Table of contents

In this study, the wafer warpage resulting from common source line tungsten (CSL W) is investigated in 3D NAND flash memory. It is found that the warpage is related to the annealing conditions after CSL W deposition, and it reduces exponentially with increasing annealing temperature or linearly with increasing annealing time. This result shows that the effect of annealing temperature on warpage is greater than that of time. Consequently, spike annealing with a low thermal budget is proposed to achieve the desired reduction of warpage as long as the annealing temperature is adequate. This work provides an effective approach to solve the wafer warpage problem in 3D NAND flash memory manufacturing.

Experimental and theoretical studies as well as the optimization of silicon epitaxy in deep trenches are reported. Detailed measurements of the trench profile as functions of time and trench location on the wafer are performed. Modeling analysis is based on a multiscale approach that accounts for the transport and chemical reactions both in the reactor and in the trenches. Simulations have revealed the mechanism of direct or reverse growth starvation (higher growth rate at the trench top or bottom, respectively) and explain the evolution of the trench shape observed in the experiment. Eventually, the thickness uniformity within the trench and wafer has been considerably improved with the proper choice of process parameters.

We report on the investigations of AlGaN/GaN field effect transistors with two lateral Schottky barrier gates on the sides of the two-dimensional electron gas. This kind of 'EdgeFET' allowed us to efficiently control the current flow in the 2DEG conduction channel. Moreover, due to depletion, regions at a certain range of reverse biasing form a nanowire, which is beneficial for the adjustable resonant THz detection. Our studies of DC characteristics and photoresponse in the sub-terahertz frequency confirm the validity of the approach.

Luminescent properties of self-assembled Ge(Si)/SOI nanoislands embedded in two-dimensional photonic crystal (PhC) slabs with and without L3 cavities were studied with PhC period a varied between 350 and 600 nm. For small periods (a ≤ 450 nm), the nanoisland luminescence, which spans over the wavelength range from 1.2 to 1.6 μm, overlaps with the PhC bandgap resulting in a coupling with the localized modes of an L3 cavity. It is shown that for larger periods (a > 450 nm), nanoisland emission couples to the radiative modes above the bandgap located in the vicinity of the Г-point of the photonic crystal Brillouin zone and is characterized by the low group velocity. In this case, a significant (up to 35-fold) increase in the PL intensity was observed in a number of PhCs without a cavity. From a technological point of view, the latter result makes such types of photonic crystal structures particularly promising for the realization of Si-based light emitters operating in the telecommunication wavelength range because, firstly, their manufacture does not require a precise cavity formation and, secondly, they provide a much larger area for the radiating region, as compared with PhC cavities.

Photoconductive antennae are one of the most common sources of terahertz radiation. Although extensive experimental and simulation work has been done to analyze the radiated fields, very limited theoretical analysis has been done to build elegant and compact models. This work presents the derivation of the radiated fields from a photoconductive antenna incorporating the effect of the semiconductor substrate. Additionally, an improved model of the semiconductor carrier dynamics has been proposed to analyze the radiation characteristics of the photoconductive antenna. The proposed carrier dynamics includes a novel analysis of the transient behavior of the substrate temperature and its effect on the parameters of the substrate material. It also incorporates the effect of the lifetime, polarization and velocities of the carriers as well as the effects of the radiated near zone fields and antenna geometry. The derived formulae and the proposed carrier dynamics model have been validated by comparing the results with the available experimental results in the literature. It has been found that the analytically calculated results closely follow the experimental results reported in the literature.

We demonstrate photodetectors sensitive to ultraviolet light entirely developed by means of inkjet printing technique and based on titanium dioxide and PEDOT:PSS. Devices have a lateral architecture and are realized on a plastic substrate, thanks to the low thermal budget production process. Pure titania devices behave as standard photodetectors, increasing their conductivity by more than four orders of magnitude upon UV light exposure. Bilayers of PEDOT:PSS and titania show an inverted behavior, with a high conductivity in the dark which drops by seven orders of magnitude upon light exposure: this is likely due to the fast recombination of PEDOT:PSS holes with photogenerated TiO₂ electrons. The series connection of pure TiO₂ and of PEDOT:PSS/TiO₂ bilayer is suggested as the basis for the development of low-power, complementary-like, photosensitive voltage dividers.

Interdot carrier and spin dynamics were studied in a two-dimensional array of high-density small quantum dots (SQDs) of InGaAs with an average diameter of 16 nm and a sheet density of 1.2 $\times \,{10}^{11}\,{\mathrm{cm}}^{-2}$, in which 24 nm diametric large QDs (LQDs) were embedded as local potential minima. We observed a delayed photoluminescence (PL) rise from the lower-lying LQD states and a considerably faster PL decay from the higher-lying SQD states, indicating carrier transfer from the two-dimensionally coupled SQDs into the LQDs. In addition, inverse carrier tunneling from the LQDs into the SQDs was thermally induced, which is characterized by the thermal activation energy between the LQDs and SQDs. Moreover, circularly polarized transient PL behavior from the SQD states exhibits a suppression of the spin polarization decay in the initial time region, depending on the excited spin density. This tentatively suppressed spin relaxation can be quantitatively explained by selective interdot transfer of minority-spin electrons from the SQDs into LQDs, when the majority spin states in both QDs are sufficiently populated by excited spins. These findings indicate that the high-density SQDs behave as the main emitters with suppressed spin relaxation, while the scattered LQDs with lower potential behave as the receivers of minority-spin electrons.

Iridates have attracted immense interest since their strong spin–orbit coupling (SOC) can lead to rich exotic phenomena such as a J_eff = ½ Mott insulating state. Here we report a novel iridate discovered in our efforts which aimed to synthesize Ba₂IrO₄ thin films. Through systematic transmission and scanning transmission electron microscopy studies, we have shown this new compound possesses a layered orthorhombic structure with the composition of Ba₇Ir₃O_13+δ (BIO). This material is an insulator with an optical band gap of ∼1.3 eV. Furthermore, we found that the thin films of this material can be grown on differently orientated perovskite substrates or MgO substrates. Although all these films maintain an identical crystallographic orientation, i.e. its c-axis perpendicular to the substrate surface, they form various domain structures dependent on the substrate. When (001)-oriented LaAlO₃ and (111) oriented SrTiO₃ perovskites are used as substrates, the domains show 12 fold and 6 fold symmetry respectively, and the domain orientations are highly coherent and the domain-walls are atomically sharp. However, the films on the (110) oriented MgO substrates feature much less coherent domain walls and thread dislocations occur at the domain boundaries. These findings not only reveal a new playground for the study of the novel SOC physics of iridates, but also provide a route to tailor the domain wall structure via epitaxial lattice mismatch in films.

We report on the large area stress free single crystal silicon (Si) layer transfer onto glass substrate using hydrogen ion (H⁺) implantation and heterogeneous direct wafer bonding (DWB) (ion-cut process). Si wafers were implanted with H⁺ ions at room temperature followed by the DWB between the Si wafer and glass substrate. Post-implantation annealing studies were performed at different temperatures. The root mean square surface roughness of the H-implanted Si wafer and glass substrate was measured to be 0.3 nm and 0.5 nm, respectively. At an annealing temperature of 330 °C, a thin layer of the Si wafer was transferred onto the glass substrate. The average thickness of the transferred layer was found to be ∼605 nm with surface roughness of 4 nm. Raman spectroscopy confirmed that the transferred Si layer was stress free and retained its crystallinity. Large area transfer of high quality stress free Si-on-glass substrate is demonstrated using heterogeneous DWB and ion-cut process.

Drain-induced-barrier-lowering (DIBL) is one of short-channel-effects (SCEs) of conventional field effect transistors (FET), which results in lowering the threshold voltage of short-channel transistors. To properly control the DIBL of FinFETs, it is well known that the fin width of the FinFET can be adjusted (usually, a narrower fin width to enhance gate-to-channel coupling). In this work, the fin width effect of FinFETs (vs. ferroelectric-gated FinFETs, a.k.a., negative capacitance (NC) FinFETs) on the DIBL has been investigated and compared. As a result, it is experimentally verified that the NC-FinFET has superior gate-to-channel controllability and better subthreshold swing due to the negative capacitance effect of ferroelectric capacitor.

In this study, detailed temperature dependent simulations for absorption and photogenerated recombination of hot electrons are compared with experimental data for an InAs/AlAsSb multi-quantum well. The simulations describe the actual photoluminescence (PL) observations accurately; in particular, the room temperature e1-hh1 simulated transition energy of 805 meV closely matches the 798 meV transition energy of the experimental PL spectra, a difference of only 7 meV. Likewise, the expected energy separations between local maxima (p1–p2) in the simulated/experimental spectra have a difference of just 2 meV: a simulated energy separation of 31 meV compared to the experimental value of 33 meV. Utilizing a non equilibrium generalized Planck relation, a full spectrum fit enables individual carrier temperatures for both holes and electrons. This results in two very different carrier temperatures for holes and electrons: where the hole temperature, T_h, is nearly equal to the lattice temperature, T_L; while, the electron temperature, T_e, is 'hot' (i.e., T_e > T_L). Also, by fitting the experimental spectra via three different methods a 'hot' carrier temperature is associated with electrons only; all three methods yield similar 'hot' carrier temperatures.

(Al_xGa_1−x)₂O₃/Ga₂O₃ metal-oxide semiconductor field effect transistors are emerging as candidates for rf and power electronics, but a drawback is the high contact resistances on these wide bandgap semiconductors. A potential solution is to use narrower gap transparent conducting oxides such as IZO and ATO to reduce the interfacial resistance. In this paper, we report the measurement of the valence band offset of the AZO/(Al_0.14Ga_0.86)₂O₃ and ITO/(Al_0.14Ga_0.86)₂O₃ heterointerfaces using x-ray Photoelectron Spectroscopy. The single-crystal β-(Al_0.14Ga_0.86)₂O₃ was grown by molecular beam epitaxy. The bandgaps of the sputter-deposited AZO and ITO were determined by reflection electron energy loss spectroscopy to be 3.2 ± 0.20 and 3.5 ± 0.20 eV, respectively, while high resolution XPS data of the O 1s peak and onset of elastic losses was used to establish the (Al_0.14Ga_0.86)₂O₃ bandgap as 5.0 ± 0.30 eV. The valence band offsets were −0.59 eV ± 0.10 eV and −1.18 ± 0.20 eV, respectively, for AZO and ITO. The conduction band offsets were 1.21 ± 0.25 eV for AZO, and 0.32 ± 0.05 eV for ITO. Both were of the straddling gap, type I alignment on β-(Al_0.14Ga_0.86)₂O₃ and all the offsets are negative, consistent with achieving improved electron transport across the heterointerface.

As a solution to future non-volatile random access memories (NV-RAMs) with large scale integration, many researchers are studying a crossbar array built with many resistive switching cells because of its simple architecture and scalability. However, a sneak current in the crossbar array that energizes memory cells not explicitly selected, raises various issues that need to be addressed for commercialization. Hence, we clearly describe the dominant causes and considerations for a passive crossbar array based on memristors. We discover that such resistive switches under reverse bias along the sneak current paths can play an important role to minimize the overall sneak current by blocking their reverse current. To develop commercializable passive crossbar resistive switching arrays, we demonstrate a read error bound and propose a new design parameter (R_RL/R_FL as ratio of reverse low resistance state and forward low resistance state) using the switching resistances on the forward and reverse current paths.

Cadmium sulfide (CdS) thin films as n-type semiconductor material were deposited by pulsed laser deposition by varying the argon pressure at room temperature. The structural, morphological and stoichiometric characteristics of the CdS films were studied as a function of the deposition pressure. The results show that the argon deposition pressure had a dramatic impact on the CdS film properties. The CdS electrical resistivity increased 10⁴ times when argon pressure was increased from 8 to 10.66 Pa. These films were employed on fully-patterned thin film transistors (TFTs) fabricated by a photolithography-based process, and their electrical characteristics were measured. The TFTs electrical performance achieved a mobility of ∼24 cm²/V-s with a threshold voltage from 1.5 to 12.6 V after testing. The deposition pressure of CdS for transistors fabrication, which optimizes the resulting electrical characteristics, was determined from this study.

The electrical properties of metal-oxide-semiconductor (MOS) capacitors with Al₂O₃/GaN interfaces formed by atomic layer deposition at various deposition temperatures (T_d) were systematically investigated through comparison with the interface properties of a Si MOS capacitor. Although interface trap densities (D_it) for the GaN MOS capacitors are almost the same as for the Si MOS capacitor, surface potential fluctuation (σ_s) for the GaN MOS capacitors formed at various T_d are much larger than for the Si MOS capacitor. Comparison between a theoretical calculation related to σ_s and the measured results clarified that the energy dependences of σ_s in the bandgap of GaN cannot be explained by electron trapping to the interface traps with the single acceptor nature (0/−). We found that the result is experimental evidence of the existence of multiple charge states at the Al₂O₃/GaN interfaces. The existence of multiple charge states suggests that it may be difficult to obtain good carrier transport in future GaN devices with a MOS structure.

In this work, we report on laser diodes with intermixed InAlGaAs multi-quantum well (MQW) as the active gain medium. Using an impurity-free vacancy-enhanced disordering technique, several silicon nitride-capped InAlGaAs MQW samples are intermixed to different levels by varying the rapid-annealing temperatures from 730 °C to 830 °C. The intermixed laser diodes operate at wavelengths spanning from 1520 nm to 1400 nm, while exhibiting an increase in threshold current from 25 mA (as-grown) to 45 mA, and a decrease in slope efficiency from 0.101 W A⁻¹ (as-grown) to 0.068 W A⁻¹ as the intermixing extent is increased. For a sample that was annealed at 770 °C while only covered with a silicon dioxide film, the lasing wavelength is 1543 nm compared to 1550 nm for the as-grown laser wavelength, and the lasing threshold current is 26 mA with a slope efficiency of 0.08 W A⁻¹.

Dopant diffusion of indium (In) in single crystal zinc oxide is studied by secondary ion mass spectrometry and is interpreted using a reaction–diffusion model that invokes predictions from density functional theory (DFT). An apparent activation energy of 2.2 eV is obtained for the diffusion of In, when the local Fermi-level position is about 0.2 eV below the conduction band edge. The diffusion of In is found to be significantly faster that that reported for the other group III donors, aluminum and gallium, with several orders of magnitude higher effective diffusivities, that can be assigned to a lower migration barrier for the diffusion of In. Furthermore, our results reveal self-consistency in previous DFT results of defect formation- and migration energies. From this, the diffusion of In is suggested to occur through mobile charged zinc vacancies ${{V}}_{\mathrm{Zn}}^{2-}$ that form intermediate mobile (${\mathrm{In}}_{\mathrm{Zn}}{{V}}_{\mathrm{Zn}}$)⁻ pairs. The pairs in turn dissociate rather readily at the studied temperatures (850 °C–1150 °C), which results in distinct and abrupt diffusion fronts for the In depth distribution profiles.

Quantum dot (QD) laser diodes, where the active region consists of wetting-layer-free InAs QDs, were demonstrated by block copolymer (BCP) lithography and subsequent selective area metal-organic vapor phase epitaxy, which results in the QD density of ∼4 × 10¹⁰ cm². In this work, we show that an In_0.1Ga_0.9As QW located in close proximity to the wetting-layer-free InAs QDs leads to an enhanced carrier injection into the QDs, allowing for lasing at room temperature (RT). Devices employing InAs QDs grown with and without In_0.1Ga_0.9As QW carrier collection layer exhibit lasing at 80 K, while only the lasers grown with the In_0.1Ga_0.9As QW exhibits lasing at RT. Driving current-dependent electroluminescence measurements reveals a low carrier injection into the QDs in the device grown without In_0.1Ga_0.9As QW carrier collection layer. In addition, the appearance of lasing emission, corresponding to high energy transitions, indicates a relatively flat gain profile for this InAs QD active region grown by SA-MOVPE. In addition, EL measurements at 80 K on devices employing varying thickness (d_InAs) QDs indicate the peak emission wavelength varies from 920 nm (d_InAs ∼ 1.5 nm) to 974 nm (d_InAs = 3 nm), although the devices employing d_InAs ∼ 3 nm did not exhibit lasing, possibly because of a significant degree of strain relaxation for this QD thickness.

In this work, we successfully synthesized Cr³⁺ doped ZnO nanoparticles using a sol-gel method, and elucidated how Cr³⁺ dopant is critical for enhancing photocatalytic activity. The nature of the point defect analyzed by electron spin resonance (ESR), and photoluminescence (PL) emission revealed the role of the Cr³⁺ dopant. When introducing Cr³⁺ ions in ZnO, the PL emission intensity decreased, indicating a reduction of the radiative recombination rate due to the heterojunction formation between the dopant and the host. The Cr³⁺ doped ZnO nanostructures showed that the typical ESR signal with g-factor value ∼1.96 was completely passivated, indicating the diffusion of electrons near the conduction band into the dopant ions. The doped Cr³⁺ ion acts as an electron trap in the ZnO crystal described as $C{r}^{3+}+{e}^{-}\to C{r}^{2+}.$ The mechanism for enhancing the photocatalytic activity of heterogeneous ZnO:Cr³⁺ was proposed in respect of point defect evolution through the manner of Cr³⁺ doping. As a result, the photocatalytic efficiency investigated by measuring methylene blue degradation under 210 min of direct sunlight irradiation reached 93.5% for 1 at % Cr³⁺ doped ZnO, which was significantly improved compared to 59.8% of the pure ZnO.

Tunnel junctions are indispensable elements of multi-junction solar cells. The fabrication of InGaN tunnel junctions requires the growth of degenerately doped n- and p-type layers. While highly doped n-type InGaN films have been demonstrated, the growth of degenerately p-doped InGaN films and the fabrication of high indium fraction InGaN tunnel junctions is still to be demonstrated. We present an investigation of the effect of Mg doping on the InGaN crystal properties over a large range of Mg fluxes and InN mole fractions in the range from 30% to 40%, using multiple characterization techniques. InGaN thin films were grown on GaN/sapphire templates and doped with Mg using plasma-assisted molecular beam epitaxy (PAMBE). We have found that the Mg concentration in the film increases linearly with the Mg beam equivalent pressure (BEP) at first, followed by a saturation at ∼4 × 10²¹ cm⁻³ similar to the Mg doping behavior reported for GaN. The growth rate of the alloy changes by more than 50% with the changes in the surface availability of Mg. These effects can be explained through the saturation of the atomic sites available for incorporation in the case of Mg concentration saturation and by the passivation of the free nitrogen radicals in the case of the growth rate variation. The incorporation of In and Ga depends on the flux ratio (Φ_In + Φ_Ga)/(Φ_Mg) at the growth surface and it is shown that the decrease of this ratio below a threshold of ∼2000 causes the almost complete loss of In and the formation of a new quaternary wide band gap semiconductor alloy (InGaMg)N.

We investigated the effects of electron-beam-irradiation (EBI) on amorphous indium gallium zinc oxide (IGZO) films deposited by RF magnetron sputtering. The device performance of thin film transistors (TFTs) fabricated from these films was also evaluated. We conducted transmission electron microscopy and x-ray photoelectron spectroscopy to analyze the microstructure and chemical state of the synthesized IGZO. The EBI-treated IGZO TFTs achieved higher carrier mobility and on/off ratio than conventionally annealed IGZO TFTs. In addition, we found that a dual-channel structure shows electrical characteristics superior to those of a single-channel structure, with a carrier mobility of 18.1 cm²/V · s.

A new numerical method for determining the reverse transition voltage between thermionic and tunneling mechanisms has been performed for 4H-SiC Schottky barrier diodes. The idea of this method is based on equality between thermionic emission and the tunneling process and both are combined with the barrier lowering model. Application of this method shows a strong discrepancy between our results and those deduced from Padovani-Stratton conditions. The reverse transition voltage versus the temperature plot exhibits an unexpected peak at low temperatures which means that the thermionic emission mechanism predominates at low temperatures. The reverse transition voltage increases linearly with the increase in barrier height, the effective mass and the inverse of doping concentration. In order to predict the reverse transition voltage as a function of temperature, doping concentration and barrier height for a 4H-SiC Schottky barrier diode, an analytical model has been proposed.