Ammonium Sulfide Passivation Research Articles

Experimental demonstration of the strong impact of plasma excitation frequency range on electronic properties of silicon nitride/GaAs interfaces

Passivation/encapsulation of III–V electronic and photonic circuits and devices is often performed using plasma-enhanced chemical vapor deposited SixNy with a standard high frequency excitation of 13.56 MHz. The hypothesis of a possible higher passivation process efficiency for low frequency SixNy deposition on GaAs has been suggested by comparison with published work on high frequency. In this work, we report a direct experimental demonstration of the effect of the RF plasma source frequency on electronic properties of GaAs/SixNy interfaces during a passivation process. GaAs substrates from the same wafer are processed in the same reactor using the same plasma conditions except for the plasma excitation frequency and MIS structures are fabricated to probe interface electronic properties. Comparing interfaces obtained with a plasma source frequency below or over the ion transit frequency (∼ 2 MHz) shows drastic difference in electrical behavior. While Fermi level is pinned when a high frequency plasma (13.56 MHz) is used, a good modulation of surface potential is observed with a low frequency plasma (380 kHz), indicating an interface states density lower by orders of magnitude. X-ray photoelectron spectra analysis is performed in order to investigate GaAs/SixNy interface formation in relation with plasma excitation frequency. We show that a short exposure to the low frequency plasma results in the almost complete removal of native oxides species. To the contrary, a significant amount of oxides remains on structures treated by the high frequency technique. The low frequency plasma technique could be a better alternative to the standard high frequency, as it is less reliant on wet chemical treatments like ammonium sulfide passivation and could allow for processing at lower temperatures.

GaAs Semiconductor Passivated by (NH4)2Sx: Analysis of Different Passivation Methods Using Electrical Characteristics and XPS Measurements

Some of the III–V semiconductor used in various devices suffer from the surface high density of states limiting their application. This study compares and evaluates five different ammonium sulfide passivation methods on GaAs surface with the aim to enhance the electrical characteristics of Au|n-GaAs Schottky junction. Wet chemical passivation of the n-GaAs surface was carried out by dipping the samples in saturated ammonium sulfide solutions at various temperatures and for various times. We also used acidic cleaning to improve the device performance. Our investigation shows a noticeable improvement in the electrical characteristics of the device reported here using acidic cleaning and ammonium sulfide passivation methods. A 23% increase in Schottky barrier height is found, which is much higher than that reported in the literature. Further, we measured a reduction of around three orders of magnitudes in saturation current as well as improvement in ideality factor to 1.23 for the best conditions of surface acidic cleaning and passivation. X-ray photoelectron spectroscopy study revealed a suppression of oxide layer by introduction of sulfide species in GaAs surface after the passivation. The lowest concentration of oxygen was found on the surface of the sample passivated under the optimum condition.

Characteristics of In0.7Ga0.3As MOS Capacitors Obtained using Hydrochloric Acid Treatment, Ammonium Sulfide Passivation, Methanol Treatment, and Forming Gas Annealing

This paper proposes a chemical pre-treatment process to reduce sulfate prior to atomic layer deposition of Al2O3 gate dielectrics on n-doped In0.7Ga0.3As layers to form metal oxide semiconductor capacitors (MOSCAPs). Cleaning with methanol was done after hydrogen chloride etching and ammonium sulfide passivation. X-ray photoelectron spectroscopy performed on the sample with methanol cleaning indicated that its native oxide regrowth was effectively prevented on the sulfur-passivated In0.7Ga0.3As surface under air exposure. By analysis using atomic force microscopy of the sample cleaned using methanol, no further increase of surface roughness was observed after 5 months of air exposure. In addition, a mid-gap interface defect density (Dit) of 1.3 × 1012 cm−2 eV−1 was obtained from this sample. After forming gas annealing (FGA), the Dit was improved significantly, to a value below 1012 cm−2 eV−1. The obtained data indicate that the combination of chemical pre-treatment and FGA is advantageous to passivating the trap states on In0.7Ga0.3As MOSCAPs.

Stability of SiGe(100) Surfaces after Ammonium Sulfide Passivation

Germanium and silicon germanium are attractive channel materials for transistors. Germanium has a smaller band gap and higher hole mobility than silicon, and varying the Ge concentration is one means of tuning the properties of SiGe alloys. Although these materials have many potential benefits, they are challenging to work with. The native oxide, GeO2, is water soluble and therefore cannot protect the surface during aqueous processes. One option is to deposit a layer that chemically passivates the surface after cleaning to protect it from reoxidation. Sulfur could passivate the surface due to its size and valency. Since the dominant cleaning technology is wet, aqueous ammonium sulfide could be a viable option. While ammonium sulfide processes deposit sub-monolayer coverages of sulfur on germanium, the surface reactions on SiGe surfaces are vastly different. Our study focuses on the treatment of Si-rich SiGe surfaces (Si1-xGex x=0.25) with aqueous (NH4)2S. We varied the concentration of (NH4)2S in solution with and without added acid and characterized the surfaces using X-ray Photoelectron Spectroscopy (XPS). Immersions in (NH4)2S alone oxidized the surface and did not deposit a sulfide layer. Instead the surface oxidized as a function of the (NH4)2S concentration. In order to remove the oxide and slow down or stop the oxidation, small volumes of HF and HCl were added to the mixture. Addition of halogen acids dropped the pH of the solution from 10 to 8 and sulfides were detected with XPS in the S 2p region. Sulfur coverage increased with increasing concentrations of HF and HCl, however oxides increased as well. The simultaneous deposition of sulfur and oxygen suggests that oxidized sulfur, such as sulfates and thiosulfates, deposited on the surface. In addition to chemical characterization of the surface, metal-insulator-semiconductor capacitors (MISCAPs) were fabricated after three different surface treatments: 1) clean and dry (control), 2) cleaning and immersion in (NH4)2S and 3) cleaning and immersion in (NH4)2S with added acid. Capacitance-voltage measurements were obtained for all three surface treatments. The density of interface defects (Dit) was calculated for each process from conductance and capacitance data. Processes that yielded the highest sulfur coverage (ammonium sulfide with added acid) showed a reduction in Dit (1.4 x 1012 cm-2 eV-1) compared to samples that were cleaned and dried only (6.4 x 1012 cm-2 eV-1) and samples treated with aqueous ammonium sulfide (4.4 x 1012 cm-2 eV-1).

Impact of Via Hole Integration on Multijunction Solar Cells for Through Cell Via Contacts and Associated Passivation Treatment

The impact of via etching on triple junction solar cell performance has been investigated for through cell via contact architectures. Triple junction solar cells with standard top and back contacts have been fabricated and vias have been etched through the subcells to investigate the new geometry proposed. The external quantum efficiency, the open-circuit voltage ( V oc), the fill factor (FF), and the ideality factor have been measured and compared to those of standard triple junction solar cells without vias. In this way, we evaluate the losses attributable to via etching. Small performance losses from via integration are observed, but performance can be partially restored with an ammonium sulfide passivation treatment. Furthermore, the results show that the V oc losses are almost absent at the high sun concentration that the new architecture is designed for. The source of the performance degradation is correlated with a larger total surface recombination at the edges of the device. The passivation treatment allows an effective surface passivation of the via holes.

Stability of SiGe(100) Surfaces after Ammonium Sulfide Passivation

Si1-xGex (x=0.25) substrates were treated with aqueous ammonium sulfide with the goal of passivating the surface both electrically and chemically. X-ray photoelectron spectroscopy (XPS) showed that no sulfur was deposited. Instead the surface was oxidized in solution, and the higher the ammonium sulfide concentration, the more oxide that formed. Spectroscopic ellipsometry showed that samples treated with dilute ammonium sulfide oxidized much more rapidly when exposed to air than those treated with a higher concentration. Addition of halogen acids to the sulfur solution deposited sulfides, but oxides were also present. The density of interface defects (Dit) was extracted from capacitance-voltage measurements of metal-insulator-semiconductor capacitors for the clean acid last surface and after either ammonium sulfide alone or ammonium sulfide plus acid. Samples treated in ammonium sulfide plus acid showed the smallest increase in interface defects compared to the other two surfaces.

Modulation of interfacial and electrical properties of HfGdO/GaAs gate stacks by ammonium sulfide passivation and rapid thermal annealing

Abstract In this work, the interface chemistry and the reduction of GaAs surface species by using (NH 4 ) 2 S solution prior to Gd-doped HfO 2 (HGO) thin film deposition and the removal of Ga Oxides and elemental As by rapid thermal annealing (RTA) have been investigated by X-ray photoelectron spectroscopy (XPS). Additionally, the effect of the surface passivation and rapid thermal annealing on the electrical properties of MOS capacitors based on sputtering-derived HGO as gate dielectric on GaAs substrate has been detected by means of capacitance-voltage ( C-V ) and leakage current density-voltage ( J-V ) measurements. Based on electrical analysis, it can be noted that the constantly improvement of electrical properties, such as the decreases of flat band voltage (V fb ), hysteresis (ΔV fb ), oxide charge density (O ox ), border trapped oxide charge density (N bt ) and leakage current density, have been observed. Especially, the dielectric constant of 16.72, flat band voltage V fb of 1.19 V, hysteresis ΔV fb of 0.04 V, leakage current density of 1.54 × 10 −5 A/cm 2 at bias voltage of 1 V, total positive charge density and border trap charge density of 6.09 × 10 12 cm −2 and 2.54 × 10 11 cm −2 , respectively render 600 °C-annealed HGO thin films, potential high-k gate dielectrics in future CMOS devices.

Surface Cleaning of SiGe(100) and Passivation of Ge(100) with Aqeuous Ammonium Sulfide

As scaling of microelectronic devices continues, germanium and silicon-germanium are promising materials for integration into silicon-based technology. Germanium and silicon-germanium based field-effect transistors can reduce leakage current and consume less power by means of improved carrier mobility and a narrower band gap, as opposed to traditional electronics. However, silicon-germanium surfaces are more challenging to work with because the topmost surface layer is comprised of germanium oxides that complicate cleaning and passivation processes. Here we seek to obtain a smooth, oxide free SiGe surface and hope to garner an understanding of the best conditions to effectively remove surface oxides. Additionally, we hope to further our understanding of aqueous ammonium sulfide passivation processes on Ge surfaces that will prevent re-oxidation of the surface for applications towards modern field-effect transistors (FETs). The first portion of this work focuses on the removal of oxides from the SiGe surface with 50 and 75% Ge molar ratios using a variety of aqueous processes. These processes include SC-1, HCl, HF, and HCl/HF mixtures. The SiGe surface was characterized using X-ray photoelectron Spectroscopy (XPS) after each chemical treatment. Film thicknesses were measured using spectroscopic ellipsometry. Immersion in HCl showed similar results to as shipped samples and was ineffective in removing surface oxides. Aqueous HF, however, removed both surface oxides and carbon contamination. SC-1 treatments selectively reduced the Ge in the SiGe film and left the surface enriched in Si, which oxidized as confirmed by XPS. If the Ge could be completely removed from the top-most layer, the SiO2 layer that forms could be used to passivate the SiGe surface and avoid problematic Ge oxides. Ellipsometry results showed that SC-1 solutions (NH4OH:H2O2:H2O, 1:1:100 v/v) etched 1 nm of the SiGe film, which was then oxidized. Ge 2p to Si 2p XPS peak ratios confirmed that increasing concentrations of SC-1 depleted the Ge content on the surface from 0.8 to 0.4, confirming the enrichment of the surface in Si. Ge to Si peak area ratios as a function of the SC-1 dilution factor showed a non-linear trend for SiGe 50% surfaces. Passivation of Ge with aqueous (NH4)2S was also studied. Because Ge oxides are unstable and water soluble, it is advantageous to remove them and deposit a sulfur layer to prevent further oxidation when exposed to air. Surface passivation was studied as a function of deposition time, solution concentration and the addition of H2O2. Passivation of the surface was ineffective at times below 20 min. Changing the concentration of (NH4)2S also had no noticeable effect on the surface based on XPS. Furthermore, at high ammonium sulfide concentrations the addition of H2O2 did not affect the passivation process. The addition of H2O2 to dilute (NH4)2S showed poor passivation due to oxidation of the surface by H2O2.

Photocurrent Mapping of 3D CdSe/CdTe Windowless Solar Cells

This paper details the use of scanning photocurrent microscopy to examine localized current collection efficiency of thin-film photovoltaic devices with in-plane patterning at a submicrometer length scale. The devices are based upon two interdigitated comb electrodes at the micrometer length scale prepatterned on a substrate, with CdSe electrodeposited on one electrode and CdTe deposited over the entire surface of the resulting structure by pulsed laser deposition. Photocurrent maps provide information on what limits the performance of the windowless CdSe/CdTe thin-film photovoltaic devices, revealing "dead zones" particularly above the electrodes contacting the CdTe which is interpreted as recombination over the back contact. Additionally, the impact of ammonium sulfide passivation is examined, which enables device efficiency to reach 4.3% under simulated air mass 1.5 illumination.

Strained germanium–tin (GeSn) p-channel metal-oxide-semiconductor field-effect-transistors (p-MOSFETs) with ammonium sulfide passivation

High-mobility strained Ge0.958Sn0.042 p-channel metal-oxide-semiconductor field-effect-transistors (p-MOSFETs) with ammonium sulfide [(NH4)2S] surface passivation were demonstrated. A ∼10nm thick fully-strained single crystalline GeSn layer was epitaxially grown on Ge (100) substrate as the channel layer. (NH4)2S surface passivation was performed for the GeSn surface, followed by gate stack formation. Ge0.958Sn0.042 p-MOSFETs with (NH4)2S passivation show decent electrical characteristics and a peak effective mobility of 509cm2/Vs, which is the highest reported peak mobility obtained for GeSn channel p-MOSFETs so far.

The Impact of HCl Precleaning and Sulfur Passivation on the Al2O3/Ge Interface in Ge Metal-Oxide-Semiconductor Capacitors

Surface treatment for Ge substrates using hydrogen chlorine cleaning and chemical passivation are investigated on AuTi/Al2O3/Ge metal-oxide-semiconductor capacitors. After hydrogen chlorine cleaning, a smooth Ge surface almost free from native oxide is demonstrated by atomic force microscopy and x-ray photoelectron spectroscopy observations. Passivation using a hydrogen chlorine solution is found to form a chlorine-terminated surface, while aqueous ammonium sulfide pretreatment results in a surface terminated by Ge-S bonding. Compared with chlorine-passivated samples, the sulfur-passivated ones show less frequency dispersion and better thermal stability based on capacitance-voltage characterizations. The samples with HCl pre-cleaning and (NH4)2S passivation show less frequency dispersion than the HF pre-cleaning and (NH4)2S passivated ones. The surface treatment process using hydrogen chlorine cleaning followed by aqueous ammonium sulfide passivation demonstrates a promising way to improve gate dielectric/Ge interface quality.

Surface channel current in InAs∕GaSb type-II superlattice photodiodes

We observed experimental evidences of surface channel current on the sidewall of InAs∕GaSb superlattice photodiodes. We investigated the surface channel current by measuring the current-voltage (I-V) characteristics of the diodes. The experimental data compare very well with our theoretical model before and after ammonium sulfide passivation. By using the passivation, we reduced the surface channel current by five times, which supports that the surface channel current is induced by surface carriers. We believe that the surface channel current results from the inversion layer of a p-type superlattice with surface Fermi levels pinned above the conduction-band minimum in InAs∕GaSb superlattices.

Surface recombination velocity reduction in type-II InAs∕GaSb superlattice photodiodes due to ammonium sulfide passivation

The surface recombination velocity (SRV) of minority electrons in a type-II InAs∕GaSb superlattice photodiode is quantitatively investigated using the electron beam induced current technique and its value used to evaluate the effects of two different passivation methods. Before passivation, the SRV was determined to be (5.0±0.2)×104cm∕s. The SRVs of two samples passivated at room temperature are compared with that of the unpassivated sample. One passivation method, using a neutralized (NH4)2S solution for 60min, reduces the SRV by a factor of 2. The other passivation method, using 4% (NH4)2S solution for 30min, reduces the SRV by more than one order of magnitude.

InAs/GaSb Superlattice Detectors Operating at Room Temperature

We have investigated mid-infrared detectors based on type-II transitions in InAs/GaSb strain layer superlattices (SLS). The SLS heterostructures were grown by solid source molecular beam epitaxy and were characterized by double crystal X-ray diffraction, transmission electron microscopy, photoluminescence and absorption measurements. The effect of aqueous ammonium sulfide passivation was investigated using picosecond excitation correlation (PEC) and variable array diode analysis (VADA). The surface recombination velocity was estimated to be 9.5times105 cm/s for the unpassivated sample and 4.0times105 cm/s for the passivated sample. NIP diodes were grown and fabricated with 300 periods of 8 ML InAs/8 ML GaSb SLS in the active region. Spectral response with lambdacut-off~5mum was observed at room temperature with a Johnson noise limited D* =4.6times109 cm.Hz1/2/W and a conversion quantum efficiency of 67% at Vb=-0.3 V. Scattering in the substrate was ignored for these measurements

Investigation of Surface Passivation in InAs/GaSb Strained-Layer-Superlattices Using Picosecond Excitation Correlation Measurement and Variable-Area Diode Array Surface Recombination Velocity Measurement

ABSTRACTThe effect of ammonium sulfide passivation on InAs/GaSb superlattice infrared detectors was investigated using two complementary techniques, namely, picosecond excitation correlation (PEC) measurement and variable-area diode array (VADA) surface recombination velocity (SRV) measurement. PEC measurements were conducted on etched InAs/GaSb superlattice mesas, which were passivated in aqueous ammonium sulfide solutions of various strengths for several durations. The PEC signal's decay time constant (DTC) is proportional to carrier lifetimes. At 77 K the PEC signal's DTC of the as-grown InAs/GaSb superlattice sample was 2.0 ns, while that of the unpassivated etched sample was reduced to 1.2 ns by the surface states at the mesa sidewalls. The most effective ammonium sulfide passivation process increased the PEC signal's DTC to 10.4 ns. However it is difficult to isolate surface recombination from other processes that contribute to the lifetime using the PEC data, therefore a VADA SRV measurement was undertaken to determine the effect of passivation on surface recombination. The obtained SRV in the depletion region of the InAs/GaSb superlattice and GaSb junction was 1.1×106 cm/s for the unpassivated sample and 4.6×105 cm/s for the passivated sample. At 77 K the highest R0A value measured in our passivated devices was 2540 W cm2 versus 0.22 W cm2 for the unpassivated diodes. The results of the lifetime, the SRV and the R0A measurements indicate that ammonium sulfide passivation will improve the performance of InAs/GaSb superlattice infrared detectors.

Comparison of surface properties of sodium sulfide and ammonium sulfide passivation of GaAs

Sodium sulfide has been demonstrated to significantly improve the electrical properties of the ambient-exposed (100) GaAs surface. It has also shown usefulness in several practical device structures. Ammonium sulfide has been proposed to have similar properties and has already been demonstrated to improve InP and GaAs devices. We compare the photoluminescence behavior and band bending of GaAs samples treated with these two different sulfide species. We also compare the chemical nature of the surfaces treated using these two processes by Auger electron spectroscopy. The surface recombination velocity was found to be similarly affected by the two treatments and the band bending found to be essentially identical. Sodium sulfide produces a film and interfacial region rich in oxygen due to hydration. However, the ammonium sulfide was found to leave very little oxygen on the surface, even compared to conventional GaAs cleaning and etching procedures.