Emily L. Allen

High magnetoresistance permalloy films deposited on a thin NiFeCr or NiCr underlayer

Significantly enhanced anisotropic magnetoresistance (MR) in permalloy (Ni0.81Fe0.19) films deposited on a thin (Ni0.81Fe0.19)1−xCrx or Ni1−xCrx underlayer is reported. The maximum ΔR/R enhancement was observed using the underlayer with x≈0.44 at an optimum thickness of ≈30–45 A, depending on the deposition technique. An enhancement of 75%–150% was observed for 45–430 A thick permalloy films, compared to the films deposited without this underlayer. The ΔR/R enhancement is attributed to the decrease in the resistivity ρ and the increase in Δρ of the permalloy film due to the formation of large (111) textured crystal grains in the permalloy films deposited on this underlayer, as revealed by the x-ray diffraction results obtained using synchrotron radiation.Significantly enhanced anisotropic magnetoresistance (MR) in permalloy (Ni0.81Fe0.19) films deposited on a thin (Ni0.81Fe0.19)1−xCrx or Ni1−xCrx underlayer is reported. The maximum ΔR/R enhancement was observed using the underlayer with x≈0.44 at an optimum thickness of ≈30–45 A, depending on the deposition technique. An enhancement of 75%–150% was observed for 45–430 A thick permalloy films, compared to the films deposited without this underlayer. The ΔR/R enhancement is attributed to the decrease in the resistivity ρ and the increase in Δρ of the permalloy film due to the formation of large (111) textured crystal grains in the permalloy films deposited on this underlayer, as revealed by the x-ray diffraction results obtained using synchrotron radiation.

A Comparison of the Diffusion Behavior of Ion‐Implanted Sn, Ge, and Si in Gallium Arsenide

The diffusion behavior of ion-implanted Sn, Ge, and Si in GaAs was investigated as a function of implant dose, temperature, and background doping. Sn and Ge were found to have diffusivities comparable to those of the same species int roduced by doping during growth or from surface diffusion. The diffusivity of Si and to a lesser extent that of Ge was very sensitive to implant conditions, and both exhibit dose and t ime dependence. The different dose dependencies -for the three Group IV dopants may indicate the effects of implant damage. Activation energies for diffusion were extracted for the three dopants. Transmission electron micrographs were examined and possible correlations between diffusion behavior and extended defect structure are made. Understanding the diffusion behavior of ionimplanted dopants in GaAs has become more important as device dimensions shrink. For many devices made in GaAs, dopant diffusion during the post-implant anneal is undesirable. One of the motivations for using Si implants for channel, source, and drain doping of GaAs MESFETs (1) is that Si diffuses very little during the activation anneal. However, in certain dose and energy regimes, even implanted dopants undergo diffusion during processing, particularly in doped substrates or where there is more than one implant. The effect of dislocations and extended defects on threshold voltages in MESFETs has been investigated (2), but there is little agreement on their effect on either activation or diffusion. In addition, the observation of impurityinduced superlattice disordering (3) has emphasized the need to understand diffusion mechanisms better in both GaAs and A1GaAs. In this work we examine the diffusion behavior of the n-type Group IV dopants Sn, Ge, and Si, implanted into doped and semi-insulating GaAs substrates. We measured their diffusivities and extracted activation energies for diffusion, and compare our values to literature values for the same dopants introduced into the lattice from vapor or solid sources (solid-source diffusivities). We have deliberately used high-dose implants and high-temperature, long-time furnace anneals in order to provide conditions under which these dopants exhibit significant diffusion. We present cross-sectional transmission electron micrographs (XTEM) showing the defect morphologies of some implanted GaAs substrates, and discuss possible correlations between diffusion behavior and extended defect structure. We also present the results of carrier measurements. The diffusivity of a dopant in a semiconductor depends on the diffusivity and concentration of vacancies and/or host interstitials. When the dopant is introduced into the lattice from a vapor source or a solid source such as a thin film, it is generally assumed that the concentrations of point defects are at their equilibrium values and that the diffusion process is an equilibritlm process, governed by the kinetics associated with migration of both the point defect and the dopant atom. After ion implantation, however, neither the dopant nor the point defect concentrations are at equilibrium. The diffusion process in that case may be governed by the kinetics associated with point defect production and/or annihilation. This could result in different * Electrochemical Society Active Member. diffusivities for implanted dopants vs. solid source dopants. The extended defects (dislocations) which form as a result of the implantation process are thought to act as sources and sinks of point defects (4). Our diffusion results suggest that this residual defect morphology determines the extent to which the implanted ion diffusivity approaches the solid-source ion diffusivity. The ability of the lattice and its extended defect structure to provide the necessary point defects for diffusion appears to determine the relative diffusivity of the implanted dopant. Our experiments also suggest that the excess point defects which are initially created by the implantation process recombine very rapidly in the case of n-type dopants in GaAs, so that we observe no transient diffusion effects attributable to excess point defects. This is in contrast to the transient diffusivities observed for implanted p-type dopants in GaAs (5) or the transient diffusion seen in boron-implanted silicon (6). In contrast, the effect of the residual damage (extended defects) is long-range and persists for very long t ime anneals. The diffusion behavior of solid-source Sn (7-10) and Si (11, 12), and to a lesser extent that of Ge (13, 14) has been discussed in the literature. Diffusivity values for ionimplanted Si have also been reported (15). However; most of the available quantitative diffusivity values are for dopant introduced either from solid or vapor sources. To our knowledge, these are the first reported diffusivity values and activation energies for ion-implanted Ge and Sn in GaAs.

Reduction of Resistivity in Cu Thin Films by Partial Oxidation: Microstructural Mechanisms

We report on the electrical resistance and microstructure of sputter deposited copper thin films grown in an oxygen containing ion-beam sputtering atmosphere. For films thinner than 5 nm, 6%–10% oxygen causes a minimum in film resistivity, while for thicker films, there is a monotonic increase in resistivity. X-ray reflectivity measurements show significantly smoother films for these oxygen flow rates. X-ray diffraction shows that the oxygen doping causes a refinement of the copper grain size and the formation of cuprous oxide. We suggest that the formation of cuprous oxide limits copper grain growth, which causes smoother interfaces, and thus reduces resistivity by increasing specular scattering of electrons at interfaces.

Effect of implant temperature on dopant diffusion and defect morphology for Si implanted GaAs

Experimental observations of dopant diffusion and defect formation are reported as a function of implant temperature in Si implanted GaAs. The diffusion of Si during post‐implant annealing decreases by a factor of 2.5 as the implant temperature increases from −2 to 40 °C. In this same temperature range, the maximum depth and density of extrinsic dislocation loops increase by factors of 3 and 4, respectively. Rutherford backscattering channeling measurements indicate that Si implanted GaAs undergoes an amorphous to crystalline transition at Si implant temperatures between −51 and 40 °C. A unified explanation of the effects of implant temperature on both diffusion and dislocation formation is proposed based on the known differences in sputter yields between crystalline and amorphous semiconductors. The model assumes that the sputter yield is enhanced by amorphization in the lower temperatures, thus increasing the excess vacancy concentration. Estimates of excess vacancy concentration are obtained by simulat...

Teaching design of experiments and statistical analysis of data through laboratory experiments

A new laboratory course at San Jose State University, Advanced Thin Film Processes, integrates fabrication of thin films with design of experiment and statistical analysis of data. In the laboratory section of this course, students work through six multi-week modules that increase in the complexity of design of experiment and statistical analysis of data. The six modules have been developed with a standard format that includes learning objectives, background on the specific thin film process, theory of design of experiment principles, instructor notes, dry lab exercises, experimental plan worksheets, and assessment tools. While the modules were developed for a semiconductor processing class, they can easily be implemented in other engineering classes. The modules have been developed with a robust framework that allows the instructor to teach design of experiments and statistical analysis of data along with the specific engineering principles related to their class.

Microstructural comparisons of ultrathin Cu films deposited by ion-beam and dc-magnetron sputtering

We report and contrast both the electrical resistance and the microstructure of copper thin films deposited in an oxygen-containing atmosphere by ion-beam and dc-magnetron sputtering. For films with thicknesses of 5 nm or less, the resistivity of the Cu films is minimized at oxygen concentrations ranging from 0.2% to 1% for dc-magnetron sputtering and 6%–10% for ion-beam sputtering. Films sputtered under both conditions show a similar decrease of interface roughness with increasing oxygen concentration, although the magnetron-deposited films are smoother. The dc-magnetron-produced films have higher resistivity, have smaller Cu grains, and contain a higher concentration of cuprous oxide particles. We discuss the mechanisms leading to the grain refinement and the consequent reduced resistivity in both types of films.

Microelectronics process engineering program at SJSU

At present, there is a need for engineers with CMOS processing knowledge, statistical process control (SPC) skills, and the ability to work in an interdisciplinary team environment and assume leadership roles. San Jose State University are developing an interdisciplinary lab-based microelectronics process engineering program that introduces SPC and DOE to students in a microelectronics manufacturing environment. At the heart of the program are three courses, each of which is imagined to be a division of a fictitious semiconductor fabrication company (Spartan Semiconductor Services, Inc., or S3i). The divisions are: Digital NMOS division (MatE/EE129: Introduction to IC Fabrication), Thin Film Research Division (MatE/ChE 166: Advanced Thin Film Processes) and CMOS Division and SPC task force (MatE/EE 167: Microelectronics Manufacturing Methods). Several unique features of the program are its introduction of SPC in a microelectronics manufacturing environment, the inclusion of design of experiments (DOE) topics, and the faculty-faculty, faculty-student and student-student interaction among the three courses (divisions). Ultimately, we are trying to provide a learning environment that will allow our students to be immediately productive in an IC production facility, to be able to communicate with IC process engineers, and to be prepared for graduate school programs.

Oxygen-enhanced IrMn spin valves deposited by ion-beam and magnetron sputtering

Enhancement of giant magnetoresistance properties of single (bottom) and dual IrMn-based spin valves through exposure of part of the CoFe pinned layer to O2 is reported. Under optimal conditions, a ΔR/R of 10.4% [Hua=460 Oe, Hf=5.1 Oe, and Hc=4.7 Oe for a free and pinned layer thickness (permalloy equivalent) of 50 A each] for an ion beam sputtered single spin valve, and a ΔR/R of as high as 20.5% for a magnetron sputtered dual spin valve having a 30 A thick CoFe free layer are observed, compared to a value of 6.5% and 10.6% for the corresponding spin valve without O2 exposure, respectively. Transmission electron microscopy results reveal the presence of a thin (10 A) crystalline oxygen-containing layer near the IrMn–CoFe pinned layer interface as a result of O2 exposure. X-ray reflectivity data show smoother interfaces for the spin valves subjected to O2 exposure, consistent with the lower Hf and smaller sheet resistance observed for these samples. The enhanced ΔR/R thus can be attributed to improved gro...

Microelectronics Process Engineering: A Non-Traditional Approach to MS&E

Materials Science and Engineering straddles the fence between engineering and science. In order to produce more work-ready undergraduates, we offer a new interdisciplinary program to educate materials engineers with a particular emphasis on microelectronics-related manufacturing. The bachelors level curriculum in Microelectronics Process Engineering (μProE) is interdisciplinary, drawing from materials, chemical, electrical and industrial engineering programs and tied together with courses, internships and projects which integrate thin film processing with manufacturing control methods. Our graduates are prepared for entry level engineering jobs that require knowledge and experience in microelectronics-type fabrication and statistics applications in manufacturing engineering. They also go on to graduate programs in materials science and engineering. The program objectives were defined using extensive input from industry and alumni. We market our program as part of workforce development for Silicon Valley and have won significant support from local industry as well as federal sources. We plan to offer a vertical slice of workforce development, from lower division engineering and community college activities to industry short courses. We also encourage all engineering majors to take electives in our program. All our course and program development efforts rely on clearly defined learning objectives.

Using learning styles preferences data to inform classroom teaching and assessment activities

This paper describes the results to date, and our ongoing efforts, to develop, adapt, disseminate and assess teaching techniques, methods and exercises which address specific learning styles of students. Our research shows that students in the engineering college showing strong preferences for the active, sensing visual and sequential learning styles outweigh students with strong preferences for the reflective, intuitive, verbal and global learning styles by significant factors. As part of our faculty instructional development program, we have been using techniques such as cooperative and active learning as well as entrepreneurial environments in the laboratory to help move the instructional experience of students away from the traditional lecture mode and into a more student-centered mode.