Tso-Chng Wu ,Smeehn S.Joshi,b ,Yee-Hsien Ho,b ,Mngesh V.Pntwne,b ,Subhsis Sinh,c ,Nrendr B.Dhotre,b,*

aDepartment of Materials Science and Engineering,University of North Texas,3940 N Elm St,Denton,TX 76207,USA

b Center for Agile and Adaptive Additive Manufacturing,University of North Texas,3940 N Elm St,Denton,TX 76207,USA

c Department of Metallurgical Engineering,Indian Institute of Technology (BHU),Varanasi,Uttar Pradesh 221005,India

Abstract The present work explored effects of laser surface melting on microstructure and surface topography evolution in AZ31B magnesium alloy.Thermokinetic effects experienced by the material during laser surface melting were simulated using a multiphysics finit element model.Microstructure and phase evolution were examined using scanning electron microscopy,X-ray diffraction,and electron back scatter diffraction.Surface topography was evaluated using white light interferometry.The interaction of surface melted samples with simulated body flui was monitored by contact angle measurements and immersion studies up to 7 days.Laser surface melting led to formation of a refine microstructure with predominantly basal crystallographic texture.Concurrently,the amount of β phase (Mg17Al12) increased with an increase in the laser fluence β phase preferentially decorated the cell boundaries.In terms of topography,the surface became progressively rougher with an increase in laser fluence As a result,upon immersion in simulated body fluid the laser surface melted samples showed an improved wettability,corrosion resistance,and precipitation of mineral having composition closer to the hydroxyapatite bone mineral compared to the untreated sample.

Keywords: Laser surface melting;Magnesium alloy;Laser surface engineering;Biomineralization.

1.Introduction

Biodegradable magnesium (Mg) alloys have gained considerable interest for consumable implant application due to their appropriate properties,such as low density (1.74 g/cm3),elastic modulus similar to natural bone,and biocompatibility[1-6].Recently,Mg alloys have been developed for several implant devices such as cardiovascular stents,bone fixation and porous scaffolds [7,8].Although,biodegradable Mg alloys are potentially attractive for the orthopedic field there are number of issues which limit the development of Mg based biomaterials.The primary issue with Mg based materials is their propensity to accelerated corrosion,especially in chloride ion containing physiological environment [9-15],resulting in the deterioration of mechanical properties of implant before the host bone tissue has been sufficientl healed.Furthermore,the evolved hydrogen bubbles during corrosion reaction can form gas pockets in the vicinity of the implant,leading to separation of tissue layer [16].Moreover,the released Mg ions dissolved into the surrounding blood plasma can induce localized alkalization near the implant leading to damage of the cells and the regenerated tissue [17].These phenomena become a challenge and demand quite an effort to resolve the associated drawbacks in order to increase the success rate of implantation.Hence,it is critical to reduce and control the degradation rate of Mg based implant in physiological environment.

It has been reported that the corrosion resistance of engineering alloys,including Mg alloys,is significantl influence by the microstructural features,such as grain size,crystallographic texture (orientation),and chemical composition [18].Besides the improved corrosion properties,a favorable surface (physical texture/roughness) to enhance cell interaction in physiological environment is also essential for a successful implant [19].In order to efficientl reduce degradation rate and improve interfacial biocompatibility of Mg based implants,various surface modification have been developed.The methods of surface modificatio include chemical modification such as plasma electrolytic oxidation [20] as well as microarc oxidation [21],and physical modification such as heat treatment [22],friction stir processing [23-25],and laser processing [26-29].Among these methods,laser offers a rapid and reproducible processing solution for the implant material.Laser provides a non-contact and clean high energy density source by focusing high power concentrated into a small diameter (few microns) spot on material surface.Laser processing can treat site specifi surfaces to have a spatial control on the surface properties without disturbing the bulk material.In addition,due to use of lasers such as Nd-YAG laser with an optical fibe based beam delivery,laser processing is highly compatible with automation and remote processing in intended manufacturing environment and can treat the contoured surfaces further enhancing these advantages.Laser beam generates a small-scale localized interaction zone with the material (few hundredμm depth) which leads to a self-quenching effect (104-105K/s) occurring due to rapid heat extraction from the bulk of the substrate producing a refine microstructure [30,31].These thermophysical effects can be employed for various end effects such as material heat treatments via phase transitions,surface melting (LSM)for producing physical texture,and material removal for machining [4,31-33].By controlling various parameters,such as power,scanning rate,and laser track overlap;the process can be optimized to produce a microstructure and suitable surface roughness on the implants to improve the biocompatibility and biointegration characteristics [27,34].Several studies revealed the refine grain structure within Mg alloys surfaces after laser surface melting (LSM) [27,28,35-37].Liu et al.used 10 kW CO2laser to carry out surface melting of AM60B Mg alloy and reported the homogeneous distribution of secondary-phase with refineα-Mg grains[35].In previous studies from the present research group [27,28],the optimal laser fluenc was investigated to produce enriched Mg17Al12βphase at theα-Mg grain/cell boundaries in AZ31B Mg alloy for the enhanced corrosion resistance.Rakesh et al.studied[38] the positive correlation between surface roughness and wettability of laser processed Mg-Zn-Dy alloy.These find ings from the literature indicated that laser process provides multiple advantages to produce Mg based implant with suffi cient corrosion resistance and biocompatibility.

Even though the previous studies pointed towards potential of LSM for Mg based implant materials,there were gaps such as understanding about evolution of crystallographic texture and exact nature of mineral deposited upon immersion in simulated body flui (SBF).As a continuation of the investigation and development of Mg based bioimplant materials by the present research group [24,25,27,28,39],the current work focused on probing into the thermokinetic effects and the influenc of laser processing on microstructure evolution in AZ31B Mg alloy using an integrated experimental and computation approach.The thermal profile experienced by the material were simulated using a multiphysics computational model.The phase and microstructure evolution were probed in the laser surface melted zone using analytical techniques to get an idea about microstructure and phase evolution,development of crystallographic texture,and surface roughness in the LSM samples.LSM samples were subjected toin-vitroevaluation in the SBF using contact angle measurements and immersion studies.Post immersed samples were observed using optical and electron microscopy.Furthermore,to get an idea about precipitated mineral compounds upon immersion,X-ray photo electron spectroscopy (XPS) was utilized.The present study shed a light on combined effects of various aspects associated with the laser processing to deepen the understanding about utilizing laser as an effective processing tool to produce improved Mg based bioimplant materials.

2.Experimental procedures

2.1.Materials and sample preparation

Commercially available AZ31B Mg alloy (composition:3.0 wt% Al,1.0 wt% Zn,0.5 wt% Mn,and Mg balance) was used for this study.Prior to laser surface modification the AZ31B Mg sheets were cut using slow speed diamond saw into the rectangular blocks of dimension 50 mm × 50 mm ×7 mm (width × length × thickness) (Fig.1).All the coupons were then ground using silicon carbide (SiC) abrasive papers from 60 grit to 400 grit to remove the oxide layer and contamination.The coupons were then cleaned by rinsing with water and methanol.

2.2.Laser surface melting (LSM)

A 3 kW diode pumped Ytterbium fibe laser in continuous mode (IPG Photonics,model# YLS-3000,wavelength of 1064 nm) was employed during this study to conduct LSM.The laser beam had a Gaussian profil (TEM00mode) with a diameter of 0.6 mm on the sample surface.The focal length of the laser was 120 mm.Such a long laser focal length ensured minimal/no deviation in spot size which remained fully focused on the sample surface.The laser scanning speed of 500 mm/s and laser track fil spacing (distance between centers of consecutive neighboring track) of 0.15 mm were kept constant during all the laser processing experiments while the laser power was varied in the range of 250-750 W (Table 1).These parameters were chosen based on several trials and optimization studies previously published by the present research group [27,28,39].Ar(g)was used as the shielding gas during all the laser processing experiments.The processed samples were then be used for microstructure characterization,contact angle measurements and immersion studies in SBF.

Fig.1.Schematic of laser surface melting (LSM) process.

Table 1 LSM parameters used to process AZ31B Mg alloy.

2.3.Phase identificatio

Phases in untreated and the LSM samples were identifie using X-ray diffraction (XRD) technique.A Rigaku Ultima diffractometer with Cu Kαradiation (wavelength -0.15418 nm),operating at 40 kV and 40 mA was employed to carry out XRD analysis.Samples were scanned in the 2θrange of 20?to 90?using a step size of 0.025 and a scan rate of 1.5?/mm.Phase identificatio was carried out by comparing the XRD spectra with standard International Center for Diffraction Data (ICDD) file obtained from the Joint Committee of Powder Diffraction Standards (JCPDS).

2.4.Microstructure characterization

Microstructure examination was conducted in the crosssection perpendicular to the laser track and laser treated surface.The cross-sections of samples were cut into smaller rectangular blocks (5 mm × 3 mm × 7 mm) using a slow speed diamond wafer cutting machine with oil based lubricant to achieve stress free cut.The cut samples were mounted in epoxy molds.These mounted samples were ground using series of (120-1200 grit) SiC abrasive papers,followed by fin surface polishing in 1-0.03 μm alumina (Al2O3) water-based abrasive suspension.A smooth mirror finis on the polished surface was obtained by employing 0.02 μm colloidal silica(SiO2) suspension for 40 min.The finishe polished samples were then cleaned by ultrasonic setup in succession with deionized water and methanol as the media.Prior to the microstructural investigation,the prepared samples were etched by the acetic picric solution (5 ml acetic acid+6 g picric acid+100 ml ethyl alcohol+10 mL H2O) to denote the microstructural features.The etched coupons were characterized using scanning electron microscope (SEM) equipped with Energy-Dispersive X-ray Spectroscopy (EDS) (FEI Quanta 200 environmental SEM).The crystallographic orientation of laser treated samples on the surface and in the cross-section were observed and analyzed by using FEI Nova NanoSEM 230 equipped with a Hikari super electron backscatter diffraction (EBSD) detector (operated at 20 kV).The EBSD scanning step size was 0.5 μm.TSL OIM Analysis 8.0 software was used to obtain microstructure maps and pole figure from the EBSD data.

2.5.Numerical modeling

The main purpose of implementing an integrated experimental and computational approach during the current work was to realize the microstructure evolution in LSM AZ31B alloy and its effect on the response of the LSM AZ31B Mg alloy in SBF environment.Laser is a rapid process characterized by μs-ms laser beam-material interaction times and small laser spot sizes in the range of μm-mm.Complex physical phenomena such as plasma and plume formation occur during these time frames in a small region making it difficul to monitor the laser beam-material interaction and carry out experimental measurements of temperature evolution experienced by the material.Material rapidly undergoes solid state heating,melting,vaporization,and sever cooling within the short time frames (μs-ms) which leads to a refine and novel microstructure.In light of this,authors modeled the LSM process using a multiphysics process model developed on a COMSOLT Mplatform to gather better understanding about the rapid time-temperature cycles experienced by AZ31B material while undergoing LSM treatment.The model mainly dwells on thermokinetics involved during laser-material interaction while considering the complex hydrodynamics associated with the melt pool.The development of the model detailing the governing equations,assumptions,boundary conditions,meshing,and model configuratio is adopted from the earlier reports by the authors [27,40].

2.6.Topographic characterization

The surface roughness of untreated and LSM AZ31B Mg samples were measured with 10X objective using white light by the Rtec Universal Tribometer (Rtec Instrument,San Jose,CA,USA).The detected surface area was 1.3 mm× 1.6 mm.The surface profil data was fi as per the standard protocol according to the International Organization for Standardization [41].Advanced data processing was conducted by a modular program software,Gwyddion 2.50.As a result,sample surface roughness was presented in form of the arithmetic mean height (Ra,μm),the root mean square height (Rz,μm),as well as the topographic maps.

2.7.Wettability

The surface wettability of untreated Mg and LSM samples was evaluated by a static liquid sessile drop method using CAM-Plus? contact angle goniometer (Cheminstruments Inc.Fairfiled OH),equipped with fibe optic light source.Freshly prepared SBF was used as testing medium,and the preparation process has been described in the previous publication [39].A liquid droplet of volume 3 μL (drop diameter of～1 mm) was placed on the thoroughly cleaned sample by a hypodermic syringe.The liquid droplet was placed on the surface for approximately 10 s to stabilize before the reading was taken.The measurement was carried out at room temperature (25.0?C) and a minimum of ten contact angle readings were taken on each sample at various locations to minimize errors in the measurement.In order to eliminate the error associated with the arbitrary tangential alignment,the contact angle was measured based on the patented half angle method(US Patent 5268733).

2.8.Biomineralization evaluation

To examine the biomineralization behavior,the untreated and LSM AZ31B Mg samples were cut into coupon of 5 mm× 5 mm and mounted in epoxy.The cut coupons were immersed in SBF maintained at a constant temperature of 36.5?C with the aid of constant temperature water bath equipped with a temperature sensor and digital temperature read out.The samples were placed vertically in a 250 ml beaker with SBF.The volume of SBF for the immersion study was determined by Vs=Sa/10,where Vsis the volume of SBF (mL) and Sais the surface area of the coupon (mm2).The periods of immersion time were setup as 1 day,3 days,5 days,and 7 days.In order to maintain a pH value of the SBF solution at 7.4,flui was refreshed every 24 h.After immersion,the samples were removed from SBF solution and rinsed with distilled water and then dried at room temperature.The morphological and elemental characterization of the surface mineralization was conducted by SEM and EDS,and the chemical compound analysis was conducted using Xray photoelectron spectroscopy (XPS) (make:VersaProbeTM5000).The monochromatic X-ray beam source at 1486.6 eV,49.3 W,and 200 μm beam diameter was operated in the vacuum (5 x 10-6Pa) to perform scans on the mineralized surface for the analysis.

2.9.Corrosion behavior

The corrosion behavior of untreated and LSM AZ31B Mg samples was studied using weight loss measurement.It is worth mentioning here that the electrochemical corrosion behavior was investigated in detail in previous report published by the current research group[28].The experimental setup for weight loss measurement was the same as that used in biomineralization evaluation described above.However,in this case,post-experimental sample cleaning process was different from that in the biomineralization testing.The immersed samples were firs cleaned with boiling chromic acid (180 g/l) to remove mineralization products and surface corrosion layer.Then rinsing with adequate methyl alcohol was performed to wash out the residual acid followed by air drying at room temperature.The weight loss is deduced as per the following Eq.(1) [42,43]:

According to the measured weight loss (W),the corrosion rate can be further derived as expressed in Eq.(2):

WhereApresents the original surface area of the sample(cm2),Tis immersion time (h),andDis the density of the sample (g/cm3).

3.Results and discussion

3.1.Phase identificatio

Fig.2.X-ray diffraction spectra for untreated and LSM AZ31B Mg samples.

Phase analysis conducted using XRD revealed the peaks corresponding to solid solution and intermetallic phases in untreated and LSM AZ31B Mg Samples (Fig.2).Theα-Mg phase with hexagonal close packed (HCP) structure was dominant phase whereas minor peaks corresponding to an intermetallicβphase (Mg17Al12) were also observed in both untreated and LSM AZ31B (Fig.2).Theβphase peaks with higher intensity were observed among the laser processed samples.With an increase in the laser fluence there was a progressive increase in the intensity of peaks associated withβphase as reported in previous papers by the authors[27,28].Additionally,the peaks corresponding to theα-Mg phase were shifted toward higher angles by～0.2?for LSM samples.Furthermore,an increase in peak width was observed for the laser treated samples suggesting a refinemen of the microstructure[44].The change in the intensity ofβphase was likely to be an outcome of changing volume fraction ofβphase,which,in turn,can be due to changing composition and thermokinetics of the melt pool with varying laser fluence This increase in volume fraction ofβ-Mg phase was presented in previous publications by the authors along with the detailed methodology of the calculations [27,28,45].The calculated volume fraction ofβphase did increase with an increase in the laser fluenc (volume percent range<0.5% to 3.3%).To further probe into these initial observations about microstructure refinemen and phase evolution,it was necessary to understand the thermokinetic effects experienced by the samples during LSM.Thus,a transient temperature distribution was computationally predicted and analyzed.

The transient temperature computationally probed at the surface of the melt pool corresponding to the different laser fluenc varying from 1.06 to 3.18 J/mm2(250 to 750 W) at a constant scanning rate (500 mm/s) is presented in Fig.3.The thermokinetic parameters such as heating rate,cooling rate,and peak temperature obtained from these temperature-time plots are listed in Table 2.Due to increased laser fluence the heating rate can be seen to increase (0.748 to 1.952 ×105K/s)with laser fluenc (Table 2).The peak temperature of the melt pool surface was reached at the end of residence timetr=D/V=1.2 ms,where D is the laser beam diameter on the sample surface and V is the laser beam scanning velocity.As the laser fluenc increased,the peak temperature increased from 1359 K (at 1.06 J/mm2) to 2955 K (at 3.18 J/mm2).Vaporization temperatures of Mg,Al,and Zn which are the key constituent elements of AZ31B alloy,are 1364 K,2743 K,and 1180 K respectively[46].The processing condition corresponding to laser fluenc of 1.06 J/mm2led to surface melting of the alloy and reaching the temperature for very short duration of time (<0.5 ms) above the vaporization temperature Zn (Fig.3).Thus,the composition likely remained unaffected as the material experienced rapid thermokinetics driven melt pool dynamics.At the laser fluenc of 2.12 J/mm2,the peak temperature of 2192 K surpassed vaporization temperature of both Zn and Mg (Fig.3).Finally,for the laser fluenc of 3.18 J/mm2,the temperature exceeded vaporization points of Zn,Mg,and Al (Fig.3).

Fig.3.Computationally predicted variation in time-temperature profile for all LSM AZ31B Mg samples.Vaporization temperature (Tv) for Zn,Mg,and Al is also shown.In addition,the region of mushy zone between solidus(Tsolidus) and liquidus (Tliquidus) is shown.

Table 2 Computationally derived thermokinetic parameters.

Fig.4.SEM BSE mode micrographs in cross sectional view (XZ plane) of (a) Untreated AZ31B Mg,(b)-(c) low magnificatio views corresponding to 3.18 J/mm2 revealing a cellular morphology and (d) high magnificatio view corresponding to 3.18 J/mm2 sample showing underlying equiaxed grain structure as a result of channeling contrast.

To analyze the peak shifts observed in XRD spectra(Fig.2),the finding from the model were coupled with the experimental observations to better realize the phase evolution during LSM of AZ31B Mg alloy.The lattice parameters forα-Mg andβphase were calculated from the XRD spectra of LSM and untreated AZ31B Mg alloy.The values of lattice parameters were reported in Fig.2.As a firs observation,the value of lattice parameter ofβphase remained identical in all the cases.βphase is an intermetallic compound with a fi ed composition at room temperature[47].However,the lattice parameters ofα-Mg appeared to increase for 1.06 J/mm2LSM sample compared to the untreated AZ31B Mg alloy(Fig.2).The lattice parameters then steadily decreased at the laser fluenc of 2.12 J/mm2and became comparable to the untreated AZ31B Mg alloy for the sample processed with laser fluenc of 3.18 J/mm2(Fig.2).Such a variation in lattice parameter can be attributed to factors such as solute concentration of alloying elements in the matrix and micro residual stresses which can change the inter planar spacing [48].For LSM sample processed with 1.07 J/mm2,as the melting occurred with temperature exceeding vaporization of Zn for a short time (Fig.3),it is possible that density driven segregation of heavy elements such as Zn would occur towards the depth of the melt pool of liquid AZ31B Mg alloy.Simultaneously,minor quantities of Zn would vaporize from the surface layer.This would leaveα-Mg matrix lean of Zn in the surface region.It has been reported that the lattice parameters of Mg increase with a decrease in solute contents of Zn and Al [49].With an increase in laser fluenc to 2.12 J/mm2the AZ31B Mg alloy experienced temperature well above vaporization of Zn and Mg(Fig.3).This would have possibly led to higher lattice parameters forα-Mg compared to the untreated AZ31B Mg alloy[49]as shown in Fig.2.Lastly,for the LSM AZ31B Mg sample corresponding to 3.18 J/mm2laser flu ence,the temperature exceeded vaporization of Zn,Mg,and Al(Fig.3).The computed lattice parameters for this condition were similar to the untreated AZ31B Mg alloy (Fig.2).It is likely that effect of localized variation in chemistry occurring due to redistribution and evaporation of constituent elements driven by thermokinetics and complex physical phenomena such as Marangoni convection during LSM can affect the lattice parameters ofα-Mg phase,which is reflecte as the shift(～0.2?) in the peaks associated withαphase in the XRD spectra of the LSM samples (Fig.2).

The cooling rate was mainly computationally determined in the solidificatio range (839 K-905 K) which dictates the size of the microstructural features.The highest cooling rate of 5.283 ×104K/s was computed at the laser flu ence of 1.06 J/mm2,which decreased up to 4.481×104K/s at 3.18 J/mm2.The increased size of the melt pool with increased laser fluenc lowers the rate of heat extraction by the surrounding solid mass,causing reduction in cooling rate[32,40].Such a rapid thermokinetics,in addition to the unique combination of other auxiliary parameters such as thermal gradient,and solidificatio rate up to 500 mm/s,produced distinct microstructural morphology which was further probed using SEM and EBSD.

The SEM micrographs of untreated AZ31B mg in back scatter electron (BSE) mode mainly revealed the coarse equiaxedα-Mg grains with an average grain size of 10±1.2 μm (Fig.4(a)).On the other hand,the LSM AZ31B Mg viewed in XZ plane (Fig.4 (b)-(d)) mainly displayed the occurrence of solidificatio under cellular mode throughout the melt pool.The cellular structure was highly refine compared to the microstructure of the untreated sample (Fig.4(a)).Interestingly,even with the presence of cellular structure,the underlying grain structure was clearly visible as a result of channeling contrast when observed carefully at a higher magnificatio (Fig.4 (c)) for the LSM AZ31B Mg samples.Moreover,the EDS elemental point probe indicated the segregation of Al along the intercellular boundaries as reported in previous papers by the authors [27,28].The Al concentration in this intercellular region was in the range of 3.5 to 4 wt%,which decreased to～1 to 2 wt% in the intracellular region.Thus,βphase is likely to occur in the intercellular or aluminum rich region.As the cooling rate decreased with increased laser fluence the more aluminum likely to diffused in the intercellular region thereby increasing theβphase formation (Fig.2).Additional insights into grain morphology and crystallographic texture were gained using EBSD.

Fig.5.EBSD results showing (a) inverse pole figur map for top surface (b) 0001 pole figur for the top surface (c) image quality map for the top surface,(d) inverse pole figur map for cross-section,(e) 0001 pole figur for cross section,and (f) image quality map for the cross-section.

The EBSD analysis performed on top surface and in the cross-section uncovered the nature of grain structure and the orientation effects in these planes (XY and XZ planes respectively)(Fig.5).The top surface grain structure had an average grain size of 2.5 μm as indicated by the inverse pole figur(IPF) maps from the XY plane is nearly 5 times smaller than grain size (～10 μm) of untreated AZ31B Mg material (Fig.5(a)).Corresponding 0001 pole figur texture plot and image quality (IQ) map are presented in Fig.5 (b)-(c) respectively.On the same lines,the IPF maps from the cross-sectional XZ plane with corresponding 0001 pole figur texture plot and image quality (IQ) map are presented in Fig.5 (d) and (f) respectively.Although,the average grain size of LSM AZ31B Mg in this section (XZ and XY plane) is smaller than that in untreated region (Fig.5 (a) and (d)),the grains are relatively smaller in the top region compared to those in the bottom region near the interface with the underlying base microstructure.This is likely as the thermokinetic parameters vary within the melt pool,the solidificatio rate near the bottom of the melt pool is lowest (1-10 mm/s) and increases rapidly towards the top and attains its maximum solidificatio rate(500 mm/s)near the surface.The laser processing usually results in high thermal gradients (105-107K/m) mainly directed towards laser heat source (along Z direction) [40].This produces strong crystallographic texture in LSM region.In the bottom region of the LSM (Fig.5 (d)-(e)),with competitive epitaxial growth accompanied by the highest thermal gradient along Z direction,the grains grow in the<0001>direction,which is the highest thermal conductivity direction for HCP structure of AZ31B Mg.Consequently,prismatic plane direction appears to lie in the cross sectional XZ plane.This is further confirme by 0001 pole figure (Fig.5 (d)-(e)) that clearly show a fibe texture with peak intensity away from the basal in the cross-section (XZ plane) and corresponding higher intensity of basal poles in the top section (XY plane).Along the same lines,majority of grains on XY plane were aligned close to basal plane (0001) (Fig.5 (a)-(b)).Further evaluations indicated that 40% of the grains in the XY planes had (0001) texture.Therefore,the results demonstrated that the laser surface modificatio did not only evolve a distinct cellular grain morphology but it also influence crystallographic texture in AZ31B Mg alloy.Such a distinct crystallographic texture with unique microstructural morphology is likely to have an effect on various properties such as corrosion resistance and biomineralization,which is discussed in the subsequent sections.Apart from crystallographic texture,the surface topography/texture also influence the response of an implant material to a physiological environment.In light of this,as one of the preceding step to thein-vitroevaluation,surface roughness and topography of the samples were examined.

3.2.Topographic characterization

The three-dimensional surface morphologies of untreated AZ31B coupon and three AZ31B Mg samples processed with various laser process conditions are presented in Fig.6.The smoothest surface corresponded to the lowest laser fluenc sample(Fig.6(a)).The roughness information of samples was expressed in area root mean square height (Rz)and area arithmetic mean height

Fig.6.Three-dimensional surface morphology of (a) untreated AZ31B Mg,(b) 1.06 J/mm2, (c) 2.12 J/mm2, and (d) 3.18 J/mm2 LSM AZ31B Mg sample.

(Ra).The overall surface roughness of the samples increased with an increase in the laser fluenc (Fig.6).The higher laser fluenc incident on the surface results in deeper melting depths.In addition,the temperature profil via computational modeling further indicated the higher vaporization occurrence with higher laser power treatment (Fig.3).The peak temperatures of 2057 K and 2728 K for 3.18 J/mm2and 2.12 J/mm2laser fluence respectively were higher than the evaporation temperature (Tv) of Mg (1373 K) in Fig.3.Therefore,a portion of material will be evaporated from the surface thereby affecting both the roughness and chemical composition within the surface/subsurface region via physical phenomena such as flui fl w and recoil pressure effects[33,50].The maximum temperature of 1285 K for the lowest laser fluenc processing (1.06 J/mm2) was lower than the Tv,hence generating lowest roughness profil among the three different laser processing conditions employed in the present work.All the other conditions experienced temperature higher than Tv.Evolution of such a rough surface topography is expected to add to the surface energy.High energy surfaces are deemed suitable for the bioactivity of the implant as they can have better affinit to the plasma.In order to evaluate the combined effect of crystallographic texture,phase evolution,and surface texture;the samples were subjected toin-vitroevaluation as explained in the following subsections.

3.3.Wettability in SBF

The contact angle in SBF steadily decreased with an increase in the laser fluenc (Fig.7).Moreover,there was an inverse correlation between surface roughness (Ra) and SBF contact angle (Fig.7).These observations indicated that surface became hydrophilic for higher laser fluence Such as modifie surface would be suitable for bio-integration and cell attachment [27,51,52].It is worth mentioning here that the surface wetting behavior is dictated by surface energy of the sample under consideration.Surface energy also bares an inverse relationship with contact angle wherein a hydrophilic(low contact angle) sample possesses a relatively higher surface energy [27].Roughness,chemical composition,and microstructure (grain size,crystallographic texture,and phase make up) are some of the key factors which influenc the surface energy and thus the wetting behavior of a material[25,27,29].Based on the discussion presented in previous sections,for all the LSM AZ31B Mg samples the grain size(～1.5 μm) and crystallographic orientation (predominantly basal plane on the top surface) were nearly same due to partially narrow range of thermokinetic conditions experienced by these samples under processing conditions employed in the present work.However,there were two key differentiating factors among the LSM samples:i) surface roughness and ii) fraction ofβphase both of which increased as a function of laser fluenc likely due to composition variation (Al enrichment) experienced at higher laser fluences

Fig.7.SBF contact angle and roughness parameter as functions of laser processing condition for both untreated and LSM AZ31B Mg samples.

Fig.8.Surface morphology of untreated and LSM AZ31B Mg samples after various immersion periods in SBF.

For a biomaterial,SBF contact angle (wettability) is related to cell attachment and growth.Reports indicated that the cell adhesion and growth is not favorable for both highly hydrophobic and hydrophilic surface [52,53].Arima et al.reported the optimal contact angle for cell adhesion is in the range 40?-60?[51].Webb et al.pointed out the hydrophilic surfaces with contact angle 20?-40?promoted the highest levels of cell attachment [54].It is clear that cell growth and adhesion are favored with higher surface wettability.To examine the influenc of wetting behavior,samples were subjected to immersion studies in SBF and were observed at regular intervals as presented in following subsections.

3.4.Biomineralization

Examination of nature of mineral precipitated on a material upon immersion in SBF is a good indicator of biointegration tendency of an implant material.As a firs level of observations,surface morphology of samples was examined optically and differences were noted as a function of immersion time in SBF (Fig.8).Untreated AZ31B Mg after 1-day immersion displayed several localized pits due to corrosion.Furthermore,the precipitated products did not have a uniform distribution onto the surface.In addition,the severity of pitting increased with time leading to separation of the precipitates from the surface dissolving back into the surrounding SBF solution(Fig.8).

On the contrary,LSM AZ31B Mg samples experienced much lesser pitting compared to the untreated AZ31B Mg sample during all the time periods of immersion (Fig.8).In addition,the precipitates appeared to remain adhered to the surface with greater stability.While comparing all laser processed samples together the samples processed with higher laser fluence appeared to mineralize better (Fig.8).It has been reported that a refine microstructure apparently stabilizes the layers of deposits [55].Furthermore,the increased fraction of grain boundaries would also contribute to the higher surface energy as indicated in the previous publication by current research group [27].

The nature of mineral deposited is also a critical parameter which would facilitate cell attachment and growth on an implant surface.The mineralized surfaces were observed using SEM+EDS and XPS techniques.Observations of SEM+EDS suggested predominant presence of Mg and O on the as mineralized surface of untreated AZ31B Mg sample (Fig.9 (a)).In contrast,the peaks of Ca and P are strong for LSM AZ31B Mg samples indicating a shift in composition of the mineralized products (Fig.9(b-d)).As observed in previous studies,the Mg rich deposit is associated with a plate like morphology,and the Ca and P rich deposit with a globular morphology [27,29].

It is worth mentioning here that SEM+EDS is a semi quantitative technique and cannot accurately determine the exact nature of the precipitated mineral.This was also a gap in the previously published work related to laser treated Mg based implant materials by the current group [27,29] where in SEM+EDS was used to semi quantitatively evaluate the mineralized surfaces.To fil in this gap and expand the understanding about mineralized products further,XPS technique was utilized to obtain the exact nature of the mineral products (Fig.10).The Ca element split into two peaks of 2P1/2and 2P3/2attributed to the spin orbital splitting,and attributed to compound Ca10(PO4)6(OH)2(calcium hydroxyapatite,HA) [56] (Fig.10(a)).Even though,this peak splitting of Ca was observed for all the samples,the two peaks for higher laser fluence were fully resolved (Fig.10(a)).These peaks matched well with Ca2+from HA mineral as shown by fitte results in Fig.10(b).However,peaks shifted toward lower binding energy as observed for both untreated and 1.06 J/mm2laser fluenc samples,suggesting precipitation of multiple types of calcium apatite (eg.octacalcium phosphate and dicalcium phosphate) [57,58].

Fig.9.SEM surface morphologies of the (a) untreated AZ31B Mg,(b) 1.06 J/mm2, (c) 2.12 J/mm2, and (d) 3.18 J/mm2 after 7 days immersion in SBF along with corresponding EDS spectra.

Fig.10.XPS analysis of apatite precipitated on surface of both untreated AZ31B Mg and all LSM samples after 7-days immersion in SBF:(a) Ca 2P (b)Curve fittin of Ca 2P showing matching with calcium phosphate Ca bonds,(c) O 1s peaks (d) curve fittin and deconvolution of O 1S showing calcium phosphate O-2 bonds for high laser fluenc samples and OH- bond for untreated AZ31B Mg sample indicating hydroxide formation,(e) P 2P peaks (f) curve fittin P 2P peaks showing P bonds matching with calcium phosphate for LSM AZ31B Mg samples,and (g) Ca/P atomic ratios for various samples.

Oxygen appeared as twin peak for untreated AZ31B Mg samples revealing the presence of a shoulder peak(Fig.10(c)).On the other hand,such a shoulder peak was absent for LSM samples AZ31B samples (Fig.10(c)).The broader peak for untreated sample was deconvoluted to resolve the nature of oxygen bonding (Fig.10(d)).Observation of the resolved peak for untreated AZ31B Mg sample indicated presence of both OH-bonding state of oxygen and the presence of O2-found in HA mineral (Fig.10(d)) [59].The precipitates in LSM AZ31B Mg samples predominantly possessed O2-bonding state (as an example for 3.18 J/mm2in Fig.10(d)).

The bonding state of element P in precipitates of the untreated AZ31B Mg sample is associated with the other apatite other than HA mineral (Fig.10(c)).On the contrary,with presence of a small shoulder peak with the bonding phases of P,both HA-like and non-HA like phases appeared to be present in the precipitates of the AZ31B Mg samples treated with low laser fluenc of 1.06 J/mm2(Fig.10(b)).AZ31B Mg sample processed with 3.18 J/mm2laser fluenc upon deconvolution of the P peak revealed the combinations predominantly consisting ofalong with presence(Fig.10(d)) [59].This is an indication that the mineralized surface of laser melted samples primarily consisted of HA along with minor amounts of other types apatite.

As a concluding examination of XPS results,Ca/P peak intensity ratio was calculated from the spectra.The AZ31B Mg samples treated at 2.12 J/mm2and 3.18 J/mm2were associated with a much closer value to stoichiometric HA mineral(Ca/P:1.67) confirmin the individual elemental peak analyses.On these lines,the untreated AZ31B Mg sample had a much lower value of Ca/P ratio illustrating formation of nonstoichiometric HA like minerals and hydroxides upon immersion in SBF solution.It is worth noting here that apart from HA there are 7 different types of apatite minerals with Ca/P ratio in the range of 1-2 [57,58].

Apart from observations of the mineralized surface and understanding the chemical nature of the deposited products,measuring the weight loss as a function of time is an important indicator of the implant performance in the physiological environment.The mineralized samples were cleaned using chromic acid and their weight loss was measured to calculate the corrosion rate.In case of untreated AZ31B Mg sample,a continuous weight loss was due to the hydration of surface via formation of a hydroxide/oxide layer that in turn simultaneously continued to dissolve in surrounding SBF (Fig.11).It has been reported that the Mg hydroxide is especially unstable in chloride ion solution such as SBF [10].On the other hand,the corrosion resistance of laser processed AZ31B Mg samples was enhanced with immersion period as indicated by a negative slope (Fig.11).This behavior can be attributed to the formation of apatite layer on the surface (as identifie by XPS analysis in Fig.10).Apatite layer is thought to protect the underlying Mg surface from corrosion due to its extremely low solubility in SBF (-log(Ksp)25?C)=116.8) [60].

Fig.11.Corrosion rate as a function of immersion time for untreated AZ31B Mg and LSM AZ31B Mg samples in SBF as measured by weight loss technique.

The results indicate that laser surface modificatio facilitate precipitation of apatite,possibly due to higher energy state of the fin grain boundary region acting as active nucleation site.Combining with the improvement in wettability (lower SBF contact angle) represents the accelerated mineralization rate for higher laser fluenc treated samples.Additionally,several reports [61,62] indicated the closed pack crystallographic planes showed higher biocompatibility for HCP biomaterials.The results suggest that the substantial increase in the volume fraction (40%) of the crystallographic orientation of (0001)basal plane of HCP structure of Mg in laser treated surface in this work was highly beneficia to biocompatibility (Fig.5).Furthermore,the basal texture has also been reported to be beneficia for corrosion resistance [63].According to the equation proposed by Song et al.[63],the electrochemical dissolution rate of basal plane (0001) was 18～20 times slower thanplanes of AZ31B Mg alloy.

where n is the number of electrons involved in the electrochemical reaction;k is a reaction constant;F,R,T and E are Faraday constant,gas constant,absolute temperature and the electrode potential,respectively.αis a transit coefficient Q is an activation energy for a metallic ion to escape from the metal lattice and dissolve into the solution.

Due to the high value of Q for a densely packed atomic plane,the dissolution rate of basal plane was slower compared to any other loosely packed atomic plane.In the present work,the～40% volume fraction of (0001) plane in 3.18 J/mm2LSM AZ31B Mg sample led to reduced corrosion rate of LSM samples as a result of enhanced growth of mineral apatite.

4.Conclusions

The LSM process with various laser fluenc (1.06,2.12,and 3.18 J/mm2) led to a reduction in grain size from 10 μm for untreated sample to 2.5 μm for LSM AZ31B Mg samples.The grain refinemen was accompanied by formation of a cellular microstructure withβphase decorating the cell boundaries.EBSD analysis uncovered presence of 40%area fraction of the basal(0001)crystallographic texture on the surface and sub surface regions.Correspondingly,in the cross-section,the grain structure was associated predominantly prismatic crystallographic texture.From topography point of view,the surface roughness (Ra) of LSM AZ31B Mg samples steadily increased as a function of laser fluenc from 1.34 μm for 1.06 J/mm2to 3.95 μm for 3.18 J/mm2laser fluence The microstructure and surface texture significantl influence thein-vitroevaluation of the LSM AZ31B Mg samples.The contact angle in SBF steadily reduced from 65?for untreated sample to 43?for 3.18 J/mm2LSM AZ31B Mg sample.Furthermore,the Ca/P ratio in the mineral phase(1.683-1.691)in LSM AZ31B Mg samples was close to the stoichiometric HA mineral (1.67),whereas,the untreated sample was associated with a Ca/P ratio of 1.526.Corrosion resistance measured via weight loss revealed a steady increase in corrosion rate from 4 to 8 mm/year as a function of time for the untreated AZ31B Mg sample.For LSM AZ31B Mg samples,the corrosion rate steadily decreased from 2.5 to less that 0.5 mm/year as a function of time.Therefore,the combined effects of crystallographic texture,grain refinement phase evolution,and surface texture induced by LSM process led to a significan improvement inin-vitroresponse of AZ31B Mg alloy demonstrated by a low contact angle in SBF,formation of HA like mineral upon 7 day immersion in SBF,and a concurrent reduction in corrosion rate.Present work helped to shed light on effects if LSM process and further reinforced its suitability for improving performance of Mg based consumable implant materials.

Declaration of Competing Interest

None.

Acknowledgment

The authors acknowledge the Materials Research Facility(MRF) at the University of North Texas for access to microscopy and phase analysis facilities.Authors thank Dr.Sundeep Mukherjee for the access to the optical profilomete.