Journal of Non-Crystalline Solids 355 (2009) 874–879

Contents lists available at ScienceDirect

Journal of Non-Crystalline Solids

journal homepage: www.elsevier .com/ locate / jnoncrysol

Local surface damage and material dissolution in 45S5 bioactive glass:Effect of the contact deformation

Ding Li a, Mimi X. Yang b, Pavitra Muralidhar b, Connie Wu b, Fuqian Yang a,*

a Department of Chemical and Materials Engineering, University of Kentucky, 161 Anderson Hall, Lexington, KY 40506, United Statesb MSTC Program, Paul Laurence Dunbar High School, Lexington, KY 40513, United States

a r t i c l e i n f o

Article history:Received 13 March 2008

PACS:61.43.Fs62.20.mm81.70.q

Keywords:BioglassCrack growthIndentation, Microindentation

0022-3093/$ - see front matter � 2009 Elsevier B.V. Adoi:10.1016/j.jnoncrysol.2009.04.026

* Corresponding author. Tel.: +1 859 257 2994; faxE-mail address: fyang0@engr.uky.edu (F. Yang).

a b s t r a c t

Bioactive glasses react with the human physiological solution in control of their biofunctionality. Thestress state in bioactive glasses determines the chemomechanical reaction and their biofunctionality.Using the microindentation technique, the effect of the indentation deformation on the surface damageand material dissolution of 45S5 bioglass was investigated. The indentation-induced local surface dam-age, including ring cracks and radial cracks, was revealed before and after the immersion tests in phos-phate buffer solution (PBS). There existed a critical load for the formation of the radial cracks, whichemanated from the periphery of the outmost ring crack. The growth of the radial cracks in the PBS solu-tion displayed the stress-corrosion behavior with the crack-growth speed being a linear function of theindentation load. Fast dissolution occurred at the edges of the surface-damaged zone due to stress-assisted dissolution under the action of local tensile stress, which was different from the dissolutionbehavior of stress-free 45S5 bioglass.

� 2009 Elsevier B.V. All rights reserved.

1. Introduction of adult dogs and used scanning electron microscope to in vivo

The control of material deformation and dissolution is one ofthe most concerns for the biomedical applications of bioactivematerials in the reconstruction and substitution of bone and tooth.Localized stresses/strains through contact between materials are ofbiomedical and technological importance because their influenceon local chemomechanical behavior affects the service life and per-formance of bioactive materials and the growth of bioactive struc-tures such as a biologically active apatite layer. The understandingof the deformation-induced chemomechanical behavior of bioac-tive materials plays an important role in the processing and man-ufacturing of bioactive materials, which generally requires thatbioactive materials maintain reasonably mechanical strength andbiofunctionality when subjected to external mechanical loading[1].

The biofunctionality of bioactive materials generally can becharacterized in the phosphate buffer solution (PBS) and the simu-lated body fluid (SBF) and be determined by examining the growthrate of the apatite layer. Fazan and Marquis [2] evaluated thechanges in crystallinity and surface topology of plasma-sprayedhydroxyapatite coatings over pure titanium and determined thedissolution behavior of the hydroxyapatite coating. Xue et al. [3]coated hydroxyapatite over the cortical bone, muscle, and marrow

ll rights reserved.

: +1 859 323 1929.

examine the dissolution of the hydroxyapatite coatings. Most workreported in literature focused on either the dissolution or themechanical response of bioactive materials. There are only a fewstudies on the dissolution of bioactive materials under the actionof mechanical stresses. Han et al. [4] studied the effect of residualstresses on the dissolution of hydroxyapatite coatings on a tita-nium alloy. Reis et al. [5] observed the degradation of the hydroxy-apatite coatings on Ti–6Al–4V substrate through surface crackingand faster dissolution when subjected to cyclic bending. Bloyeret al. [6] evaluated the stress-corrosion crack growth of Si-basedbioactive glasses in the SBF solution. Li et al. [1] recently studiedthe stress effect on the material dissolution of 45S5 bioglass inthe PBS solution using the Vickers indentation.

It is known that localized plastic deformation in a material canbe created by the sharp-instrumented indentation. Such a localizeddeformation can change local chemomechanical response of a bio-active material and alter the local dissolution and biofunctionalityof the material. It is the purpose of this work to study the indenta-tion deformation of 45S5 bioglass using a spherical, diamond in-denter and to examine the effect of the indentation deformationon the material dissolution of 45S5 bioglass in the PBS solution.Ring cracks, radial cracks and local surface damage due to the con-tact between the indenter and the 45S5 bioglass are revealed. Thedependence of the growth of the radial cracks on the indentationload is also discussed.

D. Li et al. / Journal of Non-Crystalline Solids 355 (2009) 874–879 875

2. Experimental

Small discs of 3 mm in thickness were prepared from an an-nealed cast-rod of 45S5 bioglass (45 wt% SiO2, 24.5 wt% Na2O,24.5 wt% CaO, and 6 wt% P2O5) of 10 mm in diameter (MO-SCI Spe-cialty Products, L.L.C., Rolla, MO, 65402). The discs were groundand polished mechanically in isopropyl alcohol to obtain flat sur-face for the tests. The discs were regularly cleaned by isopropylalcohol (IPA) during the grinding and polishing and were cleanedafter the polishing in an ultrasonic bath for 20 min each in acetone,then IPA, and finally ethyl alcohol. The polished samples then wereheat-treated in air at 350 �C for 2 h and furnace-cooled to roomtemperature to remove the residual stresses generated in thegrinding and polishing before indentation.

Microindentations were made on the polished surface of thediscs using a diamond, spherical indenter of 50 lm in radius on aMicro-Combi Tester (CSM Instruments, Needham, MA). The peakindentation load applied to the indenter was in the range of500–3000 mN. The indentation loading rate and unloading ratewere 500 mN/min; and the holding time was 5 s at the peak load.Arrays of indents were created with a minimum of five indents foreach indentation load, and the spacing between two nearby in-dents was at least three times larger than the size of the indents.The impression marks were measured using an optical microscopeand a scanning electron microscope (SEM). After the indentations,the specimens were immediately stored in a desiccator.

The indented-samples later were placed into a PBS solution of�1 L in an agitated water bath with the indented surface up. ThePBS solution used was made from 5 mL of magnesium chloridesolution (3.8 g MgCl2 in 100 mL of de-ionized water (DI-H2O))and 1.25 mL of potassium dihydrogen phosphate solution (3.4 gKH2PO4 in 50 mL DIUF-H2O, diluted with 50 mL DI-H2O to make100 mL, with trace 1 N sodium hydroxide to maintain pH of 7.2)added to 1 L DI-H2O. The PBS solution was not agitated and wassealed to prevent from contamination and evaporation. The tem-perature was maintained at 37 �C to mimic the body temperature.The specimens were immersed into the PBS solution over the peri-ods of 1, 3, 5, 7, and 9 days, respectively.

3. Results and discussion

3.1. Microindentation

Fig. 1 depicts the SEM micrographs of the impression marksover a 45S5 bioglass by the indentation loads of 800 and3000 mN prior to the immersion test. Both ring cracks and radialcracks are observed around the indents, depending on the indenta-tion load. This is in contrast to the indentations made by a Vickersindenter, in which only radial cracks were observed [1]. There isonly one ring crack for the indentation made by the indentationload of 800 mN. In contrast, there are multiple ring cracks and

Fig. 1. Surface cracks created by the spherical indentations, (a) a ring crack (indentation3000 mN).

six radial cracks, as shown in Fig. 1(b), emanating from the periph-ery of the indentation mark/the outmost ring crack for the inden-tation made by the indentation load of 3000 mN. This resultsuggests that the ring cracks were first formed at small indentationloads and the radial cracks were formed after the formation of thering cracks at large indentation loads. The formation of the radialcracks was likely due to the stress concentration at the damagedsurface in the 45S5 bioglass around the ring cracks, although therewere no sharp edges on the surface of the spherical indenter. Thestress concentration caused the formation and propagation of localradial cracks.

Fig. 2 shows the dependence of the size of the impression markson the indentation load. The indentation size varied linearly withthe indentation load. Such a behavior is different from the para-bolic relation for the indentation of ductile materials and likely isdue to the formation and propagation of the surface damage overthe 45S5 bioglass during the indentation. More mechanical workwas dissipated for the development of anelastic deformation. Fromthe size of the impression marks, d, one can calculate the indenta-tion hardness, H, as

H ¼ 4P

p�d2; ð1Þ

where P is the peak indentation load and �d is the average size of theimpression marks. The dependence of the indentation hardness onthe indentation load is depicted in Fig. 3. The indentation hardnessgradually increases from 3.5 GPa at the indentation load of 800 mNto 6.1 GPa at the indentation load of 2000 mN. For the indentationloads larger than 2000 mN, there is only a slight change in theindentation hardness, and the indentation hardness can be approx-imated as a constant of 6.2 GPa. This value is slightly less than7.47 ± 0.51 GPa obtained from the indentations made by a Vickersindenter [1]. Such a difference is due to the effect of the indenter-surface profile, which produces different stress distribution under-neath the indentation.

For brittle materials, the indentation-induced surface cracks in-clude the median crack, radial crack, and lateral crack. Approximat-ing the indentation-induced median-radial cracks as center-loadedhalf-penny cracks, Lawn et al. [7] obtained the relationship be-tween the size of a median-radial crack and the indentation loadfor sharp indenters as

P ¼ vc3=2; ð2Þ

where c is the distance between the center of the indentation markand the crack tip, and v is a constant depending on the mechanicalproperties of the material and the indenter profile. Eq. (2) has beensupported by various experimental results. Recently, Dériano et al.[8] found that their experimental results of the indentation fractureof silica-rich glasses deviated from Eq. (2) and suggested that thedeviation was due to the densification underneath the indentationand the formation of ring cracks.

load: 800 mN), and (b) multiple ring cracks and six radial cracks (indentation load:

10

15

20

25

0 0.5 1 1.5 2 2.5 3 3.5

Inde

ntat

ion

size

(µm

)

Indentation load (103 mN)

Fig. 2. Dependence of the indentation size on the indentation load.

0

1

2

3

4

5

6

7

0 0.5 1 1.5 2 2.5 3 3.5

Hra

dnes

s (G

Pa)

Indentation load (103 mN)

Fig. 3. Dependence of the indentation hardness on the indentation load.

0

0.5

1

1.5

2

2.5

3

3.5

0 20 40 60 80 100 120 140

Inde

ntat

ion

load

(10

3 mN

)

(c-d/2)3/2 (µm3/2)

Fig. 4. Dependence of the size of the indentation-created cracks on the indentationload.

0.01

0.1

1

10

0.1 1 10

Ein

(µJ

)

Indentation load (10 3 mN)

n=1.5

n=2.6

Fig. 5. Dependence of the inelastic energy on the indentation load.

876 D. Li et al. / Journal of Non-Crystalline Solids 355 (2009) 874–879

Consider that the radial cracks started at the periphery of thering cracks in the indentation of the 45S5 bioglass. It is reasonableto assume that the radial cracks can be approximated as quadrant-penny cracks with the center of the cracks being at the periphery ofthe outmost ring crack and are subjected to concentrated force atthe center of the cracks due to the indentation deformation. Thus,the size of the radial crack becomes ðc � d=2Þ; and one can modifyEq. (2) as

P ¼ vðc � d=2Þ3=2: ð3Þ

Fig. 4 shows the dependence of the crack size on the indentationload. The indentation load is proportional to the 3/2 power ofðc � d=2Þ, supporting Eq. (3). This trend suggests that the ringcracks hindered the propagation of the radial cracks likely due tothe crack branch and local densification underneath the indenta-tion. Under such a condition, it is unclear if the fracture toughnessof the 45S5 bioglass can be evaluated from the radial cracks cre-ated by spherical indentations.

The inelastic energy, Ein, dissipated in an indentation cycle is thearea enclosed by the loading–unloading curve and the displace-ment axis, which represents the energetic contribution to the for-mation and propagation of surface damage and local inelastic

deformation. Using dimensional analysis, Chen et al. [9] showedthat the inelastic energy is a power-function of the indentationload as, Ein / Pn (n is a force exponent). Fig. 5 shows the depen-dence of the dissipated inelastic energy on the indentation load.In contract to the indentation of ductile materials [9–11], thereare two force exponents. For the indentation loads less than orequal to 1000 mN, the force exponent is 1.5, which is consistentwith the dimensional analysis for perfect plastic materials. Forthe indentation loads larger than 1000 mN, the force exponent is2.6, which suggests that more mechanical work was dissipated inthe indentation cycle. The difference is likely due to the formationand propagation of the radial cracks. At large indentation loads, theradial cracks were created as indicated in Fig. 4, which producednew crack surfaces in addition to local surface damage and inelas-tic deformation. The formation of the radial cracks required thecompensation of surface energy from the mechanical work. Thisresulted in more dissipation of mechanical work and the high forceexponent. Obviously, there exists a critical indentation load, atwhich the radial cracks will start to nucleate during the indenta-tion of the 45S5 bioglass.

D. Li et al. / Journal of Non-Crystalline Solids 355 (2009) 874–879 877

3.2. Material dissolution

The effect of the indentation deformation on the dissolutionbehavior of the 45S5 bioglass was studied following the indenta-tions. The indented 45S5 bioglass were immersed in a PBS solutionat temperature of 37 �C to mimic the body temperature. Fig. 6shows the sequence of the surface evolution around the indentsby an indentation load of 3000 mN in the PBS solution. The regionaround the cracks first underwent the material dissolution asshown in Fig. 6(b), which suggests the stress-assisted reaction forthe material dissolution. The material dissolution widened thegaps between the crack surfaces. Severe dissolution occurred overthe contact zone, which revealed multiple ring-type grooves anddebris. This is consistent with the observation of the ring cracks.The presence of the debris indicates that local subsurface damageand densification were produced during the indentation. Continu-ous exposure to the PBS solution led to more material dissolutionover the damaged area and deep dissolution along the cracks, asshown in Fig. 6(c) and (d). In the meantime, the exposure to thePBS solution also caused random dissolution of materials overthe surface away from the area surrounding the indents as shownin Fig. 6(e). A typical surface morphology of a 45S5 bioglass is

Fig. 6. SEM micrographs of the morphological evolution around the indents (indentadissolution, (d) 9-days’ dissolution, (e) an enlarged view of the surface topology around

shown in Fig. 6(f) after continuous exposure to the PBS solutionfor 9 days. The entire surface experienced the material dissolutionand became rough except the area surrounding the indents pro-duced by the indentation loads larger than 2000 mN, which be-came much smooth. This is likely due to the residualcompressive stresses created by the indentations, which hinderedthe dissolution of the 45S5 bioglass in the PBS solution over a longperiod of the immersion. It should be pointed out that the de-formed surfaces created by the indentation loads less than1500 mN were completely dissolved after 1 day.

Fig. 7 shows the material dissolution at an crack tip after 1 day’simmersion in the PBS solution for the indentation made by anindentation load of 3000 mN. The crack tip blunted and the gap be-tween the crack surfaces became wide. This is due to local residual,tensile stress, which assisted the material dissolution. All of theseresults suggest that the dissolution behavior of the 45S5 bioglassis a function of local stress state.

Fig. 8 shows the dependence of the distance of the crack tip tothe indentation center on the immersion time for the radial cracks.The crack size increases with the increase in the immersion time.The growth speed for the cracks made by large indentation loadsis higher than that for the cracks made by small indentation loads

tion load: 3000 mN); (a) as-indented surface, (b) 1-day’s dissolution, (c) 3-days’the indent, and (f) the global view over the indentation region.

Fig. 7. Material dissolution at the crack tip for an indent created by an indentationload of 3000 mN (immersion time: 1 day).

20

40

60

80

100

120

0 2 4 6 8 10

32.521.5

c (µ

m)

Immersion Time (Day)

Indentation load (10 3 mN)

Fig. 8. Dependence of the distance of the crack tip to the indentation center on theimmersion time for the radial cracks.

1

1.5

2

2.5

3

3.5

1 1.5 2 2.5 3 3.5

Cra

ck p

ropa

gatio

n sp

eed

(µm

/day

)

Indentation load (103 mN)

Fig. 9. Dependence of the crack propagation speed on the indentation load.

878 D. Li et al. / Journal of Non-Crystalline Solids 355 (2009) 874–879

since large indentation load created higher local residual stress atthe crack tip than that made by small indentation load. It is worthpointing out that the radial cracks produced by the indentationload of 1000 mN were completely dissolved after 1 day’s immer-sion. This suggests that the radial cracks did not pass through theindentation-created plastic zone.

It is believed that water reacts with glasses in the vicinity of acrack tip and weakens the bond strength. This allows the propaga-tion of cracks in aqueous environment when subjected to localresidual–tensile stress, i.e., stress-assisted cracking. The stress-as-sisted cracking process is complex, which can be divided into threeregions [12]. In region I, the crack growth is activated through thechemomechanical interaction between water and the material andthe growth speed is a function of local stress. In region II, the crack-growth speed is much faster than the reaction rate for the chemo-mechanical interaction and the growth behavior is independent ofatomic motion at the crack tip. In region III, the crack growth is re-lated to the bond rupture.

Several models [13–18] have used the stress-assisted chemicalreaction theory to analyze the stress-assisted crack growth andassociated the crack-growth speed with the stress intensity factoror the strain energy release rate. From the rate theory [19], Sergoet al. [20] suggested that the stress-assisted dissolution is a func-

tion of the hydrostatic stress. Thus, following the analysis givenby Pollet and Burns [18], one can express the dependence of thecrack-growth speed, V, on the local stress in the vicinity of a cracktip in the region of the stress-dependent crack growth as

V ¼ A exp�QRT

� �sin h

rhX� RE

RT

� �; ð4Þ

where A is a constant dependent upon the lattice-vibration fre-quency and atomic space, Q is the activation energy of the adsorp-tion–desorption procession, R is the gas constant, T is the absolutetemperature, X is an activation volume, and RE is two time of thesolid–liquid interface energy. rh is the trace of the stress tensor inthe front of the crack tip, which is positive for tensile stress andnegative for compressive stress. For rhX� RE << RT, Eq. (4) canbe simplified as

V ¼ A exp�QRT

� �rhX� RE

RT: ð5Þ

In general, one can assume rh / P if the characteristic time forthe relaxation of the local stress in the vicinity of the crack tip ismuch larger than the immersion time. Thus, Eq. (5) becomes

V ¼ A exp�QRT

� �aP � RE

RT: ð6Þ

The growth speed of the radial cracks is a linear function of theindentation load. From Fig. 8, one can calculate the crack-growthspeed through the curve fitting. Fig. 9 shows the dependence ofthe crack-growth speed on the indentation load. A linear relationis observed between the crack-growth speed and the indentationload, which is in good accordance with Eq. (6). The result suggeststhe presence of the stress-assisted crack propagation in the 45S5bioglass, when the indented 45S5 bioglass was immersed in thePBS solution. The crack growth is associated with the stress-assistedmaterial dissolution in the vicinity of the crack tip. The local stresscan enhance or suppress the dissolution of material and the propa-gation of surface cracks in the PBS solution, dependent upon thestress state.

4. Concluding remarks

The biofunctionality of bioactive glasses depends on the chemo-mechanical interaction in the human physiological solution. Usingthe microindentation technique, the local deformation of the 45S5bioglass was studied and the behavior of the material dissolution

D. Li et al. / Journal of Non-Crystalline Solids 355 (2009) 874–879 879

around the indents was investigated. For the indentations made bythe indentation loads of larger than 1000 mN, the radial crackswere created in addition to the multi-ring cracks over the contactzone. The indentation hardness gradually increased with theindentation load and reached a constant of 6.2 GPa for the indenta-tion load larger than or equal to 2000 mN. An empirical relationwas proposed to correlate the size of the radial cracks with theindentation load, which was supported by the experimental re-sults. The inelastic energy dissipated in an indentation cycle wasa power-function of the indentation load with a transition from asmall force exponent of 1.5 to a large force exponent of 2.6 atthe indentation load of 1000 mN. The transition likely corre-sponded to the formation of the radial cracks, which emanatedfrom the periphery of the ring cracks.

The effect of the indentation deformation on the dissolutionbehavior of the 45S5 bioglass in the PBS solution was examined.The bioglass surrounding the indents experienced different disso-lution behavior from the surface away from the indents. Fast disso-lution occurred over the indentation-created damaged zone,including the edges of the cracks. Multiple small debris and ring-type grooves were observed. This suggested that, in addition tothe ring cracks over the contact zone, the indentation created localsubsurface damage underneath the indentation. The depth andwidth of the grooves and the cracks increased with the immersiontime. Smooth surface was formed over a long period of the immer-sion surrounding the indents, which was likely due to the confine-ment of the material dissolution controlled by local compressive,residual stress. In contrast, the bioglass away from the indenta-tion-influenced zone showed random dissolution in a short timeperiod and eventually grew over the entire stress-free surface.

The growth of the radial cracks in the PBS solution showed thestress-corrosion behavior, falling into the category of the region I. Asimple relation was proposed between the crack-growth speed andthe indentation load from the rate theory. The experimental resultsfor the crack growth of the radial cracks supported this relation-

ship. Local material dissolution and the tip blunting were alsodemonstrated. All of these results indicate the potential of using lo-cal stresses to control the biofunctionality of bioactive glasses inthe human physiological solution.

Acknowledgements

This work is supported by the National Science Foundationthrough the grant CMMI-0800018. The authors are grateful forTed Day at MO-SCI Corporation for providing the bioglass rods,material data, and materials processing information.

References

[1] D. Li, F.Q. Yang, J. Nychka, Eng. Fract. Mech. 75 (2008) 4898.[2] F. Fazan, P.M. Marquis, J. Mater. Sci. Med. 11 (2000) 787.[3] W.C. Xue, X.Y. Liu, X.B. Zheng, C.X. Ding, J. Biomed. Mater. Res. A 74A (2005)

553.[4] Y. Han, K.W. Xu, J. Lu, J. Biomed. Mater. Res. 55 (2001) 596.[5] R.L. Reis, F.J. Monteiro, G.W. Hastings, J. Mater. Sci.-Mater. Med. 5 (1994) 457.[6] D.R. Bloyer, J.M. McNaney, R.M. Cannon, E. Saiz, A.P. Tomsia, R.O. Ritchie,

Biomaterials 28 (2007) 4901.[7] B.R. Lawn, A.G. Evans, D.B. Marshell, J. Am. Ceram. Soc. 63 (1980) 574.[8] S. Dériano, A. Jarry, T. Rouxel, J.-C. Sangleboeuf, S. Hampshire, J. Non-Cryst.

Solids 344 (2004) 44.[9] R. Chen, F.Q. Yang, P.K. Liaw, G.J. Fan, H. Choo, Mater. Trans. 48 (2007) 1743.

[10] F.Q. Yang, W.W. Du, K. Okazaki, J. Mater. Res. 20 (2005) 1172.[11] F.Q. Yang, L.L. Peng, K. Okazaki, Metall. Mater. Trans. A 35 (2004) 3323.[12] D.J. Green, An Introduction to the Mechanical Properties of Ceramics,

Cambridge University Press, 1998. p. 264.[13] R.J. Charles, J. Appl. Phys. 29 (1958) 1651.[14] S.M. Wiederhorn, J. Am. Ceram. Soc. 50 (1967) 407.[15] S.M. Wiederhorn, S.W. Freiman, E.R. Fuller, C.J. Simmons, J. Mater. Sci. 17

(1982) 3460.[16] R. Thomson, J. Mater. Sci. 15 (1980) 1014.[17] E.R. Fuller, R. Thomson, J. Mater. Sci. 15 (1980) 1027.[18] J.-C. Pollet, S.J. Burns, Int. J. Fract. 13 (1977) 667.[19] H. Eyring, J. Chem. Phys. 4 (1936) 283.[20] V. Sergo, O. Sbaizero, D.R. Clarke, Biomaterials 18 (1997) 477.