High-resolution transmission electron microscopy (HRTEM) images of the annealed GeTe film are shown in Fig. 1 (e) and (f). The insets show the Fast Fourier transform (FFT) patterns of the GeTe film. Indices of crystal planes are indicated on the images. According to these results, the annealed GeTe film exhibited good crystallinity. Fig. 1 (g) shows the line profiles of the lattice fringes shown in Fig. 1 (e) and (f). The top and bottom line profiles of Fig. 1 (g) corresponds to (202) and (220) crystal planes of GeTe film, which has a lattice fringe separation of 0.294 and 0.209 nm, respectively. Schematic diagram of the GeTe lattice structure is illustrated in Fig. 1 (h). Fig. 1 (i) and (j) show crystal plane models of GeTe as observed in Fig. 1 (e) and (f), respectively.
Raman spectroscopy was performed to study the structure of the GeTe films before and after annealing using a Renishaw inVia Raman microscope equipped with an argon-ion laser operating at an excitation wavelength of 514 nm. Fig. 2 (a) and (b) show the normalized Raman spectra of as-deposited and annealed GeTe films, respectively. The results are in good agreement with the literatures [27, 28]. There were three distinctive bands between 100 and 300 cm-1 as shown in Fig. 2 (a). These bands were situated at 124.8, 161.8 and 223.5 cm-1, namely band B, C and D, respectively. After annealing, there was a significant reduction in band D and also an appearance of band A situated at 108.1 cm-1 as shown in Fig. 2 (b). Band B, C and D were also red-shifted by 1.1, 5.3, and 21.9 cm-1,respectively. These are attributed to structural transformation of the GeTe film resulting in reduction in the degree of disorder (e.g. ratio of intermolecular to intramolecular interactions) [27].
To investigate the optical properties of the GeTe films before and after annealing, UV-Vis-NIR absorption spectroscopy were performed using a Horiba iHR 320 spectrometer. Fig. 2 (c) shows the UV-Vis-NIR absorption spectra obtained from both films. An absorption peak at 600 nm was apparent after annealing. The absorption coefficient of annealed GeTe film was significantly larger than that of unannealed film. Furthermore, a decreasing trend in the absorption coefficient was observed for an increasing wavelength in the infrared band. Bandgap energy (Eg) of the films can be determined using the following formulae [29, 30]: (see Formula 1 in the Supplementary Files)
where hν is energy of incident photon, α is optical absorption coefficient associated with hν, and C is a constant. The direct optical bandgap of the GeTe films can be estimated from the curve of α2 vs. photon energy (hv) as shown in the inset of Fig. 2 (c). It can vary greatly depending on the experimental conditions and theoretical models [31]. In this work, the estimated Eg of the GeTe films before and after annealing was 0.85 and 0.70 eV, respectively. This is in good agreement with previous work performed by others, which reported an optical bandgap of ~ 0.85 eV for an amorphous GeTe film and ~ 0.73 - 0.95 eV for crystalline film [32]. A reduction of Eg was reported after annealing because of long-range ordering of the lattice.
Atomic force microscopy (AFM) was carried out to determine the thickness of the films using AFM (SPA-400). Photoresist mask was used to prepare the sample for AFM measurements. Fig. 2 (d) shows an optical image of the prepared sample for AFM with an obvious boundary between GeTe film and substrate. Fig. 2 (e) reveals a film thickness of 33 ± 1.5 nm on Si substrates after annealing. Annealing has a little effect on the root-mean-square (RMS) surface roughness of the GeTe thinfilms, the RMS surface roughness decreased from 2.1 nm (as-deposited GeTe) to 1.4 nm (annealed GeTe).
The effect of annealing on the structure of GeTe nanofilms was further investigated using X-ray diffraction (XRD). Fig. 2 (f) shows the XRD spectra of the as-deposited (blue) and annealed (red) GeTe nanofilms. Two strong diffraction peaks at 29.9 ° and 43.2 °, which corresponded to (202) and (220) lattice planes respectively, appeared after annealing. In addition, two weak diffraction peaks at 26.0 ° and 53.5 °, which corresponded to (021) and (042) lattice planes respectively, also appeared in the spectrum. When combined with the above TEM results, it is evident that the GeTe nanofilm preferentially ordered along (220) and (202) lattice planes during the annealing process. Compared to the as-deposited GeTe films, the annealed GeTe has a drastic change in the crystal phase, the difference in the structure-related optical properties (absorption spectra) is shown in Fig. 2 (f) and Fig. 2 (c).
Elemental composition and chemical bonds at the surface of annealed GeTe nanofilms were studied by X-ray photoelectron spectroscopy (XPS) using AlKα radiation with energy of 1486.6 eV. XPS spectra of Ge 2p, Ge 3d and Te 3d core level peaks of the annealed GeTe film are shown in Fig. 2 (g), (h) and (i), respectively. The Ge 2p core level consisted mainly of Ge 2p3/2 (1220.1 eV) and Ge 2p1/2 (1251.1 eV) doublet peaks. The Ge 3d core level was deconvoluted into two components, namely Ge-Te and Ge-O at binding energy of 30.0 and 32.8 eV, respectively. The Te 3d core level consisted of Ge-Te, Te-O and Te-Te components. The Te-O (Te4+) peaks at 576.5 eV (Te 3d5/2) and 587.0 eV (Te 3d3/2) in Fig. 2 (i) were associated with TeO2 [33, 34]. Both Ge 3d and Te 3d core levels of annealed GeTe nanofilm exhibited oxygen-related components as shown in Fig. 2 (h) and (i), respectively. However, there was no oxygen-related component at Ge 2p core level, which was at greater penetration depth, as shown in Fig. 2 (g). Furthermore, GeO2 and TeO2 were absent from the XRD and TEM characterizations, hence this suggests that the oxidation of Ge and Te atoms were primarily localized at the surface of the film by atmospheric oxygen during the transfer and annealing processes [34] and the oxide layer was
A prototype photovoltaic detector based on p-GeTe/n-Si heterojunction was fabricated to explore the use of the material in the field of optoelectronics. The device fabrication processes are illustrated in Fig. 3 (a). Fig. 3 (b) depicts the structure of the photodetector. The thickness of the GeTe film and Al electrodes was 33 and 100 nm, respectively. Fig. 3 (c) and (d) show the response time of the device. The rise time (tR) is defined as time taken for the current to increase from 10 to 90 % of the peak, while the decay time (tD) is time taken for current to decrease from 90 to 10 %. As shown, the rise and fall time were symmetrical with a response time (τ) of 134 ms (e.g. (tR+ tD)/2).
Photoresponse of the device was evaluated from J-V measurements using Keithley 2400 sourcemeter under light illumination. The log J vs V characteristics of the device irradiated by λ=850 nm light at different densities of 20, 53, and 90 μW∙cm-2 and under dark condition performed at room temperature are shown in Fig. 3 (e). It can be seen from Fig. 3 (e) that the voltage corresponding to the minimum value of Jopt (i.e. photocurrent density) deviated by 0.1 V from the voltage corresponding to the minimum value of JD (i.e. dark current density) in the direction of positive bias, and that the photogenic voltage was generated under the light conditions. Therefore, the p-GeTe/n-Si heterojunction has demonstrated its potential application in infrared detection.
Two important figures of merit for photodetector, such as responsivity (R) and detectivity (D*), were determined using the following equations [35, 36]: (see Formulas 2 and 3 in the Supplementary Files)
where Ip is photocurrent that equals to absolute value of current under irradiation subtracting that in the dark, A is the effective area of the device, Popt is incident optical power, Id is dark current and q is unit charge (1.6×10-19 C).
The values of R and D* were 6 - 15 A/W and 1 - 8×1011 Jones (1 Jones =1 cm∙Hz1/2W-1) as obtained from Fig. 3 (f) and (g), respectively. The device was evaluated at room temperature, unpackaged and without optimization. Table 1 lists the responsivity and detectivity of some infrared photodetectors based chalcogenide/Si heterojunction, it can be seen that GeTe/Si shows a relatively higher performance at room temperature, which maybe due to the big absorption coefficient and the direct band gap of GeTe.
Table 1 Comparison of responsivity and detectivity of infrared photodetectors based on other materials forming heterojunction with Si
Heterojunction
|
Wavelength (nm)
|
R (A/W)
|
D* (Jones)
|
Ref.
|
GeTe/Si
|
850
|
6 - 15
|
1 - 8×1011
|
This work
|
Bi2Se3/Si
|
808
|
24.28
|
4.4× 1012
|
[37]
|
SnS/Si
|
850
|
0.083
|
5.3 × 109
|
[38]
|
MoS2/Si
|
300-1100
|
0.0119
|
2.1 × 1010
|
[39]
|
WS2/Si
|
400-1100
|
1.11
|
5 × 1011
|
[40]
|
WS2/Si
|
Near-infrared
|
3.7-4.5
|
|
[41]
|