ALD/ALE 2026 Session AA2-TuM: ALD Dielectric Applications

Tuesday, June 30, 2026 10:45 AM in Room Ybor Salons I-IV
Tuesday Morning

Session Abstract Book
(452 KB, Mar 16, 2026)
Time Period TuM Sessions | Abstract Timeline | Topic AA Sessions | Time Periods | Topics | ALD/ALE 2026 Schedule

Start Invited? Item
10:45 AM AA2-TuM-12 Sub-5 nm Indium Oxynitride Channel in Top-gated FETs Fabricated by PEALD for High-Performance 3D Transistor
Doo San Kim, Minjong Lee, Soham Shirodkar (The University of Texas at Dallas); Min Gyeong Jo (Hanyang University); Thi Thu Huong Chu, Dushyant Narayan, Dan Le (The University of Texas at Dallas); Youngbae Ahn, Ja-Yong Kim, Seung Wook Ryu (SK hynix); Jiyoung Kim (The University of Texas at Dallas)

As silicon scaling approaches its physical limits, indium-based oxide semiconductors have emerged as promising channel materials because they maintain high carrier mobility and drive current even at channel thickness below 5 nm, enabling superior scalability. Among them, In2O3 exhibits excellent drive current; however, hydrogen incorporation leads to hydroxyl (OH) formation, resulting in device instability and threshold voltage shifts. Incorporating nitrogen to form indium oxynitride (InON) is expected to enhance mobility while effectively suppressing the formation of hydrogen induced OH bonds, thereby improving device stability.

Despite these advantages, direct deposition of InON via atomic layer deposition (ALD) is challenging because the bonding dissociation energy of In-O (~346 kJ/mol) is significantly stronger than that of In-N (~186 kJ/mol), making it difficult to incorporate sufficient nitrogen during the simultaneous supply of oxygen and nitrogen reactants.[1,2] In this work, we propose a two-step channel formation strategy: the initial deposition of an indium nitride (InN) thin film followed by an oxidation process to convert it into an InON channel layer.

In this presentation, we report the deposition of sub-5 nm conformal InN films via Hollow-Cathode Plasma (HCP)-enhanced ALD at 240 °C. To improve the structural and electrical stability of the channel, an annealing-based oxidation process was applied, converting InN into the more robust InON and evaluating their integration into top-gated (TG) thin-film transistor (TFT) devices. We investigated the influence of process parameters such as plasma power, process pressure, and deposition temperature for the film characteristics. The 5 nm InON layer was deposited at 240 °C, followed by the deposition of a 5 nm hafnium zirconium oxide (HZO) dielectric layer, and finally a TG-FET was demonstrated using Ni metal contacts. We presented InON TG-FET results, including transfer and output characteristics, temperature dependence, and channel length scaling. Our findings demonstrate that InON-based TFTs with a 100nm channel length achieve an on-current of ~1.2mA/µm at VD = 1 V, an Ion/Ioff ratio exceeding 109, and a contact resistivitywith Ni of 4 x 10-8Ω·cm2. In addition, the △Vth under positive and negative bias stress is 90 and 15 mV, at Vov = ±1 V.

This work was supported by SK hynix Inc.

[1] Imran, A., et al., Adv. Mater. Interfaces 10, 2200105 (2022).

[2] Luo, Y. R., "Bond Dissociation Energies" in CRC Handbook of Chemistry and Physics. (2010).
11:00 AM AA2-TuM-13 VFb Control Technique of SiO2/Dipole/HfZrO2 Stack Structure Using New LaTiO and AlTiO Dipoles
Tomomi Sawada, Toshihide Nabatame, Hiromi Miura, Manami Miyamoto, Takashi Onaya, Kazuhito Tsukagoshi, Naoki Fukata, Wipakorn Jevasuwan (National Institute for Materials Science (NIMS)); Shinji Migita (National Institute of Advanced Industrial Science and Technology (AIST))

Dipole technique has been widely investigated to control the threshold voltage (Vth) of metal/High-k CMOS with GAA gate stack in 50 mV increments. La2O3 and Al2O3 are generally employed as n- and p-type dipoles. However, it remains a big issue of a large Vth shift in the sub-nm region of the dipole layer. In this paper, we investigated flatband voltage (Vfb) shift for SiO2/dipole/HfZrO2 stack structure using new LaTiO and AlTiO as n- and p-type dipoles.

p-Si/SiO2/dipole/HfZrO2/Au capacitor was fabricated at a maximum process temperature of 400 °C as follows: La2O3, TiO2 and Al2O3 dipole layers were deposited on p-Si/SiO2 (3.6 nm) by ALD at 250 °C with H2O and La(iPrCp)3, TDMAT, and TMA, respectively. LaTiO and AlTiO dipoles were also prepared by changing each ALD cycle of La2O3 and TiO2, and Al2O3 and TiO2 by ALD at 250 °C with H2O, respectively. Next, an HfZrO2 (2 nm) film was deposited on dipole layer by ALD at 250 °C using TEMA(Hf/Zr)(Hf/Zr = 1/1) cocktail and H2O. Au gate electrode was deposited on HfZrO2 film to form capacitor. Finally, forming gas annealing was carried out at 400 °C in 3%H2.

The La2O3 and Al2O3 dipole capacitors exhibited negative and positive Vfb shifts as the La2O3 and Al2O3 dipole thickness increased, respectively. The Vfb shifts for the La2O3 and Al2O3 dipole capacitors were almost saturated, showing -0.35 and +0.32 V, respectively, when the dipole thickness was 0.8 nm. On the other hand, the TiO2 dipole capacitor exhibited a Vfb shift of nearly 0 V regardless of the dipole thickness. This is due to differences in the strength and direction of the dipoles of La2O3, Al2O3 and TiO2 at the SiO2/dipole interface. The LaTiO (La = 0.46) dipole capacitors reduced the negative Vfb shift from -0.29 to -0.11 V compared to the La2O3 dipole capacitors when the dipole thickness was 0.2 nm. With a 0.2 nm thick dipole, the AlTiO (Al = 0.62) dipole capacitors also reduced the positive Vfb shift by approximately 25% compared to the Al2O3 dipole capacitors. This indicates that the dipole strength of La2O3 and Al2O3 can be effectively reduced by adding TiO2 without dipole effect. In the LaTiO and AlTiO dipole capacitors, La and Ti atoms, and Al and Ti atoms, respectively, were piled up at the SiO2/HfZrO2 interface using STEM and EDS analysis. By employing new LaTiO and AlTiO dipoles in addition to La2O3 and Al2O3, Vfb controllability could be enhanced in the sub-nm region of the dipole layer. This is because TiO2 without the dipole effect was added. This presentation is partially based on the results of the post-5G project (JPNP20017), entrusted to LSTC, commissioned by NEDO.
11:15 AM AA2-TuM-14 Low-Temperature High-Pressure Deuterium Annealing for Defect Passivation in ALD-Deposited HfO2 High-k Film
Ji-Yeon Park, Seok-Won Lim, Gi-Beom Park, Chang-Kyun Park, Jin-Seong Park (Hanyang University, Korea)
Atomic layer deposition (ALD) is a key process technology for integrating high-k gate dielectrics into advanced semiconductor devices, providing excellent thickness control, compositional uniformity, and conformality. As conventional SiO2 gate dielectrics reach their physical scaling limits due to increased direct tunneling leakage, high-k materials such as HfO2 have been widely adopted [1]. Although ALD enables the formation of high-quality HfO2 thin films, device performance is strongly influenced by defect states formed at the HfO2/Si interface [2].In this study, ALD-deposited HfO2 thin films grown using a Cp-Hf/O3 process were investigated with a focus on the correlation between HfO2/Si interfacial defect and electrical characteristics. Post-deposition deuterium treatment was employed not to modify the intrinsic quality of the ALD film, but as a post-ALD interface engineering and analytical tool to evaluate and improve defect states located at or near the interface. Electrical characterization using C–V and J–E measurements showed reductions in interface trap density, improved flat-band voltage stability, decreased leakage current density, and enhanced breakdown field after deuterium treatment, suggesting effective passivation of interface-related defects. Secondary ion mass spectrometry (SIMS) and X-ray photoelectron spectroscopy (XPS) analyses were conducted to examine deuterium incorporation behavior and changes in chemical bonding states, and their relationship to the observed electrical improvements was analyzed. These results indicate that the electrical performance of ALD-grown HfO2 high-k dielectrics is governed primarily by interface quality rather than bulk film limitations. This work highlights the importance of post-ALD interface engineering and provides insight into ALD-compatible strategies for achieving reliable high-k/Si interfaces in next-generation semiconductor devices.Reference1) Wilk, Glen D., Robert M. Wallace, and JáM Anthony. "High-κ gate dielectrics: Current status and materials properties considerations." Journal of applied physics 89.10 (2001): 5243-5275. 2) E. P. Gusev,et al " Ultrathin high-k gate stacks for advanced CMOS devices," Microelectronic Engineering 69 (2003): 145-151. View Supplemental Document (pdf)
11:30 AM AA2-TuM-15 Infiltration of Porous SiOCH Thin Films by High-k Materials: Toward Nanocomposites with Enhanced Dielectric Properties
Julie Chaussard, Stéphane Cadot, Marc Veillerot, Hélène Coudert-Alteirac, Nicolas Gauthier, Nicolas Bernier, Chloé Guérin, Aude Lefèvre (CEA/LETI-University Grenoble Alpes, France); Patrice Gonon (Université Grenoble Alpes, CNRS); Vincent Jousseaume (CEA/LETI-University Grenoble Alpes, France)

Dielectric materials, especially insulating polymers, are crucial for various technological applications due to their high breakdown voltage and low cost [1]. However, they generally exhibit low dielectric constant and thermal stability, limiting their use in field as power electronics.

To enhance their dielectric constant, recent studies have explored the elaboration of nanocomposites by adding alumina nanoparticles into a polymer matrix, resulting in significant improvements in dielectric constant and breakdown voltage even at low concentrations [2]. Despite these promising results, the elaboration methods used were incompatible with standard microelectronic processes.

This work studies the elaboration of nanocomposite thin films using microelectronic materials stable up to 400°C, deposited by vacuum deposition techniques. To this end, 100 nm porous SiOCH were deposited by PE-CVD on Si wafers using a porogen approach [3]. Then high-k material was infiltrated into this porous thin film using ALD. The filling process for different materials and precursors was characterized using physicochemical methods (XRF, ellipsometry, XRR, ToF-SIMS, FTIR, TEM). Then, the electrical properties were studied (dielectric constant, dielectric losses, breakdown voltage). In the case of trimethylaluminum/O3, the results highlight that the filling of the nanometric pores of the porous SiOCH occurs in the first ALD cycles (<5). Using an optimized recipe, a maximum filling of ≈50% of the porosity was obtained. As the number of cycles increases, high-k growth continues on the surface at a rate similar to that observed on Si. FTIR analysis reveals the formation of -OH groups in the first cycles, which increases dielectric constant and dielectric losses. A similar behavior is observed when using H2O as oxidant. However, after 25 ALD cycles, the disappearance of -OH groups is observed for the TMA/O3 process. One hypothesis is that the alumina layer formed on top of the porous structure becomes thick enough to limit H2O adsorption into the matrix. This leads to the elaboration of composites with a dielectric constant 42% higher than that of the porous matrix while maintaining low dielectric losses. Finally, infiltration of porous SiOCH with HfO2 and HfZrO2 (from chlorinated precursors and water), which have higher dielectric constant than Al2O3, also enables enhanced dielectric properties without the presence of -OH groups. But in this case the filling rate is slightly lower than that observed with alumina.

In conclusion, this work demonstrates an easy approach for the elaboration of nanocomposites using standard microelectronic materials to achieve a broader range of electrical properties.

[1] Thakur, Y. et al.; Nanoscale 2017, 9 (31), 10992–10997.

[2] Zhang, T. et al.; Sci. Adv. 2020, 6 (4), eaax6622.

[3] Grill, A. et al.; Appl. Phys. Lett. 2001, 79 (6), 803–805.

11:45 AM AA2-TuM-16 Electrical Characterization of High-k ALD TiO2 on AlGaN/GaN HEMT Structures
Neeraj Nepal, James G. Champlain, Vikrant J. Gokhale, Peter M Litwin, Brian P Downey, Virginia D Wheeler (U.S. Naval Research Laboratory)

Transport and breakdown field characteristics of GaN-based HEMT technologies are limited by high and non-uniform peak electric fields at the drain-edge of the gate. The non-uniform peak field causes premature electric field induced breakdown limiting the device performance [1]. These deleterious effects can be mitigated by integrating films with high dielectric constants (high-k, >20), such as TiO2 [2-4]. For HEMTs, increased k leads to increased power handling capability without significant reduction in operation frequency. Thus, the goal is to integrate the highest k material possible. ALD provides flexibility in device design, integration and back-end-of line compatibility, making it a promising route for these high-k films. However, ALD high-k dielectrics with ideal interfaces and reduced trapping on AlGaN/GaN HEMTs have not been reported. In this talk, we report on the optimization and electrical characterization of the ALD TiO2 interface with AlGaN-barrier HEMT structures.

Previously optimized deposition parameters [3] served as the baseline to further optimize electrical performance by pretreating the dielectric-barrier interface of Al0.25Ga0.75N/GaN HEMT structures. Ex-situ pretreatments including NH4OH, UV O3+BOE, piranha and O2 plasma descum were studied. Similarly, in-situ UHV annealing, plasma treatments and their combinations were also studied. The optimum interface was found to be a combination of ex-situ O2-plasma descum and in-situ H2 and N2 plasma treatments. Initial results using this preparation, a 20nm TiO2 film demonstrated a reduction in gate leakage by 105 compared to a Schottky gate,zero CV hysteresis, a dielectric constant ≥ 50, and no significant change in 2DEG density. Even though the ALD TiO2 has a narrower bandgap than AlGaN-barrier material, its large dielectric constant provides a pathway forward for dielectric permittivity engineering within the device structure. This engineering approach can improve breakdown voltage, lowers gate leakage and minimizes non-ideality over the state-of-the-art.

References:

  1. Turuvekere et al., IEEE Trans. Electr. Dev. 60, 3157 (2013).
  2. Rahman et al., Appl. Phys. Lett. 119, 193501 (2021).
  3. Nepal et al., APL Electr. Dev. 1, 036102-1 (2025).
View Supplemental Document (pdf)
Session Abstract Book
(452 KB, Mar 16, 2026)
Time Period TuM Sessions | Abstract Timeline | Topic AA Sessions | Time Periods | Topics | ALD/ALE 2026 Schedule