Titanium nitride (TiN) exhibits extremely high wear resistance and corrosion resistance, leading to a broad range of applications. Despite the extensive theoretical efforts on the structural stability of TiN under pressure, a few experimental observations of the pressure-induced phase transition through x-ray diffraction are still in debate. Here, the Raman spectra of TiN under pressure are studied in a diamond anvil cell up to 36 GPa. The discontinuity in the pressure dependent phonon frequencies suggests a phase transition at around 5 GPa with the bonding between Ti and N hardly changing. The phase transition is ascribed to an isostructural transition or a vacancy filling mechanism.

Transition metal nitrides are a class of metallic compounds with a simple NaCl-type crystal structure (B1), which have become promising materials owing to their extremely high melting point, hardness, and metallic conductivity.1,2 Especially, titanium nitride (TiN), characterized by its extremely high wear resistance and corrosion resistance, has drawn intensive attention for its potential application as a corrosion resistant coating, microelectronic apparatus, and cutting tool.3–7 The high-pressure technique is a useful tool to understand the high hardness and bulk modulus of TiN, which is essential for its practical applications related to the mechanical properties of TiN, such as load deflection, internal strain, sound velocities, and fracture toughness. Furthermore, new phases of TiN with different mechanical and electrical properties may present at high pressure. The earlier experiments and theoretical calculations on TiN have revealed a more incompressible high-pressure CsCl (B2) structure with a larger bulk modulus and higher electrical conductivity, which are not obtainable at ambient pressure.8–14 Zhao et al. performed in situ angle dispersive synchrotron x-ray diffractions (XRDs) on TiN.8 They found that although the XRD patterns are similar with an increase in pressure, the volume of TiN changed discontinuously at around 7 GPa, suggesting that an isostructural phase transition originated from the change in the electron topological morphology. Wang et al. compressed the size-dependent nanocrystalline TiN in a diamond anvil cell and studied the structural stability by means of XRD, but no phase transition was found.9 Chen et al. studied the strength and elastic moduli of TiN by using synchrotron radial x-ray diffraction (RXRD) in a diamond anvil cell up to 45 GPa.10 They found that the uniaxial stress component of TiN samples reaches 8.6 GPa. To get a better understanding of the structure and phase transition of TiN, a series of theoretical calculations for TiN on elastic constants, stability, and thermodynamic and optical properties under high pressure were carried out. Ahuja et al. predicted that the phase transition of TiN from the B1 structure to the B2 structure happened when the pressure reached 370 GPa.11 Chauhan et al. found that the structural phase transition pressure was at 310 GPa by using the three-body potential model.12 Liu et al. determined the pressure for phase transition from the B1 structure to the B2 structure to be 364.1 GPa for generalized gradient approximation (GGA) and 322.2 GPa for local density approximation (LDA) from equal enthalpies.13 Sun et al. predicted a new phase of anti-TiP structure with the space group P63/mmc at 270 GPa.14 

One can find a huge discrepancy between the phase transition pressure of 7 GPa observed in the experiment and that of around 300 GPa found in the theoretical calculations. Moreover, the earlier high pressure XRD experiments from different researchers suggest ambiguous claims on the phase transition of TiN. Thus, other techniques reflecting high-pressure phase transitions are needed to further investigate the phase transition of TiN under pressure. In this work, we present the high-pressure Raman study up to 36 GPa using a diamond anvil cell (DAC) on TiN to unveil the phase transition in a low pressure region.

TiN powder was purchased from Sigma-Aldrich (purity >99.99%). The XRD pattern in Fig. 1(b) suggests the sample is well-crystallized with a pure phase. High-pressure Raman spectra experiments were performed in diamond anvil cells (DACs) with culets of 300 µm in diameter, as shown in Fig. 1(a). A central hole with a diameter of 100 µm was drilled in a T301 stainless steel gasket pre-indented to a thickness of around 20 µm by laser drilling as the sample chamber. The TiN powder was loaded with silicone oil as the pressure transmitting medium for experiments. A ruby ball with a diameter of around 3 µm was loaded with the TiN sample to calibrate the pressure because of its intense and sharp florescence line, sensitivity to pressure changes, and chemical stabilities with the sample (TiN) and silicone oil. The pressure was controlled through the mechanical feedthroughs connected to four loading DAC screws. The whole experiment and data measurement process was accomplished by a high-pressure Raman integration device.15 We adopted a backscattering geometry for measurements with incident laser wavelengths of 532 nm. The laser energy is around 5 mW at the sample to avoid overheating. A single stage imaging spectrograph SP2500 (Acton) with a focal length of 500 mm was equipped with a thermoelectrically cooled CCD detector array (Princeton eXcelon). Because of the weak signal for the conductive TiN, the integration time is 10 minutes for each Raman measurement.

FIG. 1.

(a) Configuration of two opposing diamond anvils in the DAC; the sample chamber is a hole in the gasket, together with ruby, silicone oil, and TiN powder, and (b) the XRD pattern of TiN powder under ambient conditions.

FIG. 1.

(a) Configuration of two opposing diamond anvils in the DAC; the sample chamber is a hole in the gasket, together with ruby, silicone oil, and TiN powder, and (b) the XRD pattern of TiN powder under ambient conditions.

Close modal

The Raman spectra study on TiN is difficult because of its high conductivity. To get a good signal to noise ratio, high laser energy is needed, which will result in the oxidation of the TiN sample due to overheating. As shown in the inset of Fig. 2, the Raman signal indicates that TiO2 is formed from the TiN sample due to the heating of laser during measuring in air. When the TiN sample is loaded in a DAC separated from air, the Raman spectra collected with the same laser energy and integration time become notably different from the ones measured in air. Figure 2 shows the Raman spectra of TiN at 3.2 GPa. The waveform of the correlation peak was fitted using a curve fitting method to get phonon frequencies. The phonon frequencies at approximately 228 cm−1, 310 cm−1, 452 cm−1, and 570 cm−1, are related to transverse acoustic (TA), longitudinal acoustic (LA), second-order acoustic (2A), and transverse optical (TO) modes of TiN, respectively. The peak at around 845 cm−1 originated from the Raman scattering of acoustical plus optical (A + O) modes.16 The good signal to noise ratio of Raman spectra for TiN in our setup allows us to systematically investigate the structural stability under pressure.

FIG. 2.

Raman spectrum of TiN at 3.2 GPa during decompression. The black hollow circles indicate the data from Raman detection, the solid lines is the multi-peak fitting to the Raman spectrum, and the dotted lines are the fitted Raman peaks to calculate phonon frequencies. The inset is the Raman spectrum measured under ambient conditions without the DAC using the same laser energy and integration time for reference.

FIG. 2.

Raman spectrum of TiN at 3.2 GPa during decompression. The black hollow circles indicate the data from Raman detection, the solid lines is the multi-peak fitting to the Raman spectrum, and the dotted lines are the fitted Raman peaks to calculate phonon frequencies. The inset is the Raman spectrum measured under ambient conditions without the DAC using the same laser energy and integration time for reference.

Close modal

Figure 3 depicts the Raman spectra of TiN from ambient pressure to 36 GPa. Upon compression, all the Raman peaks shift to higher frequencies as the pressure-induced shrinkage of the lattice enhances the phonon energy. The phonon frequencies were fitted using the Lorentzian equation, as shown in Fig. 2. The pressure dependency of the phonon frequency of TiN is illustrated in Fig. 4. The phonon frequencies increase almost linearly with an increase in pressure up to 36 GPa, except a kink at the low-pressure region in Fig. 4(a). The linear dependence of phonon frequencies on pressure is in accordance with the linear relation between elastic moduli and applied pressure in the earlier report of the elastic properties of TiN.10 The discontinuities seen in the enlarged view of the phonon frequencies of 845 cm−1 and 570 cm−1 in Fig. 4(b) suggest a phase transition at around 5.0 GPa near the earlier reported phase transition pressure of 7 GPa with XRD. The three peaks at low frequency show relatively weak changes at around 5.0 GPa due to worse fitting accuracy, as shown in Fig. 4(c). The origin of phase transition is discussed as follows.

FIG. 3.

Raman spectra of TiN at elevated pressures.

FIG. 3.

Raman spectra of TiN at elevated pressures.

Close modal
FIG. 4.

(a) Pressure dependent phonon frequencies of TiN deduced from the Raman spectra shown in Fig. 3. The enlarged view of phonon frequencies around (b) 845 cm−1 and 570 cm−1 and (c) 228 cm−1, 310 cm−1, and 570 cm−1 between 0 GPa and 20 GPa. The lines are the linear fits to experimental data.

FIG. 4.

(a) Pressure dependent phonon frequencies of TiN deduced from the Raman spectra shown in Fig. 3. The enlarged view of phonon frequencies around (b) 845 cm−1 and 570 cm−1 and (c) 228 cm−1, 310 cm−1, and 570 cm−1 between 0 GPa and 20 GPa. The lines are the linear fits to experimental data.

Close modal

The observed phase transition in TiN by Raman spectra supports the phase transition observed in the earlier XRD study.8 Both the current Raman study and the earlier XRD observed no new peaks but a discontinuity or a kink in the pressure dependencies, implying the bonding between Ti and N hardly changed at phase transition. This phase transition probably originates from an isostructural transition, in which the crystalline structure does not change but the electronic topological transition rises.8 Isostructural transition has been observed in various systems under pressure. For example, synchrotron x-ray diffraction and x-ray absorption spectroscopy studies on the Fe-based superconductor FeSe uncovered an isostructural transition at ∼2.8 GPa, based on the axial ratio c/a with a finer pressure step.17 In a metal–organic framework, isostructural phase transition rises at about 6 GPa followed by negative compressibility along the b axis.18 

Vacancy filling may also lead to discontinuity in phonon frequency of TiN under pressure. TiN contains large amounts of vacancies like the other Ti based compounds. TiN1−x has a remarkably broad range of non-stoichiometry ranging from 0.26 to 1.16, which enables a large proportion of nitrogen vacancy in the lattice.19 Hojo et al. investigated the defect structure of TiN powders and demonstrated that TiN contained large numbers of vacancies on both Ti and N sublattices.20 Pressure-induced vacancy filling has been reported in TiO with ∼15% vacancies in both Ti and O sublattices and results in the discontinuity in phonon frequencies without new Raman peaks.15 As TiN shows a crystal structure and vacancy concentration similar to TiO, the phase transition at around 5 GPa may also be ascribed to a vacancy filling mechanism.

The pressure-induced phase transition at around 5 GPa in TiN is unlikely to be a NaCl (B1) structure to CsCl (B2) structure transition under pressure, which is commonly adopted for materials with a B1 structure.21,22 The theoretically calculated B1 to B2 structural phase transition pressure in TiN at hydrostatic pressure conditions are around 300 GPa, which is far beyond the current experimental observation of 5 GPa.12,14,23 Recent calculations suggest a specific nonhydrostatic pressure could substantially reduce the stress required for the B1 to B2 structural transformation from 347 GPa to about 40 GPa.23 Nevertheless, the current Raman experiments are performed in quasi-hydrostatic pressure conditions. Moreover, the earlier XRD studies under a nonhydrostatic compression of up to 45 GPa reports no structural phase transitions.8 Further experiments at higher pressure and different nonhydrostatic conditions are needed to investigate the B1 to B2 structural phase transition of TiN in the future.

In conclusion, the structural stability of TiN is investigated using high pressure Raman spectra up to 36 GPa. The phonon frequencies calculated from Raman spectra increase almost linearly with an increase in pressure, except a discontinuity at around 5 GPa. The linear dependence of phonon frequencies implies the linear increase of elastic moduli on pressure, and the discontinuity suggests a phase transition. The phase transition at around 5 GPa is explained by isostructural transition or vacancy filling.

This work was supported by the National Natural Science Foundation of China (Grant Nos. 51672279, 11774354, 11874361, and 51727806), the Science Challenge Project (Grant No. TZ2016001), and the CASHIPS Director’s Fund (Grant No. YZJJ201705).

1.
L. E.
Toth
,
Transition Metal Carbides and Nitrides
(
Academic Press
,
New York
,
1971
).
2.
I. M.
Vinitskii
and
G. V.
Samsonov
,
Handbook of Refractory Compounds
(
Plenum
,
New York
,
1980
).
3.
L.
Hultman
,
U.
Helmersson
,
S. A.
Barnett
,
J. E.
Sundgren
, and
J. E.
Greene
,
J. Appl. Phys.
61
,
552
(
1987
).
4.
A.
Wisbey
,
P. J.
Gregson
, and
M.
Tuke
,
Biomaterials
8
,
477
(
1987
).
5.
B.
Zega
,
M.
Kornmann
, and
J.
Amiguet
,
Thin Solid Films
45
,
577
(
1977
).
6.
M.
Wittmer
,
B.
Studer
, and
H.
Melchior
,
J. Appl. Phys.
52
,
5722
(
1981
).
7.
R.
Buhl
,
H. K.
Pulker
, and
E.
Moll
,
Thin Solid Films
80
,
265
(
1981
).
8.
J.-G.
Zhao
 et al,
Chin. Phys. Lett.
22
,
1199
(
2005
).
9.
Q.
Wang
,
D.
He
,
F.
Peng
,
L.
Xiong
,
J.
Wang
,
P.
Wang
,
C.
Xu
, and
J.
Liu
,
Solid State Commun.
182
,
26
(
2014
).
10.
H.
Chen
,
F.
Peng
,
H.-k.
Mao
,
G.
Shen
,
H.-P.
Liermann
,
Z.
Li
, and
J.
Shu
,
J. Appl. Phys.
107
,
113503
(
2010
).
11.
R.
Ahuja
,
O.
Eriksson
,
J. M.
Wills
, and
B.
Johansson
,
Phys. Rev. B
53
,
3072
(
1996
).
12.
R.
Chauhan
,
S.
Singh
, and
R. K.
Singh
,
Cent. Eur. J. Phys.
6
,
277
(
2008
).
13.
K.
Liu
,
X.-L.
Zhou
,
H.-H.
Chen
, and
L.-Y.
Lu
,
J. Therm. Anal. Calorim.
110
,
973
(
2011
).
14.
X.
Sun
,
C.
Liu
,
Y.
Guo
,
D.
Sun
, and
X.
Ke
,
Phys. Lett. A
382
,
656
(
2018
).
15.
J.
Ding
,
T.
Ye
,
H.
Zhang
,
X.
Yang
,
H.
Zeng
,
C.
Zhang
, and
X.
Wang
,
Appl. Phys. Lett.
115
,
101902
(
2019
).
16.
W.
Spengler
,
R.
Kaiser
,
A. N.
Christensen
, and
G.
Müller-Vogt
,
Phys. Rev. B
17
,
1095
(
1978
).
17.
Z.
Yu
 et al,
J. Alloys Compd.
767
,
811
(
2018
).
18.
M.
Zhou
,
K.
Wang
,
Z.
Men
,
C.
Sun
,
Z.
Li
,
B.
Liu
,
G.
Zou
, and
B.
Zou
,
CrystEngComm
16
,
4084
(
2014
).
19.
S.
Xu
,
M.
Wang
,
L.
Qiao
,
Y.
Ye
,
G.
Jia
, and
Y.
Zhao
,
Adv. Appl. Ceram.
114
,
256
(
2015
).
20.
J.
Hojo
,
O.
Iwamoto
,
Y.
Maruyama
, and
A.
Kato
,
J. Less-Common Met.
53
,
265
(
1977
).
21.
A.
Jain
and
R. C.
Dixit
,
AIP Conf. Proc.
1953
,
040013
(
2018
).
22.
H.
Liu
,
H.-k.
Mao
,
M.
Somayazulu
,
Y.
Ding
,
Y.
Meng
, and
D.
Häusermann
,
Phys. Rev. B
70
,
094114
(
2004
).
23.
S. S.
Bhat
,
U. V.
Waghmare
, and
U.
Ramamurty
,
J. Appl. Phys.
113
,
133507
(
2013
).