Phosphate Cement with Calcium Phosphate ... - Wiley Online Library

J. Am. Ceram. Soc., 98 [12] 4028–4035 (2015) DOI: 10.1111/jace.13858 © 2015 The American Ceramic Society

Journal

Reinforcement of a Magnesium-Ammonium-Phosphate Cement with Calcium Phosphate Whiskers Katharina Zorn,‡ Elke Vorndran,§ Uwe Gbureck,§ and Frank A. M€ uller‡,† ‡

Otto Schott Institute of Materials Research (OSIM), Friedrich Schiller University Jena, 07743 Jena, Germany

§

Department for Functional Materials in Medicine and Dentistry, University of W€ urzburg, 97070 W€ urzburg, Germany However, it appears that both properties are the real limiting factors for an application of phosphate cements in load-bearing bone repair.35 To achieve improvements, the reinforcement with fibers seems to be the most promising method.36 Large sized calcium-deficient hydroxyapatite [CDHA, Ca9(HPO4)(PO4)5OH] whiskers can be prepared by a hydrothermal treatment of calcium tripolyphosphate gel at pH 4–637 and in a previous study, we reported the transformation of such CDHA whiskers into polycrystalline b-tricalcium phosphate [b-TCP, Ca3(PO4)2] short fibers.38 In this study, we report on the incorporation of both types of fibers in a magnesium- ammonium-phosphate hexahydrate (struvite, MgNH4PO4∙6H2O) cement. Struvite is pathological calcification found in kidney stones after bacterial infection and struvite forming cements have been demonstrated to have antimicrobial properties and a high bonding strength to dentine.39 In contrast with other studies about mechanical properties of phosphate cements, this study goes beyond measuring only strength. It also includes examination of reliability and toughness and relates both to microstructural observations at the whisker–matrix interface, which also have been rarely reported so far.40

Self-setting resorbable phosphate cements are characterized by an excellent biocompatibility and bioactivity. However, poor mechanical properties restrict their application. Most studies which characterize phosphate cements mechanically focus on strength measurements. Examinations of mechanical reliability and facture toughness were hardly performed. In this study, calcium phosphate whisker-reinforced magnesium-ammoniumphosphate (struvite) cements were examined at the whisker– matrix interface and the measured strength, reliability and toughness values were correlated to these observations. Moreover, the toughening mechanisms were evaluated. It was shown that whisker incorporation is not beneficial for material strength. It led to a strength decrease from 29.8 to 21.8 MPa by the incorporation of 15 vol% calcium-deficient hydroxyapatite (CDHA) whiskers compared to the pure struvite cement. Weibull statistics and microstructural observations revealed that this is caused by the whisker–matrix interface, which acts as a flaw. In contrast with that, the reliability increases upon whisker incorporation. Furthermore, the critical stress intensity factor KIC as well as the work-of-fracture cwof increase from 0.52 to 0.60 MPam1/2 and from 9.5 to 12.9 J/m² by the addition of 15 vol% CDHA whiskers compared to the original struvite cement. It was shown that whisker pull-out and crack deflection are the main mechanisms responsible for this increase.

I.

II.

Materials and Methods

(1) Materials Processing To prepare the struvite cement a magnesium phosphate powder [farringtonite, Mg3(PO4)2] was synthesized by heating mixtures of magnesium hydrogen phosphate trihydrate (MgHPO4∙3H2O; Riedel-de-Haen, Seelze, Germany) and magnesium hydroxide [Mg(OH)2; Fluka, Steinheim, Germany] to 1100°C for 5 h, as described earlier.41 Milling was performed for 20 min in a planetary ball mill (PM400; Retsch, Haan, Germany) at 200 rpm with 500 mL agate jars, four agate balls (30 mm) and a load of 125 g powder per jar. For the preparation of the struvite cement, 0.0342 mol of the farringtonite powder and 0.0208 mol of diammonium hydrogen phosphate [(NH4)2HPO4; Carl Roth, Karlsruhe, Germany] were mixed. Subsequently, water was added at a farringtonite/water-ratio of 2.0. Then the cement paste was mixed for 30 s on a glass slab. Finally, the mass was transferred into silicon rubber molds to produce bending bars of 3 mm 9 4 mm 9 45 mm. The whiskers and short fibers were synthesized by a hydrothermal synthesis at 140°C for 24 h using 40 mL of a calcium nitrate solution, 50 mL of a sodium tripolyphosphate solution, and 60 mL of 2-propanol as described in previous papers.37,38 After filtration and drying of the product, the monocrystalline CDHA whiskers were heat treated at 1000°C for 16 h to receive polycrystalline b-TCP short fibers. To investigate the whisker-matrix interactions, 10 vol% of the farringtonite powder were replaced by 10 vol% of CDHA whiskers or 10 vol% of b-TCP fibers (Table I). To investigate the mechanical properties, 0, 5, 10, and 15 vol% CDHA whiskers, respectively, were introduced replacing the

Introduction

T

HE first patents on self-setting phosphate cements for clinical use were obtained by Driskell et al.1 and Brown and Chow in the mid 1980’s.2 One decade later, the first medical phosphate cements were commercially available.3,4 Since then, the interest in phosphate cements has grown rapidly.5 The reason for this is related to their excellent biological acceptance and their good handling properties in surgical application.6 However, due to their poor mechanical properties, the use of phosphate cements is limited to bone repair in nonload bearing areas. Thus, attempts were made to improve these properties. This was performed by incorporating other phases7–20 or by reducing the porosity of the cements with liquefiers and/or by compacting.21–24 As a result, strengths comparable to human cortical bone (130–180 MPa compressive strength, 135–193 MPa flexural strength) could be obtained.12,22,23,25 In contrast with the efforts and improvements concerning strength; reliability26–28 and toughness29–34 have hardly been considered and, when considered, only been observed phenomenologically.

G. Fischman—contributing editor

Manuscript No. 36894. Received May 13, 2015; approved August 2, 2015. † Author to whom correspondence should be addressed. e-mail: frank.mueller@ uni-jena.de

4028

December 2015 Table I.

Nomenclatures, Compositions and Examination Methods of the Prepared Cements Examination of

Sample name

MP MP5CDHA MP10CDHA MP15CDHA MP10TCP

4029

Reinforced Struvite Cements

Farringtonite CDHA TCP Fiber–Matrix Mechanical (vol%) (vol%) (vol%) Interactions properties

100 95 90 85 90

0 5 10 15 0

0 0 0 0 10

– – x – x

x x x x –

bearing span in mm, w the width of the bending bar in mm and b the thickness of the bending bar in mm. The data were statistically analyzed using the Weibull approach (DIN EN ISO 6872). For this purpose, the bending strength values were arranged in ascending order. Then the failure probability was assigned according to the rank using Eq. (2): Pf ¼

farringtonite. The original matrix cement is abbreviated MP and the resulting CDHA composites are named MP5CDHA, MP10CDHA, and MP15CDHA respectively. The b-TCP composite is abbreviated MP10TCP. For the mechanical tests, the bending bars were grinded on SiC paper (up to a grain size of 4000) to receive plan-parallel surfaces. For the toughness measurements, notching was performed additionally. The v-notch, which had a depth of 1 mm and a root radius of 20 lm, was cut in the 3 mm side in the center of the long axis of the bending bar. This was accomplished by using a notching machine with a razor blade with saw profile and a diamante suspension up to 0.25 lm. Before mechanical testing, all samples were stored at 36°C for 20 h.

i 0:5 N

(2)

where i is the rank and N the number of specimens of the control sample. Afterwards, the variables Pf and r were transformed into lnln[1/(1Pf)] as ordinate and lnr as abscissa. Finally, the data points were fit by a straight line using linear regression. (b) Critical Stress Intensity Factor: The critical stress intensity factor KIC was calculated according to Eq. (3)44 (DIN CEN/TS 14425-5): pffiffiffi pffiffiffi F S1 S2 3 a Y KIC ¼ r aY ¼ pffiffiffiffi w b w 2ð1 aÞ1:5

(3)

where Y ¼ 1:9887 1:326a

ð3:49 0:68a þ 1:35a2 Það1 aÞ ð1 þ aÞ2 (4)

(2) Characterization (A) Structural Characterization: The reactions at the interface between the whiskers, respectively, short fibers and the cement as well as the toughening mechanisms were characterized using a scanning electron microscope (SEM) (S440i; Leica, Wetzlar, Germany). For the examination of interface reactions by EDS analysis, two kinds of preparation techniques were used. In the first one, the samples were cut, embedded in epoxy resin and polished. In the second one, fractured samples were used. Afterwards both types of samples were coated with carbon. For a further examination of interface reactions, whiskers/fibers were treated with 1M phosphoric acid for 5 min. Treated and untreated whiskers/ fibers were superimposed on a polished aluminum sample plate and were sputtered with gold. For the analysis of the toughening mechanisms, fractured samples were sputtered with gold. (B) Mechanical Characterization: Mechanical testing was performed in a four-point bending configuration with a lower bearing span of 39 mm and an upper bearing span of 13 mm. This configuration is less sensitive to positioning errors and ensures that the beam overhang is large enough.42,43 An universal testing machine (Z020; Zwick, Ulm, Germany) with a crosshead speed of 0.5 mm/min was used. The preload was set to 0.5 N. The bending strength measurements were carried out with 15 samples of each composition. These samples were tested with the 3 mm side parallel to the load. Determination of toughness (critical stress intensity factor KIC and workof-fracture cwof) was carried out with five notched samples, with the 4 mm side parallel to the load and the notches heading downward (tension side). (a) Bending Strength: The bending strength r of the samples was calculated according to Eq. (1): r¼

Mb P2 L3 PL ¼ wb2 ¼ 2 wb W 6

(1)

where Mb is the bending moment in Nmm, W the modulus of resistance in mm³, P the fracture load in N, L the outer

and a¼

a w

(5)

pffiffiffiffi where KIC is the fracture toughness in MPa m, r the bending strength in MPa, F the fracture load in MN, b the width of the bending bar in m, w the thickness of the bending bar in m, S1S2 the distance between the mounting rollers in m, a the arithmetic mean depth of the v-notch in m, a the relative depth of the v-notch and Y the stress intensity form factor. (c) Work-of-Fracture: The work-of-fracture cwof was calculated from the area under the experimental load–displacement curve divided by twice the projected fracture surface area of the unnotched ligament shown in Eq. (6):45 cwof ¼

A 2bðw aÞ

(6)

where cwof is the work-of-fracture in J/m², A the area under the load-deflection curve in Nm, b the width of the bending bar in m, w the thickness of the bending bar in m, and a the arithmetic mean depth of the v-notch in m.

III.

Results and Discussion

(1) Whisker/Fiber-Matrix Interactions The whisker/fiber–matrix interfaces of the composite cements were analyzed via EDS. Figure 1 shows the EDS line scans of the cross sections of the CDHA whisker- and b-TCP fiber-reinforced struvite cements respectively. The scans started and ended on a struvite crystal and crossed the whisker/fiber. For both samples it can be seen that magnesium and phosphor are present in the same amounts in the area of the struvite crystals. As expected, no calcium can be found in this area. On the other hand, no magnesium is present in the

4030

Journal of the American Ceramic Society—Zorn et al.

Vol. 98, No. 12

(a)

(b)

Fig. 1. EDX line scan on the polished surface of a (a) CDHA whisker reinforced struvite cement (MP10CDHA) and a (b) b-TCP fiber reinforced struvite cement (MP10TCP).

whiskers/fibers. Instead calcium and phosphorous can be found. A difference between both samples can be seen in the decrease of the calcium content at the whisker/fiber–matrix interface. There is a sharp drop of the calcium content in the vicinity of the CDHA whisker [Fig. 1(a)], while there is a smoother decline in the vicinity of the b-TCP fiber [Fig. 1(b)]. Furthermore, in both cases an increase of the phosphorous content is visible at the interface between the struvite crystal and the calcium phosphate whisker/fiber. The reason for that is not clear at first sight, but an explanation can be given by looking at the setting reaction of the cement:46 2 Mg3 ðPO4 Þ2 þ 3 ðNH4 Þ2 HPO4 þ 36 H2 O ! 6 MgNH4 PO4 6H2 O þ H3 PO4

(7)

It becomes evident that the byproduct phosphoric acid was formed during setting. Hence, the increase of the phosphorous content between the struvite crystals is caused by the deposition of phosphoric acid during the setting reaction. The smoother decrease of the calcium content in the vicinity

of the b-TCP fibers could be explained by the fact that they dissolve in contact with the phosphoric acid. A reason for this could be the polycrystalline grain structure of the b-TCP fibers which leads to grain etching along grain boundaries. This cannot occur on the CDHA whiskers due to their single crystal characteristic. To verify this, the CDHA whiskers and the b-TCP fibers where soaked in a 1M phosphoric acid solution for 5 min. Figure 2 shows the results of the test. The CDHA whiskers [Fig. 2(c)] are hardly destroyed by the acid treatment. Etching only occurred at flaws which resulted from synthesis [Fig. 2(a)]. On the other hand, the polycrystalline b-TCP fibers show severe etch marks after acid treatment [Fig. 2(d)]. These result from grain boundary etching and affect the whole fiber. Additionally, EDS line scans were also carried out on fractured surface (Fig. 3) to get an impression how the etching influences the interface reaction between whisker/fiber and matrix. For this purpose, analyses were performed in areas where whiskers/fibers were in contact to a large amount of phosphoric acid. In the EDS line scans these areas are characterized by a large phosphorous concentration. Figure 3(a) illustrates that the defect-free CDHA whisker surfaces do not

(a)

(b)

(c)

(d)

Fig. 2. (a) CDHA whisker and (b) b-TCP fiber before the soaking in H3PO4, and (c) CDHA whisker and (d) b-TCP fiber after the storage in 1M H3PO4 for 5 min.

December 2015


4031

(a)

(b)

Fig. 3. EDX line scan on the fractured surface of a (a) CDHA whisker reinforced struvite cement (MP10CDHA) and a (b) b-TCP fiber reinforced struvite cement (MP10TCP).

adhere to the matrix due to an incomplete reaction with the phosphoric acid. In contrast with that, an interface layer formed around the b-TCP fibers which led to a strong bonding to the matrix [Fig. 3(b)]. The presence of the phosphoric acid [Eq. (7)] and the decreasing calcium concentration at the boarders of the fibers [Fig. 1(b)] lead to the conclusion that the interface layer must be a calcium phosphate which has a lower pH than b-TCP and a lower molar Ca/P ratio. Most probably dicalcium phosphate (DCPD, CaHPO4∙2H2O) formed by reaction (8): Ca3 ðPO4 Þ2 þ H3 PO4 þ 6H2 O ! 3CaHPO4 2H2 O

(8)

This reaction is also used for the preparation of brushite cements.47 Furthermore, it is well-established that DCPD is stabilized by magnesium,48 which is abundant in the struvite matrix cement. In the case of the CDHA whiskers the reaction at whisker defects could be explained by Eq. (9): Ca9 ðHPO4 ÞðPO4 Þ5 OH þ 3H3 PO4 þ 17H2 O ! 9CaHPO4 2H2 O

(9)

This reaction also leads to a DCPD interface bonding to the struvite matrix cement. To examine the mechanical properties of composites, the CDHA whiskers were chosen to be incorporated into the struvite cement, since they are single crystals that are expected to be less sensitive to etching and fracture compared to the b-TCP fibers.

(2) Mechanical Properties (A) Strength and Reliability: Figure 4 shows the measured bending strengths r and, for comparative purposes, the critical stress intensity factors KIC for the samples MP to MP15CDHA. While the toughness of the struvite cement increases from 0.52 to 0.60 MPam1/2 by the incorporation of 15 vol% CDHA whiskers, the average strength decreases from 29.8 to 21.8 MPa.

Fig. 4. Bending strength r and critical stress intensity factor KIC (mean SD) of struvite cements as a function of the amount of incorporated CDHA whiskers.

To analyze the observed strength decrease, in Fig. 5 the measured strength values of every specimen are plotted in a Weibull scale. It can be seen that the strength values of the samples MP and MP5CDHA can be fitted quite well by a straight line. Thus, it can be concluded that the strength values are Weibull distributed. However, the strength values of MP10CDHA and MP15CDHA strongly deviate from the Weibull line at lnr = 3. Deviations from Weibull statistics are very well-described by Danzer et al.49 They showed that materials containing flaws of a narrow shaped size distribution have a step function of stress in the Weibull plots. This step function can be clearly seen in the graphs of sample MP10CDHA and sample MP15CDHA. This leads to the conclusion that these samples contain, in addition to the naturally occurring flaws, flaws which are artificially introduced by the whisker incorporation. It seems likely that these flaws are generated by the unbonded whisker–matrix interface. To proof this observation, the critical flaw size of sample MP15CDHA was calculated using the Irwin equation [Eq. (3)]. As a result of this calculation, the determined criti-

4032

Vol. 98, No. 12


when adding 15 vol% CDHA whiskers compared to the unreinforced struvite cement. Characteristic strength–displacement diagrams of the matrix cement MP and the reinforced cement MP15CDHA are shown in Fig. 6. Both samples deviate from linear-elastic fracture mechanics (LEFM). An analysis of the curves was performed by dividing the plots into different regions. The contribution of the different regions 0A, AB, BC, CD, and DE to cwof are shown in Table II. From Fig. 6 and Table II it becomes evident that the regions BC and DE have the strongest impact on the increase of cwof. From literature about civil engineering concretes, it is well-known that the deviation from LEFM in the prepeak region (region AB) can be attributed to the formation of microcracks during loading.52 Microcrack formation is a result of internal stresses caused by thermal expansion or elastic stiffness differences of second phase particles.53 In the here examined struvite cement, microcracking occurs due to the fact that it is—even without whisker incorporation—a composite, since it consists of farringtonite and struvite grains. Differences in elastic stiffness and the existence of low fracture energy interfaces (grain boundaries) provide good conditions for microcracking.36 This is the reason why a nonlinear prepeak response can be observed in both—the unreinforced cement MP and the reinforced cements MP5CDHA-MP15CDHA. It has to be noted that—despite the observed prepeak nonlinearity—the assumptions of LEFM are still fulfilled sufficiently to legitimate calculation of KIC as described above.54 The sharper decrease of load (region BC) of the unreinforced cement MP compared to reinforced cement MP15CDHA indicates the quick coalescence of the prior formed microcracks to a rapid growing macrocrack.52 This leads to catastrophic failure, which can only be absorbed by frictional forces (region DE) to some extent.52 In contrast with that, in the reinforced samples a catastrophic failure is inhibited (region BC) by the incorporated whiskers, which

Fig. 5. Weibull plots of reinforced struvite cements.

cal flaw size is 230 20 lm, which lies in the range of the maximum whisker size.38 Thus, it can be concluded that the decrease in bending strength is caused by the whisker incorporation due to the poor interface bonding between the fibers and the matrix. In addition to these observations, a drift of the strength values of the reinforced samples in the low probability range (low lnln[1/(1Pf)]) to higher values (higher lnr) seems to occur which leads to a kinking of the curves in the low probability range. This kind of deviation from Weibull distribution was also observed for other fiber composites50 and might result from an increasing crack resistance curve.49 Furthermore, Fig. 5 also shows that the strength distributions, i.e., the ranges of lnr values, become smaller by the incorporation of whiskers, which means that the samples fail with less variation in strength. Thus, they have a higher reliability. Commonly, Weibull modulus (i.e., the slope of the Weibull lines) is used to describe the reliability. However, in this study the Weibull moduli (as well as the characteristic Weibull strengths) are not named because the strength values of MP10CDHA and MP15CDHA are not Weibull distributed. Nevertheless, the strength plots (Fig. 5) clearly show the smaller variation in strength and with this the increased reliability with increasing whisker incorporation. Summing up the results of the strength measurements, it can be said that the strength of the struvite cement drops due to the incorporation of unbonded whiskers. Nevertheless, reliability increases by incorporating whiskers. (B) Fracture Toughness: In addition to the fracture toughness parameter KIC (Fig. 4), the determination of the work-of-fracture cwof was performed (Table II, last column). Whereas KIC represents the critical stress intensity at peak load, cwof is an averaged fracture energy, i.e., the total energy to produce one unit area of fracture surface during a complete fracture.51 Thus, the latter also includes energy consumption processes, which operate after peak load is reached. Table II shows that cwof increases from 9.5 to 12.9 J/m²

Table II.

Sample name

MP MP5CDHA MP10CDHA MP15CDHA

Fig. 6. Characteristic stress–displacement diagrams of the notched samples MP and MP15CDHA.

Determined Work-of-Fracture cwof-Values of the Prepared Cements and the Fracture Energies c Assigned to the Different Regions of the Stress–Displacement Diagram [mean (SD)] c0A (J/m²)

7.4 6.7 7.5 8.1

(1.7) (0.9) (0.7) (0.8)

cAB (J/m²)

1.3 1.3 1.6 1.9

(0.3) (0.5) (0.2) (0.9)

cBC (J/m²)

0.4 0.6 0.8 1.3

(0.2) (0.1) (0.4) (0.4)

cCD (J/m²)

0.3 0.3 0.3 0.4

(0.1) (0.1) (0.1) (0.1)

cDE (J/m²)

0.2 0.4 0.8 1.2

(0.1) (0.2) (0.4) (0.2)

cwof (J/m²)

9.5 9.4 11.1 12.9

(1.2) (0.5) (0.8) (1.1)

December 2015

4033


can bridge the crack.55 A bridging of the crack faces is possible, when the interfacial fracture energy is much lower than that of the matrix or the whisker,55 so that the crack can proceed along a whisker without provoking whisker fracture or crack deflection. The increased “tail” of the stress– displacement curve of MP15CDHA (region DE) can be explained by an increase of frictional effects,52 which result from fiber pull-out. For a verification of the processes described above, the fracture surfaces were evaluated (Fig. 7). It can be observed that the surface consists of numerous microcracks. Furthermore, many pull-out marks and detachments [white arrows in Figs. 7(a) and (b)] can be found. These exactly resemble the whiskers, which were present before pull-out or detachment. No surface roughening and microstructural changes compared to the matrix microstructure can be seen. Additionally, in contrast with observations of M€ uller et al. on HA whiskers reinforcing a calcium phosphate cement,15 no interface layer on pulled-out whiskers can be observed in the CDHA-struvite system [Fig. 7(a)]. That leads to the conclusion that there is no interface bonding between these pulledout whiskers and the matrix. Beyond that, on other spots [black arrows in Figs. 7(a) and (b)] fiber breakage and crack formation around the interface layer were observed. These mechanisms indicate a very strong interface bonding in these areas. In the here examined CDHA-struvite cement, this strong bonding is the result of a reaction of phosphoric acid with defects on CDHA whiskers. From these observations it can be concluded that the fiber–matrix interface is of alternating strength. In some region it is very strong [black arrows in Fig. 7(a)] and in other regions it is just frictional [white arrows in Fig. 7(a)] in nature. From the explanations above, crack propagation in the CDHA whisker reinforced struvite cement can be explained as follows (Fig. 8): Under an applied load elastic deformation occurs first [Fig. 8(a)]. After that, upon stress increase, microcracks start to form on grain boundaries, which are abundant in both—the unreinforced and the fiber-reinforced struvite—cements [Fig. 8(b)]. Then, at a particular stress the formed microcracks coalescence to a macrocrack. In the whisker reinforced cements rapid macrocrack propagation is inhibited by whisker bridging [Fig. 8(c)]. However, after a

certain macrocrack propagation distance the bridging force of the whiskers is not large enough to carry the applied load and the macrocrack starts to propagate rapidly [Fig. 8(d)]. Finally, the bridging whiskers are pulled out [Fig. 8(e)]. In the CDHA whisker reinforced sturvite cement, whisker bridging and pull-out [Fig. 7(d/I)] are possible in areas where the whiskers did not react with the phosphoric acid, because the fracture energy of the interface there is much lower than that of the matrix or the whisker.55 However, if there is a strong whisker–matrix interface on the way of crack propagation along the whisker, crack defection [Figs. 7(b), (c), and (d/II)] can occur. In some cases, the crack may also be trapped [Fig. 7(c)], when a whisker is surrounded by an alternating weak and strong interface. That can additionally lead to crack branching [Fig. 7(c)] or whisker breakage [Fig. 7(a)] depending on the size of the interface layer. From the operating fracture mechanisms the increase in work-of-fracture due to crack branching can be neglected, since it was rarely observed. The increase in work-of-fracture due to whisker fracture can also be neglected, because the introduced whiskers are brittle.51 The crack face bridging generated contribution is involved in and eventually converted to pull-out energy.51 Thus, the main mechanisms, which lead to an increase in work-of-fracture of the here observed CDHA whisker–struvite cement, are whisker pullout and crack deflection. The increase in work-of-fracture due to pull-out is the result of an interfacial shear stress s between whisker and matrix. This interfacial shear stress can be described by the Coulomb friction law36,56 and depends on the friction coefficient along the whisker–matrix interface and the cement setting shrinkage. The increase in work-of-fracture due to crack deflection can be attributed to the longer path the crack has to take and the reduced stress intensity at the crack tip, since the crack is no longer perpendicular to the tensile stress.57

IV.

Conclusions

It was shown that phosphoric acid which is formed as a byproduct of the struvite setting reaction leads to an

(a)

(c)

(b)

(d)

(i)

(ii)

Fig. 7. (a)–(c) Fractured surface of the sample MP15CDHA, indicating the alternating fiber–matrix interface strength which is either very strong (black arrows) or weak (white arrows), and (d) a scheme illustrating the mechanisms which lead to the increase in fracture toughness.

4034

Vol. 98, No. 12


Fig. 8. Schematic illustration of the crack propagation in a CDHA whisker reinforced struvite cement and its effect on the measured stress–displacement graph.

interface reaction of incorporated b-TCP fibers or CDHA whiskers to the struvite matrix cement. In the case of the CDHA whiskers, this reaction only takes place at defects. As a consequence of this, the bonding of the CDHA fibers to the matrix is either very strong or just frictional in nature. The partially poor whisker-matrix bonding is the reason for a decrease in bending strength from 29.8 to 21.8 MPa by the incorporation of 15 vol% CDHA whiskers compared to the pure struvite cement, because the introduced whiskers act as flaws. However, the alternating bonding conditions lead to an increase in fracture toughness via crack deflection and frictional pull-out during fracture. An increase of the workof-fracture from 9.5 to 12.9 J/m² and in the critical stress intensity factor from 0.52 to 0.60 MPam1/2 was achieved through these mechanisms by the addition of 15 vol% CDHA whiskers to the struvite cement.

Acknowledgment The authors acknowledge financial support from the Deutsche Forschungsgemeinschaft (DFG) under grant nos. Gb1/11-2 and Mu1803/7-2.

References 1

T. D. Driskell, A. L. Heller, and J. F. Koenigs, “Dental Treatments”; US patent No. 3,913,229, 1975. 2 W. Brown and L. Chow, “Dental Restorative Cement Pastes”; US Patent No. 4518430, 1985. 3 D. B. Kamerer, B. E. Hirsch, C. H. Snyderman, P. Costantino, and C. D. Friedman, “Hydroxyapatite Cement–A New Method for Achieving Watertight Closure in Transtemporal Surgery,” Am. J. Otol., 15 [1] 47–9 (1994). 4 B. R. Constantz, et al., “Skeletal Repair by In-Situ Formation of the Mineral Phase of Bone,” Science, 267 [5205] 1796–9 (1995). 5 M. Bohner, U. Gbureck, and J. E. Barralet, “Technological Issues for the Development of More Efficient Calcium Phosphate Bone Cements: A Critical Assessment,” Biomaterials, 26 [33] 6423–9 (2005).

December 2015


6 A. J. Ambard and L. Mueninghoff, “Calcium Phosphate Cement: Review of Mechanical and Biological Properties,” J. Prosthodont., 15 [5] 321–8 (2006). 7 E. F. Burguera, H. H. K. Xu, S. Takagi, and L. C. Chow, “High Early Strength Calcium Phosphate Bone Cement: Effects of Dicalcium Phosphate Dihydrate and Absorbable Fibers,” J. Biomed. Mater. Res. A, 75A [4] 966–75 (2005). 8 T. Christel, M. Kuhlmann, E. Vorndran, J. Groll, and U. Gbureck, “Dual Setting Alpha-Tricalcium Phosphate Cements,” J. Mater. Sci. Mater. Med., 24 [3] 573–81 (2013). 9 N. J. S. Gorst, et al., “Effects of Fibre Reinforcement on the Mechanical Properties of Brushite Cement,” Acta Biomater., 2 [1] 95–102 (2006). 10 W. J. E. M. Habraken, et al., “Introduction of Enzymatically Degradable Poly(Trimethylene Carbonate) Microspheres Into an Injectable Calcium Phosphate Cement,” Biomaterials, 29 [16] 2464–76 (2008). 11 M. Kon, L. M. Hirakata, Y. Miyamoto, H. Kasahara, and K. Asaoka, “Strengthening of Calcium Phosphate Cement by Compounding Calcium Carbonate Whiskers,” Dent. Mater. J., 24 [1] 104–10 (2005). 12 R. Kruger, J. M. Seitz, A. Ewald, F. W. Bach, and J. Groll, “Strong and Tough Magnesium Wire Reinforced Phosphate Cement Composites for LoadBearing Bone Replacement,” J. Mech. Behav. Biomed. Mater., 20 [3] 6–44 (2013). 13 Q. Lian, D. C. Li, J. K. He, and Z. Wang, “Mechanical Properties and In-Vivo Performance of Calcium Phosphate Cement-Chitosan Fibre Composite,” P. I. Mech. Eng. H, 222 [H3] 347–53 (2008). 14 J. L. Moreau, M. D. Weir, and H. H. K. Xu, “Self-Setting Collagen-Calcium Phosphate Bone Cement: Mechanical and Cellular Properties,” J. Biomed. Mater. Res. A, 91A [2] 605–13 (2009). 15 F. A. Muller, U. Gbureck, T. Kasuga, Y. Mizutani, J. E. Barralet, and U. Lohbauer, “Whisker-Reinforced Calcium Phosphate Cements,” J. Am. Ceram. Soc., 90 [11] 3694–7 (2007). 16 I. S. Neira, Y. V. Kolen’ko, K. P. Kommareddy, I. Manjubala, M. Yoshimura, and F. Guitian, “Reinforcing of a Calcium Phosphate Cement With Hydroxyapatite Crystals of Various Morphologies,” ACS Appl. Mater. Interfaces, 2 [11] 3276–84 (2010). 17 Z. H. Pan, P. P. Jiang, Q. Y. Fan, B. Ma, and H. P. Cai, “Mechanical and Biocompatible Influences of Chitosan Fiber and Gelatin on Calcium Phosphate Cement,” J. Biomed. Mater. Res. B, 82B [1] 246–52 (2007). 18 Y. H. Shen, J. H. Liu, K. L. Lin, and W. B. Zhang, “Synthesis of Strontium Substituted Hydroxyapatite Whiskers Used as Bioactive and Mechanical Reinforcement Material,” Mater. Lett., 70 [7] 6–79 (2012). 19 H. H. K. Xu, F. C. Eichmiller, and A. A. Giuseppetti, “Reinforcement of a Self-Setting Calcium Phosphate Cement With Different Fibers,” J. Biomed. Mater. Res., 52 [1] 107–14 (2000). 20 F. C. M. Driessens, M. G. Boltong, R. Wenz, and J. Meyer, “Calcium Phosphates as Fillers in Struvite Cements”; pp. 161–4 in Bioceramics, Vol. 17, Edited by P. Li. Trans Tech Publications Ltd., Uetikon-Zuerich, 2005. 21 J. E. Barralet, L. M. Grover, and U. Gbureck, “Ionic Modification of Calcium Phosphate Cement Viscosity. Part II: Hypodermic Injection and Strength Improvement of Brushite Cement,” Biomaterials, 25 [11] 2197–203 (2004). 22 J. E. Barralet, M. Hofmann, L. M. Grover, and U. Gbureck, “HighStrength Apatitic Cement by Modification With Alpha-Hydroxy Acid Salts,” Adv. Mater., 15 [24] 2091–4 (2003). 23 U. Gbureck, J. E. Barralet, K. Spatz, L. M. Grover, and R. Thull, “Ionic Modification of Calcium Phosphate Cement Viscosity. Part I: Hypodermic Injection and Strength Improvement of Apatite Cement,” Biomaterials, 25 [11] 2187–95 (2004). 24 R. I. Martin and P. W. Brown, “Mechanical-Properties of Hydroxyapatite Formed at Physiological Temperature,” J. Mater. Sci. Mater. Med., 6 [3] 138– 43 (1995). 25 A. J. W. Johnson and B. A. Herschler, “A Review of the Mechanical Behavior of CaP and CaP/Polymer Composites for Applications in Bone Replacement and Repair,” Acta Biomater., 7 [1] 16–30 (2011). 26 J. E. Barralet, T. Gaunt, A. J. Wright, I. R. Gibson, and J. C. Knowles, “Effect of Porosity Reduction by Compaction on Compressive Strength and Microstructure of Calcium Phosphate Cement,” J. Biomed. Mater. Res., 63 [1] 1–9 (2002). 27 J. E. Barralet, L. Grover, T. Gaunt, A. J. Wright, and I. R. Gibson, “Preparation of Macroporous Calcium Phosphate Cement Tissue Engineering Scaffold,” Biomaterials, 23 [15] 3063–72 (2002). 28 J. P. Morgan and R. H. Dauskardt, “Notch Strength Insensitivity of SelfSetting Hydroxyapatite Bone Cements,” J. Mater. Sci. Mater. Med., 14 [7] 647–53 (2003). 29 W. Z. Liu, J. T. Zhang, P. Weiss, F. Tancret, and J. M. Bouler, “The Influence of Different Cellulose Ethers on Both the Handling and Mechanical Properties of Calcium Phosphate Cements for Bone Substitution,” Acta Biomater., 9 [3] 5740–50 (2013). 30 E. F. Morgan, D. N. Yetkinler, B. R. Constantz, and R. H. Dauskardt, “Mechanical Properties of Carbonated Apatite Bone Mineral Substitute: Strength, Fracture and Fatigue Behaviour,” J. Mater. Sci. Mater. Med., 8 [9] 559–70 (1997).

4035

31 R. M. O’Hara, J. F. Orr, F. J. Buchanan, R. K. Wilcox, D. C. Barton, and N. J. Dunne, “Development of a Bovine Collagen-Apatitic Calcium Phosphate Cement for Potential Fracture Treatment Through Vertebroplasty,” Acta Biomater., 8 [11] 4043–52 (2012). 32 M. D. Weir and H. H. K. Xu, “Osteoblastic Induction on Calcium Phosphate Cement-Chitosan Constructs for Bone Tissue Engineering,” J. Biomed. Mater. Res. A, 94A [1] 223–33 (2010). 33 J. T. Zhang, F. Tancret, and J. M. Bouler, “Fabrication and Mechanical Properties of Calcium Phosphate Cements (CPC) for Bone Substitution,” Mat. Sci. Eng. C Mater., 31 [4] 740–7 (2011). 34 Y. Zhang, H. H. K. Xu, S. Takagi, and L. C. Chow, “In-Situ Hardening Hydroxyapatite-Based Scaffold for Bone Repair,” J. Mater. Sci. Mater. Med., 17 [5] 437–45 (2006). 35 J. T. Zhang, W. Z. Liu, V. Schnitzler, F. Tancret, and J. M. Bouler, “Calcium Phosphate Cements for Bone Substitution: Chemistry, Handling and Mechanical Properties,” Acta Biomater., 10 [3] 1035–49 (2014). 36 A. G. Evans, “Perspective on the Development of High-Toughness Ceramics,” J. Am. Ceram. Soc., 73 [2] 187–206 (1990). 37 Y. Mizutani, M. Hattori, M. Okuyama, T. Kasuga, and M. Nogami, “Large-Sized Hydroxyapatite Whiskers Derived From Calcium Tripolyphosphate gel,” J. Eur. Ceram. Soc., 25 [13] 3181–5 (2005). 38 K. Zorn, D. Mitr o, E. Vorndran, U. Gbureck, and F. A. M€ uller, “Hydrothermal Synthesis of Calcium-Deficient Hydroxyapatite Whiskers and Their Thermal Transformation to Polycrystalline Beta-Tricalcium Phosphate Short Fibers,” Bioinspired, Biomimetic and Nanobiomaterials, 2 [1] 11–9 (2012). 39 G. Mestres, et al., “Antimicrobial Properties and Dentin Bonding Strength of Magnesium Phosphate Cements,” Acta Biomater., 9 [9] 8384–93 (2013). 40 R. Kruger and J. Groll, “Fiber Reinforced Calcium Phosphate Cements On the way to Degradable Load Bearing Bone Substitutes?” Biomaterials, 33 [25] 5887–900 (2012). 41 E. Vorndran, A. Ewald, F. A. Muller, K. Zorn, A. Kufner, and U. Gbureck, “Formation and Properties of Magnesium-Ammonium-Phosphate Hexahydrate Biocements in the Ca-Mg-PO4 System,” J. Mater. Sci. Mater. Med., 22 [3] 429–36 (2011). 42 T. Lube, M. Manner, and R. Danzer, “The Miniaturisation of the 4-Point-Bend Test,” Fatigue Fract. Eng. M, 20 [11] 1605–16 (1997). 43 F. I. Baratta, W. T. Matthews, and G. D. Quinn, Errors Associated with Flexure Testing of Brittle Materials. U. S. Army Materials Technology Laboratory, Watertown, 1987. 44 D. Munz and T. Fett, Mechanisches Verhalten Keramischer Werkstoffe– Versagensablauf, Werkstoffwahl, Dimensionierung. Springer, Berlin, Heidelberg, New York, London, Paris, Tokyo, Hong Kong, 1989. 45 M. Sakai and H. Ichikawa, “Work-of-Fracture of Brittle Materials With Microcracking and Crack Bridging,” Int. J. Fract., 55 [1] 65–79 (1992). 46 U. Klammert, E. Vorndran, T. Reuther, F. A. Muller, K. Zorn, and U. Gbureck, “Low Temperature Fabrication of Magnesium Phosphate Cement Scaffolds by 3D Powder Printing,” J. Mater. Sci. Mater. Med., 21 [11] 2947– 53 (2010). 47 M. Bohner, J. Lemaitre, and T. A. Ring, “Effects of Sulfate, Pyrophosphate, and Citrate Ions on the Physicochemical Properties of Cements Made of Beta-Tricalcium Phosphate-Phosphoric Acid-Water Mixtures,” J. Am. Ceram. Soc., 79 [6] 1427–34 (1996). 48 M. H. Alkhraisat, J. Cabrejos-Azama, C. R. Rodriguez, L. B. Jerez, and E. L. Cabarcos, “Magnesium Substitution in Brushite Cements,” Mat. Sci. Eng. C Mater., 33 [1] 475–81 (2013). 49 R. Danzer, P. Supancle, J. Pascual, and T. Lube, “Fracture Statistics of Ceramics - Weibull Statistics and Deviations From Weibull Statistics,” Eng. Fract. Mech., 74 [18] 2919–32 (2007). 50 Z. P. Bazant and J. Planas, Fracture and Size Effect in Concrete and Other Quasibrittle Materials. CRC Press, Roca Raton and London, 1998. 51 S. M. Barinov and M. Sakai, “The Work-of-Fracture of Brittle Materials – Principle, Determination, and Applications,” J. Mater. Res., 9 [6] 1412–25 (1994). 52 B. L. Karihaloo, Fracture Mechanics and Structural Concrete. Longmann Scientific & Technical, New York City, New York, 1995. 53 J. B. Wachtman, W. R. Cannon, and M. J. Matthewson, Mechanical Properties of Ceramics, 2nd edition. John Wiley & Sons, Inc., Hoboken, New York, 2009. 54 J. R€ osler, H. Harders, and M. B€aker, Mechanisches Verhalten der Werkstoffe, 4th edition. Springer Vieweg, Wiesbaden, 2012. 55 P. F. Becher, C. H. Hsueh, P. Angelini, and T. N. Tiegs, “Toughening Behavior in Whisker-Reinforced Ceramic Matrix Composites,” J. Am. Ceram. Soc., 71 [12] 1050–61 (1988). 56 U. Lohbauer, R. Frankenberger, A. Clare, A. Petschelt, and P. Greil, “Toughening of Dental Glass Ionomer Cements With Reactive Glass Fibres,” Biomaterials, 25 [22] 5217–25 (2004). 57 K. N. Xia and T. G. Langdon, “The Toughening and Strengthening of Ceramic Materials Through Discontinuous Reinforcement,” J. Mater. Sci., 29 [20] 5219–31 (1994). h