Fatigue behavior of Ti6Al4V and 316 LVM blasted with ceramic particles of interest for medical devices.

journal of the mechanical behavior of biomedical materials 30 (2014) 30 –40

Available online at www.sciencedirect.com

www.elsevier.com/locate/jmbbm

Research Paper

Fatigue behavior of Ti6Al4V and 316 LVM blasted with ceramic particles of interest for medical devices S. Barriusoa, J. Chaoa, J.A. Jimeńeza, S. Garcıáb, J.L. Gonza´lez-Carrascoa,c,n a

Centro Nacional de Investigaciones Metalúrgicas (CENIM-CSIC), Avda Gregorio del Amo n1 8, 28040 Madrid, Spain Surgival, SL, Valencia, Spain c Centro de Investigación Biomédica en Red en Bioingeniería, Biomateriales y Nanomedicina (CIBER-BBN), Madrid, Spain b

ar t ic l e in f o

abs tra ct

Article history:

Grit blasting is used as a cost-effective method to increase the surface roughness of

Received 24 July 2013

metallic biomaterials, as Ti 6Al 4V and 316 LVM, to enhance the osteointegration, fixation

Received in revised form

and stability of implants. Samples of these two alloys were blasted by using alumina and

10 October 2013

zirconia particles, yielding rough (up to Ra 8 μm) and nearly smooth (up to Ra 1 μm)

Accepted 13 October 2013

surfaces, respectively. In this work, we investigate the sub-surface induced microstructural

Available online 25 October 2013

effects and its correlation with the mechanical properties, with special emphasis in the

Keywords:

fatigue behavior. Blasting with zirconia particles increases the fatigue resistance whereas

Grit blasting

the opposite effect is observed using alumina ones. As in a conventional shot penning

Ti 6Al 4V

process, the use of rounded zirconia particles for blasting led to the development of

316 LVM

residual compressive stresses at the surface layer, without zones of stress concentrators.

Mechanical behavior

Alumina particles are harder and have an angular shape, which confers a higher capability

Fatigue strength

to abrade the surface, but also a high rate of breaking down on impact. The higher

Biomaterials

roughness and the presence of a high amount of embedded alumina particles make the blasted alloy prone to crack nucleation. Interestingly, the beneficial or detrimental role of blasting is more intense for the Ti 6Al 4V alloy than for the 316 steel. It is proposed that this behavior is related to their different strain hardening exponents and the higher mass fraction of particles contaminating the surface. The low value of this exponent for the Ti 6Al 4V alloy justifies the expected low sub-surface hardening during the severe plastic deformation, enhancing its capability to soft during cyclic loading. & 2013 Elsevier Ltd. All rights reserved.

1.

Introduction

A goal of the implantology research is to design devices that induce controlled, guided, and rapid integration into surrounding tissues. cpTi, Ti 6Al 4V alloy and austenitic stainless steel 316 LVM are the materials of choice for many load bearing surgical implants because combine mechanical n

Corresponding author at: Tel.: þ34 915538900; fax: þ34 915347425. E-mail address: [email protected] (J.L. González-Carrasco).

1751-6161/$ - see front matter & 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jmbbm.2013.10.013

strength, excellent corrosion resistance and good biocompatibility. Considering that events leading to integration of the implant will take place largely at the tissue–implant interface, they are often additionally coated or simply surface modified by blasting with ceramics such as angular shaped particles (grit blasting), or rounded shaped particles (shot peening). The consequences of a given surface treatment are defined


mainly by the nature of the particles, but the target and the parameters used in the process also determine the changes experienced by the surface. For instance, the angular particles lead to more material removal than rounded ones and the incident angle influences noticeably the depth of the residual stress generated (Hong et al., 2008). Therefore, the appropriated procedure should be employed accordingly with the specific application. In particular, grit blasting with oxide particles has emerged as a low cost solution to improve the osteointegration, fixation and stability of the implants by increasing the surface area available for bone/implant apposition (Wennerberg et al., 1996; Goldberg et al., 1995; Aparicio et al., 2004). Typical applications are surface modifications of Ti and Ti 6Al 4V alloys for orthopaedic and dental implants. Grit blasting of 316 LVM is also considered an attractive modification of intramedullary nails for the proximal femur and diaphysary fractures, providing an optimal combination between high resistance during the consolidation period and a minimal invasive geometry. Besides the type of materials, success or failure of the devices involves implant-related factors, such as shape, topography, and surface chemistry, but also surgical technique and patient variables, such as bone quantity and quality. Influence of the coating/surface modification on the mechanical behavior of the implant becomes obviously essential for load bearing applications. This fact determines the emphasis of this paper since although in vivo experience with grit blasted implants is sound, some interrogations exist about the understanding of the blasting induced effects on mechanical behavior (Multigner et al., 2009a). The overall picture of the consequences resulting from blasting with ceramic particles should, therefore, be evaluated for materials use and in particular in bio-medical applications. However, such studies have rarely been done. Furthermore, it will be shown that conventional blasting with ceramic particles may lead to microstructures and mechanical properties which are close to those achieved by more sophisticated versions of this process, such as surface severe plastic deformation (S2PD) processes and references therein (Ortiz et al., 2008; Zhang et al., 2003; Wang and Li, 2003). In nearly every case, grit blasting rough the surface and yields a severe plastic deformation that induces microstructural changes in a narrow zone beneath the blasted surface. Depending of the substrate, grain size refinement and formation of new phases may occur (Multigner et al., 2009b, 2010). The net effect is a subsurface hardening with the highest hardness values beneath the blasted surface. In addition, compressive residual stresses with a maximum value close to the surface are often reported (Multigner et al., 2009c, 2010; Leinenbach and Eifler 2006). Influence of such changes on the fatigue behavior is well documented for Ti and TiAl4V alloys, although some inconsistencies are found. For instance, blasting of Ti 6Al 4V leads to a pronounced decrease of the fatigue strength during either axial or bending fatigue tests, which contrast with the higher fatigue resistance of shot penned Ti 6Al 4V (Wagner and Lüetjering, 1982). This different behavior is associated to the presence of microcracks at the surface and to the pronounced stress concentration in the vicinity of the particles that remain embedded onto the grit blasted surface (Leinenbach

31

and Eifler, 2006). Fatigue limit decreases with increasing roughness; about 20% decrease for fine blasting (Ra 2 μm) and up to about 40% for specimens with coarse blasting (Ra7 μm) (Baleani et al., 2000). This fatigue behavior of grit blasted Ti 6Al 4V is somewhat puzzling since blasting of cp Ti, causing a similar microstructural damage, enhances the fatigue properties (Gil et al., 2007). Therefore, conclusions about the reasons for the decrease in the fatigue strength of grit blasted Ti 6Al 4V alloy remain inconclusive. On the other hand, information on the influence of grit blasting on the fatigue behavior of 316 LVM steel is scarce despite the interest to improve fatigue strength of intramedullary nails for the proximal femur and diaphysary fractures. In this work we investigate the fatigue behaviour of Ti 6Al 4V and 316 LVM steel blasted with Al2O3 and ZrO2 particles, which are widely used to develop rough or nearly smooth surfaces, respectively. Fatigue behaviour will be investigated by rotating bending tests. During this testing, alternative tensile and compressive stresses are applied with the maxima values at the surface. Therefore, the surface and subsurface blasting induced effects will play a critical role. Special emphasis is devoted to correlate the microstructural changes at the blasted affected zone with the mechanical properties of the bulk.

2.

Materials and methods

Hot rolled and annealed (700 1C/1 h) bars (ϕ 25 mm) of Ti 6 Al 4V ELI (Extra Low Interstitial), and hot rolled and quenched bars (ϕ 30 mm) of austenitic stainless steel 316 LVM (Low Vacuum Melting) were supplied by the implant manufacturer (SURGIVAL SL, Valencia, Spain). Discs of about 20 mm in diameter and 2 mm thick, and rods to machine the fatigue specimens were removed from the bars. For the rotating bending fatigue tests, a set of specimens of 80 mm in length with 5.885 mm diameter in the gauge length was machined from the rods. Blasting of the specimens was performed by the implant manufacturer applying a jet of oxide particles under a pressure of 350 kPa for 1 min and with a distance between the nozzle and the target surface of 20 cm. A first set of specimens, hereafter BL-AlO, was blasted with angular particles of corundum of 2100 HV in hardness, apparent density of 1.65–1.80 g/cm3, and sized between 1 mm and 2 mm (Fig. 1a). The second set of specimens, hereafter named BL-ZrO, was blasted using micro spheres of ZrO2 crystals coated with silice, of 500–800 HV in hardness and apparent density of 3.76 g/cm3, sized between 125 μm and 250 μm (Fig. 1b). Whereas discs were fully blasted, fatigue specimens were only blasted at the gauge length. All specimens were cleaned following standard procedures and passivated in acid citric before delivered by the implant manufacturer. Quantitative surface roughness was determined with a profilometer Mitutoyo Surftest 401. The measurements were obtained from line profiles along a 4 mm length. The surface roughness was characterized by average surface roughness (Ra) in micrometer at a high sensitivity setting (0.01 μm).

32


a

10 m

b

1 m

Fig. 1 – BEI images obtained on non-etched 316 LVM (a) and Ti 6Al 4V (b) samples in the as-polished conditions. Microstructural characterization of the surface morphology and cross sections of selected specimens was carried out by scanning electron microscopy (SEM), using a field emission gun scanning electron microscope (JEOL-6500F) equipped with an energy dispersive X-ray (EDX) system for chemical analysis. Contrast of backscattered electron images (BEI) obtained on fresh polished samples at low voltages can be related to differences in the grain orientation or local composition. Discrimination between both effects can be easily obtained by tilting the samples during SEM examination, since contrast associated to grains will change. Thus simultaneous information on grain size and phases can be obtained without etching. Average grain size was determined on a set of representative fields of view at 2000 magnifications. X-ray diffraction (XRD) measurements were carried out with a Bruker AXS D8 diffractometer equipped with Co tube and Goebel mirror optics to obtain a parallel and monochromatic X-ray beam. A current of 30 mA and a voltage of 40 KV in both, grazing incidence condition and conventional θ–2θ scan were used to analyze depth profiling of the phase composition and crystal microstructure. The thickness at which 90% of the X-ray are scattered, t0.9, in both cases were calculated using the AbsorbDX software by Bruker AXS. For Co Kα radiation and an incidence angle of 11, t0.9 is 0.3 mm and 0.4 mm for the Ti 6Al 4V and the 316 stainless steel, respectively. In the case of θ–2θ scan, t0.9 values are equal to 3.28 mm and 5.96 mm for the diffraction peak (1 0 1) and

(2 0 1) for the Ti 6Al 4V alloy, and 5.10 mm and 9.76 μm for the diffraction peak (1 1 1) and (3 1 1) for the stainless steel, respectively. Operational conditions were selected to obtain X-ray diffraction diagrams with sufficient counting statistics and narrow peaks. XRD data were collected in both type of scans over a 2θ range of 30–1101 with a step width of 0.031. The phase present in the XRD patterns were identified using the JCPDS data base and the DIFFRACplus EVA software by Bruker AXS. It is well known that the Rietveld method is a powerful tool for calculation of structural parameters from diffraction patterns. In this work, instrument functions were empirically parameterized for both, grazing incidence and θ 2θ scans, from the profile shape analysis of a corundum plate sample. The version 4.2 of Rietveld analysis program TOPAS (Bruker AXS) for the XRD data refinement was used in this study. The refinement protocol included also the major parameters like, background, zero displacement, the scale factors, the peak breath, the unit cell parameter and texture parameters. The room temperature structures used in the refinement were an appropriated combination of the phases present in the base materials (α-Ti and β-Ti for the Ti 6Al 4V alloy and austenite for the 316 stainless steel) and oxide particles used for blasting ( α-Al2O3 or ZrO2). The α-alumina or corundum has a trigonal structure, with space group R3c and cell parameters a ¼0.4758 nm and c¼ 1.2992 nm. The rhombohedral cell contains two formular units where oxygen anions occupy 18c Wyckoff positions (x¼ 0.3062, y¼ 0, z¼ 1/2), whereas the aluminum cations are located at 12c positions (x ¼0, y¼ 0, z ¼0.3522) (Sawada, 1994). Monoclinic ZrO2 or baddeleyite has the space group P21/c and cell parameters a ¼0.5145 nm, b ¼0.5207 nm, c ¼0.5311 nm and β¼ 99.231. This cell contain four ZrO2 units cell where all the atoms occupy 4e Wyckoff positions, with Zr at x ¼ 0.2758, y¼ 0.0411, z¼ 0.2082; OI at x ¼ 0.0703, y¼ 0.3359, z ¼0.3406; and OII x ¼0.4423, y¼ 0.7549, z¼ 0.4789 (Smith and Newkir, 1965). The hardness distribution beneath the surface was measured on cross-sections of blasted specimens previously coated with a relatively thick layer of Cu obtained by electrolytic deposition. This coating preserves the blasted surface during preparation of the cross section but also allows the approach to the initial blasted interface with the indenter. To avoid interactions between adjacent indentations the measurements were performed perpendicularly to the blasted interphase drawing a zigzag path. The Vickers microhardness measurements were performed in a Wilson equipment using 15 s of dwell time. Load was very small (98 mN) to obtain indentations of small size and then approach to the blasted surface as much as possible. Fatigue specimens were tested in a rotating bending fatigue machine. The specimens were tested with constant stress amplitude (R ¼ 1) at different load levels. The fatigue strength was established as the stress after 107 cycles without failure (infinite life). For comparative purposes of the blasting induced effects on Ti 6Al 4V and 316 steel, the applied stress was normalized with regards to their yield strength. The curves of the applied stress (sa) versus number of cycles to failure (Nf) curves for each surface condition were plotted based on these tests.


33

Table 1 – Surface parameters and mass variations of investigated materials in the un-blasted and blasted conditions.

316 (LVM)

Ti 6Al 4V (ELI)

Surface condition

Ra (μm)

Rz (μm)

Mass variation (g)

Lost volume (cm3)

Polished BL-ZrO BL-AlO

0.00670.001 1.1970.10 8.2770.93

– 7.4770.15 48.0076.00

– þ0.0047 0.0076

– 0.95 10 3


0.00670.001 0.9570.04 5.1370.45

– 6.8470.06 29.0073.03

– 0.0005 0.0053

0.11 10 3 1.18 10 3

3.

Results

3.1.

Microstructural characterization

Fig. 1 (new) illustrates typical BEI images of the as-received materials in the as-polished condition. Microstructure of the 316 VM samples, Fig. 1a, reveals equiaxed grains of about 25 mm thick with numerous annealing twins inside thermally induced during processing. Ti 6Al 4V samples, Fig. 1b, show a typical duplex microstructure consisting of primary α grains (dark contrast) of about 4 mm thick and Widmanstatten αþβ colonies. SEM examination of the grit blasted specimens reveals that the impact with zirconia or alumina particles causes a severe surface plastic deformation of the substrate and produces an irregular rough surface morphology. Average Ra and Rz values are summarized in Table 1. As can be seen, blasting with alumina yields roughness values higher than those with zirconia, irrespectively of the substrate. These differences should be understood within the framework of the specific features that distinguish between both abrasive particles. As shown in Fig. 2, the alumina particles, Fig. 2a, are much larger and have a rough topography characterized by edge-like facets, whereas the zirconia particles, Fig. 2b, are rounded. The angular shaped alumina particles generate irregular protrusions and intrusions with sharp ridges, forming at some various places cracks-like defects. However, the zirconia particles produce a more homogeneous deformation without grinding down the material. As shown in Fig. 3, SEM examination of grit blasted specimens reveals the presence of dark zones containing large and heterogeneous sized particles, often broken, of the abrasive used, as confirmed by X-ray elemental mapping (not shown). However, the direct quantification of the particles remaining on blasted surfaces by SEM is difficult since small particles are hard to differentiate from the bulk material (Schuh et al., 2005). For this reason, X-ray diffraction patterns were used to characterize and quantify the contamination with alumina or zirconia particles with more accuracy. As shown in Tables 2 and 3, quantitative phase analysis by the Rietveld refinement reveals the presence of a higher mass fraction of embedded particles when alumina instead zirconia particles are used as abrasive for both, the Ti 6Al 4V alloy and the stainless steel. However, the mass fraction of Al2O3 in the Ti alloy (Table 2) is about 10 times higher than in the stainless steel (Table 3). On the other hand, these tables also show that for both alloy substrates, the difference between the mass fraction of

Fig. 2 – Secondary electron images of particles used for blasting. (a) Alumina and (b) Zirconia.

oxide particles calculated from the grazing incidence and θ 2θ X-ray diffraction scans is much lower for the alumina than for the zirconia. This means that alumina particles are able to penetrate to greater depths, creating a larger contaminated zone under the surface. The net effect is similar to that from a mechanical alloying of the material at the surface. The abrasive role of the blasting particles is manifested by the mass loss of the specimens, which is more severe when blasting with the angular alumina particles (Table 1). Mass loss is higher for the 316 LVM steel. However, considering that density of steel (8 g/cm3) is much higher than that of Ti 6Al 4V (4.5 g/cm3), the volume loss during blasting would be a more realistic factor to take into consideration. As a mater of fact, this parameter indicates a

34


Fig. 3 – BEI images of the surfaces of 316 LVM (a) and (b) and Ti 6Al 4V (c) and (d) blasted with particles of zirconia (a) and (c) and alumina (b) and (d). Table 2 – Depth profiling of phases present at the outermost surface of Ti 6Al 4V for the un-blasted and blasted condition as determined by X-Ray diffraction using (a) grazing incidence and (b) conventional θ–2θ scan.

(a) Grazing incidence Composition% (m/m)

(b) Conventional θ–2θ scan Composition % (m/m)



α-Phase

β-Phase

ZrO2

Al2O3

86.8 61.0 11.7

13.2 11.9 1.9

– 27.1 –

– – 86.4

α-Phase

β-Phase

ZrO2

Al2O3

86.1 78.1 32.4

13.9 12.4 4.9

– 9.5 –

– – 62.7

Table 3 – Depth profiling of phases present at the outermost surface of 316 LVM for the un-blasted and blasted condition as determined by X-Ray diffraction using (a) grazing incidence and (b) conventional θ–2θ scan.

(a) Grazing incidence Composition % (m/m)

Conventional θ–2θ scan Composition% (m/m)



more severe abrasion during blasting of the Ti 6Al 4V alloy. This feature is likely related to its lower ductility rather than to its higher hardness (see next sections).

γ-Phase

ZrO2

Al2O3

100 94.3 92.1

– 5.7 –

– – 7.9

γ-Phase

ZrO2

Al2O3

100 98.2 93.6

– 1.8 –

– – 6.4

Cross sectional examinations, Fig. 4, reveals that in addition to the irregular rough surface morphology, grit blasting produces significant subtle microstructural variations that


10 µm

35

10 µm

10 µm

10 µm

Fig. 4 – BEI images obtained on cross sectional views of non-etched 316 LVM (a) and (b) and Ti 6Al 4V (c) and (d) blasted with particles of alumina (a) and (c) and zirconia (b) and (d).

are consequence of the severe plastic deformation through cold work. Bottom part of Fig. 4 provides evidence of the bulk microstructure, thus the blasting affected zone can be envisaged. In the case of the 316 LVM steel, Fig. 4a and b, three zones can be distinguished without a clearly defined borderline. The zone just beneath the surface is characterized by an ultrafine microstructure containing randomly distributed nanometer scale grains. This zone seems to be rather larger for the BL-AlO ( 30 μm thick), Fig. 4a, than for the BL-ZrO specimens ( 15 μm thick), Fig. 4b. The next zone (about 50 μm depth) presents highly deformed grains, as well as twins and martensite needles without well defined grain frontiers. The third and deepest zone shows a progressive change in the backscattered signal, which shows a slight change in the crystallographic orientation. The grains are not altered for depth of about 100 μm and 200 μm for zirconiaand alumina grit blasted specimens, respectively. In the case of the blasted Ti 6Al 4V alloy, Fig. 4c and d, SEM examination of the αþβ colonies beneath the blasted surface provides evidence of a uniform narrow band severely deformed during blasting of 12 mm and 8 mm in thickness for the BL-AlO, Fig. 4c, and BL-ZrO samples, Fig. 4d, respectively. This deformation is more evident in the BL-AlO specimen than in the BL-ZrO one. The morphology of the β-phase contrasts with the rather homogeneous distribution in the bulk, where it forms a cellular-type structure that encloses quite equiaxed α-grains. The severe deformation beneath the surface is also manifested at higher magnifications (not shown) by the presence at the α-phase of an ultrafine/ deformed grained structure of 12 mm and 8 μm in thickness for the BL-AlO and BL-ZrO samples, respectively.

3.2.

Subsurface hardening

Microhardness measurements along a line perpendicular to the blasted 316 LVM surfaces, Fig. 5, reveals a gradient in hardness with a maximum of 370 HV0.01 close to the surface, irrespectively the particle used for blasting. In both cases, hardness decreases with increasing depth, achieving a near constant value of about 180 HV0.01 at a depth of 200 μm and 100 μm into the bulk, depending on whether the blasting was performed with particles of Al2O3 or ZrO2, respectively. Hardness measurements were also performed on the cross sections at blasted zones of the Ti 6Al 4V alloy, Fig. 5b. Since the primary α grains are generally softer than the Widmanstatten αþβ colonies, scattering of results was somewhat rather high. In comparison with the blasted steel surfaces, the blasting affected zones are thinner ( r30 mm) and the gradients in hardness are smaller, irrespectively the particles used for blasting, which reveals a lower hardening capacity of the Ti 6Al 4V alloy. Taking into consideration the higher scattering of results on the alumina blasted affected zones, it can be established than blasting with Al2O3 particles yields a more severe but less homogeneous plastic deformation, which in comparison with the zirconia blasted surfaces is manifested by a wider blasting affected zone (30 μm versus 20 μm) and a higher hardness increase (17% versus 7%).

3.3. Tensile and compressive behavior of non-blasted specimens In order to determine mechanical behavior of bulk alloys in the as-received condition, tensile and compressive tests were

36


values deduced from these curves. In the case of the nonblasted 316 LVM steel, Fig. 7a, it is shown that fatigue strength ( 400 MPa) is very close to the yield strength value. Interesting, whereas blasting with zirconia particles causes a moderated increase in the fatigue strength ( 420 MPa), blasting with alumina yield a significant decrease ( 370 MPa) till about 85% of the yield strength. The effect of blasting on the fatigue strength of Ti 6Al 4V is more noticeable. As shown in Fig. 7b, fatigue strength of non-blasted specimens ( 650 MPa) represents about 70% of the yield strength. Blasting with zirconia particles increases the fatigue strength ( 710 MPa) by about a 10%, whereas impact with alumina particles causes a significant decrease ( 485 MPa), which correspond to about a 50% of the yield strength. Fig. 8 shows typical fractures of broken specimens of 316 LVM and Ti 6Al 4V blasted with alumina particles. As expected, fracture initiates at the blasted affected zones where stresses were higher. For seek of clarity it should be mentioned that such fractures correspond to specimens broken at stresses above the fatigue strength. Details on the crack initiation place were not obtained since these fractures are damaged during the tensile-compressive cycles, but likely

performed in the non-blasted state. The relevant parameters are included in Table 4. Differences in the yield strength observed during tensile or compressive tests are related to the strength differential effect (Chait, 1973). As can be seen, Ti 6Al 4V is much more resistant than 316 LVM steel and consequently less ductile, as expected. For illustrative purposes, Fig. 6 shows the true stress versus true strain during compression. Relevant for the present study is that the strain rate exponent, determined in the strain range of 0.02 to 0.1 during the compression tests is 0.29, for the 316 LVM steel, and 0.073, for the Ti 6Al 4V alloy. This result is consistent with the lower hardness gradient of Ti 6Al 4V alloy with regards to that of 316 LVM steel.

3.4.

Fatigue behavior

Fig. 7 shows the sa–Nf curves for the blasted and non-blasted specimens and Table 5 summarizes the fatigue strength

BL-ZrO

360

BL-AlO

320 280

1200

True stress (MPa)

Microhardness HV 0.01

400

240 200 160 0

100

200

300

400

500

Depth (µm)

1000 800 600 400 ys

200

n (0,02-0,1) = 0,29

0 0.00

0.04

0.08

0.12

True strain (%)

460

BL-ZrO BL_AlO

440

1200

420

True stress (MPa)

Microhardness HV 0.01

= 430 MPa

400 380 360 340 320 0

20

40

60

80

900

1000 800 600 400 ys

200

n (0,02-0,1) = 0,073

0

1000

0.00

Depth (µm)

= 930 MPa

0.02

0.04

0.06

0.08

0.10

True strain (%)

Fig. 5 – Microhardness as a function of depth for (a) 316 LVM and (b) Ti 6Al 4V alloy blasted with alumina and zirconia particles.

Fig. 6 – True stress versus true strain under compression for (a) 316 LVM and (b) Ti 6Al 4V alloys.

Table 4 – Mechanical properties of as-received materials. Hardness (HV 0.01)

316 LVM Ti 6Al 4V

180 350

Yield strength (0.2%) (MPa) Tensile

Compression

405 911

437 925

Tensile strength (MPa)

Elongation (%)

700 999

68 20

Stress amplitude (%

0.2)


37

110 100 90 80 70 60

Polished BL-ZrO BL-AlO Empty symbols (not broken)

50 4 10

10

5

10

6

10

7

10

8

500 m

Stress amplitude (%

0.2)

Cycles to failure (Nf) 110 Polished BL-ZrO BL-AlO Empty symbols (not broken)

100 90 80 70 60 50 4 10

10

5

10

6

10

7

10

1 mm

8

Cycles to failure (Nf) Fig. 7 – Stress amplitude versus number of cycles to failure during rotating bending tests of (a) 316 LVM steel and (b) Ti 6Al 4V alloys. s 5 – Fatigue strength for the investigated materials in the non-blasted and after blasting with zirconia and alumina particles.

316 LVM Ti 6Al 4V

Un blasted

BL-ZrO

BL-AlO

400 640–660

400–420 700–720

370 480–490

they are connected with microcrack-like defects, pits, and/or embedded particles acting as stress concentrator.

4.

Discussion

Grit blasting of 316 LVM and Ti 6Al 4V alloy with alumina and zirconia particles is known to be an effective method to increase the roughness of load bearing components of interest in orthopaedic and dental applications. However, this process leaves contamination from the blasting material on the surfaces consisting of particles of the abrasive used. Besides the larger particles reported everywhere, our XRD results confirm the presence of a huge amount of small particles, which are originated in the blasting process. Penetration of all these particles into the substrate depends on the substrate hardness and coating material density, and the particle velocity. Because of the differences in size and mass between the alumina and zirconia particles used for blasting, both effect, the roughening of the substrate surface and the

Fig. 8 – SEI of fatigue fractures corresponding to alumina blasted specimens of (a) 316 LVM and (b) Ti 6Al 4V broken under 430 MPa (Nf of 95,000) and 600 MPa (Nf of 160,000), respectively. The arrows indicate place of the crack initiation.

size of zone under the surface, are more pronounced when alumina particles are used. Besides the biological effects of remnants of aluminum oxide on grit-blasted surfaces (Canabarro et al., 2008), implications of this surface modification in the fatigue behavior, which is the main cause of premature failure of metallic implants, become relevant. This investigation has shown that fatigue resistance increases when blasting with zirconia particles but considerably decreases when using alumina ones. Although the tendency is similar for both types of alloys, the beneficial or detrimental role of blasting is more intense for the Ti 6Al 4V alloy. For a meaningful understanding of this different fatigue behavior, the blasting induced effects and the mechanical properties of the bulk should be taken into consideration.

4.1.

Fatigue behavior of 316 LVM

316 LVM steel exhibits a high strain hardening exponent (n¼ 0.29) that leads to a high capability for strain hardening (Fig. 6a). This result has been associated in other austenitic steels to the formation of mechanical twinning (TWIP-effect) during deformation (Bäumer et al., 2010). As the volume fraction of twins increases with strain, the mean free path of dislocations decreases rapidly, causing high levels of strain hardening in the material. Therefore, the severe plastic deformation occurring during blasting causes a significant hardness increase, Fig. 5a, achieving values of hardness that beneath the blasted

38


surface are nearly twice than this of the bulk value ( 180 HV0.01), irrespectively the blasting particle. Interestingly, such hardness increase is significant if it is compared with those obtained with sophisticated techniques such as surface mechanical attrition treatment (SMAT) (Roland et al., 2007). From the analysis of the hardness gradient, Fig. 5a, it follows that the blasting affected zone for the BL-AlO specimens is larger than for the BL-ZrO specimens. Moreover, it can also be observed that for the same depth, hardness is higher for the BLAlO sample, except in the very near surface region where values are similar. These differences should be understood within the framework of the specific features that distinguish the ZrO2 and Al2O3 particles. Despite their lower density, the larger and harder alumina particles are more energetic and, therefore, lead to a more accumulative deformation that introduces a more extensive subsurface modification, but also a greater hardening. The fact that hardness beneath the blasted surface is similar for the BL-AlO and BL-ZrO specimens is explained when considering that the rounded ZrO2 particles barely lead to grinding down the material, whereas abrasion of alumina particles, characterized by edge-like facets, is more effective, as demonstrated when considering the volume lost (Table 1). The remnant material that was grinded with alumina particles has obviously experienced a lower number of impacts. Hardening at the blasted affected zone results from work hardening associated to the accumulative plastic deformation but also from the grain size refinement (Fig. 4). Additional contribution to hardening results from the presence of compressive residual stresses of up to about 500 MPa and 700 MPa for the BL-AlO and BL-ZrO specimens, respectively (Multigner et al., 2010). Compressive residual stresses were found to be the main hardening factor, irrespective of the particles used for blasting (Frutos et al., 2010). Subsurface hardening will contribute to a local increase of the yield strength (Pang et al., 2013), which in addition to the ultrafine-grained subsurface microstructure and the compressive residual stress would be beneficial for the fatigue resistance by impeding dislocation movement and delaying crack initiation (Roland et al., 2006). On the other hand, the roughness increase and the high mass fraction of grit embedded particles would play a detrimental role by acting as stress concentrators favoring crack initiation. The balance between beneficial and detrimental factors is positive for the BL-ZrO specimens because the roughness increase is moderated (Ra 1 μm) and the oxide particles becomes adhered rather than incrusted, as deduced by the mass increase after blasting (Table 1). The net effect is a moderated increase in the fatigue resistance from 400 MPa, for the un-blasted specimens, to about 420 MPa for the BL-ZrO specimens. In the case of blasting with alumina, however, the noticeable roughness increase (Ra 8 μm) makes a negative balance that is manifested by a decrease in the fatigue resistance in a nearly 10% with regards to the non-blasted condition (400 MPa). Interestingly, hardening of the surface during the successive impacts of the oxide particles decreased the amount of embedded particles, as shown in Table 3.

4.2.

Fatigue behavior of Ti 6Al 4V alloy

Fatigue resistance of the alloy represents about a 70% of the yield strength, increasing up to about 80% when blasting with

zirconia particles but considerably decreasing up to about a 55% when using alumina particles. The beneficial role of the zirconia particles are closely related on the one hand, to the moderated increase in roughness (Ra 1 μm) and adhered rather than embedded particles, and on the other hand to the development of compressive residual stresses (Multigner et al., 2009c) and moderated hardness increase ( 7%) in a narrow zone beneath the blasted surface. This beneficial role of the zirconia particles is consistent with previous results for shot penning of this alloy (Wagner and Lüetjering, 1982). Blasting with alumina particles, however, play a detrimental role despite the development of significant compressive residual stresses (Multigner et al., 2009c) and a somewhat higher hardness increase ( 17%) beneath the blasted surface. The loss of fatigue strength following blasting has been often associated to the presence of microcracks at the surface and to the pronounced stress concentration in the vicinity of the particles that remain embedded onto the surface (Leinenbach and Eifler, 2006; Baleani et al., 2000). The elevated roughness (Ra 5 μm) and high mass fraction of grit embedded particles (60–80%) would account for the loss of fatigue strength ( 25%). This finding is consistent with previous works that demonstrated that fatigue limit decreases with increasing roughness; about 20% decrease for fine blasting (Ra 2 μm) and up to about 40% for specimens with coarse blasting (Ra 7 μm) (Baleani et al., 2000). For the analysis it should be taken into consideration that blasting of cp Ti with Al2O3 or SiO2 particles, however, improves the fatigue life, irrespectively of its microstructure (Gil et al., 2007; Jiang et al., 2006). It has been argued that superficial cracks due to grit blasting have been observed in Ti 6Al 4V but not in cpTi, likely due to the higher ductility (Conforto et al., 2004). Other authors, however, have also observed such surface defects on blasted titanium, being the net effect beneficial (Gil et al., 2007). Comparative analysis with the fatigue behavior of the alumina blasted 316 LVM, processed under the same experimental conditions, reveals that the lost of fatigue strength is only about 5%, which contrast with the 20% found for the alumina blasted Ti 6Al 4 V. Therefore, the negative influence of stress raisers associated to the presence of microcracks, pits, or embedded ceramics particles cannot be discarded, but it not seem to be the only reason for the strong decrease in the fatigue limit of the alumina blasted Ti 6Al 4V. An interesting microstructural feature following sandblasting is that hardness was highest in the severely deformed near surface layer of cp Ti (Gil et al., 2007) and austenitic stainless steel 316L (Multigner et al., 2009b, 2010), but not in the Ti 6Al 4V alloy (Multigner et al., 2009a, 2009c). This study shows that the very small hardening at the outermost blasted affected zone of the Ti 6Al 4V alloy could be easily related to the low capability for strain hardening (n ¼0.073). This strain hardening exponent is similar to that of about 0.048 previously reported (Boller and Seeger, 1987). Such values contrast with the higher values determined for cp Ti (about 0.5) (Boller and Seeger, 1987) and 316L alloy (about 0.41) (Byun et al., 2001). The high levels of strain hardening in these alloys have been associated to the development of mechanical twinning during deformation (Bäumer et al., 2010).


Another feature that must be taken into consideration is the potential induced softening associated with the fatigue process at the surface layer. Let us consider that the compressive residual stresses at the surface play a beneficial role by increasing the tensile stress during fatigue. Similarly to the shot-peening process (Wagner and Lüetjering, 1984), sandblasting can be considered as a low cycle fatigue process to the surface layer in which blasting pressure can be described as the stress amplitude and the exposure time as the number of cycles in a fatigue test. Since residual stresses have been observed to decay with increasing the number of fatigue cycles (Wagner and Lüetjering, 1984), the negligible hardening will highlight the residual stresses relaxation. Relevant for this discussion is that whereas 316 LVM hardens when is cyclically loaded, as occur with the cp Ti (Boller and Seeger, 1987), Ti 6Al 4V would experience an opposite behavior (Conforto et al., 2004). Therefore, softening of the Ti 6 Al 4V alloy could be expected to occur during the fatigue process. It seems that the moderated subsurface hardening of the grit blasted Ti 6Al 4V alloy, the beneficial effect of the compressive residual stresses at the earlier stages of the fatigue process is offset by the presence of microcrack-like defects, pits, and the huge amount of embedded particles all them acting as severe notches. It has been also showed that microstructure of the material can play an important influence on the fatigue behavior of Ti 6Al 4V (Chan, 2010; Freudhental, 1974). As a matter of fact, Neal and Blenkinshop (1976) found subsurface flat α-facets along the cleavage { 1 0 1̄ 7} plane at the crack origin of Ti 6Al 4V fatigue specimens. These facets were originated as consequence of the constraint induced by the harder Widmanstatten αþβ colonies on the softer α-phase, as recently was modeled (Chan and Lee, 2008).

5.

Concluding remarks

From the above discussion it is concluded that fatigue resistance strongly depend on the topography and microstructure of the blasted affected zone, but also on the mechanical properties of the bulk. A high roughness and the presence of embedded particles enhance crack initiation, whereas crack propagation would depend on the effective tensile strength at the blasted affected zone resulting from work hardening and grain size refinement. Residual compressive stresses have little effect on crack initiation but drastically will contribute to retard the earlier stage of crack propagation. Softening of these residual stresses during cycling could be balanced by the capability of the alloy to harden during the alternative tensile/compressive stages.

Acknowledgements The authors are thankful for financial support from MICINN (MAT2009-14695-C04) and MINECO (MAT 2012-37736-C05-01). S.Barriuso to JAE predoc grant of CSIC. Technicians of Laboratories of X-Ray Diffraction, Mechanical Testing, and Microscopy (CENIM) are specially acknowledged.

39

r e f e r e nc e s

Aparicio, C., Gil, F.J., Thams, U., Muñoz, F., Padrós, A., Planell, J.A., 2004. Osseointegration of grit-blasted and bioactive titanium implants: histomorphometry in minipigs. Key Eng. Mater. 254–256, 737–740. Baleani, M., Viceconti, M., Toni, A., 2000. The effect of sandblasting treatment on endurance properties of titanium alloy hip prostheses. Artif. Organs 24 (4), 296–299. Bäumer, A., Jiménez, J.A., Bleck, W., 2010. Effect of temperature and strain rate on strain hardening and deformation mechanisms of high manganese austenitic steels. Int. J. Mater. Res. 101, 705–714. Boller, Chr., Seeger, T., 1987. Materials Data for Cyclic Loading. Part. D: Aluminium and Titanium Alloys. Materials Science Monograph, 42 D. Elsevier, Amsterdam127. Byun, T.S., Farrel, K., Lee, E.H., Hunn, J.D., Mansur, L.K., 2001. Strain hardening and plastic instability properties of austenitic stainless steels after proton and neutron irradiation. J. Nucl. Mater. 298 (3), 269–279. Chait, R., 1973. The strength differential of steel and Ti alloys is influenced by test temperature and microstructure. Scr. Metall. 7, 351–354. Chan, K.S., 2010. Roles of microstructure in fatigue crack initiation. Int. J. Fatigue 32, 1428–1447. Chan, K.S., Lee, Y.-D., 2008. Effects of deformation induced constraint on high-cycle fatigue in Ti alloys with duplex microstructure. Metall. Mater. Trans. A 39A, 1665–1675. Canabarro, A., Diniz, M.G., Paciornik, S., Carvalho, L., Sampaio, E.M., Beloti, M.M., Rosa, A.L., Fischer, R.G., 2008. High concentration of residual aluminium oxide on titanium surface inhibits extracellular matrix mineralization. J. Biomed. Mater. Res. Part A 87, 588–597. Conforto, E., Aronsson, B.O, Salito, A., Crestou, C., Caillard, D., 2004. Rough surfaces of titanium and titanium alloys for implants and prostheses. Mater. Sci. Eng., C 24, 611–618. Freudhental, A.M., 1974. New aspects of fatigue and fracture mechanics. Eng. Fract. Mech. 8 (6), 775–793. Frutos, E., Multigner, M., González-Carrasco, J.L., 2010. Novel approaches to determine residual stresses by ultramicroindentation techniques: application to sand blasted austenitic stainless steel. Acta Mater. 58, 4191–4198. Gil, F.J., Planell, J.A., Padrós, A., Aparicio, C., 2007. The effect of shot blasting and heat treatment on the fatigue behavior of titanium for dental implant applications. Dent. Mater. 23 (4), 486–491. Goldberg, V.M., Stevenson, S., Feighan, J., Davy, D., 1995. Biology of grit-blasted titanium alloy implants. Clin. Ortho. Relat. Res. 319, 122–129. Hong, T, Ooi, J.Y., Shaw, B., 2008. A numerical simulation to relate the shot peening parameters to the induced residual stresses. Eng. Fail. Anal. 15, 1097–1110. Jiang, X.P., Wang, X.Y., Li, J.X., Li, D.Y., Man, C.-S., Shepard, M.J., Zhai, T., 2006. Enhancement of fatigue and corrosion properties of pure Ti by sandblasting. Mater. Sci. Eng., A 429 (1–2), 30–35. Leinenbach, C., Eifler, D., 2006. Fatigue and cyclic deformation behaviour of surface modified titanium alloys in simulated physiological media. Biomaterials 27, 1200–1208. Multigner, M., Frutos, E., Mera, L., Chao, J., González-Carrasco, J.L., 2009a. Interrogations on the sub-surface strain hardening of grit blasted Ti-6Al-4V alloy. Surf. Coat. Technol. 203, 2036–2040. Multigner, M., Frutos, E., González-Carrasco, J.L., Jiménez, J.A., Marín, P., Ibáñez, J., 2009b. Influence of the sandblasting of 316 LVM steel on the subsurface microstructure: implications on

40


the magnetic and mechanical properties. Mater. Sci. Eng., C 29, 1357–1360. Multigner, M., Fernández-Castrillo, P., Ferreira-Barragán, S., González-Doncel, G., González-Carrasco, J.L., 2009c. Influence of sandblasting of Ti 6Al 4V alloy on the subsurface hardness and residual stresses state. Rev. Metal. 45, 52–55. Multigner, M., Ferreira, S., Frutos, E., Jaafar, M., Ibáñez, J., Marín, P., Pérez-Prado, T., González-Doncel, G., Asenjo, A., González-Carrasco, J.L., 2010. Superficial severe plastic deformation of 316 LVM stainless steel through grit blasting: effects on its microstructure and subsurface mechanical properties. Surf. Coat. Technol. 205, 1830–1837. Neal, D.F., Blenkinshop, P.A., 1976. Internal fatigue origins in α β Ti alloys. Acta Metall. 24, 59–63. Ortiz, A.L., Tian, J.W., Villegas, J.C., Shaw, L.L., Liaw, P.K., 2008. Interrogation of the microstructure and residual stress of a nickel-base alloy subjected to surface severe plastic deformation. Acta Metall. 56, 413–426. Pang, J.C., Li, S.X., Wang, Z.G., Zhang, Z.F., 2013. General relation between tensile strength and fatigue strength of metallic materials. Mater. Sci. Eng., A 564, 331–341. Roland, T., Retraint, D., Lu, K., Lu, J., 2006. Fatigue life improvement through surface nanostructuring of stainless steel by means of surface mechanical attrition treatment. Scr. Mater. 54, 1949–1954. Roland, T., Retraint, D., Lu, K., Lu, J., 2007. Enhanced mechanical behavior of a nanocrystallised stainless steel and its thermal stability. Mater. Sci. Eng., A 445–446, 281–288.

Sawada, H., 1994. Residual electron-density study of alphaaluminum oxide through refinement of experimental atomic scattering factors. Mater. Res. Bull. 29 (2), 127–133. Schuh, A., Uter, W., Kachler, W., Göske, J., Zeiler, G., Lill, C., 2005. Comparative surface examinations on corund blasted titanium implants and explants in total hip arthroplasty. Arch. Orthop. Trauma. Surg. 125, 676–682. Smith, D.K., Newkir, H.W., 1965. The crystal structure of baddeleyite (monoclinic ZrO2) and its relation to the polymorphism of ZrO2_database_code_amcsd 0009296. Acta Crystallogr. 18, 983–991. Wagner, L., Lüetjering, G., 1982. Influence of Shot Peening on the Fatigue Behaviour of Titanium Alloys. In: Shot Peening, A. (Ed.), Niku-Lari. Pergamon Press, pp. 453–460. Wagner, L., Lüetjering, G., 1984. Influence of the Shot Peening Parameters on the Surface Layer Properties and the Fatigue Life of Ti–6Al–4V, ICSP-2, G.P. Balear and H.O. Fuchs (Eds.), Chicago, IL, USA, 195. Wang, X.Y., Li, D.Y., 2003. Mechanical, electrochemical and tribological properties of nano-crystalline surface of 304 stainless steel. Wear 225, 836–845. Wennerberg, A., Albrektsson, T., Andersson, B., 1996. Bone tissue response to commercially pure titanium implants blasted with fine and coarse particles of aluminium oxide. Int. J. Oral Maxillofac. Implants 11, 38–45. Zhang, H.W., Hei, Z.K., Liu, G., Lu, J., Lu, K., 2003. Formation of nanostructured surface layer on AISI 304 stainless steel by means of surface mechanical attrition treatment. Acta Mater. 51, 1871–1881.

The influence of grain size on the fatigue behavior of annealed 316 LVM stainless steel.

Fatigue performance of medical Ti6Al4V alloy after mechanical surface treatments.

Opportunities for Fluorochlorozirconate and Other Glass-Ceramic Detectors in Medical Imaging Devices.

Investigation of the fatigue behavior of adhesive bonding of the lithium disilicate glass ceramic with three resin cements using rotating fatigue method.

A Review of Simulators with Haptic Devices for Medical Training.

Medical journals and conflicts of interest.

Centrifugal blood pump for temporary ventricular assist devices with low priming and ceramic bearings.

Fracture load of ceramic restorations after fatigue loading.

Potential conflicts of interest for academic medical center leaders.

Fatigue of Ti6Al4V Structural Health Monitoring Systems Produced by Selective Laser Melting.

Potential conflicts of interest for academic medical center leaders.

Interest and Inflation Risk: Investor Behavior.

Potential conflicts of interest for academic medical center leaders--reply.

Comparison of fatigue resistance and failure modes between metal-ceramic and all-ceramic crowns by cyclic loading in water.

An in vitro investigation of the anodic polarization and capacitance behavior of 316-L stainless steel.

Silicone surface with drug nanodepots for medical devices.

Development of nanotoxicology: implications for drug delivery and medical devices.

Conflict of interest in medical journals.

Prevention and control of bacterial infections associated with medical devices.

Metrology and Standards Needs for Some Categories of Medical Devices.

Laser heating of dielectric particles for medical and biological applications.

Levator syndrome: an analysis of 316 cases.

Sorting particles with nanoscale thermophoretic devices: how efficient is it?

Low-friction hydrophilic surface for medical devices.