Glass reinforced hydroxyapatite for hard tissue surgery—Part 1: mechanical properties

Glass reinforced hydroxyapatite for hard tissue surgery—Part 1: mechanical properties

Biomaterials 22 (2001) 2811}2815 Glass reinforced hydroxyapatite for hard tissue surgery*Part 1: mechanical properties G. Georgiou, J.C. Knowles* Dep...

127KB Sizes 0 Downloads 127 Views

Biomaterials 22 (2001) 2811}2815

Glass reinforced hydroxyapatite for hard tissue surgery*Part 1: mechanical properties G. Georgiou, J.C. Knowles* Department of Biomaterials, Eastman Dental Institute, University College London, 256 Gray+s Inn Road, London WC1X 8LD, UK Received 31 August 2000; accepted 9 January 2001

Abstract Commercial hydroxyapatite (HA) was reinforced by adding 2.5 and 5 wt% of a Na O}CaO}P O glass and then sintered. The    resulting composites have chemical compositions that are similar to the inorganic constituent of the mineral part of bone, and are closely related to the trace elements that are present, in this case Na. X-ray di!raction showed no decomposition of HA to secondary phases; however, the glass reinforced-HA composites contained a HA phase and variable amounts of tricalcium phosphate phase, depending on the sintering temperature and the amount of glass added. The HA-composite material exhibited higher #exural strength overall compared to sintered HA. The presence of secondary phases - and -tricalcium phosphate in the microstructure of the composites has a major in#uence on the mechanical properties. Additionally, the presence of porosity also has a bearing on the mechanical properties of the material.  2001 Elsevier Science Ltd. All rights reserved. Keywords: Hydroxyapatite; Decomposition; Tricalcium phosphate; Mechanical properties

1. Introduction Hydroxyapatite (HA), at present, is of great interest as a material for surgical implants, due to its biocompatibility with living tissue. Connective tissue, which holds together the di!erent structures in the body, is inherent in bone and contains collagen "bres, mineral part and ground substance. The exact composition and relative proportions of collagen depend upon numerous factors, including the location and loading requirements of bone [1}3]. Structural and chemical analyses of the inorganic part of bone have shown that ionic substitutions may occur within the HA lattice [2]. Substitutions include CO\ for OH\ or PO\, Mg> and Na> for Ca>, and   F\ for PO\ [2]. Trace ions, particularly Na>, Mg>,  K> and F\ are well known to be prevalent in the inorganic part of bone [1]. In addition, the incorporation of a CaO}P O glass as a sintering aid facilitates the de  composition of HA to -tricalcium phosphate (TCP). -TCP can undergo ionic substitutions, which involve the substitution of Ca> ions for ions such as Mg> in

* Corresponding author. Tel.: #44-1207-915-1189; fax: #44-1207915-1189. E-mail address: [email protected] (J.C. Knowles).

the lattice structure. Due to the smaller ionic radius of Mg> relative to Ca>, the lattice experiences a decrease in the lattice parameters, thus increasing the lattice enthalpy and hence the stability of the -TCP phase [4]. Thus, it is important to incorporate these trace elements in orthopaedic implants, such as HA, because the biocompatibility of apatites is closely dependent on their composition [5]. The incorporation of a P O }CaO   glass system with addition of Na, Mg and K oxides allows these substitutions within HA to occur, giving nominal elemental compositions of trace elements in the composites that are analogous to that of the inorganic constituent of the mineral part of bone [6]. Therefore, the glasses within this system are considered to have enormous potential as biomaterials. Synthetic HA is limited in its use as a biomaterial, primarily due to its low load bearing capacity which can be illustrated by the relatively poor mechanical properties compared with bone [6}11]. However, improvements on its mechanical properties have been undertaken with the inclusion of soluble phosphate based glasses (P O }CaO}Na O) [6}11]. Phosphate glasses, when in   corporated into HA, melt at a lower temperature compared to HA and can act to increase density by enhancing the sintering mechanisms which greatly enhances the mechanical properties. Furthermore,

0142-9612/01/$ - see front matter  2001 Elsevier Science Ltd. All rights reserved. PII: S 0 1 4 2 - 9 6 1 2 ( 0 1 ) 0 0 0 2 5 - 4


G. Georgiou, J.C. Knowles / Biomaterials 22 (2001) 2811}2815

decomposition of HA into secondary phases, - and -TCP, can occur. This is related to the compositional e!ects, in particular the Ca/P ratio [7]. It can also be seen as a function of sintering temperature [6}11]. Another physical property observed to change is also a signi"cant decrease in the porosity of the "nal glass reinforced HA (GR-HA) from 12003C to 12503C. This paper will investigate the e!ect of phases on the mechanical properties along with the porosity and signi"cance of each of these in relation to the mechanical properties.

2.4. X-ray diwraction Tested samples were ground to a "ne powder, placed in a specimen holder and then analysed on a Philips PW1780 di!ractometer with Ni "ltered Cu K radiation  (K  "1.5406 As , K  "1.5444 As ) at 40 kV and 30 mA. ? ? The data were collected with a scintillation counter between 10(2(90 with a step size of 0.023 and a count time of 12 s using #at plate geometry. 2.5. Structure rexnement

2. Materials and methods 2.1. Specimen preparation The glass composite was produced with a composition of 32 mol% CaO, 23 mol% Na O and 45 mol% P O .    The starting reagents used to prepare the glass consisted of NaHPO , P O and CaCO . These were mixed thor    oughly and placed in a platinum crucible, melted at 10003C for 1 h, and then poured onto a steel plate and allowed to cool. The resulting glass was then ground to a "ne powder using an agate grinder. The glass (5 or 10 g) powder was placed into a porcelain mill pot and milled dry for 24 h. HA (supplied by Plasma Biotal Ltd. UK) was then added to the mill pot at either 195 or 190 g to give glass additions of 2.5% and 5% glass additions, respectively. Methanol (300 ml) was also added and the mixture was then wet milled for a further 24 h. The resulting slip was then dried at 703C and the dry powder was then sieved to 75 m. Four grams of powder was placed into a steel die and uniaxially pressed at 20 tons using a hydraulic press to give 30 mm discs. The discs were "red at a heating rate of 43C min\ to 12003C, 12503C, 13003C and 13503C. The specimens were held at the appropriate temperature for 1 h and then furnace cooled. 2.2. Measured density determination Density measurements were performed on each sintered specimen using the Archimedes principle. The measurements were carried out in distilled water. These were then compared with the theoretical densities calculated from the Rietveld analysis to give the porosity.

The structure re"nement was carried out using General Structure Analysis Software (GSAS, Los Alamos National Laboratory). A standard model for each of the three phases was used for the re"nement of the samples and they were determined from the Daresbury Crystal Structure Database. The HA model was based on the single crystal structure determination, with P63/m space group and lattice parameters of a"9.45 As and c"6.88 As . The second phase -TCP consists of a R3CH space group and lattice parameters of approximately a"10.4 As and c"37.4 As . The standard model used to re"ne -TCP was based on the P21/a space group and unit cell dimensions of a"12.887 As , b"27.28 As , c"15.219 As , and "126.23. The peak shapes for the re"nement were modelled on a pseudo Voigt distribution. Background parameters, a scale factor, four peak shape variables, an asymmetry factor, cell parameters, and a zero point correction were all re"ned. Isotropic thermal parameters and atom positions were re"ned for all HA atoms. For -TCP and -TCP all the parameters were re"ned; however, the thermal parameters and atom positions were "xed. From the data output from the GSAS list "le, a number of other parameters are calculated and these data are presented. The software calculates the phase weight percentage and the theoretical density (assuming no porosity). The calculation for these numbers does not include any error determination. From the data for the measured density and the theoretical density, the porosity is calculated using recorded density/theoretical density;100 (%). This again does not have an associated error with it.

2.3. Mechanical testing

3. Results

The test carried out was the #exural bend strength test method using a concentric ring testing jig with a loading ring of 10 mm and outer supporting ring of 20 mm. Ten specimens were tested for each of the "ring temperatures on an Instron machine at a crosshead speed of 5 mm min\ to failure. The #exural bend strength was determined from the load displacement graph.

Fig. 1 shows the variation of #exural bend strength with change in sintering temperature for all three materials. For HA there is a slight increase in the mean #exural bend strength, which decreases at 13503C; however, this is not statistically signi"cant. The GR-HA with 2.5 wt% glass (HA2.5) addition exhibits a gradual increase in #exural bend strength with

G. Georgiou, J.C. Knowles / Biomaterials 22 (2001) 2811}2815


Fig. 1. E!ect of "ring temperature on #exural bend strength for HA, HA2.5 and HA5.

Fig. 3. E!ect of "ring temperature on phase weight percentage for HA5.

increasing "ring temperature. For the GR-HA with 5 wt% glass (HA5), the #exural bend strength increases to approximately 80 MPa and then decreases again at 13003C and 13503C to around 50 MPa. Using X-ray di!raction and Rietveld analysis, quantitative phase analysis was performed. The HA showed no sign of decomposition at any "ring temperature. For HA2.5 (Fig. 2) the HA levels stay relatively constant at about 85% for all "ring temperatures. At 13503C, it is clear that the -TCP inverts to -TCP. Fig. 3 shows the phase element fractions for HA5. Between 12003C and 12503C, there is no -TCP and the -TCP levels stay relatively constant at around 35%. However at 13003C, there is further decomposition of HA, and the -TCP inverts to -TCP and this continues at 13503C. The presence or absence of porosity in ceramics signi"cantly a!ects the mechanical properties and so the dens-

ities were measured and porosity calculated. For the measured densities (Fig. 4), HA shows a rapid increase in density between 12003C and 12503C and remains relatively constant at about 3.13 g cm\ with further increase in "ring temperature. HA2.5 exhibits a similar trend to HA, but the density values are a little lower. For HA5 the density rapidly increases between 12003C and 12503C and then decreases again with increasing "ring temperature. This will be explained. From the Rietveld analysis, a theoretical density may be calculated from the atom positions, taking into account the phases. The theoretical density assumes no porosity. These data are shown in Fig. 5. For HA, the values are slightly below the maximum theoretical density of 3.16 g cm\ and remain constant for all "ring temperatures. These values di!er from the theoretical density as the data has been re"ned for each particular sample and does not assume optimal atom positions and

Fig. 2. E!ect of "ring temperature on phase weight percentage for HA2.5.


G. Georgiou, J.C. Knowles / Biomaterials 22 (2001) 2811}2815

Fig. 4. E!ect of "ring temperature on measured density for HA, HA2.5 and HA5.

Fig. 5. E!ect of "ring temperature on theoretical densities (determined from Rietveld re"nement) for HA, HA2.5 and HA5.

all OH sites are fully hydroxylated. For HA2.5 there is a reduction in theoretical density with "ring temperature, but it is not so marked except at 13503C. Also the overall values are lower than the values for HA. HA5 shows a signi"cant decrease in density with increasing "ring temperature. Also all the theoretical values are signi"cantly lower than the values for both HA and HA2.5. The di!erence between the values for the measured and theoretical densities gives the porosity and this is shown in Fig. 6. For all materials, there is a sharp decrease in the amount of porosity between 12003C and 12503C. For HA and HA5 the porosity then remains fairly constant. At 13003C all the samples show similar levels of porosity. However, HA2.5 continues to show a reduction in porosity levels, so at 13503C the level of porosity approaches zero.

4. Discussion As previously mentioned, the presence or absence of porosity in ceramics signi"cantly a!ects the mechanical properties. The incorporation of glass in HA acts as a sintering aid, which enhances densi"cation, and as

Fig. 6. E!ect of "ring temperature on porosity for HA, HA2.5 and HA5.

a result reduces porosity. This occurs through a liquid phase sintering mechanism, which accelerates inter-particulate di!usion and bonding, thus causing elimination of porosity and shrinkage. However, in this system, it is further complicated by the occurrence of decomposition and phase changes that can also have a detrimental or bene"cial e!ect on the mechanical properties. HA, HA2.5 and HA5 exhibited a sharp decrease in porosity as temperature increases and these gave values of approximately 0}1% at 12503C, 13003C and 13503C. One would arrive at the assumption that since the porosity values for the specimens mentioned are comparable, then one would therefore expect similar #exural bend strength values. This is not seen however. What we observe is an improvement for glass reinforced HA relative to HA, and in particular for HA2.5. The phase element analysis from the Rietveld re"nement carried out for these specimens gives an indication of the amount of each of the phases present. For HA2.5 and HA5 there is decomposition of HA to -TCP at all four given temperatures, and furthermore, inversion of -TCP to -TCP. With HA5 signi"cantly more -TCP is found and the -TCP phase inverts more readily to -TCP at a lower temperature. This inversion is due to the origin of -TCP [7]. The decomposition of HA is attributed to the presence of a reactive glass, which can drive o! the hydroxyl groups when entering the HA structure and as a result causes decomposition. The glass may also a!ect the Ca : P ratio [7]. However, for HA5 the inversion of -TCP to -TCP, occurring at a lower temperature, is due to -TCP containing more residual ions, which destabilise -TCP and facilitates the formation of more -TCP [7]. The presence of these secondary phases has a major in#uence on the mechanical properties of the material. This can be correlated with values illustrated for #exural bend strength. The decrease in #exural strength for HA5 at 13003C and 13503C correlate with the phase weight fractions, where the -TCP phase content appears at around 15% and 19%, respectively. Alternatively, HA2.5

G. Georgiou, J.C. Knowles / Biomaterials 22 (2001) 2811}2815

displays a gradual increase as temperature increases. At 13503C, where -TCP is present at 5%, there does not appear to be a reduction in #exural strength as was observed with HA5. There is also probably a positive in#uence from the virtual complete absence of porosity. The presence of -TCP and -TCP phases in the HA structure would give a volume increase and thus can give rise to residual stress [6}8,10] . However, the presence of -TCP at high levels can disrupt the integrity of the GR-HA [6}8,10]. As a result of these "ndings, one can postulate that a signi"cant amount of -TCP in the structure can be detrimental to the mechanical properties of the material, but low amounts, as observed in HA2.5, can be advantageous. Finally, both porosity and phase decomposition are in#uential factors in the outcome of the mechanical properties. However, what the results of this study seem to suggest is an overall improvement in #exural strength for glass reinforced HA which is associated primarily with phase decomposition and subsequent phase inversions.

5. Conclusions HA undergoes signi"cant elimination of porosity during sintering, particularly between 12003C and 12503C. However, the mechanical properties are still relatively poor. The inclusion of a glass as a sintering aid signi"cantly improves the #exural bending strength. The majority of the improvement comes from the decomposition of HA to -TCP with an associated volume change, helping to improve the mechanical properties. At higher "ring temperatures, the -TCP inverts to the high temperature -TCP form. A relatively small amount of -TCP can be sustained in the composites without detriment. However, large percentages of glass can promote formation of signi"cant amounts of -TCP with a resultant decrease in #exural bend strength. The glass can also help to eliminate porosity from the structure and this may have contributed to the improvement in mechanical properties in HA2.5.


Acknowledgements The authors would like to acknowledge the support of the Engineering and Physical Sciences Research Council for "nancial support. References [1] Aoki H. Science and medical applications of hydroxyapatite. Tokyo: Takayama Press System Centre, 1991. [2] Evans G, Behiri J, Currey J, Bon"eld W. Microhardness and Young's modulus in cortical bone exhibiting a wide-range of mineral volume fractions and in a bone analog. J Mater Sci*Mater Med 1990;1(1):38}43. [3] Knowles JC. Development of a glass-reinforced hydroxyapatite with enhanced mechanical properties*the e!ect of glass composition on mechanical properties and its relationship to phase changes. J Biomed Mater Res 1993;27(12):1591}8. [4] Knowles JC. Development of hydroxyapatite with enhanced mechanical properties*e!ect of high glass additions on mechanical properties and phase stability of sintered hydroxyapatite. Br Ceram Trans 1994;93(3):100}3. [5] Knowles JC, Talal S, Santos JD. Sintering e!ects in a glass reinforced hydroxyapatite. Biomaterials 1996;17(14):1437}42. [6] LeGeros RZ, LeGeros JP. Dense hydroxyapatite. In: Hench LL, Wilson J, editors. An introduction to bioceramics. World Scienti"c: Singapore. 1993. p. 139}80. [7] Lopes MA, Santos JD, Monteiro FJ, Knowles JC. Glass-reinforced hydroxyapatite: a comprehensive study of the e!ect of glass composition on the crystallography of the composite. Biomed Mater Res 1998;39(2):244}51. [8] Rey C, Freche M, Heughebaert M, Heughebaert JC, Lacout JL, Vignoles M. Apatite chemistry in biomaterial preparation, shaping and biological behaviour. In: Bon"eld W, Hastings GW, Tanner KE, editors. Bioceramics, vol. 4. London: Butterworth Heinemann, 1991. [9] Rey C. Calcium phosphate biomaterials and bone mineral. Di!erence in composition, structure and properties. Biomaterials 2000;11:13}5. [10] Santos JD, Monteiro FJ, Knowles JC. Liquid-phase sintering of hydroxyapatite by phosphate and silicate glass additions*structure and properties of the composites. J Mater Sci*Mater Med 1995;6(6):348}52. [11] Santos JD, Silva PL, Knowles JC, Talal S, Monteiro FJ. Reinforcement of hydroxyapatite by adding P O }CaO glasses   with Na O, K O and MgO. J Mater Sci*Mater Med   1996;7(3):187}9.