Biomaferiak 17 (1996) 1437-1442 @ 1996 Elsevier Science Limited Printed in Great Britain. All rights reserved 0142-9612/96/$15.00
ELSEVIER
Sintering effects in a glass reinforced hydroxyapatite J.C. Knowles*+, S. Talal* and J.D. SantosS** ‘IRC in Biomedical Materials, Queen Mary & Westfield College, Mile End Road, London El 4NS. UK; tDepartment of Biomateriats, Eastman Dental Institute, 256 Grays Inn Road, London WClX 6LD, UK; $Department of Metallurgical Engineering, FEUP, Rua dos Bragas, 4099 Porto Codex, Portugal; ** INEB, National institute for Biomedical Engineering, Praca do Coronel Pacheco 1,400O Porto, Portugal
Glass was incorporated into hydroxyapatite (HA) to act as a sintering aid in order to improve the mechanical properties. The glass was added at 2 and 4 wt% and the powders compacted uniaxially. The compacts were sintered at a variety of temperatures (1200, 1250, 1300 and 1350°C) and the effects of the glass addition and the firing temperature were studied to determine their effects on the sintering process. It was found that the glass promoted some phase changes, related to the amount of glass added to the composite. At 1200 and 125o”C, some p-tricalcium phosphate (TCP) was found with the residual being HA. However, at 1300 and 135O”C, both /?- and cr-TCPwere formed. The 4 wt% addition, as expected, promoted higher levels of secondary phases in the composite. The flexural bend strength (FBS) showed high values and also significant differences dependent on the amount of glass added. A 2 wt% addition gave a gradual increase in FBS with temperature. With a 4 wt% addition there was an increase in FBS, but above 1300°C there was a rapid decrease. These effects were mirrored in the linear shrinkage. Both the linear shrinkage and the FBS results were attributed to both the phase changes occurring and also change in grain size with increasing firing temperature. 0 1996 Elsevier Science Limited Keywords:
Glass reinforced
hydroxyapatite,
Received 29 June 1995; accepted 20 September
(HA) is the main inorganic constituent of bone in humans’ and has the approximate chemical composition CaIo(PC)4)6(OH)22~3. Synthetic HA has found success in hard tissue surgery4* as it is capable of undergoing bonding osteogenesis and is relatively insoluble in vivo. However, its use has been limited to low load bearing applications, by its relatively poor mechanical properties7 compared to other technical ceramics such as alumina’. However, attempts have been made to improve the mechanical properties by optimising the processing9 and sintering regimes” as used with other ceramic systems. A more successful route to improving the mechanical properties has utilised a liquid phase sintering route with great success11-13. Consideration of the densification process for pure HA shows that it follows a classical sintering route. However, significant differences may be found as HA may contain a range of additional ions which may affect the sintering process. At high temperature, the HA decomposes to beta and then the high temperature form, alpha tricalcium phosphate (TCP), and a range of temperatures are quoted14; this is related again to the compositional effects, particularly the Ca/P ratio. The glass reinforced HA, however, exhibits different characteristics in the sintering process. The glass added
aid, grain growth, tricalcium
phosphate
1995
as a sintering aid causes two major effects, which may beneficially affect the mechanical properties. Firstly, dependent on the glass composition, it may cause breakdown of some of the HA to TCP. This may be beneficial because formation of TCP is accompanied by a volume change and thus a reinforcing mechanism, as seen with partially stabilised zirconia, may be responsible for improving the mechanical properties. A second effect is that the glass acts as a grain growth inhibitor, as used with the addition of magnesia to alumina. This effect of grain growth inhibition has already been measured15 and is of considerable significance. This paper documents the sintering behaviour, to try and elucidate the exact role that the above-mentioned phenomena may have on the sintering of a glass reinforced HA. This was evaluated by analysis of the flexural bend strength and linear shrinkage.
Hydroxyapatite
Correspondence Institute.
sintering
MATERIALS AND METHODS Processing of the samples Glass was produced with a composition shown in Table 2. The oxides were added as NaH,PO,, CaC03 and P205. The chemicals were weighed out in the appropriate amounts and blended for 2min in a stomacher blender. The mixed powder was placed in a platinum crucible and this was placed in a Carbolite furnace. Using a heating rate of lo”Cmin-‘, the
to Dr J.C. Knowles at the Eastman Dental
1437
Biomaterials 1996, Vol. 17 No. 14
Glass reinforced
1438 Table 1
Glass composition
in mol%
Oxide
mol%
NaaO CaO
31 24 45
w5
hydroxyapatite:
J.C. Know/es
et al.
Linear shrinkage The linear shrinkage of the specimens was calculated using the equation (I - I,)/&, where I0 is the diameter of the disc before firing and I is the diameter after firing at the appropriate temperature.
RESULTS mixture was heated to 1300°C and held at this temperature for 3 h. The crucible was removed from the furnace and the glass poured onto a steel plate to cool. Once cool, the glass transition temperature (Tg) was determined with a Perkin-Elmer DSC-7 at a heating and cooling rate of 20°C min-‘. HA was obtained from Merck. Glass was added to a porcelain mill pot at either 4 or 8g and then milled for 24 h. Following milling, either 196 or 192 g of HA was added to the mill pot to give glass additions of 2 or 4wt%, respectively. Also, 350ml of methanol was added and the slip was milled for a further 24 h. Following milling, the slip was poured and dried to remove the solvent. Following drying, the powder was pressed using uniaxial pressing; 4g of the powder was placed in a die and the powder was pressed with a pressure of 288MPa. Having produced the test pieces, the discs were fired at 4”CmiK’ to 1200, 1250, 1300 or 1350°C. Having reached the appropriate temperature, they were maintained at that temperature for 1 h, followed by furnace cooling. An air atmosphere was used.
Flexural bend strength The test used was a biaxial flexure test method using a concentric ring set upl” to test the specimens. The inner ring had a diameter of 10mm and the outer support ring had a diameter of 20mm. Either 9 or 10 specimens were tested for each data point on an In&on benchtop test machine, at a crosshead speed of 5 mm min-’ to failure, and from the resultant load displacement graph, the flexural bend strength could be calculated. Young’s Modulus may also be determined from the same load displacement graph.
Grain size measurement
Figure1 shows the heating and cooling curve for the DSC run. The Tg for this glass was calculated as 340°C. A crystallisation peak may also be seen at approximately 460°C: For the phase analysis, for a 2% addition of glass (Figure z), the primary phase is HA with very small amounts of a second phase. At 1200,125O and 13OO”C, the second phase is B-TCP, but at l35O’C it is the high temperature form c(-TCP. With a 4% addition of glass (Figure 3), considerable differences in the phase presence may be seen. At 1200 and 125O”C, the secondary phase is /?-TCP, but at a much higher level compared to the 2% composite, as expected. At 1300 and 135O”C, the composite is three phase, with HA, a-TCP and fl-TCP existing simultaneously and again at a much higher level, compared to the 2% composite. Figure 4 and Figure 5 show the variation of the a axis and c axis unit cell dimensions with temperature, respectively. For the c axis dimension, little statistically significant change in length may be seen. However, the a axis shows a significant reduction with increasing temperature. Also to be noted, initially at 1200”C, a 4% addition of glass gives a larger unit cell dimension compared to a 2% addition, but with increasing temperature, this rapidly drops. No difference in the c axis dimension related to the amount of glass added is noted. Figure 6 shows the flexural bend strength values for the 2 and 4% glass additions and the effect of firing temperature. For the 2% addition, with increasing temperature, there is a steady increase in the flexural bend strength. However, for the 4% addition, the flexural bend strength shows an increase up to 1300”C, but at 1350°C there is a decrease. Linear shrinkage measurements highlight consider-
Scanning electron microscopy was used to analyse the microstructure of the composites. Determination of the grain size was carried out using the linear intercept method after chemical etching of the sample with a 10% citric acid solution for 3-4min at room temperature.
Microstructural
characterisation
For X-ray diffraction (XRD) analysis, the samples were ground to a fine powder. A Siemens D5000 diff?actometer was used with flat plate geometry. Copper radiation was used and detection was with a linear position sensitive detector (PSD) with a nickel filter. Phase identification was performed by comparison with the JCPDS standards. Unit cell dimensions were calculated by measurement of the position of at least 10 peaks. Biomaterids
1996, Vol. 17 No. 14
20.0 10.0 0.0
I Km.0
/
.?m.o
I
I
I
300.0
400.0
MO.0
Tcmpwaturc t?C)
Figure
1
DSC curve
for the glass
used.
Glass reinforced
.r
2-Theta
J.C. Knowles
hydroxyapatite:
- Scale
1439
et al. IRC Biomedical
;I
Materials,
Dr.
J.C.
Knowles
19-Mar-1995
16:56
HFI
w~..“.““‘~“..‘.‘.“.“““.~.“~““~‘~”~”””’~’~’.”~’F
41a
37 34 3s 36 38 39 31 33 29 30 32 28 2x OF 0zCAFZ FIT 1200 (CT: 36.0r, SS:0.014dg, WL: 1.5406AoI A:\XRDA0359.RGW XRDA0359 n:\XRDA0360.RAW XRDAQ3bEI 22 OF 0zCfiFZ FIT 1250 (CT: 36.0s) SS:0.014dg, WL: 1 .S406Ao 1 fi::XRDfl0363.RAI-I XRDA'J363 2~ OF O:
XRD traces for hydroxyapatite
2-Theta :
I-----
(HA) reinforced
with 2% of glass.
IRC Biomedical
- Scale
’ ’ ”
Materials,
Dr.
’ ”
HA
”
J.C.
Knowles
”
19-Mar-1995
18
’
w~...i”..““““““...“...“.‘.““.“‘..”””....’...”.”
28 A:\XRDA036S,RAW XRDA0365 A :\XRDA0366.RAW XRDA0366 A:\XRl'PAB367,RAlJ XRDQ0%,7 C:\USERDATA\XRDAC138S.HUW Flgure 3
36 37 38 39 3s 29 31 32 33 34 30 42 OF 0zCAFZ TO 1200 (CT: 36.0~s SS:0.014dg, WL: 1.5406Ao) 42 OF 0zCAFZ TO 1250 (CT: 36.0s, SS:Q.014dg, WL: 1.5406Ao) 42 OF DzCAFt T@ 1300 (CT: 36.0sr SS:0.014dg < WI.: l.SrlBb~io~ XRDA0385 42 OF 0zCAF2 TO 1350 (CT: 36.0~s SS:Q.O14dg, WL: 1.5406Ao)
XRD traces for HA reinforced
with 4% of glass. Biomaterials1996, Vol. 17 No. 14
Glass reinforced hydroxyapatite: J.C. Knowles et al.
1440
9.425
f--
2% Glass Addition 4% Glass Addition
&
rm
9.415
~ A
18r
9.42 1
2% Mass Addition 4% Glass Addition Unreinforced HA
1
9.41 z .S 9.405 1L ki m 9.4
I J
9.395
/
/
I
9.39 L 9.385 1 1 I
1 Id_,
1150
1200
I
I .I I 1 / I
1250 1300 Temperature (‘C)
1350
Figure 4 Variation of HA a axis unit cell dimensions with firing temperature. + ---
6.905
4 t 1150
1400
Figure 7 ture.
1200
1300
1350
1400
Variation of linear shrinkage with firing tempera-
2% Glass Addition 4% Glass Addition -+-
F 6.901 i
1250
Temperature (“C)
~ +
5r
2% Glass Addition 4% Glass Addition Unreinforced HA
4 i6.8981
/:i\-
I
6.894 I-
6.89 L-
1
IL
1
t
L.LLLi_~
1200
1150
L-_~_..I~
1250
3 2 -
_L.J -1
1300
l1350
1400
Temperature(“C)
a
0
Figure 5 Variation of HA c axis unit cell dimensions with firing temperature. Figure 6 -f-
-
130F
-+-
2% Glass Addition 4% Glass Addition Unreinforced HA T
110 90 70 50 30 10 1 1150
1200
1250 1300 Temperature (“C)
1350
1400
Figure 6 Variation of flexural bend strength with firing temperature.
L
1
1150
L,
I
1200
I
/
1250
I
I1
1300
Temperature (“C)
,LLd
1350
1400
Variation of grain size with firing temperature.
sintering ceramics. The shrinkage reached a maximum at 1300°C and remained at this level at 1350°C. When a 4% glass addition is used, considerable differences are found. The compacts continue to shrink up to 1300°C as expected, but at 1350°C the samples show an expansion and this will be explained in the Discussion. Figure 8 shows the change in grain size with temperature. There are two parts to these curves, the first from 1200 to 1300°C and the second from 1300 to 1350°C. From 1200 to 13OO”C, the grain size increases as expected, with little difference seen related to the glass addition. However, above 13OO”C, the 4% glass addition can be seen to give significant grain growth at the higher sintering temperature. The 2% addition at 1350°C also seems to show an increase in grain growth denoted by the change in slope.
DISCUSSION able differences in the sintering profile, dependent on the glass percentage addition (Figure 7). When a 2% addition is used, the linear shrinkage increases with increasing temperature, as would be expected when Biomaterials 1996, Vol. 17 No. 14
The mechanical properties obtained are significantly higher than those obtained for unreinforced HA and the improvement in properties may be accounted for
Glass reinforced
hydroxyapatite:
J.C. Knowles et al.
by a number of changes occurring simultaneously. Two primary subdivisions may be made, classifying effects as having either a physical origin or a chemical origin, There are effects occurring which are classical sintering effects seen in ceramics, but there are also other effects, such as phase changes occurring and also grain growth inhibition, which significantly contribute to the enhanced mechanical properties. The Tg value for the glass used is considerably below lOOO”C, the temperature above which significant increases in the mec:hanical properties are achieved; hence, the glass must be in very low viscosity form. The first two parameters to be discussed are the phase changes that are occurring. With a 2wt% addition, virtually no secondary phase may be seen and this is unexpectl2d as previous work has always seen the appearance of a secondary phase due to glass additions. This result is encouraging as it allows us to produce a reinforced monophasic bioceramic. As expected, with a larger glass addition, more of the secondary phases occur and these are c(- and /?-TCP, as reported previously13. Also calculated from the XRD measurements are the unit cell dimensions. The c axis shows no significant change with temperature, as reported previously”. The glass does, however, cause a significant change in the Q axis. The unit cell is larger, to incorporate some of the ions from the glass, but at 1200”C, the 4% addition gives an eve.n greater increase in the Q axis at 1200°C. Sintering at higher temperatures negates this effect. The flexural bend strength should be viewed whilst considering the three dominant effects: phase changes, linear shrinkage and grain size. The glass is very reactive at high temperatures and so usually the glass causes breakdown of some of the HA to form secondary phases. This was proposed as a mechanism for reinforcement of the composite. However, for a 2 wt% addition of glass, little if any secondary phase is produced and so another mechanism must be responsible. Production of secondary phases will have an effect on linear shrinkage. Both the beta and alpha TCP phases would give a volume increase, but for the 2% addition, because no secondary phases are present, classical sintering eflects only are present. With the 4% glass addition, secondary phases appear and this effect may be seen in the linear shrinkage. From 1200 to 1300”C, there is a decrease in size, but at 1350”C, there is an increase which may be attributed to the secondary phases at the high temperatures, particularly the formation of alpha TCP. The grain size measurements should be compared with values for HA without any reinforcing phase. At firing temperatures of 1200 and 125O”C, the glass reinforced HA at both additions shows average diameters, higher than for unreinforced HA. However, at l300”C, the glass reinforced HA showed slightly smaller grain sizes compared to the unreinforced HA. This changes again at 1350”C, where the 2% addition of glass shows a grain size similar to unreinforced HA, but significantly, the 4% addition now gives grain sizes much higher than the unreinforced HA. These results correlate with the values found for the mechanical properties, where the 4% addition shows a
1441
significant decrease in flexural bend strength at 1350°C. These results are interesting in that in previous work, using HA from a different source, significant grain growth was seenI’. The HA used in this study appears to be very stable and this has been related to the purity of the precursor powder. CONCLUSIONS The grain size measurements show that the glass significantly inhibits grain growth and thus the mechanical properties are enhanced by maintaining a small grain size. This, however, is strongly linked with both the level of glass addition and also with the firing temperature; 1300°C appears to be an upper limit for firing temperatures for this system. The linear shrinkage also indicates that the glass is acting as a sintering aid and helping to eliminate porosity, thus also contributing to the increase in mechanical properties. ACKNOWLEDGEMENTS The continued support of the EPSRC for the IRC programme grant in Biomedical Materials is gratefully acknowledged. The authors also gratefully acknowledge the financial support of the JNICT PBIC/ C/CTM/1890/95 project. REFERENCES Currey JD. The Mechanical Adaptations of Bones. Princeton University Press, 1984. Brown WE, Chow L. Chemical composition of bone mineral. Ann Res Mater Sci 1976; 6: 213-226. Glimcher MJ. Recent studies of the mineral phase in bone and its possible linkage to the organic matrix by protein bound phosphate conds. Phil Trans R Sot Lond 1984; B304: 479-508. Cook SD, Thomas KA, Kay JF, Jarcho M. Hydroxyapatite-coated porous titanium for use as an orthopaedic biologic attachment system. Clin Orthop 1980; 230: 303-312.
Patka P, den Otter G, de Groot K, Driessen AA. Reconstruction of large bone defects with calcium phosphate ceramics. An experimental study. Neti I Surg 1985; 37: 38. Wolford LM, Wardrop RW, Hartog JM. Coralline porous hydroxyapatite as a bone graft substitute in orthognathic surgery. J Oml Maxillofac Surg 1987; 45: 10341042.
Santos JD, Morrey S, Hastings GW, Monteiro FJ. The production and characterisation of a hydroxyapatite ceramic material. In: Bonfield W, Hastings GW, Tanner KE, eds. Bioceramics, Vo14. Proceedings of the 4th International Symposium on Ceramics in Medicine. London: Butterworth Heinemann, 1991: 71-78. DGrre E, Dawihl W. Ceramic hip endoprostheses. In: Hastings GW, Williams DF, eds. Mechanical Properties of Biomaterials. New York: Wiley, 1980: 113-127. Murray MGS, Ponton CB, Marquis PM. Emulsion processing of hydroxyapatite. Bioceramics, Vol 4. Proceedings of the 4th International Symposium on Ceramics in Medicine. Kyoto: Kobunshi Kankokai, 1992: 15-22. Biomaterials 1996,
Vol. 17 No. 14
1442 10
11
Glass reinforced Puajindanetr S, Best SM, Bonfield W. Characterisation and sintering of precipitated hydroxyapatite. Br Cemm Trans 1994; 93: no.3, 96-99. Knowles JC, Bonfield W. Development of a glass reinforced hydroxyapatite with enhanced mechanical properties. The effect of glass composition on mechanical properties and its relationship to phase changes. I
14 15
Biomed Mater Res 1993; 27: 159-1598. 12
13
Santos JD, Knowles JC, Reis RL, Monteiro FJ, Hastings characterisation of glass GW. Microstructural reinforced hydroxyapatite composites. Biomaterials 1994; 15: no.1, 5-10. Knowles JC. Development of hydroxyapatite with enhanced mechanical properties: effect of high glass additions on mechanical properties and phase stability of sintered hydroxyapatite. Br Ceram Tmns 1994; 93: no.3, 100-103.
Biomaterials 1996, Vol.
17 No. 14
16 17
18
hydroxyapatite:
J.C. Knowles et al.
Ruys AJ, Wei M, Sorrel1 CC, Dickson MR, Brandwood A, Milthorpe BK. Sintering effects on the strength of hydroxyapatite. Biomaterials 1995; 16: 409-415. Santos JD, Reis RL, Monteiro FJ, Knowles JC, Hastings GW. Liquid phase sintering of hydroxyapatite by phosphate and silicate glass additions. Structure and properties of the composites. 1 Mater Sci Mater Med 1995: to be published. Soltesz U, Ritcher H, Keinzler R. In: Vincenzini P, ed. High Tech Cemmics. Amsterdam: Elsevier, 1987: 110. Laugier MT. New formula for indentation toughness in ceramics. I Mater Sci L&t 1987; 6: 355-356. Knowles JC, Abrahams I, Bonfield W. Effect of reinforcing glass composition on phase transformation and crystallographic parameters in hydroxyapatite. In: Ducheyne P, Christiansen D, eds. Riocemmics, Vol 6. Butterworth-Heinemann, 1993: 191-196.