Stencil mask ion implantation technology

The ion implantation process is important for the development or manufacturing of semiconductor devices, because ion implantation conditions directly influence some characteristics of semiconductor devices. Recently, we developed a new implantation technology, stencil mask ion implantation technology (SMIT). In the SMIT system, the stencil mask acts like a resist mask, and ions passing through the mask holes are implanted into selected regions of the Si substrate chip by chip. Use of SMIT has several advantages, notably lower manufacturing cost and shorter process time than in the case of conventional processing, because no photolithography process (including deposition and stripping of resist) is required. We have already demonstrated an application of SMIT to transistor fabrication, using various implanted dose conditions for the same wafer. Threshold voltage values can be controlled as effectively by implanted doses as they can by conventional implantation, and the dose dependence of the threshold voltage could be obtained from one wafer to which various implantation conditions are applied. Using SMIT, implantation conditions can be changed chip by chip without additional processes. This flexibility of implantation conditions is another advantage of SMIT. In this paper, we propose stencil mask ion implantation technology and show some fundamental results obtained by applying SMIT     

高速双极器件制造中的离子注入技术

本文讨论了高速双极器件制造中的离子注入技术。重点讨论在双极器件制造工艺中的离子注入掺杂技术,离子注入形成浅结技术,以及离子注入层的退火技术。根据这些要求,对离子注入设备应具有的技术性能作了一些预测。     

Current transport in an ion-implanted diode

In a Schottky diode, the diode saturation current is controlled by the barrier height at the metal and semi-conductor contact, assuming that the dominant current is due to thermionic emission. When ion implantation is used to increase the barrier height, both thermionic emission and drift-diffusion of carriers become important in calculating the current. Numerical methods are used in solving Poisson's equation and the current continuity equations for an ion implanted doping profile. The electron and hole current in the surface region are calculated as a function of the total implantation dosage. The results show that the decrease of saturation current and the increase of effective barrier height in an ion implanted diode is mainly due to the suppression of the thermionic emission current by the implanted impurity atoms, rendering the diode to act like a pn junction.     

Tellurium-implanted N+ layers in GaAs

Ion implantation of Te was investigated as a doping process for the fabrication of submicron n-type layers in GaAs. The implantation was performed with substrates held at 350°C. After implantation, a protective overcoat of AIN or Si₃N₄ was sputtered on the samples to prevent the GaAs from disassociating during anneal (900°C). The electrical characteristics of the n-type implants were then measured. Current-voltage and capacitance-voltage characteristics of implanted diodes indicated that the junctions were linearly graded and that there was no intrinsic layer present after anneal. Sheet resistivity and Hall effect measurements were used to determine the surface carrier concentration and effective mobility in the implanted layers. Ionized impurity profiles extending beyond the implanted junction depth were calculated by matching differential Hall effect data with junction capacitance-voltage data. A peak electron concentration of 7 × 10¹⁸ electrons/cm³ was observed. However, the profiles exhibited penetrating tails that resulted in junction depths being much deeper than the LSS range theory would predict.     

Multiple ion implantation profile using differential channel dose

Multiple ion implantations are frequently used, especially for extension regions in high-speed MOSFETs, to ensure symmetrical doping profiles. In the process simulation, each process step is treated independently and the final impurity concentration after the multiplied ion implantation is linearly multiplied by the number of first-implantation processes. However, as the channeling tail of the ion implantation profile is saturated in high dose regions, the simple multiplication of the profiles induces artificial deeper junction depths. To solve this problem, we introduced a differential channel dose, which enables us to generate accurate ion implantation profiles, and here we will demonstrate that the conventional treatment of the multiple ion implantations predicts worse short channel effects especially for nMOSFETs.     

The effect of boron,oxygen, and fluorine in ion implanted GaAs

Effects of boron, fluorine, and oxygen in GaAs have been investigated by electrical characterization using current-voltage, capacitance-voltage and deep level transient spectroscopy techniques. Ion implantation at 100 keV energy was conducted with doses of 10¹¹ and 10¹²/cm². Carrier compensation was observed in each implanted sample. The compensation effect strongly depended on ion implantation condition and ion species. More free carriers were compensated for higher dose and heavier species; however, severe surface damage would also be induced that degrade electrical performance. Rapid thermal annealing treatment showed the heavier ion implanted samples to be more thermally stable. Defect levels for each implanted species were compared and identified. A native shallow defect E4 was easily removed by ion implantation. In higher dose and heavier ion implantation, both electron and hole traps were induced. However, in some cases, heavier ion implantation also removed native defects. Acceptor-type surface states were created by implantation that degrade material electrical characteristics and also reduce the effect of compensation. The damage induced traps were mostly point-defects or point-defect/impurity complexes as evidenced by sensitivity to heat treatment.     

FEDSS—Finite-element diffusion-simulation system

The FEDSS program simulates semiconductor processes in two dimensions. An accurate model of the diffusion of impurity atoms into a substrate is necessary to assess the effects of process changes on impurity profiles. The process steps to be modeled include ion implantation, oxidation/drive-in, chemical predeposition through the surface, and oxide deposition. The finite-element method transforms the diffusion equation for impurity atoms to a simulation model at a discrete number of points. Direct techniques are used to solve the resulting matrix equations. The impurity distributions resulting from sequences of the process steps are shown.     

Hydrogen etching for semiconductor materials in plasma doping experiments

The etching effects of hydrogen plasma for semiconductor materials including single crystalline silicon, polycrystalline silicon, silicon dioxide, and aluminum in plasma immersion ion implantation (PHI) doping experiments have been investigated. Etching can alter device structure and affect implant profile and dose. The effects of varying different PIII process parameters such as pulse potential, pulse repetition frequency, and substrate temperature are presented. The experimental data show that the spontaneous etching by hydrogen radicals enhanced by ion bombardment is responsible for the etching phenomena that occurs at the material surface. A model is used to calculate the retained implant dose and impurity profile when the etching effect is considered.     

Two-dimensional low concentration boron profiles: Modeling and measurement

To model small devices successfully and to establish valid design rules for VLSI technology, two-dimensional (2-D) impurity profiles must be characterized. New analytical 2-D process models for ion implantation and impurity diffusion have been developed, and they include detailed considerations of the geometry and ambient effects. An approximate expression is derived to model the diffusion of low-concentration impurities under local-oxidation conditions, which can be conveniently used in the 2-D device analyses of small-geometry semiconductor devices. Measured results for the NMOST channel-stop (CS) diffusions show good agreement with the analytical calculations, and applications of the models to optimize 2-D profiles in NMOST are discussed.     

离子注入制备n型SiC欧姆接触的工艺研究

由于杂质在SiC中的扩散系数很低,所以制备SiC器件欧姆接触所需要的高掺杂区必须采用离子注入技术.本文采用蒙特卡罗模拟软件TRIM模拟得出不同能量不同剂量六次注入杂质的纵向分布图,根据大量文献数据和国内现有的制备水平,提出制备n型SiC欧姆接触最优的工艺流程.     

Submicrometer poly-Si CMOS fabrication with low-temperature laserdoping

The laser doping process for submicrometer CMOS devices with leakage currents as low as 10^-12 A/μm for both n-channel and p-channel devices is discussed. The I-V characteristics are comparable to those of poly-Si devices fabricated using ion implantation and high-temperature annealing processes. The laser-induced melting of predeposited impurity doping (LIMPID) process was used to fabricate submicrometer polycrystalline-Si CMOS devices. This process uses a very low temperature, so no dopant atom can diffuse along the grain boundaries in the solid region. The use of stacked Al/SiO₂ films as a protection layer made it possible to reduce the leakage current from several tens of picoamperes per micrometer to 1 pA/μm     

Optimal summation of Gaussians for ion implantation profile control

A method is described for selecting appropriate ion implantation energies and fluences to synthesize desired impurity profiles in semiconductor device fabrication. An optimization routine utilizing Rosenbrock's algorithm allows computation of a set of implantation parameters which result in close approximation to the preselected distribution.     

Experimental evaluation of depth-dependent lateral standarddeviation for various ions in a-Si from one-dimensional tiltedimplantation profiles

This paper presents an experimental evaluation of the lateral standard deviation for various ions implanted in amorphous silicon (a-Si) with a simple extraction method using no complicated structures. First, we derive a model for the tilted implantation profile as a function of both tilt angle and lateral standard deviation, assuming a Gaussian lateral distribution function. This model is based on the assumption that two-dimensional (2-D) ion implantation profiles can be constructed from lateral and vertical distribution functions which are independent of each other, and it enables us to extract lateral standard deviation by only evaluating one-dimensional (1-D) (vertical) impurity profiles. Next, we systematically measure the ion implantation depth profiles at various tilts (0-60°) with high resolution using secondary ion mass spectrometry (SIMS) and apply our proposed model for arsenic (As), phosphorus (P), antimony (Sb), and boron (B) ion implantations in a-Si over a wide energy range (20-160 keV) with a fixed dose of 1×10¹⁴ cm^-2. We successfully estimated not only average lateral standard deviation but also its depth dependence. Despite the simplicity of the model, the extracted depth-dependent lateral standard deviation shows good agreement over a wide energy range with the reported data calculated by theory or simulations. It is also shown that the lateral standard deviation has a linear depth dependence, and the lateral spread increases with the increase of depth for As, P, and Sb; on the other hand, it decreases for B, which reflects the difference of atomic mass between the incident ions and the target atoms     

Ion implantation of advanced silicon devices: Past,present and future

Ion implantation has been a key enabler, along with improvements in lithography, for the 40+ year evolution of MOS and then CMOS devices. Alterations in the channel doping levels followed the template developed by Dennard in the mid-70's from feature sizes of mm to tens of nm. When increasing channel doping in bulk planar CMOS created unacceptably high leakage current problems, ion implantation process developed past the "End of the Roadmap" to the "geometry-controlled" channels of fully-depleted finFETs and FDSOI of the present day. Future applications for nm-scale devices call for new understanding of ion damage accumulation in fin and nano-wire materials, consideration of effects of quantum confinement on channel conductivity, development of new implantation tools for efficient operation in the 100 eV range with ion and neutral species and soon after, for single-ion doping for quantum entangled, "atomic" electronics.     

Suppression of reverse biased diode leakage by MeV ion implantation

MeV Ion implantation has proven useful for many applications, such as latch-up suppression, SER reduction, and buried layer formation. MeV ion implantation can also be used to form a minority carrier diffusion barrier to reduce reverse bias diode leakage, particularly at high temperatures. The reduction in diode leakage has applications in DRAMs, CCDs, etc. The use of a MeV implanted diffusion barrier improves the ability to scale DRAM cell capacitance     

Current transport in fluorine implanted GaAs

Effects of fluorine implantation in GaAs have been investigated by electrical characterization. Ion implantation at 100 keV energy was conducted with doses of 10¹¹ and 10¹²/cm². The effect of fluorine implantation on current-voltage (I-V) characteristics of Schottky diodes was significant. Carrier compensation was observed after implantation by the improved I-V characteristics. The lower dose implanted samples showed thermionic emission dominated characteristics in the measurement temperature range of 300 to 100K. The starting wafer and the low dose implanted samples after rapid thermal annealing (RTA) showed similar I-V properties with excess current in the lower temperature range dominated by recombination. The higher dose implanted samples showed increased excess current in the whole temperature range which may result from the severe damage-induced surface recombination. These samples after RTA treatment did not recover from implantation damage as in the low dose implantation case. However, very good I-V characteristics were seen in the higher dose implanted samples after RTA. The influence of the higher dose ion implantation was to produce more thermal stability. The results show the potential application of fluorine implantation in GaAs device fabrication.     

Analysis and Improvement of Intermodulation Distortion in GaAs Power FET's

Tailoring of the doping profile is a powerful tool in reducing the intermodulation distortion (IMD) in GaAs power FET's. Reproducible and uniform preparation of the required profiles is a difficult task for epitaxial techniques. This shortcoming has motivated the present investigation of fabricating highly linear power FET's by ion implantation. An analytical divice model was developed for exploring the relationship between the active layer profile and the IMD. These calculations revealed a complex behavior in the variation of the distortion levels due to partial correlation between the contributions arising from nonlinear transconductance and output conductance. The device model was used to identify implant doses and energies for approaching an optimum active layer profile. Based on the results, a deep Se implant followed by a shallow compensating Be implant to reduce the doping level close to the surface was used in the device fabrication. The IMD of the transistors was measured by the two-tone method. Conventional epitaxial FET's with a flat doping profile were evaluated for comparison purposes. This comparison demonstrated that a 4-dB increase in the intercept point for the third-order intermodulation product can be realized by using the tailored implanted profile. The experiments demonstrated that the tuning conditions for maximum output power and minimum IMD are virtually identical for the implanted transistors, in contrast to the behavior of conventional devices with flat doping profiles. These performance advantages, coupled with the high levels of uniformity and reproducibility of doping parameters, show ion implantation to be a powerful technique in the fabrication of highly linear power FET's.     

激光束诱导电流谱技术表征红外器件几何结构

硼离子注入p型HgCdTe薄膜材料制备的光伏型红外探测器光敏元实际面积对器件结构的优化设计、器件性能研究具有重要意义．通过高分辨率、非直接接触的激光束诱导电流谱表征手段获取HgCdTe光伏型器件单元的几何结构信息,结果表明离子注入形成的n型区域的面积明显大于离子实际注入的区域面积。     

离子注入模型

本文结合我们的科研工作,阐述了半导体器件工艺模拟中的离子注入模型,并指出,由于离子注入具有能精确控制工艺参数,并能在较大面积上形成薄而均匀的掺杂层等特点,因此在工艺模拟中选用合适的离子注入模型对于提高工艺模拟的精确性是十分重要的。     

Ion-Implantation Control of Ferromagnetism in (Ga,Mn)As Epitaxial Layers

Epitaxial layers of (Ga,Mn)As ferromagnetic semiconductor have been subjected to low-energy ion implantation by applying a very low fluence of either chemically active, oxygen ions or inactive ions of neon noble gas. Several complementary characterization techniques have been used with the aim of studying the effect of ion implantation on the layer properties. Investigation of their electrical and magnetic properties revealed that implantation with either O or Ne ions completely suppressed both the conductivity and ferromagnetism in the layers. On the other hand, Raman spectroscopy measurements evidenced that O ion implantation influenced optical properties of the layers noticeably stronger than did Ne ion implantation. Moreover, structural modifications of the layers caused by ion implantation were investigated using high-resolution x-ray diffraction technique. A mechanism responsible for ion-implantation-induced suppression of the conductivity and ferromagnetism in (Ga,Mn)As layers, which could be applied as a method for tailoring nanostructures in the layers, is discussed in terms of defects created in the layers by the two implanted elements.