Mechanism of highly enhanced hydrogen storage by two-dimensional 1T′ MoS2

Junyu Chen a, Jiamu Cao *ab, Jing Zhou *a, Yufeng Zhang *ab, Mingxue Li a, Weiqi Wang a, Junfeng Liu a and Xiaowei Liu ab
aMEMS Center, Harbin Institute of Technology, Harbin, China. E-mail: caojiamu@hit.edu.cn; daxiongmao@hit.edu.cn; yufeng_zhang@hit.edu.cn
bKey Laboratory of Micro-systems and Micro-Structures Manufacturing, Ministry of Education, Harbin, China

Received 8th August 2019 , Accepted 19th October 2019

First published on 22nd October 2019


Hydrogen energy is a high-efficiency and clean energy, but the problem of storage still prevents its extensive use. Large-surface-area, two-dimensional (2D) layered materials have an advantage in hydrogen storage applications. Monolayer MoS2 is a typical 2D material that has been widely studied recently. And the 1T′ phase of MoS2 is a focus especially for studies concerning hydrogen. Here, first-principles calculations are carried out to investigate the adsorption behaviors of hydrogen molecules on 1T′ MoS2. Comparing with other MoS2-based materials, such as doped or decorated 2H-MoS2, 1T′ MoS2 has even better performance in hydrogen adsorption, and its preparation is easier. In multiple hydrogen molecule adsorption, the material shows good stability and appropriate adsorption energy while adsorbing hydrogen molecules. With the researches in this paper, the connection between the adsorption energy and hydrogen mass fraction was set up. This can provide a reference for further studies on hydrogen storage applications.


1. Introduction

The demand for green energy sources is one of the key challenges in the 21st century.1,2 Hydrogen, a high-efficiency and clean energy source, has been considered to be a primary energy carrier with extensive applications.3,4 However, due to the low ignition energy and wide range of flammability, gaseous hydrogen is difficult to store or transport on a large scale.5 Therefore, finding materials that can effectively contain hydrogen is crucial for the application of hydrogen energy.6 Recently, two-dimensional (2D) materials have been used to hold hydrogen molecules with superior performance, mainly due to their large surface-to-volume ratio and special electronic structure.7,8 The success of graphene-based materials has motivated researches on other 2D materials for hydrogen storage.9,10 In particular, monolayer MoS2-based materials have also shown superior performance in hydrogen storage. For example, Li-adsorbed MoS211 could adsorb up to 4.4 wt% H2. Though it seems inferior to the recent graphene-based materials which could accommodate over 10 wt% hydrogen,12–14 the density of MoS2 is far larger than that of graphene. The actual mass of adsorbed hydrogen on MoS2-based materials is even larger.

The hydrogen storage capacity mainly depends on the adsorption strength of the gas molecules over the carrier's surface which should be moderate.15 Too strong a force would make it hard for the target gas to break away from the carrier material while too weak a force means the carrier material cannot hold the target gas stably.16 A convenient range of average binding energy per hydrogen molecule for hydrogen storage applications is −0.2 to −0.6 eV per H2 at room temperature.17 Furthermore, the disadvantage of pristine graphene or 2H MoS2 is their too weak binding force for hydrogen molecules.18,19 Methods should be found to improve the adsorption energy effectively.

Doping and adding decoration atoms are common methods for improving materials’ properties. With rational doping or decorating processes, the surface characteristics could be changed so as to be able to achieve a moderate hydrogen adsorption energy range.20 Also, for hydrogen adsorption applications, the binding force for hydrogen molecules could be adjusted to an ideal range by these methods. However, for graphene and 2H MoS2, doped or decorated materials with considerable hydrogen storage performance are usually complicated and hard to prepare.21,22 And the stability of many doped or decorated systems in different conditions cannot be guaranteed.23 Thus, it is better to find pure materials with considerable hydrogen adsorbing performance. Phase-changed materials could be the solution and phase-changed MoS2 is the most studied one.

Metallic 1T-phase MoS2 has been more and more studied in recent years. Though it does not exist in nature, high-percentage 1T-phase MoS2 can be prepared using simple methods.24–26 Though it is found to be unstable,27 it can be applied after some stabilization methods,28,29 and the high chemical functionalization activity on the surface makes it a focus for research.30 Similar to 1T MoS2, 1T′-phase MoS2 is also a polymorph of MoS2 found in recent years.31 It evolves from 1T MoS2 but is found to be relatively stable. Also, the semimetallic character makes its application range wider. For both 1T and 1T′ MoS2, much research has shown their perfect performance in hydrogen evolution applications,28,32,33 the main reason being their surfaces’ strong adsorption strength for hydrogen atoms.34 This indicates they are likely to have stronger adsorbing force for hydrogen molecules than 2H MoS2. But their performance in hydrogen storage applications has not been studied before, and consideration should be given to this field.

In this work, we investigated the potential of 2D pristine 1T and 1T′ MoS2 as hydrogen storage materials. A systematic theoretical study of the structures was performed. Besides, adsorption energy and adsorption configuration of gas molecules were also analyzed. To associate with reality, the situations of multiple hydrogen molecules adsorbed on 1T′-phase MoS2 were also investigated. Thus, a theoretical basis for the applications of 1T′-phase MoS2 was provided in the hydrogen storage area. Also worth noting is that the unstable 1T MoS2 will be harder to change back to the 2H phase if it could change to the relatively stable 1T′ phase. And we found that 1T-phase MoS2 tended to convert to 1T′-phase MoS2 while adsorbing hydrogen molecules. Thus, this work can also provide a theoretical basis for the phase engineering of MoS2.

2. Computational details

Based on density functional theory (DFT), first-principles calculations were used. All calculations in this paper were conducted in Dmol3.35 The local density approximation (LDA) was used with the PWC function to deal with the exchange and correlation potentials. To reduce the computational cost, the single effective potential was used to replace kernel (DFT semi-core pseudopotentials). Dual numerical orbital basis set and orbital polarization function (DNP) were chosen to achieve higher accuracy. Then a convergence test was carried out. After that, the Monkhorst–Pack k-points were set as 4 × 4 × 1, and a vacuum layer of 20 Å was determined to prevent interlayer interactions. The energy convergence precision was set to 1 × 10−5 Hartree, the maximum displacement was 0.005 Å and the atomic forces did not exceed 0.002 Hartree per Å. All the later calculations follow these properties.

For both 1T-phase and 1T′-phase MoS2, the calculation models were supercells of 4 × 4 monolayer MoS2. The 1T MoS2 structure was first constructed by ourselves, then given a geometry optimization including cell optimization. While the 1T′ MoS2 structure was built on the basis of a 2 × 2 1T model, we set a single hydrogen atom connected to a S atom of 1T MoS2, then gave the system a geometry optimization. After that, the hydrogen atom was deleted and the structure optimized again, then the final normal 1T′ structure was obtained. As shown in Fig. 1(a) and (b), the models of 1T-phase MoS2 and 1T′-phase MoS2 contain 32 S atoms and 16 Mo atoms in the cell. Though 25 Mo atoms are present in the 1T-phase model, 16 of them are at the sites of edge or corner, so the valid quantity of Mo atoms in the cell is still 16. We can also see that 1T′ MoS2 is actually atypical 1T MoS2. The Mo atoms in the 1T model are equidistant, while those in the 1T′ model are not equal. Since the Mo atoms arranged in the 1T′ model look like zippers, this is also called the zipper structure in some studies. Comparing the structures, similar repeat units could be found, such as the black box in Fig. 1 shows. The blue boxes represent the repeat units for which the vertexes are all Mo atoms, the one in the 1T′ model being nearly twice as big as that in the 1T model. And the red boxes are the repeat units for which the vertexes are all S atoms, the size of them being similar, but that in the 1T model is much more regular. There are also structure units that are a mirror image to the one that the red box shows in the 1T′ model (Fig. S2(b), ESI), and that is the symmetry expressed in the 1T′ MoS2 structure.


image file: c9cp04402g-f1.tif
Fig. 1 The 4 × 4 × 1 supercell model of (a) and (c) pristine 1T-phase MoS2 and (b) and (d) pristine 1T′-phase MoS2.

To build a hydrogen adsorption model, a single hydrogen molecule is set above the MoS2 plane in the c-axis direction. But the adsorption sites have not been settled before, so we chose several sites that had high geometrical symmetry. For the case of 1T MoS2, the five sites are: right above the upper layer S atom site, right above the lower layer S atom site, right above the Mo atom site, above the middle of Mo atom and upper layer S atom, and above the middle of Mo atom and lower layer S atom (all shown in Fig. 2(a) and (c)). For 1T′ MoS2, as shown in Fig. 2(b) and (d), the six situations are all right above the atoms in the model (which are Mo-1, the upper S-1, the lower S-2, Mo-2, the upper S-2 and the lower S-1). Besides, we distinguished the posture of the hydrogen molecules adsorbed on the surface of 1T′ MoS2, and the adsorbed hydrogen molecule was set either horizontally or vertically to the plane of the material (shown in Fig. S1, ESI), and that doubled the adsorption situations presented in Fig. 2. After the geometry optimization, we chose the best adsorption site according to the results (Fig. S2, ESI). For the adsorption of hydrogen molecules, the adsorption energy is calculated by the following function: Eads = Ehyd+sysEsysEhyd, where Ehyd+sys is the total energy of 1T or 1T′ MoS2 and the adsorbed hydrogen molecules; Esys represents the total energy of pristine 1T or 1T′ MoS2; Ehyd is the total energy of a single free hydrogen molecule. On the basis of this relation, a higher absolute value of Eads means a more stable adsorption system. A positive value of Eads means repulsive force in the adsorption system, while a negative value represents an attraction force. The absolute value of Eads can reflect the force acting between the materials and target gas molecules. Although this method cannot give the exact adsorption energy, the form and strength of interaction between hydrogen and adsorbing material can be reflected.36 As introduced above, for hydrogen storage applications, the ideal adsorption energy for each hydrogen molecule is −0.2 to −0.6 eV per H2 at room temperature.17


image file: c9cp04402g-f2.tif
Fig. 2 The chosen hydrogen placement sites for (a) and (c) 1T MoS2 and (b) and (d) 1T′ MoS2.

3. Results and discussion

After the geometry optimization, for both pristine materials and adsorption systems, the structures with the lowest energy were found. The lengths of the Mo–S bonds in the monolayer pristine 1T-phase MoS2 are all 2.427 Å, while those of 1T′-phase MoS2 are unequal; there are three lengths: 2.386 Å, 2.411 Å, and 2.463 Å. Also, there are Mo–Mo bonds in the optimized 1T′ model, the length of which is 2.778 Å. Both structures are confirmed from previous studies (Table S1, ESI). Fig. 3(a) shows the band structure of the optimized monolayer pristine 1T′-phase MoS2, and that of the 1T-phase counterpart. We can see that for the 1T-phase MoS2, there is no band gap for it is metallic. While for 1T′ MoS2, a semimetallic band structure could be seen. Fig. 3(b) and (c) shows the partial density of states (PDOS) of the two MoS2 models. From the outline of the PDOS results, we can see that the S-p and Mo-d orbitals conform most to the shape of the total DOS, which means the total DOS of 1T (1T′) MoS2 is mostly offered by the S-p and Mo-d orbitals. However, we cannot see the semimetallic character of 1T′ MoS2 in the DOS results since the points in the DOS plot are not enough to meet the high accuracy requirement. But the tendency really confirms the energy band structure results, also in agreement with previous studies.37–40 For adsorption systems, the most stable positions of adsorbed hydrogen molecules on 1T and 1T′ MoS2 are obtained after comparing the adsorption energies of different cases (Fig. S2, ESI). As introduced above, they were chosen by comparing the value of Eads of the listed situations of a single hydrogen molecule adsorbed on 1T or 1T′ MoS2 models (the Eads results are listed in Table S4, ESI). For 1T MoS2, this was above the middle of Mo atom and lower layer S atom (at site no. 5 in Fig. 2(a)); while for 1T′ MoS2, it was above the lower S-1 (at site no. 6 in Fig. 2(c)).
image file: c9cp04402g-f3.tif
Fig. 3 (a) Band structure of 4 × 4 × 1 monolayer 1T (1T′) MoS2. (b) PDOS of 4 × 4 × 1 monolayer 1T′ MoS2. (c) PDOS of 4 × 4 × 1 monolayer 1T MoS2.

What is interesting is that, after a hydrogen molecule was adsorbed on the 1T MoS2 surface, the Mo–Mo bond appeared and the adjacent row of Mo atoms began to get close, which indicated that the structure of 1T MoS2 had a tendency to change to that of 1T′ MoS2. Similar conversions happened in other adsorbing cases of 1T MoS2 (Fig. S3, ESI). To validate the process of conversion, we used a transition state search to calculate the energy change and try to find the intermediate state of the conversion (Fig. S4, ESI). From the results, we found there was no energy barrier in the process which confirms the results of a previous study.41 This means that 1T MoS2 could transform into the final structure, which is similar to 1T′ MoS2, spontaneously under the conditions of hydrogen adsorption. The relaxation energy for a 4 × 4 1T MoS2 supercell to change to the 1T′-like structure is about −5 eV (Fig. S4(b), ESI). This is a non-negligible phenomenon since unstable 1T-phase MoS2 also usually converts to 2H MoS2 spontaneously. If it could change to the relatively stable 1T′ MoS2, it would not change back to 2H MoS2. This is meaningful for the phase engineering of MoS2. For hydrogen storage aspects, this indicates that 1T-phase MoS2 is not stable while adsorbing hydrogen. Besides, the calculated absolute value of Eads for 1T MoS2 cases was far larger than 0.6 eV (Table S4, ESI). This means that releasing the adsorbed hydrogen molecules from the 1T MoS2 surface is very hard. So, from the adsorption energy results, 1T MoS2 is not a good candidate for hydrogen storage applications. But the hydrogen adsorption energies for 1T′ MoS2 are moderate (Table S4, ESI), and really suitable for hydrogen storage. Furthermore, the adsorption energy results also show that 1T′ MoS2 has much better performance in hydrogen adsorption for storage applications than the pristine 2H counterpart which we have studied before (Fig. S5, Tables S2 and S4, ESI).

To further study the interaction between adsorbed hydrogen and 1T′ MoS2, a PDOS analysis was carried out, the results of which are shown in Fig. 4. Fig. 4(a) and (b) present the PDOS of two situations with the lowest adsorption energy of the hydrogen molecule, while Fig. 4(c) and (d) are the two situations with the highest adsorption energy. However, there were only small differences in adsorption energy among the first three situations shown, their absolute value being all smaller than 0.35 eV (Table S5, ESI). The PDOS results also agree with the adsorption energy results. The main peaks in the hydrogen PDOS results were all at −3 to −4 eV. While for the last case, the peak of hydrogen was wider than the other three cases, which was from −3 to −6 eV. And that resulted in a larger superposition between the PDOS of hydrogen and MoS2, indicating a stronger interaction between them. This is just the reason why the largest adsorption energy is obtained from the last situation.


image file: c9cp04402g-f4.tif
Fig. 4 PDOS of H2 adsorption system for (a) hydrogen vertical on the upper S-2, (b) hydrogen horizontal on the upper S-1, (c) hydrogen horizontal on the upper S-2 and (d) hydrogen horizontal on the lower S-1.

As shown in Fig. 5, we set different numbers of hydrogen molecules, from 2 to 64, on the surface of 1T′ MoS2. When the number of hydrogen molecules is below 16, the initial sites of all the adsorbed molecules are equivalent to the most stable position discussed above (as shown in Fig. 5(a) and (e)). However, the hydrogen molecules could be set on both sides of the MoS2 plane, and they could be either concentrated or scattered. The arrangement of hydrogen molecules may also influence the adsorption results. For these, we did two kinds of tests separately. First, we investigated the cases when hydrogen molecules were adsorbed on different sides. Specifically, when the number of hydrogen molecules is 2, the first molecule is settled on the most stable adsorption site of monolayer 1T′ MoS2. The second one could be set on the other corresponding sites (except the neighboring sites) on the same side or on the other side of the 1T′ MoS2 plane. When the number of hydrogen molecules is 4, there are three cases: four on the same side, three on one side and one on the other, and two on each side (shown in Fig. S6(a)–(e), ESI). All the adsorption sites in this test were not neighboring. The adsorption energy for the cases discussed above was calculated (Table S5, ESI). The results show that the absolute values of adsorption energy are close for different cases as long as the number of hydrogen molecules is equal. This indicates that hydrogen molecules arranged on different sides of the 1T′ MoS2 plane hardly influence the adsorption results. Then, we considered the situation when two hydrogen molecules were set on neighboring sites (shown in Fig. S6(f), ESI). We can see that the absolute value of adsorption energy becomes larger (Table S5, ESI). This is due to the interaction between the two adsorbed molecules, and the electron density difference results could reflect this point (Fig. S7, ESI). Based on these test results, we finally set hydrogen molecules as follows. When the number of molecules is below 8, the molecules are set on either side but not neighboring sites of the most stable adsorption sites. When the number of molecules is 8 to 16, the neighboring sites become inevitable. Thus, 8 of them are set on the previous separated adsorption sites, and the rest are inserted at the neighboring corresponding sites. For the situations where the number of hydrogen molecules is between 16 and 48, 16 of them are at the most stable position, the rest are set vertically on the Mo sites of 1T′ MoS2 (as shown in Fig. 5(b), (c), (f), (g)). The sites are chosen after an overall consideration of the stability of adsorption and the distance of arranged hydrogen molecules. When the number of hydrogen molecules is more than 48, the rest of them were set vertically above the lower S-1 sites. We chose these sites since they are sunken positions of the 1T′ MoS2, much easier to accommodate hydrogen molecules (explained in Fig. S8, ESI). Also, we guaranteed the distance between the hydrogen molecules to be not too close and tried to distribute evenly on both sides of the adsorbing materials.


image file: c9cp04402g-f5.tif
Fig. 5 Initial positions when multiple H2 molecules are adsorbed on 1T′-phase MoS2: (a) and (e) 16 hydrogen molecules; (b) and (f) 32 hydrogen molecules; (c) and (g) 48 hydrogen molecules; (d) and (h) 64 hydrogen molecules.

After conducting a geometry optimization, the adsorption energy of the hydrogen molecules can be calculated (Table S6, ESI). As Fig. 6(a) shows, the total adsorption energy increases nearly linearly with an increase of the number of hydrogen molecules. This means the force acting between the adsorbed molecules and material does not change much when the number of hydrogen molecules increases. The green area represents the moderate adsorption energy area. Furthermore, the average adsorption energy can better reflect this point, as shown in Fig. 6(b). The average adsorption energies for hydrogen molecules are between −0.2 and −0.6 eV when the number of hydrogen molecules is below 52 (below the red area). This means that the theoretical reasonable adsorption quantity for hydrogen on 1T′ MoS2 can be up to 3.9 wt%. The average adsorption energy decreases when the number is less than 8, then increases. It is usual that the average force acting on adsorbed gas molecules becomes weaker when the quantity of the adsorbed gas increases. But why does the average adsorption energy keep increasing when the number of hydrogen molecules is over 8? To seek the reason, PDOS studies were carried out again, as shown in Fig. 7(a) and (b).


image file: c9cp04402g-f6.tif
Fig. 6 Graphs of (a) total adsorption energy and (b) average adsorption energy as a function of the number of hydrogen molecules adsorbed on 1T′ MoS2.

image file: c9cp04402g-f7.tif
Fig. 7 PDOS results of multiple H2 molecule adsorption systems for (a) all hydrogen molecules and (b) single hydrogen molecules. Electron density difference results for adsorption system with (c) 4 H2 molecules, (d) 16 H2 molecules, (e) 32 H2 molecules and (f) 64 H2 molecules. The isosurface value is taken as 0.0025 e Bohr−3.

From the results, we can see that, with increasing number of hydrogen molecules, the total PDOS of the adsorbed hydrogen molecules become scattered (especially when the number of hydrogen molecules is more than 16). And for single hydrogen molecules in the systems, the PDOS range for them also becomes wider. But the PDOS for Mo atoms and S atoms remain unchanged (Fig. S8, ESI), indicating the stability of the MoS2 material while adsorbing hydrogen molecules. Also, as a consequence, the superposition area between the PDOS of hydrogen and MoS2 materials becomes larger, the interaction getting stronger. That is why the average adsorption energy keeps rising as the number of hydrogen molecules increases. But it is not only the interaction between the material and gas molecules but also the force acting among the gas molecules that increases the adsorption energy. The electron density difference results explained this point (Fig. S7, ESI). Fig. 7(c)–(f) shows the electron density difference of the adsorption systems with more hydrogen molecules adsorbed. Red areas represent positive areas while green areas are negative areas. From the results we can see that red areas mainly appear beside S atoms, indicating S atoms tend to obtain electrons while adsorbing hydrogen molecules. Accordingly, the electron donors are mainly hydrogen molecules, as the green areas indicate. Besides, the results also show that the molecules become concentrated with an increasing number of hydrogen molecules, and the electron transfer between hydrogen molecules and substrate material increases. That adds to the force acting in the adsorption system, leading to larger adsorption energy.

4. Conclusions

In this paper, models of hydrogen molecules adsorbed on monolayer 1T and 1T′ MoS2 were built. And with the LDA method, the potential of these types of MoS2 as hydrogen storage materials was explored. The adsorption energy for hydrogen molecules on 1T′ MoS2 is in the middle area of the ideal range (−0.2 to −0.6 eV) which is far better than that of the pristine 2H counterpart. The highly enhanced performance for hydrogen adsorption using a phase change could also provide a reference for similar studies. Besides, the multiple hydrogen molecule adsorption study also reveals that 1T′ MoS2 has good hydrogen storage ability, theoretically. Under the guarantee of the adsorption energy being in the range of −0.2 to −0.6 eV, a reasonable maximum of 3.9 wt% hydrogen adsorption ratio was obtained. This means that 1T′ MoS2 is comparable to graphene-based materials and to other decorated or doped 2H MoS2 materials for hydrogen storage applications. The value of the material also lies in the relatively simple preparation methods. Combining all the calculation results in this study, the phase-changed 1T′-phase MoS2 has great potential in hydrogen storage applications. Moreover, another intriguing discovery was that 1T MoS2 has a tendency to convert to 1T′ MoS2 after adsorbing hydrogen molecules, which could solve the stability issues of pristine 1T MoS2. This can provide a theoretical reference for the phase engineering of MoS2.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

The work described in this paper was financially supported by the Fundamental Research Funds for the Central Universities (grant no. HIT.NSRIF.2020022). The authors acknowledge the support from the State Key Laboratory of Urban Water Resource and Environment, Harbin Institute of Technology (no. 2016TS 06).

Notes and references

  1. A. Schneemann, J. L. White, S. Y. Kang, S. Jeong, L. F. Wan, E. S. Cho, T. W. Heo, D. Prendergast, J. J. Urban, B. C. Wood, M. D. Allendorf and V. Stavila, Chem. Rev., 2018, 118, 10775–10839 CrossRef CAS.
  2. D. Mellmann, P. Sponholz and H. Junge, Chem. Soc. Rev., 2016, 45, 3954–3988 RSC.
  3. Q.-L. Zhu and Q. Xu, Energy Environ. Sci., 2015, 8, 478 RSC.
  4. M. Yadav and Q. Xu, Energy Environ. Sci., 2012, 5, 9698 RSC.
  5. B. Sakintuna, F. Lamari-Darkrim and M. Hirscher, Int. J. Hydrog. Energy, 2007, 32, 1121–1140 CrossRef CAS.
  6. U. Eberle, M. Felderhoff and F. Schuth, Angew. Chem., Int. Ed., 2009, 48, 6608–6630 CrossRef CAS.
  7. M. Chhowalla, H. S. Shin, G. Eda, L.-J. Li, K. P. Loh and H. Zhang, Nat. Chem., 2013, 5, 263–275 CrossRef.
  8. Z. Yuan, J. Zhang, F. Meng, Y. Li, Y. Chang, J. Zhao, E. Han and S. Wang, IEEE Trans. Nanotechnol., 2018, 17, 1252–1258 CAS.
  9. K. S. Subrahmanyam, P. Kumar, U. Maitra, A. Govindaraj and K. P. S. S. Hembrarn, Proc. Natl. Acad. Sci. U. S. A., 2011, 108, 7 CrossRef.
  10. D. W. Boukhvalov, Phys. Rev. B: Condens. Matter Mater. Phys., 2008, 77, 035427 CrossRef.
  11. D. B. Putungan, S.-H. Lin, C.-M. Wei and J.-L. Kuo, Phys. Chem. Chem. Phys., 2015, 17, 11367–11374 RSC.
  12. H. Tachikawa and T. Iyama, J. Phys. Chem. C, 2019, 123, 8709–8716 CrossRef CAS.
  13. R. Muhammad, Y. Shuai and H.-P. Tan, Phys. E, 2017, 88, 115–124 CrossRef CAS.
  14. R. E. Ambrusi and M. E. Pronsato, Appl. Surf. Sci., 2019, 464, 243–254 CrossRef CAS.
  15. H. Lee, W. I. Choi, M. C. Nguyen, M. H. Cha, E. Moon and J. Ihm, Phys. Rev. B: Condens. Matter Mater. Phys., 2007, 76, 195110 CrossRef.
  16. V. Tozzini and V. Pellegrini, Phys. Chem. Chem. Phys., 2013, 15, 80–89 RSC.
  17. Y.-H. Kim, Y. Zhao, A. Williamson, M. J. Heben and S. B. Zhang, Phys. Rev. Lett., 2006, 96, 016102 CrossRef.
  18. I. Lopez-Corral, E. German, A. Juan, M. A. Volpe and G. P. Brizuela, J. Phys. Chem. C, 2011, 115, 4315–4323 CrossRef CAS.
  19. J. Cao, J. Zhou, Y. Zhang and X. Liu, Microelectron. Eng., 2018, 190, 63–67 CrossRef CAS.
  20. O. Faye, J. A. Szpunar, B. Szpunar and A. C. Beye, Appl. Surf. Sci., 2017, 392, 362–374 CrossRef CAS.
  21. H. G. Shiraz and O. Tavakoli, Renewable Sustainable Energy Rev., 2017, 74, 104–109 CrossRef CAS.
  22. H. Lee, J. Ihm, M. L. Cohen and S. G. Louie, Nano Lett., 2010, 10, 793–798 CrossRef CAS.
  23. R. Zhang and W. Chen, Biosens. Bioelectron., 2017, 89, 249–268 CrossRef CAS.
  24. C. Tan, Z. Luo, A. Chaturvedi, Y. Cai, Y. Du, Y. Gong, Y. Huang, Z. Lai, X. Zhang, L. Zheng, X. Qi, M. H. Goh, J. Wang, S. Han, X.-J. Wu, L. Gu, C. Kloc and H. Zhang, Adv. Mater., 2018, 30, 1705509 CrossRef.
  25. J. Cao, J. Zhou, Y. Zhang, Y. Wang and X. Liu, ACS Appl. Mater. Interfaces, 2018, 10, 1752–1760 CrossRef CAS.
  26. Q. Liu, X. L. Li, Q. He, A. Khalil, D. B. Liu, T. Xiang, X. J. Wu and L. Song, Small, 2015, 11, 556–5564 Search PubMed.
  27. K. C. Knirsch, N. C. Berner, H. C. Nerl, C. S. Cucinotta, Z. Gholamvand, N. McEvoy, Z. Wang, I. Abramovic, P. Vecera, M. Halik, S. Sanvito, G. S. Duesberg, V. Nicolosi, F. Hauke, A. Hirsch, J. N. Coleman and C. Backes, ACS Nano, 2015, 9, 6018–6030 CrossRef CAS.
  28. E. E. Benson, H. Zhang, S. A. Schuman, S. U. Nanayakkara, N. D. Bronstein, S. Ferrere, J. L. Blackburn and E. M. Miller, J. Am. Chem. Soc., 2018, 140, 441–450 CrossRef CAS.
  29. H. Li, S. Chen, X. Jia, B. Xu, H. Lin, H. Yang, L. Song and X. Wang, Nat. Commun., 2017, 8, 15377 CrossRef CAS.
  30. Q. Tang and D. Jiang, ACS Catal., 2016, 6, 4953–4961 CrossRef CAS.
  31. S. Chen, X. Chen, G. Wang, L. Liu, Q. He, X-bo Li and N. Cui, Chem. Mater., 2018, 30, 5404–5411 CrossRef CAS.
  32. Q. Liu, Q. Fang, W. Chu, Y. Wan, X. Li, W. Xu, M. Habib, S. Tao, Y. Zhou, D. Liu, T. Xiang, A. Khalil, X. Wu, M. Chhowalla, P. M. Ajayan and L. Song, Chem. Mater., 2017, 29, 4738–4744 CrossRef CAS.
  33. J. Ekspong, R. Sandstrom, L. P. Rajukumar, M. Terrones, T. Wagberg and E. Gracia-Espino, Adv. Funct. Mater., 2018, 28, 102744 CrossRef.
  34. L. Wang, X. Liu, J. Luo, X. Duan, J. Crittenden, C. Liu, S. Zhang, Y. Pei, Y. Zeng and X. Duan, Angew. Chem., Int. Ed., 2017, 56, 7610–7614 CrossRef CAS.
  35. J. Pan, R. Wang, X. Y. Zhou, J. S. Zhong, X. Y. Xu and J. G. Hu, Phys. Chem. Chem. Phys., 2017, 19, 24594 RSC.
  36. W. Zeng, T. Liu, D. J. Liu and E. J. Han, Sens. Actuators, B, 2011, 160, 455–462 CrossRef CAS.
  37. M. Pizzochero and O. V. Yazyev, Phys. Rev. B: Condens. Matter Mater. Phys., 2017, 96, 245402 CrossRef.
  38. D. Saha and S. Mahapatra, Phys. Chem. Chem. Phys., 2017, 19, 10453–10461 RSC.
  39. F. Ling, H. Jing, Y. Chen, W. Kang, W. Zeng, X. Liu, Y. Zhang, L. Fang and M. Zhou, J. Mater. Chem. C, 2018, 6, 12245–12251 RSC.
  40. D. Saha and S. Mahapatra, Phys. Chem. Chem. Phys., 2017, 19, 10453 RSC.
  41. G. Gao, Y. Jiao, F. Ma, Y. Jiao, E. Waclawik and A. Du, J. Phys. Chem. C, 2015, 119, 13124–13128 CrossRef CAS.

Footnote

Electronic supplementary information (ESI) available: Experimental details. See DOI: 10.1039/c9cp04402g

This journal is © the Owner Societies 2020