A. E. Morgan

Charging of insulators by ion bombardment and its minimization for secondary ion mass spectrometry (SIMS) measurements

The generation of charge build‐up, caused by bombardment of insulating samples with energetic particles, and its role in altering the relative secondary‐ion currents and reducing their absolute values, sometimes even to zero, are discussed. Proposed methods for charge reduction by bombardment with negative ions or with neutral particles, or by use of an auxiliary electron beam or spray gun, are shown to be not useful in every experimental situation. A further method involving introduction of an auxiliary conducting electrode when using negative primary ions is considered mechanistically, and tested by placing tantalum diaphragms onto the surface of yttrium iron garnet (YIG). The final charging values and mass spectra thus obtained are compared to those measured when using bare and metallic grid‐covered surfaces of the same YIG sample. The satisfactory results achieved with the diaphragm show that SIMS analyses of insulators can be performed in a simple, quick, and inexpensive manner.

Interactions of thin Ti films with Si, SiO2, Si3N4, and SiOxNy under rapid thermal annealing

Thin Ti films sputter deposited onto single‐crystal Si, thermal SiO2, and low‐pressure chemical vapor deposited Si3N4 and SiOxNy (x≊y≊1) substrates have been rapid thermal annealed in N2 or Ar, with and without an amorphous Si overlayer, and the reactions followed using Auger elecron spectroscopy, transmission electron microscopy, electron diffraction, and sheet resistance measurements. A multilayer film is created in practically every case with each layer containing essentially a single reaction product, viz.,TiSix, TiOx, δ‐TiN, or TiNxO1−x. The results are discussed in light of published Ti‐Si‐O and Ti‐Si‐N phase diagrams.

The influence of bombardment conditions upon the sputtering and secondary ion yields of silicon

Abstract The steady state sputtering yield S and the useful degree of ionisation β + η of silicon have been measured as functions of primary ion energy, atomic number, incident angle, species (rare gas or oxygen), and oxygen partial pressure. The optimum bombardment conditions for SIMS analyses are discussed, and a comparison of the instrumental factor η for two SIMS instruments is made. Variation of both S and β + with oxygen surface coverage, as determined by using Auger electron spectroscopy, is given. A simple model is used to describe the mutual intensity changes occuring in the Auger spectrum during the bombardment-induced oxidation of silicon.

A thermodynamic approach for interpreting metallization layer stability and thin‐film reactions involving four elements: Application to integrated circuit contact metallurgy

We propose a means of constructing simplified quaternary phase diagrams based on a thermodynamic approach using known and estimated data for Gibbs free energy of formation. Isothermal, isobaric sections of the condensed phase diagrams are built up as tie lines, tie planes, and tie tetrahedra (representing two‐, three‐, and four‐phase equilibrium, respectively) in a regular tetrahedron. This extends the now well‐established methodology for ternary systems described, for instance, by Beyers [J. Appl. Phys. 56, 147 (1984); Mat. Res. Soc. Symp. Proc. 47, 143 (1985)]. The procedure is illustrated by reference to Ti‐Si‐N‐O, Al‐Si‐N‐O, Ti‐Al‐Si‐N and Ti‐Al‐Si‐O, systems which are relevant to interactions occurring at various interfaces during the formation of contacts in integrated circuits. These phase diagrams are then used to predict the stability of—or reactions in—metallization layers and thin‐film systems involving four elements. In addition, a method is suggested to estimate unknown free energies of forma...

Shallow junction formation by preamorphization with tin implantation

Amorphizing n‐type 〈100〉 silicon by tin implantation prior to implanting 10 keV boron and then annealing for 30 min at 800 °C results in a 0.22‐μm‐deep p+/n junction. The implanted tin prevents boron channeling, enhances the quality of the solid phase epitaxial regrowth of the silicon, and shows no measurable diffusion. A discontinuous band of dislocation loops, 20–30 nm in diameter, with a density below 1010 cm−2 remains at the original amorphous‐crystalline interface after annealing. Junctions are nearly ideal and are characterized at −5 V reverse bias by an areal leakage of −5 nA cm−2 and a peripheral leakage less than −0.1 fA μm−1.

Reaction of titanium with silicon nitride under rapid thermal annealing

30–90 nm Ti films sputter deposited onto 50 nm Si3N4 have been rapid thermal annealed for 30 s in Ar and N2 ambients, and the phases formed identified using Auger electron spectroscopy, transmission electron microscopy, and electron diffraction. Reaction at 900 °C produces TiN surface and interfacial layers and an intermediate TiSi2 layer. The Ti‐Si‐N ternary phase diagram is used to explain the reaction sequence.

Formation of β‐SiC at the interface between an epitaxial Si layer grown by rapid thermal chemical vapor deposition and a Si substrate

The formation of β‐SiC during rapid thermal chemical vapor deposition of an epitaxial Si layer on a (100) Si substrate has been demonstrated using high‐resolution cross‐sectional transmission electron microscopy and secondary ion mass spectrometry (SIMS) depth profiling. Ellipsoidal β‐SiC precipitates (3×10 nm2 average size) were found with a density of (6±1)×109 cm−2 along the epilayer/substrate interface. SIMS revealed about 0.2 monolayer of carbon concentrated at the interface. The precipitates were shown to form during the hydrogen prebake as a result of improper surface cleaning. The crystalline quality of the epitaxial layer, however, was unaffected by the presence of the β‐SiC.

Effect of oxygen implantation upon secondary ion yields

The useful degree of ionization β+η of boron, present as a dopant in various substrates, has been measured using O+2 primary ions. The measurements also yielded β+η for the matrix elements and the steady‐state sputtering yield S of the substrate. Two SIMS instruments were employed in order to vary the bombardment parameters. Auger electron spectroscopy was used to measure the surface concentration of implanted oxygen. Neither this quantity nor (β+η)B are simply related to the inverse of S as has been proposed in the literature. Further, β+η of a matrix element, when referenced to the corresponding (β+η)B, does not exhibit an exponential variation with the ionization potential.

Formation and high‐temperature stability of CoSix films on an SiO2 substrate

A thin α‐Co layer with an amorphous Si underlayer has been sputter deposited onto a thermal SiO2 substrate, rapid thermal annealed in N2 at 700–1050 °C, and the phases formed examined using Auger electron spectroscopy, transmission electron microscopy, electron diffraction, and sheet resistance measurements. A CoSix film results where x is constant with depth and determined by the relative amounts of Co and Si deposited. With increasing x, phases identified are α‐ and β‐Co containing dissolved Si, Co2Si, CoSi, and CoSi2. At high temperatures, the CoSi2 film agglomerates and thins the underlying oxide probably on account of excess Si in the silicide film. Furthermore, in an N2 atmosphere, the CoSi2 globules are converted into CoSi in accordance with the phase diagram.

Annealing and oxidation behavior of low‐pressure chemical vapor deposited tungsten slicide layers on polycrystalline silicon gates

A WSix≊2.6 film was deposited by low‐pressure chemical vapor deposition at 350–420 °C onto a P‐doped polycrystalline silicon/SiO2/Si substrate. This polycide structure (with or without a subsequent As+ source/drain implant) was heat treated in the following manner: (i) 1000 °C/N2 anneal for 20 min, (ii) oxidation both in dry oxygen and in steam ambients at 920–950 °C (50 min–8 h), and (iii) N2 anneal as well as dry oxidation at 920 °C/50 min. Cross‐sectional transmission electron microscopy, Rutherford backscattering, secondary ion mass spectrometry, and sheet resistance measurements were used for characterization. The as‐deposited film was mostly amorphous with a fine grain structure. The N2 anneal caused the formation of a polycrystalline WSix>2 film containing tetragonal WSi2. Dry oxidation produced a void‐containing SiO2 layer atop the silicide film. However, the SiO2 layer was impregnated with W particles when the silicide film was steam oxidized. The role of excess Si in refractory metal silicide fi...