Passive Silicon Research Articles

State‐of‐the‐art all‐silicon sub‐bandgap photodetectors at telecom and datacom wavelengths

AbstractSilicon‐based technologies provide an ideal platform for the monolithic integration of photonics and microelectronics. In this context, a variety of passive and active silicon photonic devices have been developed to operate at telecom and datacom wavelengths, at which silicon has minimal optical absorption ‐ due to its bandgap of 1.12 eV. Although in principle this transparency window limits the use of silicon for optical detection at wavelengths above 1.1 μm, in recent years tremendous advances have been made in the field of all‐silicon sub‐bandgap photodetectors at telecom and datacom wavelengths. By taking advantage of emerging materials and novel structures, these devices are becoming competitive with the more well‐established technologies, and are opening new and intriguing perspectives. In this paper, a review of the state‐of‐the‐art is presented. Devices based on defect‐mediated absorption, two‐photon absorption and the internal photoemission effect are reported, their working principles are elucidated and their performance discussed and compared. image

Low-Loss BaTiO3–Si Waveguides for Nonlinear Integrated Photonics

Barium titanate (BaTiO3) has become an attractive material to extend the functionalities of the silicon photonics platform because of its large Pockels coefficient of more than 1000 pm/V. BaTiO3 integrated epitaxially on silicon-on-insulator substrates can be structured in passive and electro-optic silicon photonic devices using slot-waveguide geometries, both of which have been demonstrated. However, all devices demonstrated so far suffer from high optical propagation losses of ∼40–600 dB/cm, which limits their performance compared with state-of-the-art silicon photonics devices (<2 dB/cm). Here, we identify the origin of these high propagation losses and demonstrate a path to fabricate low-loss BaTiO3–Si waveguides with propagation losses of only 6 dB/cm. In particular, we identified the thin strontium titanate (SrTiO3) seed layer typically used for the epitaxial deposition of BaTiO3 on silicon as the main source of absorption: When manufacturing slot-waveguide structures, the BaTiO3/SrTiO3 layer stack ...

Effect of Different ALD Al2O3 Oxidants on the Surface Passivation of Black Silicon

We study how different oxidants in atomic layer deposition of aluminium oxide (ALD Al2O3) affect the surface passivation of black silicon. Here we show that processes using ozone cause higher fixed charge but surprisingly lead to lower lifetimes in black silicon samples as compared to water-based samples. In planar samples however, the best surface passivation is reached with O3-based processes. In case of water as oxidant, the planar wafers suffer from severe blistering and poorer surface passivation, while this seems to be the best process for black silicon. To find a reason for the lifetime differences we also study different Al2O3 stacks where both H2O and O3 are used as oxidants. In conclusion, surface texture seems to affect the optimal oxidant in the ALD process.

Improved Silicon Surface Passivation of APCVD Al2O3 by Rapid Thermal Annealing

Short-duration post-deposition thermal treatments at temperatures above those normally used for annealing activation have the potential to further improve the already excellent passivation of crystalline silicon (c-Si) achieved by Al2O3, but have so far received little attention. In this work we investigate the influence of rapid thermal annealing (RTA) on the surface passivation of c-Si by Al2O3 deposited by atmospheric pressure chemical vapour deposition (APCVD) as a function of RTA peak temperature between 500 and 900°C, and for Al2O3 deposition temperatures between 325 and 440°C. The saturation current density J0 of undiffused p-type surfaces is observed either to increase or decrease following RTA depending on the Al2O3 deposition temperature and the RTA peak temperature. The optimum deposition temperature depends on the post-deposition thermal processing to be applied. Films deposited at lower temperatures provide worse passivation after low temperature heat treatment, but maintain this passivation better at higher RTA temperatures. An exceptionally low J0 of 7 fA cm−2, due to the combination of a very low interface state density Dit and unusually high negative fixed charge density Qf, is achieved by the use of a short 500–550°C RTA combined with optimised deposition conditions.

Solution Synthesis, Surface Passivation, Optical Properties, Biomedical Applications, and Cytotoxicity of Silicon and Germanium Nanocrystals.

Silicon and germanium nanocrystals (NCs) are attractive materials owing to their unique size and surface-dependent optical properties. The optical properties of silicon and germanium NCs make them highly suitable for a range of applications, including bioimaging, light-emitting diodes, and solar cells. In this review, the solution synthesis, surface passivation, optical properties, biomedical applications, and cytotoxicity of silicon and germanium NCs are compared and contrasted. Over the last 10 years, synthetic protocols have improved considerably, with size control readily achieved. Investigations have begun into a range of silicon and germanium nanostructures, including doped, alloy, and metal-semiconductor hybrid NCs, which represent the next generation of silicon and germanium nanomaterials. Silicon and germanium NCs are actively researched for a wide array of biomedical applications, including, long-term in vivo cellular imaging, fluorescent nanocarriers for drug delivery, and as contrast agents for magnetic resonance imaging (MRI). Cytotoxicity studies have shown the low toxicity of Si NCs, while demonstrating that Ge NCs are less toxic than CdSe NCs at similar concentrations, giving these materials a strong future in nanomedicine applications.

Self-validating lab-on-a-chip for monitoring enzyme-catalyzed biological reactions

This paper reports on a miniaturized and highly integrated lab-on-a-chip (LoC) combining both optical and electrochemical transduction modes for self-validating determination of biological analytes that undergo enzyme-catalyzed-reactions. This LoC monolithically integrates a biofunctionalized microfluidic mixer, also working as bioreactor, together with a measurement chamber, both fabricated in polydimethylsiloxane (PDMS) and applied for enzyme-mediated detection of glucose in continuous flow regime. The measurement chamber combines a multiple internal reflection (MIR) photonic cuvette and a tailor-made electrochemical cell, where gold electrode areas are defined by a silicon oxide passivation layer. Having separated mixing/reaction and measurement chambers greatly facilitates the functionalization process during the microsystem fabrication and the further calibration/rinsing steps carried out during the microsystem performance. A cascade bi-enzymatic reaction involving glucose oxidase (GOx) and horseradish peroxidase (HRP) together with azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) diammonium salt (ABTS) mediator is applied to carry out the glucose dual detection. The optical and electrochemical properties of the ABTS enable simultaneous absorbance and amperometric measurements, without cross-talk. The system requires low sample volumes below 15μL and presents tunable analytical properties that can be adjusted by varying the flow rate. Linear ranges extending up to 1.6mM and 2mM glucose and limits of detection of 0.23±0.02mM and 0.064±0.001mM are achieved with the optical and electrochemical detection approaches, respectively, when operating simultaneously at 10μL/min flow rate. The high degree of integration results in a minimization of dead volumes, reagent consumption and response time, providing with a high performance self-validating structure ideal for biomedical applications.

In-situ optical emission spectroscopy diagnostic of plasma ignition impact on crystalline silicon passivation by a-Si:H films

The influence of the plasma ignition condition during PECVD deposition from a silane/hydrogen mixture on the amorphous silicon passivation of crystalline silicon surface is investigated. The changes in this process step mainly consist in varying the power density for very brief durations in between 1 s and 3 s. We find that the ignition phase contributes significantly in the film growth, especially in the a-Si:H/c-Si interface formation. In particular, the deposition rate increases with ignition power density. TEM cross-section inspection presents a rougher a-Si:H/c-Si interface with higher plasma power and thus, a tendency for nano-clusters formation caused by the crystalline nature of the substrate. In-situ plasma diagnostics reveal the gradual raise up of IHa*/ISiH* with the power density leading to worse SiH* abstraction to the surface. Whereas, time-resolved optical emission spectroscopy explains the possible recombination mechanism in the plasma due to higher-silane related reactive species (HSRS) formation via polymerization reactions. Our results point out that the ignition conditions with a rather low power for longer time give the best passivation, resulting an effective lifetime up to 9 ms.

Chip-Scale Silicon Photonic Interconnects: A Formal Study on Fabrication Non-Uniformity

Silicon photonic interconnect (SPI) is an attractive alternative for the power-hungry and low-bandwidth metallic interconnect in multiprocessor systems-on-chip (MPSoCs). When employing SPIs for wavelength-division multiplexing (WDM)-based applications, it is essential to precisely align the central wavelengths of different photonic devices (e.g., photonic switches) to achieve a reliable communication. However, SPIs are sensitive to fabrication nonuniformity (a.k.a. fabrication process variation), which results in wavelength mismatches between devices, and hence performance degradation in SPIs. This work presents a computationally efficient and accurate bottom-up approach to study the impact of fabrication process variations on passive silicon photonic devices and interconnects. We first model the impact of process variations at the component level (i.e., strip waveguides), then at the device level (i.e., add-drop filters and photonic switches), and finally at the system level (i.e., passive WDM-based SPIs). Numerical simulations are performed not only to evaluate the accuracy of our method, but also to demonstrate its high-computational efficiency. Furthermore, our study includes the design, fabrication, and analysis of several identical microresonators to demonstrate process variations in silicon photonics fabrication. The efficiency of our proposed method enables its application to large-scale passive SPIs in MPSoCs, where employing time-consuming numerical simulations is not feasible.

Ultrathin Titanium Dioxide Nanolayers by Atomic Layer Deposition for Surface Passivation of Crystalline Silicon

We demonstrate a low surface recombination velocity of 14 cm/s with only 1.5 nm thin titanium dioxide (TiO2 ) layers on undiffused 10 Ωcm p-type crystalline silicon. The TiO2 nanolayers were deposited by thermal atomic layer deposition at 150 °C and 200 °C substrate temperatures using tetrakis-dimethyl-amido titanium as the Ti precursor and water as the oxidant. The influence of a post-deposition anneal in forming gas at different temperatures was investigated. We have observed that a subsequent anneal in forming gas at 350 °C enhances the surface passivation quality of the TiO2 layers tremendously. Increasing the thickness of the TiO2 layers leads to a reduction of the surface passivation quality. Introducing a thin interfacial layer of silicon oxide (1.6 nm) grown by rapid thermal oxidation underneath the TiO2 layer improves the surface passivation of thicker TiO2 layers (5.5 and 15 nm). These results show that ultrathin TiO2 layers with a thickness of only 1.5 nm can be used to effectively passivate the c-Si surface.

Interface Processing of Amorphous–Crystalline Silicon Heterojunction Prior to the Formation of Amorphous-to-Nanocrystalline Transition Phase

Two effective interface optimizations, i.e., hydrogen-plasma treatment (HPT) and oxygen incorporation (OIn), have been investigated to optimize the interface structure and promote the formation of hydrogenated amorphous silicon (a-Si:H) with an amorphous-to-nanocrystalline transition phase based on the surface passivation of crystalline silicon. Both multistep HPT and intentional OIn are capable of producing a compact interface and a transition phase in the case of initial a-Si:H films deposited under different conditions. An appropriate multistep HPT is responsible for modifying the a-Si:H into a transition phase, whereas an effective OIn forms a disordered and more compact interface, as well as a crystallite-isolated and more transparent structure.

Passivation effect of water vapour on thin film polycrystalline Si solar cells

In this work, we investigated a passivation of both surface and bulk of polycrystalline silicon films by water vapour. To assess the passivation effect we used Suns‐VOC method to measure the open‐circuit voltage VOC of polycrystalline silicon thin film solar cells. The treatment conditions, i.e. temperature, steam pressure, duration, were systematically varied to optimise the passivation process. According to our results, there is a parameter interaction because the same passivation effect can be achieved at different treatment conditions. A sufficiently high temperature (350–450 °C) is necessary for a successful silicon passivation. The impact of this parameter cannot be replaced either by an elevated steam pressure or a prolonged exposure time. Nevertheless, the passivation effect of steam can be strengthened by an elevated steam pressure. Different gases were tested beside water vapour, e.g. H2, H2 + H2O, O2 + H2O, air, but none of them resulted in higher VOC than pure steam (360 mV from starting 220 mV). Results from Fourier transform infrared spectroscopy indicate that water vapour passivation is rather oxidation while hydrogen plays a significant supporting role in the process. We conclude that the water vapour passivation is able to passivate defects in the whole silicon volume. However, its passivation effect is not strong enough to become an adequate alternative to the plasma hydrogenation with the best result of VOC ∼497 mV. On the other hand, it provides advantage of simplicity (no vacuum system and deionised water steam as the only input).Recombination activity of defects in polycrystalline Si can be suppressed by their saturation for instance in hydrogen plasma (Si–H) or in water vapour (Si–O–Si).

Mechanically-Tunable Photonic Devices with On-Chip Integrated MEMS/NEMS Actuators.

This article reviews mechanically-tunable photonic devices with on-chip integrated MEMS/NEMS actuators. With related reports mostly published within the last decade, this review focuses on the tuning mechanisms of various passive silicon photonic devices, including tunable waveguides, couplers, ring/disk resonators, and photonic crystal cavities, and their results are selectively elaborated upon and compared. Applications of the mechanisms are also discussed. Future development of mechanically-tunable photonics is considered and one possible approach is based on plasmonics, which can confine light energy in the nano-scale space. Optomechanics is another innovation, derived from the strong coupling of optical and mechanical degrees of freedom. State-of-the-art studies of mechanically-tunable plasmonics and on-chip optomechanics are also selectively reviewed.

Formation of amorphous silicon passivation films with high stability against postannealing, air exposure, and light soaking using liquid silicon

We applied liquid-source vapor deposition (LVD), thermal CVD from the vapor of cyclopentasilane (CPS), to form amorphous silicon (a-Si) passivation films on crystalline Si (c-Si) wafers, and investigated the thermal stability of the films against postannealing. LVD a-Si passivation films showed a high initial effective minority carrier lifetime (τeff) of >300 µs and a higher thermal stability than a reference plasma-enhanced chemical-vapor-deposited (PECVD) sample. The high thermal stability of LVD a-Si passivation films may be attributed to the considerably high deposition temperature of the films at 360 °C or more. LVD a-Si passivation films were sufficiently stable also against air exposure and 1-sun light soaking. We also confirmed that the epitaxial growth of Si films does not occur on c-Si even at such high deposition temperatures, and LVD could realize the simultaneous deposition of a-Si films on both sides of a c-Si wafer.

Method for analyzing passive silicon carbide thermometry with a continuous dilatometer to determine irradiation temperature

Silicon carbide is used as a passive post-irradiation temperature monitor because the irradiation defects will anneal out above the irradiation temperature. The irradiation temperature is determined by measuring a property change after isochronal annealing, i.e., lattice spacing, dimensions, electrical resistivity, thermal diffusivity, or bulk density. However, such methods are time-consuming since the steps involved must be performed in a serial manner. This work presents the use of thermal expansion from continuous dilatometry to calculate the SiC irradiation temperature, which is an automated process requiring minimal setup time. Analysis software was written that performs the calculations to obtain the irradiation temperature and removes possible user-introduced error while standardizing the analysis. This method has been compared to an electrical resistivity and isochronal annealing investigation, and the results revealed agreement of the calculated temperatures. These results show that dilatometry is a reliable and less time-intensive process for determining irradiation temperature from passive SiC thermometry.

Numerical simulation of silicon heterojunction solar cells with Si/Si1-xGex quantum wells

Heterojunction with intrinsic thin-layer (HIT) solar cells attract attention due to their high open circuit voltage and stable performance. However, short circuit current density is difficult to improve due to light losses of transparent conductive oxide and hydrogenated amorphous silicon passivation (a-Si:H) layer and low absorption coefficient of crystalline silicon (c-Si). Silicon germanium alloy (Si/Si1-xGex) quantum wells and quantum dots are capable of improving low light utilization by strong optical absorption in the infrared region. In this article, opto-MoS2of the HIT solar cells integrated with Si/Si1-xGex quantum wells (HIT-QW) as a surface absorber are investigated by numerical simulation with Technology Computer Aided Design (TCAD). The influences of germanium content on the MoS2of HIT solar cells with long carrier lifetimes of Si1-xGex layers (p*) and defect-free a-Si:H/c-Si interface are investigated at first. The simulation results indicate that optical utilization in the infrared region is enhanced with the increase of germanium fraction, while open circuit voltage degrades due to the decreasing of the energy band gap of Si1-xGex, radiative recombination and auger recombination mechanism in the Si/Si1-xGex quantum wells. And the conversion efficiency reaches a maximum value at a germanium fraction of 0.25 then drops distinctly. When the germanium fraction increases from 0 to 0.25, the short circuit current density increases from 34.3 mA/cm2 to 34.8 mA/cm2, while the open circuit voltage declines from 749 mV to 733 mV. Hence, the conversion efficiency increases from 21.5% to 21.7% due to the fact that the enhancement of short circuit current density compensates for the reduction of open circuit voltage. When the germanium content increases to more than 50%, a serious open circuit voltage loss of more than 130 mV associated with the energy band gap loss of Si1-xGex arises in the HIT-QW solar cells, which indicates that the dominating carrier transport mechanism changes from shockley diffusion to recombination in the Si/Si1-xGex quantum wells. Subsequently, the influences of interface defects at a-Si:H/c-Si interface and bulk recombination centers in the Si/Si1-xGex quantum wells are discussed. Both interface holes at a-Si:H/c-Si interface and bulk holes in Si1-xGex quantum wells can be recombined through the interface defects at a-Si:H/c-Si interface and bulk recombination centers in the Si/Si1-xGex quantum wells, respectively, which restricts the position of hole fermi level in the open circuit condition. When the germanium fraction increases, the influence of interface defects at a-Si:H/c-Si interface becomes weak on the degradation of open circuit voltage compared with the significant influence of the bulk trap centers. Moreover, p* of longer than 510-5 s is necessary for the retention of electrical performance of HIT-QW solar cells by the simulation. Based on this research, high-efficiency HIT solar cells can be achieved by incorporating high-quality Si/Si0.75Ge0.25 quantum wells, which also requires the impactful passivation of a-Si:H/c-Si interface.

Hydrogen doping of Indium Tin Oxide due to thermal treatment of hetero-junction solar cells

Indium Tin Oxide (ITO) layers in silicon hetero junction solar cells change their electrical and optical properties when exposed to temperature treatments. Hydrogen which effuses from underlying amorphous silicon layers is identified to dope the ITO layer. This leads to an additional increase in conductivity. In this way an almost isolating ITO can become degenerately doped through temperature treatments. The resulting carrier density in the range of 1020cm−3 leads to a substantial increase in free carrier absorption, which in turn leads to an increased parasitic absorption in the cell device. Thus hydrogen effusion in silicon hetero-junction (SHJ) solar cells does not only affect the degradation of amorphous silicon (a-Si:H) passivation of crystalline silicon (c-Si), but also the electrical and optical properties of both front and back ITO layers. This leads to the further design rule for SHJ solar cells, meaning that ITO properties have to be optimized in the state after modification during temperature treatment.

Electronic band-gap modified passive silicon optical modulator at telecommunications wavelengths

The silicon optical modulator is considered to be the workhorse of a revolution in communications. In recent years, the capabilities of externally driven active silicon optical modulators have dramatically improved. Self-driven passive modulators, especially passive silicon modulators, possess advantages in compactness, integration, low-cost, etc. Constrained by a large indirect band-gap and sensitivity-related loss, the passive silicon optical modulator is scarce and has been not advancing, especially at telecommunications wavelengths. Here, a passive silicon optical modulator is fabricated by introducing an impurity band in the electronic band-gap, and its nonlinear optics and applications in the telecommunications-wavelength lasers are investigated. The saturable absorption properties at the wavelength of 1.55 μm was measured and indicates that the sample is quite sensitive to light intensity and has negligible absorption loss. With a passive silicon modulator, pulsed lasers were constructed at wavelengths at 1.34 and 1.42 μm. It is concluded that the sensitive self-driven passive silicon optical modulator is a viable candidate for photonics applications out to 2.5 μm.

Formation of Mach angle profiles during wet etching of silica and silicon nitride materials

In integrated circuit technology peeling of masking photoresist films is a major drawback during the long-timed wet etching of materials. It causes an undesired film underetching, which is often accompanied by a formation of complex etch profiles. Here we report on a detailed study of wedge-shaped profile formation in a series of silicon oxide, silicon oxynitride and silicon nitride materials during wet etching in a buffered hydrofluoric acid (BHF) solution. The shape of etched profiles reflects the time-dependent adhesion properties of the photoresist to a particular material and can be perfectly circular, purely linear or a combination of both, separated by a knee feature. Starting from a formal analogy between the sonic boom propagation and the wet underetching process, we model the wedge formation mechanism analytically. This model predicts the final form of the profile as a function of time and fits the experimental data perfectly. We discuss how this knowledge can be extended to the design and the realization of optical components such as highly efficient etch-less vertical tapers for passive silicon photonics.

Thermal stability of hydrogenated amorphous silicon passivation for p‐type crystalline silicon

The thermal stability of hydrogenated amorphous silicon‐passivated monocrystalline silicon wafers is investigated at different annealing times and temperatures. Stepwise thermal annealing was performed at temperatures ranging from 100 to 500 °C for 1, 5 and 15 min, respectively, and the effective minority carrier lifetime was measured after each annealing step using carrier density imaging. Annealing at lower temperatures up to 250 °C results in a significant improvement of the effective carrier lifetime compared with that of as‐deposited samples, increasing to a maximum value of ∼1.5 ms after annealing the sample at 200 °C for 15 min. Annealing at temperatures from 350 °C and upwards results in a reduction of the effective carrier lifetime of all samples. The temperature dependence of the effective lifetime after annealing was investigated by separating the temperature into in three separate ranges; activation‐, peak‐ and degradation. The limits of the activation and degradation temperature: were obtained through annealing experiments. The activation energy for the process responsible for the lifetime improvement was found to be 0.2 eV. An initial degradation energy around 2.1 eV was obtained for degradation process, which can be correlated with H effusion.

Influence of Chemical Composition and Structure in Silicon Dielectric Materials on Passivation of Thin Crystalline Silicon on Glass.

In this study, various silicon dielectric films, namely, a-SiOx:H, a-SiNx:H, and a-SiOxNy:H, grown by plasma enhanced chemical vapor deposition (PECVD) were evaluated for use as interlayers (ILs) between crystalline silicon and glass. Chemical bonding analysis using Fourier transform infrared spectroscopy showed that high values of oxidant gases (CO2 and/or N2), added to SiH4 during PECVD, reduced the Si-H and N-H bond density in the silicon dielectrics. Various three layer stacks combining the silicon dielectric materials were designed to minimize optical losses between silicon and glass in rear side contacted heterojunction pn test cells. The PECVD grown silicon dielectrics retained their functionality despite being subjected to harsh subsequent processing such as crystallization of the silicon at 1414 °C or above. High values of short circuit current density (Jsc; without additional hydrogen passivation) required a high density of Si-H bonds and for the nitrogen containing films, additionally, a high N-H bond density. Concurrently high values of both Jsc and open circuit voltage Voc were only observed when [Si-H] was equal to or exceeded [N-H]. Generally, Voc correlated with a high density of [Si-H] bonds in the silicon dielectric; otherwise, additional hydrogen passivation using an active plasma process was required. The highest Voc ∼ 560 mV, for a silicon acceptor concentration of about 10(16) cm(-3), was observed for stacks where an a-SiOxNy:H film was adjacent to the silicon. Regardless of the cell absorber thickness, field effect passivation of the buried silicon surface by the silicon dielectric was mandatory for efficient collection of carriers generated from short wavelength light (in the vicinity of the glass-Si interface). However, additional hydrogen passivation was obligatory for an increased diffusion length of the photogenerated carriers and thus Jsc in solar cells with thicker absorbers.