Silicon Nitride Layer Research Articles

Improvement of SiO2 surface morphology during the selective Si3N4 etching in the multi-layered 3D NAND Si3N4/SiO2 stack structures by the generation of CO2 gas through the control of redox reaction

Selective removal of silicon nitride (Si3N4) layers from multilayer Si3N4/silicon dioxide (SiO2) stack structures is essential for fabricating 3-dimensional (3D) NAND flash memory devices. Although phosphoric acid (H3PO4) selectively etches Si3N4 layers to a certain degree, the tip thicknesses of the SiO2 layers increase owing to oxide regrowth, and the SiO2-layered trenches become clogged. In particular, the oxide regrowth phenomenon worsens as it progresses to the bottom of the structure. In this study, carbon dioxide (CO2) gas was generated by adding acidic reducing agents to H3PO4 to suppress the oxide regrowth in the stack structure. CO2 gas creates a turbulent flow in the slits and SiO2/Si3N4/SiO2 trenches, promoting the diffusion of Si3N4 etch byproducts from the cul-de-sac of the SiO2/Si3N4/SiO2 trench structures to the outside of the stack structure. The addition of oxalic acid to H3PO4 generated a high concentration of CO2 gas and selectively etched Si3N4 layers without oxide regrowth in the stack structure. Finally, the selective etching of Si3N4 layers, which is faster and more uniform than that in pure 85 wt.% H3PO4, was performed without oxide regrowth in the entire 128-layer Si3N4/SiO2 multi-stack structure.

Molecular dynamics simulation of oxide-nitride bilayer etching with energetic fluorocarbon ions

In 3D NAND devices, the number of stacked layers dictates the storage capacity. In the fabrication of such devices, hole channels with a high-aspect ratio (HAR) are etched through these layers. The higher the aspect ratio becomes, the more difficulties the HAR etching faces. In this study, molecular dynamics simulation was performed to examine the etching of silicon dioxide (SiO2), silicon nitride (SiN), and oxide-nitride (ON) stacked layers by energetic fluorocarbon ions. Good agreement of etching yields obtained from the simulations and ion beam experiments was observed for the etching of SiO2 and SiN by CF3+ ions for the incident ion energy ranging from 200 to 2000 eV. As to the etching of the ON bilayer, the SiO2 and SiN layers were observed to be etched with their own etch rates at low ion incident energy. However, at sufficiently high incident ion energy, the oxide and nitride layers were mixed by energetic ion impact and the depth of the mixing layer exceeded the thickness of the top SiO2 layer thickness, resulting in a single etch rate limited by the etch rate of a much deeper underlying SiN material. This suggests that if the incident ion energy is high enough such that the thicknesses of the multilayers are lower than the ion penetration depth, the ON stacked layer exhibits a single etch rate determined by the mixed material of the oxide and nitride.

Fabrication of magnetic helical microribbons made of nickel thin films sandwiched between silicon nitride layers for microswimming applications

Helical swimming is adopted by microswimming robots since it is an efficient mechanism and commonly observed among microorganisms swimming at low Reynolds numbers. However, manufacturing of micro-helices made of sub-micron magnetic thin layers is neither straightforward nor well-established, advanced materials and methods are necessary to obtain such structures as reported in the literature. In this paper, a topological patterning method utilizing basic microfabrication methods is presented for the self-assembly of magnetic micro-helices made of a sandwiched nickel thin film (50–150 nm) between two silicon nitride layers. Strain mismatch between the thin films and the geometric anisotropy introduced by the slanted patterns on the top nitride layer result in self-rolled-up helical microribbons. Moreover, inspired by the actual release process during the wet-etching of the microribbon from the substrate, moving boundary conditions are incorporated in a numerical model to simulate the self-rolling of trilayer ribbons. The simulation results are compared and validated by experimental data within 7% error for all cases, including the geometries that do not result in a helical shape. The swimming performance of the magnetized micro-helix is demonstrated inside a capillary glass tube experimentally and cross-validated with a numerical model.

Development and Evaluation of Copper Based Transparent Heat Reflectors Obtained by Magnetron Sputtering.

Within the next few years climate change is likely to become a major concern for mankind. In addition, the current electronic components shortage crisis has led to an urgent need for alternative solutions in the main industry sectors (the raw materials, manufacturing, and construction industries). The current trends of research are focused on developing smart materials with functional properties, using abundant raw materials. The energy saving efforts are sustained in the glazing industries by several approaches based on dielectric-metal-dielectric multilayer structures. The use of silver to achieve a high reflectivity in near-infrared spectral range has been proposed and is already adopted as a commercially available solution. This work is focused on developing a transparent heat reflector (THR) with prefigured optical properties, using copper as a reflective layer, a material that is more abundant and cheaper than silver. The conductive copper layers obtained by the High Power Impulse Magnetron Sputtering (HiPIMS) method were interposed between two silicon nitride layers deposited by the Radio-Frequency Magnetron Sputtering (RFMS) technique. The structural, optical, and elemental composition of monolayers was investigated, qualifying each individual material for use in the multilayer structure. The time stability of films deposited on microscope glass substrates was also investigated, as an important criterion for the selection of monolayers. The obtained results revealed that the SiNx/Cu/SiNx with the Cu layer deposited by using a negative substrate bias of −100 V showed the most stable behavior over time. Optical modeling was performed to design a THR multilayer structure, which was successfully obtained experimentally. A maximum optical transparency as high as 75% in the visible range and a reflectivity of ~ 85% in near infrared spectral interval was confirmed for the experimentally obtained multilayer structures.

Passivation of poly-Si surface using vinyl and epoxy group additives for selective Si3N4 etching in H3PO4 solution

This work investigates the mechanism of polycrystalline silicon (poly-Si) dissolution in H3PO4. The etching rate of poly-Si was more rapid than that of silicon dioxide (SiO2) in H3PO4 solution, because poly-Si surface possesses unstable Si1+, Si2+, and Si3+ states – rather than Si4+, which is more stable. Additives such as epichlorohydrin, 1,2-epoxy-3-phenoxypropane, vinyl bromide, and phenyl vinyl ether with epoxy or vinyl groups were introduced into the H3PO4 to passivate the poly-Si surface. This significantly increased the presence of the Si4+ state and provided hydrophobic functional groups on the poly-Si surface to reduce the wettability of the poly-Si surface. The introduction of these additives significantly reduced the etching rate of poly-Si. 3-glycidyloxypropyl trimethoxysilane, 3-glycidyloxypropyl dimethoxymethylsilane, allyltrimethoxysilane, and vinyltrimethoxysilane with Si–O bonds, as well as epoxy or vinyl groups in their molecular structures, were demonstrated to suppress the etching rate of both SiO2 and poly-Si. The introduction of these additives to H3PO4 promoted the selective removal of silicon nitride (Si3N4) layers from multi-layered Si3N4/SiO2 trench structures without the loss of either poly-Si or SiO2.

Impact of dielectric film thickness on field emission in MEMS capacitive switches

This study presents experimental evidence of field emission in MEMS capacitive switches. Devices with dielectric layers of silicon nitride of different thicknesses between 50 and 200 nm were investigated by current-voltage (I-V) measurements. These measurements were performed at room temperature and under a controlled atmosphere pressure of 3 × 10−2 mbar at bias levels below breakdown and corresponding electric fields encountered in MEMS capacitive switches during pull-in (1-2 × 106 V/cm). Field emission although was not always clearly observed, it occurred in all devices and clearly manifested at electric fields larger than 106 V/cm.

Adaptive PAM-4/PAM-8 graphene-based electro-optical modulator integrated into a polymer waveguide platform for data-center communication

This paper presents a new concept of a graphene-based adaptive modulator for multilevel amplitude modulation on polymer technology platform. This component consists of an electro-optical modulator with a distributed five segments capacitive structure, based on two single-layers of graphene (SLG) and a thin layer of silicon nitride (Si3N4) in the form (SLG-Si3N4-SLG), which shows a dynamic configuration of the modulation scheme, PAM-4/PAM-8, by the applied logic voltage sequence (3Vpp) for an operating wavelength at 1550 nm. Considering an extinction ratio of 5.6 dB for PAM-4, and 6.0 dB for PAM-8, a compact device with 179.6 µm total length, and the maximum simulated cut-off frequency of 48.12 GHz with 122.9 fF of capacitance and total device resistance of 26.9 Ω, were achieved. This cut-off frequency results in a transmission rate of 96.24 Gbps at PAM-4 and 144.4 Gb/s at PAM-8. The simulated insertion loss is 12.0 dB (PAM-4) and 11.5 dB for PAM-8, and the power consumption is 0.54 pJ/bit for PAM-4 and 0.81 pJ/bit for PAM-8. Combining an innovative design with the unique electro-optical characteristics widely demonstrated by the materials used in the construction of the proposed adaptive device, allows it to be seen as a candidate to contribute to the nanotechnology integration and high-performance communication requirements.

Photonic crystal backbone for light trapping inside an ultrathin, low absorbing layer.

A few tens of nanometre thick ultrathin materials may suffer from a very low absorption at their band edges. In this work, we investigate a photonic crystal (PC) made of a lowcost, transparent patterned silicon nitride (SiNx) layer, conformally covered with an ultrathin active layer (e.g., 20 nm TiO2) in view of its use in various applications such as photocatalysis. A fair estimation of the absorption enhancement, considering the volume of the active material, is calculated using RCWA. A remarkable enhancement (more than ten-folds) in absorptance in the near UV range and a very high transmittance over the visible range are observed. A detailed modal analysis of the structures-of-interest unravels the Light Trapping (LT) mechanisms and allows the derivation of key design guidelines. Optical measurements on a patterned sample provide a first proof-of-concept of such possible photonic backbone structures suitable for highly efficient depollution and artificial photosynthesis for solar fuels production.

All-Dielectric Transreflective Angle-Insensitive Near-Infrared (NIR) Filter.

This paper presents an all-dielectric, cascaded, multilayered, thin-film filter, allowing near-infrared filtration for spectral imaging applications. The proposed design is comprised of only eight layers of amorphous silicon (A-Si) and silicon nitride (SiN), successively deposited on a glass substrate. The finite difference time domain (FDTD) simulation results demonstrate a distinct peak in the near-infrared (NIR) region with transmission efficiency up to 70% and a full-width-at-half-maximum (FWHM) of 77 nm. The theoretical results are angle-insensitive up to and show polarization insensitivity in the transverse magnetic (TM) and transverse electric (TE) modes. The theoretical response, obtained with the help of spectroscopic ellipsometry (SE), is in good agreement with the experimental result. Likewise, the experimental results for polarization insensitivity and angle invariance of the thin films are in unison with the theoretical results, having an angle invariance up to .

Omnidirectional, thin metasurface exhibiting selective absorption for un-polarized broadband incidence.

Thin devices with large areas have strong and omnidirectional absorption over a wide bandwidth and are in demand for applications such as energy harvesting, structural color, and vehicle LiDAR (laser detection and ranging). Despite persistent efforts in the design and fabrication of such devices, the simultaneous realization of all these desired properties remains a challenge. In this study, a 190-nm-thick metasurface with an area of 3 cm2, incorporating dielectric cylinder arrays, a chromium layer, a silicon nitride (SiNx) layer, and an aluminum layer is theoretically and experimentally demonstrated. The developed device achieves an average absorptivity of ∼99% (97% in the experiment) in the entire visible spectrum 400-700 nm. Moreover, it exhibits strong absorption over a wide range of incident angles (∼91% and 90% at 60° in the calculation and experiment, respectively). Importantly, the feasibility of applying the developed metasurface absorber to solar thermophotovoltaics and vehicle LiDAR (laser detection and ranging) has been explored. Moreover, the photoresist can be replaced by other glues and easily scaled up to a large area using the roll-to-roll nanoimprinting process. With the excellent spectral properties and performance, this device is promising for large-area applications.

Raised breakdown voltage in a high‐voltage recessed LDD‐SOI by submerged SI3N4

AbstractA new structure of metal‐oxide‐semiconductor field‐effect transistors (MOSFETs) with lateral double‐diffused MOSFET was introduced based on the silicon‐on‐insulator lightly doped‐drain metal‐oxide‐semiconductor field‐effect‐transistor (SOI‐LDD MOSFET) technology, called recessed silicon nitride lateral‐double diffused silicon‐on‐insulator (RS‐LDD SOI). In the proposed structure, a recession was created on the buried oxide layer and filled with a silicon nitride (SI3N4) layer. Silicon nitride makes the electric field distribution more uniform by modulating the electric field in the drift zone, thereby increasing the breakdown voltage of the proposed structure. The simulation results performed with ATLAS software demonstrated the performance improvement of the proposed structure. Increasing the breakdown voltage by 31% compared to the conventional structure was one of the most important achievements of this work. Also, design considerations were made on the submerged part in the buried oxide, which can be an excellent guide to achieving the optimal performance of the proposed structure.

Design and Fabrication of Multiple-Color-Generating Thin-Film Optical Filters for Photovoltaic Applications

The use of electric vehicles (EVs) can reduce greenhouse gas emissions, air pollution, dependency on fossil fuels, and their adverse health effects on humans. But, we can only utilize the full environmental benefits of EVs when they are charged with renewable energy sources with zero or low carbon emissions. As a solution, Mobarak et al. [1] suggested integrating low-cost, flexible, and thin-film copper indium gallium selenide (CIGS) solar cells directly onto the steel of all the upward-facing body parts of the vehicles. But, this integration of solar cells comes with an aesthetic drawback. Previously, colorful photovoltaics (PVs) have been designed with one-dimensional (1D) photonic crystals or various 1D and 2D metallic nanostructures for aesthetic building-integrated photovoltaics (BIPVs) [2, 3]. However, the functionality of our application differs from that of BIPV as we need maximum absorption of the solar spectrum to obtain maximum conversion efficiency. Thus, we propose replacing the anti-reflective coating (ARC) present in the solar cells with a notch filter (a narrow high-reflection region in the visible range along with high transmission for the rest of the solar spectrum) to obtain colors.High-performance notch filters with a narrow and ultra-steep notch are well known in literature [4, 5]. Generally, high-performance notch filters are designed with a minimum of 45 layers. It is challenging to use filters with many layers on solar cells due to fabrication and thickness complexities. Thus, we created designs with a maximum of 27 layers for possible integration with photovoltaics. We used OptiLayer [6] to simulate our designs and the gradual evolution technique was used to optimize the designs. We performed our simulations with a multilayer structure of alternating high and low refractive indices of 2.09 and 1.45, respectively, on top of a silicon substrate. We optimized this multilayer structure for three reference wavelengths (400 nm, 550 nm, and 700 nm) resembling three colors. Our designs have notch widths of less than 100 nm for all the reference wavelengths with an average of 70% reflection in the high-reflection region and less than 20% reflection in the high-transmission area.To fabricate our designs, we need materials that are transparent to the solar spectrum targeted by the active material of the solar cells. The materials also need to have refractive indices closer to our simulation. Thus, we chose the combination of silicon nitride and silicon dioxide as our high and low refractive index material, respectively [7, 8]. To better understand our designs’ optical characteristics, we fabricated a scaled-down version of our structure with 5-10 layers. We used electron cyclotron resonance plasma-enhanced chemical vapor deposition (ECR-PECVD) to deposit the multilayer structure on silicon wafers. To obtain the silicon nitride and silicon dioxide layers, we used a SiH4/N2/O2/Ar precursor mixture. By tuning the gas flow rate in the reactor chamber, we tuned the stoichiometry and obtained the required refractive index for each layer. To characterize the refractive index and thickness for each layer, we used variable angle spectroscopic ellipsometry (VASE). We made a detailed comparison of our simulation and fabrication results.

(Digital Presentation)Room Temperature Sensor Using Dielectrophoretic Trapping of Carbon Nanotubes

This abstract presents the design and fabrication of a temperature sensor based on carbon nanotubes (CNTs). The sensor is fabricated by dielectrophoretic trapping of CNTs between two nanoelectrodes. The conductance of the sensor is highly sensitive to the temperature variation which makes it suitable for applications requiring accurate measurements. In addition, fast response and good stability and durability are the important advantages of the sensors made of CNTs.The structure of the device is shown in the figure. CNTs are trapped in between nanoelectrodes using the dielectrophoresis (DEP) technique. Dielectrophoresis is a phenomenon in which a force is applied to a dielectric particle when it is exposed to a non-uniform electric field [1]. Although dielectrophoresis provides an accurate method for trapping materials in a predefined area, this method suffers from a costly fabrication process when dealing with nanomaterials. In this study, we used a standard photolithography method to fabricate low-cost devices and an electronic circuit is used to apply the required electric field and monitor the gap for the presence of CNTs.The device fabrication process starts with the patterning and depositing gold nanoelectrodes and contacts using the photolithography method. The process is followed by converting the surface of the device except for the nanoelectrodes area and connection pads with an insulation layer of Silicon nitride. Since the electrodes are thin (50 nm) and CNTs are flexible, fabrication on the flexible substrates is also feasible.

Surface Passivation of III-V GaAs Nanopillars by Low-Frequency Plasma Deposition of Silicon Nitride for Active Nanophotonic Devices.

Numerous efforts have been devoted to improve the electronic and optical properties of III-V compound materials via reduction of their nonradiative states, aiming at highly efficient III-V sub-micrometer active devices and circuits. Despite many advances, the poor reproducibility and short-term passivation effect of chemical treatments, such as sulfidation and nitridation, requires the use of protective encapsulation methods, not only to protect the surface, but also to provide electrical isolation for device manufacturing. There is still a controversial debate on which combination of chemical treatment and capping dielectric layer can best reproducibly protect the crystal surface of III-V materials while being compatible with readily available semiconductor-foundry plasma deposition methods. This work reports on a systematic experimental study on the role of sulfide ammonium chemical treatment followed by dielectric coating (either silicon oxide or nitride) in the passivation effect of GaAs/AlGaAs nanopillars. Our results conclusively show that, under ambient conditions, the best surface passivation is achieved using ammonium sulfide followed by encapsulation with a thin layer of silicon nitride by low-frequency plasma-enhanced chemical deposition. Here, the sulfurized GaAs surfaces, high level of hydrogen ions, and low-frequency (380 kHz) excitation plasma that enable intense bombardment of hydrogen, all seem to provide a combined active role in the passivation mechanism of the pillars by reducing the surface states. As a result, we observe up to a 29-fold increase of the photoluminescence (PL) integrated intensity for the best samples as compared to untreated nanopillars. X-ray photoelectron spectroscopy analysis confirms the best treatments show remarkable removal of gallium and arsenic native oxides. Time-resolved micro-PL measurements display nanosecond lifetimes resulting in a record-low surface recombination velocity of ∼1.1 × 104 cm s-1 for dry-etched GaAs nanopillars. We achieve robust, stable, and long-term passivated nanopillar surfaces, which creates expectations for remarkable high internal quantum efficiency (IQE > 0.5) in nanoscale light-emitting diodes. The enhanced performance paves the way to many other nanostructures and devices such as miniature resonators, lasers, photodetectors, and solar cells, opening remarkable prospects for GaAs active nanophotonic devices.

(Digital Presentation) Effect of N2-Anneal Temperature on Silicon Nitride Film: (II) Fine Structures of ESR Spectrum and FTIR

We performed fine analyses on ESR spectra of silicon nitride (SiN) films grown on a Si (0 0 1) substrate via low-pressure chemical vapor deposition and N2-annealed at temperatures ranging from 700°C to 1150°C. We observed two types of temperature dependency for the number of Si dangling bonds (DBs): in one case, a maximum occurred at 900°C, and in the other case, the number of DBs increased with increasing temperature over 1000°C. The former type of DBs was associated with defects of the Pb0 center at the topmost substrate, (three silicon back-bonds present), and the M signal (generated by Si suboxide with one or two oxygen back-bonds). The latter DB group was attributed to the E' center in the natural oxide layer at the interface between the grown SiN layer and the substrate and the K center in the SiN film. Charge breakdown (Qbd) of the MOS capacitor is strongly related to both DB groups and may have started at the defect associated with the Si suboxide state.

Omnidirectional and broadband photon harvesting in self-organized Ge columnar nanovoids

Highly porous Germanium surfaces with uniformly distributed columnar nanovoid structures are fabricated over a large area (wafer scale) by large fluence Sn+ irradiation through a thin silicon nitride layer. The latter represents a one-step highly reproducible approach with no material loss to strongly increase photon harvesting into a semiconductor active layer by exploiting the moth-eye antireflection effect. The ion implantation through the nitride cap layer allows fabricating porous nanostructures with high aspect ratio, which can be tailored by varying ion fluence. By comparing the reflectivity of nanoporous Ge films with a flat reference we demonstrate a strong and omnidirectional reduction in the optical reflectivity by a factor of 96% in the selected spectral regions around 960 nm and by a factor of 67.1% averaged over the broad spectral range from 350 to 1800 nm. Such highly anti-reflective nanostructured Ge films prepared over large-areas with a self-organized maskless approach have the potential to impact real world applications aiming at energy harvesting.

High-efficiency grating coupler for an ultralow-loss Si3N4-based platform.

Integrated silicon nitride waveguides of 100 nm height can achieve ultralow propagation losses below 0.1 dB/cm at the 1550 nm wavelength band but lack the scattering strength to form efficient grating couplers. An enhanced grating coupler design based on an amorphous silicon layer on top of silicon nitride is proposed and demonstrated to improve the directionality of the coupler. The fabrication process is optimized for a self-alignment process between the amorphous silicon and silicon nitride layers without increasing waveguide losses. Experimental coupling losses of 5 dB and a 3 dB bandwidth of 75 nm are achieved with both regular and focusing designs.

A study on passivation improvement in n-passivated emitter rear totally diffused solar cell using rapid thermal annealing

An investigation on enhanced surface passivation in the existing industrial process line of large area n-type silicon (Si) Passivated Emitter Rear Totally diffused ( n- PERT) solar cell has been performed. The Rapid Thermal Process (RTP) optimization for 20 min is conducted in the temperature range of 500–900°C and device evaluation is carried out with respect to regularly processed n-PERT solar cell. The impact of pre-metallization annealing is studied with the support of cell parameters like shunt resistance, reverse saturation current density determined from current-voltage measurements. The enhanced surface passivation via hydrogenation from silicon nitride (SiNx) layer during annealing is established with the help of external quantum efficiency, spectral response measurements and Fourier transform infrared spectroscopy analysis. The addition of optimized annealing resulted in improvement by 550% (from 38 to 247 µs), 7.73% (from 630.7 to 678.8 mV) and 84.77% (from 223.3 to 34 cm/s) in effective minority carrier lifetime, implied open circuit voltage and surface recombination velocity respectively. Finally, RTP technique for optimized process line has been successfully incorporated in industrial high-volume batch of 140898 CZ n-type Si wafers, which predicts conceptual validation of the study in mass scale production line with an increment in average efficiency of the device by 0.35%.

On the kinetics of the exchange of hydrogen between hydrogen–boron pairs and hydrogen dimers in crystalline silicon

Silicon samples after fast-firing with a hydrogen-rich silicon nitride layer on their surfaces can contain high concentrations of hydrogen (up to 6 × 1015 cm−3 in this study). Directly after fast-firing, this hydrogen is mostly present in a neutral dimeric state. Subsequent dark annealing applied in a temperature range between 140 and 175 °C leads to a conversion of dimers into HB pairs, which means that a significant number of boron acceptors are electrically inactive. The concentration of inactive boron, the hydrogen–boron (HB) pair concentration [HB], can thus be determined by measuring the change in specific resistivity before and after annealing. In our study, after the initial anneal for HB pair formation, the same samples are subsequently annealed at stepwise increasing higher temperatures, which leads to a partial backward conversion of HB pairs into neutral hydrogen dimers. This is indicated by a gradual reduction of the resistivity upon increasing the annealing temperature. By measuring the transient curves [HB](t) during each temperature step on samples with different boron content, we extract the parameters for the exchange between the HB pairs and the hydrogen dimers within the framework of our presented physical model. Apart from the backward formation of hydrogen dimers, from HB pairs upon increasing temperature, prolonged annealing at the same temperature leads to a transition of HB pairs into a more stable form of hydrogen dimers, compared to the initial form after fast-firing. This transition is also included in our proposed defect model.

The effect of upper layer in optical capacitor based on SiO2/SiNx/SiO2/Si structure

The SiNx/SiOx and SiOx/SiNx/SiOx structures were fabricated on silicon substrates by plasma-enhanced and low-pressure chemical vapour deposition. It was shown that upper silicon oxide layer enhances photoluminescence yield from silicon nitride layer in three times. Furthermore, top silicon oxide layer protects underlying nitride layer from un-intended oxidation during rapid thermal annealing in inert ambient (1100 °C, 3 min). The role of silicon oxide unintentionally formed during SiNx annealing and the specially deposited top SiO2 layer on SiNx photoluminescence have been discussed.