A novel atomic layer deposition (ALD) process was developed for low-resistivity molybdenum (Mo) from molybdenum dichloride dioxide (MoCl2O2) and atomic hydrogen (at-H). A wide ALD window of self-limiting growth was observed between 150 and 450 °C. No film deposition occurred with molecular hydrogen (H2), demonstrating the necessity to have at-H to efficiently reduce the MoCl2O2 precursor. At 350 °C and above, the film composition was determined at approximately 95 at. % of Mo and 3.5 at % of oxygen (O), with trace amounts (i.e., <1 at. %) of carbon (C), chlorine (Cl), hydrogen (H), and nitrogen (N). The growth per cycle (GPC) was roughly 0.022 nm/cycle. No substrate selectivity or pronounced nucleation delay was observed on silicon (Si), silicon dioxide (SiO2), silicon nitride (Si3N4), silicon carbide (SiC), aluminum oxide (Al2O3), hafnium dioxide (HfO2), and low-k dielectric (SiOC). Film uniformity and conformality were ±5% and ±10%, respectively, while resistivity approached a bulk value of 18.6 μ Ω cm at 24 nm. At 250 °C and below, increased levels of oxygen (up to 33 at. % at 150 °C) and chlorine (2.7 at. % at 150 °C) were detected in the film. This trend coincided with an increase in the GPC, a change in optical properties, a decrease in film density and crystallinity, and an increase in resistivity. While self-limiting growth was observed through the entire ALD window of 150–450 °C, the temperature (T) range for depositing low-resistivity Mo deposition was narrower at T ≥ 250 °C.
Hot‐wire assisted atomic layer deposition (HWALD) is a novel energy‐enhancement technique. HWALD enables formation of reactive species (radicals) at low substrate temperatures, without the generation of energetic ions and UV photons as by plasma. This approach employs a hot wire (tungsten filament) that is heated up to a temperature in the range of 1300–2000 °C to dissociate precursor molecules. HWALD has the potential to overcome certain limitations of plasma‐assisted processes. This work investigates the ability of a heated tungsten filament to catalytically crack molecular hydrogen or ammonia into atomic hydrogen and nitrogen‐containing radicals. The generation of these radicals and their successful delivery to the wafer (substrate) surface are experimentally confirmed by dedicated tellurium‐etching and silicon‐nitridation experiments. It further reports on deposition of low‐resistivity oxygen‐free tungsten films by using HWALD, as well as on the effect of hot‐wire‐generated nitrogen radicals and atomic hydrogen in deposition of aluminum nitride and boron nitride films. In parallel, this work provides important illustrative examples of using in situ real‐time monitoring of deposition and etching processes, together with extracting a variety of film properties, by spectroscopic ellipsometry technique.
Transmission line method (TLM) structures are often employed to extract contact resistivity between a metal and a doped semiconductor region. In this article we treat the situation where the doped region is junction-isolated from the substrate. The junction isolation may be leaky resulting in erroneous parameter extraction. The effect of junction leakage is treated both theoretically and through exemplary wafer-level CTLM measurement results on photovoltaic cells (solar cells) and epi-wafer samples. This paper describes how reliable contact resistivity values can be obtained using the transmission line method on junction isolated structures.
The cover depicts a reactor utilized for hot-wire assisted atomic layer deposition (Hot Wire ALD), enabling formation of reactive species (radicals) at low substrate temperatures, without the use of plasma. A tungsten filament heated up to1300–2000 °C is installed on the side. The reactor is equipped with an in situ spectroscopic ellipsometer, as reported by Alexey Y. Kovalgin and co-workers in article number 1700058.
A heated tungsten filament (wire) is well known to generate atomic hydrogen (at-H) by catalytically cracking molecular hydrogen (H-2) upon contact. This mechanism is employed in our work on hot-wire (HW) assisted atomic layer deposition (HWALD), a novel energy-enhancement technique. HWALD has been successfully utilized to deposit tungsten (W) films using alternating pulses of WF6 and at-H. Depending on the conditions, either low-resistivity alpha- or higher-resistivity beta-crystalline phases of W can be obtained. This work aims to clarify (i) which factors are decisive for the formed crystal phase and (ii) the role of the residual gases in the film growth mechanism. In this light, the effects of adding impurities (N2O, O-2, NH3 and H2O) were investigated. Oxidizing species have a retarding effect on W growth but the process can be re-initiated after stopping their supply. In contrast, nitridizing species have a permanent inhibition effect. Further, the effects of WF6 overdose were studied. The surplus of WF6 appeared to be crucial for the process: in many cases this led to the formation of beta-phase Winstead of the a-phase, with a memory effect lasting for several deposition runs. Extra fluorine-containing species were thus identified as the likely cause of beta-phase formation. (c) 2017 The Electrochemical Society. All rights reserved.
In this work, we investigated an approach of hot-wire assisted ALD (HWALD), utilizing a hot (up to 2000 degrees C) tungsten (W) wire. Tungsten films were deposited by this method using alternating pulses of WF6 gas and atomic hydrogen (at-H). The latter was generated by catalytic dissociation of molecular hydrogen (H-2) upon the hot-wire. The W films were grown on a 100-nm thick thermal SiO2. The growth process was monitored in real time by an in-situ spectroscopic ellipsometer (SE). The real-time SE monitoring revealed the coexistence of three processes: CVD, etching, and ALD of the W film. WF6 could back-stream diffuse to the hot-wire, resulting in WF6 decomposition and generation of a flux of fluorine (F). The latter caused etching of the grown W film and the filament, and provided extra tungsten supply, which might cause CVD. Higher pressure and higher carrier gas flow rate were found to largely suppress the back-stream diffusion of WF6, which efficiently limited CVD. By controlling the dose of WF6 and process pressure, the etching had also been minimized. X-ray photoelectron spectroscopy of optimized HWALD grown W revealed 99 at% of W; concentrations of oxygen and fluorine were lower than 1%, below the detection limit. (C) 2015 WILEY - VCH Verlag GmbH & Co. KGaA, Weinheim
In this work, the authors developed hot-wire assisted atomic layer deposition (HWALD) to deposit tungsten (W) with a tungsten filament heated up to 1700–2000 °C. Atomic hydrogen (at-H) was generated by dissociation of molecular hydrogen (H2), which reacted with WF6 at the substrate to deposit W. The growth behavior was monitored in real time by an in situ spectroscopic ellipsometer. In this work, the authors compare samples with tungsten grown by either HWALD or chemical vapor deposition (CVD) in terms of growth kinetics and properties. For CVD, the samples were made in a mixture of WF6 and molecular or atomic hydrogen. Resistivity of the WF6-H2 CVD layers was 20 μΩ·cm, whereas for the WF6-at-H-CVD layers, it was 28 μΩ·cm. Interestingly, the resistivity was as high as 100 μΩ·cm for the HWALD films, although the tungsten films were 99% pure according to x-ray photoelectron spectroscopy. X-ray diffraction reveals that the HWALD W was crystallized as β-W, whereas both CVD films were in the α-W phase.
This work discusses the design, finite element method modeling (FEM), fabrication and characterization of a silicon-based, catalytic micro calorimetric sensor. The sensing area is comprised of two titanium silicide (TiSi2) – polysilicon (poly-Si) resistive temperature sensors symmetrically positioned relative to a poly-Si heater on which an oxidation promoting catalyst is deposited. The resistive structures are located on a suspended, thereby, thermally isolating, low mechanical stress membrane and integrated into a glass flow channel. The micro-calorimetric sensor is applied for measuring propane and hydrogen concentrations in air.
This paper describes the potential of tunable strain in field-effect transistors to boost performance of digital logic.Voltage-controlled strain can be imposed on a semiconductor body by the integration of a piezoelectric material improving transistor performance.In this paper, we derive the relations governing the subthreshold swing in such devices to improve the understanding.Using these relations and considering the mechanical and technological boundary conditions, we discuss possible device architectures that employ this principle.Further, we review the recently published experimental and modeling results of this device, and give analytical estimates of the power consumption.
Earlier theoretical reports predicted that the usage of a piezoelectric stressor layer around the FinFET, i.e., the PiezoFET, offers a great potential for steep subthreshold slope devices. For the first time, we analyzed the practical realization of such PiezoFETs comprising a piezoelectric stressor layer, lead-zirconate-titanate (PZT), and aluminum-nitride (AlN) deposited on n-type silicon FinFETs. A high-piezoelectric response in the range of 100 pm/V has been obtained for the PZT PiezoFET evidencing the converse piezoelectric effect in the device. The piezoelectric response for the AlN device was much less (13 pm/V) as expected. Underlying device properties, such as subthreshold swing (SS) and low-field electron mobility have been significantly affected by the presence of the PZT stressor. A 20%-50% change in the mobility and a change in the SS (about 5 mV/decade) have been observed. The change can be attributed to the strain induced reduction of the interface trap density at the Si/SiO2 interface. This strain is partly formed by the bias over the piezoelectric layer, which indicates the converse piezoelectric effect related tunable strain in both the silicon channel and gate oxide.
We realized metal-ferroelectric-metal (MFM) capacitors comprising high-quality ferroelectric lead zirconate titanate (Pb(Zr0.52Ti0.48)O-3 or PZT) thin films on an LaNiO3/poly-Si/titanium nitride (TiN)/SiO2 integrated on a 100 mm Si wafer. Promising effective piezoelectric coefficient and remnant polarization of 53 pm/V and 19.2 mu C/cm(2), respectively, are obtained for the 100 nm-PZT/20 nm-LNO stack. Further analysis of the samples indicates the presence of a passive layer, possibly near the Ti/PZT interface at the top electrode. A leakage current model has been used to explain the obtained current density electric field curves. In this model, diffusion limited transport has been assumed in which the injection is interface-controlled. Based on the capacitance and the leakage current measurements, the thickness and dielectric constant values of the passive layer are estimated to be 2.1 nm and 23, respectively. The observed apparent low barrier height value of 032 eV is attributed to ferroelectric polarization related phenomena. A good agreement between measurement and leakage current model is obtained. (C) 2014 Elsevier B.V. All rights reserved.
Metal induced strain in the channel region of silicon (Si) fin-field effect transistor (FinFET) devices has been characterized using Raman spectroscopy. The strain originates from the difference in thermal expansion coefficient of Si and titanium-nitride. The Raman map of the device region is used to determine strain in the channel after preparing the device with the focused ion beam milling. Using the Raman peak shift relative to that of relaxed Si, compressive strain values up to – 0.88% have been obtained for a 5 nm wide silicon fin. The strain is found to increase with reducing fin width though it scales less than previously reported results from holographic interferometry. In addition, finite-element method (FEM) simulations have been utilized to analyze the amount of strain generated after thermal processing. It is shown that obtained FEM simulated strain values are in good agreement with the calculated strain values obtained from Raman spectroscopy.
Using low pressure atomic layer deposition, ultra-thin continuous TiN films were prepared. The temperature coefficient of resistance (TCR), resistivity and field effect properties of these films were investigated. With decreasing film thickness, a positive-to-negative transition of TCR and a steep increase of resistivity were observed. This is attributed to the metal-semimetal transition of the TiN films. We demonstrate appreciable field-induced current modulation up to 11% in a 0.65 nm TiN film. The field effect is remarkably independent of temperature. A polarity asymmetry of the current-voltage characteristics is found, attributed to the interplay between different types of the carriers.
We present the generation of atomic hydrogen made by the dissociation of molecular hydrogen upon collision with a tungsten (W) filament kept at a high temperature (T approximate to 1600-1900 degrees C). We demonstrate the ability to create atomic hydrogen and to introduce it in short pulses in experiments on etching of tellurium (Te) films. We further utilize the generated atomic hydrogen (H) to explore its impact on surface reactions in the TiCl4/NH3 precursor system. Atomic hydrogen is introduced in pulses additionally to TiCl4 and NH3 with different pulse sequences. For the TiCl4/NH3/H sequence, there is no influence on the process compared to the ALD without H-pulses. The growth rate remains at 0.02 nm/cycle and the oxygen (residual gas) content - at 3-5 at%. For the TiCl4/H/NH3 pulse sequence, the growth rate decreases to 0.01 nm/cycle and the oxygen content increases to 30-35 at%. Only TiCl4/H pulses result in no growth after the formation of approximately one monolayer. Similar effect occurs after introducing NH3 via the hot filament, pointing to the decomposition of NH3 and the formation of atomic hydrogen. (C) 2013 The Electrochemical Society. All rights reserved.
This work reports on the determination and comparison of the resistivity of ultra-thin atomic layer deposited titanium nitride films in the thickness range 0.65–20nm using spectroscopic ellipsometry and electrical test structures. We found that for films thicker than 4nm, the resistivity values obtained by the two techniques are in good agreement. However, below 4nm, the comparison shows an increasing difference with decreasing film thickness. A difference with a factor of 3 was found at 1.8nm and increased up to hundreds at 0.65nm. We attribute this significant difference to the electron scattering effects at grain boundaries and interfaces which can not be fully taken into account by spectroscopic ellipsometry measurements.
Test structures for the electrical characterization of ultrathin conductive (ALD) films are presented based on buried electrodes on which the ultrathin film is deposited. This work includes test structure design and fabrication, and the electrical characterization of ALD TiN films down to 4 nm. It is shown that these structures can be used successfully to characterize sub 10 nm films.
We present a novel method for determining the temperature budget of the process side of silicon substrates and chips, based on well-known silicide formation reactions of metal-Si systems and (four-point probe) resistance measurements. In this paper, we focus on the Pd-Si system that is most temperature sensitive in the range from 100°C to 200°C. A variety of test structures is introduced to exploit the specific properties of the diffusion-limited reaction between Pd and Si. Among others, this resulted in gap-based layouts that facilitate an extension of the temperature range to 350°C. Designs and measurement results are presented, indicating the practicality and the robustness of the proposed technique.