Type-II InAs/GaSb strained layer superlattices grown on GaSb (111)B substrate

Plis, E.,Klein, B.,Myers, S.,Gautam, N.,Rotter, T.J.,Dawson, R.L.,Krishna, S.,Lee, S.J.,Kim, Y.H. American Vacuum Society 2013 Journal of vacuum science and technology. material Vol.31 No.3

Task-specific feature extraction and classification of fMRI volumes using a deep neural network initialized with a deep belief network: Evaluation using sensorimotor tasks

Jang, Hojin,Plis, Sergey M.,Calhoun, Vince D.,Lee, Jong-Hwan Elsevier 2017 NeuroImage Vol.145 No.2

<P><B>Abstract</B></P> <P>Feedforward deep neural networks (DNNs), artificial neural networks with multiple hidden layers, have recently demonstrated a record-breaking performance in multiple areas of applications in computer vision and speech processing. Following the success, DNNs have been applied to neuroimaging modalities including functional/structural magnetic resonance imaging (MRI) and positron-emission tomography data. However, no study has explicitly applied DNNs to 3D whole-brain fMRI volumes and thereby extracted hidden volumetric representations of fMRI that are discriminative for a task performed as the fMRI volume was acquired. Our study applied fully connected feedforward DNN to fMRI volumes collected in four sensorimotor tasks (i.e., left-hand clenching, right-hand clenching, auditory attention, and visual stimulus) undertaken by 12 healthy participants. Using a leave-one-subject-out cross-validation scheme, a restricted Boltzmann machine-based deep belief network was pretrained and used to initialize weights of the DNN. The pretrained DNN was fine-tuned while systematically controlling weight-sparsity levels across hidden layers. Optimal weight-sparsity levels were determined from a minimum validation error rate of fMRI volume classification. Minimum error rates (mean±standard deviation; %) of 6.9 (±3.8) were obtained from the three-layer DNN with the sparsest condition of weights across the three hidden layers. These error rates were even lower than the error rates from the single-layer network (9.4±4.6) and the two-layer network (7.4±4.1). The estimated DNN weights showed spatial patterns that are remarkably task-specific, particularly in the higher layers. The output values of the third hidden layer represented distinct patterns/codes of the 3D whole-brain fMRI volume and encoded the information of the tasks as evaluated from representational similarity analysis. Our reported findings show the ability of the DNN to classify a single fMRI volume based on the extraction of hidden representations of fMRI volumes associated with tasks across multiple hidden layers. Our study may be beneficial to the automatic classification/diagnosis of neuropsychiatric and neurological diseases and prediction of disease severity and recovery in (pre-) clinical settings using fMRI volumes without requiring an estimation of activation patterns or ad hoc statistical evaluation.</P> <P><B>Highlights</B></P> <P> <UL> <LI> Deep neural network (DNN) was proposed to classify fMRI volume of sensorimotor tasks. </LI> <LI> DNN weights were optimized via non-zero value percentage in cross-validation (CV) framework. </LI> <LI> Classification performance was superior from three-layer DNN than one-/two-/four-layer DNNs. </LI> <LI> Weight/hidden representations were highly task-specific from higher than lower hidden layers. </LI> <LI> Sparsity of weights between the input and first hidden layer enhanced the performance. </LI> </UL> </P>

Type II InAs/GaSb Superlattice Infrared Detectors with P-on-N Polarity Structures

이상준,김준오,노삼규,최정우,E. Plis,S. Krishna 한국진공학회 2007 한국진공학회 학술발표회초록집 Vol.16 No.2

Quantum Size Effects on the Chemical Sensing Performance of Two-Dimensional Semiconductors

Nah, Junghyo,Kumar, S. Bala,Fang, Hui,Chen, Yu-Ze,Plis, Elena,Chueh, Yu-Lun,Krishna, Sanjay,Guo, Jing,Javey, Ali American Chemical Society 2012 The Journal of Physical Chemistry Part C Vol.116 No.17

<P>We investigate the role of quantum confinement on the performance of gas sensors based on two-dimensional InAs membranes. Pd-decorated InAs membranes configured as H<SUB>2</SUB> sensors are shown to exhibit strong thickness dependence, with ∼100× enhancement in the sensor response as the thickness is reduced from 48 to 8 nm. Through detailed experiments and modeling, the thickness scaling trend is attributed to the quantization of electrons which favorably alters both the position and the transport properties of charge carriers; thus making them more susceptible to surface phenomena.</P><P><B>Graphic Abstract</B> <IMG SRC='http://pubs.acs.org/appl/literatum/publisher/achs/journals/content/jpccck/2012/jpccck.2012.116.issue-17/jp300446z/production/images/medium/jp-2012-00446z_0002.gif'></P><P><A href='http://pubs.acs.org/doi/suppl/10.1021/jp300446z'>ACS Electronic Supporting Info</A></P>

III–V Complementary Metal–Oxide–Semiconductor Electronics on Silicon Substrates

Nah, Junghyo,Fang, Hui,Wang, Chuan,Takei, Kuniharu,Lee, Min Hyung,Plis, E.,Krishna, Sanjay,Javey, Ali American Chemical Society 2012 Nano letters Vol.12 No.7

<P>One of the major challenges in further advancement of III–V electronics is to integrate high mobility complementary transistors on the same substrate. The difficulty is due to the large lattice mismatch of the optimal <I>p</I>- and <I>n</I>-type III–V semiconductors. In this work, we employ a two-step epitaxial layer transfer process for the heterogeneous assembly of ultrathin membranes of III–V compound semiconductors on Si/SiO<SUB>2</SUB> substrates. In this III–V-on-insulator (XOI) concept, ultrathin-body InAs (thickness, 13 nm) and InGaSb (thickness, 7 nm) layers are used for enhancement-mode <I>n</I>- and <I>p</I>- MOSFETs, respectively. The peak effective mobilities of the complementary devices are ∼1190 and ∼370 cm<SUP>2</SUP>/(V s) for electrons and holes, respectively, both of which are higher than the state-of-the-art Si MOSFETs. We demonstrate the first proof-of-concept III–V CMOS logic operation by fabricating NOT and NAND gates, highlighting the utility of the XOI platform.</P><P><B>Graphic Abstract</B> <IMG SRC='http://pubs.acs.org/appl/literatum/publisher/achs/journals/content/nalefd/2012/nalefd.2012.12.issue-7/nl301254z/production/images/medium/nl-2012-01254z_0002.gif'></P>

InAs/GaSb 제2형 응력 초격자 nBn 장적외선 검출소자 설계, 제작 및 특성평가

Kim, Ha Sul,Lee, Hun,Klein, Brianna,Gautam, Nutan,Plis, Elena A.,Myers, Stephen,Krishna, Sanjay 한국진공학회 2013 Applied Science and Convergence Technology Vol.22 No.6

InAs/GaSb 제2형 응력 초격자(strained layer type II superlattice, T2SL)을 이용한 nBn 구조 장적외선 검출소자의 설계 및 제작을 하였다. InAs와 GaSb 두께에 따른 T2SL 구조의 장적외선 밴드갭 에너지를 Kronig-Penney 모델을 이용하여 계산하였다. 소자의 암전류 밀도를 줄이기 위해서, nBn 구조에서 장벽층인 $Al_{0.2}Ga_{0.8}Sb$ 성장 중에 Te 보상도핑(compansated doping)을 하였다. 온도(T) 80 K 및 인가전압($V_b$) -1.5 V에서, 반응스펙트럼 측정을 통한 소자의 차단파장은 ${\sim}10.2{\mu}m$ (~0.122 eV)로 나타났다. 또한 온도 변화에 따른 암전류 측정으로부터 도출된 활성화 에너지는 0.128 eV로 계산 되었다. T=80 K 및 $V_b$=-1.5 V에서 암전류는 $1.0{\times}10^{-2}A/cm^2$으로 측정되었다. 흑체복사 적외선 광원을 이용한 반응도(Responsivity)는 소자 온도 80 K 및 인가전압 -1.5 V의 조건에서 0.58 A/W로 측정되었다. Long-wave infrared detectors using the type-II InAs/GaSb strained superlattice (T2SL) material system with the nBn structure were designed and fabricated. The band gap energy of the T2SL material was calculated as a function of the thickness of the InAs and GaSb layers by the Kronig-Penney model. Growth of the barrier material ($Al_{0.2}Ga_{0.8}Sb$) incorporated Te doping to reduce the dark current. The full width at half maximum (FWHM) of the $1^{st}$ satellite superlattice peak from the X-ray diffraction was around 45 arcsec. The cutoff wavelength of the fabricated device was ${\sim}10.2{\mu}m$ (0.12 eV) at 80 K while under an applied bias of -1.4 V. The measured activation energy of the device was ~0.128 eV. The dark current density was shown to be $1.0{\times}10^{-2}A/cm^2$ at 80 K and with a bias -1.5 V. The responsivity was 0.58 A/W at $7.5{\mu}m$ at 80 K and with a bias of -1.5 V.

Self-Aligned, Extremely High Frequency III–V Metal-Oxide-Semiconductor Field-Effect Transistors on Rigid and Flexible Substrates

Wang, Chuan,Chien, Jun-Chau,Fang, Hui,Takei, Kuniharu,Nah, Junghyo,Plis, E.,Krishna, Sanjay,Niknejad, Ali M.,Javey, Ali American Chemical Society 2012 Nano letters Vol.12 No.8

<P>This paper reports the radio frequency (RF) performance of InAs nanomembrane transistors on both mechanically rigid and flexible substrates. We have employed a self-aligned device architecture by using a T-shaped gate structure to fabricate high performance InAs metal-oxide-semiconductor field-effect transistors (MOSFETs) with channel lengths down to 75 nm. RF measurements reveal that the InAs devices made on a silicon substrate exhibit a cutoff frequency (<I>f</I><SUB>t</SUB>) of ∼165 GHz, which is one of the best results achieved in III–V MOSFETs on silicon. Similarly, the devices fabricated on a bendable polyimide substrate provide a <I>f</I><SUB>t</SUB> of ∼105 GHz, representing the best performance achieved for transistors fabricated directly on mechanically flexible substrates. The results demonstrate the potential of III–V-on-insulator platform for extremely high-frequency (EHF) electronics on both conventional silicon and flexible substrates.</P><P><B>Graphic Abstract</B> <IMG SRC='http://pubs.acs.org/appl/literatum/publisher/achs/journals/content/nalefd/2012/nalefd.2012.12.issue-8/nl301699k/production/images/medium/nl-2012-01699k_0001.gif'></P><P><A href='http://pubs.acs.org/doi/suppl/10.1021/nl301699k'>ACS Electronic Supporting Info</A></P>

Bayesian brain source imaging based on combined MEG/EEG and fMRI using MCMC

Jun, Sung C.,George, John S.,Kim, Woohan,Paré,-Blagoev, Juliana,Plis, Sergey,Ranken, Doug M.,Schmidt, David M. Elsevier 2008 NeuroImage Vol.40 No.4

<P><B>Abstract</B></P><P>A number of brain imaging techniques have been developed in order to investigate brain function and to develop diagnostic tools for various brain disorders. Each modality has strengths as well as weaknesses compared to the others. Recent work has explored how multiple modalities can be integrated effectively so that they complement one another while maintaining their individual strengths. Bayesian inference employing Markov Chain Monte Carlo (MCMC) techniques provides a straightforward way to combine disparate forms of information while dealing with the uncertainty in each. In this paper we introduce methods of Bayesian inference as a way to integrate different forms of brain imaging data in a probabilistic framework. We formulate Bayesian integration of magnetoencephalography (MEG) data and functional magnetic resonance imaging (fMRI) data by incorporating fMRI data into a spatial prior. The usefulness and feasibility of the method are verified through testing with both simulated and empirical data.</P>

InAs/GaSb 제2형 응력 초격자 nBn 장적외선검출소자 설계, 제작 및 특성평가

김하술,이훈,Brianna Klein,Nutan Gautam,Elena A. Plis,Stephen Myers,Sanjay Krishna 한국진공학회 2013 Applied Science and Convergence Technology Vol.22 No.6

Long-wave infrared detectors using the type-II InAs/GaSb strained superlattice (T2SL)material system with the nBn structure were designed and fabricated. The band gap energy of the T2SL material was calculated as a function of the thickness of the InAs and GaSb layers by the Kronig-Penney model. Growth of the barrier material (Al0.2Ga0.8Sb) incorporated Te doping to reduce the dark current. The full width at half maximum (FWHM) of the 1st satellite superlattice peak from the X-ray diffraction was around 45 arcsec. The cutoff wavelength of the fabricated device was ∼10.2 μm (0.12 eV) at 80 K while under an applied bias of –1.4 V. The measured activation energy of the device was ∼0.128 eV. The dark current density was shown to be 1.0×10-2 A/cm2 at 80 K and with a bias −1.5 V. The responsivity was 0.58 A/W at 7.5 μm at 80 K and with a bias of −1.5 V. InAs/GaSb 제2형 응력 초격자(strained layer type II superlattice, T2SL)을 이용한 nBn 구조 장적외선 검출소자의 설계 및제작을 하였다. InAs와 GaSb 두께에 따른 T2SL 구조의 장적외선 밴드갭 에너지를 Kronig-Penney 모델을 이용하여 계산하였다. 소자의 암전류 밀도를 줄이기 위해서, nBn 구조에서 장벽층인 Al0.2Ga0.8Sb 성장 중에 Te 보상도핑(compansated doping)을 하였다. 온도(T) 80 K 및 인가전압(Vb) –1.5 V에서, 반응스펙트럼 측정을 통한 소자의 차단파장은 ∼10.2 μm (∼0.122 eV)로 나타났다. 또한 온도 변화에 따른 암전류 측정으로부터 도출된 활성화 에너지는 0.128 eV로 계산 되었다. T=80 K 및 Vb=–1.5 V에서 암전류는 1.0×10-2 A/cm2으로 측정 되었다. 흑체복사 적외선 광원을 이용한 반응도(Responsivity)는 소자 온도 80 K 및 인가전압 –1.5 V의 조건에서 0.58 A/W로 측정되었다.

Ultrathin compound semiconductor on insulator layers for high-performance nanoscale transistors

Ko, Hyunhyub,Takei, Kuniharu,Kapadia, Rehan,Chuang, Steven,Fang, Hui,Leu, Paul W.,Ganapathi, Kartik,Plis, Elena,Kim, Ha Sul,Chen, Szu-Ying,Madsen, Morten,Ford, Alexandra C.,Chueh, Yu-Lun,Krishna, Sanj Nature Publishing Group, a division of Macmillan P 2010 Nature Vol.468 No.7321

Over the past several years, the inherent scaling limitations of silicon (Si) electron devices have fuelled the exploration of alternative semiconductors, with high carrier mobility, to further enhance device performance. In particular, compound semiconductors heterogeneously integrated on Si substrates have been actively studied: such devices combine the high mobility of III??V semiconductors and the well established, low-cost processing of Si technology. This integration, however, presents significant challenges. Conventionally, heteroepitaxial growth of complex multilayers on Si has been explored??but besides complexity, high defect densities and junction leakage currents present limitations in this approach. Motivated by this challenge, here we use an epitaxial transfer method for the integration of ultrathin layers of single-crystal InAs on Si/SiO<SUB>2</SUB> substrates. As a parallel with silicon-on-insulator (SOI) technology, we use ??XOI?? to represent our compound semiconductor-on-insulator platform. Through experiments and simulation, the electrical properties of InAs XOI transistors are explored, elucidating the critical role of quantum confinement in the transport properties of ultrathin XOI layers. Importantly, a high-quality InAs/dielectric interface is obtained by the use of a novel thermally grown interfacial InAsO<SUB>x</SUB> layer (~1?nm thick). The fabricated field-effect transistors exhibit a peak transconductance of ~1.6?mS?쨉m<SUP>??1</SUP> at a drain??source voltage of 0.5?V, with an on/off current ratio of greater than 10,000.