The decomposition of tetraethoxysilane (TEOS) on a thermally grown SiO2 surface has been studied using molecular beam scattering, with mass spectrometric detection and time of flight techniques. Decomposition was characterized over a range of temperatures from ambient to 1100 K using a mass spectrometer having a line of sight to the sample surface. Time of flight measurements were used to characterize parent and product species and to determine surface residence time. The initial step in the decomposition process at low temperatures is the loss of one ethoxy ligand to produce triethoxysilane, ethylene and possibly water as gas phase products. At temperatures above 850 K there is a change in the rate limiting process and an abrupt increase in the decomposition probability. Observed kinetics are explained in terms or an adsorbed molecular precursor mechanism.

The reaction of Si2H6 with the Si(100) surface has been examined via supersonic molecular beam scattering techniques. The effects of incident translational energy, incident angle, mean vibrational energy and surface temperature have been considered explicitly. It is found that the reaction probability is most sensitive tothe incident translational energy, varying nearly linearly with increasing energy. The effect of incident angle θi can be accounted for by an energy scaling law that is given approximately by <Etr>cosnθi, where n ∼ 1.

Films of tantalum pentoxide (Ta2O5) have been fabricated by use of different precursor materials, deposition techniques and annealing procedures. Several analytical methods were appliedto study the layers. Fundamental properties and new data are reportedand related to practical features that are of importance in device design and manufacturing of advanced, highly integrated devices. This overview may facilitate the choice of an optimal combination of precursor, deposition technique and corresponding annealing procedure for aspecific application of these films in microelectronics, since the electrical properties reveal the potential of Ta2O5 films for the use in 64Mbit and 256Mbit DRAM devices as high dielectric constant material.

A precursor for the LPCVD of silicon oxide films has been developed that extends the low temperature deposition range to 100°C. The chemical, 1,4 disilabutane (DSB), produces silicon oxide depositions similar to those of the higher temperature silane and diethylsilane (DES) processes. Optimum DSB processes require pressures below 300 mTorr, similar to silane, in contrast to DES pressures above 600 mTorr at 350°C. This results in poorer conformalities than those of DES, but the step coverages are still superior to those from silane oxides. The DSB films are low stress, carbon-free oxide layers that are suitable for temperature-sensitive underlayers and substrates such as photoresist, plastics, GaAs, and HgCdTe.

The kinetics of deposition of SiO2 by the reaction of tetramethylsilane (TMS) with ozone (O3) has been studied over the temperature range 180 – 380° C and compared with available data for the same process using tetraethoxysilane (TEOS). Both processes exhibit the same activation energy (17 kcal/mole) below 300 ° C which falls-off at higher temperatures due to transport limitations. Transition from first- to zero-order kinetics occurs with increasing concentrations of TMS and O3, which gives an overall O3/TMS consumption ratio of 10 at 258° C and5 at 325° C. TEOS is estimated to be 5 times more reactive than TMS above 300° C and over 10 times more reactive in the kinetically-limited regime below 300° C. Results suggest that O3-induced SiO2 deposition proceeds via surface reactions and is limited by heterogeneous decomposition of ozone.

We report in-situ Fourier transform infrared (FTIR) reflection-absorption studies of curing chemistry of polyimide thin films on Crand Cu surfaces, and of the thermal stability of the resulting thin film interfaces when exposed to air at elevated temperatures. The polyimide investigated is based on 4,4'-(hexafluoroisopropylidene)bis(phthalicanhydride)-4,4'-bis(4-aminophenoxy)-biphenyl. The imidization process takes place at temperatures higher than 90°C anda small amount of anhydride is generated during curing. This by-product is converted to imide at temperatures above 250°C. Complete imidization is achieved after curing at 300°C on Cr substrates, while evidence for incomplete curing on Cu is observed under the same conditions. Thermal stability studies with Cr and Cu substrates show that thermal decomposition of thin (∼1000Å) polyimide filmsoccurs on Cu when the film is exposed to air at 200°C, while the polyimide is stable on Cr.

The object of this work is to identify the most important structure features in the bonding of polyimide/silicon during the curing anddegradation. An essential tool for this work is a quantitative adhesion measurement which can relate the macroscopic adhesion strength to the microscopic interfacial structures. We apply the blister test forthis purpose. In particular, the test is applied to examine the effect of thermal history on the adhesion of PMDA-ODA polyimide on siliconsurface. Our studies show that the test is accurate and sensitive. The adhesion data reveal that increases in adhesion strength coincide with the formation of certain thermal degradation products as revealedin the earlier thermal degradation studies. This suggests a mechanismfor increases in adhesion strength with curing temperature, based on chemical and physical modification of the interface.

Parylene films, depending on the type, are thermally stable up to 530°C and have low dielectric constants ranging from 2.35 to 3.15. One of the most interesting properties of this material is its vapor depositability. Conventional vapor deposition involves cracking the parylene dimers at temperatures from 600 to 730 °C and polymerizingthe monomers at - 35 °C to RT. We have developed a simpler and less expensive technique that directly uses the precursors from which the dimers are made. This method requires the use of metal catalysts to produce parylene films. We have used the precursors α,α'-dibromo-p-xylene and dibromotetrafluoro-p-xylene to produce N-type and F-type parylene films. FTIR, XPS, thermal stability, and electrical studies show that the F-type parylene films grown from the precursors are comparable to, or sometimes better than, the films grown from dimer, and have potential microelectronics applications.

We have investigated the feasibility of depositing GaN, AIN and InN from nitrogen plasmas by electron cyclotron resonance-metal organic molecular beam epitaxy (ECR-MOMBE). Growth rate, morphology, and resistivity were evaluated as function of growth temperature and group IB flux. It was found that stoichiometric materials could be deposited at reasonable growth rates on either GaAs or sapphire substrates. Low contact resistance, ∼5 × 10−7 Ω-cm2, can be obtained on In due to the high carrier concentrations, 1020 cm−3 obtained in this material.

The gas-phase chemical reactions in the Metallorganic Chemical Vapor Deposition (MOCVD) of A1N and TiN have been studied using IR spectroscopy. The products formed from the reaction of trimethyl aluminum (TMA) and NH3 were compared to those from the reaction of TMAwith NF3 using a static gas-phase IR cell. Reaction with NH3 is rapid at 25 °C, and the IR spectrum of the product is consistent with the acid-base adduct (CH3)3Al-NH3. At 25 °C, no reaction between TMA and NF3 was observed. However, at 58 °C a slow reaction occurredto give (CH3)2AlF. The reaction of Ti(N(CH3)2)4 with NH3 was also studied using a flow-tube reactor with a sliding injector port that provides control over the reaction time between two reactive flows. By monitoring the disappearance of Ti(N(CH3)2)4 as a function of NH3 partial pressure and reaction time, we have obtained a preliminary estimate of the rate constant as ∼ 10−16 cm3 molecule−1 s−1 at 25 °C. This result confirms that the reaction is rapid even at room temperature and demonstrates the utility of the flow-tube reactor and FTIR spectrometer for studies of MOCVD chemistry.

A molecular beam mass spectrometer coupled to the inlet portion of a hot-wall CVD reactor has been used to measure the time-of-flight (TOF) distribution of molecules sampled from the reactor. The molecular weights of the species in the reactor can be determined from test TOF distributions. This technique has been used to identify the gas phasespecies present in a hot-wall reactor during the CVD of A1N using bis-dimethylaluminum deutero-amide,[(CH3)2AlND2]3 (I), as the precursor. Evidence has been found for: deuterated methane, CH3D, trimeric form of the precursor I, and its dimeric form. TOF evidence also supports the presence of an oligomer of I with a pentameric structure, as well as a number of substituted versions of trimer and dimer. Mechanisms have been formulated that explain methane elimination from the precursor as well as the formation of all species observed.

The etching of polymeric films in O2 plasmas containing small amounts of CF4 has been studied at low pressures (10–60 mtorr) in an RIE reactor. The relationship between etch rate, ion flux, and O atom concentration is investigated for photoresist, polyimide, and amorphous carbon. The applicability of a surface-chemical model proposed by Joubert, et al. (J. Appl. Phys. 65(12) 1989, 5096) is explored. Modifications to the model are made to include physical sputtering and direct reactive ion etching. The modified model provides an improved description of the process, showing the various contributions of the ions and O atoms to the overall etching.

This paper presents results on PECVD of low refractive index silicon nitride in a large area system. Film deposition characteristics, suchas deposition rate and thickness uniformity, were investigated over awide range of process parameters, such as gas composition, power, pressure, and N2 pressure. Film properties, such as RI, absorbance, stress, and etch rate, have also been studied. A general model, which includes both the deposition and etching mechanisms, has been developed to explain these results. This model could explain the film uniformity issue in the process.

Spectroscopic evidence is presented that shows that boron nitride andaluminum oxide can be synthesized by exposing a condensed layer of molecular precursors to synchrotron radiation. In the A12 O3 circumstance a condensed layer of trimethylaluminum (TMA)and water at 78 K on a silver substrate produces pure layers of aluminum oxide. Using the same condensed layer technique boron nitride is produced by exposing a solid matrix of diborane and ammonia to synchrotron radiation. Near edge x-ray absorption fine structure and core level photoelectron spectroscopies are used to characterize the A12 O3 and BN layers, which were typically 30Å thick. During the formation of aluminum oxide the carbon component inthe alkylaluminum precursor is completely removed during irradiation as a volatile methane product which was detected by mass spectrometry. In the absence of synchrotron radiation the molecular precursors in both the aluminum oxide and boron nitride systems show evidence of some interactions within the solid, but upon warming to near room temperature (260 K) the layers desorb from the substrate.

TiN thin films were deposited by reactive sputtering onto Si substrates. TiO2 films were formed by subsequent oxidation of the TiN films, using either conventional furnace heating or rapid thermalprocessing (RTP). The materials properties of the resulting films were characterized by x-ray diffraction and Rutherford backscattering, and indicate conversion of the TiN to fully-oxidized rutile TiO2 by a diffusion-limited process. Electrical measurements of the insulating properties of the TiO2 films indicated a relative dielectric constant greater than 100, although the leakage current was greater than optimum. A fully in-situ process for the fabrication of microcapacitors is proposed, which involves the sputter deposition of TiN, the formation of TiO2 by RTP, and the deposition of a top TiN counterelectrode. This can be carried out under conditions that are compatible with Si microelectronic device processing.

Stable zirconium carbide thin films have been deposited from a singlesource organometallic precursor, tetraneopentyl zirconium, at substrate temperatures above 500 °C. Materials deposited above this temperature are crystalline by X-ray diffraction. A metal to carbon ratioof 1:2 is observed by Auger electron spectroscopy depth profiling. X-ray photoelectron spectroscopy indicates the zirconium is single phase. The observed spectra correspond well to spectra for zirconium carbide standards. Carbon XPS reveals carbidic and graphitic or hydrocarbon species with a third unknown carbon species. Elastic recoil detection finds a large, up to 16%, hydrogen content in the thin film.

Chemical spray pyrolysis (CSP) is a technique in which compounds of the constituents of the thin film to be fabricated are dissolved in an aqueous solution which is subsequently sprayed onto a heated substrate using nitrogen as the atomizing gas. At relatively low substrate temperatures (150–400°C) chemical reactions take place in which film formation of the desired compound occurs concomitantly with the release of volatile chemical reactants. The technique has advantages that include: (1) simplicity, (2) low cost and simple equipment, (3) the ability to prepare films over large areas with various shapes with relative ease and (4) the possibility of varying the physical properties through chemical means in known ways at modest temperatures. Results on the preparation of CuInSe2 which is a defect dominated semiconductor of technological interest are presented. Both n- and p- type materials were prepared with resistivities varying from 10−2 – 104 ohm-cm illustrating the use of the control of solution chemistry to produce films with reproducibly controlled properties.

The chemical vapor deposition of cubic gallium sulfide and tetragonalindium sulfide films is reported. The structure of the deposited films was demonstrated to be defined not solely by thermodynamics, but bythe predesigned molecular motif of the precursor molecules. Analysis of the deposited films has been obtained by transmission electron microscopy (TEM), with associated energy dispersive X-ray analysis (EDX) and X-ray photoelectron spectroscopy (XPS).

We have recently reported the growth of ZnTe and CdTe thin films from the novel volatile single-source precursors Zn(sitel)2 and Cd(sitel)2 [sitel = TeSi(TeSi(siMe3)3]. In an effort to understand the role played by the bulky substituents, we have prepared a number of related compounds and have investigated their potential for the growth of II-VI materials. We now describe the synthesis and characterization of M[EX(SiMe3)3]2 (M = Zn, Cd, Hg for E = Te, X = C; M = Zn for E = Se, X = C, Si) and their use as precursors molecules for the growth of ZnSe, ZnTe, CdTe and HgTe. The thin films were characterized by Rutherford Back scattering spectrometry (RBS) and X-ray diffraction. Pyrolysis reaction conditions for the alkyl and silyl chalcogenolates are also compared.

A novel method for chemical vapor deposition and atomic layer epitaxyusing radical precursors under medium vacuum conditions is being developed. Fluorine atoms are generated by thermal dissociation in a hot tube and abstract hydrogen atoms from precursor molecules injected immediately downstream of the source, generating radicals with completechemical specificity. The radical precursors are then transported to the growing substrate surface under nearly collision-free conditions. To date we have grown diamond films from CCl3 or CH3 radicals together with atomic hydrogen, generated by injecting CHCI3 or CH4 and H2 into the F atom stream at reactor pressures between 10−4 and 10−2 Torn This approach should be ideal for low-temperature growth and atomic layer epitaxy: growth rates remain relatively high because activation energies for radical reactions are typically small and because the cycle times for atomic layer epitaxy can be reduced to die msec range by fast gas-stream switching, and contamination and segregation are minimized by keeping the surface “capped” by chemisorbed intermediates.