Open Access
Add to BibSonomy
Abstract
New opportunities for high-performance CMOS-compatible optoelectronic devices have accelerated the interest in vertically configured device topologies that enable next-generation photonic technologies. Lately, TiN has been identified as a promising refractory metal–ceramic for the hybrid integration of emerging semiconductor materials on a variety of substrates, including Si, MgO, and sapphire. Among these, Si is the least expensive and most commonly used element and substrate material in the semiconductor device industry. Following these examples, a hybrid oxide–nitride–Si stack is proposed and thoroughly investigated herein for its potential use in DUV optoelectronic device applications. The stack comprises β-Ga2O3 thin films grown heteroepitaxially on TiN/Si platforms, wherein the TiN interlayers were heteroepitaxially grown on bulk (100)-oriented Si and act as lattice-mismatched templates and bottom device electrodes. Albeit the relatively large lattice mismatch between Si and TiN, a low in-plane rotation of 3$^\circ$ revealed that the TiN layers continued to grow as a bulk crystal, paving the way for heteroepitaxial β-Ga2O3 thin films being grown without exhibiting amorphous and metastable phases. DUV photodetectors based on this optoelectronic heterostructure exhibited average peak spectral responsivity and external quantum efficiency levels as high as 249 A/W and 1.23 × 105%, respectively, in the ultraviolet-C regime at an illuminating power density of around 12 µW/cm2.
© 2021 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
In the quest for developing modern optoelectronic devices, group III–oxide alloys are of interest as they are inexpensive to synthesize and have advantageous optical and electronic properties [1,2]. These materials are currently used for device applications such as light detection in the ultraviolet (UV) portions of the optical spectrum. They are chemically and thermally robust, exhibit long carrier lifetimes, operationally stable, and, besides group–III nitrides, are known to have wide and direct bandgaps and are wavelength-tunable within the UV regime of operation (from around 200 to 400 nm) [3]. The successful growth and heterogeneous integration of various forms of inorganic semiconductor materials into one electronic system enables efficient carrier injection processes and assist in the realization of excellent device performance characteristics [4]. The main causes of low efficiency parameters are the high density of threading dislocations (TDs) extending from the surface of a strained layer system, which causes internal structural cracking and the subsequent increase in nonradiative recombination channels within the device active regions. These issues arise mainly from the lattice and thermal mismatches between the grown material and the substrate [5–7].
Although there are attempts to use thick buffer layers between silicon (Si) substrates and heteroepitaxially grown semiconductor materials for reducing the strain between them, the fact that such buffer layers have thicknesses between 1 and 2 $\mu$m may negatively impact the optical properties of the fabricated semiconductor devices [8]. Examples of systems that need to be integrated on the same substrate, i.e., both electronic devices (e.g., transistors) and optical devices (e.g., optical emitters or receivers), include nanomechanical optical detection devices [9], solid-state neutron detection devices [10], piezoelectric resonators and electrical and harmonic generators [11], strain-gated transistors [12], single-photon emission devices [13], white light generation from light-emitting diodes (LEDs) and from laser diodes (LDs) [14], high-electron-mobility transistors (HEMTs) [15], and III–V lasers and photodetectors integrated on Si within complementary metal–oxide–semiconductor (CMOS) compatible frameworks [16,17]. However, the heteroepitaxial growth of monocrystalline compound oxide semiconductor thin films on monocrystalline Si substrates without intermediary layers that minimize strain is a difficult task given the formation of amorphous silicon oxide (SiO$_2$) interfacial layers [18,19]. Given the relative ease of growing high quality refractory transitional metal–ceramics (TMCs) on Si, and the relative lattice match between these refractory TMCs and group III–oxides, thin monocrystalline refractory TMC interlayers having a thickness greater than 100 nm facilitate the growth of group III–oxides on Si platforms (e.g., (100)- and (111)-oriented Si) without thick buffer layers. Furthermore, given the high thermal conductivities of the refractory TMCs and Si [20], electronic devices fabricated based on such platform are expected to demonstrate remarkable heat dissipation characteristics that will aid in the realization of reliable power electronic devices. As a wide bandgap semiconductor, $\beta$-Ga$_2$O$_3$ exhibits large breakdown fields (approximately 8 MV cm$^{-1}$ [21]), the issue of poor heat dissipation in power electronic devices that use such semiconductors has not been critically and practically addressed in the field.
In this work, we propose and investigate the growth of a hybrid oxide–nitride–Si stack comprising a beta-polymorph gallium oxide ($\beta$-Ga$_2$O$_3$) thin film on a TMC interlayer, namely titanium nitride (TiN), that acts as a growth template on a (100)-oriented Si substrate and a conductive electrode for a vertically configured DUV optoelectronic device. Previously, a gallium nitride (GaN) layers have been grown on silicon carbide (SiC) substrates using epitaxial scandium nitride (ScN) buffer layers [22], and on (111)-oriented Si substrates using aluminum nitride (AlN) nucleation layers [23]. The structure discussed here is different from the existing ones in the fact that it utilizes high-quality monocrystalline thin films of refractory transitional metal conductive ceramics, grown on (100)-oriented Si, as templates for the growth of monocrystalline $\beta$-Ga$_2$O$_3$ thin films, and hence opening a new paradigm in integrated circuit (IC) systems to transform them into more powerful resources for state-of-the-art Si-integrated electronics.
The heteroepitaxial growth of monocrystalline group III–oxide materials on (100)-oriented Si wafers, without resorting to growing significantly thick buffer layers as the traditional methods do, has not been previously reported. In other words, this approach of growing $\beta$-Ga$_2$O$_3$ thin films on Si substrates makes possible the integration of optoelectronic devices with conventional CMOS electronics because of the abundance and availability of such substrates. Moreover, of these refractory TMCs, polycrystalline TiN has widely been used as a diffusion barrier in microelectronic devices [24–26]. Monocrystalline TiN exhibits a free carrier concentrations ($N$) level in the order of $10^{22}$ cm$^{-3}$ compared to $10^{20}$ cm$^{-3}$ of ScN [22,27]. As monocrystalline TiN growth requires high temperatures that may not be compatible with some CMOS back-end-of-line (BEOL) technology processes [28,29], here we demonstrate that it is possible to first grow TiN thin films on Si substrates, and then to grow other layers, to not expose the stack to the high temperatures required by TiN growth. This integration step is therefore compatible with CMOS front-end-of-line (FEOL) process requirements.
Several characterization techniques were employed to investigate the Si-integrated heterostructure, including X-ray diffraction (XRD), scanning and high-resolution transmission electron microscopy (STEM and HRTEM), and energy-dispersive X-ray (EDX) spectroscopy. EDX analysis combined with STEM and high-angle annular dark-field (HAADF) micrography were used to confirm the chemical composition of grown films. As exhibited in Section 3, out-of-plane XRD measurements revealed the relationships between the $\{100\}$ $\beta$-Ga$_2$O$_3$, TiN, and Si family of crystal planes, whereas $\phi$-scan skewed asymmetric XRD measurements revealed the relationships between the $\{110\}$ TiN and Si family of planes. Combined together, the out-of-plane and $\phi$-scan skewed asymmetric XRD measurements showed that (100) Si and (100) TiN planes are parallel, implying that (010) Si and TiN are parallel to each other as well. Moreover, it is manifested that the (420) $\beta$-Ga$_2$O$_3$ asymmetric Bragg reflections are separated from the (220) Si and TiN plane reflections by 45$^\circ$. Given that the (200) $\beta$-Ga$_2$O$_3$ plane is parallel to (200) Si and TiN, it is implied that the (020) $\beta$-Ga$_2$O$_3$ plane is rotated about 45$^\circ$ with respect to the (020) Si and TiN planes. Therefore, we concluded that the (020) $\beta$-Ga$_2$O$_3$ plane is parallel to the (022) Si and (022) TiN planes.
Next, as detailed in Section 4, we observed the sharp layer transitions and high quality of interfaces in the heterostructure stack via HRTEM imaging. Because $\beta$-Ga$_2$O$_3$ crystallizes in the monoclinic phase, it is hypothesized that its nucleation and growth process on cubic phase TiN is disordered in nature, whereby the two unit cell configurations of the $\beta$-Ga$_2$O$_3$ crystal, exhibiting rotational twin domains, grew alternately side by side in a semiperiodic manner on TiN while maintaining a single crystal phase. This particular phenomenon was observed and detailed in our previous work where a hybrid stack similar in structure, but fundamentally different in properties, was grown on a magnesium oxide (MgO) substrate [27,30]. This film formation mechanism contributed to the introduction of additional defects in the $\beta$-Ga$_2$O$_3$ lattice (Taylor’s dislocations in addition to growth-induced vacancies). While the lattice mismatch between a grown TiN thin film and a MgO substrate is around 0.95%, it is relatively small compared to that between a TiN thin film and a Si substrate (around 22%). Therefore, the epitaxial growth of TiN thin films on Si substrates should lead to higher packing of dislocations at the TiN/Si interface. This explains the improvement in $\beta$-Ga$_2$O$_3$ crystal quality farther away from the TiN/$\beta$-Ga$_2$O$_3$ interface as it is grown thicker.
It is paramount to note that our proof-of-concept study focuses on demonstrating the growth of monocrystalline $\beta$-Ga$_2$O$_3$ on monocrystalline conductive TiN for potential use in vertically configured deep-ultraviolet (DUV) optoelectronic applications on CMOS-compatible platforms. This is a distinguishing feature that set our work apart from our previous report. More importantly, albeit the relatively large lattice mismatch between Si and the heteroepitaxially grown TiN, the proposed TiN/Si templates are shown to serve as a robust platform for the heteroepitaxial growth of $\beta$-Ga$_2$O$_3$ thin films without exhibiting amorphous and metastable phases, and the resulting lattice fit and dislocation types are revealed and examined.
Finally, Section 5 presents the characterization of solar-blind DUV photodetectors based on the proposed hybrid oxide–nitride–Si stack. The DUV photodetectors exhibited average peak spectral responsivity and external quantum efficiency (EQE) levels as high as 249 A/W and $1.23 \times 10^{5}\%$, respectively, in the ultraviolet-C (UVC) regime at an illuminating power density of around 12 $\mu$W/cm$^2$.
2. Experimental section
2.1 Thin film growth and device fabrication
Figure 1 shows a schematic illustration of the hybrid oxide–nitride–Si optoelectronic heterostructure synthesis flow. A 1 cm $\times$ 1 cm, (100)-oriented Si substrate was ultrasonically cleaned in a bath of acetone and isopropyl alcohol (IPA), then introduced into a magnetron sputtering processing chamber for the subsequent deposition of a 180 nm-thick TiN thin film. The substrate was first degassed for 60 minutes at 300 $^\circ$C, etched for 30 minutes in an Ar radiofrequency (RF) plasma of 50 W, and then annealed in vacuum at 800 $^\circ$C for another 30 minutes. The TiN film was deposited at a substrate temperature and bias of 800 $^\circ$C and $-120$ V, respectively, using RF magnetron co-sputtering of two Ti targets operated at a power of 180 W each, in an Ar-N$_2$ reactive atmosphere of 5 mTorr working pressure fed by 18.5 sccm of Ar and 1.5 sccm of N$_2$. The TiN deposition rate was about 1.5 nm/minute. Two Ti targets were used to increase the deposition rate and enhance the film uniformity. A TiN thickness of 180 nm corresponds to a deposition duration of about 2 hours and is high enough to ensure adequate electrical conductivity given that the resistivity of metallic thin films with thicknesses lower than 100 nm is dominated by interface scattering effects [31–33]. TiN thicknesses above 250 nm were observed to exhibit stacking faults that lead to the loss of in-plane orientation and hence we limited the film thickness to 180 nm to suppress the formation of such defects [27].
Fig. 1. DUV $\beta$-Ga$_2$O$_3$ photodetectors on TiN/Si—Fabrication process: (A) Si wafer preparation, (B) high-temperature magnetron sputter deposition of TiN, (C) area-selective PLD growth of $\beta$-Ga$_2$O$_3$, and (D) electron beam deposition and lift-off of Au/Ti metal contacts.
Next, a Neocera Pioneer 180 PLD system with a COMPex 205 excimer laser (KrF, 248 nm) and a maximum pulse energy of 750 mJ (as measured at low repetition rate) was used to deposit unintentionally-doped $\beta$-Ga$_2$O$_3$ thin films. The system has a maximum repetition rate of 50 Hz with an average power of 33 W. Furthermore, the system exhibits an energy stability (1 sigma) of $\leq 0.75\%$ at 248 nm, with a pulse duration of 25 ns. The beam dimensions (V $\times$ H, FWHM) were $24 \times 10$ mm$^2$, with a beam divergence of (V $\times$ H, FWHM) $\leq 3 \times 1$ mrad$^2$ at 248 nm. The beam pointing stability (1 sigma) at shutter plane over 2000 pulses was $\leq 50$ $\mu$rad. The unintentionally-doped $\beta$-Ga$_2$O$_3$ thin films were grown at a substrate temperature of 640 $^{\circ}$C, an oxygen working pressure of 5 mTorr, a laser pulse frequency of 5 Hz, an energy per pulse of 200 mJ, and a laser fluence of 2 J/cm$^2$. The films were deposited with a target-to-substrate distance of 80 mm and 30k pulses; their thicknesses were estimated at 400 nm from a previously calibrated deposition rate and confirmed using TEM imaging. Mesa regions exposing the conductive TiN films were defined using a shadow mask during the $\beta$-Ga$_2$O$_3$ film deposition by PLD. Figure 2 shows atomic force microscopy (AFM) images of grown TiN and $\beta$-Ga$_2$O$_3$ surfaces. The root mean square (RMS) roughnesses of the TiN and $\beta$-Ga$_2$O$_3$ layers are 0.702 nm and 5.12 nm, respectively. No evidence of relaxation fault formation in the TiN layer is observed. These results show that our $\beta$-Ga$_2$O$_3$ layer is relatively more rough compared to our previously reported growth on TiN/MgO [27], which can be attributed to the increased roughness of the TiN layer grown on Si given the relatively large lattice mismatch between the two crystals. In our previous report, the TiN and $\beta$-Ga$_2$O$_3$ layers had measured RMS roughness levels of 0.451 nm and 1.74 nm, respectively, with an average $\beta$-Ga$_2$O$_3$ grain diameter of 66 nm. Here, an average grain diameter of 40 nm in the $\beta$-Ga$_2$O$_3$ crystal grown on TiN/Si is observed. This indicates that the deposited $\beta$-Ga$_2$O$_3$ layer has smaller grain sizes compared to what we have previously reported using the TiN/MgO platform. We conjecture that this smaller grain size is caused by the increased roughness of the TiN surface on Si, which led to the creation of more $\beta$-Ga$_2$O$_3$ crystal nucleation centers. This is expected to have caused further increase in carrier scattering events during electrical excitation, which may have affected the carrier collection processes that are crucial to the photodetection process and thus had a negative impact on our photodetector device performance.
Fig. 2. Atomic force microscopy imaging—2D and 3D AFM measurements of the grown (A) TiN and (B) $\beta$-Ga$_2$O$_3$ layers.
Finally, a top electrode design was patterned on the $\beta$-Ga$_2$O$_3$ thin film with five interconnected Au/Ti (150 nm/50 nm in thickness) parallel fingers with 50 $\mu$m spacing acted a contact electrode to the $\beta$-Ga$_2$O$_3$ film while Au/Ti thin films were deposited directly on the TiN film through the mesa, constituting a vertically configured DUV photodetector. This device architecture whereby on oxide layer is grown on Si-based platform using an intermediary metallic interlayer is unique and has not been reported in the literature as a patent is still pending [34]. The Au/Ti metal layers were deposited using an Equipment Support Company (ESC) sputtering system (Ti: 25 sccm Ar flow at 5 mTorr and 400 W for 286 seconds at room temperature; and Au: 25 sccm Ar flow at 5 mTorr and 400 W for 310 seconds at room temperature). These metal pads were patterned through a lift-off process using a 1.6 $\mathrm{\mu}$m-thick AZ 5214 photoresist exposed using Heidelberg Instruments $\mu$PG501 optical direct-write lithography system. This design allows for the photogenerated electron-hole pairs created by ultraviolet light to be efficiently separated and transported to the metal contacts through the 400 nm-thick $\beta$-Ga$_2$O$_3$ layer in contrast to the less efficient transport that occurs in a laterally configured devices with electrodes that are several tens of microns apart.
2.2 X-ray diffraction crystallography
Crystal structure properties of the optical heterostructure were examined by Bruker D8 Advance (out-of-plane incidence diffraction, without a monochromator given the high sensitivity of the set up) and Bruker D8 Discover (asymmetric XRD diffraction, equipped with a monochromator) X-ray diffractometers using Cu $K\alpha$ ($\lambda = 1.5405$ Å) radiation.
2.3 Transmission electron microscopy
HRTEM micrographs and fast Fourier transform (FFT) patterns were acquired using a FEI Titan ST transmission electron microscope operating at 300 keV. The EDX detector configuration requires the TEM specimen to be rotated 15$^\circ$ to receive signals during EDX mapping. Moreover, analysis of HRTEM and STEM micrographs, including FFT masking/filtering and inverse FFT were carried out using Gatan DigitalMicrograph. The TEM specimens were prepared using a FEI Helios G4 dual-beam focused ion beam-scanning electron microscope (DBFIB-SEM) system equipped with an OmniProbe micromanipulator and a Ga ion source.
2.4 Device measurements
The photoelectrical performance of the photodetectors was tested under DUV illumination using a 500 W broadband mercury–xenon arc lamp (Newport’s 66142 Hg(Xe) Arc Lamp). Before illuminating the photodetector, the broadband light passed through an Oriel Cornerstone 260 monochromator fitted with a Newport 74060 diffraction grating. The light intensity was precalibrated using a Si-based photodetector and controlled using a set of neutral-density filters. The I–V characteristics were extracted using Kelvin (four-wire) resistance measurement setup and an Agilent 4156C semiconductor parameter analyzer.
3. X-Ray crystallography characterization
A 2$\theta$-scan out-of-plane XRD pattern was acquired from the $\beta$-Ga$_2$O$_3$/TiN/(100)-oriented Si optoelectronic heterostructure, as shown in Fig. 3(A), whereby parallel diffracting planes are detected irrespective of their rotations. The pattern in Fig. 3(A) indicates how the crystal axes are aligned with respect to each other in terms of normal vectors and a family of lattice planes, and one can deduce the following c-axis crystallographic plane relationship between the grown films and the Si substrate,
Fig. 3. X-ray crystallography—(A) 2$\theta$-scan out-of-plane XRD patterns of the grown optoelectronic heterostructure. (B) $\phi$-scan skewed asymmetric XRD measurements for the (100)-oriented Si substrate (top), thin monocrystalline TiN film (middle), and the $\beta$-Ga$_2$O$_3$ film (bottom), of the heterostructure. (C) Illustration of the atomic unit cell configurations in the stack, reflecting relative crystallographic alignment. (D) XRD RC measurements of the (400) [I] and (420) [II] $\beta$-Ga$_2$O$_3$ and (200) [III] and (220) [IV] TiN Bragg reflections.
Figure 3(D-I) and Fig. 3(D-II) show the XRD rocking curve (RC) measurements of the (400) and (420) $\beta$-Ga$_2$O$_3$ plane Bragg reflections, respectively. We used a Gaussian distribution to fit the XRD RC reflection profiles. The (400) and (420) $\beta$-Ga$_2$O$_3$ plane Bragg reflections exhibited FWHM values of around $1.57^\circ$ and $2.42^\circ$. The (400) and (420) $\beta$-Ga$_2$O$_3$ plane Bragg reflections exhibited relatively broad reflection profiles given the heteroepitaxial growth nature of the $\beta$-Ga$_2$O$_3$ thin films on TiN/Si templates. This relatively broad reflection profile can reasonably be associated to the double twofold crystallographic symmetry. We conjecture that because the monoclinic $\beta$-Ga$_2$O$_3$ layer grown on cubic TiN has two unit cell configurations, it has become intrinsically highly defective. Growth-induced oxygen vacancies and carbon impurities from the PLD target also contributed to such higher levels of defect densities. Finally, Fig. 3(D-III) and Fig. 3(D-IV) depict the XRD RC measurements of the (200) and (220) TiN/Si Bragg reflections The (200) TiN plane Bragg reflection exhibited an FWHM of around 0.92$^\circ$, confirming a high-quality epitaxial TiN thin film.
Next, the hybrid structure was investigated through $\phi$-scan skewed asymmetric XRD measurements. Figure 3(B) presents the $\phi$-scan skewed asymmetric XRD measurements for the (100)-oriented Si substrate, TiN thin film, and the $\beta$-Ga$_2$O$_3$ thin film, from top to bottom, respectively. Parallel diffracting planes were detected with respect to their rotations in Fig. 3(B). These measurements reveal how the crystal axes are aligned azimuthally with respect to each other. The results in Fig. 3(B) show that the (220) TiN and Si planes are aligned with each other and that the following crystallographic relationship can be confirmed by combining the results from out-of-plane and in-plane XRD projections:
4. Electron microscopy characterization
Figure 4(A-I) shows a cross-sectional TEM micrograph of an epitaxially grown $\beta$-Ga$_2$O$_3$/TiN/Si hybrid stack for the $\{110\}$ Si axis, confirming the high interface qualities exhibited at the $\beta$-Ga$_2$O$_3$/TiN and TiN/Si interfaces through the sharp layer transitions, whereas the HRTEM micrographs and FFT patterns in Fig. 4(A-II) and Fig. 4(A-III), which come from the orange- and red-colored areas, respectively, highlighted in Fig. 4(A-I), confirm the structural integrity and symmetry of the lattice on the cross sectional TiN/Si [(200), (220), and (020) FFT plane spots] and $\beta$-Ga$_2$O$_3$ [(400), (020), and (210) FFT plane spots] HRTEM micrographs, respectively. The purpose of the marked glue protection layer, composed mainly of platinum (Pt), was to protect the TEM sample during focused ion beam (FIB) milling. Figure 4(B) displays a cross-sectional STEM micrograph of the sample; the inset shows an HAADF micrograph with EDX mapping that confirm each layer’s composition and thickness and the low thermally induced interdiffusion characteristics during layer growth. In the EDX mapping shown in the inset of Fig. 4(B), the N emission peak $K_{\alpha _{1}}$ is centered at 0.392 keV, which is close to the detection limit of our TEM-EDX detector. That is why the N signal is weak and blurred in the EDX map. Moreover, the O mapping appears to show a sign of interdiffusion into TiN layer. This is an artifact as the O emission peak $K_{\alpha _{1}}$ is centered around 0.523 keV, while the Ti emission peaks are situated around 4.931 ($K_{\beta _{1}}$), 4.510 ($K_{\alpha _{1}}$), 0.458 ($L_{\beta _{1}}$), and 0.452 ($L_{\alpha _{1}}$) [37]. Our EDX detector cannot distinguish well between Ti $L_{\beta _{1}}$/$L_{\alpha _{1}}$ and O $K_{\alpha _{1}}$, causing the digital micrograph to take Ti $L_{\beta _{1}}$/$L_{\alpha _{1}}$ signals as O $K_{\alpha _{1}}$ signals.
Fig. 4. Electron microscopy imaging—TEM, HRTEM, and FFT: (A) Cross-sectional TEM micrograph of the $\beta$-Ga$_2$O$_3$/TiN/Si hybrid structure (I), and HRTEM micrographs captured at the $\beta$-Ga$_2$O$_3$/TiN (II) and TiN/Si (III) interfaces with FFT patterns. STEM-HAADF: (B) STEM cross-sectional micrograph of the $\beta$-Ga$_2$O$_3$/TiN/Si hybrid stack with EDX spectra.
Because the $\beta$-Ga$_2$O$_3$/TiN heterostructure was grown on (100)-oriented bulk Si, analysis of HRTEM micrographs for the zone axis with FFT patterns in Fig. 4(A-II) and Fig. 4(A-III) reveal the zone axis of Si and TiN to be ($0\bar {1}\bar {1}$). In each FFT pattern, the upward and rightward directions correspond to the directions of the c- and a-axis, respectively, given the mathematical descriptions that govern FFT spectra because physical and inverse coordinates are correlated. Therefore, the orientation relationships between $\beta$-Ga$_2$O$_3$ and TiN were determined to be
Fig. 5. Electron microscopy image analysis and crystallographic defects—Cross-sectional HRTEM micrographs of the (A) grown $\beta$-Ga$_2$O$_3$ layer taken at a low (I) and elevated (II) thickness levels with corresponding FFT pattern (III and IV).
5. Photocurrent Characteristics and DUV Photodetection
The DUV photodetection characteristics of a photodetector based on the proposed oxide–nitride optoelectronic heterostructure are investigated. The fabricated solar-blind photodetector exhibited a peak $\mathcal {R}_\lambda$ of 249.84 A/W at 12 V reverse bias, for $\lambda _\textrm {in} = 250$ nm and $P_\textrm {in} = 12$ $\mu$W/cm$^2$ ($0.12$ pW/$\mu$m$^2$). Ultraviolet-to-visible rejection ratios ($\mathcal {R}^{15\textrm { V}}_{250\textrm { nm}}/\mathcal {R}^{15\textrm { V}}_{400\textrm { nm}}$) of $1.53 \times 10^{3}$ and $4.26 \times 10^{3}$ were attained for $P_\textrm {in} = 4.04 \pm 0.42$ pW/$\mu$m$^2$ (high $P_\textrm {in}$ level) and $P_\textrm {in} = 12 \pm 1.26$ $\mu$W/cm$^2$ (low $P_\textrm {in}$ level), respectively. The DUV characteristics of the photodetector were measured at low and high $P_\textrm {in}$ levels to broadly analyze its sensitivity and the transport efficiency of photogenerated charge carriers as a function of incident illuminating power. When measured at $\lambda _\textrm {in} = 250$ nm, the photodetector’s sensitivity decreased exponentially as the optical power was increased beyond 12 $\mu$W/cm$^2$, providing DUV photodetection with a self-limiting gain mechanism to avoid electronic saturation at high impinging power levels. Figure 6 depicts measured photocurrent density vs. $V_\textrm {bias}$ ($J_\textrm {ph}$–$V_\textrm {bias}$) for the Si-integrated photodetector at low $P_\textrm {in}$ levels ( Fig. 6(A)) and high $P_\textrm {in}$ levels ( Fig. 6(B)); measured $P_\textrm {in}$-dependent $J_\textrm {ph}$–$V_\textrm {bias}$ curves for the photodetector at an $\lambda _\textrm {in}$ = 250 nm are shown in Fig. 6(C). The Si-integrated device exhibited dark-current densities ($J_\textrm {d}$) of $1.30 \times 10^{-10}$ A/cm$^2$ at zero bias, $1.74 \times 10^{-5}$ A/cm$^2$ at 6 V reverse bias, and $4.79 \times 10^{-5}$ A/cm$^2$ at 12 V reverse bias. From Fig. 6(C), one can observe that the photodetector exhibited photo-to-dark-current ratios of above $5 \times 10^3$.
Fig. 6. Measured $J_\textrm {ph}$–$V_\textrm {bias}$ curves for a representative $\beta$-Ga$_2$O$_3$/TiN/Si photodetector at various wavelengths at an (A) $P_\textrm {in} = 12.25 \pm 1.26$ $\mu$W/cm$^2$ (low $P_\textrm {in}$ levels) and (B) $P_\textrm {in} = 4.04 \pm 0.42$ pW/$\mu$m$^2$ (high $P_\textrm {in}$ levels). (C) Measured $P_\textrm {in}$-dependent $J_\textrm {ph}$–$V_\textrm {bias}$ curves for the photodetector at an $\lambda _\textrm {in}$ = 250 nm.
Given a reverse-biased heterojunction under a uniform incident-light illumination of a particular wavelength $\lambda _\textrm {in}$ and optical power density $P_\textrm {in}$, the spectral responsivity $\mathcal {R}^{V_\textrm {bias}}_{\lambda _\textrm {in},P_\textrm {in}}$ of the resultant photodetector is a parameter that quantizes the internal quantum efficiency ($\eta ^{V_\textrm {bias}}_{\lambda _\textrm {in},P_\textrm {in}}$) and the photoelectric gain ($g$) of that photodetector, which are determined by the generated number of carriers per incident photon as a result of the absorption of incident light and the number of carriers that conduct current through the electrical contact per generated electron–hole pair [38,39] and can be estimated using
Figure 7 plots the evolution in performance of a representative photodetector at low ( Fig. 7(A)) and high ( Fig. 7(B)) $P_\textrm {in}$ levels. In both cases, the photodetector performance peaked at $\lambda _\textrm {in} = 250$ nm. The high responsivity observed in the Si-integrated $\beta$-Ga$_2$O$_3$ photodetector can be primarily attributed to the minority carrier trapping caused by inherent deep-level vacancy states as additional photoexcited electrons can be created from these defect states [41,42]. Through an increased bias voltage, these trapped carriers were subsequently released from the defect-related trap centers resulting in higher photoconductive gain values (EQE $> 10^{5}\%$) as compared to lower bias voltages [43,44]. The possible presence of a trap-related deep-acceptor level at the metal–semiconductor interface provides another hypothesis for such high gain values as this may have led to a reduction in the barrier height (owing to the space charge neutrality principle) and thus increased extracted photocurrent levels [45].
Fig. 7. Photodetector performance metrics—(A) $P_\textrm {in}$-dependent evolution of calculated $\mathcal {R}_\lambda$ (I), $\mathcal {D}^*$ (II), and EQE (III) values for a representative DUV, Si-integrated $\beta$-Ga$_2$O$_3$ photodetector at low $P_\textrm {in}$ levels ($P_\textrm {in} = 12.25 \pm 1.26$ $\mu$W/cm$^2$) at various wavelengths. (B) Corresponding plots for the calculated $\mathcal {R}_\lambda$ (I), $\mathcal {D}^*$ (II), and EQE (III) values at high $P_\textrm {in}$ levels ($P_\textrm {in} = 4.04 \pm 0.42$ pW/$\mu$m$^2$).
Figure 8(A) plots the evolution of average $\mathcal {R}_\lambda$ values with increasing $P_\textrm {in}$ levels up to 73.6 pW/$\mu$m$^2$ at an $\lambda _\textrm {in}$ = 250 nm and a $V_\textrm {bias} = \textrm {4 and 12 V}$. Beyond 12 $\mu$W/cm$^2$, $\mathcal {R}_\lambda$ values converge to 18.09 and 110.04 A/W after peaking at 33.55 and 249.84 A/W at 4 and 12 V, respectively, indicating a limit for UVC photons to excite the photoelectrons from the valence band to the conduction band in the $\beta$-Ga$_2$O$_3$ film. The photodetectors’ average $\mathcal {R}_\lambda$ values decreased in a nonlinear manner, demonstrating nonlinear absorption characteristics. We postulate that this is attributed to the dampening of the net built-in electric field by opposite fields as a result of the increasing number of accumulated photogenerated charges, which caused less efficient photogenerated carrier separation because $P_\textrm {in}$ increased and manifested a self-limiting gain mechanism to prevent electronic saturation at higher impinging power levels. As such, we argue that our photodetectors exhibit high sensitivity to ultralow-power DUV radiation levels that are remarkably lower than what we have previously reported using the TiN/MgO platform [27]. Figure 8(B) and Fig. 8(C) plot the calculated average $\mathcal {D}^*$ and EQE values, respectively, for $\lambda _\textrm {in}$ = 250 nm and $V_\textrm {bias} = \textrm {4 and 12 V}$. The photodetectors exhibited peak $\mathcal {D}^*$ and EQE values of $6.38 \times 10^{13}$ Jones and $1.23 \times 10^{5}\%$ for $\lambda _\textrm {in}$ = 250 nm and $P_\textrm {in}$ = 12 $\mu$W/cm$^2$ at 12 V reverse-bias.
Fig. 8. $P_\textrm {in}$-dependent evolution of calculated (A) $\mathcal {R}_\lambda$, (B) $\mathcal {D}^*$, and (C) EQE values for fabricated Si-integrated $\beta$-Ga$_2$O$_3$ photodetectors at an $\lambda _\textrm {in}$ of 250 nm at various $P_\textrm {in}$ levels.
6. Conclusion
In this work, monocrystalline oxide–nitride $\beta$-Ga$_2$O$_3$/TiN/Si heterostructures were presented and investigated as potential vertically structured optoelectronic device platforms that are compatible with CMOS FEOL processes. The orientation relationships between the grown thin films and the monocrystalline Si substrate as determined from XRD and TEM analysis, were found to be (400) $\beta$-Ga$_2$O$_3$ $\parallel$ (200) TiN $\parallel$ (400) Si and (020) $\beta$-Ga$_2$O$_3$ $\parallel$ (022) TiN $\parallel$ (022) Si. Including a wafer-scalable conductive ceramic interlayer of monocrystalline TiN is hypothesized to have ensured high carrier transport and antioxidation properties for next generation oxide-based optoelectronic device platforms. The vertically structured photodetector realized based on the hybrid oxide–nitride heterostructure exhibited high responsivity and EQE levels of up to 249 A/W and $1.23 \times 10^{5}\%$, respectively, which are ideal performance levels for highly sensitive DUV photodetection applications such as early missile threat detection and interception, indoor and outdoor fire alarm and surveillance systems, and UVC solar radiation observation in ozone-monitoring stations.
Funding
Ministry of Education and Research, Romania (18N/08.02.2019); King Abdullah University of Science and Technology (BAS/1/1614-01-01).
Acknowledgment
The authors acknowledge the access of the Nanofabrication Core Lab as well as the Imaging and Characterization Core Lab facilities at KAUST.
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. N. Alfaraj, K.-H. Li, C. H. Kang, L. Braic, T. K. Ng, and B. S. Ooi, “Epitaxial growth of β-Ga2O3/ɛ-Ga2O3 polymorphic heterostructures on c-plane sapphire for deep-ultraviolet optoelectronics,” Proc. SPIE 11281, 112810G (2020). [CrossRef]
2. S. J. Pearton, J. Yang, P. H. Cary IV, F. Ren, J. Kim, M. J. Tadjer, and M. A. Mastro, “A review of Ga2O3 materials, processing, and devices,” Appl. Phys. Rev. 5(1), 011301 (2018). [CrossRef]
3. N. Alfaraj, J.-W. Min, C. H. Kang, A. A. Alatawi, D. Priante, R. C. Subedi, M. Tangi, T. K. Ng, and B. S. Ooi, “Deep-ultraviolet integrated photonic and optoelectronic devices: A prospect of the hybridization of group III–nitrides, III–oxides, and two-dimensional materials,” J. Semicond. 40(12), 121801 (2019). [CrossRef]
4. N. Alfaraj, A. M. Hussain, G. A. Torres Sevilla, M. T. Ghoneim, J. P. Rojas, A. B. Aljedaani, and M. M. Hussain, “Functional integrity of flexible n-channel metal–oxide–semiconductor field-effect transistors on a reversibly bistable platform,” Appl. Phys. Lett. 107(17), 174101 (2015). [CrossRef]
5. N. Alfaraj, W. Alghamdi, M. Alawein, I. A. Ajia, D. Priante, B. Janjua, H. Sun, T. K. Ng, B. S. Ooi, I. S. Roqan, and X. Li, “Time–energy quantum uncertainty: Quantifying the effectiveness of surface defect passivation protocols for low-dimensional semiconductors,” ACS Appl. Electron. Mater. 2(2), 409–418 (2020). [CrossRef]
6. I. A. Ajia, D. Almalawi, Y. Lu, S. Lopatin, X. Li, Z. Liu, and I. S. Roqan, “Subquantum-well influence on carrier dynamics in high efficiency DUV dislocation-free AlGaN/AlGaN-based multiple quantum wells,” ACS Photonics 7(7), 1667–1675 (2020). [CrossRef]
7. N. Alfaraj, S. Mitra, F. Wu, I. A. Ajia, B. Janjua, A. Prabaswara, R. A. Aljefri, H. Sun, T. Khee Ng, B. S. Ooi, and X. Li, “Photoinduced entropy of InGaN/GaN pin double-heterostructure nanowires,” Appl. Phys. Lett. 110(16), 161110 (2017). [CrossRef]
8. J. Wang, M. Feng, R. Zhou, Q. Sun, J. Liu, Y. Huang, Y. Zhou, H. Gao, X. Zheng, M. Ikeda, and H. Yang, “GaN-based ultraviolet microdisk laser diode grown on Si,” Photonics Res. 7(6), B32–B35 (2019). [CrossRef]
9. T. Bagci, A. Simonsen, S. Schmid, L. G. Villanueva, E. Zeuthen, J. Appel, J. M. Taylor, A. Sørensen, K. Usami, A. Schliesser, and E. S. Polzik, “Optical detection of radio waves through a nanomechanical transducer,” Nature 507(7490), 81–85 (2014). [CrossRef]
10. K. Ahmed, R. Dahal, A. Weltz, J. J.-Q. Lu, Y. Danon, and I. B. Bhat, “Solid-state neutron detectors based on thickness scalable hexagonal boron nitride,” Appl. Phys. Lett. 110(2), 023503 (2017). [CrossRef]
11. D. Alden, T. Troha, R. Kirste, S. Mita, Q. Guo, A. Hoffmann, M. Zgonik, R. Collazo, and Z. Sitar, “Quasi-phase-matched second harmonic generation of UV light using AlN waveguides,” Appl. Phys. Lett. 114(10), 103504 (2019). [CrossRef]
12. W. Hu, C. Zhang, and Z. L. Wang, “Recent progress in piezotronics and tribotronics,” Nanotechnology 30(4), 042001 (2019). [CrossRef]
13. A. Ishii, X. He, N. F. Hartmann, H. Machiya, H. Htoon, S. K. Doorn, and Y. K. Kato, “Enhanced single-photon emission from carbon-nanotube dopant states coupled to silicon microcavities,” Nano Lett. 18(6), 3873–3878 (2018). [CrossRef]
14. C. Lee, C. Shen, C. Cozzan, R. M. Farrell, J. S. Speck, S. Nakamura, B. S. Ooi, and S. P. DenBaars, “Gigabit-per-second white light-based visible light communication using near-ultraviolet laser diode and red-, green-, and blue-emitting phosphors,” Opt. Express 25(15), 17480–17487 (2017). [CrossRef]
15. S. Yang, S. Huang, J. Wei, Z. Zheng, Y. Wang, J. He, and K. J. Chen, “Identification of trap states in p-GaN layer of a p-GaN/AlGaN/GaN power HEMT structure by deep-level transient spectroscopy,” IEEE Electron Device Lett. 41(5), 685–688 (2020). [CrossRef]
16. Y. Baumgartner, D. Caimi, M. Sousa, M. Hopstaken, Y. Salamin, B. Baeuerle, B. I. Bitachon, J. Leuthold, J. Faist, B. J. Offrein, and L. Czornomaz, “High-speed CMOS-compatible III-V onSi membrane photodetectors,” Opt. Express 29(1), 509–516 (2021). [CrossRef]
17. B. Szelag, K. Hassan, L. Adelmini, E. Ghegin, P. Rodriguez, F. Nemouchi, P. Brianceau, E. Vermande, A. Schembri, D. Carrara, P. Cavalié, F. Franchin, M.-C. Roure, L. Sanchez, C. Jany, and S. Olivier, “Hybrid III–V/silicon technology for laser integration on a 200-mm fully CMOS-compatible silicon photonics platform,” IEEE J. Sel. Top. Quantum Electron. 25(5), 1–10 (2019). [CrossRef]
18. Y. Berencén, Y. Xie, M. Wang, S. Prucnal, L. Rebohle, and S. Zhou, “Structural and optical properties of pulsed-laser deposited crystalline β-Ga2O3 thin films on silicon,” Semicond. Sci. Technol. 34(3), 035001 (2019). [CrossRef]
19. N. Alfaraj, K.-H. Li, C. H. Kang, D. Priante, L. Braic, Z. Guo, T. K. Ng, X. Li, and B. S. Ooi, “Electrical characterization of solar-blind deep-ultraviolet (Al0.28Ga0.72)2O3 Schottky photodetectors grown on silicon by pulsed laser deposition,” in Conference on Lasers and Electro-Optics (CLEO), (Optical Society of America, 2019), p. SF2O.1.
20. H. Reddy, U. Guler, Z. Kudyshev, A. V. Kildishev, V. M. Shalaev, and A. Boltasseva, “Temperature-dependent optical properties of plasmonic titanium nitride thin films,” ACS Photonics 4(6), 1413–1420 (2017). [CrossRef]
21. P. Ranga, A. Bhattacharyya, A. Chmielewski, S. Roy, R. Sun, M. A. Scarpulla, N. Alem, and S. Krishnamoorthy, “Growth and characterization of metalorganic vapor-phase epitaxy-grown β-(AlxGa1 − x)2O3/β-Ga2O3 heterostructure channels,” Appl. Phys. Express 14(2), 025501 (2021). [CrossRef]
22. J. Casamento, J. Wright, R. Chaudhuri, H. Xing, and D. Jena, “Molecular beam epitaxial growth of scandium nitride on hexagonal SiC, GaN, and AlN,” Appl. Phys. Lett. 115(17), 172101 (2019). [CrossRef]
23. S. A. Kukushkin, A. M. Mizerov, A. V. Osipov, A. V. Redkov, and S. N. Timoshnev, “Plasma assisted molecular beam epitaxy of thin GaN films on Si (111) and SiC/Si (111) substrates: Effect of SiC and polarity issues,” Thin Solid Films 646, 158–162 (2018). [CrossRef]
24. M. N. Gadalla, K. Chaudhary, C. M. Zgrabik, F. Capasso, and E. L. Hu, “Imaging of surface plasmon polaritons in low-loss highly metallic titanium nitride thin films in visible and infrared regimes,” Opt. Express 28(10), 14536–14546 (2020). [CrossRef]
25. M. N. Gadalla, A. S. Greenspon, M. Tamagnone, F. Capasso, and E. L. Hu, “Excitation of strong localized surface plasmon resonances in highly metallic titanium nitride nano-antennas for stable performance at elevated temperatures,” ACS Appl. Nano Mater. 2(6), 3444–3452 (2019). [CrossRef]
26. J. Libardi, K. G. Grigorov, M. Massi, A. S. da Silva Sobrinho, R. S. Pessoa, and B. Sismanoglu, “Diffusion of silicon in titanium dioxide thin films with different degree of crystallinity: Efficiency of TiO2 and TiN barrier layers,” Vacuum 128, 178–185 (2016). [CrossRef]
27. N. Alfaraj, K.-H. Li, M. Alawein, C. H. Kang, L. Braic, N. C. Zoita, A. E. Kiss, T. K. Ng, and B. S. Ooi, “Heteroepitaxial β-Ga2O3 on conductive ceramic templates: Towards ultrahigh gain deep-ultraviolet photodetection,” Adv. Mater. Technol. 6(9), 2100142 (2021). [CrossRef]
28. C. C.-C. Lin, P.-H. Chang, and A. S. Helmy, “Supermode hybridization: A material-independent route toward record Schottky detection sensitivity using <0.05 µm3 amorphous absorber volume,” Nano Lett. 20(12), 8500–8507 (2020). [CrossRef]
29. C.-C. Chang, J. Nogan, Z.-P. Yang, W. J. Kort-Kamp, W. Ross, T. S. Luk, D. A. Dalvit, A. K. Azad, and H.-T. Chen, “Highly plasmonic titanium nitride by room-temperature sputtering,” Sci. Rep. 9(1), 15287–15289 (2019). [CrossRef]
30. N. Alfaraj, “Deep-ultraviolet optoelectronic devices enabled by the hybrid integration of next-generation semiconductors and emerging device platforms,” Ph.D. thesis, King Abdullah University of Science and Technology (2019).
31. N. C. Zoita, M. Dinu, A. E. Kiss, C. Logofatu, and M. Braic, “A comparative investigation of hetero-epitaxial TiC thin films deposited by magnetron sputtering using either hybrid DCMS/HiPIMS or reactive DCMS process,” Appl. Surf. Sci. 537, 147903 (2021). [CrossRef]
32. V. Timoshevskii, Y. Ke, H. Guo, and D. Gall, “The influence of surface roughness on electrical conductance of thin Cu films: An ab initio study,” J. Appl. Phys. 103(11), 113705 (2008). [CrossRef]
33. R. C. Munoz, G. Vidal, M. Mulsow, J. G. Lisoni, C. Arenas, A. Concha, F. Mora, R. Espejo, G. Kremer, L. Moraga, R. Esparza, and P. Haberle, “Surface roughness and surface-induced resistivity of gold films on mica: Application of quantitative scanning tunneling microscopy,” Phys. Rev. B 62(7), 4686–4697 (2000). [CrossRef]
34. N. Alfaraj, K.-H. Li, L. Braic, N. C. Zoita, and A. E. Kiss, “Heteroepitaxial growth method of compound semiconductor materials on multi-oriented semiconductor substrates and devices,” PCT international application no. PCT/IB2021/050172. International publication no. WO2021144681A1 (2021).
35. R. R. Sumathi, “Status and challenges in hetero-epitaxial growth approach for large diameter AlN single crystalline substrates,” ECS J. Solid State Sci. Technol. 10(3), 035001 (2021). [CrossRef]
36. R. R. Sumathi, R.-U. Barz, P. Gille, and T. Straubinger, “Influence of interface formation on the structural quality of AlN single crystals grown by sublimation method,” Phys. Status Solidi C 8(7-8), 2107–2109 (2011). [CrossRef]
37. P. Patnaik, Dean’s Analytical Chemistry Handbook (McGraw-Hill Education, 2004).
38. Y. Fang, A. Armin, P. Meredith, and J. Huang, “Accurate characterization of next-generation thin-film photodetectors,” Nat. Photonics 13(1), 1–4 (2019). [CrossRef]
39. A. Rose, Concepts in Photoconductivity and Allied Problems, 19 (Interscience Publishers, 1963).
40. A. Axelevitch, B. Gorenstein, and G. Golan, “Investigation of optical transmission in thin metal films,” Phys. Procedia 32, 1–13 (2012). [CrossRef]
41. Z. Zhang, E. Farzana, A. R. Arehart, and S. A. Ringel, “Deep level defects throughout the bandgap of (010) β-Ga2O3 detected by optically and thermally stimulated defect spectroscopy,” Appl. Phys. Lett. 108(5), 052105 (2016). [CrossRef]
42. K. Irmscher, Z. Galazka, M. Pietsch, R. Uecker, and R. Fornari, “Electrical properties of β-Ga2O3 single crystals grown by the Czochralski method,” J. Appl. Phys. 110(6), 063720 (2011). [CrossRef]
43. Y. Qin, L. Li, X. Zhao, G. S. Tompa, H. Dong, G. Jian, Q. He, P. Tan, X. Hou, Z. Zhang, S. Yu, H. Sun, G. Xu, X. Miao, K. Xue, S. Long, and M. Liu, “Metal–semiconductor–metal ɛ-Ga2O3 solar-blind photodetectors with a record-high responsivity rejection ratio and their gain mechanism,” ACS Photonics 7(3), 812–820 (2020). [CrossRef]
44. D. Guo, Z. Wu, P. Li, Y. An, H. Liu, X. Guo, H. Yan, G. Wang, C. Sun, L. Li, and W. Tang, “Fabrication of β-Ga2O3 thin films and solar-blind photodetectors by laser MBE technology,” Opt. Mater. Express 4(5), 1067–1076 (2014). [CrossRef]
45. H. Zhou, L. Cong, J. Ma, B. Li, M. Chen, H. Xu, and Y. Liu, “High gain broadband photoconductor based on amorphous Ga2O3 and suppression of persistent photoconductivity,” J. Mater. Chem. C 7(42), 13149–13155 (2019). [CrossRef]
