AVS1997 Session TF2-ThM: Ex-Situ Characterization II
Time Period ThM Sessions | Abstract Timeline | Topic TF Sessions | Time Periods | Topics | AVS1997 Schedule
Start | Invited? | Item |
---|---|---|
8:20 AM |
TF2-ThM-1 UV Raman Scattering Studies of Tetrahedral Amorphous Carbon.
V.I. Merkulov, J.S. Lannin (Pennsylvania State University); C.H. Munro, S.E. Asher (University of Pittsburgh); V.S. Veerasamy, W. Milne (University of Cambridge, United Kingdom) We report ultraviolet (UV) Raman scattering studies of hydrogen-free, diamond-like amorphous carbon thin films with a wide range of tetrahedral bonding. The UV Raman spectra are shown to provide direct evidence for the presence of sp3-bonded C atoms in these materials. The experimental results are found to be in excellent agreement with theoretical predictions and contribute to improved understanding of the mechanism by which the diamond-like fraction develops within the amorphous carbon network. |
|
8:40 AM |
TF2-ThM-2 Crystalline Orientation and Structure of WS2 Thin Films by MOCVD
J.W. Chung, Z.R. Dai, F.S. Ohuchi (University of Washington) The metal-organic chemical vapor deposition (MOCVD) of WS2 by the reaction of H2S with W(CO)6 was determined to be thermodynamically favored over a wide range of temperature, pressure, and precursor concentration conditions. Various degrees of crystalline WS2 films were obtained on Si and Al2O3, and their crystalline orientation, structure and morphology of thin films were studied in relation to the deposition conditions using x-ray diffraction, Raman spectroscopy, and transmission electron microscopy. WS2 thin films grew with their basal planes parallel to the interface (c(=)) for the first 20-30 nm near the interface, then further growth resulted in the formation of non-parallel crystallites (c(//)). A macroscopic growth mechanism can be explained by simple kinematical and geometrical considerations. High resolution lattice images from cross sections revealed a unique structure of lattice bending as well as stacking sequence in the transition region in which c(//) was developed from c(=). Raman spectroscopy exhibits both first and second order peaks having narrow FWHM in these films, suggesting a high degree of crystallinity. Because of the unique structure of thin films, an unusually large activity of the second order Raman processes was observed at 416 cm-1 for the first time. This is believed to be due to coupling oftransverse acoustic (TA) and longitudinal acoustic (LA) phonons at the K point of WS2. |
|
9:00 AM |
TF2-ThM-3 Deposition and Characterization of Amorphous Diamond Films on Mo Tips.
M.Q. Ding, A.F. Myers, J.J. Cuomo, J.J. Hren (North Carolina State University) To study field emission properties of amorphous diamond (a-D) and nitrogen containing a-D films, these films were deposited on sharp Mo tips by pulsed laser ablation and characterized by high resolution transmission electron microscopy (HRTEM) and electron energy loss spectroscopy (EELS). A KrF excimer laser ( 248 nm) was employed to deliver a power density of 25 J/cm2 to a graphite target at a repetition rate of 10 Hz and a duration of 20 ns. The preparation of amorphous diamond was conducted in a base pressure of 7.5 x 10-6 Pa, while nitrogen containing a-D films were deposited under nitrogen pressures of 0.067 and 0.67 Pa. Scanning TEM showed that all these films exhibited a smooth surface, whereas HRTEM revealed that these films had a unique nano-columnar structure, which varied from the shank to the apex. Interestingly, EELS data from a-D films indicated a continuous increase in sp3 fraction from the shank to the apex, which will be explained by relaxation effect of energetic ions/atoms associated with the incident angle of depositing flux to the surface normal. For N containing a-D films, EELS studies showed sp2 dominated nature, which may result from destruction of sp3 network due to the incorporation of high concentration nitrogen, as suggested by McKenzie et al. |
|
9:20 AM |
TF2-ThM-4 Mechanical, Optical and Structural Properties of MgF2 and TiO2 Thin Films and their Multilayers Deposited by Plasma Ion Assisted Deposition
G. Atanassov, J.K. Fu (Singapore Productivity and Standards Board); J.E. Turlo (KLA-Tencor); Y.S. Dai, F.H. Tan, Z.Q. Mo (Singapore Productivity and Standards Board) A study is done on TiO2 and MgF2 thin films and their multilayers deposited by plasma ion assisted deposition (PIAD) and conventionally in Leybold APS 904 vacuum system. Experimental results for stress measurements of single and multilayer films and AFM examination of film surface structure are reported. The experimental results for the studied films and coatings are discussed in terms of film columnar structure, surface free energy and sorption processes at the grain boundaries and columns. Both TiO2 and MgF2 deposited by PIAD have higher refractive indices and packing densities. MgF2 films deposited by PIAD have higher optical absorption, which can be substantially reduced by adding oxygen during deposition. Fluxes as low as 5 sccm have a better effect on optical performance than higher fluxes. AFM study corroborates that TiO2 film deposited by PIAD and conventionally have different structure. In both cases, however, the crystallization is incomplete, so annealing to 350 oC leads to further crystallization which is found to be more advanced in the PIAD case. All TiO2 films have tensile stress. The stress of TiO2 films deposited conventionally is influenced more rigorously by the air exposure than the PIAD films. Stress relaxation and stress-temperature behaviour are explained through a grain boundary mechanism. Immediately after the deposition, all MgF2 films show tensile stress. The stress value reduces gradually with time and after two days the stress of conventionally deposited MgF2 thin films becomes compressive, while the stress of all PIAD MgF2 films remains tensile. MgF2 thin films also crystallise at temperatures above 250 oC. The total stress in the multilayers depends in a complex way on the single film stress and on the coating design. When the total stress exceeds a value of about 350 MPa, the coatings tend to crack. By PIAD deposition, one can achieve lower total stress in the coating than by conventional deposition. |
|
9:40 AM | Invited |
TF2-ThM-5 Multi-Technique Characterization of WSix Films
S.M. Baumann (Evans Texas); C.J. Hitzman, R.H. Brigham, P. Lindley (Charles Evans & Associates); R.B. Ortega (AMIA Laboratories, Inc.) WSix films are used extensively for contact, interconnect, and in some cases, diffusion and Schottky barriers in semiconductor devices1. The electrical, adhesive, and barrier properties of these films are affected by a variety of factors, such as film stoichiometry, morphology, impurities, surface contamination, etc. This paper will address the capabilities and limitations of a variety of techniques which can be used to characterize WSix films. Techniques which were studied include: Dynamic and Static Secondary Ion Mass Spectrometry (SIMS), Rutherford Backscattering Spectrometry and Elastic Recoil Detection (RBS/ERD), Auger Electron Spectroscopy (AES), Field Emission Scanning Electron Microscopy (FE-SEM), Total Reflection X-Ray Fluorescence (TXRF), Atomic Force Microscopy (AFM), X-Ray Diffraction (XRD), and X-Ray Photelectron Spectroscopy (XPS). Film characteristics which were studied include surface morphology, grain structure and orientation, film stoichiometry and phase identification, bulk and interfacial impurity concentrations including metallic, atmospheric, and dopant impurities. Cross correlation between the techniques was performed whenever possible in order to compare the relative accuracy of the techniques and to identify error inducing artifacts associated with some of the measurement techniques.
|
10:20 AM |
TF2-ThM-7 Investigation of Borophosphosilicate Glass Defects with the Atomic Force Microscope
S.K. Tang (Nanyang Technological University, Singapore); V.Y. Vassiliev (Chartered Semiconductor Manufacturing Ltd., Singapore); S. Mridha (Nanyang Technological University, Singapore); L.H. Chan (Chartered Semiconductor Manufacturing Ltd., Singapore) Borophosphosilicate glass (BPSG) films have been widely used in the fabrication of Ultra Large Scale Integrated (ULSI) semiconductor devices as an interlayer dielectric between the gate and first metal level due to its excellent planarization and gettering properties1,2,3,4 . It is well known that BPSG films are prone to contain surface defects. These are believed to be due to the process chemistry and instability of the complex glass material. These defects are normally known as BPO4 crystals although many different shapes and sizes were observed5,6,7,8. In our previous works9,10 devoted to BPSG defect formation, we concluded that surface defects can exist in the liquid state and have proposed the formation and growth mechanism of these defects. Detailed investigation of both liquid and solid state defects in the first stage of formation is necessary in order to understand the nature and mechanism of BPSG defect formation. Conventional laser-based particle detectors cannot detect the quantity of surface defects below 0.2 microns. This method is used only for surface defects count and it is unable to yield results about the state and shape of small defects. On the other hand, detecting such defects with a scanning electron microscope (SEM) will not yield any results if such defects are in the liquid state. In this study, the capability of the atomic force microscope (AFM) to detect and characterize surface defects is demonstrated for the first time for both liquid and solid state defects on BPSG films. Defect sizes below 0.2 micron are successfully identified using the AFM in BPSG films of various dopant compositions deposited using sub-atmospheric chemical vapor deposition (CVD), atmospheric pressure CVD, low pressure CVD and plasma-enhanced CVD. The possibility of quantitative characterization of defects and their changes with time are studied. Issues regarding the application of the atomic force microscope on BPSG liquid and solid state defects and BPSG surface roughness will be discussed.
|
|
10:40 AM |
TF2-ThM-8 Atomic Force Microscopy Studies of Substrate-Dependent Nucleation and Selective Deposition of PECVD Silicon
L.L. Smith, W.W. Read, E. Srinivasan, G.N. Parsons (North Carolina State University) Nucleation of microcrystalline silicon (µc-Si) during a selected-area, plasma-enhanced chemical vapor deposition (PECVD) process was investigated using atomic force microscopy (AFM). Selective deposition of µc-Si is accomplished by using a time-modulated pulsed gas process, in which the deposition of thin layers of silicon alternates with exposure to hydrogen plasma. It is believed that the substrate selectivity results from the etching and removal of weak, strained Si nuclei by the hydrogen plasma. In this paper, surface morphology in the early stages of growth is characterized by AFM on a variety of substrate surfaces. Three-dimensional nucleation of µc-Si with nuclei approximately 120 Å in size are clearly observed after 15 s of deposition, whereas nucleation on glass is much less regular and shows a more 2-D structure. A significant problem for device applications of selective µc-Si deposition is that amorphous silicon (a-Si) is typically etched by atomic hydrogen exposure, making selective deposition of µc-Si on a-Si impossible by the pulsed-gas technique. We have developed a technique involving Mo metallization that stabilizes the a-Si surface with respect to hydrogen plasma exposure and allows selective µc-Si deposition on a-Si in device structures. Surfaces and subsequent selective growth were characterized using AFM, x-ray photoelectron spectroscopy (XPS), and Auger electron spectroscopy (AES), which revealed the presence of Mo incorporation in the a-Si surface remaining after complete removal of the metal layer. The roles of surface modification and surface energetics in preferential etching and selective deposition will be discussed. |
|
11:00 AM |
TF2-ThM-9 STM of GaAs Heteroepitaxy on Ge(001)
S. Gan, L. Li, B.K. Han, R.F. Hicks (University of California, Los Angeles) Understanding and controlling heteroepitaxy of GaAs on Ge(001) is important to solar cell manufacture. Using combined STM with XPS and LEED, we have studied the epitaxial process of GaAs on Ge: substrate cleaning, As-passivation, and GaAs film growth. Vicinal Ge(001) (9° off axis) substrates employed in fabricating GaAs/Ge solar cells were first cleaned in HF and H2O2 solutions. It is found that this treatment can easily introduce carbon impurities to the surface which produces rough morphology unsuitable for GaAs heteroepitaxy although it can also generate a single domain structure after removing all the carbon. Our STM results have shown that the carbon on the surface pinned steps to create a faceted surface covered with v-shape ridges. To obtain an atomically clean and smooth surface, we have developed a new procedure: annealing germanium substrates in flowing tertiarybutylarsine(TBAs) and H2 at 948K for 30min in an MOVPE reactor after chemical cleaning. Atomic resolution STM images of such surfaces revealed a single domain structure passivated by As with arsenic dimer rows running parallel to step edges, which is ideal for further GaAs film growth. Finally, GaAs film of 0.5 um thick was grown on this As-passivated surface. Depending on growth conditions, single crystal films with different surface morphology have been obtained. |
|
11:20 AM |
TF2-ThM-10 Comparison of Mass Spectroscopy of Recoiled Ions (MSRI) with SIMS for Trace Analysis of Na, Al, Cu and Fe Contaminants on Si(100) and Ge(100) Surfaces
M. VanStipdonk, J.A. Schultz, K. Eipers-Smith, K. Waters (Ionwerks) A newly developed reflectron time of flight mass spectrometer has been shown useful for either detection of 10 eV secondary ions (SIMS) or of keV directly recoiled surface atoms ionized in direct binary collision sequences with the impinging primary ion beam. The resulting mass spectrum of recoiled ions (MSRI) comprise elemental ions and their isotopes and are devoid of isobaric molecular ions. Resolutions are typically around 200 which is sufficient for trace element analysis since no interferring molecular isobars survive at the keV energies of the recoils. Our data on Fe from silicon is particularly illustrative of this point since the silicon dimer ion is destroyed in MSRI whereas in SIMS it is a major interference eliminated only with high resolution mass spectrometers. Data will show the powerful symbiosis of the two techniques. |
|
11:40 AM |
TF2-ThM-11 Fourier Deconvolution of the Spectral Components of Electromodulation Spectra
R.T. Holm, O.J. Glembocki, B.T. Jonker, R.J. Wagner, D.S. Katzer (Naval Research Laboratory) Electromodulation spectroscopies such as photoreflectance (PR), contactless electroreflectance (CER) and electroreflectance (ER) are playing an increasingly important role in the optical characterization of the performance of GaAs based devices, such as Schottky contacts, heterojunction bipolar transistors (HBT’s) and high electron mobility transistors (HEMT’s). Because of their contactless nature, PR and CER can be applied at various stages of device processing as well as to the final device structure. Modern GaAs based devices typically utilize epitaxial growth of several different layers to achieve the desired performance. Because of overlapping optical transitions, the optical electromodulation spectrum of such a structure can be complicated even in the simplest of cases. In this paper, we examine these effects on the Franz Keldysh oscillations (FKO) of undoped GaAs layers grown on n+GaAs and their Schottky contacts with Fe. The FKO, arising from transitions at the band gap, are typically used in measuring internal fields and allow one to optically determine the barrier height in Schottky contacts. In high quality materials, however, the FKO can extend to very high energies, well beyond the onset of the next highest optical transition. The resulting overlap and contributions from the heavily doped GaAs substrate complicate the FKO analysis, introducing uncertainties in the determination of the electric fields. We show that Fourier decomposition of the ER and PR data can be utilized to isolate the various components of the electromodulation spectrum, simplifying subsequent analysis. Applications to more complex systems, such as HBT’s and HEMT’s will be discussed. |