Bandgap recovery of monolayer MoS₂ using defect engineering and chemical doping

Abstract

Two-dimensional transition metal dichalcogenide materials have created avenues for exciting physics with unique electronic and photonic applications. Among these materials, molybdenum disulfide is the most known due to extensive research in understanding its electronic and optical properties. In this paper, we report on the successful growth and modification of monolayer MoS₂ (1L MoS₂) by controlling carrier concentration and manipulating bandgap in order to improve the efficiency of light emission. Atomic size MoS₂ vacancies were created using a Helium Ion Microscope, then the defect sites were doped with 2,3,5,6-tetrafluro7,7,8,8-tetracyanoquinodimethane (F4TCNQ). The carrier concentration in intrinsic (as-grown) and engineered 1L MoS₂ was calculated using Mass Action model. The results are in a good agreement with Raman and photoluminescence spectroscopy as well as Kelvin probe force microscopy characterizations.

---

1 Introduction

Recent advances in materials research have focused on novel, atomically thin semiconducting materials. Starting with gapless graphene,¹ a broad family of two-dimensional (2D) dichalcogenide materials have been fabricated in monolayer and few-layers form.^2–4 These materials possess unique properties enabling creation of a new generation of flexible and ultrathin optoelectronic devices. Among this family, direct bandgap semiconducting monolayer molybdenum disulfide (1L MoS₂) is the most popular one.^5,6 There are numerous reports of successful Chemical Vapor Deposition (CVD) synthesis processes. Unfortunately, CVD grown 2D MoS₂ always contains sulfur vacancies resulting in the creation of unsaturated chemical bonds at the molybdenum atoms. Sulfur vacancies represented by dangling bonds may serve as active sites for chemical adsorption of species. They will also result in quenching and broadening of MoS₂ photoluminescence (PL) by bringing additional defective states that are not available in the perfect system.^7–12 These vacancies influence drastically the carrier concentration and exciton generation in a sample. In a perfect system two exciton states are formed at the K and K′ point of MoS₂ Brillouin zone – exciton A (around 1.8 eV) and exciton B (around 2 eV). In the presence of sulfur vacancies, interaction of exciton and free carrier gives rise to the formation of a many-body bound state, a negatively charged exciton, also referred as a trion. It is already known that the PL of monolayer MoS₂ can be enhanced by reducing concentration of the excessive electrons and related shift of oscillator strength between trions and excitons.^13,14 Various techniques have been employed to modulate optical properties of 1L MoS₂ by controlling free carrier density, namely electrical gating, gas physisorption,^13,15 and chemical doping.^16–19

In this study, we first investigated intrinsic doping of as-grown monolayer MoS₂. We later modified monolayer MoS₂ using focused helium ion irradiation, consequently introducing monosulfur and disulfur vacancies predominately. The carrier concentration in as-grown and defected samples was linked with 2D MoS₂ optical response. Finally, we restored/enhanced PL of defected monolayer MoS₂ via chemical p-type doping.

2 Experimental procedures

2.1 Sample preparation

High quality 2D MoS₂ samples were grown at 650 °C on Si/SiO₂ substrates using a home-built CVD setup with a one-inch quartz tube fitted in Lindberg furnace equipment. Fig. 1a–d shows that the morphology and shape of 2D MoS₂ varies depending on the stoichiometric ratio of Molybdenum to sulfur : hexagon shape flakes grow when the ratio is 1 : 1, an uneven ratio will give rise to triangles. The SEM and Raman analysis confirm that the majority of MoS₂ flakes are monolayers, however, some flakes, especially hexagons, have small islands of second- and third-layer.

Fig. 1 (a and b) Helium Ion Microscopy of as-grown MoS₂; (c and d) SEM images of MoS₂ transferred to conductive substrate for CAFM characterization; Raman map of 1LMoS₂: (f) E_2g at 383 cm⁻¹, and (e) A_1g at 402 cm⁻¹. The scale bars in (a, d, e and f) are 5 μm, in (b) 50 μm, in (c) 500 nm.

MoS₂ samples were irradiated in a Zeiss Helium Ion Microscope operating at accelerating voltage of 30 kV with doses of (1) 1 × 10¹³, (2) 5 × 10¹⁴, (3) 9 × 10¹⁴, (4) 1 × 10¹⁵, (5) 5 × 10¹⁵, (6) 9 × 10¹⁵, and (7) 1 × 10¹⁶ He⁺ ions per cm². The exposure was performed using patterning software to raster the focused helium ion beam over a large area (up to 50 μm). A beam current of 5 pA was used for all exposures with dosage being controlled by exposure time. This resulted in the introduction of defects with inter-defect distances varying from 10.3 nm to 3.7 nm as we reported elsewhere.²⁰ Prior to doping, MoS₂ samples were placed in the middle of a CVD tube furnace and annealed at 250 °C for 1 hour. The samples were then soaked in a solution of 0.3 μmol ml⁻¹ of 2,3,5,6-tetrafluro7,7,8,8-tetracyanoquinodimethane (F4TCNQ) in Chloroform for 12 hours, then rinsed to remove unbounded F4TCNQ as suggested in.²¹ Thereafter, the samples were heated on a hot plate at 100 °C for 30 minutes.

2.2 Characterization techniques

Photoluminescence (PL) and Raman spectra were recorded with a Horiba XploRa Confocal Raman Microscope at 532 nm of excitation. The laser power was kept below 0.5 mW to avoid thermal damage due to heating. Scanning Probe Microscopy was performed on an Oxford Asylum Research Atomic Force Microscope (AFM) MFP-3D Infinity.

Surface potential imaging was performed in amplitude-modulated mode with conductive Si tips coated with Pt/Ir (EFM-20, NanoWorld). The Asylum's ORCA™ module was used for conductive mode (CAFM). For this characterization, we transfered MoS₂ samples to a Si substrate coated with a 50 nm gold film using the wet KOH transfer method.²² The I–V curves were recorded before and after modification of the MoS₂ samples, with the CAFM tip (CDT-NCHR-SPL) being engaged very gently with only a few nN of force being applied until the current signal stabilized. All above measurements were performed at ambient conditions. The X-ray Photoelectron Spectroscopy (XPS) data were obtained on a Thermo Fisher ESCALAB 250 Xi. The atomically resolved images of the MoS₂ sample were recorded using the Scanning Transmission Electron Microscope Nion Ultra HAADF-STEM 100 at Oak Ridge National Laboratory.

3 Results and discussions

By adjusting proper synthesis conditions, we grew 2D MoS₂ flakes with the area up to 270 μm² as shown on Fig. 1a–d. Scanning Electron and Helium Ion Microscopy allow for visualization of shape, lateral dimensions, and homogeneity of grown MoS₂ samples. However, these techniques do not provide quantitative information for the number of MoS₂ layers in a single flake. Confocal Raman spectroscopy is the most popular tool for reliable identification the MoS₂ layer number, as well as for doping and strain. The two prominent Raman peaks located around 402 cm⁻¹ and 383 cm⁻¹ assigned to the out-of-plane (A_1g) and in-plane (E_2g) phonon vibration modes respectively. The band separation Δf = 19 − 20 cm⁻¹ corresponds to 1L MoS₂. With the increasing number of layers, the difference Δf will also increase, for instance, Δf = 21 − 22 cm⁻¹ corresponds to bilayer of MoS₂.

Our typical results of Raman mapping show spatial homogeneity of E_2g (Fig. 1f) and A_1g (Fig. 1e) over the MoS₂ flake area: the intensity and variation for both peaks were negligible, so we have identified as grown flakes as pristine monolayers. The pristine 1L MoS₂ exhibits a strong PL peak centered at 1.83 eV. However, the efficiency and broadening of MoS₂ emission is far from the theoretical limits of the direct band gap semiconductor material. It is already known that p-doping can positively influence the optical properties of 1L MoS₂.^23–25 To prove this, we started with conducting ambient surface potential measurements using two-pass AM-KPFM on the as-grown sample. The work function value Φ_sample and, consequently, Fermi energy shift were calculated by equation:

Φ_sample = e × V_CPD − Φ_probe (1)

where V_CPD is the charge potential difference between the sample (MoS₂) and the AFM probe, e is elemental charge, and Φ_probe is the work function of the KPFM probe. The probe was calibrated with highly oriented pyrolytic graphite (HOPG) having a work function of 4.65 eV. The spatial distribution of Φ_1LMoS2 for the as-grown sample is presented on Fig. 2a. It should be noted that for monolayer MoS₂ (no doping) the theoretical value of Φ_1LMoS2 is equal to 5.1 eV ²⁶ and Fermi energy is in the middle of bandgap. The measured value for as-grown 1L MoS₂ showed shift of Fermi energy towards conduction band by 450 meV (Fig. 2c) indicating intrinsic n-doping of as-grown sample. These results are in agreement with previous reports on CVD grown 2D MoS₂ on Si/SiO₂, and are attributed mainly to the existence of natural defects such as sulfur vacancies.^27,28 As expected, a small increase in the local work function was found after incubation of as-grown sample with p-dopant F4TCNQ, such that the average Fermi energy shifted towards the center of bandgap by only 200 meV (Fig. 2b, c). According to DFT calculations,²⁹ the F4TCNQ molecule has a relatively weak adsorption on pristine MoS₂ surface. Therefore, we attribute this shift of Fermi energy to the interaction of F4TCNQ with the MoS₂ defect sites as further corroborated by TEM images in ESI (Fig. S5).† To have better control over optoelectronic properties of MoS₂, we propose here to combine doping and defect engineering. Our earlier work has demonstrated the ability to introduce defects by irradiating MoS₂ samples using Helium Ion Microscope,²⁰ although tuning of MoS₂ optical response had not been investigated earlier.

Fig. 2 The MoS₂ work function map acquired (a) before and (b) after p-doping with F4TCNQ; simplified sketch of MoS₂ electronic structure before doping (c) and after doping (d).

Multiple doses of Helium ion irradiation (from 1 × 10¹³ to 1 × 10¹⁶ He⁺ ions per cm²) were used to modify 1L MoS₂ electronic structure by generating sulfur vacancies. An AFM image of the irradiated sample is presented in Fig. 3c. The high-resolution HAADF-STEM image of the sample irradiated with the dose of 1 × 10¹⁴ He⁺ ions per cm² shows details of alternating symmetry of atoms arranged in hexagonal rings (Fig. 3a red arrows). Over this area, four defect sites are seen that would be available for adsorption and doping. There is also a MoS₂ grain boundary that is stitching together three parts of the 1L MoS₂ flake. Notably, there are no sulfur vacancies associated with broken symmetry near the grain boundaries.

Fig. 3 The 1L MoS₂ irradiated at 1 × 10¹⁴ He⁺ ions per cm² dose. (a) HAADF-STEM image; scale bar is 4 nm; the red arrows indicate single and double sulfur terminations. (b) Fourier transform of STEM image. (c) AFM topography; scale bar is 4 μm.

Despite of the fact, that the Fourier Transform of the STEM image (Fig. 3b) confirms preserved of the Brillouin zones hexagonal geometry, PL intensity of this sample decreased by 30%, and for the next dose it dropped down further by 60% when compared to pristine 1L MoS₂ sample (Fig. 4a black-red-blue curves). To calculate the carrier concentration in the as-grown, defected, and defected-doped sample we applied the Mass Action model as introduced by J. Ross et al.³⁰ for a three level system.¹⁴ According to this model, the relations between concentrations of excitons (N_X), trions (N_X⁻) and the excess electrons (n_e) in transition metal dichalcogenides can be written as:

where, T, k_B, E_b are the temperature, the Boltzmann constant, and the trion binding energy (about 20 meV (ref. 31)) respectively. The effective mass of electrons is m_e = 0.35 m₀ and the effective mass of holes is m_h = 0.45 m₀. The m₀ is the free electron mass. The effective masses of an exciton and trion can be calculated as m_X = m_e + m_h = 0.8 m₀, and m_X⁻ = 2 m_e + m_h = 1.15 m₀. Then, the trion PL intensity weight is related to the excess electron concentration n_e:here β is a constant, equal to 4 × 10⁻¹⁴ cm². Using this model, we can calculate the electron concentration before and after doping:

Fig. 4 (a) MoS₂ PL degradation with an increase of ion irradiation dose; and (c) Raman spectra before and after F4TCNQ doping (no ion irradiation) (b and d) PL and Raman spectra defected and doped MoS₂; (e) PL Intensity of pristine (*), defected (**), and defected and doped (***) MoS₂. Doses 1 × 10¹³ and 1 × 10¹⁴ provide perfect bandgap recovery and PL enhancement.

We found that the trion PL weight in irradiated samples increases from 0.38 to 0.46 and correspondingly electron concentration of doped 1L MoS₂ increases from 1.56 × 10¹³ 1 cm⁻² to 2.2 × 10¹² 1 cm⁻² (Table 1). Indeed, 1 × 10¹⁴ dose causes negligible damage. In contrast, 1 × 10¹⁶ dose disrupts MoS₂ lattice symmetry completely.

Table 1 Calculations of carrier concentration for untreated and helium ion treated 1L MoS₂

Applied doses (ions per cm^2)	Trion weight	Electr. conc. (cm^−2)
0	0.38	1.56 × 10^13
1 × 10^13	0.383	1.53 × 10^13
1 × 10^14	0.390	1.62 × 10^13
5 × 10^14	0.410	1.72 × 10^13
1 × 10^15	0.460	2.2 × 10^13

---

A simple and effective chemical p-doping method was used to tune the carrier concentration in 1L MoS₂. The F4TCNQ molecule can easily bind to the sulfur vacancies and act as an acceptor. As we discussed above, Raman spectroscopy could be used to trace the doping effect in 1L MoS₂. In the case of p-doping, frequency of E_2g peak should remain constant, but the frequency of A_1g peak is blue-shifted.^32,33 Raman characterization of the F4TCNQ doped as-grown sample (untreated with ion beam) shows that the A_1g peak position shifts by 1.2 cm⁻¹ while E_2g peak position remains the same (Fig. 4c). However, after irradiation, E_2g/A_1g modes are both red/blue shifted, and widened due to the appearance of new phonon modes. There is a shoulder (at 362 cm⁻¹) to the left of E_2g mode, and one (at 415 cm⁻¹) to the right of A_1g mode, which are assigned to a defect mode.^33–35 Both peaks are significantly enhanced after irradiation with helium ions.

After p-doping of defected samples, we observe a gigantic enhancement of MoS₂ PL for the dose of 1 × 10¹³ and 1 × 10¹⁴ (compare green and red curves in Fig. 4a and b), and PL recovery for the dose 5 × 10¹⁴ (blue curves in Fig. 4a and b). There is no PL recovery for the doses 1 × 10¹⁵ and higher. We attribute it to substantial damage in the MoS₂ lattice by the ion bombardment. The spectral shape of the PL peak is also changing with dosage. It slightly shifts towards visible range with increase of dosage (from 1.83 eV to 1.81 eV) due to decrease of the trion weight. It is slightly lower than reported for p-doped pristine 1L MoS₂ previously.¹⁴

We apply the Mass Action model to calculate carrier concentration after defect engineering and doping. Excess electron concentration decreased in comparison to the pristine sample from 1.56 to 1.13 × 10¹³ cm⁻² (Table 2). Fig. 4e presents data for pristine, defect, and defect-doped samples. One can see that PL efficiency is increased for the 1 × 10¹³ sample by 230% and for the 1 × 10¹⁴ sample by 90%. We attribute it to a sufficient reduction of excess electron concentration and as a result, increase of exciton oscillator strength.

Table 2 Calculations of carrier concentration for defect-doped samples 1L MoS₂

Applied doses (ions per cm^2)	Trion weight	Electr. conc. (cm^−2)
1 × 10^13	0.370	1.40 × 10^13
1 × 10^14	0.306	1.13 × 10^13
5 × 10^14	0.310	1.12 × 10^13

---

For I–V characterization, 1L MoS₂ was transferred on the conductive substrate (Fig. 5a). The XPS characterization confirmed elemental composition Mo : S – 1 : 2 (Fig. 5b). The bias applied to the sample caused current flow into the conductive AFM tip, which was then recorded by a current amplifier. Fig. 4c shows the I–V graphs of as-grown, irradiated, and irradiated-doped 1L MoS₂ measured by CAFM. The CAFM voltage was in the range of −1 V to 1 V. All samples: as-grown, doped, and defected and doped exhibited a clear non-linear slope of the I–V curve, which is characteristic of a metal/semiconductor junction. A clear difference is observed between the I–V curves recorded on a pristine, irradiated, irradiated-doped samples. Particularly, in the forward bias regime, higher current response is measured at the same voltage (−0.65 V) for the irradiated compared to the as-grown and irradiated-doped MoS₂. After irradiation at ion dose of 1 × 10¹⁴ and doping, defected-doped 1L MoS₂ showed a characteristic I–V curve of an ideal Schottky diode compared to the as-grown sample and irradiated MoS₂ sample. The results of transport measurements are consistent with the Raman and Photoluminescence characterization confirming that the F4TCNQ molecules in solution can strongly withdraw electrons from MoS₂ layers.

Fig. 5 (a) AFM topography of transferred MoS₂ on conductive substrate; (b) XPS core-level spectra of molybdenum 3d and sulfur 2 s; (c) I–V characterization for pristine and doped, defected and doped MoS₂.

4 Conclusions

In conclusion, we performed a systematic modification of the electronic structure of 1L MoS₂ by (a) gentle defect introduction using helium ion microscope with doses ranging from 1 × 10¹³ to 1 × 10¹⁶ ion per cm², and (b) subsequent p-doping of defected MoS₂ sites. Reduced PL intensity was restored and drastically enhanced due to the adsorption of p-type dopants (F4TCNQ molecules) by 1L MoS₂. We explain the MoS₂ bandgap recovery is due to decrease of excess electron concentration resulting in an increase of exciton oscillator strength. Our findings can be useful for improvement in the efficiency of light emission devices based on two-dimensional materials beyond MoS₂.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

Part of this work was performed at the Joint School of Nanoscience and Nanoengineering (JSNN), a member of the Southeastern Nanotechnology Infrastructure Corridor (SENIC) and National Nanotechnology Coordinated Infrastructure (NNCI), which is supported by the National Science Foundation (Grant ECCS-1542174). Some MoS₂ samples for this publication was provided by The Pennsylvania State University Two-Dimensional Crystal Consortium – Materials Innovation Platform (2DCC-MIP), which is supported by NSF cooperative agreement DMR-1539916. Scanning Transmission Electron Microscope (STEM) imaging was conducted at the Center for Nanophase Materials Sciences, which is a DOE Office of Science User Facility.

References

1. A. K. Geim and K. S. Novoselov, Nat. Mater., 2007, 6, 183–191.
2. W. Choi, N. Choudhary, G. H. Han, J. Park, D. Akinwande and Y. H. Lee, Mater. Today, 2017, 20, 116–130.
3. S. Manzeli, D. Ovchinnikov, D. Pasquier, O. V. Yazyev and A. Kis, Nat. Rev. Mater., 2017, 2, 17033.
4. B. M. Bersch, S. M. Eichfeld, Y.-C. Lin, K. Zhang, G. R. Bhimanapati, A. F. Piasecki, M. Labella and J. A. Robinson, 2D Materials, 2017, 4, 025083.
5. K. F. Mak, C. Lee, J. Hone, J. Shan and T. F. Heinz, Phys. Rev. Lett., 2010, 105, 136805.
6. D. Jariwala, V. K. Sangwan, L. J. Lauhon, T. J. Marks and M. C. Hersam, ACS Nano, 2014, 8, 1102–1120.
7. P. K. Chow, R. B. Jacobs-Gedrim, J. Gao, T.-M. Lu, B. Yu, H. Terrones and N. Koratkar, ACS Nano, 2015, 9, 1520–1527.
8. J. Li, S. Hu, S. Wang, H. Kang, Z. Chen, S. Zhao, Y. Zhang, Y. Sui and G. Yu, Mater. Lett., 2021, 290, 129421.
9. M. Erementchouk, M. A. Khan and M. N. Leuenberger, Phys. Rev. B, 2015, 92, 121401.
10. H. M. Oh, G. H. Han, H. Kim, J. J. Bae, M. S. Jeong and Y. H. Lee, ACS Nano, 2016, 10, 5230–5236.
11. B. Chen, H. Sahin, A. Suslu, L. Ding, M. I. Bertoni, F. M. Peeters and S. Tongay, ACS Nano, 2015, 9, 5326–5332.
12. T. Völzer, F. Fennel, T. Korn and S. Lochbrunner, Phys. Rev. B, 2021, 103, 045423.
13. S. Tongay, J. Zhou, C. Ataca, J. Liu, J. S. Kang, T. S. Matthews, L. You, J. Li, J. C. Grossman and J. Wu, Nano Lett., 2013, 13, 2831–2836.
14. S. Mouri, Y. Miyauchi and K. Matsuda, Nano Lett., 2013, 13, 5944–5948.
15. T. Granzier-Nakajima, K. Fujisawa, V. Anil, M. Terrones and Y.-T. Yeh, Nanomaterials, 2019, 9, 425.
16. D. Pierucci, H. Henck, Z. Ben Aziza, C. H. Naylor, A. Balan, J. E. Rault, M. G. Silly, Y. J. Dappe, F. Bertran, P. Le Fèvre, F. Sirotti, A. T. C. Johnson and A. Ouerghi, ACS Nano, 2017, 11, 1755–1761.
17. A. Molle, F. Fabbri, D. Campi, A. Lamperti, E. Rotunno, E. Cinquanta, L. Lazzarini, D. Kaplan, V. Swaminathan, M. Bernasconi, M. Longo and G. Salviati, Adv. Electron. Mater., 2016, 2, 1600091.
18. Y. Cai, H. Zhou, G. Zhang and Y.-W. Zhang, Chem. Mater., 2016, 28, 8611–8621.
19. M. Yarali, Y. Zhong, S. N. Reed, J. Wang, K. A. Ulman, D. J. Charboneau, J. B. Curley, D. J. Hynek, J. V. Pondick, S. Yazdani, N. Hazari, S. Y. Quek, H. Wang and J. J. Cha, Adv. Electron. Mater., 2021, 7, 2000873.
20. F. Aryeetey, T. Ignatova and S. Aravamudhan, RSC Adv., 2020, 10, 22996–23001.
21. D. M. Sim, M. Kim, S. Yim, M.-J. Choi, J. Choi, S. Yoo and Y. S. Jung, ACS Nano, 2015, 9, 12115–12123.
22. H. Van Ngoc, Y. Qian, S. K. Han and D. J. Kang, Sci. Rep., 2016, 6, 33096.
23. K. H. Kim, K. S. Kim, Y. J. Ji, I. Moon, K. Heo, D.-H. Kang, K. N. Kim, W. J. Yoo, J.-H. Park and G. Y. Yeom, J. Mater. Chem. C, 2020, 8, 1846–1851.
24. K. Cho, M. Min, T.-Y. Kim, H. Jeong, J. Pak, J.-K. Kim, J. Jang, S. J. Yun, Y. H. Lee, W.-K. Hong and T. Lee, ACS Nano, 2015, 9, 8044–8053.
25. J. Yang, H. Kawai, C. P. Y. Wong and K. E. J. Goh, J. Mater. Chem. C, 2019, 123, 2933–2939.
26. L. Britnell, R. M. Ribeiro, A. Eckmann, R. Jalil, B. D. Belle, A. Mishchenko, Y.-J. Kim, R. V. Gorbachev, T. Georgiou, S. V. Morozov, A. N. Grigorenko, A. K. Geim, C. Casiraghi, A. H. C. Neto and K. S. Novoselov, Science, 2013, 340, 1311–1314.
27. I. Sharma and B. R. Mehta, J. Alloys Compd., 2017, 723, 50–57.
28. W. H. Chae, J. D. Cain, E. D. Hanson, A. A. Murthy and V. P. Dravid, Appl. Phys. Lett., 2017, 111, 143106.
29. O. K. Le, V. Chihaia, V. Van On and D. N. Son, RSC Adv., 2021, 11, 8033–8041.
30. J. S. Ross, S. Wu, H. Yu, N. J. Ghimire, A. M. Jones, G. Aivazian, J. Yan, D. G. Mandrus, D. Xiao, W. Yao and X. Xu, Nat. Commun., 2013, 4, 1474.
31. K. F. Mak, K. He, C. Lee, G. H. Lee, J. Hone, T. F. Heinz and J. Shan, Nat. Mater., 2013, 12, 207–211.
32. B. Chakraborty, A. Bera, D. V. S. Muthu, S. Bhowmick, U. V. Waghmare and A. K. Sood, Phys. Rev. B, 2012, 85, 161403.
33. H. Li, Q. Zhang, C. C. R. Yap, B. K. Tay, T. H. T. Edwin, A. Olivier and D. Baillargeat, Adv. Funct. Mater., 2012, 22, 1385–1390.
34. S. Mignuzzi, A. J. Pollard, N. Bonini, B. Brennan, I. S. Gilmore, M. A. Pimenta, D. Richards and D. Roy, Phys. Rev. B, 2015, 91, 195411.
35. Y. Chen, S. Huang, X. Ji, K. Adepalli, K. Yin, X. Ling, X. Wang, J. Xue, M. Dresselhaus, J. Kong and B. Yildiz, ACS Nano, 2018, 12, 2569–2579.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/d1ra02888j

---