Harland G. Tompkins

Titanium nitride oxidation chemistry: An x‐ray photoelectron spectroscopy study

We report a study of the oxidation of TiN. In previous work, the oxidation kinetics for 350–450 °C were reported and an initiation time prior to fast oxidation was identified. In this study, x‐ray photoelectron spectroscopy was used to investigate the oxidation mechanisms at 350 °C during this initiation time period. The oxide thickness increases slowly with oxidation time and the film appears to change from an amorphous TiO2 layer to a crystalline TiO2 layer. Spectral features which are intermediate between TiO2 and TiN are reported and a model involving grain boundary oxidation is proposed. One of the thicker oxides studied was annealed in vacuum to 700 °C. Following oxidation, some of the capping oxide and much of the intermediate material is no longer in the analysis volume and we suggest that the oxygen and nitrogen is being dissolved into the bulk in much the same way that nonevaporable getters are activated before use.

DETERMINING THICKNESS OF THIN METAL FILMS WITH SPECTROSCOPIC ELLIPSOMETRY FOR APPLICATIONS IN MAGNETIC RANDOM-ACCESS MEMORY

Metal thicknesses can be determined using optical methods such as spectroscopic ellipsometry when the thickness of the film is small enough that the light penetrates to the bottom of the film. In order to use ellipsometry, it is necessary to determine the optical constants of the film material accurately. In this work, we describe the determination of the optical constants and thicknesses of single metal layers of NiFe, Cu, Co, Ta, FeMn deposited with two different deposition methods. The thicknesses range from about 6 to about 20 nm thick. We have also investigated the ex situ measurement of multiple metal layers, and discuss the conditions where the thicknesses of the individual layers can be determined and the conditions where they cannot be distinguished from each other.

Properties of WSix using dichlorosilane in a single‐wafer system

Chemical vapor deposition of tungsten silicide (WSix) from WF6 and SiH2Cl2 [JB Price, S. Wu, Y.Chow, and J. Mendonca, Semicon West (1986)] at higher deposition temperatures (450–650 °C) than the conventional WF6 and SiH4 (250–400 °C) process has been characterized using a plasma enhanced, single‐wafer, cold‐wall, radiantly heated system with temperature control utilizing a thermocouple in contact with the backside of the wafer. Film properties such as silicon to tungsten ratio, fluorine and chlorine concentration, resistivity, and film stress were studied as a function of substrate temperature, reactant composition, and flow rates. The film composition was measured by Rutherford backscattering spectrometry. The silicon to tungsten ratio is a function of deposition temperature at a fixed flow (x varying from 2.0–2.8 through the temperature range of 450–650 °C). The as‐deposited resistivity is also a strong function of deposition temperature. The chlorine and fluorine distributions in the WSix film were mea...

ANALYSIS OF SILICON OXYNITRIDES WITH SPECTROSCOPIC ELLIPSOMETRY AND AUGER SPECTROSCOPY, COMPARED TO ANALYSES BY RUTHERFORD BACKSCATTERING SPECTROMETRY AND FOURIER TRANSFORM INFRARED SPECTROSCOPY

This work addresses the issues of whether spectroscopic ellipsometry, using the effective medium approximation (SE-EMA), may be used meaningfully to analyze plasma-enhanced chemical vapor deposition silicon nitride films. We use Rutherford backscattering spectrometry and Fourier transform infrared spectroscopy as reference methods and compare the results to the results of SE-EMA analyses and Auger analyses. The results are that Auger analysis, using properly determined sensitivity factors, gives compositions which are within the uncertainty of the reference methods. SE-EMA, on the other hand, always overestimates the oxide contribution and underestimates the nitride contribution. Probable causes are discussed.

Spectroscopic ellipsometry measurements of thin metal films

Optical methods are used to determine the thickness of thin metal films, with emphasis on spectroscopic ellipsometry and transmission. We discuss the conditions where this is possible and how to determine the optical constants for the material. The determination of the thickness of each of two metals in a bimetallic stack is discussed. Finally, by measuring thickness with these methods and measuring weight gain, we determine the density of platinum deposited by evaporation and deposited by a simple sputter deposition method. Explicitly, one of the results is that the determined optical constants depend on the deposition method. This implies that, with only a few exceptions, one must determine the optical constants of the material of interest, and that it is inappropriate to use values from another source such as a handbook or from another investigator. The resulting optical constants in this work suggest that the microstructure of the platinum films from the two different methods will not be the same, and x-ray diffraction and sheet resistance measurements verify that this is the case. Specifically, the significantly lower extinction coefficient of the sputter-deposited films correlates with a higher sheet resistance.

Kinetics and mechanism for desorption of H2O from spin‐on‐glass

In this study, the rate equation is considered for outgassing of H2O from spin‐on‐glass during heating cycles. Also considered is the identity of the film species which desorbs. Previous studies have provided processing engineering information about how much the samples must be heated to desorb the water and how long the samples can be exposed to room temperature before readsorption. In this study, residual gas analyzer thermal analysis data are analyzed using methods suggested recently in the technical literature and it is determined that the desorption reaction is second order with an activation energy of ∼19 kcal/mol. Using Fourier‐transform infrared, it is shown that the desorption reaction which gives off the H2O during heating involves the reduction of near‐neighbor hydrogen bonded silanols to produce Si–O–Si and a water molecule.

Nitrogen-doped plasma enhanced chemical vapor deposited (PECVD) amorphous carbon: processes and properties

In this work we discuss thin film amorphous carbon, which is deposited in a dual frequency, plasma enhanced chemical vapor deposition (PECVD) system in such a manner that it contains a small amount of nitrogen. Unlike most carbon films deposited using PECVD, the films in this study were deposited on the grounded electrode and therefore, subject to little energetic bombardment during growth. Methane was used as the carbon-containing precursor. We illustrate some potential applications for this type of film and discuss the effect of various process parameters on resultant film properties such as optical constants, resistivity, stoichiometry, and chemical bonding and structure.

OXIDATION OF TIN IN AN OXYGEN PLASMA ASHER

In previous work we have studied the thermal oxidation of titanium nitride. This involved temperatures of 350 °C and greater, and times of hours. In this work, we study the oxidation of titanium nitride in a downstream plasma asher, where the sample temperature is about 200 °C. For comparison to an older technology, we also include samples oxidized in a barrel asher where the sample temperature was about 100 °C. Oxidation times of 16 min and less were studied. Thickness and optical constants were measured using ellipsometry along with Auger electron spectroscopy with argon ion etching for depth profiles. The growth kinetics are shown to be logarithmic. In the down‐stream asher, films form in 90 s which would require 40 min at 450 °C if grown thermally. We show that only the thinner films can be etched off in 50:1 HF.

An investigation of the oxidation of Ti:W

The alloy Ti:W was oxidized in 100% O2 at temperatures ranging from 375 to 505 °C. The resulting oxides were analyzed with Auger electron spectroscopy, x‐ray photoelectron spectroscopy, Rutherford backscattering spectroscopy, and ellipsometry. The resulting oxide structure contains both Ti and W and appears to be simply a combination of TiO2 and WO3. The indices of refraction of the TiW metal and the oxide were determined to be Ns=2.84–3.08j and Nf=2.25–0.12j, respectively. Film thicknesses of the oxides formed at various temperatures and times were measured with ellipsometry and the resulting kinetics are presented. The growth kinetics were determined to be parabolic with an activation energy of 33.9 kcal per mole (1.47 eV).

Tightly bound H2O in spin‐on‐glass

The purpose of this work is to measure the amount of H2O which is bound tightly in spin‐on‐glass (SOG) and the amount of reabsorption as a function of room air exposure time. Both results are needed in determining bake cycles and wait times prior to metal deposition when SOG is used in the dielectric stack of a multilevel metal system. About a tenth of the amount of gas which remains after a 2 h 150 °C bake is still present after a 2 h 350 °C bake. About half the amount of the tightly bound gas which is desorbed will be reabsorbed upon exposure to air in a period of 4 h or less. The time for half of the reabsorption to occur appears to be of the order of 20 to 60 min.