Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Jash Mehta has a total of 31 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Assessment of AlGaN channel high-electron mobility transistors on silicon for high-temperature power electronics

    by Julien Bassaler, Jash Mehta, Stéphanie Rennesson, Fabrice Semond, Sylvie Contreras, Julien Pernot, Yvon Cordier, Farid Medjdoub, Philippe Ferrandis

    Wide and ultrawide bandgap semiconductors present novel opportunities for advancing power electronics, particularly in domains where high-voltage operation and elevated temperature stability are required. In this context, AlGaN/AlGaN heterostructures emerge as promising candidates due to their better intrinsic material properties compared to AlGaN/GaN heterostructures. Furthermore, the tunability of the aluminum molar fraction within AlGaN provides more possibilities for semiconductor engineering. This investigation focuses on evaluating the temperature operation of AlGaN channel high-electron mobility transistors grown on silicon substrates with different aluminum contents. The objective is to assess the thermal reliability of these devices for high-temperature power applications. The results highlight the low off-state current density, a more abrupt subthreshold slope, and a reduction in the temperature-induced decrease in on-state current density in Al-rich AlGaN channel high-electron mobility transistors grown on silicon substrates, positioning them as a promising solution for high-temperature power electronics.

  • N-polar AlGaN HEMTs on silicon for reduced contact resistance

    by A. Barbier-Cueil, J. Mehta, A. Pédèches, F. Medjdoub, F. Semond

    Nitrogen-polar (N-polar) AlGaN-channel high electron mobility transistors (HEMTs) offer a promising pathway to overcome the limitations of equivalent metal-polar (M-polar) devices, particularly the high contact resistance associated with increased aluminum content. In this work, we report the growth and fabrication of N-polar AlGaN-channel HEMTs on a silicon substrate by ammonia molecular beam epitaxy (NH3-MBE). N-polarity is achieved using an epitaxial NbN polarity-inversion layer, enabling the growth of a buffer stack with adequate structural and electrical quality. Two heterostructure designs are investigated. Capacitance–voltage and Hall measurements confirm the formation of a high-density two-dimensional electron gas with sheet carrier densities up to 2.4×1013 cm−2. Owing to the elimination of the Al-rich barrier between the surface and the channel (present in M-polar configuration) and improving surface morphology, encouraging low-resistance Ohmic contacts are achieved, with contact resistance reduced below 1 Ω mm on the Al0.2Ga0.8N channel. These results demonstrate the viability of N-polar AlGaN-channel HEMTs on silicon and highlight their potential for next-generation ultra-wide-bandgap power devices on Si.

  • Development of FETs based on ultra-wide band gap materials for high voltage power electronics

    by Jash Rinku Mehta

    Développement de FET basés sur des matériaux à bande interdite ultra-large pour l'électronique de puissance haute tension: Les semi-conducteurs à large bande interdite (WBG) tels que le SiC et le GaN présentent des propriétés physiques supérieures à celles du silicium. Cependant, les matériaux ou diélectriques à bande interdite encore plus large, tels que le nitrure d'aluminium et de gallium (AlGaN) ou le nitrure d'aluminium (AlN), permettent de créer des hétérojonctions tout en élargissant considérablement les possibilités de conception des dispositifs pour les systèmes de conversion d'énergie à moyenne/haute puissance. Cette thèse se concentrera sur la démonstration de semi-conducteurs à bande passante ultra-large (UWBG) pour le développement de transistors à effet de champ (FET) à haute tension. Dans ce contexte, une approche basée sur la réduction de l'épaisseur du canal GaN a été développée sur des substrats d'AlN. L'évolution de la densité de courant à l'état passant et de la tension de claquage a été réalisée avec différentes épaisseurs de canaux GaN. Un champ de claquage latéral élevé de l'AlN allant jusqu'à 10MV/cm a été démontré expérimentalement. D'autre part, l'intégration de substrats d'AlN dans les fonderies de silicium actuelles est un défi. On peut donc dire que "ce qui peut être fait sur le silicium sera fait sur le silicium". Pour relever ce défi, nous avons adopté une approche visant à démontrer de nouveaux HFET à base de canaux AlGaN sur substrat de silicium pour des applications à haute tension. L'impact de la composition en Al des couches de barrière et du canal sur les performances électriques et thermiques de divers HFET AlGaN/AlGaN a été étudié. Par la suite, nous avons démontré la robustesse des HFET à canal AlGaN sur AlN massif, appropriés à l'électronique de puissance en conditions extrêmes et une tension d'opération supérieure à 2 kV. Le principal défi pour les HFET à base d'AlGaN est de minimiser les résistances de contacts de source et de drain. Nous avons pu démontré une densité de courant élevée (> 0,2 A/mm) ainsi que des transistors à canal AlGaN riche en Al sur substrat de silicium avec un champ électrique de claquage moyen supérieur à 4 MV/cm. . Ces résultats mettent en évidence une approche prometteuse qui pourrait potentiellement conduire à des solutions pour une électronique de puissance plus durable et plus efficace avec des dimensions de dispositifs réduites ainsi que des capacités de fonctionnement à haute tension/haute température accrus.

  • Open Access

    Effects of GaN channel downscaling in AlGaN–GaN high electron mobility transistor structures grown on AlN bulk substrate

    by Reda Elwaradi, Jash Mehta, Thi Huong Ngo, Maud Nemoz, Catherine Bougerol, Farid Medjdoub, Yvon Cordier

    In this work, two series of AlGaN/GaN/AlN high electron mobility transistor (HEMT) heterostructures have been grown by molecular beam epitaxy on AlN bulk substrates. The effects of reduction in the GaN channel thickness as well as the AlGaN barrier thickness and composition on structural and electrical properties of the heterostructures have been studied. The material analysis involved high-resolution x-ray diffraction, atomic force microscopy, and cross-sectional transmission electron microscopy. In a first series of HEMT structures grown with an aluminum content of 30% in the AlGaN barrier, the channel downscaling results in a reduction in the GaN strain relaxation rate but at the expense of degradation in the mean crystal quality and in the electron mobility with a noticeable increase in the sheet resistance. An opposite trend is noticed for the three-terminal breakdown voltage of transistors, so that a trade-off is obtained for a 50 nm width GaN channel HEMT, which exhibits a sheet resistance of 1700 Ω/sq. with transistors demonstrating three-terminal breakdown voltage up to 1400 V for 40 μm gate to drain spacing with static on resistance Ron = 32 mΩ cm2. On the other hand, a second series of HEMT structures with high aluminum content AlGaN barriers and sub-10 nm GaN channels have been grown perfectly strained with high sheet carrier densities allowing to preserve sheet resistances in the range of 880–1050 Ω/sq.


  • Open Access

    High Al-content AlGaN channel high electron mobility transistors on silicon substrate

    by J. Mehta, I. Abid, J. Bassaler, J. Pernot, P. Ferrandis, M. Nemoz, Y. Cordier, S. Rennesson, S. Tamariz, F. Semond, et al.


  • Open Access

    A Federated Approach to Predicting Emojis in Hindi Tweets

    by Deep Gandhi, Jash Mehta, Nirali Parekh, Karan Waghela, Lynette D'Mello, Zeerak Talat

    The use of emojis affords a visual modality to, often private, textual communication. The task of predicting emojis however provides a challenge for machine learning as emoji use tends to cluster into the frequently used and the rarely used emojis. Much of the machine learning research on emoji use has focused on high resource languages and has conceptualised the task of predicting emojis around traditional server-side machine learning approaches. However, traditional machine learning approaches for private communication can introduce privacy concerns, as these approaches require all data to be transmitted to a central storage. In this paper, we seek to address the dual concerns of emphasising high resource languages for emoji prediction and risking the privacy of people's data. We introduce a new dataset of $118$k tweets (augmented from $25$k unique tweets) for emoji prediction in Hindi, and propose a modification to the federated learning algorithm, CausalFedGSD, which aims to strike a balance between model performance and user privacy. We show that our approach obtains comparative scores with more complex centralised models while reducing the amount of data required to optimise the models and minimising risks to user privacy.

  • Open Access

    Low Buffer Trapping Effects above 1200 V in Normally off GaN-on-Silicon Field Effect Transistors

    by Idriss Abid, Youssef Hamdaoui, Jash Mehta, Joff Derluyn, Farid Medjdoub

    We report on the fabrication and electrical characterization of AlGaN/GaN normally off transistors on silicon designed for high-voltage operation. The normally off configuration was achieved with a p-gallium nitride (p-GaN) cap layer below the gate, enabling a positive threshold voltage higher than +1 V. The buffer structure was based on AlN/GaN superlattices (SLs), delivering a vertical breakdown voltage close to 1.5 kV with a low leakage current all the way to 1200 V. With the grounded substrate, the hard breakdown voltage transistors at VGS = 0 V is 1.45 kV, corresponding to an outstanding average vertical breakdown field higher than 2.4 MV/cm. High-voltage characterizations revealed a state-of-the-art combination of breakdown voltage at VGS = 0 V together with low buffer electron trapping effects up to 1.4 kV, as assessed by means of substrate ramp measurements.


  • Open Access

    AlGaN Channel High Electron Mobility Transistors with Regrown Ohmic Contacts

    by Idriss Abid, Jash Mehta, Yvon Cordier, Joff Derluyn, Stefan Degroote, Hideto Miyake, Farid Medjdoub

    High power electronics using wide bandgap materials are maturing rapidly, and significant market growth is expected in a near future. Ultra wide bandgap materials, which have an even larger bandgap than GaN (3.4 eV), represent an attractive choice of materials to further push the performance limits of power devices. In this work, we report on the fabrication of AlN/AlGaN/AlN high-electron mobility transistors (HEMTs) using 50% Al-content on the AlGaN channel, which has a much wider bandgap than the commonly used GaN channel. The structure was grown by metalorganic chemical vapor deposition (MOCVD) on AlN/sapphire templates. A buffer breakdown field as high as 5.5 MV/cm was reported for short contact distances. Furthermore, transistors have been successfully fabricated on this heterostructure, with low leakage current and low on-resistance. A remarkable three-terminal breakdown voltage above 4 kV with an off-state leakage current below 1 μA/mm was achieved. A regrown ohmic contact was used to reduce the source/drain ohmic contact resistance, yielding a drain current density of about 0.1 A/mm.