1 Introduction

Two-dimensional (2D) materials, such as graphene [1] and group III chalcogenides [2], have become worldwide research hot spots due to its advanced functional properties that can be useful for a wide range of applications [3,4,5]. What the most interested is the electric properties of 2D materials that have high carrier mobility under low temperature; such unusual combination opens exciting opportunities for their emerging applications in electronics and photonics [6]. The atomic thin graphene [7] combined with atomically smooth surface substrate [8] and edge-contact geometry [9] makes the electric properties of crystal approach its theoretical phonon-scattering limit. The surface potential of 2D material is related to its Fermi level, which is vital to its electric properties, such as the contact barriers and carrier concentration, so the surface potential of 2D material is crucial for the booming van der Waals heterostructure engineering that aligns the 2D material surfaces one by one vertically [10].

Indium selenide (InSe) is a member of layered group III chalcogenide semiconductors [11]. Recent study shows that 2D electron gas induced by the field effect on the surface of few-layer InSe showed high carrier mobility up to 1 × 103 and 1 × 104 cm2·V−1·s−1 at room and liquid helium temperatures [12], respectively, accompanying with fully developed quantum Hall effect. The few-layer InSe also shows impressive high Young’s modulus and intrinsic strength of 101.73 and 8.68 GPa [13], respectively. Meanwhile, black phosphorous/InSe-based photodetector shows fast response (~ 22 ms) and high responsivity for visible and near-infrared light [14]. However, the surface potential of few-layer InSe still lacks attention due to its air sensitivity and researcher’s partiality on other 2D materials, especially for MoS2 [15]. The surface potential and work function of MoS2 are well understood, and researchers explored the effect of factors of surface potential such as layer count, light intensity, electric bias, methods of sample preparation and different substrates [16,17,18,19]. Mechanisms for the change of work function of MoS2 were also proposed, such as interlayer screening effects [16], the adsorption of water/oxygen molecules [17] and charge distribution [18]. For InSe, the theoretical study on work function of InSe was very recently proposed based on density functional theory, presenting that the work function of InSe decreases from 5.22 to 4.77 eV when the InSe stacks together from monolayer to bulk [20]. And the experimental study for the surface potential of few-layer InSe is still absent.

Here, the exfoliated few-layer InSe was studied by Kelvin probe force microscopy (KPFM) [21, 22]. Firstly, few-layer InSe flakes were exfoliated from bulk InSe crystals and on Si/SiO2 with 300-nm-thick thermal oxide (supported InSe) and Si/SiO2 with 300-nm-thick thermal oxide and prefabricated micro-wells with 3 μm in diameter (suspended InSe). Then, the surface potential of InSe flakes was measured by KPFM. To determine the exact value of surface potential, the surface potential maps were divided into several parts based on their layer count, and then, the corresponding work functions were determined by treating SiO2 as a reference. These results show that the work function of InSe decreases with the increase in the layer count of both suspended InSe and supported InSe. By illuminating light, the surface potential (SP) and surface potential shift (∆SP) between light and dark states of supported InSe were measured and deduced, respectively; these results give the evidence that how the light affects the surface potential of supported InSe, finally.

2 Experimental

2.1 Sample preparation

Few-layer InSe flakes were exfoliated from InSe crystals and on Si/SiO2 with 300-nm-thick thermal oxide and Si/SiO2 with 300-nm-thick thermal oxide and prefabricated micro-wells with 3 μm in diameter. The Scotch tape was used as the transfer medium. More details can be found in our previous works [13, 23, 24]. Specially, due to the instability of InSe in air, all samples were stored in glove box with a nitrogen atmosphere.

2.2 Raman spectroscopy

Raman scattering spectroscopy was performed on a Horiba high-resolution confocal Raman microscope integrated with a 532-nm laser as the excitation source. The characteristic vibrations frequency were used to confirm the optical properties of InSe which is related to its layer count [25]. The LabSpec was used to control the XYZ axes motorized sample stage, which can move sample accurately, and the output power was controlled by a series of neutral density filters. Here, a standard D1 (10%) filter was used to attenuate the power intensity of 532-nm laser (~ 88 mV); as a result, the output power of Raman is less than 9 mW.

2.3 KPFM

Theoretically, by applying an alternating current (AC) voltage (VACsin(ωt)) and a direct current (DC) voltage (VDC) to the conductive cantilever, the total voltage applied in tip (Vtip) is

$$V_{\text{tip}} = V_{\text{DC}} { + }V_{\text{AC}} { \sin }(\omega t)$$
(1)

The tip and sample can be modeled as a parallel plate capacitor, and then, the relation between the total potential difference and the electrostatic force (F) can be stated as

$$F = \frac{1}{2}\frac{\partial C}{\partial z}V^{2}$$
(2)

where C and z are the capacitance and normal distance between tip and sample, respectively; V is the total potential difference between sample and tip, which can be given by

$$V = V_{\text{tip}} - V_{\text{SP}} = (V_{\text{DC}} - V_{\text{SP}} ) + V_{\text{AC}} { \sin }(\omega t)$$
(3)

where VSP is equal to the contact potential difference between cantilever tip and sample, also termed as surface potential (SP) here [22, 26]

$$V_{\text{SP}} = \frac{{\varPhi_{\text{tip}} - \varPhi_{\text{sample}} }}{ - \text{e}}$$
(4)

where \({{\varPhi_{\text{tip}}}}\) and \({{\varPhi_{\text{sample}}}}\) are work function of the tip and sample, respectively, while e is the electric charge carried by a single electron. By inserting Eqs. (3) into (2), the electrostatic force has the form

$$F = \frac{1}{2}\frac{\partial C}{\partial z}\left[ {(V_{\text{DC}} - V_{\text{SP}} ) + V_{\text{AC}} { \sin }(\omega t)} \right]^{2} = F_{\text{DC}} + F_{\omega } + F_{2\omega }$$
(5)

where FDC, Fω and F2ω are the DC, ω and 2ω component of force which can be given as

$$F_{\text{DC}} = \frac{1}{2}\frac{\partial C}{\partial z}\left[ {(V_{\text{DC}} - V_{\text{SP}} )^{2} + \frac{1}{2}V_{\text{AC}}^{2} } \right]$$
(6)
$$F_{\omega } = \frac{\partial C}{\partial z}\left[ {(V_{\text{DC}} - V_{\text{SP}} )V_{\text{AC}} { \sin }(\omega t)} \right]$$
(7)
$$F_{2\omega } = - \frac{1}{4}\frac{\partial C}{\partial z}V_{\text{AC}}^{2} \text{cos}(2\omega t)$$
(8)

From Eq. (7), if we can nullify the ω component of force (Fω) by adjusting the applied DC voltage, the applied DC voltage (VDC) is equal to the surface potential (VSP). To protect the InSe samples, all KPFM measurements were taken in nitrogen atmosphere.

3 Results and discussion

3.1 A few-layer InSe flake

Monolayer InSe has a honeycomb lattice that consists of four covalently bonded Se–In–In–Se atomic layers, as shown in Fig. 1a, and the layers are held together by vdW interactions with an interlayer distance of d ≈ 0.8 nm. Few-layer InSe flakes are micro-mechanically exfoliated from bulk InSe crystals and on Si/SiO2 with 300-nm-thick thermal oxidized and Si/SiO2 with 300-nm-thick thermal oxide and prefabricated micro-wells with 3 μm in diameter. The optical image of few-layer InSe on Si/SiO2 is shown in Fig. 1b, wherein few-layer InSe can be judged from their unique optical contrasts [24, 27], and the layer counts were measured by atomic force microscopy (AFM) scans in contact mode, with the resulting topography profiles inserted in Fig. 1b, showing thickness corresponding to 2L, 3L and 7L (L: layers) InSe. The Raman spectra of few-layer InSe are presented in Fig. 1c; four characteristic peaks are observed, which are A1′ (~ 114 cm−1) peak, E′ (~ 180 cm−1) peak, A1-LO (~ 200 cm−1) peak and A1 (225 cm−1) peak; for 7L InSe, these peaks fade away when the layer count decreases from 7 to 2, which are consistent with the data reported in previous studies [25, 28]. In order to protect the air-sensitive InSe flakes, the samples are stored in glove box with a nitrogen atmosphere. More details of sample preparation and Raman measurement are given in Sect. 2.

Fig. 1
figure 1

a Crystal structure of InSe; b optical image of a few-layer InSe flake deposited on Si/SiO2 and corresponding topography profiles; c Raman spectra of 2L, 3L and 7L InSe in range of 100–260 cm−1; d schematic of light-excited KPFM

Full size image

3.2 Work function of a few-layer InSe

Work function of 2D materials can be influenced by many factors, including substrate, crystal face and surface contamination. In order to explore the intrinsic work function of InSe, the supported and suspended InSe were prepared, and their work functions were measured by KPFM, as schematically shown in Fig. 1d. An integrated 405-nm (3.06 eV) laser was also used to excite the samples to study their responses under light illumination. During KPFM measurement, an AC voltage was applied to the scanning tip to produce an oscillation of cantilever, while a DC voltage was applied to suppress such oscillation. Consequently, the applied DC voltage is equal to the contact potential difference between cantilever tip and sample, also termed as surface potential. To prevent degeneration of InSe in air, a steady nitrogen flow was pumped into Cypher ES chamber to maintain a nitrogen atmosphere. More details of KPFM are given in Sect. 2.

Experimentally, Fig. 2a, b presents the topography and surface potential mappings of a few-layer supported InSe from 2L to 7L, and the corresponding results for a few-layer suspended InSe are shown in Fig. 2c, d. Owing to instability of InSe in air, the topography of fresh InSe flake, which is smooth and flat as shown in Fig. 2a, c, is compared with that of degenerated InSe sample which has some bubble-like islands [29], to make sure that the InSe surface is fresh and clean without any degeneration. For supported InSe, it is interesting that the 2L InSe has surface potential (and thus work function) very close to SiO2, this property is different from MoS2 [18] and the surface potential of InSe decreases monotonously with the increase in the layer count. For suspended InSe, the surface potential of InSe also decreases monotonously with the increase in the layer count, as shown in Fig. 2d.

Fig. 2
figure 2

Surface potential of few-layer InSe flakes: a topography and b surface potential images of supported InSe; c topography and d surface potential images of suspended InSe

Full size image

To extract the exact surface potential of supported and suspended InSe, the surface potential maps in Fig. 2 were divided into several parts based on their layer counts, as shown in Fig. 3. Specifically, Fig. 3a shows 2L, 3L and 7L surface potential maps and histograms of supported InSe which give corresponding surface potential values of (482.62 ± 13.47), (445.80 ± 18.18) and (263.18 ± 19.10) mV based on Gaussian fitting, respectively; Fig. 3b illustrates 3L, 10L and 11L surface potential maps and histograms of suspended InSe which give related surface potential values of (622.04 ± 22.36), (615.40 ± 13.97) and (575.77 ± 34.69) mV based on Gaussian fitting, respectively; Fig. 3c presents the surface potential maps and histograms of SiO2 substrate of supported and suspended samples which give correlated surface potential values of (482.48 ± 19.28) and (645.16 ± 21.67) mV based on Gaussian fitting, respectively. Based on Eq. (4) and the work function of SiO2 (5.00 eV) [30], the work functions of tips for supported and suspended InSe samples can be deduced, separately, and the work functions of supported and suspended InSe can be determined. Detailedly, the work function of supported InSe drops from (5.00 ± 0.01) to (4.78 ± 0.02) eV when its thickness increases from 2L to 7L, and that of the suspended InSe decreases from (4.97 ± 0.02) to (4.93 ± 0.03) eV when its thickness increases from 3L to 11L, as shown in Fig. 3d. These results indicate that the work function of InSe decreases with the increase in the layer count of both supported and suspended InSe, which is consistent with DFT calculation [20].

Fig. 3
figure 3

Work function of few-layer InSe flakes: a surface potential maps and histograms of supported InSe for 2L, 3L and 7L, respectively; b surface potential maps and histograms of suspended InSe for 3L, 10L and 11L (L: layers), respectively; c surface potential maps and histograms of SiO2 substrate for supported and suspended samples; d work function versus layer count of supported and suspended InSe

Full size image

3.3 Light-excited surface potential of a few-layer InSe

To explore the influence of light intensity on the surface potential of InSe, the supported InSe is illuminated by an integrated 405-nm (3.06 eV) laser with tunable power. The band gap of a few-layer InSe is ~ 1.40 eV [25], and thus, the laser is expected to excite the electron from highest energy state in the valence band to the lowest energy state in the conduction band. And the work function and band gap of SiO2 are around 5.00 and 9.00 eV, respectively [30]. The surface potential of InSe under light intensities of 0, 0.045 and 0.09 W was measured, as shown in Fig. 4a-c. The increased light intensity is accompanied by decreased surface potential. The surface potential of supported InSe also decreases with the increase in the layer count under illumination. Figure 4d shows the surface potential of InSe in dark, under 0.045 and 0.090 W illumination for different layer counts, and the surface potential shift between light and dark states (∆SP =SPlight − SPdark) of multiple-layer InSe under 0.045 and 0.090 W illumination. For 2L InSe, the surface potential decreases from (482.62 ± 13.47) to (408.38 ± 8.01) mV when the light intensity increases from 0 to 0.090 W. For 3L and 7L InSe, surface potential decreases from (445.80 ± 18.18) to (375.49 ± 8.23) mV and from (263.18 ± 19.1) to (225.67 ± 7.26) mV, respectively, when the light intensity increases from 0 to 0.090 W. It is evident that light excites electrons, resulting in decreased surface potential, and the amount of electrons excited is correlated with light intensity. Finally, ∆SP between light and dark states decreases from 67.20 to 33.20 mV for 0.045 W light intensity and from 74.20 to 37.50 mV for 0.090 W light intensity with layer count increasing from 2L to 7L, as shown in Fig. 4d, which suggests that the influence of light illumination decreases with the increase in the layer count of supported InSe.

Fig. 4
figure 4

Light-excited surface potential of supported InSe: a surface potential maps of InSe a in dark, under illumination with intensity of b 0.045 mW and c 0.09 mW; d surface potential (right, black) and ∆SP (left, red) between light and dark states of supported InSe versus its layer count for different light intensities

Full size image

4 Conclusion

By investigating the exfoliated few-layer InSe deposited on Si/SiO2 with 300-nm-thick thermal oxide and Si/SiO2 with 300-nm-thick thermal oxide and prefabricated micro-wells with 3 μm in diameter, the surface potential of few-layer InSe flakes was measured via KPFM integrated with 405-nm (3.06 eV) laser. Based on work function of SiO2, the work function of supported and suspended InSe are determined which drops from (5.0 ± 0.01) to (4.78 ± 0.02) eV when its thickness increases from 2L to 7L, and that of the suspended InSe decreases from (4.97 ± 0.02) to (4.93 ± 0.03) eV when its thickness increases from 3L to 11L. Conclusively, the work function of InSe decreases with the increase in its layer count of both suspended and supported InSe, and 2L supported InSe has work function very close to SiO2. By introducing laser illumination, the surface potential and the surface potential shift (∆SP) between light and dark states of supported InSe were measured and deduced, respectively. These results show that the surface potential of supported InSe decreases with the increase in the light intensity and also decreases with the increase in the layer count. It is evident that light excites electrons, resulting in decreased surface potential, and the amount of electrons excited correlates with light intensity. And, ∆SP between light and dark states decreases with the increase in the layer count, suggesting that the influence of light illumination decreases with the increase in the layer count of few-layer InSe flakes.