Search

Optical phase-change materials have enabled nonvolatile programmability in integrated photonic circuits by leveraging a reversible phase transition between amorphous and crystalline states. To control these materials in a scalable manner on-chip, heating the waveguide itself via electrical currents is an attractive option which has been recently explored using various approaches. Here, we compare the heating efficiency, fabrication variability, and endurance of two promising heater designs which can be easily integrated into silicon waveguides—a resistive microheater using n-doped silicon and one using a silicon p-type/intrinsic/n-type (PIN) junction. Raman thermometry is used to characterize the heating efficiencies of these microheaters, showing that both devices can achieve similar peak temperatures but revealing damage in the PIN devices. Subsequent endurance testing and characterization of both device types provide further insights into the reliability and potential damage mechanisms that can arise in electrically programmable phase-change photonic devices.

Homogeneous ternary oxides of silicon-, niobium-, and molybdenum-aluminate were deposited by plasma-enhanced ALD using sequential metal precursor pulses prior to the oxidation step, to reduce interfacial defects usually observed in nanolaminate growth. The growth kinetics can be understood in terms of competitive adsorption. Trimethyl aluminum (TMA) is strongly chemisorbed to the growth surface and does not permit coadsorption of any of the other precursors; when we lead with a TMA pulse, the resulting film is always Al2O3. When we lead with the Si or Nb precursors, the growth surface is partially saturated, but open sites are available for TMA coadsorption. The Mo precursor is weakly chemisorbed and is largely displaced by a subsequent TMA dose. As compared to nanolaminate films of the constituent binary oxides, the interface state density is reduced by up to a factor of 5.

We report integration of chalcogenide phase change materials (GSST and Sb2Se3) with doped silicon-on-insulator heater arrays fabricated in a 90-nm silicon foundry. Reversible electrothermal switching of transmissive metasurfaces and integrated photonic devices are demonstrated.

Optical phase change materials (PCMs) are a unique class of materials which exhibit extraordinarily large optical property change (e.g. refractive index change > 1) when undergoing a solid-state phase transition, and they have witnessed increasing adoption in active integrated photonics and metasurface devices in recent years. Here we report integration of chalcogenide phase change materials in the Lincoln Laboratory 8-inch Si foundry process and the demonstration of electrothermally switched phase-change photonic devices building on a wafer-scale silicon-on-insulator heater platform.

The quest for efficient programmable photonic circuits requires active optical components with low insertion loss, small form factor, and low electrical power consumption [1] . The most common low-loss phase shifter – the key building block – employs electrically driven heaters to exploit thermo-optic effects. However, these devices require a constant power supply, suffer from thermal crosstalk, and their sizes are typically >100 µm. Here, we demonstrate a novel alternative that meets all the aforementioned requirements. We use silicon-on-insulator (SOI) waveguides with patterned n ++ - n - n ++ doped-silicon microheaters to electrothermally switch, with a single pulse, a cell of Sb 2 Se 3 between its two optically distinct, nonvolatile states: amorphous and crystalline. Sb 2 Se 3 is a phase change material (PCM) that offers a ∆n ≈ 0.77 upon switching between states with vanishingly small losses at 1550nm [2] . This unique set of properties allows for phase-only modulation [3] , which sets a precedent in a growing PCM photonics field familiar with PCMs for amplitude modulation at the telecom band [4] .

Energy-efficient programmable photonic integrated circuits (PICs) are the cornerstone of on-chip classical and quantum optical technologies. Optical phase shifters constitute the fundamental building blocks which enable these programmable PICs. Thus far, carrier modulation and thermo-optical effect are the chosen phenomena for ultrafast and low-loss phase shifters, respectively; however, the state and information they carry are lost once the power is turned off-they are volatile. The volatility not only compromises energy efficiency due to their demand for constant power supply, but also precludes them from emerging applications such as in-memory computing. To circumvent this limitation, we introduce a novel phase shifting mechanism that exploits the nonvolatile refractive index modulation upon structural phase transition of Sb$_{2}$Se$_{3}$, a bi-stable transparent phase change material. A zero-static power and electrically-driven phase shifter was realized on a foundry-processed silicon-on-insulator platform, featuring record phase modulation up to 0.09 $\pi$/$\mu$m and a low insertion loss of 0.3 dB/$\pi$, which can be further improved upon streamlined design. We also pioneered a one-step partial amorphization scheme to enhance the speed and energy efficiency of PCM devices. A diverse cohort of programmable photonic devices were demonstrated based on the ultra-compact PCM phase shifter., Comment: 15 pages with 6 figures and 1 table

We demonstrate a nonvolatile, small-form-factor (20 m) phase-shifter based on low-loss phase-change materials (Ge2Sb2Se4Te1 and Sb2Se3) on SOI waveguides. We achieve continuous optical phase modulation and an MZI switch with 30 dB extinction ratio.

Reconfigurable photonic systems featuring minimal power consumption are crucial for integrated optical devices in real-world technology. Current active devices available in foundries, however, use volatile methods to modulate light, requiring a constant supply of power and significant form factors. Essential aspects to overcoming these issues are the development of nonvolatile optical reconfiguration techniques which are compatible with on-chip integration with different photonic platforms and do not disrupt their optical performances. In this paper, a solution is demonstrated using an optoelectronic framework for nonvolatile tunable photonics that employs undoped-graphene microheaters to thermally and reversibly switch the optical phase-change material Ge$_2$Sb$_2$Se$_4$Te$_1$ (GSST). An in-situ Raman spectroscopy method is utilized to demonstrate, in real-time, reversible switching between four different levels of crystallinity. Moreover, a 3D computational model is developed to precisely interpret the switching characteristics, and to quantify the impact of current saturation on power dissipation, thermal diffusion, and switching speed. This model is used to inform the design of nonvolatile active photonic devices; namely, broadband Si$_3$N$_4$ integrated photonic circuits with small form-factor modulators and reconfigurable metasurfaces displaying 2$\pi$ phase coverage through neural-network-designed GSST meta-atoms. This framework will enable scalable, low-loss nonvolatile applications across a diverse range of photonics platforms.

O-PCMs uniquely offer exceptionally large refractive index modulation with minimal loss penalty, made possible through a dielectric-dielectric structural transition. Here we discuss our recent work on reconfigurable integrated photonics and metasurface based on O-PCMs. © 2020 The Author(s)

The H-terminated surface of diamond when activated with NO2 produces a surface conduction layer that has been used to make field effect transistors (FETs). Previous reports have suggested that during NO2 exposure (NO2-activation), NO2− forms on the diamond surface and generates positive carriers (holes) in the diamond, making the diamond surface conductive. We report here on X-ray-photoelectron-spectroscopy (XPS) surface characterization of single crystal diamonds and on infrared absorption of diamond powder. After activation, XPS showed the presence of N atoms on the diamond surface, but infrared absorption found no evidence of NO2−, but instead NO3− is present on the diamond surface. Two wet chemistry techniques determined the concentration of NO3− per milligram of diamond powder. With the powder's surface area measured by the BET technique, the surface NO3− concentration was measured to be between 6.2 × 1013 and 8.2 × 1013 cm−2. This is in the same range as the carrier densities, 3 × 1013 to 9 × 1013 cm−2, determined by Hall mobility and surface conductivity measurements of single crystal diamonds. Using similar techniques, the concentration of NO2− was determined to be Both the surface conductance and the surface H atoms are stable in dry nitrogen, with or without NO2-activation, but the surface conductance, the concentrations of H atoms both with and without activation and NO3− decrease when exposed to laboratory air over a period of hours to days. Infrared absorption measurements showed the reduction of surface NO3− and H atoms during laboratory air exposure, but gave no indication of what reactions are responsible for their loss in laboratory air.

The H-terminated surface of diamond when activated with NO2 produces a surface conduction layer that has been used to make FETs. Variations in processing can significantly affect this conduction layer. This article discusses the effect of diamond surface preparation and H termination procedures on surface conduction. Surface preparations that generate a rough surface result in a more conductive surface with the conductivity increasing with surface roughness. We hypothesize that the increase in conductance with roughness is the result of an increase of reactive sites that generate the carriers. Roughening the diamond surface is just one way to generate these sites and the rough surface is believed to be a separate property from the density of surface reactive sites. The presence of C in the H2 plasma used for H termination decreases surface conductance. A simple procedure for NO2 activation is demonstrated. Interpretation of electrical measurements and possible alternatives to activation with NO2 are discussed. Using Kasu's oxidation model for surface conductance as a guide, compounds other than NO2 have been found to activate the diamond surface as well.

A lack of inversion symmetry coupled with the presence of time-reversal symmetry endows 2D transition metal dichalcogenides with individually addressable valleys in momentum space at the K and K' points in the first Brillouin zone. This valley addressability opens up the possibility of using the momentum state of electrons, holes, or excitons as a completely new paradigm in information processing. The opportunities and challenges associated with manipulation of the valley degree of freedom for practical quantum and classical information processing applications were analyzed during the 2017 Workshop on Valleytronic Materials, Architectures, and Devices; this Review presents the major findings of the workshop.

Two scanning electron beam lithography (SEBL) patterning processes have been developed, one positive and one negative tone. The processes feature nanometer-scale resolution, chemical amplification for faster throughput, long film life under vacuum, and sufficient etch resistance to enable patterning of a variety of materials with a metal-free (CMOS/MEMS compatible) tool set. These resist processes were developed to address two limitations of conventional SEBL resist processes: (1) low areal throughput and (2) limited compatibility with the traditional microfabrication infrastructure.

A tunable electron beam generated plasma system has been developed for selective cold plasma treatment of dilute concentrations (1-3,000 ppm range) of hazardous compounds in gaseous waste treatment. This system, referred to as the Tunable Hybrid Plasma (THP), has shown a high degree of efficiency and effectiveness in both laboratory and field tests. Decomposition energy requirements are in the 100 eV per molecule range for treatment of carbon tetrachloride and 10 eV for treatment of trichloroethylene. A cost comparison has been made between the Tunable Hybrid Plasma (THP) technology and three conventional technologies used for emission control of volatile organic compounds (VOCs): granular activated carbon, thermal incineration, and catalytic oxidation. In addition to its environmentally attractive features, THP technology has the potential to be lower cost than other technologies over a range of concentrations and flow rates. Cost projections for the THP system for decomposition of trichloroet-hylene are around 50 cents/lb for initial concentrations in the few hundred ppm range and flow rates of 5,000 cfm or greater and around $1/lb for 1,000 cfm flow rates. Cost projections for carbon tetrachloride and trichloroethane decomposition using the THP technology are several dollars per pound. The costs for THP treatment are generally significantly lower than costs for use of granular activated carbon and are also quite competitive with costs for thermal incineration and catalytic oxidation.

Fully depleted silicon-on-insulator transistors coated with gadolinium oxide are shown to be effective thermal neutron dosimeters. The theoretical neutron detection efficiency is calculated to be higher for Gd2O3 than for other practical converter materials. Proof-of-concept dosimeter devices were fabricated and tested during thermal neutron irradiation. The transistor current changes linearly with neutron dose, consistent with increasing positive charge in the SOI buried oxide layer generated by ionization from high energy 157Gd(n,γ)158Gd conversion electrons. The measured neutron sensitivity is approximately 1/6 the maximum theoretical value, possibly due to electron–hole recombination or conversion electron loss in interconnect wiring above the transistors.

A novel load–displacement sensing instrument has been designed and fabricated to characterize the fracture properties of brittle thin films at low temperature (approximately −30 °C) and pressure (1.6e-4 Pa) environments. In this study, the instrument was used to investigate the effects of harsh environments on the fracture behaviors of organosilicate glass (OSG) and silicon carbonitride (SiCN) thin films under four-point bend loading. Experimental results showed that the fracture strengths of film stacks are the highest when the environment contains a very low water molecule concentration. This condition can be achieved by purging the testing chamber with pure nitrogen or reducing the chamber pressure to less than 1 Pa. In contrast, cracks propagated readily along OSG/SiCN interfaces when experiments were performed in deionized water. The effects of low temperature (approximately −30 °C) and pressure on thin film fracture were also studied, and the results demonstrated that there is no observed degradation of the OSG fracture properties. X-ray photoelectron spectroscopy (XPS) technique was used to identify the chemical composition of the fracture surfaces.

The effective work function of a reactively sputtered TiN metal gate is shown to be tunable from 4.30 to 4.65 eV. The effective work function decreases with nitrogen flow during reactive sputter deposition. Nitrogen annealing increases the effective work function and reduces Dit. Thinner TiN improves the variation in effective work function and reduces gate dielectric charge. Doping of the polysilicon above the TiN metal gate with B or P has negligible effect on the effective work function. The work-function-tuned TiN is integrated into ultralow-power fully depleted silicon-on-insulator CMOS transistors optimized for subthreshold operation at 0.3 V. The following performance metrics are achieved: 64-80-mV/dec subthreshold swing, PMOS/NMOS on-current ratio near 1, 71% reduction in Cgd, and 55% reduction in Vt variation when compared with conventional transistors, although significant short-channel effects are observed.

Ultralow-power electronics will expand the technological capability of handheld and wireless devices by dramatically improving battery life and portability. In addition to innovative low-power design techniques, a complementary process technology is required to enable the highest performance devices possible while maintaining extremely low power consumption. Transistors optimized for subthreshold operation at 0.3 V may achieve a 97% reduction in switching energy compared to conventional transistors. The process technology described in this article takes advantage of the capacitance and performance benefits of thin-body silicon-on-insulator devices, combined with a workfunction engineered mid-gap metal gate.

Energy-efficient computing is critical for the distributed wireless sensor nodes and computing hubs which will form the Internet of Things (IoT). To meet the power requirements of IoT systems it is required to scale the operating voltage aggressively into the subthreshold/near-threshold region. To combat the performance penalty and increased sensitivity to PVT variation of subthreshold operation computational systems must be optimized at the architectural, circuit, and device level. Subthreshold circuit design techniques, low-voltage transistors based on FDSOI, and high-density 3D integration form a synergistic ecosystem which will enable useful computational complexity within small, energy-starved systems.

We describe a robust level converter that is capable of converting up to 1.2V from a subthreshold input voltage of 60mV. Measurement results are given for test circuits fabricated in a 90nm FDSOI subthreshold optimized process.

The “ouzo effect” enables one to create a dispersion of small droplets in a surrounding liquid phase without the use of surfactants, dispersing agents, or mechanical agitation: a phenomenon which can be of value in many disciplines. In the quantitative studies presented here, dispersions of oil droplets in water are formed by the addition of water to a solution of the oil dissolved in a solvent. This causes the oil to supersaturate and then nucleate into small droplets. The mean droplet diameter is a function only of the oil-to-solvent ratio at a given temperature. The number density of droplets formed can be controlled independently from the droplet diameter by changing the amount of water added. Smaller droplets are formed by using more hydrophilic cosolvents. The droplet size distribution is typically log−normal. The width of the distribution can be narrowed by mixing the components at an elevated temperature and then allowing the dispersion to cool.

Oxide etching yield has been measured directly with inductively coupled fluorocarbon plasmas. The yields measurement technique of this work can provide useful information for feature profile evolution modeling, which is essential to understand various issues in oxide etching such as reactive ion etching (RIE) lag, inverse RIE lag, etch stop, microtrenching, bowing, etc. Etching and deposition yields per ion were measured using quartz crystal microbalance (QCM) as a function of ion bombardment energy, ion-to-neutral flux ratio, and ion-impinging angle. C2HF5, C2F6, C2H4F2, and C4F8 were used for the oxide etching. Oxide etching mechanism with those gases is complex because etching and deposition are involved at the same time. In highly selective processes fluorocarbon deposition plays important role in determining etching characteristics. Two fluorocarbon deposition mechanisms are identified in this work: neutral deposition and ion-enhanced deposition. The low-energy ions are believed to enhance the deposi...

The chemistry and kinetics of alternative etching chemistries for low-k dielectric materials are explored to improve the anisotropy of the etching process and to reduce the problems associated with postetch clean-up. Etching rates, selectivities, and etching yields of Black Diamond and Coral organosilicate glasses (OSGs) have been measured. Black Diamond and Coral are etched rapidly in F2, Cl2, and HBr high density plasmas, and Cl2+HBr plasmas have been identified as a viable process chemistry with several advantages over traditional fluorocarbon plasmas. The OSG films are not spontaneously etched by F2, Cl2, HBr molecules, Cl, or Br atoms, however, F atoms etch the OSGs spontaneously. F, Cl, and H atoms extract a substantial amount of carbon from the films, but Cl and H do not attack the OSG oxide matrix. The Coral films are more strongly depleted of carbon after halogen plasma etching than the Black Diamond. In addition, oxygen atoms extract nearly all of the carbon and nitrogen from the OSGs, leaving a...

Using a simulation-based approach, the energy efficiency of the MIT Lincoln Laboratory xLP process was benchmarked against two commercial low power process technologies for a ∼40,000 gate DES enryption circuit to demonstrate the benefits of subthreshold-optimized transistors for minimum energy circuit design. Because the transistors are tuned for low voltage operation, the xLP process performs best relative to other technologies at subthreshold supply voltages, where most circuits achieve their minimum energy point. With 90nm feature sizes, the xLP process at showed a 57% percent energy efficiency improvement vs. IBM 90nm technology and a 9% energy efficiency improvement vs. IBM 65 nm. Scaling the xLP process to 65 nm should provide further energy efficiency benefit.

Etching yields of silicon in F2, Cl2, Br2, and HBr high density plasmas have been measured as a function of ion bombardment energy, ion bombardment angle, and plasma composition. This information contributes to a database of experimental values needed for feature profile evolution modeling. For all plasma chemistries, the etching yield increases approximately with the square root of ion energy. Pure Cl2 and pure HBr plasmas have very similar etching yields. Silicon etching rates are lower in HBr plasmas than in Cl2 plasmas due to lower ion fluxes, not lower etching yields. The dependence of the etching yield on ion bombardment angle is significantly different for Cl2 and HBr plasmas. The etching yield in Cl2 plasmas decreases rapidly for ion angles above 60° (measured from the surface normal), which results in significant ion scattering from the sidewalls, and may cause the sidewall bowing and microtrenching seen when patterning polysilicon with Cl2 plasmas. The etching yield in HBr plasmas decreases more...

The etching chemistry of benzocyclobutene (BCB) low-k dielectric films was studied in a high density plasma etcher using F2+O2 and Cl2+O2 plasmas. The etching rate in F2+O2 plasmas exceeded 1.2 μm/min with selectivity over oxide and nitride of 16 and 32, respectively. The etching rate in Cl2+O2 plasmas exceeded 0.6 μm/min with selectivity over oxide and nitride of 40 and 80, respectively. BCB films do not etch in pure Cl2 or pure O2 plasmas without ion bombardment, but etching rates of 1000 A/min were observed using F2 plasmas without ion bombardment. The ion flux in F2+O2 plasmas is primarily O2+ and O+, whereas in Cl2+O2 the dominant ion is ClO+. BCB etching yields in F2+O2 plasmas were measured with a plasma beam/quartz crystal microbalance system. The etching yields suggest that the neutral fluxes and surface chemistry control the etching rates under these conditions, not the ion flux. Using x-ray photoelectron spectroscopy, it was determined that oxygen plasmas preferentially remove the carbon conten...

An electron beam generated plasma reactor is used to decompose low concentrations of ethyl chloride (EC) and vinyl chloride (VC) in atmospheric pressure air streams. The energy requirements for 90% decomposition of each compound are reported as a function of inlet concentration. VC decomposition is enhanced by chlorine radical reactions. Both the EC and VC reactions seem to be inhibited by electron scavenging decomposition products. A simple analytic expression is used for fitting decomposition data where inhibition effects are important.

An electron beam generated plasma reactor is used to decompose low concentrations (100–3000 ppm) of 1,1-dichloroethane and 1,1-dichloroethene in atmospheric pressure air streams. The energy requirements for 90% and 99% decomposition of each compound are reported as a function of inlet concentration. Dichloroethene decomposition is enhanced by a chlorine radical propagated chain reaction. The chain length of the dichloroethene reaction is estimated to increase with dichloroethene concentration from 10 at 100 ppm initial dichloroethene concentration to 30 at 3000 ppm. Both the dichloroethane and dichloroethene reactions seem to be inhibited by electron scavenging decomposition products. A simple analytic expression is proposed for fitting decomposition data where inhibition effects are important and simple first order kinetics are not observed.

The effect of a carbon-carbon double bond on the energy required for decomposition in an electron beam-generated plasma reactor is studied by comparing the decomposition of trichloroethylene and 1,1,1-trichloroethane. A reaction mechanism for TCE decomposition based on a chlorine radical chain reaction is presented which accounts for the formation of all of the experimentally observed reaction products. TCE decomposition is autocatalyzed by reaction products, whereas TCA decomposition is inhibited. The rate expression for the decomposition of TCE in the reactor is determined to be r=−[T](15.07[T0]−0.40+0.006{[T0]−[T]}), where [T] and [T0] are both in ppm, and r is in ppm Mrad−1. The energy expense ɛ for TCE decomposition is determined as a function of inlet concentration. For 99% decomposition of 100 ppm TCE in air, ɛ=28 eV/molecule, and ɛ=2.5 eV/molecule at 3000 ppm. This is only 2.5–5% of the amount of energy required to decompose a similar amount of TCA as reported by the authors in a previous study. By comparing the energy requirements for TCE decomposition to those for TCA decomposition, the TCE reaction chain length is determined to increase from approximately 20 at 100 ppm initial TCE concentration, to 40 at 3000 ppm.

Solar cells embedded in the SOI substrate were successfully used as the sole energy source to power a ring oscillator fabricated using an ultra-low-power fully depleted SOI process on the same wafer. The speed of the ring oscillator increased with increasing light intensity and showed a fastest oscillation with a 4.5 ns stage delay and 0.26 fJ power-delay product. The maximum power generated by the solar cell was 9.6 mW/cm2 with an efficiency of 11.6%.

A workfunction-tuned TiN metal gate is integrated into ultra-low-power FDSOI CMOS transistors, optimized for subthreshold operation at 0.3 V. The workfunction of the TiN metal gate is tunable across the mid-gap range, by adjusting deposition parameters and post-deposition annealing. The transistors show 71% reduction in C gd and 55% reduction in V t variation, compared to conventional FDSOI transistors of the same size. A 59% decrease in switching energy and a 91% decrease in stage delay is demonstrated in ring oscillators fabricated with the subthreshold-optimized FDSOI transistors when compared to commercial bulk silicon devices.

Fully depleted silicon-on-insulator transistors were fabricated using two different metal gate deposition mechanisms to compare plasma damage effects on gate oxide quality. Devices fabricated with both plasma-enhanced atomic-layer-deposited (PE-ALD) TiN gates and magnetron plasma sputtered TiN gates showed very good electrostatics and short-channel characteristics. However, the gate oxide quality was markedly better for PE-ALD TiN. A significant reduction in interface state density was inferred from capacitance-voltage measurements as well as a 1200× reduction in gate leakage current. A high-power magnetron plasma source produces a much higher energetic ion and vacuum ultra-violet (VUV) photon flux to the wafer compared to a low-power inductively coupled PE-ALD source. The ion and VUV photons produce defect states in the bulk of the gate oxide as well as at the oxide-silicon interface, causing higher leakage and potential reliability degradation.

An electron beam generated plasma reactor was used for the decomposition of carbon tetrachloride, 1,1,1-trichloroethane, trichloroethylene, 1,1-dichloroethane, toluene, and Freon 113. 99% decomposition of most compounds is achieved in the reactor. The energy requirements for decomposition of these compounds are given, and these results are discussed in terms of the electron capture rate coefficients and the plasma chemistry of the molecules.

Etching of indium tin oxide (ITO) thin films in high density chlorine plasmas is studied, with the goal of increasing the etching selectivity to photoresist. The ITO etching rate increases with ethylene addition, but is not affected by BCl3 addition. ITO exhibits a threshold energy for ion etching, whereas the photoresist etches spontaneously in chlorine plasmas. The ITO:photoresist selectivity increases with BCl3 addition, ion bombardment energy, and C2H4 addition. It is proposed that the ITO etching rate is limited by desorption of InClx products, and that ethylene addition assists in scavenging oxygen from ITO leaving loosely bound In, which is more easily removed by physical sputtering.

An electron beam generated plasma reactor (EBGPR) is used to determine the plasma chemistry kinetics, energetics and decomposition pathways of six chlorinated ethylenes and ethanes: 1,1,1-trichloroethane, 1,1-dichloroethane, ethyl chloride, trichloroethylene, 1,1-dichloroethylene, and vinyl chloride. A traditional chemical kinetic and chemical engineering analysis of the data from the EBGPR is performed, and the following hypothesis was verified: The specific energy required for chlorinated VOC decomposition in the electron beam generated plasma reactor is determined by the electron attachment coefficient of the VOC and the susceptibility of the molecule to radical attack. The technology was demonstrated at the Hanford Reservation to remove VOCs from soils.

Atomic layer deposition (ALD) of high-quality gadolinium oxide thin films is achieved using Gd(iPrCp)3 and O2 plasma. Gd2O3 growth is observed from 150 to 350 °C, though the optical properties of the film improve at higher temperature. True layer-by-layer ALD growth of Gd2O3 occurred in a relatively narrow window of temperature and precursor dose. A saturated growth rate of 1.4 A/cycle was observed at 250 °C. As the temperature increases, high-quality films are deposited, but the growth mechanism appears to become CVD-like, indicating the onset of precursor decomposition. At 250 °C, the refractive index of the film is stable at ∼1.80 regardless of other deposition conditions, and the measured dispersion characteristics are comparable to those of bulk Gd2O3. XPS data show that the O/Gd ratio is oxygen deficient at 1.3, and that it is also very hygroscopic. The plasma etching rate of the ALD Gd2O3 film in a high-density helicon reactor is very low. Little difference is observed in etching rate between Cl2 a...

Ultralow-power electronics will expand the technological capability of handheld and wireless devices by dramatically improving battery life and portability. In addition to innovative low-power design techniques, a complementary process technology is required to enable the highest performance devices possible while maintaining extremely low power consumption. Transistors optimized for subthreshold operation at 0.3 V may achieve a 97% reduction in switching energy compared to conventional transistors. The process technology described in this article takes advantage of the capacitance and performance benefits of thin-body silicon-on-insulator devices, combined with a workfunction engineered mid-gap metal gate.

Etching of TiN metal gate materials as a part of an integrated flow to fabricate fully depleted silicon-on-insulator ultralow-power transistors is reported. TiN etching is characterized as a function of source power, bias power, gas composition, and substrate temperature in a high density inductively coupled plasma reactor. Under the conditions used in this work, the TiN etch rate appears to be ion flux limited and exhibits a low ion enhanced etching activation energy of 0.033eV. Notching of the polysilicon layer above the TiN may occur during the polysilicon overetch step as well as the TiN overetch step. Notching is not significantly affected by charging of the underlying gate dielectric under the conditions used. By optimizing the plasma etch process conditions, TiN:SiO2 selectivity of nearly 1000:1 is achieved, and a two-step TiN main etch and TiN overetch process yields well-defined metal gate structures without severe gate profile artifacts.