
in 1995 [1]. In contrast to carbon nanotubes, BNNTs have
semiconducting properties, with a large, near-constant band
gap of ,5.5 eV, which is independent of the diameter and
chirality of the nanotube [2]. A material such as this, with a
high structural stability that is also energetically stable, is of
interest for many different technological applications, which
can exploit its special physical and chemical properties.
Recently, we carried out a large-scale synthesis of BNNTs
using a chemical vapour deposition (CVD) process that was
specially designed for evaporating boron and magnesium
oxide [3]. Using X-ray diffraction (XRD) and transmission
electron microscopy (TEM) [3], we showed that BNNTs
contain two undissolved mixed phases: a hexagonal and a
rhombohedral phase.
there has been no previous study carried out on the
1
B
1
NMR of BNNTs, although B NMR studies on boron
nitride compounds have been performed up to now [4–6].
1
In this work, B NMR and the spin lattice relaxation
1
times ðT Þ in BNNTs samples have been measured for the
1
first time in the temperature range 180–430 K. We
anticipate that the results provide the information on the
phase and relaxation mechanism in BNNTs.
2
. Experimental
A mixture of boron and magnesium oxide located in a
BN-made reaction tube was heated to 1,300 8C. At this
2 2
temperature, the boron reacted with the MgO to form B O
1
1
We have employed B nuclear magnetic resonance
NMR) for the phase identification and appropriate
and Mg vapour. The vapours were argon-transported into a
reaction chamber maintained at a temperature of ca.
1,100 8C, and a flow of ammonia gas was then introduced
into the chamber. BN was synthesized from the reaction of
(
*
Corresponding author. Tel.: þ82-428-685-152; fax: þ82-42-
68-5018.
E-mail address: jkjung@kriss.re.kr (J.K. Jung).
8
2 2
the B O and ammonia vapours. After fully evaporating the
0038-1098/$ - see front matter q 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ssc.2004.01.012

46
J.K. Jung et al. / Solid State Communications 130 (2004) 45–48
mixture, the white product could be collected from the BN
wall of the chamber. Details of the growth mechanism, and
scanning electron micrographs (SEM) and high-resolution
transmission electron micrographs (HRTEM) can be found
in Ref. [3].
4.7 T (see Fig. 1(a)). On the other hand, the NMR spectrum
at 14.1 T is unresolved (see solid line in Fig. 1(b)), because
the magnetic line broadening effect due to different
1
1
chemical shifts of the surrounding atoms around the
B
nucleus site at 14.1 T is larger than that at the lower
magnetic field of 4.7 T. The dotted and dashed lines of the
two peaks shown in Fig. 1(a) and (b) represent the resolved
spectra fitted to a Gaussian lines. The narrow peaks on the
left-hand side (dotted lines) and the broad peaks on the
right-hand side (dashed lines) arise from the hexagonal
BNNTs (h-BNNTs) and rhombohedral BNNTs (r-BNNTs)
phases, respectively. a and b are used to denote the fractions
of the h-BNNTs and r-BNNTs phases in BNNTs, respect-
ively. The values of a and b were estimated to be 0.67 and
0.33 for 14.1 T, and 0.71 and 0.29 for 4.7 T, respectively,
from the fit to the NMR line intensity. The fractions a and b
at other temperatures investigated, were almost the same as
those estimated at 297 K. Using XRD and TEM analysis, the
a and b were also estimated to be 0.58 and 0.42,
respectively, which was calculated from comparing the
intensities of the h-BNNT(100) and r-BNNT(101) peaks,
unresolved of two mixed phases. These two results using
different methods are quantitatively consistent to each other.
The estimated fractions, a and b; are summarized in Table 1.
The line widths of h-BNNTs and r-BNNTs in Fig. 1(a)
and (b) are about 8 and 10 kHz, respectively. The centre of
gravity for the h-BNNTs and r-BNNTs peaks in magnetic
fields of 14.1 and 4.7 T was 22.7 and 6.8 kHz, and 211.6
and 7.4 kHz, respectively. It is of interest that the centre of
gravity of the h-BNNTs peaks at both magnetic fields were
shifted in the down-field direction, and the centre of gravity
1
1
The B NMR spectra were obtained using two different
NMR spectrometers operating at Larmor frequencies of
1
6
2
92.493 MHz ðB ¼ 14:1 T; Varian, UnityNOVA 600) and
0
4.164 MHz ðB ¼ 4:7 T; Varian, UnityINFINITYplus
0
00). The NMR experiments were performed at the Daegu
Branch of the Korean Basic Science Institute. The spectra
were obtained using a solid echo ðp=2 2 t 2 p=2Þ or a Hahn
echo ðp=2 2 t 2 pÞ sequence, where the p/2 and p pulse
lengths were 2.5 and 5 ms, respectively. The repetition delay
time was 50 s. All the NMR spectra were referenced to an
aqueous saturated H BO solution.
3
3
1
1
The T1 of B NMR was measured using the same
spectrometers in the temperature range 180–430 K. The
temperature stability was kept within ^0.05 K for the T1
measurement times. The T pulse sequence was an inversion
1
recovery pulse sequence using solid echo, p 2 t2
ðp=2 2 t 2 p=2Þ: The pulse separation between the solid
echo, t; was 25 ms.
3
. Results and discussion
1
The B NMR spectra of BNNTs in two magnetic fields
1
at 297 K are shown in Fig. 1. The line shape and line width
of the spectrum in Fig. 1 is similar to that reported by
Kanehashi et al. [6]. The resolved two peaks are observed at
11
Fig. 1. The B NMR spectra of BNNTs at 297 K in a magnetic field of: (a) 4.7 T and (b) 14.1 T. The dotted and dashed lines at two magnetic
fields represent the decomposed fitted Gaussian lines for h-BNNTs and r-BNNTs, respectively.

J.K. Jung et al. / Solid State Communications 130 (2004) 45–48
47
Table 1
The values of fractions a and b corresponding to h-BNNT and r-BNNT, respectively, according to the different experimental methods
Method used
NMR
Applied field (T)
4.7
Measuring quantity
Peak
a
b
Reference
0.71
0.29
Present work
Present work
1
4.1
Peak
0.67
0.64
0.33
0.36
T1
XRD
–
Peak
0.58
0.42
[3]
of the r-BNNTs peaks at both magnetic fields were shifted in
the up-field direction. A detailed investigation into the
origin of these different shifts in direction, i.e. of the
opposite magnetic environment of the two phases, is
presently underway.
respectively. The values of 1=T were also obtained in the
1
same manner for the other temperatures investigated.
On the other hand, in the case of an applied magnetic
field of 14.1 T, the T was measured for the undissolved
1
spectrum arising from the contributions of the two phases.
The two exponential function shown in the inset of Fig. 2
correspond to one exponential function arising from each
phase. One exponential behaviour in a high magnetic field
The T1 was measured in the temperature range 180–
30 K in two magnetic fields of 4.7 and 14.1 T. A recovery
4
function for the central line with the dominant quadrupole
relaxation is given in the following equation for the case of
an inversion recovery pulse sequence.
may be understood by considering the following. W and W2
1
are angular dependent with respect to the applied magnetic
field. In the powder samples with quadrupole relaxation at
yðtÞ ¼ ½1 2 MðtÞ=Mð1ފ=2
higher magnetic field usually W ø W2 after an orienta-
1
tional average and then one exponential is essentially
¼
0:5exp ð22W tÞ þ 0:5exp ð22W tÞ
ð1Þ
1
2
observed [7]. Thus, W ¼ W ¼ W; the recovery function of
1
2
where MðtÞ is the nuclear magnetization at time t; and W1
and W2 are the transition probabilities corresponding to
Dm ¼ ^1 and Dm ¼ ^2; respectively. Thus, the recovery
curve is not a single exponential function, and the spin
Eq. (1) is written by the form of exp ð22WtÞ ¼ exp ð2t=T1Þ;
where 2W ¼ 1=T1: Therefore, the recovery data shown in
the inset of Fig. 2 can be fitted to the following two
exponential functions.
lattice relaxation rate, 1=T ; is given by
1
yðtÞ ¼ ½1 2 MðtÞ=Mð1ފ=2
1
=T ¼ 2ðW þ 4W Þ=5
ð2Þ
Fig. 2 shows the T recovery data of peaks for h-BNNTs and
1
1
2
¼
a exp ð2t=T1h-BNNTsÞ þ b exp ð2t=T1rBNNTs
Þ
ð3Þ
1
r-BNNTs in 4.7 T at 297 K, which is well fitted using Eq.
where T1h-BNNTs and T1r-BNNTs are the spin lattice relaxation
times for the h-BNNTs and r-BNNTs phases, respectively.
The a and b are the weighting factors (or fractions) of the h-
BNNTs and r-BNNTs phases, respectively, for the total
BNNTs spectrum. We also estimated the fractions a and b
were 0.64 and 0.36, respectively, which are almost the same
values as those obtained from the spectrum intensity fit of
(
to be 0.53 s and 0.58 s for h-BNNTs and r-BNNTs,
1). Using Eq. (2), the values of 1=T at 297 K were obtained
1
2
1
21
Fig. 1(b). The fitting of the T data at other temperatures
1
gave almost the same fraction values. The a and b estimated
from the spectrum analysis and T fitting are summarized in
Table 1.
1
Fig. 3 shows the temperature dependence of 1=T for the
1
two phases in two different magnetic fields. The exper-
imental results show that the temperature dependence of
2
=T1 for r-BNNTs phase is proportional to T ; which is
1
shown by the fitted solid line in Fig. 3, although the data
fluctuation is quite large. An exponential recovery and no
Larmor frequencies dependence of 1=T (see Fig. 3) for r-
1
BNNTs phase was also observed. These features indicate
that the relaxation process of r-BNNTs phase in BNNTs is
characteristic of a Raman relaxation mechanism driven by
lattice vibration [8]. Meanwhile, the origin of relaxation
Fig. 2. The recovery trace of the magnetization at 4.7 T for BNNTs
at 297 K. The inset shows the recovery data at 14.1 T at the same
temperature.
1
1
mechanism for h-BNNTs is being study. In summary,
B

48
J.K. Jung et al. / Solid State Communications 130 (2004) 45–48
technique is very powerful to study the local structure and
the relaxation mechanism.
Acknowledgements
This work was supported by the Centre for Nanoscale
Mechatronics and Manufacturing.
References
[
[
1] N.G. Chopra, R.J. Luyken, K. Cherrey, V.H. Crespi, M.L.
Cohen, S.G. Louie, A. Zettl, Science 269 (1995) 966.
2] A. Rubio, J.L. Corkill, M.L. Cohen, Phys. Rev. B49 (1994)
5081.
Fig. 3. The temperature dependence at two different magnetic fields,
.7 and 14.1 T, for h-BNNTs and r-BNNTs. The solid line is fitted
4
[3] C. Tang, Y. Bando, T. Sato, K. Kurashima, Chem. Comm. 12
(2002) 1290.
2
using the relationship, 1=T / T :
1
[
[
[
4] A.H. Silver, P.J. Bray, J. Chem. Phys. 32 (1960) 288.
5] M. Fanciulli, M. Corti, Phys. Rev. B52 (1995) 11872.
6] K. Kanehashi, K. Saito, J. Mol. Struct. 602–603 (2002) 105.
NMR technique was successfully employed to resolve the
two local phases existed in BNNTs by line intensity analysis
and T data fit, unlikely XRD technique. Furthermore, the
1
[7] A. Rigamonti, Adv. Phys. 33 (1984) 115.
1
1
B relaxation mechanism was discovered by temperature
dependence of 1=T : Consequently, solid-state NMR
[8] A. Abragam, The Principles of Nuclear Magnetism, Oxford
University Press, London, 1961, p. 112.
1