Temperature-Dependent Optical Absorption and DLTS Study of As-Grown and Electron-Irradiated GaSe Crystals

Abstract

Optical absorption spectra of 9 MeV electron-irradiated GaSe crystals measured at temperatures in the range from 9.5 to 300 K were analyzed. The absorption spectra with features caused by Ga vacancies in two charge states and direct interband transitions were fitted by a model equation. Temperature dependencies of the defect concentrations and optical transition energies, as well as of the GaSe band gap, were determined. Current- and capacitance-voltage characteristics and DLTS spectra were measured for as-grown and electron-irradiated GaSe slabs with Sc (barrier) and Pt (ohmic) contacts. An experimental Sc/GaSe Schottky barrier height of 1.12 eV was determined in close agreement with a theoretical estimate. The activation energy and the hole capture cross-section deduced from the DLTS data are 0.23 (0.66) eV and 1.5 × 10⁻¹⁹ (2.3 × 10⁻¹⁵) cm⁻² for the supposed $V_{\mathrm{Ga}}^{-1}$ ($V_{\mathrm{Ga}}^{-2}$) defect. For the electron-irradiated GaSe crystals, the found activation energies are close to the values inferred from the optical measurements.

1. Introduction

Gallium selenide (GaSe) is an important member of layered post-transition metal chalcogenides, which are suitable for the formation of van der Waals heterointerfaces [1] and fabricating various quasi-two-dimensional structures [2,3,4,5,6], with high potential for application in electronic and optoelectronic devices. The interlayer binding energy of GaSe is calculated to be in the range of 14–16 meV/Ų [7]. The layered nature has the possibility of producing ultrathin sheets through exfoliation, as well as of introducing impurity atoms into the interlayer space (intercalation) for the preparation of structures with new properties [8,9]. The main application of GaSe crystals is in nonlinear optics as a medium for converting the frequency of laser radiation into the IR and terahertz frequency ranges [10,11]. Currently, it is one of the promising materials for obtaining a high repetition rate IR and THz optical parametric oscillators/amplifiers (OPOs/OPAs) [12] and high peak power laser beam frequency conversion to achieve GW-level (multi)terahertz pulses [11].

One of the areas of research concerning GaSe is the study of the properties of nanolayered A³B⁶ compounds, methods for their preparation, and applications in photodetectors [4,13] and gas sensors [3,14]. It is established that nanolayered materials with a small area of the working surface often significantly change their properties, primarily conductivity, under the influence of optical radiation [3,13]. This leads to high values of photosensitivity. For example, in GaS, this is associated, among other things, with an increase in mobility and absorption coefficients due to modification of the electronic structure during the transition from bulk to nanosized form [4]. Thus, the final parameters may depend on the method of obtaining the structures, the presence of defects and chemical impurities, and the thickness of the layers. In particular, the presence of mechanical strains, for example, when the semiconductor layers are deformed on a flexible substrate [15], can have a significant effect on the performance characteristics.

According to the available data, GaS and GaSe compounds possess the highest photosensitivity values among other representatives of the A³B⁶ family [3,4]. It is also known that high-energy particle bombardment can be used to modify the electrical and optical properties of semiconductors [16,17]. In particular, this can be employed to obtain modified 2D structures, either specially grown or cleaved from the irradiated bulk materials. At present, the physical properties of GaSe under the influence of high-energy particles have not been sufficiently studied. Recently, the optical and dielectric characteristics of GaSe crystals modified by irradiation with high-energy electrons were analyzed [18]. The appearance of “steps” in the optical transmission spectra was reported. It should be pointed out that (i) similar “steps” were previously observed for neutron-irradiated GaS [17], and (ii) the arising radiation defect absorption can increase the spectral range of photosensitivity of bulk and nanolayered A³B⁶ photodetectors. The investigation of the defect energy levels in irradiated semiconductors can be accomplished using the deep level transient spectroscopy (DLTS) technique [19,20]. DLTS, in combination with the measurements of current-voltage and capacitance-voltage dependencies, is known to be a powerful tool to determine the concentration of defects and their activation energies.

In the present paper, the results of the study of electron-irradiated GaSe samples by DLTS method are reported. First, we analyze the optical absorption spectra measured at various temperatures to obtain the temperature-dependent optical transition energies. Then, we employ DLTS together with I–V and C–V measurements to obtain electrical characteristics of the defects: activation energies, concentrations, and hole capture cross-sections.

2. Experimental Methods

2.1. Sample Preparation

The GaSe crystals with p-type conductivity were grown by the vertical Bridgman method from high-purity initial components (Ga and Se of 6N grade) [18]. The slabs irradiated with high-energy electrons (9 MeV) at the nuclear reactor of the Karpov Institute of Physical Chemistry (Obninsk, Russia) with fluences Φ of 1 × 10¹⁷ and 5 × 10¹⁷ cm⁻² were taken for measurements.

The GaSe slabs, both as-grown and irradiated with a dose of 1 × 10¹⁷ cm⁻², were used to fabricate metal-semiconductor-metal structures for I–V, C–V, and DLTS measurements. Before applying metal contacts, several upper layers were peeled off from both sides of each slab to remove the oxide and obtain a clean surface. Then, magnetron sputtering of platinum with a thickness of about 200 nm was performed. During the sputtering process, the substrates were heated to about 300 °C. This may be the reason for some decrease in the defect concentration relative to its initial level in the Φ = 1 × 10¹⁷ cm⁻² irradiated GaSe sample. The sputtered platinum layer served as an ohmic contact. Next, to create a barrier contact, scandium was magnetron-sputtered from a tungsten crucible. The rare-earth metal was sputtered to a thickness of about 50 nm. As a result, “vertical” structures were obtained, rectangular samples with a thickness of about 500 μm and an area of about 6 × 6 mm² with metal contacts on opposite sides (i.e., the direction of the current flow was perpendicular to the layer planes).

2.2. Optical Absorption Measurements

The optical absorption spectra, measured using a USB 4000+ spectrometer (Ocean Optics, Largo, FL, USA) in the spectral range 350–950 nm, were analyzed. The low-temperature measurements were done using a closed-cycle cryostat Model CCS-300S/204-HT (Lake Shore Cryotronics (Janis), Woburn, MA, USA). The power absorption coefficients α(λ) were derived from the measured intensity transmission spectra T(λ) as it was described in ref. [18].

The following equation was used to fit the experimental absorption spectra at each temperature [18]:

$$\alpha (\hslash \omega )=A_1{\displaystyle \frac{{(\hslash \omega -E_{A1})}^{1/2}}{{(\hslash \omega )}^3}}+A_2{\displaystyle \frac{{(\hslash \omega -E_{A2})}^{1/2}}{{(\hslash \omega )}^3}}+A_3{(\hslash \omega -E_g)}^{1/2}$$

(1)

where A₁–A₃ are constants, E_A1 and E_A2 are threshold energies for optical transitions, and E_g is the GaSe band gap.

The last term in Equation (1) describes direct interband transitions. The first two terms are based on the model for photoionization cross-section spectral dependence σ(ћω) proposed by Lucovsky [21,22] to explain optical absorption features in semiconductors with optical transitions involving deep levels (induced by introduced impurity atoms or structural defects), for which the hydrogen model (for shallow defect states) does not give a satisfactory description of the experimental results. Taking into account Equation (1) and assuming a 100% photoionization yield, the absorption produced by each of the defect states ($V_{Ga}^{-1}$ and $V_{Ga}^{-2}$) can be written as

$${\alpha }_{\mathrm{j}}(\hslash \omega )={\sigma }_{\mathrm{j}}(\hslash \omega )N_{\mathrm{j}}=A_{\mathrm{j}}{\displaystyle \frac{{(\hslash \omega -E_{\mathrm{Aj}})}^{1/2}}{{(\hslash \omega )}^3}}$$

(2)

where j equals 1 or 2, and N_j and σ_j are concentration and photoionization cross-sections for a defect of type j, whereas A_j and E_Aj are parameters for absorption by this defect from Equation (1). After finding the A₁ and A₂ values from fitting the experimental optical absorption spectra using Equation (1), we used Equation (2) and the Lucovsky formula for σ(ћω) [21,22] (with substituted m* = 0.17 m₀ [23], effective field ratio (E_eff/E₀) = 2, and ordinary refractive indices from ref. [24]) to estimate the defect concentrations in both charge states (see also the Supplementary Materials to ref. [18]).

2.3. I–V, C–V, and DLTS Measurements

Electrical measurements were performed on a setup based on a phase-sensitive lock-in amplifier MFLI 5 MHz (Zurich Instruments, Zurich, Switzerland) with an installed MF-IA option. Also used were a Linkam T96-S temperature controller (Linkam Scientific Instruments Ltd., Redhill, UK) and a THMS600 stage (Linkam Scientific Instruments Ltd., UK) equipped with BNC connectors for electrical connections to the sample.

The DLTS measurements were carried out with a four-terminal connection of the sample to the MFLI inputs at heating/cooling rates of 3 K/min. The test signal frequency of the impedance meter was 100 kHz, and the test signal amplitude was 100 mV. The sampling frequency was 20 kSamples/s, the bandwidth was 20 kHz. A reverse bias of −2 V was used with a filling pulse voltage of 0 V (+2 V relative to the reverse bias).

3. Results and Discussion

In our previous research, it was found that two bands appear in the optical absorption spectra of GaSe crystals after irradiation with 9 MeV electrons [18]. These absorption bands were associated with gallium vacancies in two charge states, $V_{Ga}^{-1}$ and $V_{Ga}^{-2}$, inducing levels with energy positions in the forbidden gap of the crystal, respectively, E_v + 0.23 eV and E_v + 0.61 eV at T = 295 K, where E_v is the valence band maximum (VBM).

It should be noted that a defect state with an energy level about E_v + 0.2 eV was found in a number of Hall effect, DLTS, thermally stimulated current (TSC), and photoluminescence studies of GaSe crystals and usually attributed to a gallium vacancy (for example, [25,26,27,28]). This conclusion finds support in the results of first principles DFT calculations [29,30]. The computer simulation of point defects in the β-polytype of GaSe in ref. [29] with large (up to 400 atoms) supercells and using semilocal density functional revealed the presence of an energy level E_v + 0.26 eV, generated by the gallium vacancy. The corresponding defect state has a moderately deep character and splits off the valence band. An unoccupied gap state with an energy level E_v + 0.92 eV at the Г point was also obtained in ref. [29] for GaSe with a Ga vacancy. In recent theoretical studies of electronic and structural properties of GaSe with intrinsic defects [30], an optimized HSE functional allowing better reproducing excited states was employed. The (0/−) charge transition level for a Ga vacancy was calculated to be E_v + 0.1 eV. Note that the authors of ref. [30] did not find the (−/−2) charge transition level for V_Ga with the Fermi level lying within the band gap. In an even more recent paper [31], devoted solely to monolayers of GaSe, some results of ref. [30] were disputed and, at least for the monolayer GaSe, the gallium vacancy was related to deep states.

In the present work, we have performed fittings with Equation (1) for all the measured absorption spectra for a GaSe sample irradiated with a fluence of 5 × 10¹⁷ cm⁻². The obtained data are presented in Figure 1. First, the examples of fitting curves for the spectra measured at 9.5, 30, 60, 140, and 300 K are shown in Figure 1a; the data for only five temperature points are chosen for simplicity. In order to better illustrate the contribution to the resulting absorption spectra of each of the defect states, the model curves resolved into contributions from each of three terms in Equation (1) are provided in Figure 1b for the spectra at temperatures of 9.5 and 300 K.

It can be seen that the model for Equation (1) provides a satisfactory description of the observed spectral features (Figure 1). The main discrepancy is between the experimental and the model curves in the spectral range of 1.8–2.09 eV for T = 9.5 K. Using the dispersion relation n_o(λ) [24] when calculating the absorption coefficients α(λ) from the measured transmission spectra T(λ) can be among the possible reasons. This dispersion relation is determined for room temperature conditions and therefore may deviate from real dispersion of the refractive index at low temperatures, especially in the range above 1.8 eV, close to the room temperature band gap of GaSe. The main source of the discrepancy is the inherent limitation of the Lucovsky model, where δ function-like potential is taken for the localized deep state [21,22]. If a non-zero radius of the localized potential were introduced into the model, the high-photon-energy tail after the absorption maximum would become shorter, and this would better match the spectral features observed in Figure 1. Also, the energy interval between the optical transition threshold E_A2 and the fundamental absorption edge becomes larger at lower temperatures, and a dip in the experimental spectrum is observable.

In the present work, we index the defect activation energies (charge transition levels) ΔE_aj and threshold energies for optical transitions E_Aj in ascending values. Thus, for $V_{Ga}^{-1}$, the associated defect is ΔE_a1 = E_A2 − E_g, while, for $V_{Ga}^{-2}$, the associated defect is ΔE_a2 = E_A1 − E_g. The physical sense of these quantities is also clear from Figure 2a.

The obtained temperature-dependent optical transition energies E_A1 and E_A2 and related defect activation energies ΔE_a1 and ΔE_a2 are displayed in Figure 2b. It is seen that the $V_{Ga}^{-1}$ induced energy level has a tendency to move closer to the VBM with the increasing temperature (shifts downward by 0.17 eV when the temperature changes from 9.5 to 300 K), while the $V_{Ga}^{-2}$ associated energy level shifts slightly upward.

From a theoretical point of view, the prediction and interpretation of temperature-driven changes in the defect level position is a complex problem. In ref. [32], a theoretical approach to assessing the temperature shift of the charge transition level was developed, in which the main role was played by changes in the local volume around the defect and band edge positions. According to the rules formulated in ref. [32], and taking into account the observed ΔE_a1 and ΔE_a2 temperature dependencies (Figure 2b), the charge transition −/−2 induces a small change in the local volume, while this change is larger for the transition 0/−1, which leads to level shifts in opposite directions. Note that a large lattice reconstruction around the V_Ga defect is predicted in refs. [29,31]. It can be supposed that, at further ionization, the lattice reconstruction will not be that large. Here, we also assume that GaSe is a semiconductor in which VBM rises with temperature (type I semiconductor [32]). This is partly supported by the results of our studies of the influence of hydrostatic pressure on the interatomic distances and interband transitions in GaSe [7,33]. Qualitatively, the contraction of the crystal lattice with decreasing the temperature can be compared with that after the application of pressure. The contraction of the interlayer space increases the overlapping of Se p_z orbitals, which form the highest valence band and therefore pushes the VBM up in energy.

The temperature-dependent concentrations of the optically active defects calculated using the found values of fitting parameters A₁ and A₂ and Equation (2) are given in Figure 2c. It is seen that, with the increasing temperature, the concentration of $V_{Ga}^{-1}$ defects is reduced, while the concentration of $V_{Ga}^{-2}$ is increased. At temperatures above 225 K, the majority of the optically active Ga vacancy defects are expected to be in a −2 charge state. The found temperature dependence of the GaSe band gap, E_g, is shown in Figure 2d. The experimental data were fitted by the Varshni formula [34]:

$$E_{\mathrm{g}}(T)=E_{\mathrm{g}}(0)-{\displaystyle \frac{\gamma T^2}{T+\beta }}$$

(3)

where E_g(0) is the band gap value at T = 0 K, and β and γ are constants. The found fitting parameters are given in Figure 2d. The obtained E_g(0) value is close to the one reported in ref. [35].

As is known, I–V characteristics of a Schottky diode can carry information about the barrier height and the series resistance of the sample, R_s. The latter parameter is important for determining the measurement mode using DLTS. In the case of large R_s values, measurements may become impossible or require compensation for the high bulk resistance of the sample when processing the measurement data.

In this work, the experimentally obtained I–V characteristics of the Sc/GaSe Schottky barrier (Figure 3a) were approximated by the well-known equation (for example, [36])

$$I=I_0\exp (e{U}_{\mathrm{b}}/nkT)(1-\exp (-e{U}_{\mathrm{b}}/kT))$$

(4)

where

$$U_{\mathrm{b}}=U-R_{\mathrm{s}}I$$

(5)

is the voltage at the metal-semiconductor junction, U is the applied external voltage, e is the electron charge, k is the Boltzmann constant, T is the sample temperature, n is the ideality factor, and I is the current flowing through the sample. The saturation current I₀ is described by the expression

$$I_0=A^{*}{T}^{2}S\exp (-e{\phi }_{\mathrm{b}}/kT)$$

(6)

where S is the contact area, φ_b is the Schottky barrier height, and

$$A^{*}=4\pi m^{*}e{k}^2/h^3$$

(7)

is the effective Richardson constant, and h is Planck’s constant. The effective mass for holes in GaSe m* = 0.5 m₀ was assumed [37], where m₀ is the free electron mass.

The energy barriers for electrons ${\mathsf{\Phi }}_{\mathrm{b}\mathrm{S}}^{\mathrm{n}}$ and holes ${\mathsf{\Phi }}_{\mathrm{b}\mathrm{S}}^{\mathrm{p}}$ are estimated using the charge neutrality level concept (CNL) as [38]

$${\Phi }_{\mathrm{b}\mathrm{S}}^{\mathrm{n}}=\left(E\mathrm{g}-CNL\mathrm{v}\right)+S\left(\Phi \mathrm{m}-CNL\mathrm{a}\mathrm{b}\mathrm{s}\right),{\Phi }_{\mathrm{b}\mathrm{S}}^{\mathrm{p}}=(E\mathrm{g}-{\Phi }_{\mathrm{b}\mathrm{S}}^{\mathrm{n}})$$

(8)

Here, the energy position of the charge neutrality level CNL_abs = (E_g − CNL_v + EA) is determined relative to the vacuum level. The parameter S is calculated according to the expression

$$\mathrm{S}=1/[1+0.1({\epsilon }_{\mathrm{\infty }}^{\mathrm{e}\mathrm{f}\mathrm{f}}-1{)}^2]$$

(9)

which takes into account the screening of the interface dipole induced by the metal in the semiconductor. In Equation (9), ${\epsilon }_{\mathrm{\infty }}^{\mathrm{e}\mathrm{f}\mathrm{f}}$ = (${\epsilon }_{\mathrm{\infty }}^{┴}$ × ${\epsilon }_{\mathrm{\infty }}^{\left|\right|}$)^1/2. For ε-GaSe, ${\epsilon }_{\mathrm{\infty }}^{\mathrm{e}\mathrm{f}\mathrm{f}}$ = 6.55, which gives S ≈ 0.25. The other physical parameters used to estimate the barrier heights ${\Phi }_{\mathrm{b}\mathrm{S}}^{\mathrm{n}}$ and ${\Phi }_{\mathrm{b}\mathrm{S}}^{\mathrm{p}}$ on the metal/GaSe(0001) interfaces are the band gap E_g = 2 eV at T = 300 K, the electron affinity EA = 3.6 eV, and the charge neutrality level position with respect to the VBM: CNL_v = E_v + 0.83 eV (CNL_abs = 4.77 eV) [38]. According to the calculations using Equation (8), the barrier height for holes at the Sc/GaSe interface constitutes 1.12 eV, assuming the work function Φ_m = 3.5 eV for Sc. For the Pt/GaSe interface (the Pt work function is Φ_m = 5.65 eV), ${\Phi }_{\mathrm{b}\mathrm{S}}^{\mathrm{p}}$= 0.6 eV.

The measured I–V characteristics and the results of the fittings by Equations (4)–(7) are shown in Figure 3a. It can be noted that the obtained data for the barrier height at the Sc/GaSe interface are in good agreement with the calculation result using Equation (8), especially for the case of the unirradiated sample. It can also be noted that the electrical resistance of the sample, R_s, increased after electron irradiation. This agrees with the CNL position calculated in ref. [38], since, at irradiation, the Fermi level moves towards CNL, and as in the unirradiated GaSe crystals, it is usually located at 0.1–0.3 eV above the VBM, and the acceptor compensation increases.

The measured C–V characteristics of the GaSe samples under study are presented in Figure 2b. In the case of a Schottky barrier, the C–V curves are described by the equation [36]

$$U_{\mathrm{c}}-U=e\epsilon {\epsilon }_0{N}_a{S}^2/2C^2$$

(10)

where N_a is the acceptor concentration, ε₀ is vacuum permittivity, ε is the low-frequency relative permittivity of the crystal, U_c is the contact potential difference, U is the applied voltage, and C is the Schottky barrier capacitance. Figure 3b shows the fittings to the experimental data by Equation (10).

As in the case of the I–V characteristics, good agreement can be noted between the Schottky barrier height determined from Equation (8) and that found from the experimental data fitting by Equation (10) (for the Sc/as-grown GaSe interface). At the same time, the calculated acceptor concentrations in the irradiated sample are much lower than the corresponding values obtained from the optical data (cf. the N values in Figure 2c and Figure 3b). Partly, this can be explained by the lower electron fluence 1 × 10¹⁷ cm⁻² for the sample used in electrical measurements and by an unintentional annealing of defects during a contact deposition at a temperature about 300 °C, but the predominant reason is probably due to the inherent differences and limitations of the optical and electrical techniques used, as well as of the employed models. Nevertheless, the obtained N_a values were further used, as is usually done, to estimate the defect concentrations responsible for the appearance of DLTS peaks.

DLTS is a commonly used technique for determining the electrical parameters of defects in semiconductors. Therefore, it is reasonable to compare the results of a study using optical absorption spectroscopy with DLTS data. Standard DLTS analysis is based on registering capacitance transients after the end of the filling voltage pulse, and the emission rate changes with the temperature (determining the emission rates at a number of temperatures allows to obtain an Arrhenius plot like in Figure 4). This allows determining the defect energy level with respect to the conduction or valence band. It is clear that the Arrhenius plot shows the ΔE_a position, which is assumed to be temperature-independent [19,20]. Like any experimental method, DLTS has its limitations. It is known that it, for example, hardly resolves closely lying energy levels with close emission rates. Laplace DLTS analysis can help in this case [20]. Maybe a set of capacitance transients measured in a narrow temperature range would allow to relate the corresponding ΔE_a value to a distinct temperature. More reliable interpretation of DLTS activation energies requires first principles calculations to reveal the possible contribution from lattice reconstruction and surmounting of temperature-dependent capture barriers [39].

The time constants τ, as is known [19,20], are inversely proportional to the emission rate of the charge carriers (holes):

$$e_{\mathrm{p}}=1/\tau =\ln (t_1/t_2)/(t_1-t_2)$$

(11)

where t₁ and t₂ are time delays from the back edge of the filling pulse (t₁ < t₂).

The obtained DLTS spectra are shown in Figure 4a,b. It is possible to note the presence of a broad peak in the temperature range of 200–220 K for the unirradiated GaSe. In the DLTS spectra of the GaSe crystal irradiated with high-energy electrons, this peak shifts toward higher temperatures and becomes distorted. At the same time, a peak of lower intensity is observed at a temperature of about 370 K, which is clearly not visible in the spectra of the unirradiated crystal.

When processing the data following the conventional procedure [19,20,37], it was assumed that

$$e_{\mathrm{p}}={\sigma }_{\mathrm{p}}{v}_{\mathrm{p}}{N}_{\mathrm{v}}\exp (-\Delta E/kT)$$

(12)

where ν_p is the thermal velocity of holes (~T^1/2), σ_p is the hole capture cross-section, N_v is the effective density of the states for the top of the valence band (~T^3/2), and ΔE is the activation energy of the trap. The defect concentrations were estimated as [19,20,37]

$$N_{\mathrm{t}}=2N_{\mathrm{a}}(\Delta C/C_0)$$

(13)

where ΔC is the difference in the structure capacitance at the beginning and the end of the capacitance transient, and C₀ is the structure capacitance during the filling pulse. The results of the DLTS data processing are presented in Figure 4c,d. For the irradiated GaSe sample, the obtained defect activation energies are in good agreement with the energy positions of the levels deduced from the optical transmission measurements at 300 K (Figure 1b and ref. [18]). The physical meaning of the determined defect-associated “energy positions” is, in the case of DLTS, closely related to the charge transition levels, while the optical transition energies are more directly related to the energy intervals between the levels in the forbidden gap. Nevertheless, our measurement values are close.

It should be noted that the DLTS spectra obtained differed from those reported earlier in other studies. The most similar results were obtained in ref. [25] at photoexcitation of the sample. The spectrum registered in ref. [25] for a specially undoped GaSe sample contained two peaks, corresponding to the defect levels at E_v + 0.2 eV and E_v + 0.8 eV. This resembles the spectrum obtained in our work for the electron-irradiated crystal. At the same time, in ref. [28], using a similar technique, five peaks were found for high-resistivity GaSe samples. The DLTS spectra for GaSe:Cd crystals demonstrated in ref. [37] contained a single broad peak in the temperature range of 250–270 K, which was similar to the data presented in Figure 4a. This peak was attributed by the authors to the presence of defects including a Cd impurity, but its origin could also be connected with the formation of intrinsic lattice defects as a result of the introduction of cadmium atoms. Finally, four well-resolved DLTS peaks were observed for a GaSe:In sample in the temperature range from 100 to 350 K in ref. [27]; the defect level with the energy position E_v + 0.22 eV was associated with the presence of a Ga vacancy, while the defect level at E_v + 0.44 eV was attributed to a complex of V_Ga and an indium impurity atom.

A detailed analysis of the possible nature and differences of DLTS peaks in various GaSe samples was provided in ref. [28], where a peak related to Ga vacancy appeared at 0.17 eV above the VBM. The same peak was observed in ref. [25]. An additional peak at 100–115 K was recorded at 0.13 eV below the conduction band minimum. The used method (with photoexcitation) did not allow distinguishing between traps for electrons and holes, but the author supposed that this was a trap for minority charge carriers (electrons). Here, it is important to note that a similar peak was seen in our samples when the filling pulse changed the sign of the total applied voltage (its amplitude was higher than that of the reverse bias pulse). On the other hand, the amplitude of this peak was small, and its maximum did not show clear rate window dependence. Therefore, this data were not analyzed in the present paper. In ref. [28], was is also stated that the defect level at 0.53 eV was related to defect complexes of V_Ga-metal impurity atoms, justifying this by the fact that close values were found in metal-doped GaSe crystals [27,37,40,41]. It is also mentioned in ref. [25] that this peak was not detected, which was explained by a higher resistivity of the sample (less amount of uncontrolled impurity elements). In ref. [28], the defect level at energy of 0.75 eV, which was close to 0.68 eV found in the present study, was associated with extended defects, like dislocations or stacking faults. It was justified by high values of captured cross-sections.

Thus, it can be concluded that the origin of crystals and post-growth treatment have a strong impact on their defect structure (registered DLTS spectra). It can be also pointed out that a similar picture of the DLTS peaks related to the divacancies in two charge states was previously observed in high-energy particle irradiated Si [42].

4. Conclusions

In conclusion, optical absorption spectra of 9 MeV electron-irradiated GaSe crystals were analyzed. By fitting the experimental data measured at various temperatures in the range 9.5–300 K with the model equation, taking into account spectral dependencies of deep-level photoionization cross-section and fundamental direct interband transitions, the temperature-dependent defect ionization energies, concentrations of the Ga vacancies in two charge states, and GaSe band gap were determined.

DLTS spectra were measured on as-grown and electron-irradiated GaSe slabs with barrier (Sc) and ohmic (Pt) contacts. We estimated the barrier height for the Sc/GaSe interface, as well as defect activation energies, defect concentrations, and hole capture cross-sections. For the electron-irradiated GaSe crystals, activation energies quite close to the values determined from optical measurements were obtained.

Thus, in the present work, inferred from the optical absorption measurements, the presence of defect energy levels in electron-irradiated GaSe samples at 0.23 and 0.66 eV above the VBM is confirmed by DLTS. The optical absorption measurements allow determining the threshold energies for optical transitions, which can take place both from the valence band to a defect level or from a deep level to a conduction band. The optical absorption data, together with DLTS data, are helpful to clear the real picture of optical transitions. Note that DLTS technique does not allow to see the temperature dependence of the defect charge transition levels. In this regard, the optical absorption measurements are more informative. Employing the latter method, the clear temperature dependence of the energy levels related to $V_{Ga}^{-1}$ and $V_{Ga}^{-2}$ defects in electron-irradiated GaSe was revealed. The observed behavior of temperature-dependent energy positions of the defect levels agreed with the “rules” provided in ref. [32], assuming GaSe to be a semiconductor with a valence band increasing in energy at decreasing temperatures and low (large) change of the local volume around a defect for the charge transition −/−2 (0/−1).

In future work, a Laplace DLTS technique is also planned to be used for a more detailed study of broad “distorted” DLTS peaks observed in the irradiated GaSe samples.