Hydrogenation-controlled phase transition on two-dimensional transition metal dichalcogenides and their unique physical and catalytic properties.

Two-dimensional (2D) transition metal dichalcogenides (TMDs) have been widely used from nanodevices to energy harvesting/storage because of their tunable physical and chemical properties. In this work, we systematically investigate the effects of hydrogenation on the structural, electronic, magnetic, and catalytic properties of 33 TMDs based on first-principles calculations. We find that the stable phases of TMD monolayers can transit from 1T to 2H phase or vice versa upon the hydrogenation. We show that the hydrogenation can switch their magnetic and electronic states accompanying with the phase transition. The hydrogenation can tune the magnetic states of TMDs among non-, ferro, para-, and antiferro-magnetism and their electronic states among semiconductor, metal, and half-metal. We further show that, out of 33 TMD monolayers, 2H-TiS2 has impressive catalytic ability comparable to Pt in hydrogen evolution reaction in a wide range of hydrogen coverages. Our findings would shed the light on the multi-functional applications of TMDs.

Extensive attention has been drawn to two-dimensional (2D) transition metal dichacogenides (TMDs) because of their unique chemical, mechanical, electronic and magnetic properties, multi-functional applications in various fields of science and technology from spintronics, optoelectronics, sensors, catalysts to energy harvesting and storage1234567891011121314, and easier fabrication215161718. Depending on the point-group symmetries (D6h and D3d), these 2D monolayers with the formula of MX2 can have 1T (Fig. 1a) or 2H phase (Fig. 1d)2, where M is the transition metal element and X is a chalcogen element (S, Se, and Te). These 2D TMDs show rich physical and chemical properties and can be metallic, half-metallic, semiconducting, magnetic, and catalytic, which can be tuned by phase transition, composition engineering, surface functionalization, and external fields (strain and electrical field)291920212223242526272829303132. For example, 2H MoS2 and WS2 monolayers are semiconductor, while their 1T phases are metallic323334. Hydrogenated MoS2 monolayer can be non-magnetic and ferromagnetic tuned by tensile strain23. Magnetic evolution from non-magnetism, to anti-ferromagnetism, via paramagnetism, then to ferromagnetism accompanying with electronic switching from semiconductor to metal, then to half-metal was achieved on VX2 monolayers by hydrogenation and tensile strain1124. Vanadium disulfide monolayer showed better catalytic performance than its selenides and tellurides counterparts25 and tensile strain can enhance the ability dramatically26. The catalytic performance of MX2 monolayers can be strongly improved by phase transition303132333435363738. For example, the 1T phase MoS2 and WS2 nanosheet was proven to be more catalytically active in facilitating hydrogen production in electrolysis of water than their 2H counterpart, although 2H phase is more stable than its 1T phase30. Theoretically, it was reported that hydrogen-functionalization can trigger the 2H to 1T phase transition of MoS228 and the catalytic activity of 1T-MoS2 mainly arises from its affinity for binding H at the surface S sites28394041. Recently, 1T phase domains were formed in 2H-MX2 monolayer by creating X vacancies and these 2H-1T mixed monolayer showed ferromagnetism31. Although oxidization is a simple way to create the vacancy for the phase transition, it may also result in a lot of defects in the monolayers. In this work, we present a general method – hydrogenation – to realize the phase transition and tune the physical and chemical properties of 33 different MX2 monolayers. We find that 1T and 2H phases can transform to each other upon hydrogenation, depending on the transition metal elements in MX2 monolayers. Accompanying with the phase transition, their electronic properties switch among semiconducting, metallic, and half metallic, and magnetic ground states among nonmagnetic, ferromagnetic, paramagnetic and anti-ferromagnetic states. We further predict that TiS2 monolayer in 2H phase shows effective catalytic performance for hydrogen evolution reaction in a wide range of hydrogen coverages with neutral thermal Gibbs free energies.

Figure 1

The representative unit cells of MX2 monolayers in 1T phase: (a) MX2-0HC, (b) MX2-1HC, (c) MX2-2HC. The representative structures of MX2 monolayers unit cell in 2H phase: (d) MX2-0HC, (e) MX2-1HC, (f) MX2-2HC. (g) the periodic table with highlighted transition metal atoms (M) and chalcogen atoms (X) considered in our calculations.

Results and Discussion

Structural Properties

In our calculations, we focus on 2D transitional metal dichalcogenides (MX2) with M from group IV (Ti, Zr, and Hf), group VI (Cr, Mo, and W), group VII (Tc and Re), and group VIII (Ni, Pd, and Pt), as shown in Fig. 1g. The hydrogenation of these TMDs is realized by putting hydrogen atoms directly on the tops of X atoms41424344. The MX2 monolayers with and without hydrogenation is referred as MX2-nHC, where n equals to 0 (no hydrogenation; Fig. 1a for 1T phase and Fig. 1d for 2H phase, respectively), 1 (one surface fully covered by hydrogen atoms; Fig. 1b for 1T phase and Fig. 1e for 2H phase, respectively), and 2 (two surfaces fully covered by hydrogen atoms; Fig. 1c,f for 1T and 2H phases, respectively). In order to identify the phase transition of 2D TMDs, all the MX2 unit cells with and without hydrogenation in both 1T and 2H phases (Fig. 1) are firstly optimized to obtain the lattice parameters and formation energies. The phase transition can be identified by the energy differences (ΔE1T−2H) between 1T and 2H phases of MX2 as calculated from the following equation:

where E1T and E2H are the total energies of a MX2-nHC in 1T and 2H phases, respectively. If ΔE1T−2H is negative, the 1T phase is more stable, otherwise 2H phase is more stable.

The negative ΔE1T−2H shows that pure MX2 monolayers with metal elements from group IV (M = Ti, Zr, and Hf) are stable in 1T phase (Fig. 2a). When their surfaces are hydrogenated, the 2H phases become stable, as indicated by the positive ΔE1T−2H. We also note that 1T transition metal disulfides (MS2) are more stable than their selenides (MSe2) and tellurides (MTe2) with the same phase due to their larger negative energy differences. Differently, the MX2 monolayers with the metal elements from group VI (M = Cr, Mo, and W) experience a phase transition of 2H → 2H → 1T as the hydrogenation progresses from 0HC, 1HC, to 2HC (Fig. 2b). Our result on MoS2 is consistent with literatures353637. From literatures353637, we also see that pure 2H-MoS2 is more stable than pure 1T-MoS2. As hydrogen coverage increasing, 1T-MoS2 becomes stable28. If the metal element is from group VII, the MX2 (M = Tc and Re) monolayers go through a phase transition of 2H → 1T → 1T as the hydrogenation increases from 0HC, 1HC, to 2HC (Fig. 2c). Totally different from the above three groups, all of the MX2 monolayers with metal elements from group X (M = Ni, Pd, and Pt) are stable in 1T phase regardless of hydrogenation due to large negative energy differences. Our calculations show that pure MX2 monolayers are stable in either 1T or 2H phase depending on the metal elements in their composition, and hydrogenation can trigger the phase transition from one to another effectively, except the MX2 monolayers with M from group X. The MX2 monolayers, where M atoms are in the same group, follow the same trend of phase transition under hydrogenation.

Figure 2

Calculated total energy differences between 1T and 2H phase of MX2 monolayers with and without hydrogenation for metal elements from: (a) group IV: M = Ti, Zr & Hf, (b) group VI: M = Cr, Mo & W, (c) group VII: M = Tc & Re, (d) group X: M = Ni, Pd & Pt.

In the following discussion, the pure and hydrogenated MX2 monolayers with the stable phases are investigated (Table 1 and Fig. 2). The optimized geometries show that the hydrogenation results more or less in the changes of their lattice constants. For MX2 with M from group IV, their lattice constants (a) and the X-M bond length remain unchanged or slightly decrease under hydrogenation. For example, the lattice constant (a) are 3.47 Å for both TiS2-1HC and TiS2-2HC, slightly larger than that of TiS2-0HC (3.42 Å). The lattice constant of pure TiS2 monolayer is consistent with literature45. Normally, the thickness (the vertical distance between two chalcogen atoms) increases upon hydrogenation in all considered MX2 systems. For MX2 monolayers with M from group VI, VII, and X, both the lattice constant (a) and X-M bond length increase as hydrogenation increases from 0HC, 1HC, to 2HC. For example, the lattice constants are extended by 8–9%, 1–9%, and 0–10% for CrX2, MoX2, and WX2, respectively, under hydrogenation.

Table 1

Stable phases and lattice parameters of MX2 monolayers without hydrogenation (0HC), with one-surface full hydrogenation (1HC) and two-surface full hydrogenation.

Group IV	Condition	Stable Phase	a (Å)	X – M (Å)	Thickness (Å)	Group VI	Condition	Stable Phase	a (Å)	X – M (Å)	Thickness (Å)
TiS2	0HC	1T	3.42	2.43	2.83	CrS2	0HC	2H	3.04	2.29	2.94
	1HC	2H	3.47	2.45	4.18		1HC	2H	3.31	2.31	3.97
	2HC	2H	3.47	2.45	5.57		2HC	1T	3.43	2.34	5.25
TiSe2	0HC	1T	3.54	2.56	3.10	CrSe2	0HC	2H	3.21	2.43	3.15
	1HC	2H	3.64	2.58	4.50		1HC	2H	3.49	2.45	4.32
	2HC	2H	3.65	2.58	6.01		2HC	1T	3.67	2.48	5.62
TiTe2	0HC	1T	3.75	2.78	3.49	CrTe2	0HC	2H	3.47	2.64	3.42
	1HC	2H	3.93	2.79	4.97		1HC	2H	3.75	2.65	4.78
	2HC	2H	3.93	2.78	6.66		2HC	1T	3.99	2.66	6.10
ZrS2	0HC	1T	3.69	2.57	2.89	MoS2	0HC	2H	3.19	2.41	3.13
	1HC	2H	3.67	2.59	4.34		1HC	2H	3.22	2.44	4.57
	2HC	2H	3.60	2.59	5.83		2HC	1T	3.44	2.44	5.62
ZrSe2	0HC	1T	3.80	2.70	3.16	MoSe2	0HC	2H	3.32	2.54	3.34
	1HC	2H	3.82	2.71	4.66		1HC	2H	3.38	2.57	4.92
	2HC	2H	3.77	2.71	6.26		2HC	1T	3.65	2.58	5.99
ZrTe2	0HC	1T	3.97	2.92	3.61	MoTe2	0HC	2H	3.56	2.74	3.62
	1HC	2H	4.08	2.91	5.12		1HC	2H	3.90	2.74	4.86
	2HC	2H	4.04	2.90	6.87		2HC	1T	4.04	2.75	6.35
HfS2	0HC	1T	3.64	2.55	2.89	WS2	0HC	2H	3.19	2.42	3.14
	1HC	2H	3.62	2.56	4.32		1HC	2H	3.19	2.44	4.64
	2HC	2H	3.56	2.56	5.81		2HC	1T	3.40	2.44	5.68
HfSe2	0HC	1T	3.77	2.68	3.14	WSe2	0HC	2H	3.32	2.55	3.36
	1HC	2H	3.78	2.69	4.64		1HC	2H	3.34	2.57	4.99
	2HC	2H	3.70	2.68	6.30		2HC	1T	3.61	2.57	6.11
HfTe2	0HC	1T	3.98	2.90	3.52	WTe2	0HC	2H	3.56	2.74	3.63
	1HC	2H	4.04	2.89	5.11		1HC	2H	3.92	2.74	4.83
	2HC	2H	3.99	2.87	6.88		2HC	1T	4.06	2.76	6.40
Group VI						Group X
TcS2	0HC	2H	3.28	2.38	2.89	NiS2	0HC	1T	3.36	2.26	2.32
	1HC	1T	3.68	2.41	3.60		1HC	1T	3.46	2.31	3.69
	2HC	1T	3.63	2.41	5.16		2HC	1T	3.63	2.37	4.96
TcSe2	0HC	2H	3.42	2.51	3.10	NiSe2	0HC	1T	3.54	2.39	2.48
	1HC	1T	3.85	2.53	3.93		1HC	1T	3.66	2.44	3.96
	2HC	1T	3.87	2.54	5.45		2HC	1T	3.81	2.49	5.37
TcTe2	0HC	2H	3.67	2.69	3.33	NiTe2	0HC	1T	3.79	2.58	2.73
	1HC	1T	4.05	2.70	4.41		1HC	1T	3.90	2.61	4.39
	2HC	1T	4.12	2.70	5.98		2HC	1T	4.05	2.66	5.99
ReS2	0HC	2H	3.31	2.40	2.88	PdS2	0HC	1T	3.55	2.40	2.48
	1HC	1T	3.32	2.42	4.35		1HC	1T	3.72	2.46	3.75
	2HC	1T	3.68	2.42	5.07		2HC	1T	3.94	2.53	4.96
ReSe2	0HC	2H	3.46	2.52	3.08	PdSe2	0HC	1T	3.74	2.52	2.61
	1HC	1T	3.86	2.54	3.93		1HC	1T	3.91	2.58	4.01
	2HC	1T	3.89	2.54	5.43		2HC	1T	4.10	2.64	5.36
ReTe2	0HC	2H	3.71	2.70	3.30	PdTe2	0HC	1T	4.03	2.70	2.75
	1HC	1T	4.09	2.71	4.37		1HC	1T	4.14	2.75	6.44
	2HC	1T	4.15	2.70	5.97		2HC	1T	4.31	2.80	6.01
						PtS2	0HC	1T	3.58	2.40	2.46
							1HC	1T	3.93	2.58	3.99
							2HC	1T	4.14	2.64	5.29
						PtSe2	0HC	1T	3.75	2.53	2.62
							1HC	1T	3.93	2.58	3.99
							2HC	1T	4.13	2.64	5.30
						PtTe2	0HC	1T	4.02	2.71	2.78
							1HC	1T	4.15	2.74	4.44
							2HC	1T	4.30	2.79	6.02

Magnetic Properties

Our calculations show that the hydrogenation results in the phase transition of TMD monolayers, which may also tune other physical properties. In this section, we focus on the effect of hydrogenation on the magnetic properties of the TMD monolayers. To find the magnetic ground state, a supercell with 2 × 2 × 1 unit cells for each MX2 system is constructed (referred as 221 supercell in Supporting Data S1). As an indication of stable ground state, the exchange energy (Eex) is calculated as below:

where EAFM and EFM are the total energies of a MX2 monolayer at antiferromagnetic and ferromagnetic states, respectively, and N is the number of unit cells adopted in calculation (N = 4 in a 221 supercell). The MX2 monolayer is ferromagnetic (FM) when Eex is positive, while antiferromagnetic (AFM) when Eex is negative. We consider the systems as non-magnetism (NM) when the absolute values of Eex, EAFM − ENM, and EFM − ENM are less than 10 meV per unit cell because of possible calculation error.

The calculated exchange energies show that nonmagnetic TiX2, ZrS2 and ZrSe2 monolayers switch to ferromagnetic, then back to nonmagnetic as the hydrogenation progresses from 0HC, 1HC, to 2HC (Fig. 3a). Whereas, ZrTe2 and HfX2 monolayers are nonmagnetic regardless of the hydrogenation. The Curie temperatures (TC) of ferromagnetic systems can be estimated from KBTC = (2/3)Eex based on the mean field theory and Heisenberg model46, which are 448, 297, 149, 90, and 91 K for TiTe2-1HC, TiSe2-1HC, TiS2-1HC, ZrS2-1HC and ZrSe2-1HC, respectively (Table 2).

Figure 3

Calculated exchange energies between antiferromagnetic and ferromagnetic states of MX2 monolayers with and without hydrogenation for metal elements from: (a) group IV: M = Ti, Zr & Hf, (b) group VI: M = Cr, Mo & W, (c) group VII: M = Tc & Re, (d) group X: M = Ni, Pd & Pt. The grey region indicates the non-magnetic states.

Table 2

Electronic, and magnetic ground states and magnetic moments of MX2-nHC (n = 0, 1, and 2).

Group IV	Structure	Conductivity	Magnetic state	Magnetic moment of M atom (μB)	Magnetic moments of X atom & X atom with H (μB)	Curie Temperature (K)	Exchange Energy (meV)
TiS2	1T-0HC	semiconductor	NM	—	—	—	0
	2H-1HC	semiconductor	FM	0.46	0.006 & 0.007	149	19
	2H-2HC	semiconductor	NM	—	—	—	0
TiSe2	1T-0HC	metal	NM	—	—	—	0
	2H-1HC	semiconductor	FM	0.62	0.007 & 0.020	297	38
	2H-2HC	semiconductor	NM	—	—	—	0
TiTe2	1T-0HC	metal	NM	—	—	—	0
	2H-1HC	semiconductor	FM	0.71	0.015 & 0.030	448	58
	2H-2HC	semiconductor	NM	—	—	—	0
ZrS2	1T-0HC	semiconductor	NM	—	—	—	0
	2H-1HC	half-metal	FM	0.40	0.027 & 0.003	90	12
	2H-2HC	semiconductor	NM	—	—	—	0
ZrSe2	1T-0HC	semiconductor	NM	—	—	—	0
	2H-1HC	half-metal	FM	0.42	0.013 & 0.002	91	12
	2H-2HC	semiconductor	NM	—	—	—	0
ZrTe2	1T-0HC	metal	NM	—	—	—	0
	2H-1HC	semiconductor	NM	—	—	—	0
	2H-2HC	semiconductor	NM	—	—	—	0
HfS2	1T-0HC	semiconductor	NM	—	—	—	0
	2H-1HC	metal	NM	—	—	—	0
	2H-2HC	semiconductor	NM	—	—	—	0
HfSe2	1T-0HC	semiconductor	NM	—	—	—	0
	2H-1HC	semiconductor	NM	—	—	—	0
	2H-2HC	metal	NM	—	—	—	0
HfTe2	1T-0HC	metal	NM	—	—	—	0
	2H-1HC	metal	NM	—	—	—	2
	2H-2HC	metal	NM	—	—	—	0

Non-magnetism, ferromagnetism, and anti-ferromagnetism are denoted as NM, FM and AFM, respectively.

For MX2 monolayers with M from group VI, the magnetic switching under hydrogenation is slightly complicated. CrS2 and CrSe2 switch following NM → AFM → FM with the hydrogenation increasing (Fig. 3b). CrTe2-0HC and CrTe2-1HC are antiferromagnetic, but CrTe2-2HC is ferromagnetic. MX2 (M = Mo and W) monolayers with and without hydrogenation are nonmagnetic, except MoTe2-2HC and WTe2-2HC are ferromagnetic because their exchange energies are positive (Fig. 3b) and 1T-MoSe2-2HC is paramagnetic because the energies at its magnetic states are lower than that at non-magnetic state (Supporting data, Table S1). Importantly, the Curie temperatures of CeS2-2HC, CrSe2-2HC, MoTe2-2HC, and WTe2-2HC are above 1000 K (Table 3).

Table 3

Electronic, and magnetic ground states and magnetic moments of MX2-nHC (n = 0, 1, and 2).

Group VI	Structure	Conductivity	Magnetic state	Magnetic moment of M atom (μB)	Magnetic moments of X atom & X atom with H (μB)	Curie Temperature (K)	Exchange Energy (meV)
CrS2	2H-0HC	semiconductor	NM	—	—	—	0
	2H-1HC	metal	AFM	2.27	0.044 & 0.021	—	−202
	1T-2HC	half-metal	FM	1.97	0.018	1317	170
CrSe2	2H-0HC	semiconductor	NM	—	—	—	0
	2H-1HC	metal	AFM	2.51	0.051 & 0.006	—	−137
	1T-2HC	metal	FM	2.11	0.039	1012	131
CrTe2	2H-0HC	semiconductor	AFM	1.42	0.056	—	−50
	2H-1HC	metal	AFM	2.67	0.065 & 0.027	—	−90
	1T-2HC	metal	FM	2.62	0.089	596	77
MoS2	2H-0HC	semiconductor	NM	—	—	—	0
	2H-1HC	semiconductor	NM	—	—	—	2
	1T-2HC	metal	NM	—	—	—	0
MoSe2	2H-0HC	semiconductor	NM	—	—	—	0
	2H-1HC	semiconductor	NM	—	—	—	0
	1T-2HC	metal	PM	1.05	0.008	—	−6
MoTe2	2H-0HC	semiconductor	NM	—	—	—	0
	2H-1HC	metal	NM	—	—	—	6
	1T-2HC	half-metal	FM	1.61	0.042	1096	12
WS2	2H-0HC	semiconductor	NM	—	—	—	0
	2H-1HC	semiconductor	NM	—	—	—	0
	1T-2HC	metal	NM	—	—	—	0
WSe2	2H-0HC	semiconductor	NM	—	—	—	0
	2H-1HC	semiconductor	NM	—	—	—	0
	1T-2HC	metal	NM	—	—	—	0
WTe2	2H-0HC	semiconductor	NM	—	—	—	0
	2H-1HC	metal	NM	—	—	—	5
	1T-2HC	half-metal	FM	1.48	0.046	1137	147

Non-magnetism, paramagnetism, ferromagnetism, and anti-ferromagnetism are denoted as NM, PM, FM and AFM, respectively.

For MX2 monolayers with M from VII, we see that MTe2 (M = Tc and Re) are nonmagnetic regardless of hydrogenation (Fig. 3c), except 1T-TcTe2-2HC is paramagnetic because its magnetic states are more stable than its non-magnetic state (Supporting Data, Table S1). For systematic study, Tc is considered although it is radioactive. MSe2 switches from non-magnetism to ferromagnetism after hydrogenation. MS2-0HC and MS2-1HC are non-magnetic, while MS2-2HC monolayers are ferromagnetic (Fig. 3c and Table 4). The Curie temperatures range from 117 to 613 K for ferromagnetic systems in this group.

Table 4

Electronic, and magnetic ground states and magnetic moments of MX2-nHC (n = 0, 1, and 2).

Group VII	Structure	Conductivity	Magnetic state	Magnetic moment of M atom (μB)	Magnetic moments of X atom & X atom with H (μB)	Curie Temperature (K)	Exchange Energy (meV)
TcS2	2H-0HC	metal	NM	—	—	—	0
	1T-1HC	metal	NM	—	—	—	0
	1T-2HC	half-metal	FM	0.94	0.002	541	70
TcSe2	2H-0HC	metal	NM	—	—	—	1
	1T-1HC	half-metal	FM	1.66	0.052 & 0.045	613	80
	1T-2HC	half-metal	FM	0.88	0.012	117	15
TcTe2	2H-0HC	metal	NM	—	—	—	4
	1T-1HC	metal	NM	—	—	—	6
	1T-2HC	metal	PM	0.61	0.004	—	−2
ReS2	2H-0HC	metal	NM	—	—	—	2
	1T-1HC	metal	NM	—	—	—	0
	1T-2HC	half-metal	FM	0.89	0.001	454	59
ReSe2	2H-0HC	metal	NM	—	—	—	4
	1T-1HC	half-metal	FM	1.51	0.051 & 0.036	537	69
	1T-2HC	half-metal	FM	0.84	0.009	204	26
ReTe2	2H-0HC	metal	NM	—	—	—	1
	1T-1HC	metal	NM	—	—	—	1
	1T-2HC	metal	NM	—	—	—	−3

Non-magnetism, paramagnetism, ferromagnetism, and anti-ferromagnetism are denoted as NM, PM, FM and AFM, respectively.

For MX2 with M from group X, NiS2 and NiSe2 show the same magnetic evolution as NM → FM → AFM with the hydrogenation increasing (0HC → 1HC → 2HC), whereas NiTe2 is nonmagnetic regardless of hydrogenation (Fig. 3d and Table 5). The ground states of MS2 and MSe2 (M = Pd and Pt) switch as NM → FM → NM with the hydrogenation increasing (0HC → 1HC → 2HC). However, MTe2 keeps NM unchanged under hydrogenation. The Curie temperatures of ferromagnetic systems in this group varies from 171 to 233 K.

Table 5

Electronic, and magnetic ground states and magnetic moments of MX2-nHC (n = 0, 1, and 2 & M is from group X).

Group X	Condition	Conductivity	Magnetic state	Magnetic moment of M atom (μB)	Magnetic moments of X atom & X atom with H (μB)	Curie Temperature (K)	Exchange Energy (meV)
NiS2	1T-0HC	semiconductor	NM	—	—	—	0
	1T-1HC	half-metal	FM	0.59	0.176 & 0.065	228	30
	1T-2HC	metal	AFM	1.04	0.067	—	−73
NiSe2	1T-0HC	semiconductor	NM	—	—	—	0
	1T-1HC	semiconductor	FM	0.41	0.108 & 0.026	233	30
	1T-2HC	metal	AFM	0.51	0.011	—	−14
NiTe2	1T-0HC	metal	NM	—	—	—	1
	1T-1HC	metal	NM	—	—	—	8
	1T-2HC	metal	NM	—	—	—	8
PdS2	1T-0HC	semiconductor	NM	—	—	—	0
	1T-1HC	semiconductor	FM	0.33	0.220 & 0.054	197	25
	1T-2HC	metal	NM	—	—	—	0
PdSe2	1T-0HC	semiconductor	NM	—	—	—	0
	1T-1HC	semiconductor	FM	0.24	0.170 & 0.031	171	22
	1T-2HC	metal	NM	—	—	—	0
PdTe2	1T-0HC	semiconductor	NM	—	—	—	0
	1T-1HC	metal	NM	—	—	—	2
	1T-2HC	metal	NM	—	—	—	0
PtS2	1T-0HC	semiconductor	NM	—	—	—	0
	1T-1HC	semiconductor	FM	0.29	0.146 & 0.040	195	25
	1T-2HC	metal	NM	—	—	—	0
PtSe2	1T-0HC	semiconductor	NM	—	—	—	0
	1T-1HC	semiconductor	FM	0.29	0.146 & 0.040	195	25
	1T-2HC	metal	NM	—	—	—	0
PtTe2	1T-0HC	semiconductor	NM	—	—	—	0
	1T-1HC	metal	NM	—	—	—	1
	1T-2HC	metal	NM	—	—	—	0

Non-magnetism, ferromagnetism, and anti-ferromagnetism are denoted as NM, FM and AFM, respectively.

Our calculations show that the hydrogenation induces not only the phase transitions of MX2 monolayers, but magnetic switching. Among all 33 considered systems, CrS2-2HC has the highest Tc (1317 K). The estimated Curie temperature needs to be confirmed experimentally. The ferromagnetic MX2 systems with Tc above room temperature may find applications in spintronics. To reveal the origin of the magnetic switching, their electronic structures are calculated.

Electronic Properties

To understand the magnetic evolution of MX2 monolayers under hydrogenation, the electronic structures and magnetic moments of MX2 monolayers with and without hydrogenation are calculated (Figs 4 and 7, S2~S12, and Tables 2~5). The calculated partial density of states (PDOSs) of TiX2-nHC show that nonmagnetic TiX2-0HC and TiX2-2HC monolayers are either metallic or semiconducting (Figs 4a,c and S2a,c), while ferromagnetic TiS2-1HC and TiSe2-1HC monolayers are n-type semiconductors and TiTe2-1HC is narrow band semiconductor (Figs 4b,e,h and S2b,e,h). The d electrons of Ti atoms near the Fermi levels are spin-polarized (Fig. 4b,e,h), leading to the magnetic moments of 0.46, 0.62, and 0.71 μB/Ti in TiX2-1HC (X = S, Se, and Te), respectively (Table 2). The p electrons of X atoms in TiX2-1HC are also weakly spin-polarized (Fig. 4b,e,h), leading to smaller magnetic moments (Table 2), which are anti-parallel to the moments of Ti atoms. Therefore, the ferromagnetism of TiX2-1HC may attribute to the double-exchange111447484950. Pure ZrS2 and ZrSe2 monolayers are semiconductors and their band gaps are reduced by two-surface hydrogenation (S3). The ferromagnetic ZrS2-1HC and ZrSe2-1HC are half-metal (S3b,e). Nonmagnetic ZrTe2 switches from metal, n-type semiconductor, to intrinsic semiconductor as hydrogenation increases from 0HC, 1HC, and 2HC (S3g–i). Pure nonmagnetic HfS2 and HfSe2 monolayers are semiconductors, and switch to metal upon hydrogenation (S4b,c,e,f), except that HfS2-2HC is narrow-band semiconductor (S4c). Nonmagnetic HfTe2 monolayer keeps metallic regardless of hydrogenation (S4g,h,i).

Figure 4

Calculated partial density of states of (a) TiS2-0HC, (b) TiS2-1HC, (c) TiS2- 2HC, (d) TiSe2-0HC, (e) TiSe2-1HC, (f) TiSe2-2HC, (g) TiTe2-0HC, (h) TiTe2-1HC and (i) TiTe2-2HC monolayers. The Fermi level is at 0 eV and indicated with gray dotted line.

Figure 5

Calculated partial density of states of (a) CrS2-0HC, (b) CrS2-1HC, (c) CrS2- 2HC, (d) CrSe2-0HC, (e) CrSe2-1HC, (f) CrSe2-2HC, (g) CrTe2-0HC, (h) CrTe2-1HC and (i) CrTe2-2HC monolayers. The Fermi level is at 0 eV and indicated with gray dotted line.

Figure 6

Calculated partial density of states of (a) TcS2-0HC, (b) TcS2-1HC, (c) TcS2- 2HC, (d) TcSe2-0HC, (e) TcSe2-1HC, (f) TcSe2-2HC, (g) TcTe2-0HC, (h) TcTe2-1HC and (i) TcTe2-2HC monolayers. The Fermi level is at 0 eV and indicated with gray dotted line.

Figure 7

Calculated partial density of states of (a) NiS2-0HC, (b) NiS2-1HC, (c) NiS2- 2HC, (d) NiSe2-0HC, (e) NiSe2-1HC, (f) NiSe2-2HC, (g) NiTe2-0HC, (h) NiTe2-1HC and (i) NiTe2-2HC monolayers. The Fermi level is at 0 eV and indicated with gray dotted line.

For MX2 with M from group VI, accompanying with the magnetic switching of NM → AFM → FM, the electronic properties of CrX2 monolayers switch from semiconductor, metal, to half-metal as hydrogenation increases (0HC → 1HC → 2HC) (Figs 5 and S5). Nonmagnetic MoX2 (X = S and Se) systems switch from intrinsic semiconductor, n-type semiconductor, to metal as hydrogenation increases (S6a–f). Semiconducting MoTe2 monolayer transfers to metal upon 1HC, and to half-metal upon 2HC (S6g–i). WX2 monolayers show the same electronic switching as MoX2 upon hydrogenation (S7).

For MX2 with M from group VII, nonmagnetic TcS2-0HC and TcS2-1HC are metallic, and ferromagnetic TcS2-2HC is half-metallic (Figs 6a–c and S8). Nonmagnetic TcSe2-0HC is metal (Fig. 6d) and ferromagnetic TcSe2-1HC and TcSe2-2HC are half-metals, respectively (Fig. 6e,f). Nonmagnetic TcTe2 monolayers with and without hydrogenation are metallic (Figs 6g–i and S8). The nonmagnetic ReX2 monolayers with and without hydrogenation are metallic (S9a,b,d,g–i), while ferromagnetic counterparts are half-metallic (S9c,e,f).

For MX2 with M from group X, semiconducting NiX2 (X = S and Se) monolayers switch to half-metal upon 1HC (Figs 7b,e and S10b,e), and metallic upon 2HC (Figs 7c,f and S10c,f). Nonmagnetic NiTe2 systems keep metallic regardless of hydrogenation (Figs 7g–i and S10 g–i). Nonmagnetic MX2-0HC and MX2-2HC monolayers (M = Pd and Pt, X = S and Se) are semiconducting and metallic, respectively (S12a,c,d,f, and S13a,c,d,f), and MX2-1HC monolayers are ferromagnetic n-type semiconductors (S11b,e and S12b,e).

The calculated electronic properties clearly show that metallic or semiconducting systems are nonmagnetic/antiferromagnetic, and half metallic or doped semiconducting systems are ferromagnetic. The ferromagnetism is contributed to the carrier-mediated double exchange112447484950.

Catalytic Ability for Hydrogen Evolution

MX2 monolayers have been widely investigated as electrocatalysts for hydrogen evolution reaction (HER)252632333442434451525354555657585960. In this section, we investigate the catalytic activities of the considered 33 MX2 systems. To characterize their catalytic performances, Gibbs free energies (∆GH) are calculated based on published methods43535460. A catalyst with optimal performance needs to have near-zero ∆GH. Two HER processes, individual and collective processes, are considered. I-ΔGH and A-ΔGH are referred as the Gibbs free energies calculated from individual and collective processes, respectively2653. To investigate the catalytic activities of the 33 MX2 monolayers in HER, a supercell with 3 × 3 × 1 unit cells for each MX2-1HC monolayer, referred as 331 supercell, is constructed (S13). Hydrogen atoms are taken away one by one from 331 supercell to calculate the Gibbs free energies, where the partial hydrogen coverages is referred as (n = 1–9). Both 1T and 2H phases are considered because of the possible phase transition as discussed above.

Out of 33 MX2 monolayers, we find that TiS2 monolayer shows excellent catalytic ability at a wide range of hydrogen coverages. Our calculations show that both I-ΔGH and A-ΔGH of TiS2-1HC increase with the increment of hydrogen coverages (Fig. 8a–d), which are much closer to zero than those of TiSe2 and TiTe2 at the same hydrogen coverages, similar to VX2 monolayers25. Interestingly, we see that 2H phase shows much better catalytic performance than 1T phase because of the relatively lower overpotentials (absolute value of Gibbs free energy) at the same hydrogen coverages. As discussed above, 1T phase of pure TiX2 is more stable than its 2H phase, indicating that the metastable phase shows better catalytic performance, which is similar to MoS2 and WS2 monolayers303233. At the same time, we showed above that the phase transition from 1T to 2H can be easily realized through surface hydrogenation. Therefore, 2H-TiX2 could be achieved and stabilized during the process of HER. Importantly, I-ΔGH of 2H-TiS2 is close-to-zero (−0.14, −0.01, 0.07, and 0.09 eV) in a hydrogen coverage ranging from to in the individual process. For collective process, A-ΔGH (−0.14, −0.07, −0.02, 0.006, and 0.06 eV) is near-zero in a hydrogen density of from to . The near-zero Gibbs free energies clearly indicate that 2H-TiS2 monolayer show high catalytic ability within a wide range of hydrogen coverages. The catalytic activity of 2H-TiS2 monolayer is much better than other MX2 that only showed catalytic activity at certain hydrogen coverage. For example, VS2 monolayer was good at low-hydrogen coverages25. The Mo-edge of MoS2 was only catalytically active with an A-ΔGH of 0.06 eV at a hydrogen density of 44. 1T-WS2 monolayer has an I-ΔGH of 0.28 eV at a hydrogen density of 30. Therefore, TiS2 monolayer possesses overall excellent catalytic ability, which is comparable to Pt, due to near-zero ΔGH in wide range of hydrogen densities.

Figure 8

Calculated Gibbs free energies of TiS2 monolayer as a function of one-surface hydrogen coverages in 1T phase: (a) I-ΔGH, and (b) A-ΔGH, in 2H phase: (c) I-ΔGH, and (d) A-ΔGH.

Conclusions

In summary, we present a comprehensive first-principles calculations on the physical and chemical properties of 2D TMDs with and without hydrogenation. We find that the hydrogenation plays an important role on tuning the structural, electronic, magnetic and catalytic properties of TMD monolayers (Fig. 9). We show that pure 1T-MX2 (M in group IV) monolayers transfer into 2H phase upon hydrogenation, their ground states can be tuned from non-magnetic to ferromagnetic accompanying with electronic switching from semiconducting to half-metallic. Phase transition between 1T and 2H phases in MX2 (M is from group VI and VII), and magnetic and electronic switching in MX2 (M is from group VI, VII, X) can be realized by the hydrogenation. We further predict that 2H-TiS2 monolayer, out of 33 MX2 monolayers, has excellent catalytic ability in a wide range of hydrogen coverage and may find applications as electrocatalyst in hydrogen evolution reaction. It is expected that MX2 monolayers with controllable structure, tunable electronic, magnetic and optimized catalytic properties can find applications in catalysts, spintronics, sensors and nanodevices.

Figure 9

Summary of our findings on the structural, magnetic, electronic and catalytic properties of 33 TMDs monolayers with and without hydrogenation.

Methods

The first-principles calculations are conducted to systematically investigate the structural, electronic and magnetic properties of 2D TMDs monolayers through hydrogenation as well as their catalytic ability in hydrogen evolution reduction. Based on the density functional theory (DFT)61 and the Perdew-Burke-Eznerhof generalized gradient approximation (PBE-GGA)62, our calculations are carried out by using the Vienna ab initio simulation package (VASP)63, which is incorporated with projector augmented wave (PAW) scheme6465. An energy cut-off of 500 eV is consistently used in our calculations. Large vacuum regions of 20 Å along vertical directions are used in constructing the unit cells to avoid interaction between neighboring monolayers. The integration over the first Brillouin zone is based on the Monkhorst and Pack scheme of k-point sampling66. The 13 × 13 × 1 grid, 5 × 5 × 1 grid, and 3 × 3 × 1 grid for k-point sampling are used for geometry relaxation of unit cells, 2 × 2 × 1 supercells, and 3 × 3 × 1 supercells, respectively. Spin-polarized calculations are also employed to study the magnetic properties. Good Convergence is obtained with these parameters and the total energy is converged to 2.0 × 10−5 eV/atom.

The Gibbs free energy is an important descriptor for electrocatalyst in HER, and can be calculated as ΔGH = ΔEH + ΔEZPE − TΔSH, where ΔEH is the hydrogen chemisorption energy defined as I-ΔEH = E(MX2 + mH) − E(MX2 + (m–1)H) − E(H2)/2 for individual process and A-ΔEH = [E(MX2 + mH) − E(MX2) − mE(H2)/2 ]/m for average process, respectively. E(H2), E(MX2) and E(MX2 + mH) are the calculated total energies of a hydrogen molecule, pure MX2 and hydrogenated MX2, m is the number of hydrogen atoms adsorbed on a monolayer. ΔEZPE is the difference in zero point energy between the adsorbed and the gas phase, ΔSH is the difference in entropy, ΔEZPE − TΔSH is about 0.24 eV254243444553. Therefore, Gibbs free energy can be calculated as ΔGH = ΔEH + 0.24. Positive ΔGH of a catalyst suggests weak adsorption of protons, leading to absorbing less protons on its surface, however, negative ΔGH indicates strong binding of protons on a catalyst’s surface, resulting in difficult desorption. Electrocatalysts with neutral-thermal (close-to-zero) ΔGH exhibit optimal catalytic ability in hydrogen evolution with neither stronge nor weak binding of protons in the electrolyte.

Additional Information

How to cite this article: Qu, Y. et al. Hydrogenation-controlled phase transition on two-dimensional transition metal dichalcogenides and their unique physical and catalytic properties. Sci. Rep. 6, 34186; doi: 10.1038/srep34186 (2016).