of 7

Clarifying the effect of sintering conditions on the microstructure and mechanical properties of β-tricalcium phosphate

All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Clarifying the effect of sintering conditions on the microstructure and mechanical properties of β-tricalcium phosphate
  Clarifying the effect of sintering conditions on the microstructure andmechanical properties of   b -tricalcium phosphate Fidel H. Perera, Francisco J. Martı´nez-Va´zquez, Pedro Miranda*, Angel L. Ortiz, Antonia Pajares  Departamento de Ingenierı´ a Meca´ nica, Energe´ tica y de los Materiales, Universidad de Extremadura. Avda de Elvas s/n. 06071 Badajoz, Spain Received 22 February 2010; received in revised form 3 March 2010; accepted 22 March 2010Available online 21 May 2010 Abstract Theeffectofthesinteringconditions(temperatureandtime)onthemicrostructure(densityandgrainsize)andmechanicalproperties(hardness,elastic modulus, and strength) of   b -tricalcium phosphate ( b -TCP) bioceramics fabricated from Ca-deficient commercial powders is analyzed.Contrary to current general opinion, it is demonstrated that the optimal sintering temperature to maximize the mechanical performance of this  b -TCP material is not necessarily below the  b $ a  transformation temperature (1125  8 C). In particular, optimal performance was achieved insamplessinteredat1200  8 Cfor3 h,sinceitwasnotuntilhighertemperaturesorlongersinteringtimesthatmicrocrackingdevelopsandmechanicalpropertiesaredegraded.Itisarguedthattheresidualstressesdevelopedduringthisreversibletransformationdonotleadtomicrocrackpropagationuntil sufficiently large starting flaws develop in the microstructure as a consequence of grain growth. Implications of these findings for theprocessing routes to improve sintering of this important bioceramic are discussed. # 2010 Elsevier Ltd and Techna Group S.r.l. All rights reserved. Keywords:  A. Sintering; C. Mechanical properties; Tricalcium phosphate;  b -TCP 1. Introduction Calcium phosphate materials are widely used in biomedicalapplications due to their osteoconductive and osteoinductiveproperties. In this context, hydroxyapatite (HA,Ca 5 [OH(PO 4 ) 3 ]), the main mineral phase of bone, is an obviousfirst choice for orthopaedic implants. Nevertheless, its lowbiodegradabilityisamajorhandicapforthosetissueengineeringapplications where the implanted material is intended to beresorbed and replaced by living tissue. Tricalcium phosphate(TCP, Ca 3 (PO 4 ) 2 ), on the contrary, is progressively resorbed  invivo byacell-mediatedprocessandreplacedbynewbone[1–3],and is thus a more suitable candidate for this particular type of application. As shown in the phase diagram of  Fig. 1, TCP cancrystallizeinthreeallotropicforms: b -TCP,whichisstablefromroom-temperature up to 1125  8 C,  a -TCP, which is stable from1125  8 Cupto1470  8 C,and a 0 -TCPwhichisstablefrom1470  8 Cup to the melting point at 1810  8 C [4]. Of the three variants, b -TCP is the most desirable in terms of chemical stability andbioresorption rate –  a -TCP hydrolyses partially or completelyinto hydroxyapatite in a biological environment losing itsresorbability [5], and  a 0 -TCP cannot be stabilized at roomtemperature [4]. b -TCP is routinely used as bone replacement, especially inthe field of oral and craniofacial surgery, in the form of granules and rods [5,6] or as filler in polymeric scaffolds [7]. In bulk,  b -TCP bioceramics have mechanical properties toopoor to be used in load-bearing clinical applications [8–10],which has been attributed to the difficulties in fully densifying b -TCP powders [9,11–13]. These difficulties are associatedwith the presumption that the sintering temperature should bekept below 1125  8 C to avoid the  b $ a  phase transformationthat is considered deleterious to mechanical properties. This isattributed to spontaneous massive microcracking in thesintered body due to the expansion-contraction cyclegenerated by the differences in density between  b -TCP(3.07 g/cm 3 ) and  a -TCP (2.86 g/cm 3 ) [10,11,14]. While thisis accepted as a fact within the bioceramic community, to thebest of the authors’ knowledge no unambiguous proof thatmicrocracking inherently follows the  b $ a  transformationexists in the literature. www.elsevier.com/locate/ceramint  Available online at www.sciencedirect.com Ceramics International 36 (2010) 1929–1935* Corresponding author. Tel.: +34 924 28 96 00x86735;fax: +34 924 28 96 01. E-mail address:  pmiranda@unex.es (P. Miranda).0272-8842/$36.00 # 2010 Elsevier Ltd and Techna Group S.r.l. All rights reserved.doi:10.1016/j.ceramint.2010.03.015  With this in mind, the present study was conducted with theobjective of investigating in more detail the effect of thesintering conditions on the microstructure and mechanicalproperties of   b -TCP bioceramics, to clarify the role of the b $ a  transformation on their mechanical degradation. Theresults challenge the accepted notion that sintering of thesematerials has to be carried out below the transformationtemperature in order to obtain a good mechanical response.Based on these results, processing guidelines for fabricating b -TCP bioceramics with superior mechanical properties areproposed. 2. Experimental procedure The  b -TCP starting powder used in this study was obtainedfrom a commercial source (Fluka, Buchs, Switzerland). This b -TCP powder is Ca-deficient (Ca/P molar ratio of 1.37  0.03),and consists of micrometer particles with an average size of 1.8  0.8  m m. The CaO molar concentration corresponding tothis particular Ca/P ratio is marked in the phase diagram of Fig. 1 as a vertical light-gray line. The choice of a Ca-deficient(Ca/P < 1.5) powder is because the slow kinetics of the reverse a ! b  transformation in Ca-rich powders does not ensure thereversibility of the b $ a transformation [4,9]. Compacts weremade by uniaxial pressing (C, Carver, Inc., Wabash, IN, USA)at 32 MPa, followed by isostatic pressing (CP360, AIP,Columbus, OH, USA) at 180 MPa. Pressureless sinteringwas performed in a conventional furnace (C.H.E.S.A, Madrid,Spain) under the following conditions: temperatures of 1100,1200, and 1300  8 C, times of 1, 3, 5, and 7 h, air atmosphere andheating and cooling rates of 10  8 C/min. The sinteringtemperatures are indicated as circles on the composition lineof  Fig. 1. Surfaces for microstructural and mechanicalcharacterizations were diamond-polished to 1  m m finish.The starting powders and sintered samples were character-ized by X-ray diffractometry (XRD, PW-1800, PhillipsResearch, The Netherlands) using Cu-K  a  radiation and theirphase composition was determined by the Rietveld method.Bulk samples were imaged using scanning electron microscopy(SEM, S-3600N, Hitachi, Japan) to analyze their microstruc-tural evolution. Grain sizes were determined by routine imageanalysis (AnalySIS 1 , Olympus Soft Imaging Solutions GmbH,Germany) using random images totaling a minimum of 500grains. In addition, the density,  r , of the sintered samples wasmeasured using the Archimedes method.Berkovich indentation tests were performed on polishedsurfaces to evaluate the hardness and Young’s modulus using adepth-sensing indentation instrument (NanoTest, Micro Mate-rials, Wrexham, UK). For each sintering condition a total of 20tests were conducted under ambient conditions at 10 N, withindentation load rate of 400 mN/s and dwell time of 10 s.Hardness and Young’s modulus were evaluated from theindentation load-displacement curves using the Oliver andPharr method [15]. The indentation sites were examined underoptical microscopy (Nikon Epiphot 300, Nikon Corp., Japan)with Nomarski contrast.For selected sintering conditions, larger specimens wereprepared and cut into polished bars (25 mm  5 mm  5 mm)with chamfered edges to analyze their fracture behaviour.Flexural strength was determined using a four-point bendingdevice – with outer and inner spans of 20 mm and 10 mm,respectively – assembled onto a universal testing machine(Instron Corp, Canton, MA). Tests were carried out at ambientconditions, using a crosshead speed of 15 mm/min. 3. Results and discussion Fig. 2 shows SEM micrographs of the samples sintered at1100,1200,and1300  8 Cfor1and7 h.Theseimagesevidenceasignificant grain growth produced with increasing sinteringtemperature and time. Microcracking is also evident in thesamples sintered at 1300  8 C (marked with arrows in Fig. 2e andf),butnotinthe rest.Nevertheless,extensiveSEMobservationsrevealed the presence of a few isolated microcracks in thesamples sintered at 1200  8 C for 5 and 7 h. The absence of microcracking at 1200  8 C for 1 h and 3 h is an unexpectedresult because this temperature is above the b $ a transforma-tion temperature of 1125  8 C.Apart from the microcracks, the SEM observations alsoshow that the microstructure of the samples sintered at 1300  8 Cpresents isolated regions with much finer grains (Fig. 2e and f).As shown in Fig. 3, the compositional maps obtained by EDSreveal that these regions are deficient in Ca with respect to the b -TCP matrix. No evidence for such a new microconstituentwas observed in any case at 1100 and 1200  8 C. Taken together,the morphology, chemical composition, and temperature atwhich the new microconstituent appears suggest that itoriginates from the recrystallization of a liquid phase.According to the phase diagram of  Fig. 1, that new Fig. 1. Partial CaO–P 2 O 5  equilibrium phase diagram [19]. The followingnotation is adopted to identify the phase composition: CHP: heptacalciumphosphate (Ca 7 (P 5 O 16 ) 2 ), CPP: calcium pyrophosphate (Ca 2 O 7 P 2 ), TCP: tri-calcium phosphate (Ca 3 (PO 4 ) 2 ), and TTCP: tetracalcium phosphate(Ca 4 (PO 4 ) 2 O). The gray vertical line marks the composition of the startingpowders and the circles indicate the sintering temperatures analyzed in thiswork. F.H. Perera et al./Ceramics International 36 (2010) 1929 – 1935 1930  microconstituent would indeed result from the eutectictransformation of the liquid that forms above 1288  8 C.Fig. 4 compares the XRD patterns of the starting powdersand of the samples sintered at 1200  8 C for 7 h and at 1300  8 Cfor 1 h. The comparison reveals that the phase composition of the starting powders is maintained up to 1300  8 C. At thistemperature, however, a new phase appears, identified as  a -calcium pyrophosphate ( i.e ., a -Ca 2 P 2 O 7 , denoted a -CPP in thephasediagramof Fig.1).Therelativeproportionsof  b -TCPand a -CPP determined by the Rietveld method are 84  1 and16  1 wt.%, respectively. In addition, the phase compositionof the samples sintered at 1300  8 C is the same within theexperimental error, regardless of the sintering time (1–7 h).Logically, the  a -CPP grains are located in the Ca-deficientregions in Fig. 3.The average grain size is plotted in Fig. 5 as a function of sintering time for each sintering temperature. The symbolsrepresent experimental data, with standard deviation as errorbars, and the lines are empirical fits. It has to be noted that thedata corresponding to 1300  8 C were calculated neglecting thesmall grains within the eutectic microconstituent since theysrcinate from the recrystallization of a liquid phase duringcooling, and therefore are not affected by sintering time. As canbe seen, increasing sintering temperature and time producessignificant grain growth in this material. Clearly, the effect of sintering temperature is more marked because diffusion is athermally activated process.As a consequence of the great susceptibility of this materialto grain growth, its densification is very slow, as clearlyevidenced in Fig. 6, which shows the density data as a functionof time for the three sintering temperatures studied. Again, thesymbols represent experimental data, with standard deviationas error bars, and the lines are linear fits to the data. The densityincreases with increasing sintering temperature but neverreaches the theoretical value for  b -TCP (3.07 g/cm 3 ). Thedensification rate is virtually zero (0.004  0.003 g/cm 3 h) at1100  8 C, increases slightly at 1200  8 C (0.035  0.006 g/cm 3 h,around 1% increase in density per hour) and is slightly negative Fig. 2. SEM micrographs showing the microstructures of samples sintered for 1 h (left) and 7 h (right) at 1100  8 C (a and b), 1200  8 C (c and d), and 1300  8 C (e and f).Differences in grain size and porosity are apparent (note scales). Arrows in (e and f) mark the presence of microcracks. F.H. Perera et al./Ceramics International 36 (2010) 1929 – 1935  1931  at 1300  8 C (statistically significant as confirmed by one-wayANOVA test). This latter result is attributed to the increase inthe number and size of cracks that have developed in thematerial, together with a negligible diffusion-induced densi-fication rate with increasing sintering time, at this temperature.In any case, it is evident that a significant increase in density isachieved by increasing sintering temperature, especially up to1300  8 C because the presence of the liquid phase helps to fillthe pores, significantly enhancing densification [9].The improved densification translates into an increase in thehardness of the material with sintering temperature and time, asshown in Fig. 7a. However, elastic modulus data ( E  * =  E   / (1  n 2 )) obtained during the same Berkovich instrumentedindentation tests, which are plotted in Fig. 7b, do not show thesame trend. They first increase steadily up to 1200  8 C at 3 htreatmentandthendroptoanearlyconstantvalue.Theseresultsare explained by an increasing level of chipping around theimprint for sintering conditions above 1200  8 C for 3 h. Indeed,since the elastic modulus in instrumented indentation iscalculated from unloading data in the load  vs  penetration depthcurve, if cracks appear in the material during loading the Fig. 5. Average grain size as a function of sintering time, for the indicatedsintering temperatures. The symbols represent experimental data, with standarddeviation as error bars, and the lines are empirical best-fits.Fig. 6. Evolution of density with sintering time at the indicated temperatures.Symbols represent experimental data, with standard deviation as error bars, andlines are linear fits to data.Fig. 4. XRD patterns corresponding to  b -TCP starting powders and samplessintered at the indicated conditions. No phase transformation is observed up to1300  8 C where new peaks (arrows), corresponding to  a -CPP, appear.Fig. 3. SEM image (a) and Ca content mapping by EDS (b) of a b -TCP samplesintered at 1300  8 C for 3 h. A new microconstituent, Ca-deficient compared tothe matrix, is apparent (dark regions). F.H. Perera et al./Ceramics International 36 (2010) 1929 – 1935 1932  measured values are lower than those expected for anundamaged material.The optical images in Fig. 8 illustrate the aforementionedincrease in chipping with sintering temperature by showingrepresentative indentations on samples sintered for 3 h at thethreeselectedtemperatures.Whilechippingisnotobservableat1100  8 C (Fig. 8a) and hardly noticeable at 1200  8 C (Fig. 8b), it is extensive at 1300  8 C (Fig. 8c). Further evidence of the occurrence of chipping during indentations was the observationof an increasing number of pop-in events in the load-displacement curves as the sintering temperature and timeincreased. This increased level of chipping is attributed to thedevelopment of transformation-induced microcracking duringsintering, which hints at a significant reduction in the strengthof the material.The reduction of strength hypothesized in the precedingparagraph is clearly evidenced in the results of  Fig. 9. ThisfigureshowstheWeibullplotforthestrengthdataobtainedin4-point bending tests performed on samples sintered for 3 h at thethree selected temperatures. Samples sintered for 3 h wereselected for these tests because they exhibited the bestindentation results at each temperature. This plot shows thefailure probability,  P , as a function of applied stress,  s  . Thestraight lines are the best fits to data of the Weibull probabilityfunction [16–18], P ¼ 1  exp   s s  f    m    (1)where the Weibull modulus,  m , and central value,  s  f  , areadjustable parameters. This plot clearly shows that increasing Fig. 8. Optical images with Nomarsky contrast showing representative indenta-tion imprints on samples sintered for 3 h at (a) 1100  8 C, (b) 1200  8 C, and (c)1300  8 Candgold-coatedafterthetests.Extensivechippingisevidentat1300  8 C.Fig.7. Evolutionof(a)hardnessand(b)elasticmodulus, E  * =  E   /(1  n 2 ),of  b -TCP with temperature and sintering time. Average values from Berkovichinstrumented indentation tests, with standard deviations as error bars. F.H. Perera et al./Ceramics International 36 (2010) 1929 – 1935  1933
Related Search
Related Docs
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks