(GaN) n
Nanoscale View Article Online
Published on 08 October 2014. Downloaded by University of Texas Libraries on 06/01/2015 21:14:12.
PAPER
Cite this: Nanoscale, 2014, 6, 14733
View Journal | View Issue
Quantum state engineering with ultra-shortperiod (AlN)m/(GaN)n superlattices for narrowband deep-ultraviolet detection Na Gao,a Wei Lin,a Xue Chen,a Kai Huang,*a,b Shuping Li,a Jinchai Li,a Hangyang Chen,a Xu Yang,a Li Ji,b Edward T. Yub and Junyong Kang*a Ultra-short-period (AlN)m/(GaN)n superlattices with tunable well and barrier atomic layer numbers were grown by metal–organic vapour phase epitaxy, and employed to demonstrate narrowband deep ultraviolet photodetection. High-resolution transmission electron microscopy and X-ray reciprocal space mapping confirm that superlattices containing well-defined, coherently strained GaN and AlN layers as thin as two atomic layers (∼0.5 nm) were grown. Theoretical and experimental results demonstrate that an optical absorption band as narrow as 9 nm (210 meV) at deep-ultraviolet wavelengths can be produced, and is attributable to interband transitions between quantum states along the [0001] direction in
Received 28th July 2014, Accepted 5th October 2014
ultrathin GaN atomic layers isolated by AlN barriers. The absorption wavelength can be precisely engineered by adjusting the thickness of the GaN atomic layers because of the quantum confinement effect.
DOI: 10.1039/c4nr04286g
These results represent a major advance towards the realization of wavelength selectable and narrowband
www.rsc.org/nanoscale
photodetectors in the deep-ultraviolet region without any additional optical filters.
Introduction The deep-ultraviolet (DUV) wavelength region between 200 and 280 nm has become indispensable for protein analysis, DNA sequencing, drug discovery and detection, forensic analysis, disinfection, surface and water decontamination, as well as for many other applications.1–6 Spurred by the demand for DUV detection and measurement technology, a variety of solid-state and vacuum devices can be made sensitive to different wavelength bands using bandpass interference filters and quartz lens systems.7–11 However, these approaches can suffer from the unwanted response,12 various instabilities,13 and device complexity.14 To eliminate the filters that are widely used in photodetectors, passive filter layers have recently been integrated into device structures during growth. Although the bandwidth has narrowed from 55 nm to 30 nm to limit the short-wavelength responsivity,15 the propensity for degradation, bulkiness, and unreliability of the devices remain unsolved in these systems, which combine a series of typical filters and phosphor coatings inserted into the optical path of
a
Department of Physics, Fujian Provincial Key Laboratory of Semiconductor Materials and Applications, Xiamen University, Xiamen, 361005, China. E-mail: [email protected], [email protected]; Fax: +86-592-2187737; Tel: +86-592-2185962 b Department of Electrical and Computer Engineering, Microelectronic Research Centre, The University of Texas at Austin, Austin, Texas, 78758, USA
This journal is © The Royal Society of Chemistry 2014
the sensor system.14,16 To overcome these limitations, semiconductor nanostructures are promising because of the possibility of achieving high wavelength-selective interband absorption between quantum states. Compared to the zerodimensional (0D) structures, i.e. nanodots or one-dimensional (1D) structures, i.e. nanowires, ultrathin two-dimensional (2D) layers (one or a few atomic layers) have an additional advantage that current can easily conduct in the 2D layers because of the relatively high in-plane conductivity. Thus photodetectors using ultrathin 2D layers as absorption materials would have higher external quantum efficiency (QE). Recent experimental results demonstrated that the bandgap energy of 2D MoS2 crystals decreases with the number of the layers indicating the existence of the quantum states.17 However, the quantum state in ultrathin 2D layers can only be applied to narrowband photodetectors when the incident photons can be made to propagate perpendicularly to the ultrathin 2D layers. The optical absorption is extremely low for the incident light perpendicular to the single 2D atomic layer (∼2.3% for monolayered graphene) due to the extremely short absorption optical length,18,19 restricting the optical response of the ultrathin 2D layer photodetectors. Although the optical response can also be enhanced by grazing or scattering the light onto the ultrathin 2D layers,20 in this geometry optical absorption occurs between continuous energy bands rather than between quantum states. To increase the optical length of the incident light absorption, a number of layers can be stacked along the
Nanoscale, 2014, 6, 14733–14739 | 14733
View Article Online
Published on 08 October 2014. Downloaded by University of Texas Libraries on 06/01/2015 21:14:12.
Paper
normal plane of the ultrathin 2D structures using transparent layers to isolate the quantum state. It has been proven that three hBN atomic layers could maintain the quantized characteristics of graphene due to the sufficiently high potential barrier provided by hBN.21–23 The 2.4% lattice mismatch between GaN and AlN makes growth of coherently strained individual layers and thick multilayered structures challenging. To ensure efficient absorption of incident ultraviolet photons, a large number of superlattice periods are required. In addition to optical absorption, good electrical conductivity in the vertical direction, i.e. perpendicular to the superlattice interfaces, is a crucial issue for the electrical responsivity of photodetectors. Thus, the AlN layers must be sufficiently thin to allow for efficient carrier transport via tunnelling, while at the same time be thick enough to provide effective quantum confinement. Minimization of material defects is also essential for high performance, as defects can lead to carrier trapping and high dark currents. Avoidance of strain relaxation will therefore also limit the maximum individual layer thickness and the total superlattice thickness that can be employed. In this paper, we propose and demonstrate a unique narrowband DUV photodetector structure in which a high degree of wavelength selectivity is achieved without additional filters using interband absorption between the quantum states in ultra-short-period (AlN)m/(GaN)n superlattices (SLs), where m and n represent the number of atomic layers in the AlN and GaN regions, respectively. Through computational modeling and atomic-scale epitaxial growth, a series of metal–semiconductor–metal (MSM) DUV photodetectors with tunable detection wavelengths and bandwidths as narrow as 9 nm (210 meV) were successfully designed, fabricated, and characterized. Our results indicate that the use of ultrathin semiconductor 2D layers can provide new classes of narrowband photodetectors without additional optical filters.
Computational and experimental methods First-principles calculations of the electronic structure of ultrashort-period (AlN)m/(GaN)n SLs were carried out using the Vienna ab-initio simulation package (VASP) in the framework of density-functional theory (DFT).24,25 The exchange-correlation function was described within the generalized-gradient approximation (GGA).26 The pseudopotentials were generated by means of the projector augmented wave (PAW) method27 that allows for the accurate treatment of the valence s and p electrons as well as the semicore Ga 3d states. With an 8 × 8 × 4 Monkhorst-Pack grid of k points sampling the Brillouin zone, a 500 eV cutoff energy was used to expand the electronic wave functions, which was sufficient for the plane wave basis to obtain well-converged results. The size of the supercell was fixed and the atoms within the cells were allowed to relax to minimize the total energy of the system with a convergence criterion of 0.1 meV. By introducing a scissors correction with the shifting parameter of 1.75 eV, the underestimated bandgap
14734 | Nanoscale, 2014, 6, 14733–14739
Nanoscale
values were then adjusted. The imaginary part of the dielectric function for optical absorption was computed using Fermi’s Golden Rule for the optical transition rate.28 To determine the maximum confinement energy achievable in an ultrathin GaN layer and verify the band-like nature of the electronic structure in such a layer, two atomic layers of GaN and a bulk GaN crystal were treated as slab and crystal structures, respectively. It should be noted that, in the case of a slab, introducing vacuum is necessary. Next, a series of ultrashort-period (AlN)m/(GaN)n SLs were constructed along the [0001] direction with different AlN layer numbers m = 4, 6 and 8 in the (1a × 1a × 3c), (1a × 1a × 4c) and (1a × 1a × 5c) supercell to examine whether the barrier was sufficiently high to maintain the quantized characteristics of the wide bandgap GaN ultrathin 2D layers, respectively. Finally, by varying the atomic layer number of the GaN ultrathin 2D layers (n = 1, 2, 4) while the atomic layer number of AlN was kept approximately constant at m = 6, tunable optical absorption at different peak energies by quantum state engineering was realized in the DUV spectra range. Experimentally, a series of ultra-short-period (AlN)m/(GaN)n SLs were epitaxially grown on c-plane sapphire substrates via metal–organic vapour phase epitaxy (MOVPE) in a vertical Thomas Swan system (3 × 2 inch CCS Aixtron). The source precursors were trimethylaluminum (TMA), trimethylgallium (TMG), trimethylindium (TMI), and ammonia (NH3), and hydrogen (H2) was used as the carrier gas. An AlN buffer layer with a thickness of approximately 1.0 µm was first grown at 1100 °C, followed by 300 periods of (AlN)m/(GaN)n SLs with varying well and barrier atomic layer numbers. The atomic layer numbers were altered by changing the flow rates of source precursors. During the ultra-short-period SL growth, a brief growth interruption was adopted to separate the ultrathin GaN and AlN layers. The nominal thickness of one monolayer is 2.6 Å for GaN and 2.5 Å for AlN, respectively. The surface morphology and crystalline quality of the ultra-short-period (AlN)m/(GaN)n SLs were characterized by a combination of an atomic force microscope system (AFM, SPA400), a high-resolution X-ray diffractometer (HRXRD, PANalytical X’ Pert PRO) with an X-ray wavelength of 0.154056 nm using Cu Kα radiation, and a high-resolution transmission electron microscope system (FEI Tecnai G2 F20 S-Twin HRTEM) operating at 200 kV. Subsequently, by means of standard photolithography (Karlsuss MA6/BA6), electron-beam deposition (Temescal FC2000) and lift-off processes, MSM DUV photodetectors were fabricated on the ultra-short-period (AlN)m/(GaN)n SLs. For the interdigitated MSM structures reported in this work, the fingers were 300 µm long and 6 µm wide with a spacing of 6 µm. Titanium/gold (10/200 nm) was then evaporated to form Schottky contacts. The fabricated photodetectors were finally annealed by rapid thermal annealing (RTA 300) under a N2 atmosphere at 400 °C for 300 s. For the fabricated devices, I–V characteristics were measured using a Keithley 2410 sourcemeter and a Keithley 6514 programmable electrometer. A 450 W Xe arc lamp, a mechanical chopper and a lock-in amplifier
This journal is © The Royal Society of Chemistry 2014
View Article Online
Nanoscale
Paper
Published on 08 October 2014. Downloaded by University of Texas Libraries on 06/01/2015 21:14:12.
were used to measure the photocurrent response spectra. The system was calibrated with a UV-enhanced Si detector. The devices were all illuminated perpendicularly from the front metal/semiconductor contact side. All measurements were carried out at room temperature.
Results and discussion Quantum state engineering in GaN 2D layers and ultra-shortperiod (AlN)m/(GaN)n SLs To understand the underlying mechanism that determines the optical properties, first-principles calculations of electronic structures of ultrathin 2D GaN layers were performed and compared with that of bulk GaN. When reducing the GaN thickness to as thin as two atomic layers to form a 2D structure, along the high symmetry lines from Γ to A in the Brillouin zone corresponding to the [0001] direction, i.e. perpendicular to the ultrathin 2D layers, the electronic states near the band edge show completely discrete quantum states, as shown in Fig. 1a. In contrast, the electronic structure for the conduction (CB) and valence (VB) bands in bulk GaN, shown in Fig. 1b, exhibits the usual strong dispersion around the Γ-point. Owing to selection rules,29 electronic transitions in the 2D GaN layer can only occur between the quantum states, such as 1h–1e, thereby enabling narrowband optical absorption at a specific wavelength. Fig. 2a shows the electronic structure computed for an ultra-short-period (AlN)6/(GaN)2 SL. One can observe that the well-defined quantum states are still identified as 1h and 1e within the range of Γ to A, i.e. perpendicular to the ultrathin 2D layers. However, the energy dispersion of the SLs is slightly larger than that of GaN 2D atomic layers in a vacuum. More than one quantum states are expected to exist for both elec-
Fig. 1 Comparison of electronic structures for GaN 2D layer and bulk crystals. (a) GaN 2D structure in a vacuum. Completely discrete quantum states are shown along the symmetry lines (vertical dashed red lines) from Γ to A in the Brillouin zone, in contrast to a strong dispersion around the Γ-point in (b) GaN bulk materials.
This journal is © The Royal Society of Chemistry 2014
Fig. 2 Electronic structures and optical absorption properties of the designed ultra-short-period (AlN)m/(GaN)n SLs. (a) Typical electronic structures for the stacked (AlN)6/(GaN)2 as well as for the crystal structure of the supercell, with the metal atoms Ga and Al in green and blue, respectively. (b, c) Normalized imaginary part of the dielectric function εxx + εyy (εxx = εyy) when unpolarized light is perpendicularly incident to the 2D atomic layers (in the directions). (b) The stacked GaN 2D structure (two atomic layers) isolated by different AlN atomic layer numbers m: 4 (blue line), 6 (red line) and 8 (green line) atomic layers. (c) Stacked GaN 2D atomic layers by different atomic layer numbers n: 4 (blue line), 2 (red line) and 1 (green line); all are isolated by the same barrier of approximately 6 AlN atomic layers.
trons and holes in the GaN 2D layer because of the large bandgap difference between the AlN barrier and the GaN well.30 The dispersive quantum states in the CB are denoted as 1e, 2e, etc. While in the VB, the number of dispersionless quantum states is greater. The interband optical absorption of unpolarized light incident normal to the 2D ultrathin layers (i.e. in the directions) is proportional to the imaginary part of the dielectric function,31 εxx + εyy, plotted in Fig. 2b–2c. As shown in Fig. 2b, the peak energies of the 1h–1e transition are approximately the same among the (AlN)m/(GaN)n SLs with different AlN atomic layer numbers m. This fact clearly demonstrates that the AlN potential barriers are sufficiently high to maintain the optical absorption between the 1h–1e quantum states of the bilayer GaN 2D atomic layers, even when the AlN barrier thickness is reduced to 4 atomic layers. Additionally, on the high energy side of the 1h–1e peak, other weaker but broader peaks can be also recognized with 20–40% peak intensities of the main peak. Based on the calculated band structures, these peaks peaked at approximately 5.10 eV and 5.50 eV are mainly derived from the 2h–1e and 1h–2e transitions, respectively. One can see from Fig. 2a that the 1h is accompanied by discrete quantum states 2h and 3h as well as the crystal-field split-off-hole (ch) band, which consists of atomic p orbitals parallel to the directions.32 Therefore, the transition from ch to the 1e devotes no contribution to the optical absorption along the directions for light perpendicular to the ultrathin 2D layers. Thus the absorption associated with the transition between ch and 1e is not shown
Nanoscale, 2014, 6, 14733–14739 | 14735
View Article Online
Published on 08 October 2014. Downloaded by University of Texas Libraries on 06/01/2015 21:14:12.
Paper
in Fig. 2b. These results show that the transition probability of the 1h–1e interband transition is much higher than that of the other higher energy interband transitions. It should be noted that a slight energy dispersion along the [0001] direction of approximately 97 meV occurs in the bottom of the CB, as shown in Fig. 2a, which enables the electrons to tunnel through the isolating AlN barrier and provide a parallel connection through the stacked GaN 2D atomic layers. By altering the thickness of the AlN barrier layer, the electron transport in the direction perpendicular to the 2D atomic layers changes due to variations in the coupling of the density of states (DOS) between the adjacent GaN 2D atomic layers.33,34 Furthermore, the interband absorption 1h–1e exhibits a remarkably narrow bandwidth, with a full width at half maximum (FWHM) of approximately 230 meV for the AlN isolated atomic layer with m greater than 6. Although the FWHM of the 4 atomic layer AlN is broader, the energy dispersion along the [0001] direction in the bottom of the CB will be favourable for electron quantum tunnelling between the GaN 2D atomic layers. Because of the small changes in the absorption wavelength in response to altering the AlN layer thickness, we can more reliably tune the quantum states by changing the atomic layer number n of the GaN ultrathin 2D layers, as shown in Fig. 2c. When the atomic layer number is reduced from 4 to 1, the optical absorption energy from the 1h to 1e transition clearly increases in the DUV region. On the other hand, the associated shoulder peaks originated from the higher energy transitions, such as 2h–1e, increase when reducing the atomic layer number responded by an upshift of the 1e quantum states in the thinner GaN 2D layers. Thus, we implemented narrowband photodetection with tunable wavelengths in the DUV region by altering the atomic layer number n of the GaN 2D layers rather than that of the AlN barrier layer. Characteristic of ultra-short-period (AlN)m/(GaN)n SLs Based on these computations, a series of narrowband DUV photodetectors using ultra-short-period (AlN)m/(GaN)n SLs was grown, fabricated, and characterized. To confirm the achievement of a coherent AlN/GaN heterostructure, a series of (AlN)m/(GaN)n SLs grown on the atomic scale were characterized by AFM, HRXRD and HRTEM. An extremely smooth surface with atomic-step terraces is shown in Fig. 3a, with the surface root-mean-square (rms) roughness determined to be 1.7 Å in a 500 × 500 nm2 scan area, which indicates abrupt interfaces between the AlN/GaN heterostructures on the atomic scale. Fig. 3b displays the asymmetric (105) reciprocal space mapping (RSM) of the samples, where the two coordinates (Qx, Qy) in the map correspond to a pair of lattice constants (a, c).35 The main peaks are aligned at nearly the same Qx value (denoted by a dashed blue line), while the reciprocal vector perpendicular to the surface is resolved. This result confirms that the AlN/GaN heterostructure is almost coherently grown on the AlN/sapphire templates and is fully strained. The HRXRD patterns of (0002) ω/2θ scans further demonstrate a high degree of periodicity and excellent interface abruptness, with well-defined satellite diffraction peaks until the third
14736 | Nanoscale, 2014, 6, 14733–14739
Nanoscale
Fig. 3 Surface morphology and crystal quality of epitaxially grown ultra-short-period (AlN)m/(GaN)n SLs. (a) AFM image with 1.7 Å rms roughness and (b) the asymmetric (105) RSM of the periodic stacked (AlN)m/(GaN)n SLs grown on AlN/sapphire templates. The dashed blue line indicates that the AlN/GaN heterostructure is almost coherent. (c) The HRXRD patterns of the ω/2θ (0002) scans, with the layer thicknesses of the GaN determined to be 1 ( purple line), 2 (orange line) and 4 (blue line) atomic layers, respectively. (d) Representative cross-sectional TEM images. (e) Fourier transform (FFT) imaged along the [101¯0] zone axis, shown with the SAED patterns (f ). The inset of (f ) reveals period-perfect ultra-short-period (AlN)m/(GaN)n SLs without composition fluctuations. a, b, d, e, and f correspond to the sample with 2 atomic layers of GaN.
order, as shown in Fig. 3c. By combining Vegard’s law and angular separations between superlattice satellite peaks, the thicknesses of the GaN 2D layers were determined to be 1, 2 and 4 atomic layers, respectively.36 These values are in good agreement with the growth parameters and the design expectations, which demonstrates reliable, precise epitaxial growth on the atomic scale, even for a single monolayer of GaN, via MOVPE. To examine the crystalline features in detail, crosssectional TEM and its Fourier transform (FFT) as well as selected area electron diffraction (SAED) were carried out on the samples. As shown in Fig. 3d, distinct AlN/GaN fringes with bright/dark contrast, clear and atomically sharp interfaces and two atomic layers (∼0.5 nm) of GaN 2D layers were recogˉ0] zone nized. High-resolution FFT images taken along the [101 axis in Fig. 3e further confirmed that the stacked AlN/GaN heterointerface is almost coherently grown. Moreover, Fig. 3f revealed that the (AlN)m/(GaN)n SLs were period-perfect without composition fluctuations.37 Therefore, we concluded that the
This journal is © The Royal Society of Chemistry 2014
View Article Online
Nanoscale
GaN 2D layers could be stacked periodically and isolated coherently by AlN layers without defect states at the interfaces.
Published on 08 October 2014. Downloaded by University of Texas Libraries on 06/01/2015 21:14:12.
Narrowband DUV detection based on ultra-short-period (AlN)m/(GaN)n SLs Using the (AlN)m/(GaN)n SLs, MSM DUV photodetectors were then fabricated in the geometry shown in Fig. 4a. An optical micrograph of a representative fabricated device is shown in Fig. 4b. The typical dark current–voltage (I–V) characteristics and the I–V characteristics under illumination at room temperature exhibit symmetrical Schottky features, as shown in Fig. 4c. The dark current at 20 V is estimated to be as low as 4.8 × 10−13 A due to the UV-transparent AlN isolating layer in the device. It is worth noting that all the fabricated MSM DUV photodetectors exhibit narrowband spectral responses at relevant wavelengths. For the GaN 2D layers with atomic layer numbers of 1, 2, 4 and 6 isolated by the same barriers of approximately 6 AlN atomic layers, the corresponding wavelengths appear at 230, 240, 248 and 266 nm (Fig. 4d–g), indicating the dominantly optical transition from 1h to 1e respectively. Notably, the full width at half maximum (FWHM) was as narrow as 9 nm (210 meV) at wavelengths of 230 and 240 nm. These experimental results were similar to the calculated imaginary part of the dielectric function εxx + εyy shown in Fig. 2b–c. However, besides the optical absorption peak related to the 1h–1e interband transition, the optical absorp-
Fig. 4 (a) Three-dimensional schematic of the designed structure and (b) top-view optical micrograph of the device. (c) Typical dark current– voltage (I–V) characteristics (black line) as well as the current–voltage characteristics under light illumination (red line). (d–g) Normalized photocurrent spectra measured at room temperature as a function of wavelength under different biases with light illumination. Wavelength tunable narrowband photodetection appears at wavelengths of 230, 240, 248 and 266 nm in the DUV region, respectively.
This journal is © The Royal Society of Chemistry 2014
Paper
tion peaks related to the 2h–1e and 1h–2e transitions shown in Fig. 2b–c cannot be found in the experimental photocurrents. This behaviour has yet to be explored and might be attributed to two possible effects. First, the incident light power decreases sharply as the DUV wavelength shortens, which leads to low optical absorption. Second, it can be assumed that the binding energy of excitons increases greatly as a result of the strong confinement in such high-quality and extremely thin GaN 2D layers. It is known that the 2D exciton can greatly enhance the band edge absorption.38 Thus the experimental narrowband photodetection might be attributed to the optical absorption of the 1h–1e excitonic transition. Previous investigation on the InGaAs/GaAs MQWs indicated that interband excitonic transitions were difficult to be observed at room temperature except for the 1hh–1e excitonic transition. Therefore, this might be another reason that the shoulder peaks with higher energy were not found experimentally. We deduce the unique mechanism of generation and collection of the photo-induced carriers in the (AlN)m/(GaN)n SLs, which is described as follows. (i) As a result of quantum confinement, quantum states are generated in contrast to continuous energy bands. (ii) When light is incident on the device perpendicular to the GaN 2D atomic layers, electrons in the discrete quantum states of the VB are excited to the CB and then form photo-induced carriers. (iii) Under bias, the photo-induced carriers generate photocurrent in external circuits. Finally, we optimized the carrier transport properties of the (AlN)m/(GaN)n SL photodetectors by changing the value of m. Reducing the AlN barrier thickness in the (AlN)m/(GaN)n SLs was expected to enable more efficient vertical carrier transport. Since the influence of electron quantum tunnelling can be described by the energy dispersion along the [0001] direction in the bottom of the CB with relevant electronic structures, the dispersion energy in the (AlN)6/(GaN)2 was determined to be approximately 97 meV, in contrast to 236 meV in (AlN)4/ (GaN)2. Thus, we decreased the isolated AlN barrier to 4 atomic layers to enhance the coupling of DOS between the adjacent GaN 2D layers. One can see from Fig. 5a that the photocurrent spectrum of the photodetector fabricated by (AlN)4/(GaN)2 peaks at approximately the same wavelength as that fabricated by (AlN)4/(GaN)2. However, the photocurrent of the (AlN)4/(GaN)2 photodetector was more than 20 times higher than that of the (AlN)6/(GaN)2 photodetector when the bias voltage was 5 V. Additionally, the FWHM of the photocurrent peak broadened from 9 nm to 12 nm when the atomic layer number of AlN was reduced from 6 to 4. These results are in good agreement with the calculated optical absorption characteristics shown in Fig. 2b. Fig. 5b–c show the critical parameters of the typical MSM DUV photodetector fabricated by (AlN)4/(GaN)2, spectral responsivity and external QE. The responsivity is given by the following equation: R ¼ I ph =P ¼ ηGq=hν
ð1Þ
where Iph and P represent the photocurrent and the illumination light power,39 respectively, G is the photogain, q is the
Nanoscale, 2014, 6, 14733–14739 | 14737
View Article Online
Paper
Nanoscale
Published on 08 October 2014. Downloaded by University of Texas Libraries on 06/01/2015 21:14:12.
quantum states along the [0001] direction in (AlN)m/(GaN)n SLs. The unique narrow bandwidth (down to 9 nm) and wavelength tunable properties enable real-time, high resolution wavelength-selective imaging without additional filters in the DUV region to selectively separate the signals of biological/ chemical components. This work further makes it possible to develop multiple wavelength narrowband detection integrated on a single chip by adjusting the interband quantum states (atomic layer number) in the direction perpendicular to the 2D layers. We believe that by introducing other stacked 2D atomic layers, such as GaN/InN SLs, tunable narrow-bandwidth photodetectors in the visible region may be another promising candidate for digital cameras beyond colour charge-coupled devices (CCDs) with filters.
Author contributions N.G. and W.L. contributed equally to this work. K.H. and J.Y.K. conceived and supervised this research. N.G., S.P.L., H.Y.C. and J.Y.K. carried out the MOVPE growth. N.G. and J.C.L contributed to the sample characterization. N.G., X.C. and X.Y. fabricated the device, and N.G. X.C. and K.H. performed the measurements. W.L. and N.G. contributed to the computational simulations. All authors analysed the data and wrote the manuscript.
Acknowledgements
Fig. 5 (a) Photocurrent spectra of the typical MSM DUV photodetectors fabricated by (AlN)4/(GaN)2 (red line) and (AlN)6/(GaN)2 (black line) when the bias voltage is 5 V. (b) The spectral responsivity and (c) the external QE of the photodetector based on (AlN)4/(GaN)2 SLs at a wavelength of 240 nm with a narrow bandwidth under different biases with light illumination. The optical power density of the illumination is ∼0.013 µW mm−2.
electronic charge, h is Planck’s constant and ν is the frequency of the incident wavelength.40 Obviously, the spectral responsivity increases with increasing bias, reaching 51 mA W−1 under a bias of 40 V at a wavelength of 240 nm. The external QE η is further estimated to be up to ∼26%.
Conclusions In summary, we demonstrated MSM photodetectors in the DUV spectrum range by interband absorption between
14738 | Nanoscale, 2014, 6, 14733–14739
The authors appreciate Yanyan Fang and Changqing Chen at Huazhong University of Science Technology as well as Jiafa Cai at Xiamen University for the HRXRD support, Mutong Niu and Ke Xu at Suzhou Institute of Nano-Tech and Nano-Bionics, Chinese Academy of Sciences for TEM assistance. The authors also acknowledge Shaoxiong Wu at Xiamen University for technical support for the device measurements. This work was supported by the National Research Program of China under grant no. 2012CB619300, 2011CB301905 and 2011CB925600, the National Natural Science Foundation of China under grant no. 61227009, 61108064, 91321102, 11204254 and 11404271, the Fundamental Research Funds for the Central Universities (2011120143) and in part by the scholarship from the China Scholarship Council (CSC) under the Grant CSC no. 201306315030. This work was also supported by the Natural Science Foundations of Fujian Province (2012J01024).
Notes and references 1 J. A. Nichols and S. K. Katiyar, Arch. Dermatol. Res., 2010, 302, 71. 2 V. Nandakumar, M. Vaid, T. O. Tollefsbol and S. K. Katiyar, Carcinogenesis, 2011, 32, 597. 3 N. Ahmad and H. Mukhtar, J. Invest. Dermatol., 2004, 123, 417.
This journal is © The Royal Society of Chemistry 2014
View Article Online
Published on 08 October 2014. Downloaded by University of Texas Libraries on 06/01/2015 21:14:12.
Nanoscale
4 K. Virkler and I. K. Lednev, Anal. Bioanal. Chem., 2010, 396, 525. 5 J. Herrmann, Catal. Today, 1999, 53, 115. 6 T. Oto, R. G. Banal, K. Kataoka, M. Funato and Y. Kawakami, Nat. Photonics, 2010, 4, 767. 7 G. L. Kellogg and T. T. Tsong, J. Appl. Phys., 1980, 51, 1184. 8 F. Xie, H. Lu, D. J. Chen, P. Han, R. Zhang, Y. D. Zheng, L. Li, W. H. Jiang and C. Chen, Electron. Lett., 2011, 47, 930. 9 W. Yang, S. S. Hullavarad, B. Nagaraj, I. Takeuchi, R. P. Sharma, T. Venkatesan, R. D. Vispute and H. Shen, Appl. Phys. Lett., 2003, 82, 3424. 10 M. Liao and Y. Koide, Appl. Phys. Lett., 2006, 89, 113509. 11 T. Oshima, T. Okuno, N. Arai, N. Suzuki, S. Ohira and S. Fujita, Appl. Phys. Express, 2008, 1, 011202. 12 R. P. H. Kooyman, A. T. M. Lenferink, R. G. Eenink and J. Greve, Anal. Chem., 1991, 63, 83. 13 J. Siuzdak, G. Stepniak, M. Kowalczyk and L. Maksymiuk, IEEE Photonics Technol. Lett., 2009, 21, 993. 14 E. Muñoz, E. Monroy, J. L. Pau, F. Calle, F. Omnès and P. Gibart, J. Phys.: Condens. Matter, 2001, 13, 7115. 15 S. Ghosh, C. Rivera, J. L. Pau, E. Muñoz, O. Brandt and H. T. Grahn, Appl. Phys. Lett., 2007, 90, 091110. 16 S. K. Zhang, W. B. Wang, F. Yun, L. He, H. Morkoç, X. Zhou, M. Tamargo and R. R. Alfano, Appl. Phys. Lett., 2002, 81, 4628. 17 L. Britnell, R. V. Gorbachev, R. Jalil, B. D. Belle, F. Schedin, M. I. Katsnelson, L. Eaves, S. V. Morozov, A. S. Mayorov, N. M. R. Peres, A. H. C. Neto, J. Leist, A. K. Geim, L. A. Ponomarenko and K. S. Novoselov, Nano Lett., 2012, 12, 1707. 18 F. Bonaccorso, Z. Sun, T. Hasan and A. C. Ferrari, Nat. Photonics, 2010, 4, 611. 19 S. Song, Q. Chen, L. Jin and F. Sun, Nanoscale, 2013, 5, 9615. 20 Y. Liu, R. Cheng, L. Liao, H. Zhou, J. Bai, G. Liu, L. Liu, Y. Huang and X. Duan, Nat. Commun., 2011, 2, 579. 21 V. Bouchiat, Nat. Phys., 2012, 8, 862. 22 W. J. Yu, Z. Li, H. Zhou, Y. Chen, Y. Wang, Y. Huang and X. Duan, Nat. Mater., 2013, 12, 246. 23 X. Zhong, R. G. Amorim, R. H. Scheicher, R. Pandey and S. P. Karna, Nanoscale, 2012, 4, 5490. 24 X. Gonze, J.-M. Beuken, R. Caracas, F. Detraux, M. Fuchs, G.-M. Rignanese, L. Sindic, M. Verstraete, G. Zerah, F. Jollet, M. Torrent, A. Roy, M. Mikami, P. Ghosez, J.-Y. Raty and D. C. Allan, Comput. Mater. Sci., 2002, 25, 478.
This journal is © The Royal Society of Chemistry 2014
Paper
25 M. D. Segall, P. J. D. Lindan, M. J. Probert, C. J. Pickard, P. J. Hasnip, S. J. Clark and M. C. Payne, J. Phys.: Condens. Matter, 2002, 14, 2717. 26 G. Kresse and D. Joubert, Phys. Rev. B, 1999, 59, 1758. 27 J. Enkovaara, C. Rostgaard, J. J. Mortensen, J. Chen, M. Dułak, L. Ferrighi, J. Gavnholt, C. Glinsvad, V. Haikola, H. A. Hansen, H. H. Kristoffersen, M. Kuisma, A. H. Larsen, L. Lehtovaara, M. Ljungberg, O. LopezAcevedo, P. G. Moses, J. Ojanen, T. Olsen, V. Petzold, N. A. Romero, J. Stausholm-Møller, M. Strange, G. A. Tritsaris, M. Vanin, M. Walter, B. Hammer, H. Häkkinen, G. K. H. Madsen, R. M. Nieminen, J. K. Nørskov, M. Puska, T. T. Rantala, J. Schiøtz, K. S. Thygesen and K. W. Jacobsen, J. Phys.: Condens. Matter, 2010, 22, 253202. 28 W. Lin, W. Jiang, N. Gao, D. Cai, S. Li and J. Kang, Laser Photonics Rev., 2013, 7, 572. 29 G. S. Ofelt, J. Chem. Phys., 1962, 37, 511. 30 K. B. Nam, J. Li, K. H. Kim, J. Y. Lin and H. X. Jiang, Appl. Phys. Lett., 2001, 78, 3690. 31 K. S. Cole and R. H. Cole, J. Chem. Phys., 1941, 9, 341. 32 J. Zhang, H. Zhao and N. Tansu, Appl. Phys. Lett., 2010, 97, 111105. 33 L. Britnell, R. V. Gorbachev, R. Jalil, B. D. Belle, F. Schedin, M. I. Katsnelson, L. Eaves, S. V. Morozov, A. S. Mayorov, N. M. R. Peres, A. H. C. Neto, J. Leist, A. K. Geim, L. A. Ponomarenko and K. S. Novoselov, Nano Lett., 2012, 12, 1707. 34 J. Kang, D. Sarkar, W. Liu, D. Jena and K. Banerjee, Int. Electron Devices Meet., IEEE, 2012, 17.4.1. 35 S. Pereira, M. R. Correia, E. Pereira, K. P. O’Donnell, E. Alves, A. D. Sequeira, N. Franco, I. M. Watson and C. J. Deatcher, Appl. Phys. Lett., 2002, 80, 3913. 36 R. E. Nahory, M. A. Pollack, W. D. Johnston and R. L. Barns, Appl. Phys. Lett., 1978, 33, 659. 37 M. E. Hawkridge, Z. Liliental-Weber, H. J. Kim, S. Choi, D. Yoo, J.-H. Ryou and R. D. Dupuis, Appl. Phys. Lett., 2009, 94, 071905. 38 W. Z. Shen, S. C. Shen, W. G. Tang, S. M. Wang and T. G. Andersson, J. Appl. Phys., 1995, 78, 1178. 39 W. Choi, M. Y. Cho, A. Konar, J. H. Lee, G.-B. Cha, S. C. Hong, S. Kim, J. Kim, D. Jena, J. Joo and S. Kim, Adv. Mater., 2012, 24, 5832. 40 W. Zhang, J.-K. Huang, C.-H. Chen, Y.-H. Chang, Y.-J. Cheng and L.-J. Li, Adv. Mater., 2013, 25, 3456.
Nanoscale, 2014, 6, 14733–14739 | 14739
Published on 08 October 2014. Downloaded by University of Texas Libraries on 06/01/2015 21:14:12.
PAPER
Cite this: Nanoscale, 2014, 6, 14733
View Journal | View Issue
Quantum state engineering with ultra-shortperiod (AlN)m/(GaN)n superlattices for narrowband deep-ultraviolet detection Na Gao,a Wei Lin,a Xue Chen,a Kai Huang,*a,b Shuping Li,a Jinchai Li,a Hangyang Chen,a Xu Yang,a Li Ji,b Edward T. Yub and Junyong Kang*a Ultra-short-period (AlN)m/(GaN)n superlattices with tunable well and barrier atomic layer numbers were grown by metal–organic vapour phase epitaxy, and employed to demonstrate narrowband deep ultraviolet photodetection. High-resolution transmission electron microscopy and X-ray reciprocal space mapping confirm that superlattices containing well-defined, coherently strained GaN and AlN layers as thin as two atomic layers (∼0.5 nm) were grown. Theoretical and experimental results demonstrate that an optical absorption band as narrow as 9 nm (210 meV) at deep-ultraviolet wavelengths can be produced, and is attributable to interband transitions between quantum states along the [0001] direction in
Received 28th July 2014, Accepted 5th October 2014
ultrathin GaN atomic layers isolated by AlN barriers. The absorption wavelength can be precisely engineered by adjusting the thickness of the GaN atomic layers because of the quantum confinement effect.
DOI: 10.1039/c4nr04286g
These results represent a major advance towards the realization of wavelength selectable and narrowband
www.rsc.org/nanoscale
photodetectors in the deep-ultraviolet region without any additional optical filters.
Introduction The deep-ultraviolet (DUV) wavelength region between 200 and 280 nm has become indispensable for protein analysis, DNA sequencing, drug discovery and detection, forensic analysis, disinfection, surface and water decontamination, as well as for many other applications.1–6 Spurred by the demand for DUV detection and measurement technology, a variety of solid-state and vacuum devices can be made sensitive to different wavelength bands using bandpass interference filters and quartz lens systems.7–11 However, these approaches can suffer from the unwanted response,12 various instabilities,13 and device complexity.14 To eliminate the filters that are widely used in photodetectors, passive filter layers have recently been integrated into device structures during growth. Although the bandwidth has narrowed from 55 nm to 30 nm to limit the short-wavelength responsivity,15 the propensity for degradation, bulkiness, and unreliability of the devices remain unsolved in these systems, which combine a series of typical filters and phosphor coatings inserted into the optical path of
a
Department of Physics, Fujian Provincial Key Laboratory of Semiconductor Materials and Applications, Xiamen University, Xiamen, 361005, China. E-mail: [email protected], [email protected]; Fax: +86-592-2187737; Tel: +86-592-2185962 b Department of Electrical and Computer Engineering, Microelectronic Research Centre, The University of Texas at Austin, Austin, Texas, 78758, USA
This journal is © The Royal Society of Chemistry 2014
the sensor system.14,16 To overcome these limitations, semiconductor nanostructures are promising because of the possibility of achieving high wavelength-selective interband absorption between quantum states. Compared to the zerodimensional (0D) structures, i.e. nanodots or one-dimensional (1D) structures, i.e. nanowires, ultrathin two-dimensional (2D) layers (one or a few atomic layers) have an additional advantage that current can easily conduct in the 2D layers because of the relatively high in-plane conductivity. Thus photodetectors using ultrathin 2D layers as absorption materials would have higher external quantum efficiency (QE). Recent experimental results demonstrated that the bandgap energy of 2D MoS2 crystals decreases with the number of the layers indicating the existence of the quantum states.17 However, the quantum state in ultrathin 2D layers can only be applied to narrowband photodetectors when the incident photons can be made to propagate perpendicularly to the ultrathin 2D layers. The optical absorption is extremely low for the incident light perpendicular to the single 2D atomic layer (∼2.3% for monolayered graphene) due to the extremely short absorption optical length,18,19 restricting the optical response of the ultrathin 2D layer photodetectors. Although the optical response can also be enhanced by grazing or scattering the light onto the ultrathin 2D layers,20 in this geometry optical absorption occurs between continuous energy bands rather than between quantum states. To increase the optical length of the incident light absorption, a number of layers can be stacked along the
Nanoscale, 2014, 6, 14733–14739 | 14733
View Article Online
Published on 08 October 2014. Downloaded by University of Texas Libraries on 06/01/2015 21:14:12.
Paper
normal plane of the ultrathin 2D structures using transparent layers to isolate the quantum state. It has been proven that three hBN atomic layers could maintain the quantized characteristics of graphene due to the sufficiently high potential barrier provided by hBN.21–23 The 2.4% lattice mismatch between GaN and AlN makes growth of coherently strained individual layers and thick multilayered structures challenging. To ensure efficient absorption of incident ultraviolet photons, a large number of superlattice periods are required. In addition to optical absorption, good electrical conductivity in the vertical direction, i.e. perpendicular to the superlattice interfaces, is a crucial issue for the electrical responsivity of photodetectors. Thus, the AlN layers must be sufficiently thin to allow for efficient carrier transport via tunnelling, while at the same time be thick enough to provide effective quantum confinement. Minimization of material defects is also essential for high performance, as defects can lead to carrier trapping and high dark currents. Avoidance of strain relaxation will therefore also limit the maximum individual layer thickness and the total superlattice thickness that can be employed. In this paper, we propose and demonstrate a unique narrowband DUV photodetector structure in which a high degree of wavelength selectivity is achieved without additional filters using interband absorption between the quantum states in ultra-short-period (AlN)m/(GaN)n superlattices (SLs), where m and n represent the number of atomic layers in the AlN and GaN regions, respectively. Through computational modeling and atomic-scale epitaxial growth, a series of metal–semiconductor–metal (MSM) DUV photodetectors with tunable detection wavelengths and bandwidths as narrow as 9 nm (210 meV) were successfully designed, fabricated, and characterized. Our results indicate that the use of ultrathin semiconductor 2D layers can provide new classes of narrowband photodetectors without additional optical filters.
Computational and experimental methods First-principles calculations of the electronic structure of ultrashort-period (AlN)m/(GaN)n SLs were carried out using the Vienna ab-initio simulation package (VASP) in the framework of density-functional theory (DFT).24,25 The exchange-correlation function was described within the generalized-gradient approximation (GGA).26 The pseudopotentials were generated by means of the projector augmented wave (PAW) method27 that allows for the accurate treatment of the valence s and p electrons as well as the semicore Ga 3d states. With an 8 × 8 × 4 Monkhorst-Pack grid of k points sampling the Brillouin zone, a 500 eV cutoff energy was used to expand the electronic wave functions, which was sufficient for the plane wave basis to obtain well-converged results. The size of the supercell was fixed and the atoms within the cells were allowed to relax to minimize the total energy of the system with a convergence criterion of 0.1 meV. By introducing a scissors correction with the shifting parameter of 1.75 eV, the underestimated bandgap
14734 | Nanoscale, 2014, 6, 14733–14739
Nanoscale
values were then adjusted. The imaginary part of the dielectric function for optical absorption was computed using Fermi’s Golden Rule for the optical transition rate.28 To determine the maximum confinement energy achievable in an ultrathin GaN layer and verify the band-like nature of the electronic structure in such a layer, two atomic layers of GaN and a bulk GaN crystal were treated as slab and crystal structures, respectively. It should be noted that, in the case of a slab, introducing vacuum is necessary. Next, a series of ultrashort-period (AlN)m/(GaN)n SLs were constructed along the [0001] direction with different AlN layer numbers m = 4, 6 and 8 in the (1a × 1a × 3c), (1a × 1a × 4c) and (1a × 1a × 5c) supercell to examine whether the barrier was sufficiently high to maintain the quantized characteristics of the wide bandgap GaN ultrathin 2D layers, respectively. Finally, by varying the atomic layer number of the GaN ultrathin 2D layers (n = 1, 2, 4) while the atomic layer number of AlN was kept approximately constant at m = 6, tunable optical absorption at different peak energies by quantum state engineering was realized in the DUV spectra range. Experimentally, a series of ultra-short-period (AlN)m/(GaN)n SLs were epitaxially grown on c-plane sapphire substrates via metal–organic vapour phase epitaxy (MOVPE) in a vertical Thomas Swan system (3 × 2 inch CCS Aixtron). The source precursors were trimethylaluminum (TMA), trimethylgallium (TMG), trimethylindium (TMI), and ammonia (NH3), and hydrogen (H2) was used as the carrier gas. An AlN buffer layer with a thickness of approximately 1.0 µm was first grown at 1100 °C, followed by 300 periods of (AlN)m/(GaN)n SLs with varying well and barrier atomic layer numbers. The atomic layer numbers were altered by changing the flow rates of source precursors. During the ultra-short-period SL growth, a brief growth interruption was adopted to separate the ultrathin GaN and AlN layers. The nominal thickness of one monolayer is 2.6 Å for GaN and 2.5 Å for AlN, respectively. The surface morphology and crystalline quality of the ultra-short-period (AlN)m/(GaN)n SLs were characterized by a combination of an atomic force microscope system (AFM, SPA400), a high-resolution X-ray diffractometer (HRXRD, PANalytical X’ Pert PRO) with an X-ray wavelength of 0.154056 nm using Cu Kα radiation, and a high-resolution transmission electron microscope system (FEI Tecnai G2 F20 S-Twin HRTEM) operating at 200 kV. Subsequently, by means of standard photolithography (Karlsuss MA6/BA6), electron-beam deposition (Temescal FC2000) and lift-off processes, MSM DUV photodetectors were fabricated on the ultra-short-period (AlN)m/(GaN)n SLs. For the interdigitated MSM structures reported in this work, the fingers were 300 µm long and 6 µm wide with a spacing of 6 µm. Titanium/gold (10/200 nm) was then evaporated to form Schottky contacts. The fabricated photodetectors were finally annealed by rapid thermal annealing (RTA 300) under a N2 atmosphere at 400 °C for 300 s. For the fabricated devices, I–V characteristics were measured using a Keithley 2410 sourcemeter and a Keithley 6514 programmable electrometer. A 450 W Xe arc lamp, a mechanical chopper and a lock-in amplifier
This journal is © The Royal Society of Chemistry 2014
View Article Online
Nanoscale
Paper
Published on 08 October 2014. Downloaded by University of Texas Libraries on 06/01/2015 21:14:12.
were used to measure the photocurrent response spectra. The system was calibrated with a UV-enhanced Si detector. The devices were all illuminated perpendicularly from the front metal/semiconductor contact side. All measurements were carried out at room temperature.
Results and discussion Quantum state engineering in GaN 2D layers and ultra-shortperiod (AlN)m/(GaN)n SLs To understand the underlying mechanism that determines the optical properties, first-principles calculations of electronic structures of ultrathin 2D GaN layers were performed and compared with that of bulk GaN. When reducing the GaN thickness to as thin as two atomic layers to form a 2D structure, along the high symmetry lines from Γ to A in the Brillouin zone corresponding to the [0001] direction, i.e. perpendicular to the ultrathin 2D layers, the electronic states near the band edge show completely discrete quantum states, as shown in Fig. 1a. In contrast, the electronic structure for the conduction (CB) and valence (VB) bands in bulk GaN, shown in Fig. 1b, exhibits the usual strong dispersion around the Γ-point. Owing to selection rules,29 electronic transitions in the 2D GaN layer can only occur between the quantum states, such as 1h–1e, thereby enabling narrowband optical absorption at a specific wavelength. Fig. 2a shows the electronic structure computed for an ultra-short-period (AlN)6/(GaN)2 SL. One can observe that the well-defined quantum states are still identified as 1h and 1e within the range of Γ to A, i.e. perpendicular to the ultrathin 2D layers. However, the energy dispersion of the SLs is slightly larger than that of GaN 2D atomic layers in a vacuum. More than one quantum states are expected to exist for both elec-
Fig. 1 Comparison of electronic structures for GaN 2D layer and bulk crystals. (a) GaN 2D structure in a vacuum. Completely discrete quantum states are shown along the symmetry lines (vertical dashed red lines) from Γ to A in the Brillouin zone, in contrast to a strong dispersion around the Γ-point in (b) GaN bulk materials.
This journal is © The Royal Society of Chemistry 2014
Fig. 2 Electronic structures and optical absorption properties of the designed ultra-short-period (AlN)m/(GaN)n SLs. (a) Typical electronic structures for the stacked (AlN)6/(GaN)2 as well as for the crystal structure of the supercell, with the metal atoms Ga and Al in green and blue, respectively. (b, c) Normalized imaginary part of the dielectric function εxx + εyy (εxx = εyy) when unpolarized light is perpendicularly incident to the 2D atomic layers (in the directions). (b) The stacked GaN 2D structure (two atomic layers) isolated by different AlN atomic layer numbers m: 4 (blue line), 6 (red line) and 8 (green line) atomic layers. (c) Stacked GaN 2D atomic layers by different atomic layer numbers n: 4 (blue line), 2 (red line) and 1 (green line); all are isolated by the same barrier of approximately 6 AlN atomic layers.
trons and holes in the GaN 2D layer because of the large bandgap difference between the AlN barrier and the GaN well.30 The dispersive quantum states in the CB are denoted as 1e, 2e, etc. While in the VB, the number of dispersionless quantum states is greater. The interband optical absorption of unpolarized light incident normal to the 2D ultrathin layers (i.e. in the directions) is proportional to the imaginary part of the dielectric function,31 εxx + εyy, plotted in Fig. 2b–2c. As shown in Fig. 2b, the peak energies of the 1h–1e transition are approximately the same among the (AlN)m/(GaN)n SLs with different AlN atomic layer numbers m. This fact clearly demonstrates that the AlN potential barriers are sufficiently high to maintain the optical absorption between the 1h–1e quantum states of the bilayer GaN 2D atomic layers, even when the AlN barrier thickness is reduced to 4 atomic layers. Additionally, on the high energy side of the 1h–1e peak, other weaker but broader peaks can be also recognized with 20–40% peak intensities of the main peak. Based on the calculated band structures, these peaks peaked at approximately 5.10 eV and 5.50 eV are mainly derived from the 2h–1e and 1h–2e transitions, respectively. One can see from Fig. 2a that the 1h is accompanied by discrete quantum states 2h and 3h as well as the crystal-field split-off-hole (ch) band, which consists of atomic p orbitals parallel to the directions.32 Therefore, the transition from ch to the 1e devotes no contribution to the optical absorption along the directions for light perpendicular to the ultrathin 2D layers. Thus the absorption associated with the transition between ch and 1e is not shown
Nanoscale, 2014, 6, 14733–14739 | 14735
View Article Online
Published on 08 October 2014. Downloaded by University of Texas Libraries on 06/01/2015 21:14:12.
Paper
in Fig. 2b. These results show that the transition probability of the 1h–1e interband transition is much higher than that of the other higher energy interband transitions. It should be noted that a slight energy dispersion along the [0001] direction of approximately 97 meV occurs in the bottom of the CB, as shown in Fig. 2a, which enables the electrons to tunnel through the isolating AlN barrier and provide a parallel connection through the stacked GaN 2D atomic layers. By altering the thickness of the AlN barrier layer, the electron transport in the direction perpendicular to the 2D atomic layers changes due to variations in the coupling of the density of states (DOS) between the adjacent GaN 2D atomic layers.33,34 Furthermore, the interband absorption 1h–1e exhibits a remarkably narrow bandwidth, with a full width at half maximum (FWHM) of approximately 230 meV for the AlN isolated atomic layer with m greater than 6. Although the FWHM of the 4 atomic layer AlN is broader, the energy dispersion along the [0001] direction in the bottom of the CB will be favourable for electron quantum tunnelling between the GaN 2D atomic layers. Because of the small changes in the absorption wavelength in response to altering the AlN layer thickness, we can more reliably tune the quantum states by changing the atomic layer number n of the GaN ultrathin 2D layers, as shown in Fig. 2c. When the atomic layer number is reduced from 4 to 1, the optical absorption energy from the 1h to 1e transition clearly increases in the DUV region. On the other hand, the associated shoulder peaks originated from the higher energy transitions, such as 2h–1e, increase when reducing the atomic layer number responded by an upshift of the 1e quantum states in the thinner GaN 2D layers. Thus, we implemented narrowband photodetection with tunable wavelengths in the DUV region by altering the atomic layer number n of the GaN 2D layers rather than that of the AlN barrier layer. Characteristic of ultra-short-period (AlN)m/(GaN)n SLs Based on these computations, a series of narrowband DUV photodetectors using ultra-short-period (AlN)m/(GaN)n SLs was grown, fabricated, and characterized. To confirm the achievement of a coherent AlN/GaN heterostructure, a series of (AlN)m/(GaN)n SLs grown on the atomic scale were characterized by AFM, HRXRD and HRTEM. An extremely smooth surface with atomic-step terraces is shown in Fig. 3a, with the surface root-mean-square (rms) roughness determined to be 1.7 Å in a 500 × 500 nm2 scan area, which indicates abrupt interfaces between the AlN/GaN heterostructures on the atomic scale. Fig. 3b displays the asymmetric (105) reciprocal space mapping (RSM) of the samples, where the two coordinates (Qx, Qy) in the map correspond to a pair of lattice constants (a, c).35 The main peaks are aligned at nearly the same Qx value (denoted by a dashed blue line), while the reciprocal vector perpendicular to the surface is resolved. This result confirms that the AlN/GaN heterostructure is almost coherently grown on the AlN/sapphire templates and is fully strained. The HRXRD patterns of (0002) ω/2θ scans further demonstrate a high degree of periodicity and excellent interface abruptness, with well-defined satellite diffraction peaks until the third
14736 | Nanoscale, 2014, 6, 14733–14739
Nanoscale
Fig. 3 Surface morphology and crystal quality of epitaxially grown ultra-short-period (AlN)m/(GaN)n SLs. (a) AFM image with 1.7 Å rms roughness and (b) the asymmetric (105) RSM of the periodic stacked (AlN)m/(GaN)n SLs grown on AlN/sapphire templates. The dashed blue line indicates that the AlN/GaN heterostructure is almost coherent. (c) The HRXRD patterns of the ω/2θ (0002) scans, with the layer thicknesses of the GaN determined to be 1 ( purple line), 2 (orange line) and 4 (blue line) atomic layers, respectively. (d) Representative cross-sectional TEM images. (e) Fourier transform (FFT) imaged along the [101¯0] zone axis, shown with the SAED patterns (f ). The inset of (f ) reveals period-perfect ultra-short-period (AlN)m/(GaN)n SLs without composition fluctuations. a, b, d, e, and f correspond to the sample with 2 atomic layers of GaN.
order, as shown in Fig. 3c. By combining Vegard’s law and angular separations between superlattice satellite peaks, the thicknesses of the GaN 2D layers were determined to be 1, 2 and 4 atomic layers, respectively.36 These values are in good agreement with the growth parameters and the design expectations, which demonstrates reliable, precise epitaxial growth on the atomic scale, even for a single monolayer of GaN, via MOVPE. To examine the crystalline features in detail, crosssectional TEM and its Fourier transform (FFT) as well as selected area electron diffraction (SAED) were carried out on the samples. As shown in Fig. 3d, distinct AlN/GaN fringes with bright/dark contrast, clear and atomically sharp interfaces and two atomic layers (∼0.5 nm) of GaN 2D layers were recogˉ0] zone nized. High-resolution FFT images taken along the [101 axis in Fig. 3e further confirmed that the stacked AlN/GaN heterointerface is almost coherently grown. Moreover, Fig. 3f revealed that the (AlN)m/(GaN)n SLs were period-perfect without composition fluctuations.37 Therefore, we concluded that the
This journal is © The Royal Society of Chemistry 2014
View Article Online
Nanoscale
GaN 2D layers could be stacked periodically and isolated coherently by AlN layers without defect states at the interfaces.
Published on 08 October 2014. Downloaded by University of Texas Libraries on 06/01/2015 21:14:12.
Narrowband DUV detection based on ultra-short-period (AlN)m/(GaN)n SLs Using the (AlN)m/(GaN)n SLs, MSM DUV photodetectors were then fabricated in the geometry shown in Fig. 4a. An optical micrograph of a representative fabricated device is shown in Fig. 4b. The typical dark current–voltage (I–V) characteristics and the I–V characteristics under illumination at room temperature exhibit symmetrical Schottky features, as shown in Fig. 4c. The dark current at 20 V is estimated to be as low as 4.8 × 10−13 A due to the UV-transparent AlN isolating layer in the device. It is worth noting that all the fabricated MSM DUV photodetectors exhibit narrowband spectral responses at relevant wavelengths. For the GaN 2D layers with atomic layer numbers of 1, 2, 4 and 6 isolated by the same barriers of approximately 6 AlN atomic layers, the corresponding wavelengths appear at 230, 240, 248 and 266 nm (Fig. 4d–g), indicating the dominantly optical transition from 1h to 1e respectively. Notably, the full width at half maximum (FWHM) was as narrow as 9 nm (210 meV) at wavelengths of 230 and 240 nm. These experimental results were similar to the calculated imaginary part of the dielectric function εxx + εyy shown in Fig. 2b–c. However, besides the optical absorption peak related to the 1h–1e interband transition, the optical absorp-
Fig. 4 (a) Three-dimensional schematic of the designed structure and (b) top-view optical micrograph of the device. (c) Typical dark current– voltage (I–V) characteristics (black line) as well as the current–voltage characteristics under light illumination (red line). (d–g) Normalized photocurrent spectra measured at room temperature as a function of wavelength under different biases with light illumination. Wavelength tunable narrowband photodetection appears at wavelengths of 230, 240, 248 and 266 nm in the DUV region, respectively.
This journal is © The Royal Society of Chemistry 2014
Paper
tion peaks related to the 2h–1e and 1h–2e transitions shown in Fig. 2b–c cannot be found in the experimental photocurrents. This behaviour has yet to be explored and might be attributed to two possible effects. First, the incident light power decreases sharply as the DUV wavelength shortens, which leads to low optical absorption. Second, it can be assumed that the binding energy of excitons increases greatly as a result of the strong confinement in such high-quality and extremely thin GaN 2D layers. It is known that the 2D exciton can greatly enhance the band edge absorption.38 Thus the experimental narrowband photodetection might be attributed to the optical absorption of the 1h–1e excitonic transition. Previous investigation on the InGaAs/GaAs MQWs indicated that interband excitonic transitions were difficult to be observed at room temperature except for the 1hh–1e excitonic transition. Therefore, this might be another reason that the shoulder peaks with higher energy were not found experimentally. We deduce the unique mechanism of generation and collection of the photo-induced carriers in the (AlN)m/(GaN)n SLs, which is described as follows. (i) As a result of quantum confinement, quantum states are generated in contrast to continuous energy bands. (ii) When light is incident on the device perpendicular to the GaN 2D atomic layers, electrons in the discrete quantum states of the VB are excited to the CB and then form photo-induced carriers. (iii) Under bias, the photo-induced carriers generate photocurrent in external circuits. Finally, we optimized the carrier transport properties of the (AlN)m/(GaN)n SL photodetectors by changing the value of m. Reducing the AlN barrier thickness in the (AlN)m/(GaN)n SLs was expected to enable more efficient vertical carrier transport. Since the influence of electron quantum tunnelling can be described by the energy dispersion along the [0001] direction in the bottom of the CB with relevant electronic structures, the dispersion energy in the (AlN)6/(GaN)2 was determined to be approximately 97 meV, in contrast to 236 meV in (AlN)4/ (GaN)2. Thus, we decreased the isolated AlN barrier to 4 atomic layers to enhance the coupling of DOS between the adjacent GaN 2D layers. One can see from Fig. 5a that the photocurrent spectrum of the photodetector fabricated by (AlN)4/(GaN)2 peaks at approximately the same wavelength as that fabricated by (AlN)4/(GaN)2. However, the photocurrent of the (AlN)4/(GaN)2 photodetector was more than 20 times higher than that of the (AlN)6/(GaN)2 photodetector when the bias voltage was 5 V. Additionally, the FWHM of the photocurrent peak broadened from 9 nm to 12 nm when the atomic layer number of AlN was reduced from 6 to 4. These results are in good agreement with the calculated optical absorption characteristics shown in Fig. 2b. Fig. 5b–c show the critical parameters of the typical MSM DUV photodetector fabricated by (AlN)4/(GaN)2, spectral responsivity and external QE. The responsivity is given by the following equation: R ¼ I ph =P ¼ ηGq=hν
ð1Þ
where Iph and P represent the photocurrent and the illumination light power,39 respectively, G is the photogain, q is the
Nanoscale, 2014, 6, 14733–14739 | 14737
View Article Online
Paper
Nanoscale
Published on 08 October 2014. Downloaded by University of Texas Libraries on 06/01/2015 21:14:12.
quantum states along the [0001] direction in (AlN)m/(GaN)n SLs. The unique narrow bandwidth (down to 9 nm) and wavelength tunable properties enable real-time, high resolution wavelength-selective imaging without additional filters in the DUV region to selectively separate the signals of biological/ chemical components. This work further makes it possible to develop multiple wavelength narrowband detection integrated on a single chip by adjusting the interband quantum states (atomic layer number) in the direction perpendicular to the 2D layers. We believe that by introducing other stacked 2D atomic layers, such as GaN/InN SLs, tunable narrow-bandwidth photodetectors in the visible region may be another promising candidate for digital cameras beyond colour charge-coupled devices (CCDs) with filters.
Author contributions N.G. and W.L. contributed equally to this work. K.H. and J.Y.K. conceived and supervised this research. N.G., S.P.L., H.Y.C. and J.Y.K. carried out the MOVPE growth. N.G. and J.C.L contributed to the sample characterization. N.G., X.C. and X.Y. fabricated the device, and N.G. X.C. and K.H. performed the measurements. W.L. and N.G. contributed to the computational simulations. All authors analysed the data and wrote the manuscript.
Acknowledgements
Fig. 5 (a) Photocurrent spectra of the typical MSM DUV photodetectors fabricated by (AlN)4/(GaN)2 (red line) and (AlN)6/(GaN)2 (black line) when the bias voltage is 5 V. (b) The spectral responsivity and (c) the external QE of the photodetector based on (AlN)4/(GaN)2 SLs at a wavelength of 240 nm with a narrow bandwidth under different biases with light illumination. The optical power density of the illumination is ∼0.013 µW mm−2.
electronic charge, h is Planck’s constant and ν is the frequency of the incident wavelength.40 Obviously, the spectral responsivity increases with increasing bias, reaching 51 mA W−1 under a bias of 40 V at a wavelength of 240 nm. The external QE η is further estimated to be up to ∼26%.
Conclusions In summary, we demonstrated MSM photodetectors in the DUV spectrum range by interband absorption between
14738 | Nanoscale, 2014, 6, 14733–14739
The authors appreciate Yanyan Fang and Changqing Chen at Huazhong University of Science Technology as well as Jiafa Cai at Xiamen University for the HRXRD support, Mutong Niu and Ke Xu at Suzhou Institute of Nano-Tech and Nano-Bionics, Chinese Academy of Sciences for TEM assistance. The authors also acknowledge Shaoxiong Wu at Xiamen University for technical support for the device measurements. This work was supported by the National Research Program of China under grant no. 2012CB619300, 2011CB301905 and 2011CB925600, the National Natural Science Foundation of China under grant no. 61227009, 61108064, 91321102, 11204254 and 11404271, the Fundamental Research Funds for the Central Universities (2011120143) and in part by the scholarship from the China Scholarship Council (CSC) under the Grant CSC no. 201306315030. This work was also supported by the Natural Science Foundations of Fujian Province (2012J01024).
Notes and references 1 J. A. Nichols and S. K. Katiyar, Arch. Dermatol. Res., 2010, 302, 71. 2 V. Nandakumar, M. Vaid, T. O. Tollefsbol and S. K. Katiyar, Carcinogenesis, 2011, 32, 597. 3 N. Ahmad and H. Mukhtar, J. Invest. Dermatol., 2004, 123, 417.
This journal is © The Royal Society of Chemistry 2014
View Article Online
Published on 08 October 2014. Downloaded by University of Texas Libraries on 06/01/2015 21:14:12.
Nanoscale
4 K. Virkler and I. K. Lednev, Anal. Bioanal. Chem., 2010, 396, 525. 5 J. Herrmann, Catal. Today, 1999, 53, 115. 6 T. Oto, R. G. Banal, K. Kataoka, M. Funato and Y. Kawakami, Nat. Photonics, 2010, 4, 767. 7 G. L. Kellogg and T. T. Tsong, J. Appl. Phys., 1980, 51, 1184. 8 F. Xie, H. Lu, D. J. Chen, P. Han, R. Zhang, Y. D. Zheng, L. Li, W. H. Jiang and C. Chen, Electron. Lett., 2011, 47, 930. 9 W. Yang, S. S. Hullavarad, B. Nagaraj, I. Takeuchi, R. P. Sharma, T. Venkatesan, R. D. Vispute and H. Shen, Appl. Phys. Lett., 2003, 82, 3424. 10 M. Liao and Y. Koide, Appl. Phys. Lett., 2006, 89, 113509. 11 T. Oshima, T. Okuno, N. Arai, N. Suzuki, S. Ohira and S. Fujita, Appl. Phys. Express, 2008, 1, 011202. 12 R. P. H. Kooyman, A. T. M. Lenferink, R. G. Eenink and J. Greve, Anal. Chem., 1991, 63, 83. 13 J. Siuzdak, G. Stepniak, M. Kowalczyk and L. Maksymiuk, IEEE Photonics Technol. Lett., 2009, 21, 993. 14 E. Muñoz, E. Monroy, J. L. Pau, F. Calle, F. Omnès and P. Gibart, J. Phys.: Condens. Matter, 2001, 13, 7115. 15 S. Ghosh, C. Rivera, J. L. Pau, E. Muñoz, O. Brandt and H. T. Grahn, Appl. Phys. Lett., 2007, 90, 091110. 16 S. K. Zhang, W. B. Wang, F. Yun, L. He, H. Morkoç, X. Zhou, M. Tamargo and R. R. Alfano, Appl. Phys. Lett., 2002, 81, 4628. 17 L. Britnell, R. V. Gorbachev, R. Jalil, B. D. Belle, F. Schedin, M. I. Katsnelson, L. Eaves, S. V. Morozov, A. S. Mayorov, N. M. R. Peres, A. H. C. Neto, J. Leist, A. K. Geim, L. A. Ponomarenko and K. S. Novoselov, Nano Lett., 2012, 12, 1707. 18 F. Bonaccorso, Z. Sun, T. Hasan and A. C. Ferrari, Nat. Photonics, 2010, 4, 611. 19 S. Song, Q. Chen, L. Jin and F. Sun, Nanoscale, 2013, 5, 9615. 20 Y. Liu, R. Cheng, L. Liao, H. Zhou, J. Bai, G. Liu, L. Liu, Y. Huang and X. Duan, Nat. Commun., 2011, 2, 579. 21 V. Bouchiat, Nat. Phys., 2012, 8, 862. 22 W. J. Yu, Z. Li, H. Zhou, Y. Chen, Y. Wang, Y. Huang and X. Duan, Nat. Mater., 2013, 12, 246. 23 X. Zhong, R. G. Amorim, R. H. Scheicher, R. Pandey and S. P. Karna, Nanoscale, 2012, 4, 5490. 24 X. Gonze, J.-M. Beuken, R. Caracas, F. Detraux, M. Fuchs, G.-M. Rignanese, L. Sindic, M. Verstraete, G. Zerah, F. Jollet, M. Torrent, A. Roy, M. Mikami, P. Ghosez, J.-Y. Raty and D. C. Allan, Comput. Mater. Sci., 2002, 25, 478.
This journal is © The Royal Society of Chemistry 2014
Paper
25 M. D. Segall, P. J. D. Lindan, M. J. Probert, C. J. Pickard, P. J. Hasnip, S. J. Clark and M. C. Payne, J. Phys.: Condens. Matter, 2002, 14, 2717. 26 G. Kresse and D. Joubert, Phys. Rev. B, 1999, 59, 1758. 27 J. Enkovaara, C. Rostgaard, J. J. Mortensen, J. Chen, M. Dułak, L. Ferrighi, J. Gavnholt, C. Glinsvad, V. Haikola, H. A. Hansen, H. H. Kristoffersen, M. Kuisma, A. H. Larsen, L. Lehtovaara, M. Ljungberg, O. LopezAcevedo, P. G. Moses, J. Ojanen, T. Olsen, V. Petzold, N. A. Romero, J. Stausholm-Møller, M. Strange, G. A. Tritsaris, M. Vanin, M. Walter, B. Hammer, H. Häkkinen, G. K. H. Madsen, R. M. Nieminen, J. K. Nørskov, M. Puska, T. T. Rantala, J. Schiøtz, K. S. Thygesen and K. W. Jacobsen, J. Phys.: Condens. Matter, 2010, 22, 253202. 28 W. Lin, W. Jiang, N. Gao, D. Cai, S. Li and J. Kang, Laser Photonics Rev., 2013, 7, 572. 29 G. S. Ofelt, J. Chem. Phys., 1962, 37, 511. 30 K. B. Nam, J. Li, K. H. Kim, J. Y. Lin and H. X. Jiang, Appl. Phys. Lett., 2001, 78, 3690. 31 K. S. Cole and R. H. Cole, J. Chem. Phys., 1941, 9, 341. 32 J. Zhang, H. Zhao and N. Tansu, Appl. Phys. Lett., 2010, 97, 111105. 33 L. Britnell, R. V. Gorbachev, R. Jalil, B. D. Belle, F. Schedin, M. I. Katsnelson, L. Eaves, S. V. Morozov, A. S. Mayorov, N. M. R. Peres, A. H. C. Neto, J. Leist, A. K. Geim, L. A. Ponomarenko and K. S. Novoselov, Nano Lett., 2012, 12, 1707. 34 J. Kang, D. Sarkar, W. Liu, D. Jena and K. Banerjee, Int. Electron Devices Meet., IEEE, 2012, 17.4.1. 35 S. Pereira, M. R. Correia, E. Pereira, K. P. O’Donnell, E. Alves, A. D. Sequeira, N. Franco, I. M. Watson and C. J. Deatcher, Appl. Phys. Lett., 2002, 80, 3913. 36 R. E. Nahory, M. A. Pollack, W. D. Johnston and R. L. Barns, Appl. Phys. Lett., 1978, 33, 659. 37 M. E. Hawkridge, Z. Liliental-Weber, H. J. Kim, S. Choi, D. Yoo, J.-H. Ryou and R. D. Dupuis, Appl. Phys. Lett., 2009, 94, 071905. 38 W. Z. Shen, S. C. Shen, W. G. Tang, S. M. Wang and T. G. Andersson, J. Appl. Phys., 1995, 78, 1178. 39 W. Choi, M. Y. Cho, A. Konar, J. H. Lee, G.-B. Cha, S. C. Hong, S. Kim, J. Kim, D. Jena, J. Joo and S. Kim, Adv. Mater., 2012, 24, 5832. 40 W. Zhang, J.-K. Huang, C.-H. Chen, Y.-H. Chang, Y.-J. Cheng and L.-J. Li, Adv. Mater., 2013, 25, 3456.
Nanoscale, 2014, 6, 14733–14739 | 14739
