Steven Danyluk

Influence of fluids on the abrasion of silicon by diamond

Abstract Silicon wafers ((100) p-type) were abraded at room temperature in the presence of acetone, absolute ethanol and water by a pyramid diamond, and the resulting groove depth was measured as a function of the normal force on the diamond and the adsorbed fluids, all other experimental conditions being held constant. The rates of increase in groove depth (the depths of the groove(s)) are in the ratio 1:2:3 for water, absolute ethanol and acetone respectively for a constant normal force. The groove depth rate is lower when the normal force is decreased. The abraded surfaces were examined by scanning electron microscopy. At a constant normal force, the silicon abraded in the presence of water was chipped, as expected for a classical brittle material, while the surfaces abraded in the other two fluids showed ductile ploughing as the main mechanism for silicon removal.

Ball-on-flat reciprocating sliding wear of single-crystal, semiconductor silicon at room temperature

Abstract This paper describes the reciprocating linear sliding friction and wear experiments of a polycrystalline silicon ball on a single-crystal, silicon flat at room temperature in an argon environment. The normal loads used were 2.65, 3.59, 4.73 and 5.80 N, and the sliding velocity was 1.3 cm s −1 . The coefficient of friction at a normal load of 2.65 N initially decreased from 0.45 and stabilized at 0.23 after 400 cycles. At higher normal loads, the coefficients of friction were initially 0.45 but later fluctuated between 0.2 and 0.4. The degree of frictional fluctuation as well as the wear volume increased with load. The fluctuation of the coefficient of friction is related to the geometry of the wear debris which form into a cylindrical shape (‘rolls’). A mechanism for roll formation is described.

Microhardness of carbon-doped (111) p-type Czochralski silicon

The effect of carbon on (111) p-type Czochralski silicon is examined. The preparation of the silicon and microhardness test procedures are described, and the equation used to determine microhardness from indentations in the silicon wafers is presented. The results indicate that as the carbon concentration in the silicon increases the microhardness increases. The linear increase in microhardness is the result of carbon hindering dislocation motion, and the effect of temperature on silicon deformation and dislocation mobility is explained. The measured microhardness was compared with an analysis which is based on dislocation pinning by carbon; a good correlation was observed. The Labusch model for the effect of pinning sites on dislocation motion is given.

The influence of backgrinding on the fracture strength of 100 mm diameter (1 1 1) p-type silicon wafers

The influence of grinding geometry and damage depth on the fracture strength of 100 mm diameter (1 1 1) p-type silicon wafers has been studied. The fracture strengths were measured in a biaxial flexure test after the wafers were ground to 0.36 mm from 0.53 mm thick, in a grinding apparatus that produces a swath of swirls on the silicon wafer surfaces. Analysis of orientations of the swirl geometries and fracture probability was used to deduce the fracture strength relative to the crystallographic orientation of the wafers. Optical and scanning electron microscopy of bevelled, and cleaved and etched samples was used to measure the damage depths from selected locations on the wafers. The depth of damage and fracture strengths were correlated to the geometry of the backgrind swirl pattern and the relative position of the orientation flat. The damage depth was smaller when the swirl path was parallel or at 45° to the orientation flat as compared to the swirl paths at 90° and 135° orientations. As a result, the wafers ground in the former orientations had a higher fracture strength than those of the latter orientations (136 and 124 MPa versus 100 and 103 MPa, for the four orientations, respectively).

The wear rate of n-type Si(100)

Abstract Single-crystal Si(100) (n type) was scratched at room temperature by a single-point 90° pyramidal diamond in each of the following fluids: reagent-grade absolute ethanol, methanol, acetone and deionized water. The wear rate in each case was determined from a measurement of the cross-sectional areas of circular multiscratch grooves and the increase in this area with scratching time. These areas vary as the contact force on the diamond and the fluid covering the silicon surface during the scratching test vary. As expected, the wear rate depended on the force (dead-weight load) on the diamond (the wear rate increased as the force was increased) but fluids also had a significant effect on the wear. The wear rate of silicon scratched in ethanol is twice that in deionized water when the dead-weight load on the diamond is 0.49 N and all other experimental variables are held constant. These results were compared with two models of abrasive wear of ceramic materials.

Residual stresses of thin, short rectangular plates

The analysis of the residual stresses in thin, short rectangular plates is presented. The analysis is used in conjunction with a shadow Moiré interferometry technique by which residual stresses are obtained over a large spatial area from a strain measurement. The technique and analysis are applied to a residual stress measurement of polycrystalline silicon sheet grown by the edge-defined film growth technique.

Correlation of dynamic friction and the dislocation etch pit density surrounding annealed scratches in (1 1 1) p-type silicon

The dynamic friction coefficient between a 90° pyramid diamond and a (1 1 1) p-type silicon single crystal has been measured for linear, unidirectional scratches made in the [1 1 0] direction in laboratory air, deionized water and ethanol. The friction coefficient for grooves formed in air increased from 0,6 to 0.7 as the number of scratches increased, reaching a steady state value after ten scratches. The friction coefficient for grooves formed in deionized water and ethanol decreased from 0.7 to 0.5 and also reached a steady state value after ten scratches. Scanning electron microscopy showed that the groove morphology depended on the fluid in contact with the surface during the scratch test. The grooves formed after ten scratches were annealed to 750° C for 3600 sec and etch pits were measured as a function of distance from the groove wall. The etch pit density and their distance from the groove wall was related to the type of fluid used in the scratch test: the density was highest and the distance from the groove wall largest for the groove formed in air and lowest for the groove formed in ethanol. These results imply that the deformation mode and the magnitude of the residual stresses surrounding the groove depend on the enviromental conditions during the scratch test. The friction coefficient was found to vary linearly with the average etch pit density.

Surface damage of single-crystal silicon abraded in ethanol and deionized water

Semiconductor grade, single-crystal silicon wafers of (1 0 0) p-type were abraded by a single-point, slow speed (2.3 cm sec−1) 90° pyramid diamond in ethanol and deionized water. The scratching was carried out in each of the fluids with a load of 0.5 N on the sliding diamond. The scratching produces a groove, the depth of which depends on the number of traverses of the diamond. A measure of the cross-sectional area of the groove was used to determine the abrasion rate in ethanol, which was about 1.3 times that in deionized water. Some of the samples scratched in deionized water were annealed at 1000° C for 1 h and these samples, along with those unannealed and scratched in both fluids, were fractured perpendicular to the scratch in a three-point bend apparatus. The fracture strengths and the mirror distances obtained by scanning electron microscope (SEM) observation were used to deduce tensile residual stresses of 15.6 and 99.0 MN m−2 beneath the grooves formed in deionized water and ethanol, respectively. The SEM investigation also showed that (a) the groove surfaces contained microcracks wedged with wear debris, and (b) dislocations were generated and propagated away from the groove surfaces as a result of annealing. The relatively higher tensile residual stress produced in the presence of ethanol is consistent with the higher wear rate in this fluid.

Friction and wear of single-crystal silicon at elevated temperatures

Single-crystal silicon wafers ((1 1 1) and (1 0 0)p-type) were abraded at room temperature 300 °C, and 600 °C by a polycrystalline partially stabilized zirconia ball in a ball-on reciprocating flat geometry. The sliding direction was 〈1 1 0〉. The friction coefficient was recorded as a function of reciprocating strokes and the deformation mode of the silicon. The friction coefficient at room temperature decreased with the number of strokes, and this variation was less affected by the number of strokes at the higher temperatures. The wear track width and depth were measured at the three temperatures. Wear increases as the temperature is raised to 300 and 600 °C. Optical and scanning electron microscopy of the subsurface damage reveals that cracks are generated at RT and 300 °C and dislocations are produced at 600 °C. The change in deformation mode with temperature from brittle fracture to plastic deformation accounts for the differences in wear.

Residual stress measurement in filament-evaporated aluminium films on single crystal silicon wafers

Shadow Moire interferometry was used to determine the residual stresses in thin, filament-evaporated aluminium films deposited on (100) p-type, 10.16 cm diameter, 0.05 cm thick, circular single crystal silicon wafers. The aluminium film thicknesses ranged from 70 to 780 nm. Benchmark experiments on wafers without aluminium films showed that wafers possess in-plane residual stresses; the centre of the wafer is under tensile stresses of the order of 30 M Pa and these stresses decrease toward the edge of the wafers to approximately 10 M Pa. The deposition of aluminium films increases these tensile residual stresses by 3 to 15 M Pa depending on the film thickness. The increase in the stress is attributed to the stresses in the films.