Patent Application
20100031720
The present invention relates to a manufacturing method and apparatus for producing plates and sheets with shear texture or minimal through-thickness texture gradient, or both.
The crystallographic texture of a plate or sheet plays an important role in many applications. Crystallographic texture is crucial for the performance of the sputtering targets used to deposit thin films, due to the dependence of the sputtering rate on crystallographic texture.
The uniformity of thin films deposited from a sputtering target with non-uniform crystallographic texture is not satisfactory. Only a plate with uniform texture throughout its volume will give optimum performance.
The rate of sputtering from a grain in the target depends on the orientation of the crystal planes of that grain relative to the surface (ref. Zhang et al, Effect of Grain Orientation on Tantalum Magnetron Sputtering Yield, J. Vac. Sci. Technol. A 24(4), July/August 2006); the sputtering rate of each orientation relative to the plate normal is different. Also, certain crystallographic directions are preferred directions of flight of the sputtered atoms (ref. Wickersham et al, Measurement of Angular Emission Trajectories for Magnetron-Sputtered Tantalum, J. Electronic Mat., Vol 34, No 12, 2005). The grains of a sputtering target are so small (typically 50-100μm diameter) that the orientation of any individual grain has no significant effect. However, over a larger area (an area roughly 5 cm to 10 cm diameter) texture can have a significant effect. Thus, if the texture of one area on the surface of a target is different from the texture of any other area, the thickness of the film produced is unlikely to be uniform over the whole substrate. Also, if the texture of a surface area is different from that of the same area at some depth into the target plate, the thickness of the film produced on a later substrate (after the target is used, or eroded, to that depth) is likely to be different from that produced on the first substrate.
So long as the texture of one area, then, is similar to that of any other, it is not important what that texture is. In other words, a target plate in which every grain has a 111 orientation parallel to the plate normal direction (ND) is no better and no worse than one in which every grain has a 100 orientation parallel to ND, or than one which consists of a mix of 100, 111 and other grains, so long as the proportions of the mix remain constant from area to area.
Uniformity of film thickness is of major importance. In integrated circuits, several hundred of which are created simultaneously on a silicon wafer, for example, too thin a film at one point will not provide an adequate diffusion barrier, and too thick a film at another point will block a via or trench, or, if in an area from which it should be removed in a later step, will not be removable. If the thickness of the film deposited is not within the range specified by the designer, the device will not be fit for service, and the total cost of manufacture up to the point of test is lost, since no repair or rework is normally possible.
If the target does not have uniform texture, and thus does not provide a predictable, uniform sputtering rate, it is impossible, with state-of-the-art sputtering equipment, to control the variation of thickness from one point on the substrate to another. Partial, but not total, control of variation of thickness from substrate to substrate, and from target to target, is possible using test-pieces. Use of test-pieces, however, is time-consuming and costly.
With targets made according to the prior art, the non-uniformity of texture found in the target plate causes unpredictability or variability in the sputtering rate (defined as the average number of tantalum atoms sputtered off the target per impinging argon ion), leading to variations in the thickness of the film produced on a particular substrate, and also variations in film thickness from substrate to substrate and target to target.
Crystallographic texture also affects the mechanical behavior of a material. This is due to differences in the mechanical behavior of a single crystal of an anisotropic material when tested in different directions. Although single crystal materials are used in various applications, the majority of materials used in practice are polycrystals, which consist of many grains. If the grains forming a polycrystal have a preferred orientation (i.e. crystallographic texture), the material tends to behave like a single crystal having similar orientation. The formability of a material depends on the mechanical behavior of the material, which is a strong function of crystallographic texture.
Other material properties such as magnetic permeability are also influenced by crystallographic texture. For example, crystallographic texture is an important factor for the performance of a grain-oriented silicon steel, which is mainly used as the iron core for transformers and other electric machines. Improved magnetic properties, such as high magnetic permeability of the grain-oriented silicon steels, result in energy savings. To achieve good magnetic properties, a grain-oriented silicon steel should have strong <110>//ND and <100>//RD (rolling direction) texture (Goss orientation), which can then be easily magnetized in the rolling direction.
Crystallographic texture develops as a material is plastically deformed, and plastic deformation can only occur along certain slip systems that become active during deformation. Normal and shear strain components, along with other parameters such as temperature, determine which slip systems become active. Activation of a slip system causes grains to rotate towards a certain orientation, resulting in a crystallographic texture. The final crystallographic texture of a material is a strong function of both the starting texture and the strain induced in the material.
For example, during rolling of a plate in plane strain condition, material through the thickness of the plate is subjected to shear and normal strains simultaneously. The amount of shear strain varies significantly through the thickness of a plate. The mid-thickness of a plate is not subjected to any shear strain due to the symmetry of a conventional rolling process, whereas locations away from mid-thickness experience both shear and normal strains. Therefore, texture at the mid-thickness of a plate is considerably different than other locations.
Non-uniformity of texture through the thickness of a plate is referred to as the “through-thickness texture gradient”. Conventional rolling produces a plate or sheet with a strong through-thickness texture gradient. Neither the through-thickness texture gradient nor the main components of texture can be altered significantly by parameters which are varied and controlled in conventional rolling, such as % reduction in thickness per pass and rotation between passes.
Certain texture components, i.e. “rolling texture” components, become dominant in conventional rolling. Rolling texture components for a bcc metal are different than “shear texture” components, which form when a bcc metal is subjected to shear strain. When subjected to shear strain, the grains in a bcc metal rotate towards <110>//ND. An almost opposite behavior is observed for a fcc metal, which, when subjected to shear strain, will cause <111>//ND and <100>//ND to become the major texture components. The greater the shear strain introduced in a workpiece, the stronger the shear texture developed.
In a material (fcc or bcc) with a perfectly random texture, 10.2% of the volume (and 10.2% by number of the grains) has a <100> axis within 15-deg of ND. Another 13.6% of the volume has a <111> axis within 15-deg of ND and a further 20.4% of the volume has a <110> axis within 15-deg of ND. Therefore, a fcc material is said to have a shear texture if more than 10.2% of the volume has a <100> axis within 15-deg of ND, and more than 13.8% of the volume has a <111> axis within 15-deg of ND. A bcc material is said to have shear texture if more than 20.4% of the volume has a <110> axis within 15-deg of ND.
A higher plastic strain ratio (r-value) is known to enhance formability of a metal, and a bcc or fcc metal with a dominant <111>//ND texture component has higher plastic strain ratio (r-value). Therefore, shear texture with <111>//ND as one of the major components is desirable for improving the formability of a fcc metal.
The amount of shear strain through the thickness of a plate or sheet can be altered by switching from a conventional (symmetric) rolling to an asymmetric rolling process. The total amount of shear strain through the thickness can be increased, and more specifically, the mid-thickness can be subjected to some amount of shear strain, which is not possible in conventional rolling. Prior art asymmetric rolling methods include use of rolls with different diameters, rolls with different rotational speeds, and rolls with different surface properties that result in different friction coefficient between the top surface of a workpiece and the top roll, and the bottom surface of a workpiece and the bottom roll. Due to the difficulties in controlling the friction coefficient consistently, asymmetric rolling with different friction coefficients top and bottom is impractical and is excluded from further discussion here. These prior art methods can also be used to decrease the through-thickness texture gradient.
The application of the above-mentioned types of asymmetric rolling for introducing shear texture and minimizing texture gradient have been described in the prior art. See, e.g., Field et al., Microstructural Development in Asymmetric Processing of Tantalum Plate, J. Electronic Mat., Vol 34, No 12, 2005; Sha et al., Improvement of recrystallization texture and magnetic property in non-oriented silicon steel by asymmetric rolling, J. Magnetism and Magnetic Mat., Vol 320, 2008; Lee and Lee, Analysis of deformation textures of asymmetrically rolled steel sheets, Internat. J. Mech. Sci., Vol 43, 2001; Lee and Lee, Texture control and grain refinement of AA1050 Al alloy sheets by asymmetric rolling, Internat. J. Mech. Sci., Vol 50, 2008; Jin et al. Evolution of texture in AA6111 Al alloy after asymmetric rolling with various velocity ratios between top and bottom rolls, Mat. Sci. and Eng., Vol 465, 2007; Jin et al. The reduction of planar anisotropy by texture modification through asymmetric rolling and annealing in AA5754, Mat. Sci. and Eng., Vol 399, 2005; Kim et al. Formation of textures and microstructures in asymmetrically cold rolled and subsequently annealed aluminum alloy 1100 sheets, J. Mat. Sci., 2003; Zhang et al. Experimental and simulation textures in an as symmetrically rolled zinc alloy sheet, Scripta Materialia, Vol 50, 2004; and Kim et al. Texture and microstructure changes in asymmetrically hot rolled AZ31 magnesium alloy sheets, Mat. Lett. 59, 2005.
The asymmetric rolling methods described above introduce some amount of shear strain through the thickness of the plate by using asymmetry in the top and bottom roll diameter or the top and bottom roll speed. As the roll diameter or roll speed ratios of the top and bottom rolls increase, the shear strain introduced in the plate increases, but there are practical limits to these ratios and the amount of shear strain that can be introduced with these methods.
Accordingly, the present invention provides an apparatus and a rolling method for controlling the crystallographic texture of a material to improve the related material properties and enhance the performance of the material. The present invention allows the introduction of a controlled amount of shear strain through the thickness of a plate or a sheet, which results in plates and sheets with minimal through-thickness texture gradient. A minimal through-thickness texture gradient in sputtering targets improves the predictability and uniformity of the thickness of the films produced, and thus improves the ease of use of the targets.
The introduction of shear strain can also provide shear texture that results in better formability of materials, such as fcc metals, which increases the yield and decreases processing costs for forming operations used widely in many industries.
The improved shear texture also improves the magnetic properties (i.e. magnetic permeability) of the materials such as grain oriented silicon steel. Improved magnetic properties result in energy savings as grain oriented silicon steel is used as iron core for transformers and other electric machines.
In the present invention, the workpiece (a plate or sheet) is tilted about an axis parallel to the axis of the rolls in a rolling mill with a prescribed angle (tilt angle). The tilted workpiece is fed into the rolls and the entry tilt angle is maintained during the entire rolling pass. As used herein, this process is referred to as “tilt rolling”. The material through the thickness of the workpiece is sheared as a result of tilt rolling. The amount of shear strain can be controlled by the tilt angle along with other rolling parameters that are normally controlled in conventional rolling. Multiple passes are used to reduce the thickness of the workpiece to the desired value.
Tilt rolling can be achieved by a specially designed rolling mill with aprons that can be tilted to different angles. In an embodiment, the tilted apron is an integral part of the rolling mill. This permits utilization of a rolling mill for both conventional and tilt rolling with very quick change-over. In another embodiment, tilt rolling can also be implemented in a conventional rolling mill by means of a fixture that can be easily installed on the mill without major modifications. In this embodiment, the initial investment for equipment is smaller, and the rolling mill can be used for both conventional and tilt rolling, but the change-over time is greater than the specially designed rolling mill described above. However, in both embodiments, a relatively small change-over time between conventional and tilt rolling provides production flexibility unlike the alternative asymmetric rolling processes that require increased time for change over, resulting in greater down-times for the equipment.
Accordingly, in one aspect the present invention provides a method of rolling a metal plate or sheet, the method comprising the step of feeding the plate or sheet into rollers in a rolling mill at an angle of between 2-20 degrees above or below horizontal.
In an additional aspect, the present invention provides an apparatus for rolling a metal plate or sheet at an angle, the apparatus comprising a rolling mill having a tilted feed table inclined at an angle of between 2 and 20 degrees above or below horizontal.
These and other aspects of the invention will become more readily apparent from the following figures, detailed description and appended claims.
The invention is further illustrated by the following drawings in which:
a), 2(b) and 2(c) are diagrams depicting finite element modeling of (a) asymmetric rolling with different roll diameters, (b) asymmetric rolling with different roll speeds and (c) tilt rolling.
As used herein in the specification and claims, including as used in the examples and unless otherwise expressly specified, all numbers may be read as if prefaced by the word “about”, even if the term does not expressly appear. Also, any numerical range recited herein is intended to include all sub-ranges subsumed therein.
A—Tilt-Roll Process
The tilt-rolling process of the present invention, as shown in two embodiments in
The amount of shear strain introduced through the thickness of a material by tilt-rolling can be controlled by adjusting the parameters such as tilt-angle and % reduction in thickness after each pass, as explained below. The ability to control the amount of shear strain in a workpiece with the methods of the present invention permits achievement of two types of special texture in a plate or a sheet: 1) minimal through-thickness texture gradient, 2) shear texture throughout the thickness of the workpiece.
The angle selection depends on the primary objective of a user for using the tilt-rolling process; minimizing through-thickness texture gradient or inducing shear texture. Preferably, the angle of tilt above or below horizontal is between 2 and 20 degrees. To minimize the through-thickness gradient, the angle of tilt is preferably between 3 and 7 degrees. To increase the shear texture, preferably the angle of tilt is between 10 and 20 degrees.
As a general rule, shear texture can be more effectively introduced in a material as the tilt angle increases. However, the through-thickness texture gradient does not necessarily decrease with a larger tilt-angle. The % thickness reduction and tilt angle should be adjusted together for a given thickness of a workpiece to achieve minimal through-thickness texture gradient. The simulation methods for optimizing each important parameter are given in detail below, and one skilled in the art can adjust these parameters in additional simulations to balance the various effects and achieve the desired result, a final product in the form of a plate or a sheet with predominantly shear texture or minimal texture gradient.
The tilt angle may be above or below horizontal, depending on the pass. For rolling plate (single stand mill), the angle should be above horizontal because gravity is then used for locating the workpiece on the tilted-feed table. If a multi-stand mill is used for rolling sheet, the direction of the tilt angle is preferably alternated to save space in vertical direction and to distribute the effect of tilt rolling evenly to top and bottom halves of the sheet.
The strain in a workpiece has a direct influence on the “deformation texture”, a well-known term in the art. After a material is strained using metal working methods (rolling in this case), the workpiece is preferably annealed by increasing the temperature of the workpiece above the recrystallization temperature to achieve recrystallization, especially if the metal working process is performed cold (near or below room temperature) or warm (above room temperature and below recrystallization temperature). If the workpiece reaches a temperature above recrystallization temperature during metal working processing, dynamic recrystallization may occur and the annealing step after metal working may not be necessary. The texture of a workpiece may change during recrystallization and the resulting texture is known as “recrystallization texture”. However, the recrystallization texture of a workpiece is a strong function of the deformation texture. Therefore, the benefits of the tilt rolling methods of the present invention can be realized for cold, warm or hot rolling.
A metal plate or sheet can be passed through the rolls at a tilt more than once, in other words, 2, 3, 4, 5 or more passes. The passes are repeated until the desired thickness of the workpiece is reached. If symmetric texture about the mid-thickness of a workpiece is desired, especially for minimizing the through-thickness texture gradient, the % thickness reduction should be adjusted so that the minimum number of passes to reach the final thickness is preferably at least four or greater. Another consideration for maximum % reduction is the load on the mill. The % thickness reduction should be kept lower than a % reduction that would result in an excessive load on the mill.
Finite element simulations were used to compare the shear strain levels developed in a workpiece rolled with the tilt rolling methods of the present invention and other asymmetric rolling methods. Finite element simulation permits calculation of the amount and direction of strains in a workpiece, which is very difficult to accomplish in experiments. Finite element simulations are used as a tool here to quantify the influence of tilt rolling as compared to other asymmetric rolling methods. A finite element software package, Deform 2-D available from Scientific Forming Technologies Corp., Columbus, Ohio, was used for all the simulations.
The simulations were set up for rolling a workpiece of an initial thickness of 0.5″ in one pass.
For the simulation of rolling with different roll speeds, the diameter of top and bottom rolls was set at 16″. The rotational speed of the faster roll (1 in
The tilt rolling simulation used a roll diameter of 16″ and a roll speed of 1 radian/second (approximately 10 rpm).
The friction coefficient, roll diameter and roll speed affect the simulation results quantitatively, but the conclusions drawn from the simulation results for the qualitative evaluation of different processes is not influenced significantly by the selection of these parameters.
Tantalum, a bcc metal, was selected as the workpiece material. It is important to note that the amount of shear strain obtained in a material will be very similar in different materials for a given set of rolling parameters. However, the resulting texture due to the shear strain will vary based on the material. Therefore the simulation results for the shear strain are not influenced significantly by the material selected in the simulations.
Shear strain accumulates as a material goes through the rolls. The material is sheared in one direction at the entrance and the shear direction changes as the material passes the neutral point in rolling. The “cumulative” shear strain was calculated by the summation of the absolute values of the positive and negative shear components. The average cumulative shear strain through the thickness was calculated by averaging the shear strain of evenly spaced 5 locations from the top to the bottom surface of the workpiece.
Tilt-rolling with a tilt angle of 15-deg achieves a shear strain similar to that achieved by asymmetric rolling with roll diameter ratio of 2.
The amount of shear strain introduced by any of the asymmetric rolling methods, including tilt rolling, depends on the thickness of the workpiece and the % reduction in thickness per pass. For example, if tilt-rolling is compared to other asymmetric rolling methods for the same thickness (0.5″) and higher % reduction (for example 10%), slightly different results are obtained from the results presented in
When the % thickness reduction is 10% per pass, the amount of shear strain averaged through the thickness for 5-deg tilt-rolling was equivalent to the amount of shear strain obtained by asymmetric rolling with diameter ratio of 1.65 and a speed ratio of 4. Tilt rolling with a 10-deg tilt angle, produced shear strain similar to diameter ratio of 2.
In light of these results, it can be concluded that the tilt-rolling process introduces shear strain in a material more effectively than other asymmetric rolling methods considering the limitations of each method. A tilt angle as low as 5 degrees causes equivalent or more shear strain when compared to asymmetric rolling with diameter ratio of 1.6 or asymmetric rolling with roll speed ratio of 4. Practical difficulties for implementing the process in a rolling mill may become severe for asymmetric rolling methods with roll diameter of 1.6 or speed ratios of 4, whereas no practical difficulty is encountered for tilt-rolling up to a tilt angle of 15 or 20 degrees.
The shear strain through the thickness of a workpiece is neither uniform nor symmetric about the mid-thickness in one pass of tilt-rolling.
In order to distribute the shear strain uniformly to the top and bottom halves of a workpiece, the workpiece may be turned over after each tilt-rolling pass or at regular intervals such as after every second pass. The frequency of turn-over of the workpiece is dependent on the requirements for the uniformity of the shear strain through the thickness of the workpiece. In order to minimize the through-thickness texture gradient, the variation of shear strain through the thickness should be decreased. The average shear strain for top and bottom surface (S), top and bottom quarter (Q), and mid-thickness (M) is plotted in
Curling of a workpiece during conventional rolling can be a major problem in production if curling makes it difficult to feed the workpiece into the rolls or if the leading edge of the workpiece hits and damages the apron on the exit side of the mill. In addition to these practical difficulties, curling affects the normal strains in the workpiece and results in additional strain and texture non-uniformity. As a workpiece curls in rolling, additional strain due to curling is induced in the material. Strain due to curling reaches its maximum near the surface and decreases to zero at mid-thickness. The effect of curling on texture may be evaluated by comparing the maximum strain due to curling with the normal strain in rolling. Curling may also occur in tilt-rolling and other asymmetric rolling methods, unless minimized as follows.
It is known that curling of a workpiece may be minimized by optimizing the % reduction for a certain thickness in asymmetric rolling with different roll speeds. See, e.g., Shivpuri et al., ‘Finite element investigation of curling in non-symmetric rolling of flat stock’, Int. J. of Mech. Sci., Vol. 30, 1988; and Knight et al., ‘Investigations into the influence of asymmetric factors and rolling parameters on strip curvature during hot rolling’, J. Mat. Proc. Tech., Vol. 134, 2003.
The same concept can be applied to tilt-rolling. The simulation results presented in
As used herein, the term “substantially no curling” refers to achieving a maximum curl strain that is 10% or less of normal strain. This can be achieved by using a predetermined % reduction, as explained above.
It is also necessary to control roll roughness and lubrication to ensure that the curling of a workpiece is consistent from pass to pass or from workpiece to workpiece. If the roll roughness and lubrication of top and bottom rolls are different, the friction coefficient between the top roll and workpiece, and the bottom roll and workpiece become different. This variation of the friction coefficient causes inconsistency in curling behavior and excessive curling may occur even when the % reduction is optimized for a given tilt angle and thickness of the workpiece. The rolls and the workpiece are preferably flooded with lubrication to increase the uniformity of the friction coefficient.
Another important factor in determining final texture is the starting texture of the workpiece. If the texture of the starting workpiece is not favorable, it will be difficult to achieve the benefits of tilt rolling by the methods of the invention. For example, if the texture of the starting workpiece before rolling is non-uniform, the texture after tilt rolling is likely to be non-uniform even though the strains induced in the tilt rolling are substantially uniform.
Depending on the requirements for the final product, a workpiece may be optionally tilt-rolled in some passes and conventionally rolled in other passes. The rolling practice used in conventional rolling is preferably applied to meet additional requirements of the final product.
B—Tilt-Roll Fixture
Conventional rolling mills for rolling metal plate and/or sheet are well known in the art. In a typical rolling mill, each of the work rolls will be substantially the same diameter and operate at substantially the same rolling speed.
A conventional rolling mill may be re-designed and manufactured to permit tilting the aprons about an axis parallel to the axis of the rolls. A schematic of such a rolling mill is depicted in
As an alternative to a specially designed rolling mill, tilt rolling can be achieved by means of a tilt-roll fixture, which can be installed on a conventional rolling mill without major modifications. This gives a production facility more flexibility.
An embodiment of a tilt roll fixture that can be used in the methods of the present invention is shown in
The fixture is supported by the cross-bar (10) attached to the mill frame (6a-b) to prevent the fixture from being pulled into the work rolls. As an alternative to the cross-bar, the tilted-feed table may be bolted on the apron (8) if the apron is strongly supported structurally. The shims (11a-b) between the mill frame (6a-b) and cross-bar (10) permit adjustment of the tilted-feed table horizontally and the shims (13) between the apron (5) and tilted-feed table (4) in
A workpiece tends to curl in conventional rolling as well as tilt rolling process. If the workpiece curls in one pass, it becomes difficult to feed the workpiece into the rolls for the next pass. This may become a severe problem for both conventional and tilt rolling. This can be managed as shown in
In order to achieve the benefits of tilt rolling throughout the workpiece, the tilt angle should be retained during tilt-rolling. A workpiece tends to get pushed to horizontal once the trailing edge comes off the tilted-feed table. When this happens tilt-rolling changes to conventional rolling, and the benefits of tilt-rolling cannot be obtained in the material that is being rolled.
To minimize this effect, it is important to minimize the distance between the work rolls and the tip of the tilted-feed table (15).
The tilt-angle also cannot be retained if a workpiece is not fed into the rolls with conditions for “perfect entry”, where both top and bottom edges of the workpiece make contact with the top and bottom rolls simultaneously. When perfect entry is not established, the tilt angle of the workpiece is different than the tilt angle of the tilted-feed table.
In addition to controlling the tilt angle, perfect entry is required for maintaining a large contact area between the workpiece and the tilted-feed table. If a workpiece is not fed with conditions for perfect entry, the contact between the workpiece and the tilted-feed table is reduced from area contact to line contact, either at the tip of the table or at the trailing edge of the workpiece. Line contact may cause excessive contact pressure on the tilted-feed table or the workpiece that may cause defects in the table or the workpiece.
To achieve perfect entry, the tip of the tilted-feed table should be correctly positioned; position will vary as a function of the thickness and % reduction per pass. Once the tip of the table is positioned in the vertical direction to ensure perfect entry, the tip of the table is preferably positioned in the horizontal direction to move the tip as close to the rolls as possible. Therefore, the tilted-feed table (4) should be made adjustable to move in vertical and rolling directions. The tilted feed table (4) can be adjusted in vertical and rolling directions by changing the shim heights (13) in
Another advantage of the fixture is that it can be easily installed within 15 minutes on a conventional rolling mill. The installation requires no major modification to the rolling mill. The rolling mill can be used for conventional rolling, and then changed over to tilt-rolling without major disruption of production.
The invention is further illustrated by the following examples, which are not meant to be limiting.
In the two examples presented below, a tantalum workpiece made by powder metallurgy was used as the starting workpiece material for rolling. The texture of a workpiece produced by powder metallurgy is known to be close to random. The effects of tilt-rolling can be clearly observed if a workpiece with random texture is used as the starting material so that the effect of the prior processing can be isolated.
Three plates, 7 to 8 mm thick were produced from powder made in accordance with U.S. Pat. No. 6,521,173. The process given below (steps 1 to 6), results in a puck 165 mm diameter and 81 mm thick.
Specifically, the operations are:
1) Cold Isostatically Press (CIP) the powder to 60-90% density;
2) Encapsulate the pressed preform in a steel can and evacuate and seal the can;
3) Hot Isostatically Press (HIP) the preform to a billet with 100% density;
4) Remove the steel can;
5) Anneal the billet; and
6) Cut, using a band-saw or any similar suitable cutting equipment, into slices suitable for rolling into a plate: the slices have the shape of a hockey puck.
The pucks were rolled using conventional techniques (including an annealing step at 33 mm thickness), and finish-processed conventionally. In rolling, 15% reduction per pass and 90-deg rotations between passes were used. The workpiece was not turned over.
Samples were taken from the centre of the plate, the mid-radius of the plate and the edge of the plate (2 samples, well separated), and the texture determined by EBSD, using a 10 μm step in both horizontal and vertical directions. The average grain size was about ASTM 7 (28 microns ALI). Once the texture maps showing the texture from top to bottom surface of the sample were obtained, the texture maps were analyzed mathematically to quantify the through-thickness texture gradient as follows:
1) The maps are divided into two halves, the top half (H1) and bottom half (H2).
2) A mask, with a cut-out hole 90 μm high, but full-width (1.64 mm), is placed over the map, such that the top of the cut-out hole corresponds to the top of the map. Note that the height of the window is chosen to be approximately 3 grains, but an integral number of EBSD steps (in this case, 9 steps).
3) The percentage of the area of the cut-out hole occupied by the grains within 15-deg of <100>//ND, as is the percentage occupied by the grains within 15-deg of <111>//ND.
4) The mask is moved down by 10 μm, and the calculations repeated.
5) Operation 4 is repeated until the bottom of the cut-out hole corresponds to the bottom of the map.
6) This data is analyzed to determine, for each half of the thickness:
The results of this analysis, for both half-thicknesses of the three specimens, are:
A plate 7.5 mm thick was made, using the same powder-metallurgy process as was described above, (steps 1 to 6), resulting in a puck 165 mm diameter and 42 mm thick.
It was then rolled to thickness. A 5-degree tilt angle was used. The thickness of the piece was reduced by approximately 5-10% in each pass. The piece was rotated 45 degrees about a vertical axis after each pass. The piece was turned over after every 4 passes. The final thickness of the piece after rolling was 7.5 mm. The finish-processing (annealing etc.) was performed conventionally.
Samples were taken from the centre of the plate, the mid-radius of the plate and the edge of the plate, and the texture determined by EBSD, using a 15 μm step in both horizontal and vertical directions. The average grain size was about ASTM 6½ (32 microns ALI). The results are calculated in the same way as for Example 1.
Although the number of data points is limited, a statistical comparison of the prior art and the inventive method may be useful. In Table 4, the variation of the texture gradient for example 1 (comparative) and example 2 (inventive) are compared. The absolute value of the texture gradient values listed in Table 2 and 3, were used to obtain the min-max range, mean and standard deviation of texture gradient for plates 1, 2, 3 in Example 1, and the plate in Example 2. Table 4 shows that the method described in this invention reduced the texture gradient for both 100 and 111 components significantly.
Whereas particular embodiments of this invention have been described above for purposes of illustration, it will be evident to those skilled in the art that numerous variations of the details of the present invention may be made without departing from the invention as defined in the appended claims.
