Interfacial transport modulation by intrinsic potential difference of janus TMDs based on CsPbI3/J-TMDs heterojunctions.

Although perovskite/two-dimensional (2D) materials heterojunctions have been employed to improve the optoelectronic performance of perovskite photodetectors and solar cells, effects of the intrinsic potential difference (ΔV in ) of asymmetrical 2D materials, like Janus TMDs (J-TMDs), were not revealed yet. Herein, by investigating the optoelectronic properties of CsPbI3/J-TMDs heterojunctions, we find a reversible type-II band alignment related to the intensity and direction of ΔV in , suggesting that carrier transport paths can be reversed by modulating the contact configuration of J-TMDs in the heterojunctions. Meanwhile, the band offset, carrier transfer efficiency and optical properties of those heterojunctions are directly determined by the intensity and direction of ΔV in . Overall, CsPbI3/MoSSe heterojunction is suggested in this work with a tunneling probability of 79.65%. Our work unveils the role of ΔV in in asymmetrical 2D materials on the optoelectronic performances of lead halide perovskite devices, and provides a guideline to design high performance perovskite optoelectronic devices.

Introduction

In recent years, lead halide perovskites have attracted great attention in the application of high-performance photovoltaic and optoelectronic because of their large optical absorption coefficient, low cost, simple preparation, etc. (Gu et al., 2020; Guzelturk et al., 2021; Tao et al., 2021, 2020; Wang et al., 2021a; Wang et al., 2020) Photodetectors based on lead halide perovskites have been widely applied in diverse fields including optical communication, remote sensing, military surveillance, imaging technology, and industry automation control. However, it should be noted that the organic-inorganic lead halide perovskite possesses the poor stability which hinders the development of perovskite devices (Li et al., 2019; Zhou and Zhao, 2019). Alternatively, all-inorganic cesium lead halide perovskite (CsPbX3) not only shows a more effective stability (Jiang et al., 2018; Seth et al., 2019), but also demonstrates excellent optoelectronic properties, such as direct optical band gaps, large optical absorption, and high luminescence efficiency (Nedelcu et al., 2015; Protesescu et al., 2015). These optoelectronic properties make CsPbX3, especially CsPbI3, promising for solar cells, light-emitting diodes, lasers, and photodetectors (Liu et al., 2021; Zhang et al., 2015).

To further improve the optoelectronic performance of perovskite devices, heterojunction engineering, especially perovskite/2D material heterojunction has been becoming an effective approach (Zhang et al., 2021). For instance, CH3NH3PbI3/graphene heterojunction not only broadened the spectral photoresponsivity of perovskite photodetector, but also increased the effective quantum efficiency to 5 × 104% under an illumination power of 1 μW (Lee et al., 2015). Moreover, CsPbBr3/MoS2 heterojunction could increase the photoresponsivity and photodetectivity to 4.4 A/W and 2.5×1010 Jones, respectively, because of the high efficient photoexcited carrier separation at the interface between MoS2 and CsPbBr3 (Song et al., 2018). CsPbI3/phosphorus heterojunctions enhanced the light absorptions especially in the infrared region (Liu et al., 2018).

Compared to the symmetrical 2D materials, 2D asymmetrical polar materials Janus TMDs (J-TMDs, namely, MXY, M = Mo, W; X = S; Y=S, SE, Te in this work), which can be prepared by selenization or tellurization of symmetrical MoS2 in experiment (Lu et al., 2017; Yun et al., 2017), not only possess excellent optical absorption, tunable band gap, high carrier mobility, but also have the intrinsic potential difference (ΔV) of J-TMDs by destroying the out-of-plane mirror symmetry which can be regarded as another degree of freedom for modulating their properties (Yin et al., 2021). Moreover, such intrinsic ΔV has been found to promote the electron-hole separation, band levels variation, and mobility modulation, resulting in effective inhibition of carrier recombination and enhancement of responsivity (Dong et al., 2017; Er et al., 2018; Li et al., 2018; Lian et al., 2019; Thanh et al., 2020; Zhang et al., 2020; Zhou et al., 2019). Inspired by these, it is thought that the optoelectronic performances of perovskite devices based on perovskite/TMDs heterojunctions may be further improved by the intrinsic ΔV when the TMDs are replaced by the J-TMDs.

Herein, CsPbI3/J-TMDs (MXY, M = Mo, W; X = S; Y=S, SE, Te) heterojunctions are studied to unveil the roles of ΔV in J-TMDs on the optoelectronic properties of CsPbI3 based device. The results demonstrate that all CsPbI3/J-TMDs heterojunctions in this work have a type-II band alignment, indicating that charge transfer rather than energy transfer dominates in CsPbI3/J-TMDs heterojunctions. Moreover, a reversed type-II band alignment happened when ΔV deviates from the interface and increases further, suggesting that carrier transport paths can be reversed by modulating the contact configuration of J-TMDs in the heterojunctions. A better optoelectronic performance was found in the heterojunction based on PbI configuration rather than CsI configuration, resulting from the former higher work function and more complete structural framework. Meanwhile, the band offset, carrier transfer efficiency, and optical properties are directly dependent on both the intensity and direction of ΔV. In addition, the heterojunction with a larger ΔV pointing away from the interface possesses a smaller band offset with higher carrier transfer efficiency and optical absorption. CsPbI3/MoSSe heterojunction is recommended with a tunneling probability (P) of 79.65%. Our work unveils the role of intrinsic ΔV in asymmetrical polar 2D materials on the optoelectronic performances of perovskite devices, and provides a guideline to design high performance perovskite optoelectronic devices.

Results and discussion

Isolated CsPbI3 surfaces and monolayer J-TMDs

Before investigating the electronic structures of CsPbI3/J-TMDs heterojunctions, the electronic structures of their isolated surfaces are calculated by PBE and HSE06 functionals. In addition, the calculated lattice parameters are listed in Table S1 and consistent with previous experimental and theoretical results. The calculated projected band structures and corresponding parameters of those isolated surfaces are illustrated in Figure S1 (PBE), Figure 1 (HSE06), and Table 1. It can be found that electronic structures of those isolated surfaces calculated by PBE and HSE06 functionals are similar except for bandgaps, where the bandgaps calculated by HSE06 functional are consistent with experimental values. Therefore, HSE06 functional is mainly applied to investigate the electronic structure of CsPbI3/J-TMDs heterojunctions in this work.

Figure 1

The structure models and electronic structures od isolated CsPbI3 and J-TMDs

(A) The structure models for CsPbI3 surfaces and J-TMDs monolayers.

(B) Electronic band structures of CsI surface.

(C) Electronic band structures of PbI surface.

(D) Electronic band structures of MoS2 monolayer.

(E) Electronic band structures of MoSSe monolayer.

(F) Electronic band structures of MoSTe monolayer.

(G) Electronic band structures of WSTe monolayer.

(H) Planar-averaged electrostatics potentials of Janus TMDs along the direction normal to the interface region.

(I) Exciton binding energies and effective masses of Janus TMDs as functions of the ΔV

Table 1

Bandgaps, effective masses and exciton energies of isolated surface of CsPbI3 and J-TMDs monolayers

Isolated surface		CsPbI3		MoS2	MoSSe	MoSTe	WSTe
		CsI	PbI
Band gap/eV	HSE06	1.75	1.83	2.01	1.97	1.62	1.75
	PBE	1.34	1.43	1.53	1.32	1.04	1.21
	Experimental	1.73		1.98	–	–	–
me∗/m0		0.19	0.18	0.41	0.49	0.51	0.57
mh∗/m0		0.29	0.26	0.48	0.58	1.21	1.32
Ex/meV		106.21	77.98	262.21	247.69	238.61	222.92

Here, two types of CsPbI3 surface namely CsI and PbI surface are considered, as shown in Figure 1A. The CsI and PbI surfaces have a direct bandgap with 1.75 and 1.83 eV, respectively, which are consistent with the previous studies ( Liu et al., 2018). The conduction band minimum (CBM) and valence band maximum (VBM) of both the CsI and PbI surfaces are dominated by Pb p-orbitals and I p-orbitals, respectively, as is illustrated in Figures S2A and S2B. As for the effective masses m∗ and m∗, they are 0.19 and 0.29 m0 for CsI surface, whereas 0.18 and 0.26 m0 for PbI surface, respectively; such effective masses are consistent with previous research (He et al., 2019). Correspondingly, the exciton binding energy E calculated is 106.21 meV for CsI surface which is larger than that of 77.98 meV for PbI surface, indicating a poor carries recombination for CsI surface compared with the PbI surface. In addition, as listed in Table S2, the energies of E and E surfaces are −57.17 and −58.96 eV, respectively, suggesting that the PbI surface is much more stable than the CsI surface.

For J-TMDs monolayer, their electronic structures are displayed in Figures 1D–1G and S2C–S2F. MoS2 monolayer has a direct bandgap with 2.01 eV, and its CBM and VBM are mainly consisted of Mo-d orbitals and a little contribution of S-p orbitals, which are consistent with previous reports (Kadantsev and Hawrylak, 2012; Santos and Kaxiras, 2013). It is thought that both the electron and hole are transported at the Mo-S bonds. The calculated effective masses m∗ and m∗ of MoS2 are 0.41 and 0.48 m0, leading to the exciton binding energy of 262.21 meV, which are consistent with previous studies (Park et al., 2018; Pedersen et al., 2016). For J-TMDs, evident intrinsic potential difference (ΔV) points from the heavy atoms Y side to light atoms X side are observed, as marked in Figures 1H and S3. The ΔVs are 0.61, 1.34, and 1.38 eV for Janus MoSSe, MoSTe, and WSTe, respectively, indicating that ΔV can be enhanced significantly by altering the chalcogens (X) or enhanced slightly by altering the metals (M) of J-TMDs. As a result, the CBMs of J-TMDs are dominated by the hybrid orbitals between M d-orbitals and p-orbitals of heavy Y, whereas the VBMs are dominated by the hybrid orbitals between M d-orbitals and p-orbitals of light X, as demonstrated in Figures 1D–1G and S2C–S2F. It is thought that electrons and holes of J-TMDs are transported at the M−Y and M-X bonds, respectively, indicating an excellent ability of electron-hole separation for J-TMDs (Tao et al., 2019; Thanh et al., 2020). It is noted that MoSSe has a direct bandgap with 1.97 eV, whereas MoSTe and WSTe have an indirect bandgap with 1.62 and 1.75 eV, respectively. The bandgap of J-TMD reduces with the enlarging ΔV by altering the X of J-TMDs, whereas an opposite character occurs while altering the M of J-TMDs. On the contrary, the effective masses significantly increase with the enhancing ΔV by altering the X of J-TMDs, whereas they are slightly enlarged as the variation of ΔV induced by altering X, as listed in Table 1 and Figure 1I. It is noted that variation of m∗ and m∗ of J-TMDs are strongly and slightly tuned by the ΔV of J-TMDs. As a result, the exciton binding energies of J-TMDs are lower than that of MoS2 monolayer, indicating the weaker carrier recombination for J-TMDs. Moreover, the exciton binding energies of J-TMDs decrease monotonously with the increasing ΔV of J-TMDs. Moreover, Table S2 lists out the energies of those E; the stability of J-TMDs changes to worse as the intrinsic ΔV increases while being altered by chalcogens, and changes to better when ΔV is being altered by metals. Hence, J-MoSSe seems to be a better choice because of its perfect properties and easy selenization of MoS2 in experiment. Such dependence of electronic and transport properties of J-TMDs on the strength of intrinsic ΔV expands the tunability of optoelectronic properties and application of 2D material in CsPbI3/2D heterojunctions.

Binding energies

Figure S4 shows the structure models of CsPbI3/J-TMDs heterojunctions, and Figure S5 illustrates the binding energies (E) as functions of the interlayer separation. For CsPbI3/MoS2 heterojunction, when the interlayer separations of with CsI/MoS2 and PbI/MoS2 heterojunctions are 3.2 and 3.4 Å, respectively, binding energies of corresponding CsI/MoS2 and PbI/MoS2 heterojunctions reach to the lowest negative values of about −0.039 and −0.109 eV/atom, as listed in Table S3. It indicates that the CsPbI3/MoS2 heterojunctions, in particular the PbI/MoS2 heterojunction, with such interlayer separations are stable and possible to be fabricated in experiment, which is consistent with previous reports (He et al., 2019).Similar to the CsPbI3/MoS2 heterojunction, the larger interlayer distances and lower binding energies are observed for PbI/MoS2 heterojunctions than CsI/MoS2 heterojunctions, suggesting the more stable for PbI/MoS2 heterojunctions than CsI/MoS2 heterojunctions. Moreover, it is interesting that the interlayer separations of CsPbI3/J-TMDs heterojunctions are dependent on the interfacial contact character, where the interlayer separation of interface contacted by light atoms X is shorter than that contacted by heavy atoms Y. Such opposite phenomenon is suitable to the binding energies for CsPbI3/J-TMDs heterojunctions. For examples, the equilibrium interlayer separations and binding energies of CsI/S-MoSSe and CsI/Se-MoSSe heterojunctions are 3.2 Å and 0.060 eV/atom, and 3.4 Å and 0.038 eV/atom, respectively. Moreover, all the interlayer separations are higher than the sum of radii of Pb (or Cs) and X (or Y) atoms, suggesting the van der Waals heterojunction for CsPbI3/J-TMDs heterojunctions. Except for the interfacial contact character, the intrinsic potential differences of J-TMDs also affect the binding energies of CsPbI3/J-TMDs heterojunctions, where the E increases with the increment of ΔV, as listed in Table S3. It suggests the poor stability for the CsPbI3/J-TMDs heterojunctions with large intrinsic potential difference.

Electronic structures

Figures 2A and S6A illustrate the projected band structures of CsI/J-TMDs and CsI/J-TMDs heterojunctions, respectively. It can be found that the electronic band structures of CsI/J-TMDs heterojunctions seem to be a simple sum of electronic band structures isolated CsI surface and J-TMDs monolayer, irrespective of heterojunction configurations. For example, the bandgaps of CsPbI3 and MoS2 parts for CsI/MoS2 heterojunctions are 1.79 and 1.97 eV (see Table 2), respectively, which are close to those of the isolated CsI surface (1.75 eV) and MoS2 monolayer (2.01 eV). Moreover, as shown in Figures 3 and S7, the CBM and VBM of isolated CsI part of CsI/MoS2 heterojunctions are dominated by Pb p-orbitals and I p-orbitals, respectively; the CBM and VBM of isolated MoS2 part of CsI/MoS2 heterojunctions both are dominated by Mo d-orbitals and S p-orbitals. The effective masses and exciton binding energy of CsI part and MoS2 part of CsI/MoS2 heterojunctions are 0.26 m0 of m∗ and 101.61 meV, and 0.43 m0 of m∗ and 270.12 meV, respectively. These are close to those of isolated CsI surface and MoS2 monolayer, as listed in Tables 1 and 2. These phenomena are the typical characters for vdW heterojunction, suggesting the CsI/MoS2 vdW heterojunction. Similar characters are also observed for other CsI/J-TMDs vdW heterojunctions, no matter how the intrinsic ΔVs and contact characters of J-TMDs change, as demonstrated in Figure 2 and Table 2.

Figure 2

The electronic structures of CsI/J-TMDs heterojunctions

(A) Projected band structures of CsI/J-TMDs heterojunctions.

(B) Schematic diagram of band alignment for CsI/J-TMDs heterojunctions.

(C) Schematic diagram of type-II heterojunctions.

(D) The band offsets of CsI/J-TMDs heterojunctions as a function of intrinsic ΔV

Table 2

The detailed bandgaps, CBM, VBM and effective masses (m0) of CsPbI3/J-TMDs heterojunctions

Contacted surface		MoS2	MoSSe		MoSTe		WSTe
			S-	Se-	S-	Te-	S-	Te-
CsI-	Eg (CsI)	1.79	1.72	1.69	1.90	1.80	1.89	1.85
	Eg (MXY)	1.97	1.96	1.96	1.73	1.64	1.92	1.75
	CBM	Mo-d	Mo-d	Mo-d	Mo-d	Cs-p	Mo-d	Cs-p
	VBM	I-p	I-p	I-p	I-p	Mo-d	I-p	Mo-d
	me∗	0.43	0.59	0.61	0.55	0.20	0.55	0.15
	mh∗	0.26	0.23	0.22	0.23	1.28	0.34	1.23
PbI-	Eg (CsI)	1.76	1.80	1.88	1.96	1.94	2.04	2.02
	Eg (MXY)	2.05	1.96	1.85	1.49	1.52	1.57	1.58
	CBM	Mo-d	Mo-d	Mo-d	Mo-d	Cs-p	Mo-d	Cs-p
	VBM	I-p	I-p	I-p	I-p	Mo-d	I-p	Mo-d
	me∗	0.41	0.66	0.61	0.54	0.21	0.50	0.18
	mh∗	0.29	0.24	0.19	0.20	1.36	0.25	1.47

Figure 3

Projected density of states for CsPbI3/J-TMDs heterojunctions

(A–C) Projected density of states for the heterojunctions based on CsI configuration.

(D–F) Projected density of states for the heterojunctions based on PbI configuration.

In addition, the VBM and CBM of CsI/J-TMDs heterojunctions are located at the perovskite (or J-TMD) parts and J-TMD (or perovskite) parts of heterojunctions, indicating the type-II band characters for all CsI/J-TMDs heterojunctions, as shown in Figures 2C and S6C. For example, the VBM and CBM of CsI/MoS2 heterojunctions are located at the CsI surface and MoS2 part, respectively, which are in good agreement with previous reports (He et al., 2019). It is thought that a strong ability of this heterojunction to move free electrons from perovskite to MoS2 part, and promote hole extraction from MoS2 part to perovskite surface. The conduction band offset (CBO) and valence band offset (VBO) of CsI/MoS2 heterojunctions are 1.24 and 1.42 eV, respectively. When the MoS2 layer with symmetrical structure is replaced by the J-TMD with asymmetrical structure, both the CBO and VBO of CsI/J-TMDs heterojunctions are reduced. Moreover, the band offsets, in particular the VBO, continue to decrease as the intrinsic potential difference (ΔV) of J-TMD enlarges, as illustrated in Figure 2D. It is interesting that the decrement of CsI/J-TMDs heterojunction with heavy atom contact is stronger than that of CsI/J-TMDs heterojunction with light atom contact. As a result, the band offsets of type-II CsI/J-TMDs heterojunction with heavy atom contact are easy to turn into negative values, but the band offsets of type-II CsI/J-TMDs heterojunction with light atom contact keep positive values. It means that carrier transport paths of CsI/J-TMDs heterojunctions can be reversed by modulating the contact configuration of J-TMDs in the heterojunctions, although the CsI/J-TMDs heterojunctions remain the type-II band characters, as demonstrated in Figure S6C. Taking the CsI/MoSTe heterojunction as an example, the CBO (VBO) of CsI/Te-MoSTe heterojunction of about −0.11 eV (−0.23 eV) is smaller than that of CsI/S-MoSTe heterojunction of about 0.91 eV (0.81 eV), which are lower than that of CsI/MoS2 heterojunction of about 1.24 eV (1.42 eV). Moreover, photo-generated electron and hole of CsI/Te-MoSTe heterojunction are transport at CsPbI3 surface and MoSTe parts of heterojunctions, which are inverse to the CsI/S-MoSTe heterojunction whose photo-generated electron and hole are transport at the MoSTe and CsPbI3 parts, respectively. It is noted that the band offset of type-II heterojunction indicates the driving force and efficiency of carrier transfer at the type-II heterojunction. It suggests that the lower driving force and efficiency of electron (or hole) transfer from CsI surface (or J-TMDs layer) to J-TMDs layer (or CsI surface) for CsI/J-TMDs heterojunctions, in particular the heterojunctions with heavy atom contact, compared to those of CsI/MoS2 heterojunctions. Note that the direction of intrinsic ΔVs of J-TMD points from the heavy atomic layer to the light atomic layer of J-TMD. Thus, it is thought that the decrements of band offsets for CsI/J-TMDs heterojunctions are dependent of the direction and amount of intrinsic ΔVs of J-TMDs layers in the heterojunctions, where the CsI/J-TMDs heterojunctions show smaller band offsets when the directions of intrinsic ΔVs of J-TMDs layers point away from the contact region than those point to the contact region. In addition, because the intrinsic ΔVs of J-TMDs are significantly and gradually tuned by altering the chalcogens and metals of J-TMDs, respectively (see Figure 1), the band offsets of CsI/J-TMDs heterojunctions can be severely reduced by altering the chalcogens of J-TMDs, whereas they are slightly reduced by altering the metals of J-TMDs.

Compared to the CsI/J-TMDs heterojunctions, similar characters are observed for the PbI/J-TMDs vdW heterojunctions. However, the band offsets of PbI/J-TMDs vdW heterojunctions are larger than those of CsI/J-TMDs heterojunctions. Moreover, the variations of band offsets for PbI/J-TMDs vdW heterojunctions are more significantly tuned by the direction and strength of intrinsic potential difference than those of CsI/J-TMDs heterojunctions, as comparatively illustrated in Figures 2 and S6. In addition, it is noted that the intrinsic ΔVs of J-TMDs can result in the photo-generated electron and hole separation at the J-TMDs layer according to previous studies (Li et al., 2018; Lian et al., 2019; Thanh et al., 2020; Zhang et al., 2020). Thus, the band structures of J-TMDs can be simplified as the oblique band levels. As a consequent, the photo-generated carrier can be transferred at the CsPbI3/J-TMDs contact region induced by the band offsets and J-TMDs layer induced by the intrinsic ΔVs of J-TMDs. Thus, in order to understand the band offsets and charge transfer at the CsPbI3J-TMDs heterojunctions, the band alignment diagrams of CsI/J-TMDs and PbI/J-TMDs heterojunctions are summarized in Figures 2B and S6B.

Transport properties

To further understand the carrier transfer at the CsPbI3/J-TMDs heterojunctions, energy band diagrams of CsPbI3/J-TMDs heterojunctions are illustrated in Figure 4. The energy levels of isolated CsPbI3 surface and J-TMDs layer are referenced to the vacuum level. The calculated work function (WF) of isolated CsI surface, PbI surface, and monolayer MoS2 are 4.65, 5.23, and 5.22 eV, respectively, which are consistent with previous studies (Cao et al., 2020; Matter, 2020). The carriers will flow until Fermi levels are lined up when perovskite surface and MoS2 contact. It is thought that there is a negative vacuum barrier (ΔV) between CsI surface and MoS2 layer upon forming CsI/MoS2 heterojunction with equilibrium Fermi level. As a result, CsI surface and MoS2 layer in CsI/MoS2 heterojunction demonstrate the upward and downward bending band levels, respectively (see Figure 4A), which will induce a positive energy barrier (ΔT) at when carrier transfers from CsI surface to MoS2 layer. Such vacuum barrier ΔV can be regarded as the driving force for carrier transfer at the heterojunction; such energy barrier ΔT hampers the carrier transfer efficiency at the heterojunctions. Compared to the CsI/MoS2 heterojunction, similar ΔV and ΔT are observed for the CsI/MoSSe heterojunction, except for the lower absolute values, because the MoSSe monolayer possesses the lower WF than the MoS2 monolayer. Moreover, the CsI/Se-MoSSe heterojunction shows the lower ΔV and ΔT than the CsI/S-MoSSe heterojunction, because the WF of MoSSe monolayer at the SE atomic layer is lower than that at the S atomic layer. This inconsistent WF induces the intrinsic potential difference ΔV for MoSSe monolayer whose direction points from SE atomic layer to S atomic layer, which promotes electron transfer from the SE atomic layer to the S atomic layer of MoSSe monolayer. It means that CsI/MoSSe heterojunction, especially the CsI/Se-MoSSe heterojunction whose ΔV direction points from SE atomic layer to S atomic layer, shows the weaker carrier transfer drive force and higher carrier transfer efficiency compared to the CsI/MoS2 heterojunction. According to the above analysis, increasing the intrinsic ΔV by altering the X/Y and M atoms of MXY can further decrease the WF of J-TMD monolayer, as listed in Table S4. As a result, the WF of J-TMD, especially at the Y atomic layer, is lower than that of CsI surface. Consequently, the negative ΔV can turn into positive ΔV, and ΔT continues to decrease for CsI/J-TMD heterojunction with large ΔV, as the ΔV and ΔT for CsI/Te-SMoTe and CsI/Te-SWTe heterojunctions displayed in Figure S8A. That is why the CsI/Te-SMoTe and CsI/Te-SWTe heterojunctions show the inversely type-II band alignments. It also indicates that the carrier transfer efficiency for CsI/J-TMD heterojunction can be improved by increasing the ΔV of J-TMD.

Figure 4

Schematic energy band diagrams for CsPbI3/J-TMDs heterojunctions

(A–C) Schematic energy band diagrams for the heterojunctions based on CsI configuration.

(D–F) Schematic energy band diagrams for the heterojunctions based on PbI configuration.

Different from the CsI/MoS2 heterojunctions, PbI/MoS2 heterojunctions demonstrate a positive ΔV because the WF of PbI surface is larger than that of MoS2 monolayer (see Figure 4D and Table S4). This positive ΔV hampers the electron transfers from the perovskite surface to MoS2 layer. Meanwhile, PbI surface and MoS2 layer in PbI/MoS2 heterojunction demonstrate the downward and upward bending band levels, respectively (see Figure 4D), which will induce a much lower energy barrier ΔT when the carrier transfers from PbI surface to the MoS2 layer. Different from the MoS2 monolayer, the MoSSe monolayer with an intrinsic ΔV exhibits the lower WF than PbI surface, which resulting in the larger ΔV for PbI/MoSSe heterojunction than PbI/MoS2 heterojunction. It suggests the weaker drive force for charge transfer from perovskite to the MoSSe layer. Theoretically, the larger ΔV will induce the stronger band bending, which leads to the smaller ΔT for PbI/MoSSe heterojunction, in particular the PbI/Se-MoSSe heterojunction. As the intrinsic ΔV increases, the WFs of J-TMDs will continue to decrease, which theoretically leads to the increasing ΔV and reducing ΔT for PbI/J-TMDs heterojunction, especially with the PbI/Y-MXY configuration, as displayed in Figures 4E and 4F. It is noted that PbI surface is p-type semiconductor, suggesting that hole transfers from PbI surface to J-TMDs layer. It will raise the band levels of J-TMDs layer. Meanwhile, the intrinsic ΔV of J-TMD promotes hole transfers from the Y atomic layer to the X atomic layer of J-TMDs, which further increases the band levels of X atomic layer of J-TMDs. Consequently, the ΔT of PbI/J-TMD heterojunction with PbI/X-MXY configuration may be enlarged as the intrinsic ΔV increases.

As an evidence of the analysis above, the planar-averaged electrostatics potentials (top) and planar-averaged charge density differences (bottom) of the CsPbI3/J-TMDs heterojunctions are demonstrated in Figure 5 and detailed in Figures S9–S11. It can be found that electrons and holes are mainly accumulated at the MoS2 and CsI layers of CsI/MoS2 heterojunction, respectively, suggesting that electrons transfer from the CsI surface to the MoS2 layer of heterojunction. Moreover, the charge transfer decreases when the MoS2 layer of heterojunction is replaced by J-TMDs layer with intrinsic ΔV to form CsI/J-TMDs heterojunction, in particular the heterojunction with CsI/Y-MXY configurations, as displayed in Figures 5A–5C. The charge transfer of CsI/J-TMDs heterojunction further decreases with the enlarging intrinsic ΔV (see Figure S12. These characters further confirm the above variation of ΔV for CsI/J-TMDs heterojunctions. Different from CsI/J-TMDs heterojunction, some electrons are accumulated at the PbI surface of PbI/J-TMDs heterojunctions, especially the heterojunction with PbI/X-MXY configuration, because holes are transferred from p-type PbI surface to J-TMDs. This is consistent with the above analysis of the band diagram. Moreover, the hole (electron) transfer are enhanced (reduced) for PbI/J-TMDs heterojunctions as the intrinsic ΔV enlarges, which is in good agreement with the above ΔV for PbI/J-TMDs heterojunctions.

Figure 5

Interfacial transportation properties for CsPbI3/J-TMDs heterojunctions

(A–C) Planar-averaged electrostatic potentials and charge density differences for the heterojunctions based on CsI configuration.

(D) The tunnel barriers and tunneling probabilities of CsI/J-TMDs as a function of intrinsic ΔV

(E–G) Planar-averaged electrostatic potentials and charge density differences for the heterojunctions based on PbI configuration.

(H) The tunnel barriers and tunneling probabilities of PbI/J-TMDs as a function of intrinsic ΔV

It is noted that the above analyzed ΔT can be quantitatively marked in the electrostatics potentials of heterojunction, as seen in the yellow square in Figure 5. The detailed ΔTs of CsI/J-TMDs and PbI/J-TMDs heterojunctions are illustrated in Figures 5D and 5H. It can be found that the ΔTs of CsI/J-TMDs heterojunctions decline with the enlarging intrinsic ΔV. Moreover, the ΔT reduction of CsI/J-TMDs heterojunctions with CsI/Y-MXY configuration is more significant than that with CsI/X-MXY configuration. It is thought that the carrier transfer efficiency at the CsI/J-TMDs heterojunctions, especially the heterojunction with CsI/Y-MXY configuration, increases with the enlarging intrinsic ΔV. In contrast, the ΔTs of PbI/J-TMDs heterojunctions with PbI/X-MXY configuration enhance, whereas those of PbI/J-TMDs heterojunctions with PbI/Y-MXY configuration decline with the enlarging intrinsic ΔV, which are consistent with the above analysis of ΔT for PbI/J-TMDs heterojunctions from band diagram.

To directly show the carrier transfer efficiency at the CsPbI3/J-TMDs heterojunctions, the tunneling probability (P) is calculated as follows (Yuan et al., 2020, 2021)where m is the free electron mass of isolated CsPbI3 of about 0.18 m0, ħ is the reduced Plank's constant, and Φ and W are the height and full width at half maximum of the tunneling barrier, respectively. The tunneling probability P of CsPbI3/J-TMDs heterojunctions are detailed at Figures 5D and 5H. For the CsI/MoS2 heterojunction, the ΔT is about 3.8 eV with corresponding P of about 31.06%, which is consistent with previous studies, (He et al., 2019) indicating a poor carrier transfer efficiency at the CsI/MoS2 heterojunction. Moreover, as illustrated in Figure 5D, when intrinsic ΔV is introduced into J-TMDs to form CsI/J-TMDs heterojunctions, the ΔT decreases and P increases with the enlarging intrinsic ΔV, irrespective of the ΔV direction, indicating an improved carrier transfer efficiency. In addition, the ΔT reaches to the lowest 1.53 eV with a highest P of about 60.98% in CsI/Te-WSTe heterojunction. Better than CsI/MoS2 heterojunction, the ΔT reduces to 0.74 eV with P of about 77.26% in the PbI/MoS2 heterojunction which is about twice as much as CsI/MoS2 heterojunction. Consistent with above analysis of ΔT and ΔV in Figure 4, the ΔT (P) of PbI/J-TMD heterojunction enlarges (reduces) in PbI/X-MXY configuration, while being reduced (enlarges) in PbI/Y-MXY configuration as the intrinsic ΔV increases. In addition, the P of PbI/Te-WSTe heterojunction reaches a maximum of 81.02% and a minimum ΔT of 0.63 eV. Those dates with an enhanced P suggests an improved carrier transfer efficiency in CsPbI3/J-TMDs heterojunctions compared with CsPbI3/MoS2 heterojunctions, indicating an effectively way to improve the interfacial transportation by introduce ΔV to construct CsPbI3/J-TMDs heterojunctions.

Optical properties

It is important to study the light absorption capacity of heterojunction for its application in optoelectronic devices. The absorption spectra is calculated by the following formulawhere is the absorption coefficient, is optical frequency, and and are the real and imaginary parts of the dielectric function, respectively. Figure 6 demonstrated the calculated absorption coefficients of the CsPbI3/J-TMDs heterojunctions, and it seems a higher optical absorption of those heterojunctions with PbI configuration rather than CsI configuration at infrared-visible range, but a lower absorption in UV region, result from the former stronger carrier transfer efficiency analyzed above. Moreover, the optical absorption of those CsI/J-TMDs heterojunctions drops in infrared region with increase in visible-ultraviolet range, and that trend enlarges further as the intrinsic ΔV increases, irrespective of the configuration and the direction of ΔV, as detailed in Figure S13. In UV region, those CsI/J-TMDs heterojunctions with CsI/Y-MXY configuration have higher absorption than those CsI/J-TMDs heterojunctions with CsI/X-MXY configuration, which is in good agreement with the above carrier transfer efficiency analyses. Same characteristics are observed in PbI/J-TMDs heterojunctions. These results suggest that forming CsPbI3/J-TMDs heterojunction with ΔV has a significant enhancement for its application in photoelectric devices.

Figure 6

Calculated absorption coefficients for CsPbI3/J-TMDs heterojunctions

In summary, the optoelectronic properties of CsPbI3/J-TMDs heterojunctions have been modulated by the intrinsic ΔV here. It is found the heterojunction with a larger ΔV pointing away from the interface possesses a smaller band offset with higher carrier transfer efficiency and optical absorption, but a reversed type-II band alignment occurs when ΔV increases further, suggesting that carrier transport paths can be reversed by modulating the contact configuration of the J-TMDs in the heterojunctions. As a result, the CsPbI3/MoSSe heterojunction is suggested in this work with P of about 79.65%.

Conclusions

Here, the optoelectronic properties of CsPbI3/J-TMDs heterojunctions are studied by using the first-principle calculation to reveal the influence of intrinsic ΔV on the interfacial transport properties of optoelectronic devices. Remarkably, we found that a modulable ΔV exists in J-TMDs, which could be modulated significantly by altering chalcogens or modulated slightly by altering metals of J-TMDs. Moreover, the CsPbI3/J-TMDs heterojunctions have a type-II band alignment with the VBM and CBM located at the perovskite and J-TMDs parts, respectively, indicating an ability for separating electrons and holes continuously and effectively, similar to CsPbI3/MoS2 heterojunctions. In addition, the type-II band alignment is reversed when ΔV deviates from the interface and increases further, suggesting that carrier transport paths can be reversed by modulating the contact configuration of J-TMDs in the heterojunctions. Meanwhile, a better optoelectronic performance is found in the heterojunction based on PbI configuration rather than CsI configuration, resulting from the former higher work function and more complete structural framework. The band offset, carrier transfer efficiency and optical absorption are directly determined by the intensity and direction of ΔV. In addition, the heterojunction with a larger ΔV pointing away from the interface possess a smaller band offset with higher carrier transfer efficiency and optical absorption, and CsPbI3/MoSSe heterojunction with P of about 79.65% is suggested in this work. Overall, modulating the intensity and direction of ΔVin can facilitate to formation type-II CsPbI3/J-TMDs heterojunctions with excellent interfacial carrier transfer efficiency, our findings are vital to design and realize novel high-performance lead halide perovskite-based optoelectronic device.

Limitations of the study

In this work, we introduced the interfacial transport modulation of lead halide perovskite by the intrinsic potential difference of asymmetrical 2D materials Janus TMDs based on CsPbI3/J-TMDs heterojunctions. The heterojunctions with ΔV show excellent interfacial charge transport and absorption propriety. It would be more interesting to study the optoelectronic device based on CsPbI3/J-TMDs heterojunctions with ΔV in experiment.

STAR★Methods

Key resources table

Resource availability

Lead contact

Further information and requests for resources should be directed to and will be fulfilled by the lead contact, Jingjing Chang (jjingchang@xidian.edu.cn).

Materials availability

This study did not generate new unique reagents.

Experimental model and subject details

Our study does not use experimental models typical in the life sciences.

Method details

All calculations is performed using density functional theory (DFT), as implemented in the Vienna ab initio simulation package (VASP) code(Allouche, 2012). The projector-augmented wave (PAW) method is used to describe the interaction between ion cores and valence electrons(Bloechl, 1994). Perdew-Burk-Ernzerhof (PBE) exchange-correlation functional is employed. Van der Waals (vdW) interaction is applied total energies and forces by means of DFT-D2 method(Ilawe et al., 2015; Tran et al., 2007). The cutoff energy of plane-wave is set to be 400 eV. The convergence criterion is set to be 1 × 10−5 eV for the self-consistent process, and all atoms are allowed to be fully relaxed until the atomic Hellmann-Feynman forces are less than 0.01 eV/Å−1, respectively. A vacuum of 15 Å is considered along z direction to avoid artificial interlayer interactions. A 9 × 9×1 k-sampling generated by the Monkhorst-Pack scheme is used for the Brillouin zone integration. The hybrid functional of Heyd-Scuseria-Ernzerhof (HSE06) with the spin−orbital coupling (SOC), which adopts a mixing parameter of 0.25 and a screening parameter of 0.2 Å−1 for the Hartree-Fock exchange, is used to correct the electronic structures(Heyd et al., 2005; Heyd and Scuseria, 2004).

The CsPbI3/J-TMDs heterojunctions are composed of 1 × 1 supercell of CsPbI3 (001) surface and 2×√3 supercell of J-TMDs (001) surface. The average lattice mismatches of CsPbI3/J-TMDs heterojunctions are less than 2.39%.

To obtain the stable CsPbI3/J-TMDs heterojunctions, the binding energies (Eb) of CsPbI3/J-TMDs heterojunctions are calculated using the formula (He et al., 2019)where , and respectively represent the energies of CsPbI3/J-TMDs heterojunctions, isolated CsPbI3 surface and monolayer J-TMDs. N represents the atoms number of the corresponding CsPbI3/J-TMDs heterojunctions.

The Wannier-Mott exciton binding energy (E) estimated based on the effective mass theory can be calculated as(Guo et al., 2021; Wang et al., 2021b)where is the reduced mass given from the effective masses of electron and hole as followin which ε is the dielectric constant of surrounding material in the Coulomb interaction, m∗ and m∗ are the effective masses of electron and hole, respectively.

The charge density difference along z direction Δρ(z) was characterized as(Zhang et al., 2019)where , and are the charge densities of CsPbI3/J-TMDs heterojunctions, isolated CsPbI3 surface and monolayer J-TMDs, respectively.

Quantification and statistical analysis

Our study does not include statistical analysis or quantification.

Additional resources

Our study does not include any additional resources.

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Software and algorithms
Vienna Ab-initio Simulation Package	Xidian University	VASP 5.4.1