Formation of cBN nanocrystals by He+ implantations of hBN

R. Machaka1,2,3 , R. M. Erasmus2,3 , and T. E. Derry2,3

1
iThemba LABS (Gauteng), P Bag 11, WITS 2050, RSA
2
School of Physics, University of the Witwatersrand, P Bag 3, WITS 2050, RSA
3
DST/NRF Centre of Excellence in Strong Materials, University of the Witwatersrand,
P Bag 3, WITS 2050, RSA

Abstract
The structural modifications of polycrystalline hexagonal boron nitride implanted with He+ ion
beams at energies between 200 keV and 1.2 MeV to fluences of 1.0 × 1017 ions·cm−2 were inves-
tigated using Raman spectroscopy. Micro-Raman spectra show evidence of implantation-induced
structural transformations to nanocrystalline cubic boron nitride, rhombohedral boron nitride and
amorphous boron nitride phases. The asymmetrical broadening and general downshifting of the
first–order Longitudinal–Optical cBN phonon may be explained using the phonon confinement model.

1 Introduction
In recent years cubic boron nitride (cBN) has attracted considerable experimental and theoretical interest
due to its rather unusual combination of superior physico-chemical properties and its enormous potential
for technological applications. Several comprehensive reviews on the structures, properties and the
significant technological importance of cBN and other BN modifications are discussed elsewhere [1, 2,
3, 4, 5].
cBN does not occur in nature. Bulk cBN is synthesised at high pressures and temperatures by direct
phase transformation of the graphite–like (usually hBN) or disordered BN modifications in the presence
of a suitable catalytic solvent [1, 5, 6, 7, 8, 9, 10]. The p,T transformation conditions are dictated by
the boron nitride thermodynamic (p,T) phase diagram [9]. From the knowledge of the equilibrium (p,T)
phase diagram of BN [11], cBN is the most stable phase with respect to other BN modifications under
ambient room temperature and pressure conditions. A spontaneous hBN–to–cBN phase transformation
is hindered by the very large activation energy barrier [1]. Since its successful synthesis, efforts have
continued in searching for new routes that can be used to synthesize cBN under less extreme conditions
[10].
We here demonstrate the synthesis of nanocrystals of the cBN phase by He+ ion implantation of the
hBN phase. It has been reported before, theoretically [6, 12] and experimentally [13, 14, 15, 17], that
under the non-equilibrium conditions of ion implantation, the sp2 to sp3 phase transformation can be
carried out at zero pressure and moderate temperatures. This phenomenon has also been reported in
other graphitic carbon materials [15, 16]. The introduction of defects leads to dangling bonds between
the layers and may promote the subsequent corrugation of the basal planes [15]. Large concentrations

1
of defects (vacancies/interstitials) present in graphite could lower the transformation barrier between
diamond and graphite and permit diamond nucleation. Attempts have also been made by other research
groups to use femtosecond [18] and ultraviolet [19] laser irradiation of graphite-like BN to try and
influence the hBN–to–cBN phase transformation at ambient pressure and moderate temperatures.

2 Experimental Details
Hot–pressed polycrystalline hBN powder samples of research grade quality were supplied by Goodfel-
low UK for this study.
The samples were implanted with He+ particle beams of energies between 200 keV and 1.2 MeV to
fluences in the range of 1.0×1015 ions·cm−2 to 1.0×1017 ions·cm−2 . The implantations were done using
the iThemba LABS (Gauteng) modified 1.4 MeV pressurized Cockcroft–Walton accelerator. All of the
reported implantations were carried out at room temperature. The distributions of the radiation induced
damage and the implant depths were calculated using the SRIM2006 program [20].

35000 unimplanted
Raman Intensity (arbt. units)

30000

25000

20000

15000

10000

5000

0
1200 1250 1300 1350 1400 1450

W -1
avenumber (cm )

Figure 1: The µ-Raman spectrum of an unimplanted hBN sample.

Micro–Raman spectroscopy (µ-RS) measurements were used to characterize the samples before and
after implantation. The 514.5nm line of an Ar+ ion laser was used as an excitation source. An 1800
grooves/mm grating in the single spectrograph mode and triple subtractive configurations of the Jobin–
Yvon T64000 Raman spectrometer were used to disperse the scattered light. The measurements were
performed using a 20× Olympus objective and a the spot size diameter ∼ 1.5µm. The Raman measure-
ments were carried out at the Raman and Luminescence Laboratory at the School of Physics, University
of the Witwatersrand.

3 Results and discussions

Fig 1 shows the µ-RS spectrum (1200 - 1475cm−1 ) measured on an unimplanted sample. As widely
reported, the µ-RS spectra of nominally pure hBN exhibit an intense high–frequency E2g phonon mode
that is attributed to the B – N bond stretching vibrational mode within the basal planes appearing around
1366cm−1 [21, 22, 23, 24].

400

2400 a 1200keV b 750keV

350
Intensity (arbt.units)

300
2250

250

2100
200

150
1950

100
Raman

1800
50

0
1650

-50

1500 -100
1200 1250 1300 1350 1400 1450 1200 1250 1300 1350 1400 1450

1500

400 350keV
200keV
c 1250 d
Intensity (arbt.units)

300
1000

750
200

500
Raman

100

250

0
0

-100 -250
1200 1250 1300 1350 1400 1450 1200 1250 1300 1350 1400 1450

-1 -1
Wavenumber (cm ) Wavenumber (cm )

Figure 2: The µ–Raman spectra of hBN samples implanted to fluences of 2.0×1016 He+ ·cm−2 at ener-
gies of (a.) 1.2 MeV, (b.) 750 keV, (c.) 350 keV, and at (d.) 200 keV. Please note that the vertical scales
are arbitrary.

The µ-RS spectra shown in fig. 2a–d were measured for hBN samples that were implanted to a fluence
of 2.0×1016 He+ ·cm−2 at energies between 200 keV and 1.2 MeV (the intensity scales of the four spectra
are arbitrary). The spectra clearly reveal that after implantation with 4 He+ particles, the intensity of the
principal hBN Raman peak greatly diminishes. In addition, the spectra also reveal weak and broad
Raman features appearing around 1290cm−1 whose origins we have attributed to the formation of nano-
crystals of cBN that was influenced by He+ implantation induced hBN–to–cBN phase transformation.
Similar downshifted, weak, and asymmetrically broadened Raman features have been reported before
by other researchers for micro- and nano-crystals of cBN [21, 22], diamond [15, 16, 24], and other group
III-V semiconductor compounds [25]. The wavenumbers of the broad features correspond very well to
the values reported in references [21] and [26]. It is generally accepted that the phonon confinement
model explains the origins and the deviation of the observed nanocrystalline cBN Raman spectrum
from that obtained for ‘finite’ cBN single crystals. The theory of the model is discussed in great detail
elsewhere, see [27] and references therein.
 Gaussian peak position (cm )
0.024 1297

-1
(a.) (b.)
Lorentzian
0.020 1296

0.016
1295
/I

0.012
I / I
Gaussian

G L
1294
Exponential growth fit

0.008 Gaussian peak position

1293
Exponential decay fit
I

0.004

1292
200 400 600 800 1000 1200
200 400 600 800 1000 1200

Implantation energy (keV)

Implantation Energy (keV)

Figure 3: The dependence of (a.) the normalized intensities of the Gaussian peaks and (b.) the peak
position of the Gaussian peaks with ion implantation energy.

By curve fitting the experimental data to Lorentzian and Gaussian lineshapes, values of the peak
positions and intensities were obtained. From these values, two graphs were plotted; fig 3a which gives
the variation of the normalized intensities of the Gaussian peaks (IG /IL ) plotted against the implantation
energy and fig 3b which shows the variation of the position of the Gaussian peak with the implantation
energy. For each implantation energy, each Gaussian peak was normalized to the respective principal
hBN Raman peak that was measured outside the implanted region. The position of the Gaussian peaks
appears to be approaching a saturation as the implantation energy increases, possibly 1306cm−1 . The
possible explanations for this are either that the nanocrystals formed at the lower energies are more
disordered in structure or that the formed crystallites do not have any particular preferential orientation.
The ratio IG /IL can be linked to the change in the hybridization states from sp2 to sp3 . Fig 3b shows a
large, possibly exponential growth in sp3 hybridised states in the surface region that is being probed by
the Raman laser, for the lowest implantation energy.
SRIM calculations of the inelastic damage profiles and the He+ implant distributions as a function of
depth were carried out to explain the rapid growth of sp3 hybridised states towards the surface layers
that is observed in the lower implantation energy regime. It is noted that much of the damage profile
created at 200 keV lies within the penetration depth region (typically 1µm) of the probing Raman laser
light. It is also noted that the damage profile created at 1.2 MeV lies mostly beyond this penetration
depth region, however, the profile exhibits a long tail that does overlap this penetration depth region of
the probing laser radiation.
A critical dose of the order of 2.0×1016 He+ ·cm−2 was observed to favour the hBN–to–cBN phase
transformation; He+ implantations below the critical dose, of the order of 1015 He+ ·cm−2 , were ob-
served not to have influenced the hBN–to–cBN phase transformation in any observable way (see fig 4)
whilst implantations at doses above the critical, of the order of 1017 He+ ·cm−2 were observed only to
produce extensive radiation damage on the sample as shown by the collapse of the principal hBN line
at 1366−1 . However, the µ-Raman spectra that was measured on the 1.2 MeV He+ implanted samples
revealed in addition to the principal hBN Raman peak, a broad feature appearing around 752.9cm−1
whose origins we attribute to either crystalline/amorphous boron clusters or ion implantation-induced
hBN–to–rBN phase transformation, where rBN is an sp2 hybridized rhombohedral BN allotrope. The
 (a.) (b.)

20000
Raman Intentity / (Arbitrary units)
virgin
7
136 .3cm
-1

18000
implanted
16000
implanted

14000 implanted -1
1296.6cm

12000

10000

900 1000 1100 1200 1300 1400 1500

8000

6000
752.9cm -1

4000 760.2cm -1

2000

0
600 800 1000 1200 1400 600 800 1000 1200 1400

-1 -1
W avenumber (cm ) W avenumber (cm )

Figure 4: The µ-Raman spectra taken at different points on the sample that was implanted to a fluence
of [a] 1.0 × 1015 ions/cm2 and [b] 2.0 × 1016 ions/cm2 . The 1290cm−1 feature is becoming evident in
[b].

peak has previously been observed by other researchers [23, 28].

4 Summary and conclusions

We have presented experimental results showing the possibilities of using ion implantation into hBN to
influence a structural transformation at room temperature that results in the formation of nanocrystals
of the ultrahard cBN phase. We have also shown that the µ-Raman spectra fit the phonon confinement
model that justifies the deviation of the observed spectra from that of cBN single crystals. Results
showing, less conclusive, that implantations at higher energies irrespective of the implantation dose
might also influence an hBN–to–rBN phase transformation were also presented.

Acknowledgements
The work was funded by DST/NRF Centre of Excellence in Strong Materials and iThemba LABS (Gaut-
eng). The authors are grateful to the following individuals for the great efforts they made in making this
publication a reality; M. Rebak, J.D. Comins, J.E. Lowther, T.E. Mosuang, N.W. Makau, and G. Amolo.

References
[1] L. Vel, G. Demazeau, and J. Etourneau, Mater. Sci. Eng. B 10 (1991), 149–151
 [2] P. B. Mirkarimi, K. F. McCarty, and D. L. Merlin, Mater. Sci. Eng. R 21 (1997), 47–100

[3] Common Crystal Structures of Group III–Nitrides in Properties of Group III Nitrides, Edited by
J. H. Edgar, Kansas State University, 1994, Published by INSPEC, The Institution of Electrical
Engineers, London, pp 3–6

[4] Crystal Structure, Mechanical Properties and Thermal Properties of BN in Properties of Group
III Nitrides, Edited by J. H. Edgar, Kansas State University, 1994, Published by INSPEC, The
Institution of Electrical Engineers, London, pp 7–21

[5] G. Will and P. G. Perkins, Diamond Relat. Mater. 10 (2001) 2010

[6] T. E. Mosuang and J. E. Lowther, J. Phys. Chem. Solids 63 (2002), 363–368

[7] G. Will, G. Nover and J. von der Gonna, J. Solid State Chem. 154 (2000), 280–285

[8] Mei-Yan Yu et al., Mater. Sci. Eng. B 121 (2005), 166

[9] V. Z. Turkevich, J. Phys.: Condens. Mater 14 (2002), 10963-10968.

[10] J. Y. Huang and Y. T. Zhu, Chem. Mater. 14 (2002), 1873

[11] V. L. Solozhenko, J. Hard Mater. 6 (1995), 51

[12] E. Sandre et al., J. Phys. Chem. Solids 55 (1994), 1261

[13] F. Banhart et al., Chem. Phys. Lett. 269 (1997), 349

[14] M. Zaiser and F. Banhart, Phys. Rev. Lett. 79 (1997), 3680

[15] F. Banhart and P. M. Ajayan, Nature 382 (1996),433

[16] T. L Daulton et al., Nucl. Instr. and Meth. B 175–177 (2001), 12–20

[17] S. Talapatra et al., Nanotechnology 17 (2006), 305

[18] A. V. Kanaev, J. P. Petitet, L. Museur, V. Marine, V. L. Solozhenko, and V. Zafiropulos, J. Appl.

Phys. 96 (2004), 4483–4489

[19] M. Sokolowski et al., J. Mater. Sci. 25 (1990), 263

[20] J. F. Ziegler, Nucl. Instr. and Meth. B 219–220 (2004), 1027–1036.

[21] T. Werninghaus, J. Hahn, F. Ritcher, and D. R. T. Zahn, Appl. Phys. Lett. 70 (1997), 958–960

[22] R. J. Nemanich, S. A. Aolin, and R. M. Martin, Phys. Rev. B 23 (1981), 6348–6356

[23] H. Sachdev, R. Haubner, H. Noth, and B. Lux, Diamond Relat. Mater. 6 (1997), 286

[24] P. V. Huong, Diamond Relat. Mater. 1 (1991), 33

[25] P. Parayanthal and F. H. Pollak, Phys. Rev. Lett. 52 (1984), 1822–1825

[26] W.J. Zhang et al., Appl. Phys. A 76 (2003), 953–955

[27] J.W. Ager III et al., Phys. Rev. B 43 (1991), 6491–6499

[28] J. Liu, Y. K. Vohra, J. T. Tarvin, and S. S. Vagarali, Phys. Rev. B 51 (1995), 8591