2019 167 Moesm1 Esm

Download as pdf or txt
Download as pdf or txt
You are on page 1of 8

Supplementary Information

Unveiling the Origin for the Ultrafast Growth of Monolayer WSe2

via Chemical Vapor Deposition

Shuai Chen1*, Junfeng Gao1,2,*, Bharathi M Srinivasan1, Gang Zhang1, Viacheslav Sorkin1,

Ramanarayan Hariharaputran1, Yong-Wei Zhang1,†


1
Institute of High Performance Computing, A*STAR, 138632 Singapore
2
Key Laboratory of Materials Modification by Laser, Ion and Electron Beams, Dalian University

of Technology, Ministry of Education, Dalian 116024, People's Republic of China

SUPPLEMENTARY METHODS

KMC Algorithms

At each given state i, kMC model specifies all the possible states j that it may transit into at the

next step based on these defined events. The transition rate is calculated by the transition state

theory (TST)1:
𝐸
𝑟𝑖→𝑗 = 𝜈 exp(− 𝑘𝑖→𝑗 ) (S1)
𝑇 𝑏

𝜈 is the order of the frequency of atomic vibration (1012-1013 s-1)2. In our simulations, 𝜈 is

selected as 1012 s-1, 𝐸𝑖→𝑗 is the energy barrier for the transition from current state i to another

*
Both authors contributed equally.

Corresponding author. Email: [email protected]
1
state j, 𝑘𝑏 is the Boltzmann constant, and 𝑇 is the temperature. The sum of all the rates

corresponds to the total rate of leaving the current state, that is:

𝑅𝑖 = ∑𝑛𝑗=1 𝑟𝑖→𝑗 (S2)

To determine which state j is to be selected for the transition from state i, a random number

U1 in the range (0, 1) is generated. The state j that satisfies the following condition is chosen.

∑𝑗−1 𝑗
𝑘=1 𝑟𝑖→𝑘 < 𝑈1 × 𝑅𝑖 < ∑𝑘=1 𝑟𝑖→𝑘 (S3)

Meanwhile, the time of this transition can be calculated by

1
△ 𝑡 = − (𝑅 ) ln 𝑈2 (S4)
𝑖

𝑈2 is another random number in the range (0, 1).

Events and Energy Barriers in the kMC Model

In the simulations, the key events for growth kinetics and their corresponding energy barriers are

defined and shown in Figure S1, such as the surface diffusion of W/Se (Event 1/2), attachment of

W/Se to the edge (Event 3/4), attachment of W/Se to form a kink (Event 5/6), edge diffusion of

W/Se (Event 7/8), deposition of W/Se to the substrate surface (Event 9/10), deposition of W/Se

to a zigzag edge (Event 11/12), and deposition of W/Se to form a kink (Event 13/14).

The energy barriers for the events are set based on Gao et al.’s DFT calculations3. Gao et al.

reported that the highest energy barrier for the diffusion of Se adatom on Au (111) surface was

calculated to be 0.36 eV. Therefore, the energy barriers for surface diffusion (Event 1/2) are 0.36

eV, including the forward and reverse processes for each event. Gao et al. also calculated the

attachment of a W2 dimer to the zigzag edge and found that the two energy barriers were 0.55

and 0.54 eV for each W atom, and the energy barrier for the detachment was 2.5 eV. Based on

2
this calculation, the energy barriers for attachment to the zigzag edge (Event 3/4) are 0.55 eV for

the forward attachment and 2.5 eV for the reverse detachment. Meanwhile, Gao et al. found that

the attachment of a Se atom to form a kink was an exothermic reaction with a small energy

barrier of 0.32 eV. In this case, the energy barriers for attachment to form a kink (Event 5/6) are

0.32 eV for the forward attachment and 2.5 eV for the reverse detachment.

Supplementary Figure 1. Events and energetics in the kMC simulation model. Light blue and yellow

colored bonds represent the lattices. Purple and orange spheres represent the W and Se, respectively.

We also carried out first-principles calculations on the events that are critical to our proposed

ultrafast growth mechanism, that is, the attachment to zigzag edge contributing to the kink

3
nucleation and the surface diffusion determining the kink propagation. The minimum energy

path for the attachment of a W dimer to the zigzag edge of a WSe2 cluster on Au (111) surface is

plotted in Figure S2a, and the schematic structures corresponding to the different states are

shown in Figure S2c. The results indicate that the energy barrier is 0.56 eV, which is quite close

to Gao et al.’s10 calculated value (0.55 eV). This low energy barrier for edge attachment results

in the fast kink nucleation. The minimum energy path for the diffusion of a Se adatom on Au

(111) surface is also plotted in Figure S2b, and the schematic structures corresponding to the

different states are shown in Figure S2d. The results demonstrate that the energy barrier is 0.25

eV, which also matches with Gao et al.’s10 calculation (Gao et al.10 reported that the highest

energy barrier for surface diffusion is 0.36 eV). This lower energy barrier (in comparison with

the edge attachment) causes the ultrafast surface diffusion, which is able to assist the attachment

of adatoms to diffuse to the most energetically favorable sites, leading to the ultrafast kink

propagation.

4
Supplementary Figure 2. First-principles calculations on the minimum energy path for (a) the

attachment of a W dimer to the zigzag edge of a WSe2 cluster on Au (111) surface and (b) the diffusion of

a Se adatom on Au (111) surface. (c, d) The schematic structures corresponding to the different states in

(a, b), respectively. IS: initial state; LS: local state; TS: transition state; and FS: final state. Purple, orange

and grey spheres represent W, Se and Au atoms, respectively.

To the authors’ knowledge, there is no calculation for the energy barriers of the edge diffusion

of W/Se on Au(111) substrate (Event 7/8). In our study, initially, it is chosen to be 0.5 eV, which

is able to produce a compact triangular domain shape. We then increase it to 1.0 eV and 1.5 eV

to study the growth from the thermodynamic regime to the kinetic regime. To obtain accurate

energy barrier for edge diffusion, we perform our DFT calculations (see S1.3).

The surface depositions of W atom and Se dimer (Event 9/10) are controlled by the deposition

fluxes of FW and FS. In our kMC simulations, the values of FW and FS vary from 3×10-5 to 2×10-2

ML/s, which are comparable with the experimental growth rate. In Section 3.1, only surface

deposition is considered to study the growth from the thermodynamic regime to the kinetic

regime. In Section 3.2, the edge deposition from the vapor is also included and the edge

deposition is energetically more favorable than substrate surface deposition. The differences in

their deposition fluxes are determined by the energy barriers for attachments. The energy barriers

for attachment to a zigzag edge (Event 3/4) are 0.55 eV. Therefore, the deposition fluxes for

Event 9/10 and Event 11/12 are F and F×exp(0.55/kbT). The energy barriers for attachment to

form a kink are (Event 5/6) 0.32 eV. Therefore, the deposition fluxes for Event 11/12 and Event

13/14 are F×exp(0.55/kbT) and F×exp((0.55+(0.55-0.32)/kbT)= F×exp(0.78/kbT).

First-Principles Calculations of Energy Barriers for Edge Diffusion

5
The energies of W atom with different in-plane (cf. Figure S3a) and out-of-plane (cf. Figure S3b)

distributions at the domain edge are calculated (cf. Figure S3c). Figure S3a shows 6 different in-

plane distributions of W atom resulting from the lattice mismatch between WSe2 and Au. Figure

S3b shows 2 different out-of-plane distributions of W. Figure S3c plots the energies of W atoms

with type I and type II out-of-plane distributions located at the 6 different in-plane positions. It

shows that W atoms of type I are the same as or more stable than those of type II. Furthermore, it

indicates that the diffusion of W atom from site 1 to site 2 or site 5 to 6 should be easier than that

from site 2 to site 3, site 3 to site 4, and site 4 to site 5 because of the lower energies of site 1 and

site 6. Therefore, the diffusions of W atom from site 2 to site 3, site 3 to site 4 and site 4 to site 5

with type I out-of-plane distribution are further calculated to obtain the largest energy barriers for

the edge diffusion (cf. Figure 2f). The results show that the energy barriers for the edge diffusion

of W along Se terminated zigzag edge are ~1.4 eV. First-principles calculations on the energy

barrier for Se atom diffusion along the W-terminated zigzag edge were also performed. We

found that the W-terminated zigzag edge underwent reconstruction (cf. Figure S3d), exhibiting

substantial inward displacements of all the first-row W atoms and outward displacements of half

of the second-row Se atoms. The minimum energy path for the diffusion of a Se atom along the

reconstructed W-terminated zigzag edge is plotted in Figure S3e, and the schematic structures

corresponding to the different states are shown in Figure S3f. Our calculations show that the

forward (from IS to FS in Figure S3f) diffusion barrier is 1.08 eV, while the backward one (from

FS to IS in Figure S3f) is 2.94 eV.

6
Supplementary Figure 3. (a) In-plane and (b) out-of-plane distributions of W at the Se zigzag edge. (c)

Energies of W at the Se zigzag edge with W different distributions. (d) In-plane and out-of-plane

structures of W-terminated zigzag edge. (e) The minimum energy path for the diffusion of a Se atom

along the reconstructed W-terminated zigzag edge. (f) The schematic structures corresponding to the

different states. IS: initial state; TS: transition state; LS: local state; and FS: final state. Purple, orange and

grey spheres represent W, Se and Au atoms, respectively.

SUPPLEMENTARY MOVIES

Supplementary Movie 1 showing the evolution of WSe2 domain with edge diffusion barrier of

0.5 eV.

Supplementary Movie 2 showing the evolution of WSe2 domain with edge diffusion barrier of

1.0 eV.

Supplementary Movie 3 showing the evolution of WSe2 domain with edge diffusion barrier of

1.5 eV.

7
Supplementary Movie 4 showing the evolution of WSe2 domain with the same W and Se fluxes.

Supplementary Movie 5 showing the evolution of WSe2 domain with W flux being smaller than

Se flux.

Supplementary Movie 6 showing the evolution of WSe2 domain with W flux being higher than

Se flux.

SUPPLEMENTARY REFERENCES

[1] Nitzan, A. Chemical Dynamics in Condensed Phases: Relaxation, Transfer and Reactions in

Condensed Molecular Systems. Oxford university press (2006).

[2] Voter, A. F. Introduction to the Kinetic Monte Carlo Method. Radiation Effects in Solids 235,

1–23 (2007).

[3] Gao, Y.; Hong, Y. L.; Yin, L. C.; Wu, Z.; Yang, Z. et al. Ultrafast Growth of High-Quality

Monolayer WSe2 on Au. Adv. Mater. 29, 1700990 (2017).

You might also like