Skip to main content
Scientific Reports logoLink to Scientific Reports
. 2014 Aug 21;4:6143. doi: 10.1038/srep06143

Optically decomposed near-band-edge structure and excitonic transitions in Ga2S3

Ching-Hwa Ho 1,2,a, Hsin-Hung Chen 1
PMCID: PMC4139941  PMID: 25142550

Abstract

The band-edge structure and band gap are key parameters for a functional chalcogenide semiconductor to its applications in optoelectronics, nanoelectronics, and photonics devices. Here, we firstly demonstrate the complete study of experimental band-edge structure and excitonic transitions of monoclinic digallium trisulfide (Ga2S3) using photoluminescence (PL), thermoreflectance (TR), and optical absorption measurements at low and room temperatures. According to the experimental results of optical measurements, three band-edge transitions of EA = 3.052 eV, EB = 3.240 eV, and EC = 3.328 eV are respectively determined and they are proven to construct the main band-edge structure of Ga2S3. Distinctly optical-anisotropic behaviors by orientation- and polarization-dependent TR measurements are, respectively, relevant to distinguish the origins of the EA, EB, and EC transitions. The results indicated that the three band-edge transitions are coming from different origins. Low-temperature PL results show defect emissions, bound-exciton and free-exciton luminescences in the radiation spectra of Ga2S3. The below-band-edge transitions are respectively characterized. On the basis of experimental analyses, the optical property of near-band-edge structure and excitonic transitions in the monoclinic Ga2S3 crystal is revealed.


Ga2S3 is an important member of III-VI compounds (i.e. III: In, Ga, and VI: S, Se, Te), which may possess the widest band gap. Dissimilar to the other III-V compounds (e.g. GaAs and InP) having the strongest covalent bond, the misvalency between the III and VI atoms usually renders a III-VI compound possessing different stoichiometries, diversified crystal phases, and various lattice forms1,2,3. For the gallium chalcogenides, GaS and Ga2S3 are the general constituents existed in the gallium sulfides4,5 while the GaSe and Ga2Se3 are the common stoichiometric elements formed in the gallium selenides6,7. The GaSe and GaS compounds may crystallize in the hexagonal layered structure but possess different stacking formula (i.e. GaSe is in ε stacking phase and GaS in β polytype)8. The stacking deviation in GaX (X = S, Se) will make GaSe a direct semiconductor with an energy gap close to 2 eV whereas the GaS material becomes an indirect semiconductor with an indirect gap of ~2.53 eV4. Ga2Se3 may possess a defective zinc-blende structure in which one-third part of cation sites are vacant randomly in the lattice7. It is naturally a defect semiconductor with a direct band gap around 2–2.4 eV9,10. All the gallium chalcogenides GaSe, GaS, and Ga2Se3 have the values of band gaps below 2.55 eV, which can only be catalogued into the visible-range materials, not for the blue to UV applications. Ga2S3 is also a defect semiconductor with various existing phases of monoclinic5,11,12, hexagonal3, and cubic13,14 owing to the misvalency of the III-Ga and VI-S elements. The most stable and generally found crystal structure of Ga2S3 is the monoclinic phase. Previous studies claimed that the doped (Cr and Fe) and undoped Ga2S3 can be a luminescent material with emitting wavelengths ranging from near infrared to blue portion15,16. A gold-doped chalcogenide glass containing the Ga2S3 nanocrystals can be applied in the third-order nonlinear optics17. Ga2S3 is a wide-band-gap semiconductor, however, to date, discrepant values of band gaps ranging from 2.5 to 3.4 eV have ever been found in the literatures18,19. This result is owing to the uncertainty on the crystal quality and the lack of knowledge to the optically decomposed experimental band-edge structure of Ga2S3.

In this paper, we demonstrate the detailed characterization of near-band-edge structure (below and above band gap) of high-quality Ga2S3 crystals using optical techniques of polarized thermoreflectance (PTR), photoluminescence (PL), and optical-absorption measurements in the temperature range between 15 and 300 K. Single crystals of Ga2S3 were grown by chemical vapor transport (CVT) method using ICl3 as the transport agent. The as-grown crystals are essentially transparent and light-yellow colored. High-resolution transmission electron microscopy (HRTEM), X-ray diffraction (XRD), and Raman measurements confirmed monoclinic phase of the as-grown Ga2S3 crystals. The lattice parameters and crystal structure are determined. The below-band-edge emissions of Ga2S3 including defects, bound-exciton (BX), and free-exciton (FX) luminescences are respectively probed by PL and temperature-dependent PL measurements. The low-temperature TR spectra clearly show four band-edge transitions denoted as EA, EB, EC1, and EC2 detected in the vicinity of the band gap. The EA transition is the direct band gap of Ga2S3 which is determined to be ~3.052 eV at room temperature. The EB, EC1, and EC2 transitions are the band-edge excitons and they respectively demonstrate orientation and polarization dependences of anisotropic characters with respect to the c plane of Ga2S3. The EB exciton behaves like a free-exciton emission (PL) from the c-plane Ga2S3 while the EC1 and EC2 transitions belong to an excitonic series with an effective Rydberg constant (i.e. binding energy) of about Ryd ≈ 68 meV. Photo-voltage-current (Photo V-I) measurements of the Ga2S3 sample under different illumination conditions of dark, halgen lamp, and 405-nm laser are, respectively, carried out to evaluate the photoelectric-conversion behavior of the Ga2S3 crystal. On the basis of the experimental analyses of PL, TR, PTR, and Photo V-I measurements, the near-band-edge transitions and excitonic structures of the monoclinic Ga2S3 are characterized. The direct-gap chalcogenide can be a functional material suitable for white-light emitting device and UV solar-energy conversion cell.

Results

Crystallinity and crystal structure

Displayed on the left side of Fig. 1(a) is the HRTEM image of the as-grown Ga2S3 crystal on the c plane. Regular and clear atomic sites for the periodically-arranged lattice of the c-plane Ga2S3 can be observed in HRTEM image, indicating that excellent crystallinity of the Ga2S3 can be obtained by the CVT growth with ICl3 as the transport agent. The better crystal quality of Ga2S3 can also be evident in the upper-right side of Fig. 1(a) by a fast Fourier transformation (FFT) image of the HRTEM picture for revealing distinct and clear dot pattern (i.e. no rings formation). The selected-area electron diffraction (SAED) pattern of the Ga2S3 with the zone axis along [001] is also shown in the right side of Fig. 1(a). The apparent and obvious dot spots of the SAED pattern of Ga2S3 verifies high crystalline quality of the crystal. The dot spots also represent several sets of complete crystalline planes of Ga2S3, e.g. the (200) spot as indicated in the SAED pattern of Fig. 1(a). As shown in the HRTEM image in Fig. 1(a), the atomic arrangements of the c-plane Ga2S3 can allow us to distinguish the a and b axes of the crystal plane. Depicted by the green-arrow lines in the HRTEM image are the indications of the a and b axes on the c-plane Ga2S3. The angle lying in between the a and b axes is γ (≈140°). From the HRTEM image of Fig. 1(a), the estimate of inter-planar spacing of d(200) is about 0.55 nm (i.e. a = 1.1 nm) while the lattice spacing of the b axis can be determined to be d(010) = 2 d(020) ≈ 2 × [0.31 nm/cos(50°)] = 0.96 nm. The crystal symmetry of the Ga2S3 crystal observed by the HRTEM image of the c plane in Fig. 1(a) seems to be a monoclinic phase. It can also be verified by the XRD measurement [see supplementary information (SI)]. The lattice constants from the XRD result are determined to be a = 1.111 nm, b = 0.958 nm, c = 0.64 nm, and γ = 141.15°, respectively. The obtained values of a, b, and γ agree well with the lattice spacing determined by the HRTEM image as evident in Fig. 1(a).

Figure 1.

Figure 1

(a) HRTEM images of Ga2S3 on the c plane. The orientations of a and b axes are indicated. The overall FFT pattern and the SAED image with the zone axis of <001> are also included for comparison. (b) The representative scheme of atomic layer arrangement of the Ga sheet in the Ga2S3 defect semiconductor. The one-third Ga vacancies in the Ga sheet are also depicted. The easily forming interfacial defects (IF) along the a axis on the c plane is indicated (left side). The right side shows the schema of real tetrahedra of sp3 structure of the Ga and S bonding (i.e. GaS4 tetrahedral molecular) in Ga2S3.

The crystalline phases of Ga2S3 found in the literatures may be monoclinic, cubic, and hexagonal structures12,13,14. However, their main XRD peak positions are nearly the same with the significant difference only being in the relative intensity change of the diffraction peaks in these phases (e.g. Ga2S3 film on the GaAs substrate)13. It means that even in the monoclinic phase, Ga2S3 can also have a nearly layered type with the hexagonal (defect wurtzite) thin-film form. In fact, the III-VI Ga2S3 is a naturally defect semiconductor with one-third of cation sites are vacant (Ga vacancies), similar to that of γ-In2Se320. The left side of Fig. 1(b) depicts the representative scheme of the atomic arrangement in the metal (Ga)-sheet of the Ga2S3 crystal, where 1/3 of the Ga vacancies (“□”) appeared, and the □ repeats throughout the lattice periodically along a axis. The right side of Fig. 1(b) shows the real tetrahedra of sp3 structure of the Ga and S bonding (i.e. GaS4 tetrahedral molecular) in the Ga2S3 crystal above the Ga metal sheet. The dashed-enclosed line represents the area of the a and b axes as indicated in the left side of Fig. 1(b) by Inline graphic and Inline graphic. This kind of defect structure is easy to form an interfacial disorder of metal atoms in some specific interfaces (IF), which may present in some interfacial lines along the a axis shown in the left side of Fig. 1(b) with a blue arrow on the c plane12. The easily-forming IF along the a axis in the defect semiconductor Ga2S3 may render an in-plane anisotropy of the c-plane Ga2S3 with the presence of a polarization-dependent optical character along and perpendicular to the crystal's a axis, and we will show and discuss the experimental result later.

To further verify the structural property of the Ga2S3 crystals, Raman measurement of the c-plane Ga2S3 was implemented. Figure SI-2(a) shows the Raman spectrum of the Ga2S3 in the energy range of 200-500 cm−1. There are seven peak features at 234, 282, 309, 331, 348, 387, and 427 cm−1 detected in the Ga2S3 crystal. Most of the frequencies can be assigned in terms of internal and external vibrations of tetrahedral GaS4 groups (see SI)21,22. The Raman modes sustain the monoclinic crystal symmetry of the as-grown Ga2S3 crystals.

Luminescence properties of Ga2S3 below band edge

Shown in the left side of Fig. 2(a) is the low-temperature and wide-range PL spectrum of Ga2S3 (i.e. 1.25–3.8 eV by a lower-resolution spectrometer with the focal length of ~101 mm) on the c plane at 15 K. The excitation source is a 325-nm He-Cd laser. Three main bands respectively denoted as D1, D2, and BE emissions are detected. The D1 (~1.99 eV) and D2 (~2.79 eV) bands are the defect related luminescences originated from the donors and acceptors in the Ga2S3 crystals while the BE band (~3.36 eV) must be the band-edge emission. In comparison with the previous results of crystalline and thin-film forms of the Ga2S3 in the literatures, it is maybe a near-band-edge emission coming from the Ga2S3 crystal13,23. The appearance of the band-edge emission BE sustains a better crystalline quality of the as-grown Ga2S3 crystal. The PL intensities of the D1 and D2 emissions below band edge are much stronger than that of the band-edge emission BE observed at 15K. It is an indication that the monoclinic Ga2S3 is a naturally defect semiconductor with the existence of intrinsic defects coming from the imperfection levels consisted in the band gap. Even in such a high-quality crystal (as evident in the HRTEM result of Fig. 1 and the appearance of BE emission in Fig. 2), the defect emissions are still stronger than that of the band edge emission.

Figure 2.

Figure 2

(a) Low-temperature and wide-range PL spectrum of the monoclinic Ga2S3 on the c plane at 15 K. The temperature dependent PL spectra of Ga2S3 in the temperature range between 15 and 300 K are also shown. (b) The possible transition mechanism for the defects and band-edge emissions in Ga2S3 at low and room temperatures.

To further evaluate the physical origin of the PL emissions of the Ga2S3, temperature-dependent PL measurements were carried out. Figure 2(a) also shows a three-dimensional (3D) plot of the relative intensity change and peak-energy shift of the D1, D2, and BE emissions obtained by temperature-dependent PL measurements of Ga2S3 in the temperature range between 15 and 300 K. With the temperature increases, the amplitude of the D1 emission was gradually decreased from 15 to 200 K while the D2 emission revealed similar behavior with D1 at T < 180 K. When T > 180 K, the PL intensities of D1 and D2 are comparable from ~200 to ~240 K, and then the D2 emission may increase a little bit, and finally the D1 and D2 peaks are merged together to form a broadened peak (centered at ~2.4 eV) at room temperature. The mechanism of the defect bands inside the Ga2S3 defect semiconductor is maybe coming from the formation of sulfur vacancies (VS) by native chalcogen deficiency in the crystal growth as well as the existence of Ga vacancies (VGa) inside the imperfection crystal structure. The representative band-edge schemes of monoclinic Ga2S3 under room temperature (RT) and at low temperature (LT) are depicted in Fig. 2(b). For the below-band-edge portion, the VS states may form a defect donor band (ED) in the Ga2S3 while the VGa states may still from some acceptor levels (e.g. EA1 and EA2) existed inside the band gap. The D1 and D2 emissions are, respectively, inferred to come from the lowest ED band to the acceptor levels at higher VGa (EA2) and lower VGa (EA1) states as shown in Fig. 2(b). At low temperatures (LT), the higher EA2 level is nearly empty but the lower EA1 level is almost occupied by electrons to render higher D1 but lower D2 intensities. When the temperature increases, thermal ionization effect between the lower EA1 and higher EA2 acceptor levels causes the D1 emission decreased but the D2 enhanced a little bit as shown in the PL result and the indications in Fig. 2(b). For the band-edge emission BE, the PL intensity in Fig. 2(a) is weaker than those of D1 and D2, and the BE is not observable at T = 300 K. We will evaluate and study the band-edge emission BE by using the temperature-dependent PL measurement in a high-resolution spectrometer with a focal length of 550 mm later.

Figure 3 shows the temperature-dependent PL spectra (i.e. high resolution) of the c-plane Ga2S3 in the temperature range between 10 and 150 K for observation of the optical behavior in the band-edge BE emission. It is clear and relevant that the BE emission consists of lower bound-exciton (BX) and higher free-exciton (FX) luminescences near the band edge of Ga2S3, such as the excitonic transitions of other direct oxide semiconductor as In2O324. As shown in Fig. 3, the intensity of the BX feature is higher than that of the FX feature at low temperatures. The energy positions of the excitons are BX≈3.363 eV and FX≈3.378 eV at 10 K. When the temperature increases, the intensity of the BX peak decreases rapidly while the FX feature decays slowly. The temperature-energy shift behavior of the FX is shown faster than that of BX and finally they will merge together and dominated by the FX peak at even higher temperatures. This is a general behavior of the bound and free excitons in a direct semiconductor with high luminescent efficiency. The prominent BX feature at low temperatures is probably owing to the peak bound by a neutral donor from the defect donor band ED formed by chalcogen deficiency (i.e. VS). This situation is similar to the donor levels formed by sulfur vacancies to render a donor-bound D0X emission in a nanostructure cadmium sulfide material25. The FX feature in Fig. 3 is closely related to one of the band-edge excitons existed in the monoclinic Ga2S3 and we will evaluate and discuss the direct band-edge structure later.

Figure 3. Temperature dependence of high-resolution PL spectra of bound exciton (BX) and free exciton (FX) emissions in the Ga2S3 crystal between 10 and 150 K.

Figure 3

Discussion

Optical investigation of the band-edge structure of Ga2S3

Figure 4(a) shows the temperature-dependent (unpolarized) TR spectra of Ga2S3 on the c plane in the temperature range between 20 and 300 K. At 20 K, there are four band-edge transitions denoted as EA, EB, EC1, and EC2 can be detected in the unpolarized TR spectrum of the c-plane Ga2S3. The energy values of the band-edge transitions at each temperature can be analyzed by using least-square fits of the experimental data to a derivative Lorentzian line-shape functional form expressed as26:

graphic file with name srep06143-m1.jpg

where i = A, B, C1, and C2, and I and ϕ are the amplitude and phase of the line shape, and Ei and Γ are the energy and broadening parameter of the respective interband transition. The value of m = 2 is used for the first derivative line shape analysis of the band-edge excitons of Ga2S3. The obtained values of transition energies from the line-shape fitting of equation (1) are indicated with arrows and their variation traces of temperature-energy shift are shown by dotted lines in Fig. 4(a). The transition energies at 20 K are EA = 3.256 ± 0.010 eV, EB = 3.386 ± 0.008 eV, EC1 = 3.440 ± 0.005 eV, and EC2 = 3.491 ± 0.005 eV, respectively. With the increase of temperatures, the TR transition features demonstrate an energy red-shift behavior and a line-width broadened character such as the general semiconductor behavior for the c-plane Ga2S3. The EA feature is assigned as the direct band gap of Ga2S3, and EB and EC features are correlated with the excitonic transitions near band edge. The origins of the EA, EB, and EC transitions are different and the assignments of band-edge transitions of Ga2S3 will be verified and identified by the PTR, PL and transmission measurements later. As shown in the low-temperature TR spectra of Fig. 4(a), the most prominent feature of the c-plane Ga2S3 is the EC1 feature (n = 1) together with a higher-exciton level EC2 feature (n = 2) is also detected. The EC1 and EC2 transitions may be an excitonic series coming from the c-plane Ga2S3. The excitonic sequence of the EC features of Ga2S3 can be further analyzed by using Rydberg series27:

graphic file with name srep06143-m2.jpg

where Ryd is the effective Rydberg constant (binding energy) and E is the threshold energy of the excitonic sequence (EC). From the obtained values of transition energies of EC1 = 3.440 and EC2 = 3.491 eV at 20 K [see Fig. 1(a)], the effective Rydberg constant is determined to be Ryd = 68 meV and threshold energy is about E = 3.508 eV at 20 K. It also means that the binding energy of the EC2 (n = 2) feature is about 17 meV (i.e. Eb = 68/22 meV). This result can also be verified by the temperature dependence of the TR feature of EC2 that was thermally ionized between 180 K and 220 K as evident in Fig. 4(a) (i.e. thermal ionization energy is kT = 17 meV, and the corresponding thermal temperature is about T = 197 K).

Figure 4.

Figure 4

(a) Temperature-dependent TR spectra of the c-plane Ga2S3 near band edge between 20 and 300 K. Four band-edge transitions denoted as EA, EB, EC1, and EC2 are observed at low temperatures. (b) Temperature dependence of transition energies of the EA, EB, EC1, and EC2 features. The solid lines are the Bose-Einstein type fits to the experimental data from 20 to 300 K.

Figure 4(b) depicts the temperature dependence of transition energies of EA-EC2 features obtained by Fig. 4(a) for the Ga2S3 crystal by TR. The hollow squares, open circles, hollow triangles, and solid triangles are the energy values of EA, EB, EC1, and EC2 features and the solid lines are the least-square fits to a Bose-Einstein type expression responsible for the temperature-energy shift of the EA-EC2 in Ga2S3 expressed as Ei(T) = Ei0 - Si<ħΩi>[(coth<ħΩi/kT>) - 1], where i = A, B, C1 or C2, Ei0 is the transition energy at 0 K, Si is a dimensionless coupling constant related to the strength of electron-phonon interaction, and <ħΩi> is an average phonon energy. The obtained fitting parameters are EA0 = 3.255 ± 0.005, EB0 = 3.384 ± 0.007, EC10 = 3.435 ± 0.004, and EC20 = 3.491 ± 0.003 eV for Ei0, SA = 15 ± 5, SB = 10 ± 2, SC1 = 7 ± 2, and SC2 = 7 ± 2 for Si, and <ħΩA> = 15 ± 6, <ħΩB> = 13 ± 5, <ħΩC1> = 13 ± 5, and <ħΩC2> = 13 ± 5 meV for the <ħΩi>, respectively, available for the EA, EB, EC1, and EC2 transitions. The value of electron (exciton)-phonon interaction constant S for the EA, EB, and EC transitions shows somewhat different in Ga2S3. It implies that the variation speeds of temperature-energy shift of the EA, EB, and EC transitions are dissimilar. The EA, EB, and EC transitions near band edge are maybe coming from different origins in Ga2S3.

To further characterize optical anisotropy of the EA-EC transitions in Ga2S3, orientation-dependent and polarization-dependent TR measurements of Ga2S3 are respectively carried out. Fig. 5 shows the orientation-dependent TR spectra of (a) 20 K and (b) 300 K for the monoclinic Ga2S3 near band edge, together with a low-temperature PL spectrum at 20 K is also included in Fig. 5(a) for comparison. The orientation-dependent TR measurement was implemented with the probed monochromatic beam impinged on the c plane or on the tilt c plane of Ga2S3 as the indications shown in the insets of Fig. 5. The tilt c plane was performed by a tilt polishing of the Ga2S3 sample. As shown in Fig. 5(a), for the TR results of Ga2S3, we can separate the features into two groups for the EA, EB, EC1, and EC2 transitions. One main part consists of the EA feature and the other part contains both the EB and EC transitions. For the c-plane TR result of Ga2S3 in Fig. 5(a) (i.e. 20 K, with solid line), the transition amplitudes of EB and EC are larger than that of EA, whereas, the tilt c-plane TR intensity of the EA feature is stronger than those of the other EB and EC features as evident in Fig. 5(a). The anisotropic effect of orientation-dependent TR measurements can also apply to the room-temperature TR spectra as evident in Fig. 5(b). The EB and EC excitonic transitions are mainly dominant on the c plane and they will decrease their intensities when the c plane is tilted. In fact, the amplitudes of the EB and EC1 features in the c-plane TR spectrum are comparable, whereas the EB transition shows much smaller intensity than that of the EC1 as displayed in Fig. 5(a). This phenomenon is owing to the partially polarizing effect of the incident light coming from the grating inside the monochromator. The comparable intensity of EB and EC1 at low temperature will be verified by polarization-dependent TR measurement later. The orientation-dependent TR results in Figs. 5(a) and 5(b) provide conclusive evidence that the transition origin of EA is different from those of the other EB and EC transitions. The EB and EC are dominated on the c plane while the EA is orientation dependent presented for the monoclinic Ga2S3. As shown in the c-plane PL spectrum in Fig. 5(a), the FX emission nearly agrees well with the EB transition and the EA feature is maybe too broad to observe in the PL spectrum at 20 K. The BX is the most prominent peak at ~3.362 eV and an additional BX-LO peak (longitudinal-optical-phonon replica) can be detected at ~3.333 eV to render a LO phonon energy approximately to be 29 meV. The LO phonon energy matches well with the main Raman vibration branch of 234 cm−1 (A1 mode) as detected in Fig. SI–2.

Figure 5. Orientation dependent TR measurements on the c plane and tilt c plane (by grinding and polishing sample) for Ga2S3 at (a) 20 K and (b) 300 K.

Figure 5

A low-temperature PL spectrum at 20 K is also included in (a) for comparison.

In order to distinguish the transition natures of the EB and EC features, PTR measurements are carried out on the c plane of Ga2S3. Fig. 6 shows the PTR spectra near the vicinity of band edge at (a) 300 K and (b) 50 K, respectively. The measurement configuration of the PTR experiment is presented in Fig. 6(c). A pair of linear polarizers was used, and the PTR experiments were carried out with the linearly polarized light parallel and perpendicular to the Ga2S3 crystal's a axis (i.e. E || a and E ⊥ a). The a axis is easy to form the planar interfaces (IF) as the indication shown in Fig. 1(b). As shown in Figs. 6(a) and 6(b), the PTR measurements (done on the c plane) of Ga2S3 clearly show that EB transition is present only in the E || a polarization while the EC excitons are merely allowed at the E ⊥ a polarization. The unpolarized spectrum is approximately a random superposition of the E || a and E ⊥ a polarized spectra of Ga2S3. The intensity of the EA feature (E || a and E ⊥ a) in the PTR spectra at 300 and 50 K (see Fig. 6) shows similar. The origin of EA is different from those of the EB and EC transitions on the c plane. However, the polarization dependence of the EB and EC transitions on the c plane of Ga2S3 also indicates that EB and EC are also coming from different origins. The polarization dependency and in-plane anisotropy of the band-edge transitions along some specific crystal axis with line structures (i.e. IF along a axis) have also been found in a triclinic ReS2 layered crystal with the existing Re4 diamond chain clusters along the crystal's b axis27,28. The solid lines shown in Figs. 6(a) and 6(b) are the derivative Lorentzian line-shape fits of the PTR spectra using equation (1) and the obtained transition energies are EA = 3.052 eV, EB = 3.240 eV, and EC1 = 3.328 eV at 300 K, and EA = 3.252 eV, EB = 3.385 eV, EC1 = 3.437 eV, and EC2 = 3.488 eV at 50 K, respectively. The PTR transition amplitudes of EB and EC1 at 50 K show comparable in Fig. 6(b). It verifies that the transition probabilities of EB and EC are comparable in Ga2S3, regardless of an observed smaller intensity of EB was detected in the unpolarized TR spectra due to the partially polarizing effect of the monochromator system in Fig. 5(a). The orientation-dependent and polarization-dependent TR measurements confirmed that all the band-edge transitions of EA, EB, and EC features are coming from different origins and an asymmetric valence-band top may account for the observed optical-anisotropic effects of the monoclinic Ga2S3.

Figure 6. Polarization-dependent TR spectra of (a) 300 K and (b) 50 K for the c-plane Ga2S3 near band edge.

Figure 6

The measurements are done with the linearly polarized light parallel and perpendicular to the crystal's a axis. (c) The representative scheme for the c-plane PTR measurement of Ga2S3.

For the monoclinic Ga2S3, the highest valence band is consisted of mainly S 3p and some Ga 4p orbitals29. Unlike the other s orbital has highly spherical symmetry, the p states in the valence band of Ga2S3 are strongly oriented dump-bell shape (axial dependent) distribution, which may enhance the anisotropic character present in the optical property of the diindium trisulfide. The oriented- and polarized-TR spectra may therefore present strongly anisotropic character in the band-edge transitions of the EA, EB, and EC features as displayed in the Figs. 5 and 6. For the lowest conduction-band portion of Ga2S3, the density of states (DOS) are mainly composed of Ga 4s and a little S 3p DOS over 1.7–3.4 eV (see the calculated DOS in Fig. SI-3). For the higher-energy portion (3.4–10 eV), the conduction band is consisted of mainly Ga 4p and partially S 3p. For the lowest energy states of the Ga 4s doublet, the energy range is about from -8 to -6 eV below the main valence band. The main valence band is consisted of S 3p and Ga 4p electrons. To further indentify the band gap nature and band-edge transitions of the Ga2S3, transmission measurement with the incident light impinged on the c plane of a thin Ga2S3 sample was carried out. Figure 7(a) shows the transmittance (T) spectrum of the Ga2S3 sample with energy range close to the band-edge portion. For comparison purpose the derivative T spectrum (Derv. T) is calculated, together with the TR spectrum of the c-plane Ga2S3 is also included in Fig. 7(a) for contrast. The absorption edge in the T spectrum of Fig. 7(a) shows two stair steps positioned in between 2.8 and 3.4 eV at 300 K. It is very interesting that the EA and EB transitions in TR are matching well with the Derv. T spectrum calculated from T (i.e. TR is also derivative line shape). It suggests that EA and EB are two of the direct band-edge transitions in Ga2S3. The calculated absorption coefficient α in the inset of Fig. 7(b) also sustains that the spectral analysis of both EA and EB followed α ∝ (E-Ei)1/2, a direct allowed transition (i = A or B). The converted spectrum of (αhν)2 versus hν plot shown in Fig. 7(b) also reveals that the linear fitting results get the energy values of EA = 3.05 eV and EB = 3.24 eV for the monoclinic Ga2S3, respectively. The energy values of EA and EB match quite well with those obtained by TR and PTR measurements as evident in the Figs. 57 at room temperature. The TR technique is a powerful tool for probing direct band gap and direct inter-band transitions of semiconductors24,30. The energy positions of the band-edge transitions in TR show good agreement with the absorption edge of the transmittance spectrum of Ga2S3. This result sustains that Ga2S3 is a direct semiconductor with a direct band gap positioned at EA≈3.052 eV at room temperature. With this energy value, the Ga2S3 may be a potential material available for fabrication of white lighting device and a UV photoelectric conversion cell.

Figure 7.

Figure 7

(a) Transmittance, derivative transmittance, and TR spectra of Ga2S3 near band edge. (b) The absorption spectrum and the result of (αhν)2 vs. hν for determining the transition energies of EA and EB.

To evaluate the UV photoconduction behavior of the Ga2S3 crystal, Photo V-I measurements are implemented at 300 K. Figure 8 shows the Photo V-I measurement results of the Ga2S3 crystal under different illumination conditions of dark, halogen lamp, and 405-nm laser. The measurement setup is depicted in the lower inset of Fig. 8. The illumination power density for the halogen lamp and the 405-nm solid state laser was kept at P≈15 mW·cm−2. The scanning voltage range is from −1000 to 1000 V. Nearly a linear relationship for the Photo V-I curves was obtained (i.e. by a linear fitting). The obtained values of resistivity are ρ = 53.4 GΩ-cm for the dark condition, ρ = 30 GΩ-cm for the halogen lamp, and ρ = 1.15 GΩ-cm for the 405-nm laser illumination, respectively. The resistivities under dark, halogen lamp, and 405-nm laser illumination are respectively depicted as star, hollow diamond, and open triangle in the upper inset of Fig. 8. The halogen lamp represents a broadband blackbody radiation with a main hump peak close to approximate 1.9 eV, which also demonstrates a decreased and lowered intensity toward to the UV range. The photoconduction ratio of halogen lamp in Fig. 8 is about Δρ/ρ = 44%. However, for the 405-nm laser (with the same power density) illumination, the photoconduction ratio can reach Δρ/ρ = 98%. The photon energy of the 405-nm laser is 3.062 eV, which is larger than that of the direct gap (i.e. EA = 3.052 eV) in the Ga2S3 to render a significant photoresponse in the violet to UV region in Fig. 8.

Figure 8. The experimental results of Photo V-I measurement under different illumination conditions of dark, halogen lamp and 405-nm laser for a Ga2S3 photoconductor.

Figure 8

The photo-resistivities and measurement setup are, respectively, depicted in the upper and lower insets.

In summary, the band-edge structure, defect luminescence, and excitonic transition of Ga2S3 crystals have been detailed characterized by PL, optical absorption, orientation-dependent and polarization-dependent TR measurements. The Ga2S3 crystals have been grown by CVT method using ICl3 as the transport agent. HRTEM, XRD and Raman measurements confirmed high quality and monoclinic phase of the as-grown Ga2S3 crystals. The interfacial defects IF are easy to form on the c plane along a axis, which results in optical anisotropy of Ga2S3 with polarizations along a and perpendicular to a axis. Ga2S3 is a naturally defect semiconductor and the defects caused by VGa and VS vacancies render a visible luminescence ranging from 1.5 to 3 eV, which is detected by the PL spectra from low to room temperatures. Especially, the free exciton and bound exciton emissions are firstly observed in the Ga2S3 crystal. There are three main band-edge transitions EA, EB, and EC can be detected in the TR spectra of Ga2S3. The EA transition is the direct band gap of Ga2S3 (i.e. EA = 3.052 eV). The EB and EC transitions are dominated on the c plane and will reveal intensity degradation if the TR measurements are carried out on a tilt c plane by tilt polishing the Ga2S3 sample. The in-plane anisotropy of the c-plane Ga2S3 clear shows that the EB transition merely allows along the E || a polarization while the EC transition is only present in the E ⊥ a polarized spectra. The energy position of EB is matching well with the FX emission of Ga2S3 at low temperature. The EC is an excitonic series (observed by EC1 and EC2 levels at low temperatures) whose binding energy is determined to be Ryd = 68 meV. The highly-oriented S 3p and Ga 4p states consisted in the valence-band top renders the strong optical anisotropic behavior of the band-edge transitions in the Ga2S3. Ga2S3 is a wide band gap semiconductor and the Photo V-I measurements under dark, halogen lamp, and 405-nm laser illumination conditions confirmed blue to UV photoelectric conversion behavior of the Ga2S3 crystal. The sulfide compound is not only a white-light emitting material but also a UV optical absorber.

Methods

Crystal growth

The Ga2S3 crystals were grown by chemical vapor transport using ICl3 as a transport agent31. The growth was conducted in a horizontal three-zone tube furnace with a temperature gradient setting as 850 °C ← 920 °C → 850 °C for simultaneously growing two sealed quartz ampoules. Prior to the crystal growth, the pure elements of Ga and S with proper stoichiometry combined with a small amount of transport agent (ICl3) were put into the quartz ampoule, which was then cooled using liquid nitrogen, evacuated to approximately 10−6 Torr, and then sealed. The reaction was maintained for 200 h for growing large single crystals. After the growth was completed, the as-grown crystals exhibited white and light-yellow colored.

Optical measurements

The TR experiments were implemented using indirect heating manner with a gold-evaporated quartz plate as the heating element. The thin sample of Ga2S3 was closely attached on the heating element by silicone grease30. The on-off heating disturbance uniformly modulates the layered sample periodically. An 150 W xenon-arc lamp filtered by a PTI 0.2-m monochromator provided the monochromatic light. The incident light was focused onto the sample with a spot size less than thousand μm2. The reflected light from the sample surface was collected and detected by a Hamamatsu H3177-51 PMT module. The signal was detected and recorded via an EG&G model 7265 lock-in amplifier and a personal computer. A closed-cycle cryogenic refrigerator with a thermometer controller facilitated the temperature-dependent measurements. For PTR measurement, one pair of Glan-Taylor-prism polarizer (320 nm–2200 nm) was employed. High-resolution PL measurements were carried out in a spectral measurement system where an iHR 550 imaging spectrometer equipped with a 2400 groves/mm grating acted as the optical dispersion unit. The low-resolution PL spectra were detected by a QE65000 spectrometer. The pumping light source is a helium-cadmium laser (λ = 325 nm). Photo V-I experiments were performed using two dissimilar light sources of tungsten halogen lamp and 405-nm laser as the solar emulators. The averaged power density was adjusted and maintained at approximate 15 mW/cm2 by using the monitor of an OPHIR optical power meter equipped with a broadband high-sensitivity thermal sensor (0.15–6 μm). To prepare the sample for Photo V-I measurements, the specimen of Ga2S3 was cut into a rectangular shape with dimension of 3.15 × 2.95 × 0.3 mm3. The two ends of each specimen were then coated with Au/In, which served as the ohmic-contact electrodes by sputtering. To perform the optical measurements, the ohmic-contact electrodes on each sample were shielded with light. The Photo V-I measurements were implemented using the auxiliary of a semiconductor parameter analyzer. The voltage scanning range was set at −1000 to 1000 V.

Author Contributions

C.H.H. conceived the idea and conducted the experiments. C.H.H. also grew the crystals, analyzed (calculated) the data, and wrote the manuscript. H.H.C. performed the optical measurement and structural characterization.

Supplementary Material

Supplementary Information

Supporting Information

srep06143-s1.pdf (155.6KB, pdf)

Acknowledgments

The authors would like to acknowledge the funding support from the Ministry of Science and Technology of Taiwan under the grant No. 101-2221-E-011-052-MY3.

References

  1. Mafi E., Soudi A. & Gu Y. Electronically driven amorphization in phase-change In2Se3 nanowires. J. Phys. Chem. C 116, 22539–22544 (2012). [Google Scholar]
  2. Li Y. et al. Thermal phase transformation of In2Se3 nanowires studied by in situ synchrotron radiation X-ray diffraction. J. Mater. Chem. 21, 6944–6947 (2011). [Google Scholar]
  3. Jones C. Y. & Edwards J. G. Observation of a phase transformation of Ga2S3 in a quartz effusion cell above 1230 K by means of neutron scattering. J. Phys. Chem. B 105, 2718–2724 (2001). [Google Scholar]
  4. Ho C. H. & Lin S. L. Optical properties of interband transitions of layered gallium sulfide. J. Appl. Phys. 100, 083508-1 to 083508-6 (2006). [Google Scholar]
  5. Goodyear J. & Steigmann G. A. The crystal structure of α-Ga2S3. Acta Cryst. 16, 946–949 (1963). [Google Scholar]
  6. Ho C. H., Wu C. C. & Cheng Z. H. Crystal structure and electronic structure of GaS1-xSex series layered solids. J. Cryst. Growth 279, 321–328 (2005). [Google Scholar]
  7. Peressi M. & Baldereschi A. Structure and electronic properties of Ga2Se3. J. Appl. Phys. 83, 3092–3095 (1998). [Google Scholar]
  8. Ho C. H. & Huang K. W. Visible luminescence and structural property of GaSe1-xSx (0 ≤ x ≤ 1) series layered crystals. Solid State Commun. 136, 591–596 (2005). [Google Scholar]
  9. Persľn M., Čelustka B., Popović S. & Persľn A. Some electrical and optical characteristics of thin films of Ga2Se3. Thin Solid Films 37, L61–L62 (1976). [Google Scholar]
  10. Yoon C. S. et al. Optical properties of Ga2Se3 and Ga2Se3:Co2+ single. crystals. J. Phys. Chem. Solids 62, 1131–1137 (2001). [Google Scholar]
  11. Collin G. et al. Preparation and structure of α-Ga2S3. Mat. Res. Bull. 11, 285–292 (1976). [Google Scholar]
  12. Tomas A., Guymont M., Pardo M. P., Guittard M. & Flahaut J. X-ray diffraction and electron microscopy studies of α- and β-Ga2S3. Phys. Stat. Sol. (a) 107, 775–784 (1988). [Google Scholar]
  13. Liu H. F. et al. Synthesis and phase evolutions in layered structure of Ga2S3 semiconductor thin films on epiready GaAs (111) substrate. ACS Appl. Mater. Interfaces 6, 3501–3507 (2014). [DOI] [PubMed] [Google Scholar]
  14. Barbouth N., Berthier Y., Oudar J., Moison J.-M. & Bensoussan M. The first steps of the sulfurization of III-V compounds. J. Electrochem. Soc.: Solid-State Sci. Technol., 1663–1666 (1986). [Google Scholar]
  15. Ren J. et al. Broadband near-infrared emission of chromium-doped sulfide glass-ceramics containing Ga2S3 nanocrystals. Opt. Lett. 37, 5043–5045 (2012). [DOI] [PubMed] [Google Scholar]
  16. Yoon C. S. et al. Blue photoluminescence of α-Ga2S3 and α-Ga2S3:Fe2+ single crystals. Appl. Phys. Lett. 83, 1947–1949 (2003). [Google Scholar]
  17. Chen F., Dai S., Lin C., Yu Q. & Zhang Q. Performance improvement of transparent germanium-gallium-sulfur glass ceramic by gold doping for third-order optical nonlinearities. Opt. Express 21, 24847–24855 (2013). [DOI] [PubMed] [Google Scholar]
  18. Shakikh H. A. E., Abdal-Rahman M., Belal A. E. & Ashraf I. M. Photoconductivity studies of gallium sesquisulphide single crystals. J. Phys. D: Appl. Phys. 29, 466–469 (1996). [Google Scholar]
  19. Kim W. T., Kim H. S., Kim Y. G. & Hahn S. R. Optical energy gaps of In2-xGaxS3 thin films prepared by spray pyrolysis. J. Mater. Sci. Lett. 6, 479–480 (1987). [Google Scholar]
  20. Ho C. H. & Chen Y. C. Thickness-tunable band gap modulation in γ-In2Se3. RSC Adv. 3, 24896–24899 (2013). [Google Scholar]
  21. Lucazeau G. & Leory J. Etude vibrationnelle de α Ga2S3. Spectrochimica Acta 34A, 29–32 (1978). [Google Scholar]
  22. Julien C. et al. Vibrational studies of copper thiogallate solid solutions. Mater. Sci. Eng. B57, 102–109 (1999). [Google Scholar]
  23. Lee J. S., Won Y. H., Kim H. N., Kim C. D. & Kim W. T. Photoluminescence of Ga2S3 and Ga2S3:Mn single crystals. Solid State Commun. 97, 1101–1104 (1996). [Google Scholar]
  24. Ho C. H., Chan C. H., Tien L. C. & Huang Y. S. Direct optical observation of band-edge excitons, band gap, and Fermi level in degenerate semiconducting oxide nanowires In2O3. J. Phys. Chem. C 115, 25088–25096 (2011). [Google Scholar]
  25. Liu B. et al. Exciton-Related Photoluminescene and Lasing in CdS Nanobelts. J. Phys. Chem. C 115, 12826–12830 (2011). [Google Scholar]
  26. Aspnes D. E. Modulation Spectroscopy/Electric Field Effects on the Dielectric Fuction of Semiconductors, edited by Balkanski, M. Handbook on Semiconductors 2, Chap. 4A, 109–154 (North Holland: Amsterdam., 1980). [Google Scholar]
  27. Ho C. H., Yen P. C., Huang Y. S. & Tiong K. K. Photoreflectance study of the excitonic transitions of rhenium disulphide layer compounds. Phys. Rev. B 66, 245207-1 to 245207-5 (2002). [Google Scholar]
  28. Tongay S. et al. Monolayer behavior in bulk ReS2 due to electronic and vibrational decoupling. Nat. Commun. 5, 4252 (2014). [DOI] [PubMed] [Google Scholar]
  29. Zhang M. J., Jiang X. M., Zhou L. J. & Guo G. C. Two phases of Ga2S3: promising infrared second-order nonlinear optical materials with very high laser induced damage thresholds. J. Mater. Chem. C 1, 4754–4760 (2013). [Google Scholar]
  30. Ho C. H., Lee H. W. & Cheng Z. H. Practical thermoreflectance design for optical characterization of layer semiconductors. Rev. Sci. Instrum. 75, 1098–1102 (2004). [Google Scholar]
  31. Ho C. H. Growth and characterization of near-band-edge transitions in β-In2S3 single crystals. J. Cryst. Growth 312, 2718–2723 (2010). [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Information

Supporting Information

srep06143-s1.pdf (155.6KB, pdf)

Articles from Scientific Reports are provided here courtesy of Nature Publishing Group

RESOURCES