Laser-Produced Porous Surface

Abstract

The present invention disclosed a method of producing a three-dimensional porous tissue in-growth structure. The method includes the steps of depositing a first layer of metal powder and scanning the first layer of metal powder with a laser beam to form a portion of a plurality of predetermined unit cells. Depositing at least one additional layer of metal powder onto a previous layer and repeating the step of scanning a laser beam for at least one of the additional layers in order to continuing forming the predetermined unit cells. The method further includes continuing the depositing and scanning steps to form a medical implant.

Claims

1. (canceled)

2: An orthopedic implant comprising: porous metallic first and second structures, at least a portion of either one or both of the first and second structures being defined by polygonal porous cells; and an intermediate structure attached to and located between the first and the second structures, the intermediate structure having a different porosity than the first and the second structures.

3: The orthopedic implant of claim 2, wherein the first structure includes pores having a pore size that exceeds 80 μm in diameter.

4: The orthopedic implant of claim 3, wherein the pores of the first structure have a maximum pore size that is less than 800 μm in diameter.

5: The orthopedic implant of claim 3, wherein the second structure includes pores having a pore size that exceeds 80 μm in diameter, and wherein the pores of the first and the second structures have a maximum pore size that is less than 400 μm in diameter.

6: The orthopedic implant of claim 2, wherein the cells of the first structure have an irregular shape.

7: The orthopedic implant of claim 2, wherein the intermediate structure is directly attached to the first and the second structures.

8: The orthopedic implant of claim 2, wherein the intermediate structure is substantially solid.

9: The orthopedic implant of claim 2, further comprising an opening passing through a thickness of the implant, the opening having a diameter substantially larger than a diameter of pores of the first and the second structures.

10: The orthopedic implant of claim 2, wherein the first structure includes a plurality of pores and surrounds a first hole larger than each of the plurality of pores of the first structure, and wherein the first hole extends through a thickness of the first structure.

11: The orthopedic implant of claim 10, wherein the second structure includes a plurality of pores and surrounds a second hole larger than each of the plurality of pores of the second structure, and wherein the second hole extends through a thickness of the first structure.

12: The orthopedic implant of claim 2, wherein the first structure includes a plurality of pores that correspond to a computer-aided design model.

13: The orthopedic implant of claim 2, wherein the first structure has a graded porosity.

14: The orthopedic implant of claim 2, wherein the shape of at least some of the cells vary from the shape of other ones of the cells.

15: The orthopedic implant of claim 2, where the cells are tessellated unit cells.

16: The orthopedic implant of claim 15, wherein the unit cells are in the shape of a tetrahedron, dodecahedron or octahedron.

17: The orthopedic implant of claim 2, wherein the first structure, the second structure, and the intermediate structure are integrated.

18: An orthopedic implant comprising: porous metallic first and second structures, at least a portion of either one or both of the first and second structures being defined by irregular porous cells; and an intermediate structure attached to and located between the first and the second structures, the intermediate structure having a different porosity than the first and the second structures.

19: The orthopedic implant of claim 18, wherein the first structure includes pores having a pore size that exceeds 80 μm in diameter.

20: The orthopedic implant of claim 19, wherein the pores of the first structure have a maximum pore size that is less than 800 μm in diameter.

21: The orthopedic implant of claim 20, wherein the second structure includes pores having a pore size that exceeds 80 μm in diameter, and wherein the pores of the first and the second structures have a maximum pore size that is less than 400 μm in diameter.

22: An orthopedic implant comprising: porous metallic first and second structures, at least a portion of either one or both of the first and second structures being defined by polygonal porous cells having an irregular shape; and a substantially solid intermediate structure directly attached to and located between the first and the second structures.

23: The orthopedic implant of claim 22, wherein the first structure includes pores having a diameter between 80 μm and 800 μm.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Methods of forming the porous surface of the present invention can be performed in many ways and some embodiments will now be described by way of example and with reference to the accompanying drawings in which:

[0038] FIG. 1 is a diagrammatic illustration of the apparatus used to make test samples according to the processes of the present invention;

[0039] FIG. 2 is a photographic image showing an array of samples produced by the processes as performed by the apparatus of FIG. 1;

[0040] FIG. 3 is a table showing a series of parameters used for the samples of FIG. 2;

[0041] FIGS. 4A to 10F are scanning electron microscope images of the surface structure of various samples made by the method according to the invention;

[0042] FIG. 11 is a scanning electron microscope micrograph taken from a porous Ti sintered structure;

[0043] FIG. 12 is an optical image of a section through a sample showing the microstructure;

[0044] FIG. 13 is an image detailing surface structures;

[0045] FIGS. 14A-15B are non-contact surface profilimetry images detailing plan views of the samples; and

[0046] FIGS. 16A-25 are scanning electron microscope micrographs produced prior to multi-layer builds shown in FIGS. 7A-8E.

[0047] FIG. 26 indicates the metallography and spectra of a typical bond coat structure.

[0048] FIG. 27 shows the effect of line spacing on pore size.

[0049] FIG. 28A-28F are examples of typical waffle structures.

[0050] FIG. 29 is a trabecular bone-type structure constructed from a micro CT scan.

[0051] FIG. 30 shows typical freestanding structures.

[0052] FIG. 31 shows a freestanding structure built using the preferred scanning strategy.

[0053] FIG. 32A illustrates one embodiment of a unit cell of the present invention;

[0054] FIG. 32B illustrates an alternate embodiment of a unit cell of the present invention.

[0055] FIG. 32C illustrates a lattice structure formed using a plurality of unit cells illustrated in FIG. 32BB;

[0056] FIG. 33 illustrates lattice structures with and without laser beam compensation formed using the unit cells illustrated in FIG. 32B;

[0057] FIG. 34A illustrates an alternate embodiment of a unit cell of the present invention;

[0058] FIG. 34B illustrates a lattice structure formed using a plurality of unit cells illustrated in FIG. 34A;

[0059] FIG. 35 illustrates lattice structures formed with and without laser beam compensation;

[0060] FIG. 36A illustrates an alternate embodiment of a unit cell of the present invention;

[0061] FIG. 36B illustrates a lattice structure formed using a plurality of the unit cells illustrated in FIG. 36A;

[0062] FIGS. 37A and 37B illustrate actual lattice structures formed using a plurality of unit cells represented in FIG. 36A;

[0063] FIG. 38A illustrates an additional embodiment of a unit cell of the present invention;

[0064] FIG. 38B illustrates a lattice structure created using a plurality of unit cells illustrated in FIG. 38A;

[0065] FIG. 39A illustrates lattice structures created using unit cells illustrated in FIG. 38A with varying exposure time;

[0066] FIG. 39B illustrates lattice structures created using unit cells illustrated in FIG. 32A with varying exposure time;

[0067] FIG. 39C illustrates a side view of an embodiment of FIG. 39A;

[0068] FIG. 39D illustrates a side view of a lattice structure illustrated in FIG. 39B;

[0069] FIG. 40 is a representation of a lattice structure created using a plurality of the unit cells illustrated in FIG. 38A with random perturbation;

[0070] FIG. 41 illustrates graduation of a solid to a lattice build;

[0071] FIG. 42 illustrates a graduation from one lattice density to another;

[0072] FIG. 43A illustrates a femoral hip component;

[0073] FIG. 43B illustrates an exploded view of FIG. 43A;

[0074] FIG. 44 illustrates the component of FIG. 43A with a reduced sized femoral attachment;

[0075] FIG. 45 illustrates a “jacket” created by the subtraction of the embodiment of FIG. 44 from the embodiment of FIG. 43A;

[0076] FIG. 46A illustrates one embodiment of a single unit cell for use in an open cellular lattice structure;

[0077] FIG. 46B illustrates an open cellular lattice structure;

[0078] FIG. 47 illustrates the embodiment illustrated in FIG. 46B merged with the embodiment illustrated in FIG. 44;

[0079] FIGS. 48A and 48B illustrate one embodiment of a finished product;

[0080] FIGS. 49A-49C illustrate an alternate embodiment of a finished product;

[0081] FIGS. 50A and 50B illustrate an alternate embodiment of a finished product;

[0082] FIGS. 51A-51C illustrate an alternate embodiment of a finished product;

[0083] FIGS. 52A and 52B illustrate an alternate embodiment of a finished product;

[0084] FIG. 53 illustrates an alternate embodiment of a finished product;

[0085] FIG. 54 illustrates an alternate embodiment of a finished product;

[0086] FIGS. 55A and 55B illustrate an apparatus used in conjunction with the present invention;

[0087] FIG. 56 illustrates a zoomed-in view of the embodiment illustrated FIG. 55B;

[0088] FIG. 57 illustrates a zoomed-in view of the apparatus illustrated in FIG. 55B, further along in the process;

[0089] FIG. 58 illustrates a zoomed-in view of the apparatus illustrated in FIG. 55B, further along in the process;

[0090] FIGS. 59A and 59B illustrate porous surface coatings being applied to a substrate;

[0091] FIGS. 60A and 60B illustrate one embodiment of a representation of a finished product;

[0092] FIGS. 61A and 61B illustrate one embodiment of a finished product created using the present invention;

[0093] FIGS. 62A to 62D illustrate one embodiment of a finished product created using the present invention;

[0094] FIG. 63 illustrates a titanium lattice structure with hierarchical surface coating of sintered titanium satellites; and

[0095] FIGS. 64-71 illustrate the change occurring to the embodiment illustrated in FIG. 63, while the lattice is exposed to a laser at increasing time periods.

DETAILED DESCRIPTION OF THE INVENTION

[0096] The present invention relates to a method of forming porous and partially porous metallic structures which are particularly but not exclusively applicable for use in hard or soft tissue interlock structures for medical implants and prosthesis. The method makes use of laser technology by employing a variety of scanning strategies. Typical metal and metal alloys employed include stainless steel, cobalt chromium alloys, titanium and its alloys, tantalum and niobium, all of which have been used in medical device applications. The present invention can be used for such medical device applications where bone and soft tissue interlock with a component is required, or where a controlled structure is required to more closely match the mechanical properties of the device with surrounding tissue. Additionally, the present invention may be employed to enhance the biocompatibility of a porous structure with animal tissue. With these advantages in mind, a structure may be created using specific dimensions required to accommodate a particular patient.

[0097] One particular intention of the present invention is to produce a three-dimensional structure using a laser remelting process, for example, for building structures with or without a solid base or core. When applied to an orthopedic prosthesis, the three-dimensional structure could be used to provide a porous outer layer to form a bone in-growth structure. Alternatively, the porous structure, when applied to a core, could be used to form a prosthesis with a defined stiffness to both fulfill the requirement of a modulus match with surrounding tissue and provide interconnected porosity for tissue interlock. A further use could be to form an all-porous structure with grade pore size to interact with more than one type of tissue. Again, the process can be used to build on a solid base or core with an outer porous surface, the porosity of which is constant or which varies. The base or core materials to which the process is applied may be either titanium and its alloys, stainless steel, cobalt chrome alloys, tantalum or niobium as well as any other suitable material. The preferred surface coatings are titanium, cobalt chrome and tantalum but both stainless steel and niobium can also be used as well as any other suitable material. Fully porous structures may be built from any of the materials tested, with the preferred material being titanium. One intention of the present invention is to produce a method which can be exploited on a commercial basis for the production of, for example, bone interlock surfaces on a device although it has many other uses.

[0098] According to the present invention, a method of forming a three-dimensional structure includes building the shape by laser melting powdered titanium and titanium alloys, stainless steel, cobalt chrome alloys, tantalum or niobium using a continuous or pulsed laser beam. Individual layers of metal are scanned using a laser. The laser may be a continuous wave or pulsed laser beam. Each layer or portion of a layer is scanned to create a portion of a plurality of predetermined unit cells, as will be described below. Successive layers are deposited onto previous layers and also may be scanned. The scanning and depositing of successive layers continues the building process of the predetermined unit cells. As disclosed herein, by continuing the building process refers not only to a continuation of a unit cell from a previous layer but also a beginning of a new unit cell as well as the completion of a unit cell.

[0099] The method can be performed so that the structure is porous and if desired, the pores can be interconnecting to provide an interconnected porosity.

[0100] If desired, the method can include using a base or core of cobalt chrome alloy, titanium or alloy, stainless steel, niobium and tantalum, on which to build a porous layer of any one of the aforementioned metals and alloys by laser melting using a continuous or pulsed laser beam. Thus, a mixture of desired mixed materials may be employed.

[0101] The method can be applied to an existing article made from cobalt chrome, titanium or titanium alloys, stainless steel, tantalum or niobium, such as an orthopedic implant, to produce a porous outer layer from any of the aforementioned metals or alloys to provide a bone in-growth structure.

[0102] Preferably, prior to the deposition of any powder onto a substrate, a cleaning operation to ensure a contaminant-free surface may be employed. Typically, this process may include a solvent wash followed by a cleaning scan of the laser beam without the presence of particles.

[0103] In order to increase the mechanical bond between a substrate i.e., core or base, and a first layer of deposited powder a coating process may be employed. The coating process includes applying a third metal directly to the substrate, which has a higher bond strength to the substrate then does the first layer of powder. This process is particularly useful when the substrate and first powder layer are of different materials. The process of coating the substrate may be accomplished using known processes including laser deposition, plasma coating, cold gas dynamic spraying or similar techniques. One example of the coating process includes using either niobium or tantalum as an interface between a cobalt chrome alloy substrate and a first layer of titanium powder.

[0104] The present invention can include a laser melting process, which precludes the requirement for subsequent heat treatment of the structure, thereby preserving the initial mechanical properties of the core or base metal. The equipment used for the manufacture of such a device could be one of many currently available including the MCP Realiszer, the EOS M270, Trumpf Trumaform 250, the Arcam EBM S12 and the like. The laser may also be a custom produced laboratory device.

[0105] The method may be applied to produce an all-porous structure using any of the aforementioned metal or metal alloys. Such structures can be used as finished or final products, further processed to form a useful device for bone or soft tissue in-growth, or used to serve some other function such as that of a lattice to carry cells, for example.

[0106] The pore density, pore size and pore size distribution can be controlled from one location on the structure to another. It is important to note that successive powder layers can differ in porosity by varying factors used for laser scanning powder layers. As for example, a first layer of powder is placed and subsequently scanned. Next a second layer of powder is placed and scanned. In order to control porosity the second scan may be angled relative to the first scan. Additionally, the angling of the scanning as compared to previous and post scans may be maneuvered and changed many times during the process of building a porous structure. If a structure was built without alternating the angling of any subsequent scans you would produce a structure having a plurality of walls rather than one with an interconnecting or non-interconnecting porosity.

[0107] In one such method, the laser melting process includes scanning the laser beam onto the powder in parallel scan lines with a beam overlap i.e., scan spacing, followed by similar additional scans or subsequent scans at 90°. The type of scan chosen may depend on the initial layer thickness as well as the web height required. Web height refers to the height of a single stage of the porous structure. The web height may be increased by deposited additional layers of powder of a structure and scanning the laser at the same angle of the previous scan.

[0108] Further, the additional scan lines may be at any angle to the first scan, to form a structure with the formation of a defined porosity, which may be regular or random. The scan device may be programmed to proceed in a random generated manner to produce an irregular porous construct but with a defined level of porosity. Furthermore, the scan can be pre-programmed using digitized images of various structures, such as a trabecular bone, to produce a similar structure. Contrastingly, the scan may be pre-programmed using the inverse of digitized images, such as the inverse of a digitized trabecular bone to produce trabecular shaped voids. Many other scanning strategies are possible, such as a waffle scan, all of which can have interconnecting porosity if required.

[0109] The beam overlap or layer overlap may be achieved by rotation of the laser beam, the part being produced, or a combination of both.

[0110] A first method according to the present invention is intended to produce a porous structure for bone in-growth on the outer surface layer of a prosthesis. To produce a porous surface structure, the nature of the material formed as a result of laser melting of powdered beads is principally dependent on the thermal profile involved (heating rate, soaking time, cooling rate); the condition of the raw material (size and size distribution of powder particles); atmospheric conditions (reducing, inert or oxidizing chamber gas) In some instances, the nature of the material formed may be further a result of accurate control of the deposited layer thickness.

[0111] There have been a number of studies to determine the optimum pore structure for maximization of bone in-growth on prostheses. The general findings suggest that optimum porosity is between approximately 20% and 40%, and aim to mid value with a mean volume percent of voids of about 70%. The preferred pore structure is interconnected, with a minimum pore size between about 80 μm and 100 μm and a maximum pore size between 80 μm and 800 μm. The structured thickness for in-growth is 1.4-1.6 mm, but can be larger or smaller depending on the application. As for example, it may be necessary to produce even smaller pore sizes for other types of tissue interaction or specific cellular interaction.

[0112] The first phase of development of the present invention involved an investigation, designed to characterize the material transformation process and to identify the optimum parameters for processing using three substrate materials CoCr and Ti stainless steel alloys, with five powder types Ti, CoCr, Ta and Nb, stainless steel.

[0113] The initial Direct Laser Remelting trials explored a comprehensive range of process parameters during the production of a number of coated base substrates. Results from this task were evaluated using laser scanning and white light interferometry in order to define the range of process parameters that produced the optimum pore structure.

[0114] Referring to FIG. 1, there is shown the apparatus used to carry out the method which comprises an Nd; YAG industrial laser 1 manufactured by Rofin Sinar Lasers, in Hamburg, Germany, integrated to an RSG1014 analogue galvo-scanning head 2 providing a maximum scan speed of 500 mm/s. The laser beam 3 is directed into an atmospherically controlled chamber 4, which consists of two computer-controlled platforms for powder delivery and part building. The powder is delivered from a variable capacity chamber 5 into the chamber 4 and is transported by a roller 6 to a build platform 7 above a variable capacity build chamber 8. In the embodiment shown in FIG. 1, the build and delivery system parameters are optimized for an even 100 μm coating of powder to be deposited for every build layer. The metals chosen as surface materials are all difficult to process due to their affinity for oxygen. Cr and Ti are easily oxidized when processed by laser in oxygen-containing atmosphere, their oxide products have high melting points and poor flowability. For this reason, and to prevent the formation of other undesirable phases, the methods were carried out under an Argon inert atmosphere in chamber 8. Pressure remained at or below atmospheric pressure during the entire application.

[0115] The build chamber 8 illustrated in FIG. 1 and method of layering described above is suitable for test specimens and constructs such as three-dimensional freestanding structures. However, in order to build on to an existing device, such as acetabular metal shells, hip and knee femoral components, knee tibial components and other such devices, considerable changes to the powder laying technique would need to be applied.

[0116] Preliminary experiments were performed on CoCr alloy to determine the efficacy of in-situ laser cleaning of the substrate. Typical processing conditions were: Laser power of 82 W, pulse frequency of 30 KHz, and a laser scan speed of 160 mm/sec.

[0117] Preliminary experiments were performed on CoCr to assess the environment conditions within the chamber. In these examples, Co212-e Cobalt Chrome alloy was used. The CoCr was configured into square structures, called coupons. Arrays of CoCr coupons were built onto a stainless steel substrate. The Co212-e Cobalt Chrome alloy had a particle size distribution of 90<22 urn, i.e., 90% of the particles are less than 22 μm, the composition of which is shown in the table below.

TABLE-US-00001 TABLE 1 Composition of Co212-e CoCr alloy Element Cr Mo Si Fe Mn Ni N C Co Wt % 27.1 5.9 0.84 0.55 0.21 0.20 0.16 0.050 Balance

[0118] An array of nine sample coupons were produced as shown in FIG. 2, with the process of Table 2, using a maximum laser power of 78 watts (W) and laser scanning speed for each coupon varying between 100-260 mms.sup.−1. Of course a higher laser power may be employed; however, a higher laser power would also necessitate increasing the speed of the laser scan speed in order to produce the desired melting of the powder layer. A simple linear x-direction scan was used on each of the coupons. This allowed the processing parameter, beam overlap, to be used to control the space between successive scan lines. That is, with a 100 μm laser spot size, an overlap of −200% produces a 100 μm gap between scans. Although the acceptable range for the beam overlap is given at +50% to −1200% it should be duly noted that the negative number only refers to the fact the there is a gap as opposed to a beam overlap between successive scans. For instance a beam overlap of zero refers to the fact that successive scans on the same layer of powder border each other. If the beam overlap was 5% then 5% of the first scan is overlapped by the second scan. When computing the Andrew number the absolute value of the beam overlap is used. The complete set of process parameters used is shown in Table 2 below.

TABLE-US-00002 TABLE 2 Process parameters Layer Beam Scanning Power Watts Thickness Diameter Speed No. of Overlap (W) (μm) (μm) (mms.sup.−1) Atmosphere Layers (% of line width) 78 100 100 100-260 No 16 25, 50, −500

[0119] The incremental changes in scanning speed and the size of the speed range were modified as the experiments progressed. To begin with, a large range of speeds was used to provide an initial indication of the material's performance and the propensity to melt. As the experiments progressed, the range was reduced to more closely define the process window. Speed and beam overlap variations were used to modify the specific energy density being applied to the powder bed and change the characteristics of the final structure. The complete series of parameters are given in FIG. 3, the parameters sets used for the definitive samples are shaded in gray.

[0120] CoCr was the first of four powders to be examined and, therefore, a wide range of process parameters was used. In each case, laser power and the pulse repetition rate were kept constant, i.e., continuous laser pulse, to allow the two remaining parameters to be compared. Layer thickness was maintained at 100 μm throughout all the experiments described here. Layer thickness can, however, vary between 5 μm to 2000 μm.

[0121] On completion of the initial series of experiments using CoCr powder on 2.5 mm thick stainless steel substrates, basic optical analysis was conducted of the surface of the coupons to ascertain the size of the pores and degree of porosity being obtained. Once a desired pore size was obtained and the coupons had suitable cohesion, the two experiments closest to the optimum desired pore size were repeated using first CoCr and then Ti substrates. The same structure could be obtained by other parameters.

[0122] Following the conclusion of the CoCr experiments, the remaining three powders; Niobium, Tantalum and Titanium were investigated in turn. The procedure followed a simple course although fewer parameter sets were explored as the higher melting points of these materials forced the reduction in speeds compared to CoCr powder.

[0123] For Niobium, the particle size description was 80%<75 μm at a purity of 99.85%. Due to its higher melting temperature compared to that of CoCr (Nb being at about 2468° C., and CoCr being at about 1383° C.), the laser parameters used included a reduced scanning speed range and increased beam overlap providing increased specific energy density at the powder bed. In addition, the pulse repetition rate was varied from 20 kHz to 50 kHz.

[0124] On completion of a small number (four in total) of preliminary experiments of Nb on stainless steel substrate, the experiment with the most ideal parameters was repeated on both the CoCr and Ti substrates.

[0125] The Tantalum used in this study had a particular size distribution of 80%<75 μm with a purity of 99.85%. Ta has a melting point of about 2996° C. and was processed using the same laser parameters as Nb. Now confident of the atmospheric inertness, the Ta powder was melted directly onto the CoCr and Ti substrates.

[0126] Bio-medical alloys of Titanium were not readily available in powder form and so pure Ti was chosen. The particle size distribution for the Ti powder was 80%<45 μm with a purity of 99.58%. The same parameters used for Nb and Ta were also used for the Ti powder. Ti has a lower melting point than Ta or Nb, Ti being at about 1660° C., but has a higher thermal conductivity than Ta or Nb. This implies that although the powder should require less energy before melting, the improved heat transfer means a larger portion of the energy is conducted away from the melt pool.

[0127] Following the completion of samples with all four powders on the required substrates, surface analysis was conducted using optical analysis and a scanning electron microscope to obtain images of the resultant pores. Porosity was calculated using a simple image processing technique involving the setting of contrast thresholds and pixel counting. While this technique is not the most accurate method, it allows the rapid analysis of small samples produced. Techniques such as Xylene impregnation would yield more accurate results but they are time consuming and require larger samples than those produced here.

[0128] Following an extended series of experimentation, two sets of laser processing parameters were selected for the laser melting of CoCr powder. From analysis of the stainless steel substrates, it was seen that a large portion of the results fell within the pore size required for these materials, stated as being in the range of 80 μm to 400 μm.

[0129] Optical analysis of the surface structure of each of the coupons produced with CoCr on CoCr and Ti alloy substrates were initially viewed but due to problems with the depth of field associated with an optical microscope, little information could be extracted. In addition to the coupons that were produced to satisfy the project requirements, two experiments were conducted using a relatively large negative beam overlap of −250 and −500%. Optical images of the coupon's surface and in section are shown in FIG. 4. These were not the definitive parameters chosen for the final arrays on CoCr and Ti alloy substrates as the pore size exceeds the required 80 μm to 400 μm. They are shown here to display what the Direct Laser Remelting process can produce when an excessive beam overlap is used.

[0130] To provide a clearer indication of the pore size and pore density, the optical analysis was repeated using images obtained from the scanning electron microscope. FIG. 5 is an image of two coupons produced from a CoCr array on Ti alloy substrates. This array was chosen because it best satisfied the requirements of this exercise. The parameters were: laser power of 82 W continuous wave (cw); 25% beam overlap; scanning speed varied from 100 mms-.sup.1 to 260 mms-.sup.1 in 20 mm-.sup.1 increments; the images of the coupons shown here, taken from this array, were produced with scanning speeds of 180 mms.sup.−1 to 200 mms.sup.−1. The surface is comprised of fused pathways that develop a network of interconnected pores. This structure continues throughout the layer until the interface is reached. The interface is characterized by a patchwork of fusion bonds. These bond sites are responsible for securing the interconnected surface structure to the baseplate. The macroscopic structure is covered with unmelted powder particles that appear to be loosely attached. In addition, there are larger resolidified globules that may have limited bonding to the surface.

[0131] FIGS. 6 and 7 are the scanning electron microscope images produced from the Nb and Ta coupons on Ti alloy substrates. Specifically, FIGS. 6A to 6E are scanning election microscope images of the surface structure of Nb on Ti alloy substrates, produced with a laser power of 82 W cw, −40% beam overlap. The scanning speeds used were: 160 mms.sup.−1 for FIG. 6A, 190 mms.sup.−1 for FIG. 6A, 200 mms.sup.−1 for FIG. 6C, 210 mms.sup.−1 for FIG. 6D and 240 mms.sup.−1 for FIG. 6E, respectively.

[0132] FIGS. 7A to 7C are scanning election microscope images of the surface structure of Ta on Ti alloy substrates produced using the same parameters used in the Nb or Ti alloy substrates except: FIG. 7A was produced with a scanning speed of 160 mms.sup.−1; FIG. 7B's speed was 200 mms.sup.−1 and FIG. 7C's speed was 240 mms.sup.−1, respectively. An increased beam overlap was used here as Nb and Ta have high melting points, which require a greater energy density. The surfaces once again exhibit significant levels of unmelted powder particles and loosely attached resolidified beads that vary in size from a few microns to several hundred microns. All samples were loosely brushed after completion and cleaned in an ultrasonic aqueous bath. It is possible that further cleaning measures may reduce the fraction of loose particles.

[0133] FIGS. 8A to 8E are scanning electron microscope images taken from the Ti coupons on the CoCr alloy substrates. The laser processing parameters used were the same as those for the Nb and Ta powders, with once again only the speed varying. The scanning speed was varied from 160 mms.sup.−1 (FIG. 8A, 170 mms.sup.−1 (FIG. 8B), 200 mms.sup.−1 (FIG. 8C); 230 mms.sup.−1 (FIG. 8D) to 240 mms.sup.−1 (FIG. 8E). The Ti coupon on CoCr samples, (FIGS. 8A to 8C) indicate very high density levels compared to the other examples. The line-scans can be clearly seen, with good fusion between individual tracks, almost creating a complete surface layer. The surface begins to break up as the scanning speed is increased.

[0134] FIGS. 9A to 9E are scanning electron microscope images of surface structures of Ti on Ti alloy substrates produced with the same parameters used in FIGS. 8A to 8E, respectively. It is unclear why Ti should wet so well on CoCr substrates. In comparison, Ti on Ti exhibits similar characteristic patterns as with Nb, Ta, and CoCr, specifically, an intricate network of interconnected pores.

[0135] Following the completion of the multi-layer coupons, a series of 20 mm×20 mm structures were produced from Ti that utilized an X and Y-direction “waffle” scanning format using the optimum Ti operating parameters with the two scans being orthogonal to one another. The intention behind these experiments was to demonstrate the ability of the Direct Laser Remelting process to produce parts with a controlled porosity, e.g. internal channels of dimensions equal to the required pore size, e.g. 80 μm to 400 μm. To do this, a relatively large beam overlap of between −400% and −600% was used. Scanning electron microscope images of the surfaces of these structures are shown in FIGS. 10A to 10F. The scanning speed, 160 mms.sup.−1 and the laser power 72 W cw were kept constant while the beam overlaps; −400% in FIGS. 10A and 10B; −500% in FIGS. 10C and 10D and −600% in FIGS. 10E and 10F, were varied accordingly. Scanning electron microscope micrographs, taken from a porous Ti sintered structure provided by Stryker-Howmedica are shown for reference in FIG. 11.

[0136] To illustrate more clearly the interaction between the substrate/structure metallurgical interaction, the Ti on Ti substrate was sectioned, hot mounted and polished using a process of 1200 and 2500 grade SiC, 6 μm diamond paste and 70/30 mixture of OPS and 30% H.sub.2O.sub.2. The polished sample was then etched with 100 ml H.sub.2O, 5 ml NH.FHF and 2 cm.sup.3 HCl for 30 seconds to bring out the microstructure. Optical images of this sample in section are shown in FIG. 12.

[0137] FIG. 13 is an image taken from a non-contact surface profilimentry to show the surface structures obtained when using Ti, CoCr, Ta and Nb on Ti substrates. Values for Ra, Rq and Rb roughness are also shown.

[0138] From the optical and scanning election microscope analysis conducted, it is apparent that the Direct Laser Remelting process is capable of satisfying the requirements for pore characteristics, concerning maximum and minimum pore size, interconnectivity and pore density. From the initial visual analysis of the CoCr coupons, it was apparent from these and other examples, that subtle variations in pore structure and coverage could be controlled by scanning velocity and line spacing.

[0139] The key laser parameters varied for forming the three-dimensional metallic porous structures are: (a) Laser scanning speed (v.) in (mms.sup.−1), which controls the rate at which the laser traverses the powder bed; (b) Laser power, P(W), which in conjunction with the laser spot size controls the intensity of the laser beam. The spot size was kept constant throughout the experiment; (c) Frequency, (Hz) or pulse repetition rate. This variable controls the number of laser pulses per second. A lower frequency delivers a higher peak power and vice versa.

[0140] The line width can be related to the laser scanning speed and the laser power to provide a measure of specific density, known as the “Andrew Number”, where:

[00001]$\mathrm{An}=\frac{P}{b\times v}\left(J\text{/}{\mathrm{mm}}^{-2}\right)$

Where P denotes the power of the laser, v is the laser scanning speed and b denotes beam width of the laser. The Andrew number is the basis for the calculation of the present invention. The Andrew number may also be calculated by substituting the line separation (d) for beam width (b). The two methods of calculating the Andrew number will result in different values being obtained. When using line separation (d) as a factor only on track of fused powder is considered, whereas when using the beam width (b) as a factor, two tracks of fused powder are considered as well as the relative influence of one track to the next. For this reason we have chosen to concern ourselves with the Andrew number using scan spacing as a calculating factor. It can thus be appreciated, that the closer these tracks are together the greater the influence they have on one another.

[0141] Additionally, the laser power may be varied between 5 W and 1000 W. Utilizing lower power may be necessary for small and intricate parts but would be economically inefficient for such coatings and structures described herein. It should be noted that the upper limit of laser power is restricted because of the availability of current laser technology. However, if a laser was produced having a power in excess of 1000 W, the scanning speed of the laser could be increased in order that an acceptable Andrew number is achieved. A spot size having a range between 5 um (fix) to 500 um (fix) is also possible. For the spot size to increase while still maintaining an acceptable Andrew number, either the laser power must be increased or the scanning speed decreased.

[0142] The above formula gives an indication of how the physical parameters can vary the quantity of energy absorbed by the powder bed. That is, if the melted powder has limited cohesion, e.g. insufficient melting, the parameters can be varied to concentrate the energy supply to the powder. High Andrew numbers result in reduced pore coverage and an increase in pore size due to the effects of increased melt volume and flow. Low Andrew numbers result in low melt volume, high pore density and small pores. Current satisfactory Andrew numbers are approximately 0.3 J/mm.sup.−2 to 8 J/mm.sup.−2 and are applicable to many alternative laser sources. It is possible to use a higher powered laser with increased scanning speed and obtain an Andrew number within the working range stated above.

[0143] Line spacing or beam overlap can also be varied to allow for a gap between successive scan lines. It is, therefore, possible to heat selected areas. This gap would allow for a smaller or larger pore size to result. The best illustration of this is shown in FIGS. 4A to 4C where a −500% beam overlap has been applied. FIGS. 4A to 4C are scanning election microscope images of the surface structure of CoCr on stainless steel produced with a laser power of 82 W cw. FIG. 4A was produced with a laser scanning speed of 105 mms.sup.−1 and FIG. 4B was produced with a laser scanning speed of 135 mms.sup.−1. FIG. 4C is an image of the same structure in FIG. 4B, in section. There is a significant self-ordering within the overall structure. Larger columnar structures are selectively built leaving large regions of unmelted powder. It is worth noting that these pillars are around 300 μm wide, over 1.6 mm tall and fuse well with the substrate, as seen in FIG. 4C. Further analysis shows that the use of a hatched scanning format allows porosity to be more sufficiently controlled to allow the pore size to be directly controlled by the beam overlap.

[0144] The use of an optical inspection method to determine this approximate porosity is appropriate given the sample size. This method, although not accurate due to the filter selection process, can, if used carefully, provide an indication of porosity. An average porosity level of around 25% was predicted. This porosity level falls within the range of the desired porosity for bone in-growth structures. The mechanical characteristics of the porous structures are determined by the extent of porosity and the interconnecting webs. A balance of these variables is necessary to achieve the mechanical properties required by the intended application.

[0145] Increased fusion may, if required, be obtained by heating the substrate, powder or both prior to scanning. Such heating sources are commonly included in standard selective laser sintering/melting machines to permit this operation.

[0146] Following trials on the titanium build on the cobalt chromium substrate, it was determined that the interface strength was insufficient to serve the intended application. Trials were made by providing a bond coat of either tantalum or niobium on the cobalt chromium substrate prior to the deposition of the titanium layers to for the porous build. The typical protocol involved: [0147] (i) Initial cleaning scan with a scan speed between 60 to 300 mm/sec, laser power 82 watts, frequency of 30 KHz, and a 50% beam overlap. [0148] (ii) Niobium or tantalum deposition with three layers of 50 μm using a laser power of 82 watts, frequency 30 to 40 KHz, with a laser speed of between 160 to 300 mm/sec. The beam overlap was low at 50% to give good coverage. [0149] (iii) A build of porous titanium was constructed using a laser power of 82 watts, frequency between 0 (cw) and 40 KHz, scanning speed of between 160 and 240 mm/sec, and beam overlap of −700%.
The strengths of the constructs are indicated in Table 3 with a comparison of the values obtained without the base coat.

TABLE-US-00003 TABLE 3 TENSILE MAXIMUM STRENGTH SPECIMEN LOAD (kN) (MPa) FAILURE MODE Ti on CoCr 2.5 5 Interface Ti on CoCr 3.1 6.2 Interface 1 (Nb on 13.0 26.18 65% adhesive, 35% bond Co-Cr) interface 4 (Ti on Nb on 7.76 15.62 Mostly bond coat interface Co-Cr) 5 (Ti on Nb on 9.24 18.53 20% adhesive, 40% bond Co-Cr) coat, 40% porous Ti 6 (Ti on Ta on 11.58 23.33 Mostly adhesive with Co-Cr) discrete webbing weakness 8 (Ta on 13.93 27.92 60% adhesive, 40% bond Co-Cr) interface 9 (Ti on Ta on 6.76 13.62 100% bond interface Co-Cr)
FIG. 26 shows the metallography of the structures with energy dispersive spectroscopy (EDS) revealing the relative metal positions within the build.

[0150] A typical waffle build of titanium on a titanium substrate was constructed as a way of regulating the porous structure. Scanning sequences of 0° 0° 0°, 90° 90° 90°, 45° 45° 45°, 135°, 135°, 135°, of layer thickness 0.1 mm were developed to form a waffle. Three layers of each were necessary to obtain sufficient web thickness in the “z” direction to give a structure of adequate strength. Typical parameters employed were: Laser power was 82 watts, operating frequency between 0 (cw) and 40 KHz, scan speed of between 160 and 240 mm/sec with a beam overlap of −700%. FIG. 27 gives an indication of the effect of line spacing and pore size. FIG. 28A shows typical examples of the waffle structure. The magnification level changes from 10, 20, 30, 70 and 150 times normal viewing as one moves respectively from Fig. B to F. FIG. 28A more specifically shows Ti powder on a Ti substrate with a controlled porosity by varying line spacing, i.e., beam overlap.

[0151] Trabecular structures of titanium on a titanium substrate were constructed as a way of randomising the porous structures. An STL (sterolithography) file representing trabecular structure was produced from a micro CT scan of trabecular bone. This file was sliced and the slice data sent digitally to the scanning control. This allowed the layer-by-layer building of a metallic facsimile to be realised. FIG. 29 shows a cross-sectional view of such a construct.

[0152] A method for making lattice-type constructs was referred to in the relevant art. A typical example of this type of structure is shown in FIG. 30. The scanning strategy employed to form such a construct was mentioned and such a strategy could be produced within the range of Andrew numbers outlined. Table 4 shows an indication of scanning strategies and their relationships to the Andrew number.

TABLE-US-00004 TABLE 4 Ti on Ta on CoCr Experimental Procedure. Initial Tantalum Coating RELATIVE BUILD SCAN PARAMETER LAYER PLATFORM LAYER STRATEGY SET THICKNESS POSITION ADDITIONAL Zero 0 Distance Between Roller & Build Platform 0 1.sup.st layer 50 μm  −50 μm thickness set using feeler gauges but powder not laid in preparation for cleaning scan with no powder. 1 50% Beam P = 82 W Initial Overlap Qs = 30 kHz Cleaning Scan v = 60 mm/s (no powder) A.sub.n = 27.333 J/mm.sup.2 Circular profile. P = 82 W Powder laid as 5 concentric Qs = 40 kHz usual circles, 0.1 mm V = 160 mm/s offset to negate A.sub.n = effects of ‘First 5.125 J/mm.sup.2 Pulse’ 50% Beam P = 82 W Scanned on Overlap Qs = 30 kHz same powder v = 300 mm/s layer as A.sub.n = previous 5.467 J/mm.sup.2 profile scan. 2 Circular profile. P = 82 W 50 μm −100 μm Powder laid as 5 concentric Qs = 40 kHz usual circles, 0.1 mm V = 160 mm/s offset to negate A.sub.n = effects of ‘First 5.125 J/mm.sup.2 Pulse’ 50% Beam P = 82 W Scanned on Overlap Qs = 30 kHz same powder v = 300 mm/s layer as A.sub.n = previous 5.467 J/mm.sup.2 profile scan. 3 Circular profile. P = 82 W 50 μm −150 μm Powder laid as 5 concentric Qs = 40 kHz usual circles, 0.1 mm V = 160 mm/s offset to negate A.sub.n = effects of ‘First 5.125 J/mm.sup.2 Pulse’ 50% Beam P = 82 W Scanned on Overlap Qs = 30 kHz same powder v = 300 mm/s layer as A.sub.n = previous 5.467 J/mm.sup.2 profile scan.

Final Titanium Coating

[0153]

TABLE-US-00005 RELATIVE BUILD SCAN PARAMETER LAYER PLATFORM LAYER STRATEGY SET THICKNESS POSITION ADDITIONAL 0 1.sup.st layer −150 μm thickness set using feeler gauges but powder not laid in preparation for cleaning scan with no powder. 1 50% Beam P = 82 W  50 μm −200 μm Cleaning Scan Overlap Qs = 30 kHz (No powder. v = 60 mm/s A.sub.n = 27.3 J/mm.sup.2 Circular profile. P = 82 W Powder spread 5 concentric Qs = 40 kHz but build circles, 0.1 mm V = 160 mm/s platform not offset to negate A.sub.n = lowered. effects of ‘First 5.125 J/mm.sup.2 Pulse’ 50% Beam P = 82 W Scanned on Overlap Qs = 30 kHz same powder v = 300 mm/s layer as A.sub.n = previous 5.467/mm.sup.2 profile scan. 2 Circular profile. P = 82 W 100 μm −300 μm Powder laid as 5 concentric Qs = 40 kHz usual circles, 0.1 mm V =160 mm/s offset to negate A.sub.n = effects of ‘First 5.125 J/mm.sup.2 Pulse’ 25% Beam P = 82 W Scanned on Overlap Qs = 30 kHz same powder v = 300 mm/s layer as A.sub.n = previous 3.644 J/mm.sup.2 profile scan. 3 Circular profile. P = 82 W 100 μm −400 μm Powder laid as 5 concentric Qs = 40 kHz usual circles, 0.1 mm V = 160 mm/s offset to negate A.sub.n = effects of ‘First 5.125 J/mm.sup.2 Pulse’ 0% Beam P = 82 W Scanned on Overlap Qs = 30 kHz same powder v = 300 mm/s layer as A.sub.n = previous 2.733 J/mm.sup.2 profile scan. 4 Waffle 0 and 90° P = 82 W  75 μm −475 μm Powder laid as 700 μm Qs = 0 Hz (cw) usual linespacing v = 240 mm/s (−600% Beam A.sub.n = overlap) 0.488 J/mm.sup.2 5 Waffle 0 and 90° P =82 W  75 μm −550 μm Powder laid as 700 μm Qs = 0 Hz (cw) usual linespacing v = 240 mm/s (−600% Beam A.sub.n = overlap) 0.488 J/mm.sup.2 6 Waffle 0 and 90° P = 82 W  75 μm −625 μm Powder laid as 700 μm Qs = 0 Hz (cw) usual linespacing v = 240 mm/s (−600% Beam A.sub.n = overlap) 0.488 J/mm.sup.2 7 Waffle 45 and P = 82 W  75 μm −700 μm Powder laid as 135° Qs = 0 Hz (cw) usual 700 μm v = 240 mm/s linespacing A.sub.n = (−600% Beam 0.488 J/mm.sup.2 overlap) 8 Waffle 45 and P = 82 W  75 μm −775 μm Powder laid as 135° Qs = 0 Hz (cw) usual 700 μm v = 240 mm/s linespacing A.sub.n = (−600% Beam 0.488 J/mm.sup.2 overlap) 9 Waffle 45 and P = 82 W  75 μm −850 μm Powder laid as 135° Qs = 0 Hz (cw) usual 700 μm v = 240 mm/s linespacing A.sub.n = (−600% Beam 0.488 J/mm.sup.2 overlap)

Ti on Ti Experimental Procedure.

[0154] Initial Titanium Coating

TABLE-US-00006 RELATIVE BUILD SCAN PARAMETER LAYER PLATFORM LAYER STRATEGY SET THICKNESS POSITION ADDITIONAL Zero 0 Distance Between Roller & Build Platform 0 1.sup.st layer 50 μm −50 μm thickness set using feeler gauges but powder not laid in preparation for cleaning scan with no powder. 1 50% Beam P = 82 W Initial Cleaning Overlap Qs = 30 kHz Scan (no v = 60 mm/s powder) A.sub.n = 27.333 J/mm.sup.2 Circular P = 82 W Powder laid as profile. Qs = 40 kHz usual 5 concentric v = 160 mm/s circles, 0.1 mm A.sub.n = offset to negate 5.125 J/mm.sup.2 effects of ‘First Pulse’ 50% Beam P = 82 W Scanned on Overlap Qs = 30 kHz same powder v = 300 mm/s layer as A.sub.n = previous profile 5.467 J/mm.sup.2 scan. 2 Circular P =82 W 50 μm −100 μm Powder laid as profile. Qs = 40 kHz usual 5 concentric V = 160 mm/s circles, 0.1 mm A.sub.n = offset to negate 5.125 J/mm.sup.2 effects of ‘First Pulse’ 50% Beam P = 82 W Scanned on Overlap Qs = 30 kHz same powder v = 300 mm/s layer as A.sub.n = previous profile 5.467 J/mm.sup.2 scan. 3 Circular P = 82 W 50 μm −150 μm Powder laid as profile. Qs = 40 kHz usual 5 concentric V = 160 mm/s circles, 0.1 mm A.sub.n = offset to negate 5.125 J/mm.sup.2 effects of ‘First Pulse’ 50% Beam P = 82 W Scanned on Overlap Qs = 30 kHz same powder v = 300 mm/s layer as A.sub.n = previous profile 5.467 J/mm.sup.2 scan.

[0155] Final Titanium Coating

TABLE-US-00007 RELATIVE BUILD SCAN PARAMETER LAYER PLATFORM LAYER STRATEGY SET THICKNESS POSITION ADDITIONAL 1 Circular profile. P = 82 W 100 μm −250 μm Powder laid as 5 concentric Qs = 40 kHz usual circles, 0.1 mm V = 160 mm/s offset to negate A.sub.n = effects of ‘First 5.125 J/mm.sup.2 Pulse’ 50% Beam P = 82 W Scanned on Overlap Qs = 30 kHz same powder v = 300 mm/s layer as A.sub.n = previous profile 5.467 J/mm.sup.2 scan 2 Circular profile. P = 82 W 100 μm −350 μm Powder laid as 5 concentric Qs = 40 kHz usual circles, 0.1 mm V = 160 mm/s offset to negate A.sub.n = effects of ‘First 5.125 J/mm.sup.2 Pulse’ 25% Beam P = 82 W Scanned on Overlap Qs = 30 kHz same powder v = 300 mm/s layer as A.sub.n = previous profile 3.644 J/mm.sup.2 scan. 3 Circular profile. P = 82 W 100 μm −450 μm Powder laid as 5 concentric Qs = 40 kHz usual circles, 0.1mm V = 160 mm/s offset to negate A.sub.n = effects of ‘First 5.125 J/mm.sup.2 Pulse’ 0% Beam P = 82 W Scanned on Overlap Qs = 30 kHz same powder v = 300 mm/s layer as A.sub.n = previous profile 2.733 J/mm.sup.2 scan. 4 Waffle 0 and 90° P = 82 W  75 μm −525 μm Powder laid as 700 μm Qs = 0 Hz (cw) usual linespacing v = 240 mm/s (-600% Beam A.sub.n = overlap) 0.488 J/mm.sup.2 5 Waffle 0 and 90° P = 82 W  75 μm −600 μm Powder laid as 700 μm Qs = 0 Hz (cw) usual linespacing v = 240 mm/s (−600% Beam A.sub.n = overlap) 0.488 J/mm.sup.2 6 Waffle 0 and 90° P = 82 W  75 μm −674 μm Powder laid as 700 μm Qs = 0 Hz (cw) usual linespacing v = 240 mm/s (600% Beam A.sub.n = overlap) 0.488 J/mm.sup.2 7 Waffle 45 and P = 82 W  75 μm −750 μm Powder laid as 135° Qs = 0 Hz (cw) usual 700 μm v = 240 mm/s linespacing A.sub.n = (−600% Beam 0.488 J/mm.sup.2 overlap) 8 Waffle 45 and P = 82 W  75 μm −825 μm Powder laid as 135° Qs = 0 Hz (cw) usual 700 μm v = 240 mm/s linespacing A.sub.n = (−600% Beam 0.488 J/mm.sup.2 overlap) 9 Waffle 45 and P = 82 W  75 μm −900 μm Powder laid as 135° Qs = 0 Hz (cw) usual 700 μm v = 240 mm/s linespacing A.sub.n = (−600% Beam 0.488 J/mm.sup.2 overlap)

[0156] The second and preferred approach uses a continuous scanning strategy whereby the pores are developed by the planar deposition of laser melted powder tracks superimposed over each other. This superimposition combined with the melt flow produces random and pseudorandom porous structures. The properties of the final structure, randomness, interconnectivity, mechanical strength and thermal response are controlled by the process parameters employed. One set of scanning parameters used was: Scanning sequences of 0° 0° 0°, 90° 90° 90°, 45° 45° 45°, 135°, 135°, 135°, of layer thickness 0.1 mm were developed to form a waffle. Three layers of each were necessary to obtain sufficient web thickness in the “z” direction. The array of sequences was repeated many times to give a construct of the desired height. Laser power was 82 watts, operating frequency between 0 (cw) and 40 KHz, scan speed of between 160 and 240 mm/sec with a beam overlap of −700%. FIG. 31 shows such a construct.

[0157] The use of an optical inspection method to determine this approximate porosity is appropriate given the sample size. This method, although not accurate due to the filter selection process, can, if used carefully, provide an indication of porosity. An average porosity level of around 25% was predicted. This porosity level falls within the range of the desired porosity for bone in-growth structures.

[0158] In consideration of the potential application, it is important to minimize loose surface contamination and demonstrate the ability to fully clean the surface. Laser cleaning or acid etching technique may be effective. Additionally, a rigorous cleaning protocol to remove all loose powder may entail blowing the porous structure with clean dry compressed gas, followed by a period of ultrasonic agitation in a treatment fluid. Once dried, a laser scan may be used to seal any remaining loose particles.

[0159] On examination, all candidate materials and substrates were selectively fused to produce a complex interconnected pore structure. There were small differences in certain process parameters such as speed and beam overlap percentage. From FIG. 12 it can also be seen how the Ti build has successfully fused with the Ti alloy substrate using a laser power of 82 W cw, beam overlap of −40% and a laser scanning speed of 180 mms.sup.−1. With the ability to produce structures with a controlled porosity, this demonstrates how the Direct Laser Remelting process can be used as a surface modification technology. Certain metal combinations interacted unfavourably and resulted in formation of intermetallics, which are inherently brittle structures. To overcome this problem it may be necessary to use a bond coat with the substrate. It is then possible to build directly on to the substrate with a porous build. A typical example of the use of a bond coat is provided by the combination of titanium on to a cobalt chromium substrate. Tantalum also was an effective bond coat in this example. Note that the bond coat may be applied by laser technology, but other methods are also possible such as gas plasma deposition.

[0160] The non-contact surface profilimeotry (OSP) images shown in FIGS. 13A to 13D show the surface profile. In addition, the Surface Data shown in FIGS. 14A and 14B and 15A and 15B show a coded profile of the plan views of the samples. FIG. 14A shows Ti on Ti (OSP Surface Data) where v=200 mms-.sup.1, FIG. 14B shows CoCr on Ti (OSP Surface Data) where v=200 mms-.sup.1, and FIG. 15A shows Nb on Ti (OSP Surface Data) where v=200 mms-.sup.1 and FIG. 15B shows Ta on Ti (OSP Surface Data) where v=200 mms-.sup.1.

[0161] FIGS. 16A to 25 are scanning electron microscope (SEM) micrographs of a series of single layer Ti on CoCr and Ti on Ti images that were produced prior to the multi-layer builds shown in FIGS. 8 and 9 respectively and as follows.

[0162] FIG. 16A shows Ti on CoCr (Single Layer; SEM Micrograph) v=160 mms-.sup.1;

[0163] FIG. 16B shows Ti on CoCr (Single Layer; SEM Micrograph) v=160 mms-.sup.1;

[0164] FIG. 17A shows Ti on CoCr (Single Layer; SEM Micrograph) v=170 mms-.sup.1;

[0165] FIG. 17B shows Ti on CoCr (Single Layer; SEM Micrograph) v=180 mms-.sup.1;

[0166] FIG. 18A shows Ti on CoCr (Single Layer; SEM Micrograph) v=190 mms-.sup.1;

[0167] FIG. 18B shows Ti on CoCr (Single Layer; SEM Micrograph) v=200 mms-.sup.1;

[0168] FIG. 19A shows Ti on CoCr (Single Layer; SEM Micrograph) v=210 mms-.sup.1;

[0169] FIG. 19B shows Ti on CoCr (Single Layer; SEM Micrograph) v=220 mms-.sup.1;

[0170] FIG. 20A shows Ti on CoCr (Single Layer; SEM Micrograph) v=230 mms-.sup.1;

[0171] FIG. 20B shows Ti on CoCr (Single Layer; SEM Micrograph) v=240 mms-.sup.1;

[0172] FIG. 21A shows Ti on Ti (Single Layer; SEM Micrograph) v=160 mms-.sup.1;

[0173] FIG. 21B shows Ti on Ti (Single Layer; SEM Micrograph) v=170 mms-.sup.1;

[0174] FIG. 22A shows Ti on Ti (Single Layer; SEM Micrograph) v=190 mms-.sup.1;

[0175] FIG. 22B shows Ti on Ti (Single Layer; SEM Micrograph) v=200 mms-.sup.1;

[0176] FIG. 23A shows Ti on Ti (Single Layer; SEM Micrograph) v=220 mms-.sup.1;

[0177] FIG. 23B shows Ti on Ti (Single Layer; SEM Micrograph) v=230 mms-.sup.1;

[0178] FIG. 24A shows Ti on Ti (Single Layer; SEM Micrograph) v=240 mms-.sup.1;

[0179] FIG. 24B shows Ti on Ti (Single Layer; SEM Micrograph) v=240 mms-.sup.1;

[0180] The method according to the present invention can produce surface structures on all powder/baseplate combinations with careful selection of process parameters.

[0181] As described above, the process is carried out on flat baseplates that provide for easy powder delivery in successive layers of around 100 μm thickness. Control of powder layer thickness is very important if consistent surface properties are required. The application of this technology can also be applied to curved surfaces such as those found in modern prosthetic devices; with refinements being made to the powder layer technique.

[0182] The structures have all received ultrasonic and aqueous cleaning. On close examination, the resultant porous surfaces produced by the Direct Laser Remelting process exhibit small particulates that are scattered throughout the structure. It is unclear at this stage whether these particulates are bonded to the surface or loosely attached but there are means to remove the particulates if required.

[0183] The Direct Laser Remelting process has the ability to produce porous structures that are suitable for bone in-growth applications. The powdered surfaces have undergone considerable thermal cycling culminating in rapid cooling rates that have produced very fine dendritic structures (e.g. FIGS. 13A to 13D).

[0184] The Direct Laser Remelting process can produce effective bone in-growth surfaces and the manufacturing costs are reasonable.

[0185] In the preceding examples, the object has been to provide a porous structure on a base but the present invention can also be used to provide a non-porous structure on such a base to form a three-dimensional structure. The same techniques can be utilized for the materials concerned but the laser processing parameters can be appropriately selected so that a substantially solid non-porous structure is achieved.

[0186] Again, a technique can be used to deposit the powder onto a suitable carrier, for example a mold, and to carry out the process without the use of a base so that a three-dimensional structure is achieved which can be either porous, as described above, or non-porous if required.

[0187] Additionally, the porosity of successive layers of powder can be varied by either creating a specific type of unit cell or manipulating various dimensions of a given unit cell.

[0188] There have been a number of studies to determine the optimum pore structure for maximization of bone in-growth on prostheses. The general findings suggest that optimum porosity is between approximately 20% and 40%, and aim to mid value with a mean volume percent of voids of about 70%. The preferred pore structure is interconnected, with a minimum pore size between about 80 μm and 100 μm and a maximum pore size between 80 μm and 800 μm. The structured thickness for in-growth is 1.4-1.6 mm, but can be larger or smaller depending on the application.

[0189] In the present invention porous structures are built in the form of a plurality of unit cells. Many designs of unit cells are possible to give the shape, type, degree, and size of porosity required. Such unit cell designs can be dodecahedral, octahedral, diamond, as well as many other various shapes. Additionally, besides regular geometric shapes as discussed above the unit cells of the present invention may be configured to have irregular shapes where various sides and dimensions have little if any repeating sequences. The unit cells can be configured to constructs that closely mimic the structure of trabecular bone for instance. Unit cells can be space filling, all the space within a three-dimensional object is filled with cells, or interconnected where there may be some space left between cells but the cells are connected together by their edges.

[0190] The cells can be distributed within the construct a number of ways. Firstly, they may be made into a block within a computer automated design system where the dimensions correspond to the extent of the solid geometry. This block can then be intersected with the geometry representing the component to produce a porous cellular representation of the geometry. Secondly, the cells may be deformed so as to drape over an object thus allowing the cells to follow the surface of the geometry. Thirdly, the cells can be populated through the geometry following the contours of any selected surface.

[0191] The unit cell can be open or complete at the surface of the construct to produce a desired effect. For instance, open cells with truncated lattice struts produce a surface with a porosity and impart the surface with some degree of barb.

[0192] Modifying the lattice strut dimensions can control the mechanical strength of the unit cell. This modification can be in a number of key areas. The lattice strut can be adjusted by careful selection of build parameters or specifically by changing the design of the cross-section of each strut. The density of the lattice can similarly be adjusted by modification of the density of the unit cells as can the extent and shape of porosity or a combination thereof. Clearly the overall design of the unit cell will also have a significant effect of the structural performance of the lattice. For instance, dodecahedral unit cells have a different mechanical performance when compared to a tetrahedral (diamond) structure.

[0193] As shown in FIGS. 32A and 32B, in a tetrahedron 8, each point 10, 12, 14, and 16 is the same distance from the neighboring point. This structure is analogous to the arrangements of the carbon atoms in diamond.

[0194] Each carbon atom in the diamond is surrounded by four nearest neighbors. They are connected together by bonds that separate them by a distance of 1.5445 angstroms. The angles between these bonds are 109.5 degrees. As a result, the central atom and its neighbors form a tetrahedron. This geometry as in the case discussed herein may then be scaled to appropriate value for the pore construct required.

[0195] The two key parameters used to define the relations regarding height, surface area, space height, volume of tetrahedron, and the dihedral angle of a tetrahedron are the strand length of the tetrahedron and, i.e., the diameter or height and width, cross section area of the strand i.e., strut. These two parameters control the pore size and porosity of the structure. The parameter editor and relation editor within a typical CAD system can be used to control these parameters. Hence, by changing the parameters one can change the fundamental properties of the porous structure. As shown in FIGS. 32A and 32B, the diamond structure may have a circular cross-section strands or square cross-section strands. Although only two strand cross-sections are illustrated, strands having various cross-sections are possible. Further, this is true with most of the designs for the unit cell.

[0196] To create the mesh as shown in FIG. 32C, the unit cell can be instanced across the 3-D space to produce the required lattice. FIG. 33 illustrates a view of a diamond lattice structure with and without laser beam compensation. Laser beam compensation essentially allows the diameter of the beam to be taken into account. Without it the constructed geometry is one beam diameter too wide as the beam traces out the contour of the particular section being grown. When laser beam compensation is utilized, the contour is offset half a beam diameter all around the constructed geometry which is represented in the CAD file. Although various parameters may be used, the parameters employed to create the lattices of FIG. 33 include a laser power of 90.5 watts with an exposure time of 1,000 μsec from a point distance of 90 μm. Table 1 illustrates various other examples of parameters that may be used to create various unit cells.

TABLE-US-00008 TABLE 1 edge laser point length diameter power exposure distance Part build on SLM μm μm Watts μsec μm Diamond Structure 2000 200 90.5 1000 90 Diamond Structure with 2000 200 90.5 1000 90 compensation Dodecahedron Structure 1500 200 68.3 1000 90 Dodecahedron Structure 1500 200 68.3 1000 90 with compensation Modified Truncated 1500 200 90.5 1000 90 Octahedron

[0197] As shown in FIGS. 34A and 34B, the porous structure can also be created using a unit cell in the shape of a dodecahedron. The regular dodecahedron is a platonic solid composed of 20 polyhedron vertices, 30 polyhedron edges, and 12 pentagonal faces. This polyhedron is one of an order of five regular polyhedra, that is, they each represent the regular division of 3-dimensional space, equilaterally and equiangularly. This basic unit cell for a decahedron mesh can be built up in a CAD package using the following calculations and procedure. The dodecahedron has twelve regular pentagonal faces, twenty vertices, and thirty edges. These faces meet at each vertex. The calculations for a side length of a dodecahedron are given by simple trigonometry calculations and are known by those in the art.

[0198] In a method of use, a sweep feature is first used to model the dodecahedron structure by driving a profile along a trajectory curve. The trajectory curves are constructed from datum points corresponding to the vertices of the dodecahedron connected by datum curves. The type of profile remains constant along the sweep producing the model shown in FIG. 34A. The size and shape of the profile can be designed to suit the particular application and the required strut diameter. Once a particular unit cell has been designed, the cell can be instanced to produce a regular lattice as shown in FIG. 34B. As a dodecahedron is not spaced filling, meshes are produced by simple offsetting of the unit cell and allowing some of the struts to overlap. This method of overlapping may be used with the alternate shapes of the unit cell.

[0199] FIG. 35 shows a view of a dodecahedron (with and without laser beam compensation, from left to right) structure using selective laser melting process parameters. Once again, although the parameters may be varied, the lattices of FIG. 35 were created using the following parameters; a laser power of 90.5 watts, exposure of the powder for 1,000 μsec and a point distance of 90 μm.

[0200] As shown in FIGS. 36A and 36B, the unit cell of the present invention may also be constructed in the shape of a truncated octahedron. A truncated octahedron has eight regular hexagonal faces, six regular square faces, twenty-four vertices, and thirty-six edges. A square and two hexagons meet at each vertex. When the octahedron is truncated, it creates a square face replacing the vertex, and changes the triangular face to a hexagonal face. This solid contains six square faces and eight hexagonal faces. The square faces replace the vertices and thus this leads to the formation of the hexagonal faces. It should be noted here that these truncations are not regular polyhedra, but rather square-based prisms. All edges of an Archimedean solid have the same length, since the features are regular polygons and the edges of a regular polygon have the same length. The neighbors of a polygon must have the same edge length, therefore also the neighbors and so on. As with previous unit cells, various dimensions such as the octahedron height, octahedron volume, octahedron surface area, octahedron dihedral angle, and truncated octahedron volume, truncated octahedron height, truncated octahedron area, truncated octahedron volume, truncated octahedron dihedral angle can be determined by simple trigonometry and are known by those skilled in the art.

[0201] In a method of use, a CAD model of the truncated octahedron is constructed using the sweep feature and calculations and dimensions are incorporated using basic trigonometry. Two tessellate the unit cell, the unit cell is first reoriented to enable easy tessellation and to reduce the number of horizontal struts in the model. Further, the model can be modified to remove all of the horizontal struts as shown in FIG. 38A. The modified structure is reproduced in order to save file size in the Stereolithography (“STL”) format of the program. Next, in order to create the unit cells, the method of using a laser melting process is performed. In one preferred embodiment, the parameter chosen includes a laser power of 90.5 watts, an exposure of 1000 μsec with a point distance of 90 μm. FIG. 38B illustrates a lattice structure formed using a plurality of individual truncated octahedron. As discussed earlier, the removal of various struts can create a barb effect on the exterior surface of the lattice structure.

[0202] As shown in FIGS. 39A-D, it is possible to reduce the size of the unit cell geometry. Also as shown, it is possible to manufacture open cell structures with unit cell sizes below 1 millimeter. FIG. 39A illustrates truncated octahedron structures manufactured using the laser melting process. All the structures were created using a laser power of 90.5 W, and a point distance of 90 μm; however, from left to right, the exposure time was varied from 500 μsec and 100 μsec. FIG. 39B illustrates similar structures and parameters as used with FIG. 39A, however, the unit cell used to create the lattice is diamond. FIGS. 39C and 39D illustrate a side view of the truncated octahedron structure of FIG. 39A and the diamond structure of FIG. 39B, respectively. Table 2 includes various manufacturing parameters used to construct various unit cell structure.

TABLE-US-00009 TABLE 2 Length of Width of Point Part build on Strand length strand c/s strand c/s Laser Power Exposure distance SLM μm μm μm Watts μsec μm Truncated 3000 50 50 90.5 500 90 Octahedron Truncated 3000 50 50 90.5 300 90 Octahedron Truncated 3000 50 50 90.5 100 90 Octahedron Truncated 1000 50 50 90.5 500 90 Octahedron Truncated 1000 50 50 90.5 300 90 Octahedron Truncated 1000 50 50 90.5 100 90 Octahedron Diamond 700 50 50 90.5 500 90 Structure Diamond 700 50 50 90.5 300 90 Structure Diamond 700 50 50 90.5 100 90 Structure

[0203] Random representative geometries may be made from the current regular unit cells by applying a random X, Y, Z perturbation to the vertices of the unit cells. One such example can be seen in FIG. 40. In another aspect of the present invention, various freestanding constructs can be generated. In a typical manufacturing procedure for the production of a construct, in this case a femoral hip component, the laser melting of appropriate metallic powders is employed. Table 3 listed below, includes various examples of equipment and material used in the construct, as well as typical software utilized.

TABLE-US-00010 TABLE 3 Equipment/Software Description Magics V8.05 CAD software package used for manipulating (Materialise) STL files and preparing builds for Rapid Manufacture (RM) Python Programming language MCP Realizer SLM machine using 100w fibre laser 316L gas atomized Metal powder with an mean particle size of metal powder approximately 40 μm Osprey Metal Powders Ltd

[0204] In one example of this procedure an STL file of hip component 50 is loaded into an engineering design package such as Magics, as shown in FIG. 43A. The femoral attachment 51 may then be segmented from the body 52 of the construct. The femoral attachment 51 may then be scaled down to 80% of its original size and reattached to the body 52 of the implant 50 as shown in FIG. 44. This permits the implant to act as a structural core for the surface coating. The selection of the amount of scaling or indeed the design of the core allows for the production of the required structural properties of the stem. Thus, the core may either be scaled down even more or less to meet the required needs of the implant. A Boolean operation may next be performed in Magics to subtract the reduced femoral attachment from the original. This creates a “jacket” 56 i.e., mold to be used as the interconnecting porous construct as shown in FIG. 45.

[0205] Jacket 56 is processed via a bespoke application that populates STL shapes with repeating open cellular lattice structures (OCLS). The OCLS used in this instance is a repeating unit cell of size 1.25 millimeters and strand diameter 200 μm. FIG. 46A illustrates a representation of a single unit cell of the OCLS which will be used to populate jacket 56. The OCLS “jacket” 56 as shown in FIG. 46B will act as the porous surface coating of the femoral attachment 50. Once produced, the OCLS is sliced using a bespoke program written in the Python programming language with a layer thickness of 50 μm. The main body of the construct is then loaded into Fusco, a user interface for the MCP realizer. The file is then prepared for manufacture by slicing the file with a 50 μm layer thickness and applying the hatching necessary for building solid constructs. The component and OCLS femoral coating are then merged as shown in FIG. 47. The component may then be built on the SLM system as shown in FIG. 48A with typical process parameters being shown in table 4 below.

TABLE-US-00011 TABLE 4 Slice height Power Exposure Feature (μm) (watts) (μs) P.sub.dist (μm) H.sub.dist (mm) Solid layer 100 90.5 800 80 0.125 Porous layer 100 90.5 3500 N/a (spot) N/a

[0206] Although the present invention has been described with regard to the femoral hip component as shown in FIG. 48A, the present invention may also be used to construct additional elements. For example, other elements include an acetabular cup component illustrated in FIGS. 49A-49C, augments from knee and hip surgery, FIGS. 50A and 50B, spinal components FIGS. 51A-51C, maxillofacial reconstruction FIGS. 52A and 52B, part of a special Nature, FIG. 53, and other additional irregular shapes such as that shown in FIG. 54. The list of illustrative components above is only an example of various constructs which may be composed using the method as disclosed herein and should be thought of as being inclusive as opposed to exclusive.

[0207] In other aspect of the present invention an existing product may be coated with various metal layers and then scanned with a laser in order to produce a finished product. In order to apply coating to existing products having either concave and/or convex profiles the present invention i.e., SLM requires the design of a special powder lay system. One such example was conducted and is shown in FIGS. 55A-59B. Specifically, a convex surface was created by using build apparatus 60 as shown in FIGS. 55A-58. Build apparatus 60 includes a rotating piston 62 and a cylinder onto which the convex surface 64 to be coated was mounted. As the component rotates on the cylinder, it was made to drop in the Z-direction using platform 66 within the SLM machine. Powder 71 was deposited onto the side of the component using a powder hopper 68 and a wiper device 70 that runs up against the surface of the component. Once the correct amount of powder has been established a laser (not shown in the figures) in conjunction with a computer and various programming packages, including those already discussed, were used to apply a laser beam to the powder in a predetermined manner. The powder was deposited by hopper 68 and wiped to the correct height by wiper device 70. A full layer of metal powder was deposited by rotation of the cylinder through a full 360 degrees. However, the synthesis of the laser melting process and the layer production process requires that only a fraction of the circumference is layered and melted at any one time. For example, the method from production of a full layer would require that the service be built up from, possibly individual quarter revolutions and melting steps as depicted in FIG. 59A. Preferably the laser melting process is fast enough that the discreet stepping process tends to be a continuous one with melting and rotation as well as layering occurring at the same time so as to increase throughput. FIGS. 55A-58 illustrate the sequence of operations with a final coated sample being shown in FIGS. 59A and 59B. In FIG. 59A, the lattice structure was built 3 mm thick and disposed against a 70 mm diameter steel hemisphere. In FIG. 59B, the same hemisphere was used, but the lattice structure is 6 mm thick. FIGS. 60A-60B are CAD illustrations of the final assembly of a product component.

[0208] In an alternate embodiment of the present invention, the process can be parallelized by addition of many pistons and cylinder pairs around a central laser beam. Optimal laser alignment to the surface can be achieved by a number of methods, including inclining the piston and cylinder pairs so the powder surface and the part surface are correctly aligned normal to the laser beam. Typical operating parameters are shown in Table 5 below.

TABLE-US-00012 TABLE 5 Slice height Exposure (μm) Power (watts) (μs) P.sub.dist (μm) H.sub.dist (mm) 100 90.5 700 80 0.125

[0209] In another aspect of the present invention the laser produced porous structure system may be used to manufacture a porous shell which then can be inserted over a substrate and sintered in order to fix permanently to the same. Some examples include the preparation of an acetabular cup component, a tibia knee insert component, and a femoral insert as well as many additional products. In order to illustrate this aspect of the present invention, reference will be made to the outer profile of an acetabular component which serves as an inner profile of a “cap” to insure that an accurate fit is achieved when the cap is set on the substrate (acetabular shell). The cup is built to a thickness of 1.5 millimeters for example using a diamond configured construct to develop the interconnecting porosity. The metal powder used in one example is stainless steel. The processing parameters are shown in Table 6 listed below:

TABLE-US-00013 TABLE 6 Slice height Exposure (μm) Power (watts) (μs) P.sub.dist (μm) H.sub.dist (mm) 100 90.5 2000 N/a N/a
However, the process parameters are dependent on the metal used and if a different metal, say for example, titanium was used, the parameters would be different. FIGS. 61A-62B illustrate finished products manufactured by SLM.

[0210] In order to achieve a better and tighter fit of the cap over the component, some adjustments to the geometry of the cap may be considered. For example, the inclusion of a rim 70 on the inner surface of the cap that interfaces with the groove 72 on the outer surface of the acetabular cup component 68 may be included. This mechanism acts a simple lock and gives both security and extra rigidity during the sintering process. Additional modifications may be utilized to improve closeness of the fit and stability. For instance, the introduction of “snap-fits” which are apparent in everyday plastic components may be employed to provide a more reliable attachment mechanism between the two elements. Typical pads or center pads for both the femoral and tibial knee components can be produced by the SLM process and dropped or snapped fit into place to the components and then sintered to attach firmly to the underlying substrate. As previously stated, this technique can apply to other components where a porous outer surface is required to interface with either soft or hard tissue.

[0211] A further improvement in the mechanical and microstructural properties of the porous construct may be achieved by either conventional sintering under vacuum or inert atmosphere and/or hot isostatic pressing using temperature regimes known in the state of the art. As the constructs possess high density properties throughout their strands minimal degradation in the structure of the construct is apparent.

[0212] In another aspect of the present invention, the appearance of the porous construct can be changed by the alteration of the processing conditions or by the introduction of an acid etch process. For example, the laser power or laser residence time may be reduced or a combination of both which creates struts of the porous construct having a coating with layers of unmelted metal particles firmly adhered to the strut surfaces. This has the effect of producing additional porous features that offer a new dimension to the overall porous structure of the construct. Such features are able to interact with cells in a different manner than the microstructure imparted by the lattice construct and provide extra benefits. A typical example of such construct with this satellite appearance as depicted in FIG. 63 together with the processing parameters is employed. The structure illustrated in FIG. 63 was created using a laser power of 44.2 W and exposure time of 400 μsec. The metal layer thickness was 50 μm.

[0213] It is also possible to remove these satellites by an acid etching process and a strong acid. The acid may consist of a mixture of 10 milliliters of hydrogenfloride (HF), 5 milliliters of nitric acid (HNO.sub.3) and 85 milliliters of H.sub.2O. the HF and HNO.sub.3 were respectively 48% and 69% concentrated. FIGS. 64 and 71 show the effects of such an acid's etch with respect to time with the relevant conditions being noted. It can be seen clearly that the solids are moved to give a pure melted lattice construct. It is also clearly evident that the overall openness within the lattice is increased by the removal of the satellites. Additionally, prolonged exposure to the acid etch mix does result in some reduction in strut thickness which may also increase the lattice size further. This enables the production of struts having a reduced thickness to be created by the STL method. Other acid types and combination may also be applied to obtain similar results.

[0214] It will be appreciated that this method can, therefore, be used to produce an article from the metals referred to which can be created to a desired shape and which may or may not require subsequent machining. Yet again, such an article can be produced so that it has a graded porosity of, e.g., non-porous through various degrees of porosity to the outer surface layer. Such articles could be surgical prostheses, parts or any other article to which this method of production would be advantageous.

[0215] Although the invention herein has been described with reference to particular embodiments, it is to be understood that these embodiments are merely illustrative of the principles and applications of the present invention. It is therefore to be understood that numerous modifications may be made to the illustrative embodiments and that other arrangements may be devised without departing from the spirit and scope of the present invention as defined by the appended claims.