Materials Forum (1990) 14, 161 -173
THERMAL SPRAYING FOR BlOCERAMlC APPLICATIONS C. C. Berndt, * G. N. Haddadt, A. J. D. Farmer' and K. A. Gross*
Bioceramics are used in many applications and forms within the human body. Some major shortcomings of bulk ceramics concern their high variability in material properties and restriction to only limited forming operations. However, the excellent biocompatibility of many ceramics has provided a strong incentive to promote investigations for alternative manufacturing processes such as ceramic coating technology. Coatings may be applied by pyrolytic deposition, electrophoretic deposition, sintering, hot isostatic pressing, controlled oxidation, sol-gel techniques, physical vapour deposition (PVD), chemical vapour deposition (CVD), dipping processes and thermal spraying methods. The present work will focus on the thermal spraying of hydroxy- apatite (HAp) onto metallic substrates which are intended for orthopaedic applications. It is intended to present the pertinent literature on HAp as it relates to thermal spraying, and also briefly detail some thermal spraying experiments. The thermophysical processes which occur during thermal spraying will be related to the structure/property relationships of the resultant coating.
1 INTRODUCTION Hench classifies ceramics into three classes Materials used for prosthetic applications include which exhibit different rea~tivities.~Figure 1 shows polymers, ceramics, metals and composites. The the relative reactivity of bioresorbable, surface active present work is concerned with orthopaedic appli- and inert ceramics. The biodegradation or bio- cations and, therefore, the focus will be directed resorption curves can change depending upon the towards materials that are analogous to the com- physical and chemical properties of the implant as position of bone. Osseous tissue contains both a well as biological relationships between the implant calcium phosphate ceramic (hydroxyapatite) and and body. Resorbability can be increased by a protein called collagen. Hydroxyapatite (HAp) porosity, reductions in crystal size, more crystal found in osseous tissue has its structure modified imperfections and smaller grain size. The chemical by monovalent and divalent ions such as Na', K', properties giving rise to the same trend, in HAp ~g*',COG- and other ions. The obvious choice for ceramics, would be ionic substitutions (for example repairing or replacing the supporting structure of COG- or ~g~'),the presence of /3-whitlockite and the body would be a calcium phosphate ceramic.' impurities. The biological factors include pH drops Table 1 details some of the ceramics containing due to cell-mediated factors (e.g. macrophages, calcium and phosphate ions, which are currently osteoclasts and fibroclasts), osteolytic infection and used as biomaterials. Some of these materials con- disease, degree of bone contact, type of bone (e.g. tain other oxides and form glasses; these ceramics skull, mandible and femur), animal species, age, are referred to as 'bioglasses' but should not be sex, hormone levels and genetic predisposition.' confused by a patented ceramic composition which The time scale for bioresorption and the maximum is known as 'Bioglass'. The implant must not only resorbability within the body can change signifi- possess chemical cornpati bility but also good bulk cantly if all these factors are taken into considera- @ properties and exhibit a strong bond with the tion. Figure 1 adapted from Hench is thus a very surrounding tissue in applications where they simplified picture, which should only be used in experience high load. Bioglass and HAp are rarely pointing out the major distinctions between the T used as stress members in bulk form since their three classes of bioceramic~.~It should be noted intrinsic strength is poor. Instead they find appli- that in vivo assessment of biodegradation includes cations where the mechanical requirements are low ascertaining the reduction in implant size, any (Table 2). It should be noted that ceramics can be changes in pH and histomorphometry (measure- designed for a range of applications. ment of pore size variation). In vitro assessment is much simpler and includes measuring the specimen size and any surface degradation via ' Department of Materials Engineering, Monash University, Clayton, Vic. 3168, Australia. microscopy. t Division of Applied Physics, CSIRO, Lindfield, Vic. 2070, Inert bioceramics of current interest include Australia. Manuscript received 3 April 1990; in final form 19 July 1990. alumina, partially stabilized zirconia, titania, silicon C. C. Berndt et al. e
TABLE 1 Bioceramics containing calcium and/or phosphate ions in their structure
Bioceramic Chemical composition Trade names Reference %
Hydroxyapatite (HAp) Calo(P04)6(OH)z Durapatite, Calcitite 2 P-Whitlockite (P-TCP) p-Ca3(P04)~or 3Ca0.P2O5 Synthos, Synthograft 3,4 Tri-calcium phosphate (a-TCP) a-Ca3(PO& - 5 , Tri-sodium phosphate Na3P04 - 5 -. Calcium hydroxide Ca(0I-i)~ - 5 Calcium sulfate CaS04 - 5 Wollastonite Ca0.Si02 - 2 Bioactive glass Na~0-CaO-P~0~-Si0~system Bioglass 2 ANV glass ceramic* HAp and CaO.SiOz - 2 Bioactive glass ceramic (Na,K) Mg3(AlSi3010)F~ - 2 and HAp crystals Bioactive glass ceramic ? Ceravital 2 Bioactive composites Palavital 2
* Glass ceramic containing apatite and wollastonite.
TABLE 2 Clinical applications for HAp bioceramics
Application as stress members Artificial joints (i.e. knee, hip, shoulder, elbow, etc) Spinal fusion Low stress applications 6 2 Cranioplasty Implantation time Implantation time Middle ear implants (months) (months) Tooth implants Fig. 2. Strength-timerelationship for: (a) tibia with a bone Alveolar ridge reconstruction (orthognathic surgery) g Vascular grafts plate and (b) a femur possessing a hip prosthesis. Percutaneous devices
The in vivo time dependent behaviour of the ceramic depends on the implant material and also the Relative mechanical requirement of the implant by the bone. Reactivity For instance, a ceramic coated implant performs t most effectively where bone cannot regenerate enough to repair any previous damage. The result- ing mechanical behaviour of the coated material Sllrface active from the time of implantation would show an insig- nificant reduction in strength. The poor mechanical properties of bulk ceramics 0.1 1 10 loo loo0 Tiime (days) may be circumvented by using surface reactive w ceramics which interface with the biological Fig. 1. Comparison of the relative reactivity of biomaterials (adapted from Hen~h).~ systems. A coating can be made of this material and the bulk can be made of a tough but less nitride, silicon carbide, low temperature isotropic compatible or inert metal that provides mechanical carbon and vitreous carbon. Surface active cer- support.1° The coating is applied to a metallic amics include HAp and bioactive glasses while substrate (commonly Ti-6AI-4V) to enhance the tricalcium phosphate is used for bioresorbable fixation of the implant to the osseous tissue." This ceramics.' coated implant can then take on the characteristics A ceramic can be tailored to have the properties of both metallic and ceramic materials. Results from 9 suitable for each type of implant. One example of plasma spraying of HAp show that bone bonding this is in fracture healing of bones. The pin used to HAp coatings is essentially the same as to HAp to stabilize fractures will slowly dissolve and weaken bulk implants.12 HAP, bioactive glasses and bio- if it can be manufactured from an appropriate active glass ceramics can develop a bond strength material. Thus, the bone experiences an increasing which exceeds the bond strength of metals to bone load until it is restored to normal loading conditions. by a factor of three and the above ceramics are This ideal strength-time relationship is illustrated therefore mostly adopted for orthopaedic purposes. in Fig. 2a;' while Fig. 2b indicates the relationship The bond strengths of some coatinghubstrate which may exist for a hip prosthesis in a femur. systems which are attached to bone are reported Thermal spraying for bioceramic applications
TABLE 3 Bond strength of biocoatings
Coating materials Substrate material Bond strength Reference (MW Bioglass 45S5F Ti alloy 316L stainless steel Ti-6AI-4V Ti-6AI-4V Alumina Ti alloy 316L stainless steel Hydroxyapatite Ti-6AI-4V 316 stainless steel
in Table 3. It can be noted that the bond strength Key: = implant = sintered Co-Cr balls is variable and depends on, among other factors, 0 =bone = sintered Ti alloy mesh the type of bone (cortical or cancellous), the bone =bone -ceramic coating thickness and the testing conditions. cement HAp has been used to coat Ti-6AI-4V hip implants. The lower part of the stem is cylindrical for ease of bone bed preparation and implant insertion but the shape at the top deviates from a circular cross-section. If the full stem length was cylindrical the implant could rotate and the bone and repaired femur would then experience stresses which were atypical of those that are normally experienced. This could lead to premature implant failure. The coating is applied to the stem because this part interfaces with the osseous tissue and forms a bond joining the implant to the body. The coated prostheses which are commercially available show a difference in design philosophy. The stem can be totally covered with HAp or a section can be left uncoated. The uncoated section would be in the middle of the stem on the back where the implant experiences the least stress.
2 IMPLANT FIXATION AND Ti-V-AI ALLOYS Fig. 3. Fixation methods currently being used for An important reason for applying a coating or orthopaedic implants (a) bone cement, (b) sintered Co-Cr balls, (c) sintered Ti alloy mesh and (d) ceramic modifying the surface is for implant fixation to coatings. tissue. This tissue could be bone, a part of an organ, skin or flesh. Orthopaedic applications are con- cerned with bone bonding so the following fixation mechanisms examine the bonding of the bone to implant collagen bundles the implant. Each of these techniques (Fig. 3) have evolved separately and are based on different tissue mechanisms. A thermosetting cement such as PMMA is used to adhesively fix the implant in place (Fig. 3a). Disadvantages are cell necrosis from the exothermic reaction and cement fai~ure.""~The reliability of this fixation technique over periods calcified amorphous layer bonding zone (
TABLE 5 Biomaterial coatings and their method of application
Method of application Coating material Substrate material Reference
Electrophoretic deposition HAP Titanium alloy Pyrolysis Carbon Titanium alloy Hydrolysis Calcium phosphate Carbon composite Immersion Bioglass Co-Cr alloy Dipping Bioglass 316L stainless steel Bioglass Co-Cr alloy Frit enamelling HAP Co-Cr alloy Chemical vapour deposition Carbon Ti-6AI-4V alloy TiN-alumina Stainless steel Physical vapour deposition TiN Co-Cr-Mo alloy Bioglass Alumina HAP Titanium alloy Carbon Stainless steel Sintering Polysulfone Titanium alloy Plasma spraying HAP Ti-6AI-4V alloy Alumina Ti-6AI-4V alloy Bioglass 316 stainless steel Hot isostatic pressing HAP Titanium alloy
Coating Copper anode . "
Coating processes
ing B frit enamelling I I carrier gas Spray stream Electrical Electrical connection connection 1 Substrate Hot isostatic pressing kve) (+ve) . . 10' I lo 10' 14 lo4 Fig. 6. Schematic diagram of the thermal spraying Microns process. Fig. 5. Typical thicknesses of coatings applied by different processes.
Table 5. The coating thickness as seen in Fig. 5, orthopaedic purposes. De Groot has proposed a varies from < 1 pm for sol-gel coatings, to 1-40 limitation of 50 pm whereas Osborn recommends pm for PVD and CVD coatings to > 1000 pm for 200 pm. 2943 coatings which are produced by thermal spraying, dipping and sintering methods. Coating thickness 4 THERMAL SPRAY TECHNOLOGY is an important parameter in bio-applications, since It is not intended to review the thermal spraying thick coatings tend to crack and fail prematurely process; however it is useful to summarize some due to their greater bending moment. A coating features of the technology which are of special thickness > 100pm can potentially introduce fatigue emphasis to the manufacture of bioceramic coat- failure under tensile loadings. On the other hand, ings. Thermally sprayed coatings are produced by the coating must be thick enough to avoid dissolu- injecting particles of 20-120 pm in diameter into tion by the physiological solutions. Physiological the tail flame of a high temperature source (Fig. 6).44 media will have direct access to the implant-bone The source may be either from a d.c. plasma (hence interface if the osseous tissue does not interface 'plasma spraying'), an oxyacetylene gas mixture directly onto the implant; and thereby allow ('flame spraying') or an explosive mixture of degradation of the implant and hence its loosening. acetylene, oxygen and nitrogen ('detonation gun This is especially true when infection occurs spraying'). Table 6 summarizes some of the fea- because the accompanying fall in pH renders the tures of these processes. The underlying principles coating prone to dissolution. There is some dispute behind all these processes are similar. Thus, the in the open literature concerning the optimum thick- particles are heated and then driven at velocities ness of bioceramic coatings which are intended for of up to 300 m s " against a substrate. The molten 166 C. C. Berndt et al.
TABLE 6 Comparison of spraying techniques
Process Porosity Advantages Disadvantages Flame 10- 15O/0 Economic High porosity Prior substrate heat treatment Arc 10-15% High deposition rates High porosity Greater bond strength than flame spraying Prior treatment of substrate Plasma 1 - 10°/o High bond strength Costly Can use high m.p. materials Inert heat source Detonation 0-1 O/O High hardness Costly Gun Very high bond strength
Coam~Qualitv -Chemical composition Energy Mechanical properties Impurities -Gases Thermal prope&es Particle size/ distribution Nozzle type Chemical properties Morphology Powder feed Thickness Density Spray distance Porosity Phases Reliability
Heat transfer Melting point Chemical composition Heat conductivity Time in plasma Thermal expansion Thermal stability Substrate preparation
Fig. 7. Interdependencies of powder and coating quality (modified from S~hwier).~'
and semi-molten particles rapidly solidify on impact the titanium alloy) at 1000°C forms CaTiOa and against the substrate and an integral layer (- 0.5- tricalcium phosphate.16 The CaTiOs, a compound 2 mm in thickness) is built up by the overlaying formed from both the substrate and the HAp of many partic~es.~~Some droplets may not com- coating, acts as a bond layer. The formation of an pletely melt before impacting the surface due to interfacial compound is not unique to the plasma particle size, velocity and temperature distributions spraying process since this also occurs with within the stream of particles and this results in electrophoretic coatings.46Thus, the adhesion of embedded particles in the lamellar structure. HAp to Ti alloys is much stronger (- 60 MPa) than Ceramic coatings are usually not deposited bonding mechanisms, which rely on only mechani- directly onto the substrate due to their brittle nature. cal interlocking. It is important to point out that It is normal practice for these coatings to be overlaid the plasma spraying process can be carried out - onto a thermally sprayed metal (which is termed under either ambient or low pressure condition^.^^ as the 'bond coat') and this acts as a compliant Low pressure plasma spraying (LPPS) is carried layer between the substrate and the ceramic coat- out at - 6 kPa and gives rise to very dense coatings ing. However, no bond coats are used for HAp with high adhesion strengths. coatings on Ti-6AI-4V substrates. The coating The quality of thermal spray coatings is a complex adhesion to the substrate as well as the coating problem since there are various interdependencies integrity is often explained in terms of the mechani- (Fig. 7) between powder properties, the spray pro- cal interlocking of the particles when they solidify. cess and coating q~ality.~'Thermophysical prop- When HAp is deposited onto Ti-6AI-4V, a chemical erties are determined by the particle morphology reaction with titanium dioxide (the surface layer of and their composition. Melting behaviour is influ- 2 Thermal spraying for bioceramic applications
enced by particle size, thermal conductivity, melting The thermal spraying technique is widely docu- point, purity and spraying parameters. The deposi- mented for bioceramic applications and was tion of the molten particles and the final coating pioneered for biomaterial applications by Brown in quality are determined by the spray parameters the early 1980s. It was established that alumina such as the spray rate, gas type, gas flow, spraying coatings in the y-phase developed a bond strength distance and nature of the substrate. All stages of of - 25 MP~." There is, however, concern that the plasma spraying process are thus affected by plasma sprayed alumina exhibits questionable bio- the characteristics and the properties of the powder. compatibility because it reacts more readily with Transport during the thermal spraying operation physiological environments than its fused, sintered and subsequent powder properties which affect its or hot pressed counterparts. Aluminium ions are deposition onto a substrate are: bulk density, par- released and form bayerite (A1203-3H20)and other ticle shape, particle size, percentage of superfines compounds which have led to renal failure during (< 10 pm), particle surface smoothness, hydro- animal tests.49'50~he bond strength of alumina coat- scopic density (moisture pick-up), chemical stability ings decreases when they are aged in physiological and electrostatic charging effects (which may pro- media (Fig. 8). For example, the adhesion of plasma- duce particle clumping). Spherical powders of a sprayed alumina on a 316L stainless steel substrate given size distribution are important as they give is lowered from 34.5 to 10.3 MPa after ageing in a consistent melting behaviour which increases physiological medium (Medium 199) for 16 weeks.51 deposit efficiency. A uniform surface area ensures There is an initial rapid decrease in strength to 13.8 consistent melting but the melting characteristics MPa within 8 weeks. Any further loss in strength become erratic when the surface area, particle occurs at a much lower rate. dimensions and mass for similarly sized powders The plasma spraying of HAp was first docu- vary.48The optimum particle shape from the view- mented in the open literature by De Groot in the ? point of flow characteristics is spherical since then 1980s.16 This method is now being used by 'Joint the best aerodynamic performance is achieved. The Replacement Instrumentation' based in London, flight path of a particle that deviates from sphericity England to coat the femoral stem of 'The Furlong' will be more unpredictable. HAC Coated THR (total hip replacement) system'.52 The surface of the femoral stem of the hip implant is plasma sprayed to a thickness of > 200 pm and can then be pressure-inserted into a cavity slightly - 20 smaller than the implant to provide a no-motion 2 18 E fit with the surrounding bone. Within a week the 16 osseous tissue begins to grow onto the surface of 14 L the HAp coating to provide a functional interface r" 12 that can transfer stress.43 2 10 I. I 8 o 10 20 5 POWDER PRODUCTION AND COATING Ageing time (weeks) MANUFACTURE n Fig. 8. Adherence decrease of plasma sprayed alumina The calcium phosphate system although co - on Ti-6AI-4V after ageing in saline sol~tion.~' ceptually simple displays a rich and complex
TABLE 7 Sources of HAp powder
Source Name* Ca/P ratiot Surface area (m2 g) Reference -- - - Asahi Optical Co. Apaceram Euro Crystals Hy-Apatite Fischer Scientific HAP Fluka High resolution HAp Mallincrodt Corp HAP Merck HAP Monash University HAP NGK Spark Plug Co. HAp XVC-56 Plasma Biotal Ltd. HAp XVC-56 Podmore Calcined bone ash Sigma Calcium phosphate Trans-Tech, Inc. HAP Wako Pure Chemicals HAP * Terminology used by manufacturer. t The ideal ratio according to the stoichiometry of HAp is 1.67. 168 C. C. Berndt et al.
TABLE 8 Processes available for the production of HAp
Process Reactants Reference Hydrolysis c&H~(Po~)~~H~o Hydrothermal synthesis Ca(CH3C00)~H20& (C2H50)3PO Ca(N03)2 & (NH4)2HP04 CaHP04, & CaHP04.2H20 Hydrothermal exchange CaC03 & (NH4)2HP04with H20 Solid process Bovine bone Sol-gel Calcium and phosphorus alkoxides Spray pyrolysis Ca(N03)2& in Me-H20-HN03 Wet method CaC12.6H20 & (NH4)2HP04 CaC12 & Na2HP04 CaC12 & H3P04 Ca(N03)2.4H20& (NH4)2HP04 CaC03, CaO, Ca(OH)2, CaF2, CaCI2, Ca(N03)2 Wor CaC204powder with H3P04 solution Ca(N03)2 & Na2HP04 Ca3 (P04)2 & H20 Ca(OC2H5)rEtOH & H3P04-EtOH CaO (or Ca(OH)2) & CaHP04.2H20 CaHPOa.2H20 & CaCOn
TABLE 9 Dependence of morphology on the production process of HAp
Process Morphology Reference Hydrothermal synthesis Hexagonal-prism 63 Needle-like prisms or spherical particles 62 Hydrothermal exchange Variable 65 Spray pyrolysis Spherical 69 Wet method/freeze drying Plate-like 79 Wet method Plate-like 71 Jaggedhpherical 81 Jaggedkpherical 86
TABLE 10 They include hydrolysis of calcium phosphates, The Ca/P ratio of some calcium phosphates hydrothermal synthesis, hydrothermal exchange, -- Calcium solid-processes, spray pyrolysis, sol-gel and the wet phosphate Name Ca/P ratio method. The most commonly used technique is the - wet method (Table 8), and this allows a more exten- Brushite sive choice of reactants compared with the other Monetite processes. Pyrophosphate Octacalcium phosphate Variations in these routes lead to differences in Whitlockite morphology (Table 9), crystallographic structure, a-tricalcium phosphate Hydroxyapatite stoichiometry and density. In the wet process a Tetracalcium phosphate difference in composition could arise from ~~02- ~~ ~ or H20 that have a tendency to adhere to the particle^.^"^^ The density depends on the stacking of the separate crystals. The individual crystals are thermal chemistry.23The choices of reaction routes aligned more efficiently thus imparting a higher are many; each having reactants and conditions density to the powder. A process should be chosen unique to the process. Sources of HAp powders to obtain the powder with the desired charac- are detailed in Table 7. There are a total of seven HAp teristics. The overall density of the HAp is deter- production routes documented in the literature. mined by the stacking efficiency of the individual Thermal spraying for bioceramic applications
HAp crystals; so that a higher density will arise from The resulting gelatinous precipitate is allowed to efficient packing. settle and the clear liquid decanted. Crystal size is The powders used in the present study were < 1 pm;" but due to flocculation the crystals form prepared by the wet method that has been described aggregates - 20 pm in size. The size distribution by Aoki et a/.89 This method involves the addition of particles can be varied by altering the reaction of an acid (orthophosphoric acid) to a base (calcium conditions and, for instance, a larger particle size hydroxide). It is important that the reactants have results when the addition rate of the acid is lowered. the correct molar ratio of calcium and phosphorus. The temperature is allowed to settle to room Low pH conditions yield more stoichiometric pre- temperature after the reaction reaches completion cipitates, but as the pH increases less stoichiometric and the precipitate is left in the mother solution precipitates are formed." In the case of HAP, the overnight. Water can be evaporated by oven drying pH is greater than a value of 8 for the reaction. at 180°C for 36 h. This removes any absorbed water Careful monitoring of the pH level can be omitted from the external surface but is insufficient to if the reactants contain calcium and phosphorus remove water from ultrafine pores in the particles.92 in a molar ratio of 1.67 and the solution is well This water is removed during the calcining stage. mixed during the course of the reaction. A lower The HAp cake is usually processed by crushing, Ca/P ratio corresponds to a lower pH and leads grinding and sieving to the required particle size to the production of calcium monophosphate and of 25-185 pm. calcium phosphate dihydrate. These materials, Particles produced were of a roughly spherical whose formulas are shown in Table 10, are soluble morphology. At longer drying times (48 h) the caked in aqueous media and will resorb upon implantation. materials had a lower moisture content and the Both monetite and brushite formation is favoured resultant particle agglomerates were therefore by insufficient mixing of the calcium containing harder. Thus these materials, upon crushing, pro- solution as the PO:- solution is added. The PO:- duced a more angular surface morphology (Fig. 9). ion concentration builds up when it is not distributed The resulting powder was identified as amorphous homogeneously and a localized low pH is estab- HAp using X-ray diffraction. After calcining at 800°C lished. This part of the reaction vessel is richer in for 3 h the peak width decreased and the peak height PO:- and provides conditions for the formation of increased more than two-fold to 1500 cVs as shown soluble precipitates. in Fig. 10. This behaviour is characteristic for a Most wet methods entail adding the phosphate powder which has become crystalline. The three solution dropwise to the calcium ion containing major peaks indicated in Fig. 10 are located at 28 solution. The reason for this is that precipitated values of 31.8, 32.3 and 32.9 and have relative anions (PO:-) are generated slowly in solutions con- intensities of l,0.6 and 0.6 respective~y.~~Both crys- taining the metal ion (ca2') to be precipitated. This tallinity and purity of the source HAp powder are is controlled by hydrolysis of suitable compounds important for the development of a pure crystalline under the condition for formation of insoluble HAP. coating.94It is believed that an amorphous coating
Fig. 9. HAp particle morphology after: (a) a lor r7g drying time and (b) a short drying time. C. C. Berndt et al.
40.00
Fig. 10. XRD trace of HAP. will be resorbed by the body and thus a crystalline hindrance to the performance of the coating but form is most preferable. both tricalcium phosphate and tetracalcium phos- HAp plasma sprayed under ambient conditions phate may transform to HAp through hydrolysis partially transforms to tricalcium phosphate and cal- upon implantation in the human body. cium oxide. This transformation can be avoided by The transformation is accompanied by a volume keeping the plasma gas at a low hydrogen/argon change. For example, the crystal structure changes ratio during the plasma spraying process.95A low from orthorhombic (a-tricalcium phosphate) to hex- pressure plasma spraying facility was also used to agonal (HAp) which not only increases the packing apply coatings 100 pm thick. Particle sizes from efficiency of atoms but also leads to a decrease 25-185pm were used but only the particles <50pm in the unit cell volume. The manner by which this in diameter were totally molten (Fig. 11). To assist transformation influences the prosthesis fixation the melting of HAp the pressure could be raised has not been established. giving more dense, better formed coatings. In- complete melting could also be observed if the chamber pressure was too low. It was easier to melt 6 CONCLUDING REMARKS the HAp by plasma spraying under ambient con- The purpose of the present review has been to ditions; presumably because the heat transfer from outline the chemistry, processing schedules and the plasma effluent to the particles under vacuum properties of HAp materials which are intended for conditions is quite poor. bioengineering applications. The focus has been The phases which may form upon heating include on powder processing methods that enable tricalcium phosphate and tetracalcium phosphate. materials of the appropriate chemistry, particle size Whitlockite (i.e. tricalcium phosphate of the beta and particle distribution to be manufactured. Such phase) forms at 1200°C but with ~~02-it is known powders could be used for bulk ceramic processing to transform at lower temperature^.'^ At tempera- operations (e.g. pressing and sintering, etc) or to tures > 1400°C a phase transformation from the enable coating processes. The current experimental P-phase occurs and tetracalcium phosphate forms. work established that powders suitable for thermal These transformations have been reported in the spraying must be of approximately spherical plasma spraying process and might appear as a morphology, with a particle size from - 20-100 pm Thermal spraying for bioceramic applications
Fig. 11. Plasma sprayed HAP: (a) under ambient conditions and (b) under low pressure conditions.
and a mean particle size of - 55 pm. Powders of present work indicates that some thermal reduction " this specification have been successfully plasma of slightly impure HAp materials may be minimized sprayed under ambient and low pressure condi- by spraying under LPPS conditions. It was also . tions. No high quality deposits, as distinguished by established that coatings of HAp could be manu- . good adhesion to the substrate and internal coating factured with powders of finer particle size by integrity, are formed by flame spraying of the HAP. employing LPPS technology. The implication of The thermally sprayed deposits showed a range these results is that the character of the porosity of morphologies. The particles were either well- may be controlled by using a finer cut of powder melted and formed an integral part of the coating in conjunction with LPPS technology. or were only partially melted and exhibited many It is important to make some comments with high profile surface features. The partially melted regard to the mechanical properties of thermally coating would also be expected to exhibit a higher sprayed coatings. Data has been reported on bond degree of bulk porosity as well as larger sized strengths in the range of 60-70 MPa for HAP coat- porosity. This, is an important microstructural fea- ing~.'~It can be noted that such large values are ture of the coating to ascertain since the nature quite uncharacteristic for thermally sprayed ceram- of any porosity will determine the ingrowth rate of ics. For example, alumina and zirconia coatings bone tissue into the ceramic coating. It should also which are intended for aerospace applications, have be emphasized that relatively low mechanical recommended strengths of - 45 MPa. An additional properties of the coatings may not limit the utility fact is that these aerospace coatings are normally of the coating to a high degree, as they would in used in conjunction with a bond coat whereas no usual engineering applications. Thus, the usual mention of a bond coat is made for the HAp mater- requirement of obtaining dense and highly adherent ials. Thus, it seems that the presently used bio- coatings may, to a certain extent, be of less import- ceramic coatings may be over designed with respect ance than the need to control the pore size and to their intended use, and that the enhancement pore network so that any tissue/implant interactions of this material property might give something other are optimized. than the most preferred microstructure of the coating. The processing technology that has been most It has also been emphasized that HAp coatings widely documented is that of atmospheric pressure are not used in conjunction with a bond coat. Bond plasma spraying. It can be noted that traditional coats of Ti and Ni-AI alloys are normally used to thinking would propose that ceramics cannot improve the adhesion of the coating system onto benefit from LPPS since ceramics generally exhibit the substrate. However, very high strength values high melting points and are therefore relatively are reported in the literature for HAp materials and stable to thermal treatments. Thus, LPPS tech- it has been postulated that highly adherent oxide nology is normally restricted to the production of layers at the substrate surface promote this high high density coatings of reactive metals (e.g. adhesion. It can be further noted that there has titanium and nickel-chrome alloys). However, the been little work with regard to determining the C. C. Berndt et al. behaviour of coatings with respect to their fracture H. Hahn, P. J. Lare, R. H. Rowe, A. C. Fraker & F. Ordway, Corrosion and Degradation of Materials: toughness, fracture morphology and fracture mech- Second Symposium, ASTM STP 859, (eds A. C. Fraker anisms under environmental conditions. Another & C. D. Griffin), American Society for Testing and important material property which has not been Materials, Philadelphia, 1985, 179-91. # S. D. Brown, J. L. Drummond, M. K. Ferber, D. P. ascertained is the elastic modulus of the coating. Reed & M. R. Simon, Mechanical Properties of Such measurements would enable a database to Biomaterials (eds G. W. Hastings & D. F. Williams), be established so that modelling of the coating Wiley, 1980, Ch. 19. K. De Groot, Interceram, 1987,4, 38. behaviour could be performed. B. J. Anderson, M. A. R. Freeman & S. A. V. Swanson, It can be concluded that it is crucial to establish J. Bone & Joint Surg., 1972, 548, 590. E. Ebramzadeh, M. Mina-Araghi, I. C. Clarke & R. the above material properties; and, for example, Ashford, Corrosion and Degradation of Implant examine the mechanism whereby porosity improves Materials: Second Symposium, ASTM STP 859 (eds the tissue ingrowth prospects of the implant, yet A. C. Fraker & C. D. Griffin), American Society for Testing and Materials, Philadelphia, 1985,373-99. also reduces the adhesion and toughness of the A. Cullison, Welding J., 1987, 2, 51. material system. It is necessary to perform a repro- B. Pourdeyhimi & H. D. Wagner, J. Biomed. Mater. ducible test which simulates the stress and/or Res., 1989,23,63. H. W. Wevers & T. D. V. Cooke, Clin. Mater., 1987, environmental conditions under which the coating 2, 67. is subjected to during its service life. It is only by D. P. Rivero, J. Fox, A. R. Skipor, R. M. Urban & J. 0.Galante, J. Biomed. Mater. Res., 1988, 22, 191. ascertaining such processing/structure/materials M. Jarcho, Clin. Orthoped. Rel. Res., 1981, 157, 259. property relationships that these bioceramic mater- W. Rostoker & C. W. Pretzel, J. Biomed Mater. Res., ials can be further developed. 1974, 8,407. W. Bonfield, Biocompatibility of Tissue Analogs (ed. D. F. Williams), CRC Press, Boca Raton, Florida, 1985, Ch. 5. ACKNOWLEDGEMENTS A. Ravaglioli, A. Krajewski & G. de Portu, Bioceramics: Proceedings of 1st International Bioceramic Sym- The authors wish to thank the CSIROIMonash posium (eds H. Oonishi, H. Aoki & K. Sawai), lshiyaku P University Collaborative Grant Fund that has en- EuroAmerica, Inc., Tokyo, 1989,13-18. G. de With, H. J. A. van Dijk & N. Hattu, Brit. Ceram. abled much of this work to be performed. Mr Gross SOC.Proc., 1981, 31, 181. has been supported by a Monash Research Scholar- P. Sioshansi, Mater. Eng., 1987, 104, 19. P. A. Higham, Mater. Res. Soc. Symp. Proc., 1986, ship. Funds for improvement of the plasma spraying 55, 253. facility at Monash were provided by the Harold P. Ducheyne, W. Van Raemdonck, J. C. Heughebaert Armstrong Memorial Foundation; and we are grate- & M. Heughebaert, Biomaterials, 1986, 7, 97. J. H. Dumbleton & P. Higham, Metal and Ceramic ful for their support. The authors also wish to thank Biomaterials (ed. P. Ducheyne & G. W. Hastings), CRC the reviewer for many pertinent comments. Press, Boca Ratori, Florida, 1984, Ch. 5. R. Marancho, J. Chommidh, G. Constant, B. Moyen, J. J. Comtet, R. Santini & B. Buttazoni, Mater. Sci. Monogr., 1983,17,97. REFERENCES W. R. Lacefield, Bioceramics, Materials Characteriza- tion versus In Vivo Behaviour (eds P. Ducheyne & S. H. Bhaskar, D. E. Cutright, M. J. Knapp, J. D. J. E. Lemons), New York Academy of Sciences, 1988, Beasley, B. Perez & T. D. Driskell, Oral Surg., 1971, 72-80. 31, 282. S. D. Cook, F. S. Georgette, H. B. Skinner & R. J. L. L. Hench & J. Wilson, Mater. Res. Soc. Symp. Proc., Haddad, J. Biomed. Mater. Res., 1984, 18, 497. 1986, 55, 65. Y. Inoe, M. Maeda, H. Iwase, A. Taniguchi & M. Ikenaga, G. F. Howden, Mechanical Properties of Biomaterials Prosthetic implant preparation by coating stainless (eds G. W. Hastings & D. F. Williams), Wiley, New steel with titanium nitride and aluminium oxide, JP York, 1980, Ch. 36. 60,225,569, 9 Nov 1985. D. F. Williams, Biocompatibility of Tissue Analogs (ed. A. Wisbey, P. J. Gregson & M. Tuke, Biomaterials, D. F. Williams), CRC Press, Boca Raton, Florida, 1985, 1987, 8,477. Ch. 5. P. Ruzakowski, J. Wilson, J. Weigand & L. L. Hench, J. W. Boretos, Cer. Bull. 1985, 64, 1098. Transactions of the 7th Annual Meeting of the Society L. L. Hench & E. C. Ethridge, Biomaterials: An for Biomaterials, 1981, 73. InterfacialApproach, Academic Press, New York, 1982, Asahi Optical Co. Ltd., Coating of prosthetic materials Ch. 5. by hydroxyapatite sputtering, JP 58, 109,049, 29 June R. Z. Legeros, J. R. Parsons, G. Daculsi et al., 1983. Bioceramics: Material Characterization Versus In Vivo H.-s Shim & A. D. Haubold, Biomater. Med. Dev., Behaviour (eds P. Ducheyne and J. E. Lemons), The Art. Org., 1980, 8, 333. New York Academy of Sciences, 1988,268-71. N. B. Beak, D. L. McDowell & M. Spector, Int. J. A. Ravaglioli, A. Krajewski & G. de Portu, Proceedings Fatigue, 1987, 9, 211. of 1st lnternational Bioceramic Symposium (eds H. Y. Manabe, Y. Sato, Y. Masuda & Y. Nanba, Multilayered Oonishi, H. Aoki & K. Sawai), lshiyaku Euro America, composites for prosthetic implants, JP 63,147,455, 20 Inc., Tokyo, 1989, 13-18. June 1988. L. L. Hench, Ann. Rev. Mater. Sci., 1974, 5, 279. H. Heide and U. Roth, Proceedings of the lnternational S. D. Brown, Thin Solid Films, 1984, 119, 127. Conference on Hot Isostatic Pressing, Lulea, Sweden, W. R. Lacefield & L. L. Hench, Biomaterials, 1986, 7, 1987, 423-7. 104. J. F. Osborn, Biomed. Tech., 1987,32, 177. K. De Groot, R. Geesink, C. P. A. T. Klein & P. Serekian, H. Herman, Sci. Amer., 1988,259,78. J. Biomed. Mater. Res., 1987, 2l,1375. E. Pfender, Surface Coatings Technol., 1988, 34, 1. A. Krajewski, A. Ravaglioli & G. Perugini, Sci. Ceram. G. Daculsi, N. Passuti, L. Hamel & J. P. Frayet, Bio- 1981, 11, 85. ceramics: Proceedings of 1st lnternationalBioceramic Thermal spraying for bioceramic applications
Symposium, lshiyaku EuroAmerica Inc., Tokyo, 1989, 71. A. Tiselius, S. Hjerten & 0. Levin, Arch. Biochem. 375-81. Biophys., 1956, 65, 132. 47. G. Schwier, Proceedings of the Eleventh lnternational 72. G. Xie, S. Ma & Y. Wang, Huaxie Shijie, 1986, 27, 291. Thermal Spraying Conference (Welding Institute of 73. M. Jarcho, C. H. Bolen, M. B. Thomas, J. F. Bobick, Canada), Pergamon Press, Montreal, 1986, 277-86. J. F. Kay & R. H. Doremus, J. Mater. Sci., 1976, 11, 48. M. R. Dorfman & J. D. Reardon, Proceedings of The 2027. Eleventh lnternational Thermal Spraying Conference, 74. H. Aoki, CaO-P205 Apatite, JP 78,110,999, 28 Pergamon Press, Montreal, 1986, 241-9. September 1978. 49. J. L. Drummond, M. R. Simon, J. L. Woodman & S. 75. K. De Groot, Biocompatibility of Clinical lmplant D. Brown, Biomater. Med. Dev., Art. Org., 1983, 11, Materials (ed D. F. Williams), CRC Press, Boca Raton, 147. Florida, 1984, 199. 50. J. L. Drummond, M. R. Simon & S. D. Brown, Cor- 76. K. Ozaki, Preparation of hydroxyapatiteas a prosthetic rosion & Degradation of Implant Materials: Second material, JP 61,295,215, 26 December 1986. Symposium, ASTM STP 684 (eds B. C. Syrett & A. 77. Y. Nakaso & H. Nakahara Manufacture of hydroxy- Acharya), American Society for Testing and Materials, apatite, JP 61,151,010, 9 July 1986. Philadelphia, 1979, 89-104. 78. K. Hakamazuka, Manufacture of calcium phosphate- 51. J. L. Drummond, Biomaterials in Reconstructive containing hydroxyapatite, JP 62,191,410, 21 August Surgery (ed. L. R. Rubin), C. V. Mosby, St Louis, 1983, 1987. 273. 79. T. Hattori, Y. Iwadate, H. Inai, K. Sato & Y. Imai, Yogyo 52. Joint Replacement Instrumentation Ltd: The Furlong Kyokaishi, 1987, 95, 825. H-A.C Coated THR System, advertisement from J. 80. 0. Suzuki & Y. Matsuda, Continuous manufacture of Bone Joint Surg., 1987,696. hydroxyapatite, JP 63,170,205, 14 July 1988. 53. Fluka catalogue, Chemika-Biochemika, Buschs. 81. H. Tagai & H. Aoki, Bioceramics Symposium, 16 Switzerland, 1988/89. September 1978, University of Keele, UK. 54. P. Ducheyne, L. L. Hench, A. Kagan, M. Martens, A. 82. G. D. Irvine, Synthetic bone ash US 4,274,879,23 June Bursens & J. C. Mulier, J. Biomed. Mater. Res., 1980, 1981. 14, 225. 83. H. Nagai & Y. Nishimura, Hydroxyapatite, and filter 55. K. Shinjo, T. Makiyama, I. Sugiura & K. Kondo, cake, ceramic material and implant material com- Proceedings of 1st lnternational Bioceramic Sym- prising hydroxyapatite, EP 26,090, 1 April 1981. posium (eds H. Oonishi, H. Aoki & K. Sawai), lshiyaku 84. J. F. Conn & L. A. Leofwin, Hydroxyapatite, US EuroAmerica Inc., Tokyo, 1989, 124-9. 4,324,772, 13 April 1982. 56. D. Anderson, G. W. Hastings, S. Morrey & C. Rich, 85. N. Satokami & N. Matsuda, Preparation of prosthetic 2nd International Symposium on Ceramics in Medi- materials from calcium hydroxide and phosphoric cine, Heidelbera- FRG, 10-11 September, 1989, 11 .PP . . acid, JP 61,168,364, 30 July 1986. (preprint). 86. H. Aoki, M. Akao, Y. Shin, T. Tsuzi & T. Togawa, Med. 57. W. Bonfield, J. C. Behiri, C. Doyle, J. Bowman & J. Prog. Through Technol., 1987,12, 213. Abram, Biomaterials and Biomechanics (eds P. 87. R. Z. Legeros, G. Bonel & R. Legros, Calcif. Tissue Duchevne. G. Van de Perre & A. E. Auberti Elsevier Res., 1978,26, 111. scienck, Amsterdam, 1984, 421 -6. 88. H. Newesely, J. Oral Rehab., 1977, 4, 97. 58. Sigma catalogue, Biochemical and organic com- 89. H. Aoki, M. Akao, Y. Shin, Y. Tsuzi & Z. Togawa, Med. pounds for research and diagnostic clinical reagents, Prog. Through Technol., 1987, 12, 213. St Louis, USA, 1988. 90. D. F. Williams, Biocompatibility of Tissue Analogs (ed. 59. T. Hattori, Y. lwadate & T. Kato, J. Mater. Sci. Lett., D. F. Williams), CRC Press, Boca Raton, Florida, 1985, 1989, 8, 305. 43-65. 60. A. Schleede, W. Schmidt & H. Kindt, Z. Elektrochem., 91. H. Tagai & H. Aoki, Mechanical Properties of Bio- 1932, 38, 633. materials leds G. W. Hastinas- & D. F. Williamsl.,. Wilev... 61. G. H. Nancollas, Biological Mineralization (ed. G. H. 1980, 477.' Nancollas), Dahlem Konferenzen, Springer-Verlag, 92. H. Furedi-Milhofer, V. Hlady, F. S. Baker, R. A. Beebe, Berlin, 1982, 79-99. N. W. Wikholm & J. S. Kittelberger, J. Colloid Interface 62. T. Hattori, Y. lwadate & T. Kato, Adv. Ceram. Mater., Scl, 1979,70, 1. 1988, 3, 426. 93. Powder diffraction card 9-432. 63. S. Soyima, K. loku & M. Yoshimura, Mater. Sci. Forum, 94. G. W. Hastinas. D. Daillv & S. Morrev. Bioceramics. 1988, 34-36,371. Proceedings "of 1st lnt&national ~ihcerarnicsym: 64. T. Kanazawa & H. Monma, Kagaku Ryoiki, 1973, 27, posium (eds H. Oonishi, H. Aoki & K. Sawai), lshiyaku 662. EuroAmerica Inc. Tokyo, 1989,355. 65. D. L. Roy & S. K. Linnehan, Nature, 1974, 247, 220. 95. M. Higashikata, Y. Ukegawa, Y. Tsutsumi, H. Aoki & 66. B. Simons, C. Kasperk & R. Ewersk, Hydrothermal M. Akao, abstract from lnternational Meeting on synthesis of hydroxyapatitesuitable for bone implants, Advanced Materials, Tokyo, 1988. DE 3,709897,6 October 1988. 96. W. Van Raemdonck, P. Ducheyne & P. De Meester, 67. A. V. Webster, J. J. Cooper, C. J. Hampson & R. C. Metal and Ceramic Biomaterials (eds P. Ducheyne P. Cubbon, Brit. Ceram. Trans. J., 1987, 91-98, 86. & G. W. Hastings), CRC Press, Boca Raton, Florida, 68. Kobe Steel Ltd., Advanced Ceram. Rep., 1989,4, 3. 1984,143-55. 69. S. lnoue & A. Ono, Yogyo Kyokaishi, 1987, 95, 759. 70. E. Hayek & W. Stadlmann, Angew. Chem., 1955, 67, on-,