Deep-UV Transparent Conducting Oxide La-Doped SrSnO3 with a High Figure of Merit.

Perovskite stannate SrSnO3 (SSO) is attracting attention as ultraviolet transparent conducting oxides (UV TCOs) due to its ultrawide band gap and high conductivity. Here, we investigate in detail the thickness-dependent electrical, structural, and optical properties of sequentially strain-relaxed La-doped SrSnO3 (SLSO) epitaxial thin films. We find that the SLSO films grow as an orthorhombic Pnma phase with a - a - c + in the c + direction under the tensile strain. With the strain relaxation, as the films become thicker, vertical grain boundaries are created and the orthorhombic phase becomes reoriented to all three possible orientations. Simultaneously, the conductance starts to deviate from the linear behavior with increasing film thickness. Through the analysis of thickness fringes in optical transmittance, we found that a 120 nm thick nominally 4% La-doped SrSnO3 film has a figure of merit (φTC = 2.65 × 10-3 Ω-1) at λ = 300 nm in the deep-UV region, which is the highest value among the well-known candidates for UV TCOs reported to date.

Introduction

The need for deep-ultraviolet transparent conducting oxides (DUV TCOs) has been growing steadily. Accordingly, the search for high conductance in ultrawide band gap (Eg > 4.0 eV) materials becomes important. DUV TCOs are most likely to be applied as transparent electrodes in ultraviolet light-emitting diodes (UV LEDs). In the last decade, UV LEDs are gradually emerging as a new UV source and replacing the traditional UV mercury lamps since they have a number of advantages, such as free of mercury, higher energy efficiency, longer lifetime, and fast operations.[1,2] This promising alternative can be employed in a wide range of applications such as decontamination of various food products,[3] disinfection of water,[4,5] sensing for biomolecules,[6] UV curing,[7] and UV phototherapy.[8] At the same time, UV detection and spectroscopy are becoming increasingly important for space and military applications.[9−11] However, a current roadblock toward such wide applications is that the light extraction and/or detection efficiency is still low in the deeper UV range due to absorption at the top contact layer of an LED or a detector.[12,13] Therefore, it is essential to develop transparent electrode materials that are highly conductive and transparent in the UV range. Al-doped ZnO (AZO)[14] and Sn-doped In2O3 (ITO)[15] are widely used as transparent electrodes for LEDs. However, they have an insufficient transparency in the deep-UV region (λ = 300 nm) due to their small band gap: Eg ∼ 3.27 eV for AZO[16] and Eg ∼ 3.6 eV for ITO.[17]

To date, two classes of materials have been proposed as UV transparent conducting oxide candidates. One is correlated metals, which have strong electron–electron interactions, such as SrNbO3,[18,19] CaVO3,[20][20] and SrVO3.[20,21] The other class is doped semiconductors with an ultrawide band gap such as Sn-doped β-Ga2O3 (Eg ∼ 4.9 eV),[22] Sn-doped α-Ga2O3 (Eg ∼ 5.3 eV),[23] and La-doped SrSnO3 (Eg ∼ 4.6 eV).[24,25] Among the materials corresponding to the former class, CaVO3 and SrVO3, perovskite vannadates, have excellent conductivity over 104 S cm–1. This high conductivity comparable to ITO comes from the higher carrier density (n ∼ 1022 cm–3) with high effective mass due to the strong Coulomb interaction between electrons. However, their optical transmittance for energies above ∼3.3 eV is suppressed and the transmittance is under 50% in the deep-UV region.[20,26] In the case of SrNbO3, there is no direct measurement data yet for optical transmittance, although it has just been calculated to be over 80% from the absorption coefficient through the complex dielectric function obtained by spectroscopic ellipsometry.[18,19] Sn-doped α-Ga2O3 and Sn-doped β-Ga2O3 are candidates for UV TCOs belonging to the latter class. However, they have poor transparency in the deep-UV range at the Sn-doping concentration required for the large conductance.[27,28] In the case of Sn-doped β-Ga2O3, the band gap greatly decreases from 4.9 eV for undoped to 4.1 eV at the optimum doping rate of 10%, resulting in decrease of transmittance from 60 to 5% at 300 nm.[27] A similar phenomenon is observed in Sn-doped α-Ga2O3; its optical transmittance reduces from 80% for 0.005% Sn doping to 50% at the optimum doping rate of 0.1%.[28] Another disadvantage is that they have a high electrical resistivity due to the low dopant solubility limit and degradation of crystallinity with increasing dopant concentration.[27,28]

Very recently, La-doped SrSnO3 (SLSO) has emerged as one of the perovskite-structured alkaline stannates which have a potential for transparent conducting oxides due to their superior electrical conductivity and ultrawide band gap.[24] The conduction band of perovskite stannates originating mainly from the Sn-5s orbital leads to small electron effective mass and high electron mobility in spite of degenerate doping.[29] La-doped BaSnO3 single crystals and epitaxial films are shown to have an outstanding electron mobility of 320 and 150 cm2 V–1 s–1, respectively.[30,31] SrSnO3 possesses a wider band gap (Eg ∼ 4.6 eV) and smaller effective mass (me ∼ 0.23 m0)[24] than BaSnO3 (Eg ∼ 3.1 eV, me ∼ 0.42 m0).[32,33] In bulk powders, an orthorhombic phase () is known to be the stable phase at room temperature[34] with lattice constants of aortho = 5.703 Å, bortho = 5.709 Å, and cortho = 8.065 Å (its pseudocubic lattice constant apc = 4.034 Å).[35] Trivalent ions such as La3+, Ta3+, Nd3+, and Sm3+ were used as effective n-type dopants[36−46] by substituting for the Sr site without affecting the band structure.[47] Additionally, pentavalent ions such as Nb5+, Sb5+, and Pb5+ in place of Sn were studied as n-type dopants.[48−50] It is worth noting that the La dopant creates shallow donor levels in SrSnO3, which indicates that extra electrons in La ions are easily activated.[47] The recently reported 3% La-doped SrSnO3 (SLSO) film has the electron mobility as high as 56 cm2 V–1 s–1 with the carrier density (n ∼ 3.26 × 1020 cm–3), resulting in the highest conductivity (σ ∼ 3000 S cm–1) ever reported for SrSnO3.[24] The solubility of La ions in SLSO films was reported to be enhanced through the postdeposition vacuum-annealing process.

In this paper, we report systematic studies on the thickness-dependent properties of highly conductive and DUV-transparent La-doped SrSnO3 films grown on epitaxial SrHfO3 and BaHfO3 buffer layers on MgO (001) substrates, all DUV-transparent materials, by pulsed laser deposition (PLD) as illustrated in Figure . We used a sequential strain relaxation method using a BaHfO3 perovskite buffer layer, which has an ultrawide band gap (Eg ∼ 6.1 eV)[51] and a lattice constant (a = 4.171 Å),[52] which matches well with that of MgO (a = 4.21 Å).[53] SrHfO3 was chosen as the second buffer layer because it is also an ultrawide band gap (Eg ∼ 6.1 eV)[54] perovskite with a lattice constant (apc = 4.087 Å)[55] between those of BaHfO3 and SrSnO3 (apc = 4.034 Å).[35] The structural properties are examined by a high-resolution X-ray diffractometer (HRXRD) and scanning transmission electron microscope (STEM). We found that as the orthorhombic phase of SLSO films becomes thicker, they begin to change its direction and their conductance starts to deteriorate. The optical transmittance is investigated to maximize the figure of merit by analyzing the thickness fringes displayed by thin-film interference.

Figure 1

Tailoring of deep-UV transparent conducting oxide heterostructure. Schematics show sequential strain-relaxed and thickness-tuned nominally 4% La-doped SrSnO3 films, which are highly conductive and transparent in the deep-UV range. SrSnO3 (SSO) has an orthorhombic perovskite structure with the space group (a–a–c+) at room temperature with a pseudocubic lattice constant (apc = 4.034 Å) and a wide indirect band gap of 4.6 eV.

Results and Discussion

Doping Dependence of the Electrical Properties of 50 nm Thick SLSO Films

To investigate the effect of a nominal La-doping rate, we measured the electrical conductivity, the carrier density, and the Hall mobility of 50 nm thick Sr1–LaSnO3 thin films (x = 0.003, 0.005, 0.01, 0.02, 0.03, 0.04, 0.05) at room temperature. Figure a shows the current–voltage (I–V) data and the schematic for the Van der Pauw measurement used. Figure b–d shows the conductivity, Hall mobility, and carrier density, respectively. A significant increase in conductivity was observed at a 0.5% La-doping rate due to sharp increases in the carrier density (Figure d) and the Hall mobility (Figure c). As the nominal doping rate increases from 0.5 to 4%, the conductivity increases almost linearly with the nominal doping rate since there is little change in the Hall mobility, and the carrier density increases linearly. However, exceeding the nominal La-doping rate beyond 4%, the Hall mobility starts to decrease sharply, while the carrier density maintains the same trend. After a nominal 4% La-doping rate, scattering mechanisms by ionized and/or unionized La dopants seem dominant. Therefore, we found the optimum La-doping rate to be 4%; a 50 nm thick 4% SLSO film shows a high Hall mobility (μ = 44.14 cm2 V–1 s–1) and large carrier density (n3D = 4.25 × 1020 cm–3), and above all, the highest electrical conductivity (σ ∼ 3100 S cm–1) among the values of SLSO films reported to date. On the other hand, when the nominal La-doping rate drops from 0.5 to 0.3%, the Hall carrier density drops quickly from 4.01 × 1019 to 8.14 × 1018 cm–3, and it was difficult to measure in 50 nm thick 0.25% SLSO films due to its high resistance. From this, we can infer that the deep acceptor density of our SLSO films is in the same range of the La3+ ion dopant density in 0.25% SLSO films (nDA ∼ 3.81 × 1019 cm–3).

Figure 2

Doping dependence of electrical properties of 50 nm thick SLSO films grown on 33 nm thick SHO and 55 nm thick BHO buffers on the MgO (001) substrate at room temperature. (a) Voltage measured between 1 and 4 of the film by applying a current between 2 and 3 of the film. The inset is a schematic of Van der Pauw measurements. Indium was used as a contact. (b) Electrical conductivity, (c) Hall mobility, and (d) carrier density. A black thick dotted line represents the carrier density when La ions in the films are 100% activated.

Thickness Dependence of the Electrical Properties of 4% SLSO Films

To examine the electrical properties depending on the thickness of 4% SLSO films, we fabricated t nm thick 4% SLSO films (t = 22, 28, 50, 70, 105, 145, 210, 290, 350, 420) grown on 33 nm thick SrHfO3 and 55 nm thick BaHfO3 buffer layers on the MgO (001) substrate. Figure a depicts the I–V data and a schematic for our Van der Pauw measurement. The conductance is not in a linear relationship with 4% SLSO film thickness, as illustrated in Figure b. Looking at the slope of the conductance up to 145 nm, the extracted conductivity is about 3000 S cm–1. However, beyond the thickness of 145 nm, the conductance does not increase as expected and starts to deviate from the 3000 S cm–1 line. The Hall mobility in Figure c tends to decrease slightly with the increasing film thickness. The two-dimensional (2D) carrier density increases linearly with the film thickness, as illustrated in Figure d. As a result, the conductance, which is proportional to the product of mobility and 2D carrier density, deviates further from the trend when exceeding the thickness of 145 nm. It suggests that there exists a change in microstructure in films as the thickness increases, which tends to impede the current flow.

Figure 3

Thickness-dependent electrical properties of 4% SLSO films grown on 33 nm thick SHO and 55 nm thick BHO buffers on the MgO (001) substrate at room temperature. (a) Voltage measured between 1 and 4 of the film by applying a current between 2 and 3 of the film. (b) Conductance gradually deviates from the red dotted line representing the conductivity of 3000 S cm–1 with increasing film thickness. (c) Hall mobility slightly decreases as the film thickness increases. (d) Two-dimensional carrier density follows the blue dotted line representing the three-dimentional (3D) carrier density when La ions in the film are 100% activated.

Structural Properties of 4% SLSO

To study the phenomenon discussed above in detail, we investigate the structural properties of t nm thick 4% SLSO films. Figure a,b shows the results of high-resolution X-ray θ–2θ scans and rocking curves of t nm thick 4% SLSO films (t = 22, 50, 105, 145, 290). In Figure a, the 2θ peak positions of the MgO (aop = 4.217 Å) (001) substrate and the BaHfO3 (aop ∼ 4.184 Å) and SrHfO3 (aop ∼ 4.101 Å) buffer layers do not move, regardless of 4% SLSO thickness. The thickness fringes of BaHfO3 around the (002) diffraction peak can be observed. From the spacing between the fringes, the BaHfO3 thickness was calculated to be 53 nm, and it agrees with those measured by the stylus profiler and transmission electron microscopy (TEM). As the thickness of the 4% SLSO film increased, the 2θ peak position of 4% SLSO shifted to the left. It is seen clearly between t = 50 and 105 nm. Simultaneously, the symmetric peak shape of the 4% SLSO film becomes asymmetric beyond the thickness of 50 nm. Figure e–g shows the results of deconvoluting 50, 105, and 290 nm thick 4% SLSO asymmetric peaks using the Voigt functions. The 50 nm thick film peak is well fitted with only one Voigt function and its position is very close to the peak position of the 22 nm thick film in the Supporting Information (Figure S1a). However, in the case of the 105 nm and 290 nm thick films, the additional peak (peak 2) is necessary to fit the experimental data. The sums of the two peaks can be very well fitted to the experimental data (black circles). Therefore, either a new orientation or a new phase, corresponding to peak 2, starts to appear in the thickness range between 50 and 105 nm. The case of 145 nm thick film in the Supporting Information (Figure S1b) is the same as the case of 105 nm. It is highly likely that the tensile strain is relieved by creating a new orientation or a new phase as the 4% SLSO films become thicker. This is in contrast to the previously reported situation, where the compressive strain on GdScO3 (110) substrates is relaxed as the SLSO film becomes thicker.[44,45]Figure c illustrates the rocking curves for 105 nm 4% SLSO, 33 nm SrHfO3, 55 nm BaHfO3 and the MgO (001) substrate. The MgO (001) single-crystal substrate has a sharp peak shape with full width at half-maximum (FWHM) of around 0.032. The rocking curve of 55 nm BaHfO3 consists of two Gaussian functions: narrow and broad Gaussian functions, which suggests that 55 nm BaHfO3 is partially relaxed on the MgO substrate. Typically, 33 nm SrHfO3 and 105 nm 4% SLSO with an orthorhombic structure have wider rocking curves than cubic BaHfO3 and the single-crystal MgO substrate. The full width at half-maximum (FWHM) values of the rocking curves of the 4% SLSO films gradually decrease with increasing the film thickness, as illustrated in Figure d.

Figure 4

Thickness-dependent high-resolution X-ray diffraction. (a) θ–2θ scan and (b) rocking curves of t nm thick 4% SLSO films. (c) Rocking curves of 105 nm thick 4% SLSO, 33 nm thick SHO, and 55 nm thick BHO and MgO in order from the top. (d) FWHM for t nm thick 4% SLSO, 33 nm thick SHO, and 55 nm thick BHO as a function of t. (e–g) Results of the deconvolution of 2θ peaks using the Voigt functions for 50, 105, and 290 nm thick 4% SLSO films, respectively. (e) 2θ peak of 50 nm thick 4% SLSO is well fitted by only one Voigt function (green line) to the measured data (black circles). However, in the case of (f) 105 and (g) 290 nm thick 4% SLSO, the 2θ peaks are well fitted by two Voigt functions (green and red lines) to the measured data. Blue lines in panels (f, g) are the summation of peak 1 and peak 2.

To investigate the structural properties in detail, we measure the asymmetric reciprocal space mapping (RSM) for the (024) plane and (002) plane for t nm thick 4% SLSO, 33 nm thick SrHfO3, and 55 nm thick BaHfO3 and MgO. Figure a,b shows the reciprocal space mapping (RSM) results around the (024) plane for 22 and 290 nm thick 4% SLSO, respectively. MgO, BaHfO3, SrHfO3, and 4% SLSO peaks sequentially appear from the bottom. The RSM results for the (024) plane of 28, 50, 70, 105, 145, and 210 nm thick 4% SLSO films are shown in the Supporting Information (Figure S2). The schematic display of the four peaks can be shown in Figure a,b. The 290 nm thick 4% SLSO peak consists of a light and a dark green ellipse. In light of the 4% SLSO peak deconvolution results in Figure , we can interpret that the light green corresponds to peak 1, and the dark green corresponds to peak 2. The horizontal long black dotted lines are the value corresponding to the reciprocal space vector Q at each peak position. In line with the previous θ–2θ scan results, the peak positions of BaHfO3, SrHfO3, and MgO do not change, regardless of 4% SLSO thickness. Compared with the 22 and 290 nm thick 4% SLSO peaks, one can notice that their shapes and positions are changed. Even the 22 nm thick 4% SLSO does not grow coherently on the SrHfO3 buffer layer from the fact that their reciprocal space vector Q is not the same. Figure c,d shows the reciprocal space mapping (RSM) around the (002) plane for 22 and 290 nm thick 4% SLSO films, respectively. One can observe the thickness fringes of a 55 nm BaHfO3 buffer layer. The widths of the four peaks in Figure c in the reciprocal space vector Q direction seem constant. However, the width of 290 nm 4% SLSO peak in Figure d in the Q direction is definitely larger. Consequently, the (002) RSM results confirmed again that a new orientation or a new phase is formed as the 4% SLSO thickness increases. Figure e shows the in-plane (aip) and out-of-plane (aop) lattice constants for 4% SLSO calculated from the RSM (024) data as a function of the 4% SLSO film thickness. As the film thickness increases, the tensile strain by the SrHfO3 buffer gets relaxed and a noticeable drop in the in-plane lattice constant can be observed between 50 and 70 nm thickness. The in-plane and out-of-plane lattice constants eventually merge with each other as the films become thicker.

Figure 5

Thickness-dependent reciprocal space mapping (RSM). (a, b) RSM images for the (024) plane and the corresponding schematics of peak shape of 22 and 290 nm thick 4% SLSO films, respectively. (c, d) RSM images for the (002) plane of 22 and 290 nm thick 4% SLSO films, respectively. The horizontal black dotted lines represent the value of the reciprocal space vector (Q) corresponding to the four peaks in the RSM (024) and RSM (002) images for the 22 nm thick 4% SLSO film. (e) In-plane (aip), the out-of-plane lattice constants (aop), and the unit cell volume (Vunit cell) of the 4% SLSO films calculated from the RSM (024) as a function of the film thickness.

Since it is clear that the microstructure of the films changes as a function of the SLSO thickness, scanning transmission electron microscopy (STEM) experiments were conducted (see the Supporting Information, Figures S3 and S5 for basic structural and compositional characterizations of SrSnO3 films using STEM-energy-dispersive X-ray (EDX) analysis). In Figure a, a cross-sectional high-angle annular dark-field (HAADF)-STEM image of the 290 nm thick 4% SLSO film is presented, where the growth of columnar grains with the width of 10–50 nm is noted from the grain boundary contrasts. High-resolution bright-field (BF)-STEM images acquired from the top and bottom regions show distinct patterns (Figure b). At the bottom, only a lateral stripe pattern with two-unit-cell periodicity along the z-direction is seen, whereas three dissimilar patterns—lateral stripe, vertical stripe, and cross patterns—in different grains are observed in the top region. STEM images of a 22 nm thick SLSO film show only lateral patterns (see the Supporting Information, Figure S3c), implying the development of different crystal orientations in the SrSnO3 grains with the film growth. Fast Fourier transforms (FFTs) obtained from the grains with the distinct patterns are shown in Figure c, where the appearance of spots corresponding to (022) (highlighted with the green circles), (002) (yellow circles), and (020) (violet circles) planes from the cross, lateral, and vertical patterns is witnessed.

Figure 6

Atomic images of 4% SLSO films. (a) HAADF-STEM image of a 290 nm thick 4% SLSO film. The top and bottom regions in the film are indicated with pink and red rectangular boxes, respectively. The scale bar is 100 nm. (b) High-resolution BF-STEM images from top and bottom regions in the 4% SLSO film. Distinct patterns in different grains are highlighted. The scale bars are 10 nm. (c) FFTs obtained from the top region. Spots from SrSnO3 grains with different orientations are circled with color codes: green (c+a–a–), yellow (a–a–c+), and violet (a–c+a–). The scale bars are 0.2 Å–1.

At room temperature, SrSnO3 exhibits an orthorhombic phase (space group of ) with the SnO6 octahedral rotation pattern of a–a–c+ in the Glazer notation.[56]Figure shows the atomic model of the phase SrSnO3 projected in three Cartesian axes along the in-phase (c+) (Figure a) and out-of-phase (a–) (Figure b,c) octahedral rotation directions. Here, the projection direction is denoted as the x-axis, and the film growth direction is the z-axis. The projected image in the c+ direction (Figure a) shows the in-phase octahedral rotation from corresponding O displacements; however, the image from the a– direction (Figure b,c) does not since the rotations in the opposite (clockwise and anticlockwise) directions are overlapped. Atomic-resolution BF-STEM images were obtained to connect the atomic structure viewed in different orientations with the three types of grains. As presented in Figure , dark contrast appears at the atomic positions, which enables visualization of light O atoms and subtle atomic displacements in the structure. First, a BF-STEM image producing the cross pattern shows the O displacement corresponding to the octahedral rotation, suggesting that the projection direction of the STEM image is the c+ direction, i.e., c+a–a–. In the case of the stripe-patterned BF-STEM images, such O pattern is not discernable, but slight Sr displacements along the stripe lines are seen, which is the result of the atomic images projected in the a– direction. A comparison between the atomic model and BF-STEM images reveals that the stripe pattern is generated from an alternating Sr displacement pattern that is perpendicular to the c+ direction. Thus, in the case of grains displaying a lateral stripe pattern, the c+ direction is along the z-axis (growth direction), i.e., a–a–c+, while the vertical stripe pattern develops when the c+ axis is the y-axis, i.e., a–c+a. The development and distribution of the reoriented grains were inspected from the Fourier filtered images using the unique spots from each orientation as presented in the Supporting Information (Figure S4a). The reoriented grains starting from ∼50 nm above the film–substrate interface are evidenced, which agrees with the lattice parameter change evidenced in the X-ray diffraction experiments. The BF-STEM image of the boundary between the initial and oriented grains exhibits a continuous but clearly distinct boundary between two regions (see the Supporting Information, Figure S4b).

Figure 7

Atomic model and BF-STEM images of 4% SLSO. (a–c) Atomic model of the orthorhombic SrSnO3 structure viewed along the in-phase (c+) and out-of-phase (a–) octahedral rotation directions. Here, the x-direction is the viewing direction, and the z-direction is the film growth direction. Atomic-resolution BF-STEM images from grains with different orientations. Atomic models are overlaid. The O atomic positions showing the in-phase octahedral rotation are highlighted with red arrows, and the Sn atomic displacement directions are marked with green arrows.

STEM analysis concludes that the SrSnO3 films initially grow along the c+ direction under the tensile strain, and then, with the strain relaxation, grains are reoriented to all three possible orientations. Since the bulk lattice parameters of the phase are 4.035 and 4.033 Å along the a– and c+ directions, respectively,[40] the initial growth of a–a–c+ alignment with larger lattice parameters being the in-plane axes can be stabilized with the imposed tensile strain. It should be noted that while recent studies have reported strain-driven phase transition of SrSnO3 to the tetragonal I4/mcm phase with an octahedral rotation pattern of a0a0c–,[40,45] XRD result from the films studied here shows significantly smaller (∼3%) out-of-plane lattice parameters compared to those of the tetragonal phase and, therefore, excludes the possibility of tetragonal phase formation.

Optical Properties of 4% SLSO Films

Optical transmittance spectra of t nm thick 4% SLSO films (t = 50, 70, 105, 145, 210, 290) are measured as a function of wavelength in Figure a. As the film thickness increases, the absorption becomes stronger and the number of fringes by thin-film interference increases. Analyzing the number of peaks and their positions, we believe that the interference fringes are caused by the reflections at the air/SLSO interface and the BHO/MgO interface. This is not surprising since the refractive indices of MgO, BHO, SHO, and SSO are 1.71 at 550 nm,[57] 2.5 at 300 nm,[58] 2.0 at 300 nm,[59] and 2.2 at 413 nm,[60] respectively. We obtained the highest optical transmittance in the DUV range (λ = 300 nm) of about 77.5% for a 120 nm thick 4% SLSO film, which is higher than the previously reported 73% transmittance at the wavelength of 300 nm in postdeposition vacuum-annealed 112 nm thick 3% SLSO.[24]

Figure 8

Thickness-dependent optical properties of 4% SLSO films. (a) Optical transmittance spectra vs wavelength of 4% SLSO films with various thicknesses. The inset shows a magnification of the optical transmission spectra in the UV region. A black dotted vertical line represents the wavelength of 300 nm. (b) Optical indirect band gap vs photon energy of 4% SLSO films with various thicknesses. The inset is the results of indirect band gap fitting for 4% SLSO films. At all thicknesses, the indirect band gap is about 4.6 eV.

Figure b shows the (αhν)1/2–photon energy plot to estimate the indirect band gap of t nm thick 4% SLSO films, where α is the absorption coefficient, v is the frequency of the incident photon, and h is the Planck’s constant, since SrSnO3 is known to exhibit an indirect band gap.[61,62] The indirect band gap can be estimated by extrapolating the straight-line portion of the plot to the energy axis, illustrated in the inset of Figure b. Independent of the film thickness, the indirect band gap of t nm thick 4% SLSO is around 4.6 eV, which is similar to the previously reported value of 3% SLSO films estimated by assuming a direct band gap.[24] Our estimation for the direct band gap by plotting (αhν)2 vs photon energy is shown in Figure S6, which suggests that the direct band gap is about 4.9 eV.

Finally, we can calculate the figure of merit for transparent conductors[63] (φTC = T10/Rs, where T is the transmission and Rs is the sheet resistance) at a wavelength of 300 nm, as shown in the Supporting Information (Table S1), and obtain φTC = 2.65 × 10–3 Ω–1 for 120 nm thick 4% SLSO, which is the highest value reported to date, among the candidates of deep-UV transparent conducting oxides such as Sn-doped β-Ga2O3 (φTC = 7.8 × 10–7 Ω–1)[22] and 3% La-doped SrSnO3 (φTC = 1.45 × 10–3 Ω–1).[24]

Conclusions

In conclusion, we have fabricated thickness-tuned 4% SLSO films grown on 33 nm SrHfO3 and 55 nm BaHfO3 buffers on the MgO (001) substrate by employing a sequential strain relaxation method. We have found that excellent conducting properties were obtained in 50 nm thick nominally 4% La-doped films with their high conductivity (σ ∼ 3100 S cm–1), large carrier density (n = 4.25 × 1020 cm–3), and high Hall mobility (μ = 44.14 cm2 V–1 s–1). As the films become thicker, we found that the tensile strain exerted by a SrHfO3 buffer layer is relaxed by forming grains in all three directions of the orthorhombic phase. At the same time, the conductance starts to deviate from the linear relationship with the film thickness due to the vertical grain boundaries that impede the transport of electrons. Via thickness tuning for the optical transmittance at a wavelength of 300 nm, a 120 nm 4% SLSO film was found to show the highest figure of merit calculated from the sheet resistance and transmittance. We demonstrate the great potential of 4% SLSO films for the next-generation deep-UV (λ = 300 nm) transparent conducting oxide (DUV TCO).

Experimental Section

Thin-Film Growth

All La-doped SrSnO3 (SLSO) epitaxial thin films grown on 33 nm SrHfO3 and 55 nm BaHfO3 buffers on the MgO (001) single-crystal substrate by the PLD method (KrF excimer laser, λ = 248 nm, coherent). Two side-polished MgO (001) substrates (5 mm × 5 mm × 0.5 mm) were provided by MTI Korea. All targets of BaHfO3, SrHfO3, SrSnO3, and 0.5, 1, 4, 8% La-doped SrSnO3 were provided by Toshima manufacturing Co., Japan. First, a 55 nm BaHfO3 buffer layer is grown on the MgO (001) substrate by PLD (energy influence = 1.44 J cm–2 and repetition rate = 10 Hz). The substrate temperature and oxygen partial pressure in the chamber were kept at 750 °C and 100 mTorr, respectively. The distance between the target and substrate is 51 mm. Second, the 33 nm SrHfO3 buffer layer is grown on the 55 nm BaHfO3 buffer layer by PLD (energy influence = 1.44 J cm–2 and repetition rate = 10 Hz). The substrate temperature and oxygen partial pressure in the chamber were kept at 750 °C and 100 mTorr, respectively. The distance between the target and substrate is 45 mm. Finally, SLSO is grown by PLD (energy influence = 1.55 J cm–2 and repetition rate = 4 Hz). The substrate temperature and oxygen partial pressure in the chamber were kept at 750 °C and 150 mTorr, respectively. The distance between the target and substrate is 55.5 mm. After deposition, the SLSO films were cooled in vacuum to 400 °C and in 600 Torr oxygen thereafter.

Structural Property Measurement

The structural properties of the SLSO films were analyzed by θ–2θ scan, rocking curves, and reciprocal space mapping (RSM). θ–2θ scan and rocking curves were conducted using a high-resolution X-ray diffractometer (D8 Discover, Bruker) at IBS Center for Correlated Electron Systems, Seoul National University. The reciprocal space mapping (RSM) is performed using SmartLab with a Cu Kα1 source (λ = 1.5406 Å; Rigaku, Japan) at room temperature. An X-ray CBO system, a Ge (220) 2-bounce monochromator, and a one-dimensional semiconductor array detector (hybrid photon counting detector; HyPix-3000) were used for high-resolution crystalline qualities. A scanning transmission electron microscopy (STEM) experiment was conducted using a FEI Titan G2 60-300 (S)TEM equipped with a CEOS DCOR probe corrector, a Schottky extreme field emission gun, and a super-X energy dispersive X-ray (EDX) spectrometer. The STEM was operated at 200 keV with a beam current of ∼30 pA, and the probe convergence angle was 25.5 mrad. High-angle annular dark-field (HAADF)-STEM images were acquired with the annular dark-field (ADF) detector inner angle of 93 mrad, and bright-field (BF)-STEM images were acquired with the BF detector outer angle of 19 mrad.

Electrical Property Measurement

The electrical properties of SLSO films were measured using a Keithley 4200 semiconductor characterization system. Indium was used as the contact. The Van der Pauw method was used in a 4-point geometry for determining the sheet resistance of the films. The Hall carrier density and mobility were determined by applying the magnetic field perpendicular to the films while measuring the change in resistance.

Optical Property Measurement

The optical transmittance of SLSO films was measured by a grating spectrometer (Cary 5000, Bruker) over 200–2000 nm (0.6–6.2 eV). The spectrometer has a QI lamp light source for 2000–350 nm and a deuterium UV lamp light source for below 350 nm. Samples are mounted on a holder with a 3 mm diameter hole. Absorbance is calculated as the minus logarithm of the transmittance, and the absorption coefficient α is calculated by accounting for the sample thickness. We removed the absorption of the substrate and two buffers by measuring their optical absorption separately and subtracting it from the absorbance. From the optical absorption measurement, we plotted the Tauc’s plot of (αhν)1/2 vs photon energy, where α denotes the absorption coefficient.