Microstructural Studies of Thermally Sprayed Deposits by Neutron Scattering

Transcription

1 Microstructural Studies of Thermally Sprayed Deposits by Neutron Scattering T. Keller, W. Wagner Paul-Scherrer-Institute, Villigen, Switzerland J. Ilavsky University of Maryland at College Park, College Park, MD, USA N. Margadant, S. Siegmann EMPA Swiss Federal Laboratories for Materials Testing and Research, Thun, Switzerland J. Pisacka, J. Matejicek Institute of Plasma Physics, Academy of Sciences of the Czech Republic, Prague, Czech Republic G. Barbezat Sulzer-Metco AG, Wohlen, Switzerland P. Fiala Skoda Research, Plzen, Czech Republic T. Pirling Institute Laue-Langevin, Grenoble, France Abstract Technological properties of thermally sprayed deposits are to a great extent related to the underlying microstructure. The present project aims to relate macroscopic properties of metallic coatings to their microstructure. For this purpose, thermally sprayed deposits of nickel based alloys (NiCr, NiCrAlY) were manufactured by various spraying techniques - atmospheric and vacuum plasma spraying, flame spraying, high velocity oxygen fuel and water-stabilized plasma spraying. One of the key microstructural features is the void system. This system is usually characterized by the total volume of voids, the so called porosity. An additional characteristic parameter of the void system is the specific surface area. The method of anisotropic Small Angle Neutron Scattering (SANS) in the "Porod Regime" allows the determination of the anisotropic specific surface area of the complex void system that consists of intralamellar cracks and interlamellar pores [1]. In contrast to optical microscopy, the SANS technique is capable of resolving the pore structure down to the nanometer scale, and the measured specific surface area represents a statistically relevant average value for the whole illuminated sample volume which is usually a few mm 3. Besides the presence of voids and cracks the performance of thermally sprayed coatings is also significantly influenced by residual stresses. In the present work residual strains were determined by the technique of neutron diffraction as well as by bending tests, i.e. laser profilometry of the substrate before and after the spraying process. The specific surface area and the residual stresses are discussed with respect to total porosity, the presence of secondary phases like oxides and wear behavior. Special attention is drawn to the anisotropy of the apparent surface area, which is discussed with respect to the anisotropy of macroscopic properties like electrical resistance. Introduction Improved spray control allows reproducible manufacturing of deposits for different applications. Therefore generally the specific design of thermally sprayed coatings is possible, but complicated by the demand to fulfill several requirements at a time, for example wear and corrosion resistance for several applications of Ni-based alloys. Designing deposits with suitable properties requires a general understanding of the coating processes, i.e. spraying techniques and their parameters, of the microstructure and also of the relationships between microstructure and deposits performance and properties. For the development of a fundamental understanding besides the classical techniques of investigation also more sophisticated methods may need to be used. For example, neutron scattering can provide unique information. Special features of neutrons com-

2 pared to x-rays or electron microscopy are the high penetration depth and moderate needs for sample preparation. The present work reports results from measurements of the voids specific surface areas obtained by Small Angle Neutron Scattering (SANS) in NiCrAlY coatings as well as residual stress measurements on the same deposits. The anisotropy of the specific surface areas of NiCr coatings [2,3] is discussed in relation to the anisotropy of a macroscopic property like the electric resistance. Experimental 1. Feedstock Material, Spray Techniques and Parameters Nickel based alloys - NiCr (8% Ni, 2% Cr) and NiCrAlY (67% Ni, 22% Cr, 1% Al, 1% Y) - were used as feedstock materials. Commercially available powders were used with size ranges suitable for the different spraying techniques. All spraying techniques applied within this project were used to manufacture not only the standard microstructures but the widest possible range of microstructures reasonably attainable by each of the different techniques. Details on the spray parameters applied are described in [3]. 2. Small Angle Neutron Scattering Small Angle Neutron Scattering can provide information on the total area of homogeneous interfaces in a two component system in the so called "Porod Regime" [4], where the scattering cross section, i.e. the absolute intensity I(Q), decreases with the fourth power of the scattering vector Q or of its modulus, the momentum transfer Q, respectively. The latter is defined in the usual way as Q = (4sin where is half the angle between the direction of incident and scattered neutrons, and the neutron wavelength. If L is the smallest dimension of the scatterers, a necessary condition for the validity of the Porod Regime is QL>3 [1]. For thermally sprayed coatings the relevant interfaces are the surfaces between the pore system and the matrix and, if present, between the matrix and secondary phases, like oxides. The voids and the metallic matrix have a different scattering length density,, which causes some of the neutrons to be scattered at the interfaces. The measurements were performed at the SANS instrument at the neutron spallation source SINQ at Paul-Scherrer-Institute (PSI) in Villigen / Switzerland. Cold neutrons of a wavelength of.5 nm were used, selected by a mechanical velocity selector (Dornier NSG) with a resolution of / 1.. A two dimensional area sensitive detector with 128x128 pixels, each of size 7.5 x 7.5 mm 2, and 3 He as detecting gas is used for the data collection. A detailed description of the instrument can be found elsewhere [5, 6]. Figure 1 shows a photograph of the SANS instrument. To follow the intensity curve over a wide range of momentum transfer Q, the detector was placed at distances of 2 m, 4.5 m and 2 m from the sample position, combined with adapted collimation of similar lengths for optimum adjustment of the beam divergence. SANS samples were prepared by cutting cross sections of 1.9 mm in thickness, and subsequent grinding to obtain flat surfaces. The sample cross section was masked by a rectangular Cadmium slit such that only the coating was illuminated edge-on and scattering from the substrate was avoided. An electromagnet allowed to apply a horizontally oriented magnetic field to the sample to avoid magnetic refraction effects. The magnetic field at the sample position was about 1 Tesla such that magnetic saturation was almost achieved. Still, to identify and separate possible (anisotropic) magnetic scattering contributions the samples were measured in two positions, i.e. aligned horizontally and vertically with respect to the magnetic field. The two dimensional intensity curves were calibrated by the incoherent scattering of a water sample to account for the different detector pixel efficiency and to scale the data to absolute values. Figure 1: SANS instrument at PSI. In the foreground the collimation system can be seen, guiding the neutrons from the source on the right to the sample position. The vessel on the left contains the movable detector. The total length of the instrument is about 4 m. The data were averaged in sectors around the beam center in steps of 1, the width of each sector being 15, to compute the absolute intensity as a function of momentum transfer Q for different directions of the scattering vector Q. By fitting a Q -4 power law to the Porod Regime for each sector the apparent Porod constant in this direction is obtained. Assuming isotropic scattering around the axis of symmetry perpendicular to the surface plane (which was experimentally verified for selected coatings), the apparent Porod constant is known for all (3-dimensional) directions. Then the total specific surface area S total can be obtained from the 2-dimensional data by averaging the contributions for all 3 dimensions [1]. This orientational average then relates to S total as follows [4]:

3 d( Q) d ORIENTATION 2 2 S 4 Q TOTAL 2 is the neutron scattering length contrast between the pores and the matrix. 3a. Strain Determination by Profilometry To determine macroscopic strain values introduced during the thermal spraying process, bending measurements were performed by applying the technique of laser profilometry. Using an UBM laser profilometer a resolution of 3 data points per millimeter was chosen. For each sample analyzed three subsequent scans along the extended direction of the sample of size 1x25x5 mm 3 were made: One scan after the substrate was cut and annealed to relax residual stresses, one scan after grit blasting and a final scan after spraying. An average strain value is obtained by Stoney's equation [7] for thin coatings: Deposit 2 ESubst H Subst 6h ( 1 ) Deposit Subst E Subst is the elastic modulus of the substrate and Subst its Poisson ratio. is the curvature parameter as defined in [8]. It is obtained by fitting a circle to the bending data. The curvature is assumed to be introduced by the thermal spray process only. H Subst and h Deposit are the thicknesses of the substrate and the coating, respectively. The resulting Deposit is the calculated constant in-plane stress value in the coating. It has to be noted that generally the "ex-situ" curvature technique after spraying cannot account for stress gradients, and the assumption of a thin coating disregarding its elastic modulus can only give an approximate stress value. 3b. Strain Analysis by Neutron Strain Scanning Strains can be measured by diffraction from crystal planes according to the Bragg equation: 2d hkl sinwhereis the wavelength, d hkl the lattice spacing and 2 the diffraction angle. The strain is then defined as (d hkl -d hkl, )/d hkl,, where d hkl, is the lattice spacing of a stress free reference. Throughthickness strain measurements were performed at the high resolution powder diffractometer D1A at the Institute Laue- Langevin (ILL) in Grenoble / France. A newly designed setup employed two radially focusing collimators to define the gauge volume of size.6x1x1 mm 3. Bragg - peak positions as a function of coating depth were recorded by a 3 He two dimensional position sensitive detector which covers a range of about 6. From peak shifts in the coating or the substrate, residual strain values, averaged over the sampled volume, can be deduced. (1) (2) 1-2 mm thick NiCrAlY coatings deposited onto a 5 mm thick steel substrate of size 1x25 mm 2 were measured by horizontal scans through the gauge volume. The wavelength of.2994 nm was chosen using a Germanium monochromator crystal to yield an optimum 2 value near 9 for the {111} reflection of the Ni based NiCrAlY coating and the Fe {11} reflection of the substrate. To obtain the reference values an annealed free standing coating of the same spraying technique as well as annealed feedstock powder and an annealed substrate were used. The sample was mounted on a translation table and scanned through the gauge volume in steps of minimal.1 mm. One in-plane component parallel to the surface and the normal component perpendicular to the surface were measured in the transmission and the reflection setup respectively, as schematically shown in figure 2. The 2 calibration was obtained by the measurement of a stress free, annealed Fe powder. The position of the translation table holding the sample with respect to the neutron beam was determined from the integral area under the Bragg peak. Then, the effective gauge volume position, i.e. the position of the center of mass of the scattering material with respect to the scan direction, was calculated from the motor positions of the translation table as described in [9]. Near-interface corrections due to artificial peak shifts from a partially filled gauge (sampling) volume have to be applied. second measurement in reflection geometry, crystal planes parallel to surface sample scan in horizontal direction (x) first measurement in transmission geometry, crystal planes perpendicular to surface gauge volume Cd slit, width 1 mm 15 cm sample scan in horizontal direction (y) 15 cm neutrons neutrons horizontally focusing collimator, translation in horizontal direction Figure 2: Setup employed for the strain scanning at the powder diffractometer D1A at ILL in Grenoble. The position sensitive detector (PSD) collects neutrons that are Bragg - reflected from crystal planes inside the gauge volume in the coating or in the substrate, respectively. Results 1. Small Angle Neutron Scattering Figure 3 shows the scattering cross section from a NiCrAlY deposit as a function of momentum transfer measured for two P S D

4 sectors, parallel and perpendicular to the coating surface. Different heights of the two fitted lines, both representing a Q -4 behaviour, indicate different Porod constants for these two directions (at high Q the background of incoherent scattering dominates into which the Porod scattering merges). Since the value of the Porod constants obtained for each of these sectors does not by itself reflect the surface area of the scatterers in that particular direction, it often is called apparent Porod constant. The total specific surface area of the scatterers in the sample can be obtained from the apparent Porod constants by averaging over all directions, c.f. equation (1), provided that these are known. Furthermore, the anisotropy of the microstructure can be related to the angular dependence of the apparent Porod constant, even though the relationship is not simple. From the apparent Porod constant an apparent surface area in a specified direction can be defined [1]. d/d [1/cm] 1E31 1E3 1E29 1E28 1E27 main scattering contribution from surfaces perpendicular to surface plane "Porod Regime" main scattering contribution from surfaces parallel to the surface plane atmospheric plasma sprayed NiCrAlY Deposit.1 1 Q [1/nm] Figure 3: Example of the scattering intensity of an atmospheric plasma sprayed NiCrAlY coating as a function of the momentum transfer Q. Two sectors, one parallel to the coating plane and the other perpendicular, are shown to stress the anisotropy in the Q -4 "Porod regime" (in this graph, region of linear dependence). Figures 4 to 7 display the anisotropic apparent surface areas for the NiCrAlY coatings obtained by fitting a Q -4 law to the intensity I(Q) in each sector around the central beam. The results obtained by the same evaluation procedure, with the neutron beam passing perpendicular to the deposit top surface showed no anisotropy, proving that the in-plane microstructure of the investigated coatings was isotropic [1]. Measurements of samples mounted horizontally as well as vertically showed the same anisotropy in both positions, giving evidence that possible magnetic scattering contributions are negligible and do not require corrections to be applied. The low oxide content of the NiCrAlY deposits, determined to be less than 1% [3], ensures that only a negligible Porod scatter ing contributes from oxide-metal and oxide-pore interfaces. The calculated total specific surface areas of the NiCrAlY deposits for the investigated spraying techniques and spraying parameters are collected in table 1 and visualized in figure 8. Table 1: Specific surface areas obtained from the Q -4 region by averaging the Porod constants. The estimated error for the results is about 5%. Process and Parameter Specific Surface Area S total [m²/m³].91 x 1 6, a.86 x 1 6, b.71 x 1 6, c.76 x 1 6, a.77 x 1 6, b.76 x 1 6, c.71 x 1 6, a 1.2 x 1 6, b 1.12 x 1 6, c 1.11 x Density, Oxide Content and Porosity Table 2 presents the density, the porosity and the wear rate (according to ASTM standard G75-95) of the NiCrAlY coatings. The table is taken from [3], where details of the measurements are described. The neutron scattering contrast 2 was calculated on the basis of these density data. Table 2: Density, porosity and wear rates of the NiCrAlY coatings, from [3]. Process and Parameter Density [g/cm 3 ] Porosity [%] Wear Rate [g/km] 6,44 11,4,768, a 6,875 3,9,95, b 6,889 4,6,886, c 6,914 4,1,173, a 6,625 8,1,1222, b 6,494 9,9,1295, c 6,619 8,4,1366, a 6,368 11,9,58, b 6,451 1,9,711, c 6,566 9,3, Microstructure - Properties Relation Figure 9 displays the total specific surface areas of the Ni- CrAlY deposits as a function of their porosity, as listed in table 2, for the different spraying techniques. The wear rates as a function of the total specific surface area are displayed in figure 1. The data points in the graphs are labeled by the respective spraying technique.

5 Apparent Surface Area [m 2 /m 3 ] Apparent Surface Area [m 2 /m 3 ] Figure 4: Anisotropic apparent surface area as a function of the azimuthal angle around the incident beam of an - NiCrAlY coating, deduced from scattering curves as shown in figure 3 by fitting a Q -4 - power law to each sector. Figure 5: As figure 4, for a - NiCrAlY coating. 9 9 Apparent Surface Area [m 2 /m 3 ] Apparent Surface Area [m 2 /m 3 ] Figure 6: As figure 4, for a - NiCrAlY coating. Figure 7: As figure 4, for a - NiCrAlY coating. 4 Strain Measurements Figure 11 displays a through-thickness strain profile obtained by the neutron strain scanning at the instrument D1A. The coating material is atmospheric plasma sprayed NiCrAlY. For comparison, two measurements - the one from the coating and the other from the substrate - are collected in a single graph. The maximum surface error to be corrected is d/d [1-6 ]=6 [11]. The data for the substrate are not affected by the surface error, since a vertical scan was performed. The lateral resolution in this scan was 1 mm. The results of the curvature measurement for the same atmospheric plasma sprayed NiCrAlY sample are given in table 3. Different signs of the curvature imply a bending of the sample in different directions. The compressive strain introduced by the grit blasting is overcompensated by tensile "quenching" stresses induced during the deposition. H subst = 4.6 mm and h Deposit =.95 mm are the thicknesses of the substrate and the coating, respectively. The bulk modulus of the steel substrate was taken to be E subst = 29 GPa [9]. According to equation (2) a macroscopic average stress in the coating of Deposit = 221 MPa is calculated. Table 3: Curvature data of the NiCrAlY coating. after grinding and annealing after grit blasting after spraying curvature [1/m]

6 Specific Surface Area S TOTAL [m 2 /m 3 ] 1.3x x1 6 9.x1 5 7.x1 5 NiCrAlY sample number Figure 8: Total specific surface areas of the NiCrAlY coatings obtained for the different spraying techniques. of the points of each spraying technique indicates a "typical" microstructure, even though the spray parameters were varied with the intention to produce a wide range of microstructures for each spraying technique. This can be explained when each of the spray techniques produces different, but "typical" pore sizes and shapes. Some basic conclusions might be drawn about the pore sizes from figure 9. The and deposits have quite the same surface area, whereas the porosity volume is twice as high compared to the deposits. A smaller porosity volume at almost the same specific surface area could be explained by a smaller pore size with the same or similar pore microstructure. This is confirmed by the optical micrographs of these deposits as shown in [3], indicating (on average) smaller pores in the deposits. The deposits have about the same porosity as the coating, but a higher specific surface area. However, a similar comparison as for the and coatings is more difficult since the micrographs of the and deposits show quite a different microstructure. 1.3x1 6 Specific Surface Area S TOTAL [m 2 /m 3 ] 1.1x1 6 9.x1 5 7.x1 5 NiCrAlY Porosity [%] Wear Rate [g/km] NiCrAlY Specific Surface Area S TOTAL [m 2 /m 3 ] Figure 9: Relationship between the porosity volume fraction and its specific surface area for the NiCrAlY deposits for different spray techniques. Discussion 1. Small Angle Neutron Scattering The total specific surface areas of the NiCrAlY deposits displayed in figure 8 show differences between the spraying techniques within a factor of two. The scatter of total surface areas for each of the spraying techniques is due to the variation of microstructures created by varying the spray parameters. This scatter is less than the difference between the spraying techniques. This indicates that the specific surface area is a characteristic of the spraying technique rather than of the spray parameters applied for one technique. The relationship between the porosity volume fraction and the specific surface area (figure 9) shows the same principal trend. The clustering Figure 1: Relationship between the total specific surface area of NiCrAlY coatings and their wear rates according to ASTM G The estimated errors are about 1% for the wear rates [3] and about 5% for the total specific surface areas. Special attention should be drawn to the anisotropy present in these coatings. The and NiCrAlY coatings show similarly pronounced apparent surface area anisotropy (figures 4 and 5), whereas the and coatings are more isotropic (figures 6 and 7). The and apparent surface area anisotropy can well be described by a single ellipsoid, which indicates that one void system in the microstructure is dominant. The apparent surface area anisotropy of the coating (figure 6) seems to consist of two ellipsoids perpendicular to each other, suggesting that two significantly scattering void systems with the main orientations perpendicular to each other

7 are present in the microstructure. Similarly, the apparent surface area anisotropy (figure 7) seems to be better described by two ellipsoids, even though due to lower overall anisotropy it is more difficult to reliably separate them. The orientation of the ellipsoids indicates that one dominating void system in all the deposits is that of interlamellar pores oriented mostly parallel to the substrate. The weight of the other system, most likely cracks, which are dominantly found in ceramic deposits [1] and mostly oriented perpendicular to the substrate, is less relevant in the metallic NiCrAlY deposits. It should further be noted that by the nature of the surface characterization of the voids by their surface area, large globular pores with small surface area, possibly present in the microstructure, have almost no weight. These can better be characterized by the so called multiple small-angle neutron scattering method (MSANS), which is currently planned to be applied to these deposits. 2. Anisotropic Specific Surface Area and Properties Relations The work reported here is part of a larger project in progress [3]. Some of the data obtained in this project are discussed here to preliminary document possible relationships between the (anisotropic) SANS results and macroscopic properties with the aim to illustrate the importance of the presented microstructural characterization. More detailed relationships are currently being studied for all of the materials involved in this project. The SANS results for NiCr deposits [2] showed similar anisotropies as the SANS data presented in this work for NiCrAlY materials. The NiCr data are compared with ratios of the electrical resistivity parallel and perpendicular to the coating surface as presented in [3]. The results are given in table 4. The values of table 4 indicate a direct relationship between the anisotropy of the apparent Porod surface area and the anisotropy of the electrical conductivity. Investigations in the literature [12] find similar relationships between the SANS anisotropy and the elastic modulus. Understanding these relationships may, in the future, allow modeling the properties of these deposits and therefore bring about a new level of microstructure-properties understanding. Further investigations are in progress. Table 4: Ratios of electrical resistivity of NiCr deposits for different spraying methods and SANS surface area anisotropy expressed as aspect ratio of ellipsoids fitted to the apparent Porod surface area. HVOF,9 3,1 3,8 8,6 SANS aspect ratio Figure 1 displays the wear rate as a function of the total specific surface area obtained by the SANS measurements. The graph indicates that for the investigated NiCrAlY coatings low wear rates are obtained by spraying techniques like and, which produce coatings of high specific surface areas. The low oxide content in all investigated NiCrAlY deposits (determined to be less than 1%) could be a reason for the here found pronounced dependence on the specific surface area. When more (hard) oxide phases were present, they might have dominated the wear behavior and then covered the influence of the specific surface area. 3. Neutron Strain Scanning and Laser Profilometry Figure 11 displays the in-plane strain, oriented parallel to the surface, as a function of depth into the coating as well as into the substrate for the NiCrAlY sample. In the coating the strains are tensile and highest at the surface. This can be explained by the fact that the spray process generates tensile quenching stresses in the coating, and that on top no more layers under tensile strain are deposited which could introduce compressive strains in the deeper layers and partly compensate the tensile strains. The data point nearest to the interface is already in the compressive region. In the substrate compressive strain is found near the interface, whereas deeper inside, near the uncoated surface, the stresses tend to be tensile. The compressive strain near the interface can be explained by the fact that the tensile stresses in the coating must be balanced by compressive strains inside the substrate. d/d [1-6 ] 2 interface NiCrAlY coating and substrate in-plane strain component statistical error coating -2 substrate effective gauge volume position [mm] Figure 11: Through thickness strain profile in a thermally sprayed NiCrAlY coating and its substrate. The combination of tensile quenching stresses introduced in the coating by the thermal spray process and compensating compressive stresses in the substrate leads to the measured bending of the whole sample. A conventional bending test clearly cannot resolve strain gradients in a material. As seen in figure 11 the technique of neutron scattering is able to resolve existing gradients - limited by the spatial resolution of the technique.

8 Bending tests and neutron scattering results can be compared by the resulting average (macroscopic) stress in the coating, however, it has to be noted that different elastic constants appropriate for each technique have to be applied. The average macroscopic stress value obtained by the bending is Deposit = 221 MPa. For the diffraction methods, it can be assumed that the crystal planes in the sample behave like bulk material. Then, the elastic constant obtained from single crystal data for the specific {hkl} reflection is an appropriate choice. The model of Eshelby Kroener [13] further accounts for the polycrystalline nature of the coating. Using the Ni {111} values E Deposit = 233 GPa and Deposit =.29 [8] and an average strain value of d/d [1-6 ] = 65, estimated from the coating strain data displayed in figure 11, one obtains an average stress value in the coating of ' Deposit = 213 MPa from the diffraction technique. This value is in excellent agreement with the average stress Deposit = 221 MPa calculated from the bending measurements. However, with respect to the approximations made for the average strain in the coating as well as the application of the simplifying Stoney's formula one should refrain from overweighting this agreement but rather regard it as indication of principal consistency of the methods. Detailed analysis with more precise calculations is in progress. Conclusion Among other techniques Small Angle Neutron Scattering was applied to characterize the microstructure of thermally sprayed coatings. For NiCrAlY coatings the total specific surface area as well as the anisotropy of the apparent surface area are found to be different for the different spraying techniques. At the same time the porosity volume varies significantly for different spraying techniques. These results, in combination, allow a more detailed characterization of the voids microstructure in these materials. The importance of this type of characterization is reflected in a qualitative comparison between the anisotropy of the apparent surface area and macroscopic properties, here the electric resistance in the case of the NiCr coatings. This comparison reveals a direct relationship between an anisotropy in the microstructure and the anisotropy of the electric resistance. Further, the wear rates of the NiCrAlY deposits show that the total specific surface area measured in these coatings by Small Angle Neutron Scattering is a valuable parameter to characterize the underlying microstructure. Strain measurements by conventional bending tests reflect the overall stress state in a deposit, whereas neutron strain scanning allows to resolve strain gradients. For an atmospheric plasma sprayed deposit a residual strain profile in the coating as well as in the substrate obtained by the neutron strain scanning is presented and compared with the bending results. Further refinement of the data analysis (i.e. elastic constants and surface corrections) is in progress. Measurements on samples manufactured by the other spraying techniques will be per formed to analyze spray technique specific stress profiles and average stress values with the aim to relate the residual stresses to the microstructure and properties of the deposits. Acknowledgments The authors would like to acknowledge the partial Eureka / KTI grant obtained by the THERMETCOAT project. References 1. J. Ilavsky, G.G. Long, A.J. Allen, L. Leblanc, M. Prystay and C. Moreau, "Anisotropic Microstructure of Plasma- Sprayed Deposits", Proc. ITSC 1998, Nice, France, J. Ilavsky, J. Pisacka, P. Chraska, N. Margadant, S. Siegmann, W. Wagner, P. Fiala and G. Barbezat, "Microstructure - Wear and Corrosion Relationships for Thermally Sprayed Metallic Deposits", Proc. ITSC 2, Montreal, Canada, N. Margadant, S. Siegmann, J. Patscheider, T. Keller, W. Wagner, J. Ilavsky, J. Pisacka, G. Barbezat and P. Fiala, "Microstructure - Property Relationships and Cross- Property-Correlations of Thermally Sprayed Ni-Alloy Coatings", submitted to the Proc. ITSC 21, Singapure. 4. G. Porod, in "Small Angle X-ray Scattering" edited by O. Glatter and O. Kratky, Academic Press, London, (1982), W. Wagner, "New Instruments and Science around SINQ", PSI Proceedings 96-2, edited by A. Furrer, (1996), J. Kohlbrecher and W. Wagner, "The New SANS Instrument at the Swiss Spallation Source SINQ", J. Appl. Cryst. 33 (2), G. Stoney, "The Tension of Metallic Films Deposited by Electrolysis", Proc. Roy. Soc., London, A 82, (199), T.W. Clyne and S.C. Gill, "Residual Stresses in Thermal Sprayed Coatings and Their Effect on Interfacial Adhesion: A Review of Recent Work", J. Thermal Spray Technology, 5 (1996), J. Matejicek, "Processing Effects on Residual Stress and Related Properties of Thermally Sprayed Coatings", PhD Dissertation, SUNY, Stony Brook, (1999). 1. J. Ilavsky, A.J. Allen, G.G. Long and S. Krueger, "Influence of Spray Angle on the Pore and Crack Microstructure of Plasma-Sprayed Deposits", J. Am. Ceram. Soc., 8, 3 (1997), T. Pirling, "A New High Precision Strain Scanner at the ILL", Materials Science Forum, (2), J.S. Wallace, J. Ilavsky, "Elastic Modulus in Plasma Sprayed Deposits", Proc. ITSC 1997, Indianapolis, Indiana, USA, E. Kroener, Z. Physik, 151, (1958) 54.