Influence of the shot-peening intensity on the structure and near-surface mechanical properties of Ti 40 Zr 10 Cu 38 Pd 12 bulk metallic glass

Shot-peening (SP) changes the near-surface structure and mechanical properties of a Ti40Zr10 Cu 38Pd12 bulk metallic glass. Near the surface, the hardness, Young's modulus, and elastic strain limit are all reduced. Measurements of the heat of relaxation show that an exceptionally high stored energy of cold work can be induced, implying a large increase in free volume. At the highest SP intensity there is partial nanocrystallization enabled by the increased free volume and not by the increase in temperature.

Shot-peening (SP) changes the near-surface structure and mechanical properties of a Ti 40 Zr 10 Cu 38 Pd 12 bulk metallic glass.Near the surface, the hardness, Young's modulus, and elastic strain limit are all reduced.Measurements of the heat of relaxation show that an exceptionally high stored energy of cold work can be induced, implying a large increase in free volume.At the highest SP intensity there is partial nanocrystallization enabled by the increased free volume and not by the increase in temperature.V C 2013 AIP Publishing LLC.[http://dx.doi.org/10.1063/1.4833017]2][3] BMGs are of interest for sporting goods, medical and electronic devices, and defense and aerospace applications.However, BMGs exhibit poor room-temperature macroscopic plasticity compared to polycrystalline metals, due to formation and rapid propagation of shear bands. 4MG plasticity can be improved by promoting the nucleation of shear bands and hindering their propagation by a wide variety of methods 4 including surface treatments (e.g., shot peening (SP) or laser peening). 5n SP, the sample surface is repeatedly struck by beads at high velocity with impact energy sufficient to cause plastic deformation.For polycrystalline alloys, the technique is conventionally used to induce a compressive surface stress particularly effective in increasing fatigue resistance.SP also refines the microstructure. 6SP has been applied to BMGs to create multiple shear bands at the surface and to generate a compressive stress, enhancing the overall plasticity. 5][9] In a partially crystalline Zr 55 Cu 30 Al 10 Ni 5 BMG, SP can induce either amorphization (or full crystallization) at higher (or lower) temperature. 10Crystallization was attributed to increased atomic mobility due to a large induced free volume V f rather than to temperature rise.The role of V f is studied further in the present work.
Structural changes observable in shot-peened surfaces are of interest as they aid understanding of changes (such as V f creation and crystallization) seen in shear bands themselves.Thus, SP may be useful in fundamental studies of plasticity mechanisms in BMGs, as well as in developing practical beneficial surface treatments.
Ingots, nominal composition Ti 40 Zr 10 Cu 38 Pd 12 (at.%), were prepared by levitation melting a mixture of high-purity (99.9 wt.%) elements under argon.Rods (diameter 2.5 mm) were obtained by copper-mold casting.Cylindrical specimens were cut from the rod, and the top surface was ground, polished, and shot-peened.The SP intensity was assessed by the Almen method based on the curvature induced in standard (type A) steel strips. 11Two treatments were applied: S1 (S2) with steel shot of diameter 0.4 (0.7) mm, air pressure of 1.5 (4.0) bar, peening time of 10 (20) s, giving an Almen intensity of 10 A (21 A).The SP was carried out at room temperature (i.e., 298 K) until attainment of 100% coverage.
The surface morphology after SP was observed with a scanning electron microscope (SEM) (Zeiss Merlin).Roughness was measured by profilometry with a confocal 3D optical surface metrology system LEICA DCM3D using blue light for high-resolution.The average line profile roughness (R a ) and surface area roughness (S a ) were obtained from four regions.
The as-cast and shot-peened samples were characterized by X-ray diffraction (XRD) (Philips X'Pert instrument, Cu-Ka).Nanoindentation under load-control was performed using a UMIS instrument (Fischer-Cripps Laboratories) with a Berkovich pyramidal diamond tip applying a 10 mN maximum load.Thermal drift was below 60.05 nm s À1 .To study the effects of SP, the samples were cut perpendicular to the shot-peened surface, and nanoindentation tests were performed on the transverse section at different depths (distances from the surface).From the load-displacement (P-h) curves, the hardness (H), and reduced Young's modulus (E r ) were evaluated at the beginning of the unloading segments, using the method of Oliver and Pharr. 12,13Corrections for the contact area (calibrated using a fused-silica specimen), initial penetration depth, and instrumental compliance were applied.The normalized plastic energy was evaluated as the ratio of plastic and total (plastic þ elastic) energies during nanoindentation, U pl /U tot .Values of U pl were calculated as the area between the loading and unloading nanoindentation curves while U tot is the area between the loading curve and the displacement axis.The thermal stability of as-cast and shot-peened samples was investigated by differential scanning calorimetry (DSC), heating at 40 K/min in a Perkin-Elmer DSC7.
SEM and 3D profilometry images of shot-peened surfaces (Fig. 1) show that after S1 treatment the surface is relatively smooth, with uniformly distributed shallow spherical craters; regions with detached material are scarce (white arrows).Under the harsher S2 conditions, the surface is more heavily deformed, the perimeter of deep craters is well defined, and there is more detached material.The average roughness values were measured using profilometry: for S2 R a ¼ 0.99 lm and S a ¼ 1.54 lm, significantly greater than for S1 with R a ¼ 0.33 lm and S a ¼ 0.63 lm.
XRD from the S1 sample surface shows a broad halo (Fig. 2(a)), indicating that the sample remains fully amorphous.The main diffraction halo becomes sharper after SP using S2 conditions (Fig. 2(b)), suggesting that the sample surface has become partly nanocrystalline (as observed after nanoindentation 14 or severe plastic deformation 15 ).
Shot-peened surfaces were analyzed by energy dispersive X-ray (EDX) spectroscopy.There was no evidence for iron or for oxides in sample S1, but oxide particles and iron (about 1 at.%) were detected for sample S2, particularly inside the deepest craters.Such contaminants may decrease the glass-forming ability 1 and catalyze the nucleation of nanocrystals.
][18] Nanoindentation data (Fig. 3) show the profile of H, E r , U pl /U tot , and r y /E as a function of distance from the surface for S 1 and S 2 samples.Both H and E r progressively decrease, and the indentation plastic energy increases, within 50-100 lm from the shot-peened surface.The elastic strain limit, e y ¼ r y /E, can be estimated using H % 3r y and E r % E.More realistic values of r y and e y (comparable to the as-cast sample, r y % 2 GPa, and e y % 0.02) 19 are obtained once an indentation size effect of around 35% for H and 20% for E is accounted for, in good agreement with the indentation size effect in other Ti-based BMGs. 17The results so obtained are plotted in Fig. 3(d) and indicate that SP lowers r y /E, consistent with the easier onset of plasticity expected in pre-deformed BMGs.
DSC curves were measured from thin slices cut at different distances from the peened surfaces (Fig. 4).No significant variations were observed in T g ($686 K) or crystallization temperature (T x % 720 K) (Table I).The crystallization enthalpy is, however, smaller for slice S2a (H cryst ¼ À6:26 kJ mol À1 ) than for the other samples ($ À6:97 kJ mol À1 ), suggesting that partial crystallization has taken place near the peened surface, consistent with the XRD results.However, three partially overlapped exothermic events are observed in all samples by DSC, indicating that no significant changes in the crystallization mode are induced, even after the most aggressive shot peening condition.
To interpret structure and property changes, it is useful to have a measure of the degree of SP-induced damage in the glassy structure.Damage can be characterized as increased  free volume V f and the enthalpy of a metallic glass can be linearly related to the V f content. 20Thus the stored energy of cold work, manifest as exothermic relaxation below T g , is a key parameter.The enthalpy of relaxation H relax (Table I) increases dramatically after SP.This effect progressively decreases as the distance from the shot-peened surface is increased.This confirms earlier works 8,9 that SP induces an increase in V f .More V f is generated in the S2 sample, despite the greater heating presumably induced during the more intense SP.Increased V f should lead to lower hardness, yet Fig. 3 shows that the deformation-induced decreases in H and E r in S2 are smaller than for S1.While the surface oxide phases in sample S2 could contribute to the higher hardness, the most likely dominant cause is nanocrystallization, consistent with the XRD observations.The Ti 40 Zr 10 Cu 38 Pd 12 BMG crystallizes to CuTi, CuTi 2 , and CuZr intermetallic phases, giving increased H and E r compared to the fully glassy structure. 21he observation of partial crystallization in the less relaxed surface layer of the S2 sample allows us to definitively exclude the possibility that the crystallization is induced by local heating.If local heating were responsible, the residual glassy phase would be more, not less relaxed than in the S1 sample.In previous work on BMGs, crystallization induced by SP was observed in only one case, when FIG. 4. DSC curves from thin slices cut at different depths from the shotpeened surfaces of S 1 and S 2 samples.Slices S1a, S1b, and S1c were cut at 450, 610, and 900 lm, and slices S2a, S2b, and S2c at 550, 756, and 916 lm, from the end surfaces of the BMG rods.For clarity, the curves are shifted along the y-axis.TABLE I.The glass-transition temperature (T g ), crystallization temperature (T x ), relaxation enthalpy (H relax ), and crystallization enthalpy (H cryst ) for disks cut at different distances from the shot-peened surface (Fig. 4), for samples S1 and S2.H cryst was measured as the area underneath the three observed crystallization peaks (Fig. 4).there were pre-existing crystals to seed the transformation and when the SP was at 77 K.It was inferred that the further crystallization took place, not during the SP itself, but during subsequent heating back to room temperature.In the present case there is no evidence for pre-existing crystals, and the crystallization occurs at the temperature of the SP itself.The magnitude of the stored energy of cold work is useful in understanding why crystallization can be induced in the present case.

Sample
The thickness of the top slices (450 lm for S1a; 550 lm for S2a) used for DSC is much larger than the material depth affected by the SP (<100 lm, Fig. 3).The increase in H relax (i.e., stored energy of cold work) in slice S2a is À455À(À290) ¼ À165 J mol À1 when averaged over the slice.Given that only $20% of the overall slice thickness is affected by SP, the increase in H relax for the peened layer itself must be $À825 J mol À1 .The decreased crystallization enthalpy in slice S2a (Table I) suggests a crystalline volume fraction of (6970À6260)/6970 ¼ 10%.Since only $20% of the slice is affected by SP, this suggests a fraction as high as 50% in the peened layer.In that case, the increase in H relax for the residual glassy phase in that layer is estimated to be $À1650 J mol À1 (%23% of the crystallization enthalpy H cryst ).
For pre-annealed Pd 40 Cu 30 Ni 10 P 20 , the stored energy of cold work was at most À180 J mol À1 (À470 J mol À1 ) from SP at 298 K (77 K). 9 These values are comparable with the stored energies of up to À200 J mol À1 [Ref.22] and À445 J mol À1 [Ref.23] in cold-rolled ribbons of Pd-and Ni-based glasses.The value of $ À1650 J mol À1 , suggested in the present work, is much larger but still in the range of H relax found in undeformed as-cast glasses.For example, for as-cast Pd 40 Cu 30 Ni 10 P 20 BMG, H relax ¼ À580 J mol À1 [Ref.9] while for melt-spun (i.e., rapidly quenched) ribbons of Fe-based glasses, the value of H relax is as high as -5500 J mol À1 . 24he greatly increased H relax induced by the S2 treatment implies that SP induces a large increase in V f .Yet the overall V f is only comparable with, or less than that in, as-cast rapidly quenched glasses.SP-induced nanocrystallization may need to be interpreted not only in terms of the total V f , but also in terms of its distribution.This distribution can be characterized from positron annihilation lifetime spectroscopy and from the thermal relaxation spectrum seen in DSC; deformation-induced changes in the distribution are reviewed in Ref. 9.
The large stored energy found in the S2 sample may be favored by the choice of steel shot instead of the glass beads used in previous works 8,9 on SP of BMGs.In addition, inducing more damage because they are heavier, the steel shot must be more effective in extracting from the sample the heat generated during successive impacts.
To conclude, shot-peening the surface of Ti 40 Zr 10 Cu 38 Pd 12 bulk metallic glass induces heavy plastic deformation that results in softening (reduction in hardness and in the elastic strain limit).The use of steel shot (rather than glass beads as in earlier work) gives a stored energy of cold work in the peened glass estimated to be as high as À1650 J mol À1 , nearly four times the previous highest stored energy recorded in a metallic glass as a result of any mechanical treatment.The extreme deformation causes partial nanocrystallization of the surface.The coexistence of nanocrystallization and a highly unrelaxed glass means the underlying enhanced mobility that crystallization implies can be attributed to greatly increased free volume, and definitely not to heating.The onset of crystallization may be a key factor limiting the stored energy possible through cold work.
Influence of the shot-peening intensity on the structure and near-surface mechanical properties of Ti 40 Zr 10 Cu 38 Pd 12 bulk metallic glass Instituci o Catalana de Recerca i Estudis Avanc ¸ats (ICREA) and Departament de F ısica, Universitat Aut onoma de Barcelona, E-08193 Bellaterra, Spain (Received 29 August 2013; accepted 8 November 2013; published online 21 November 2013)

FIG. 3 .
FIG. 3. Mechanical properties of Ti 40 Zr 10 Cu 38 Pd 12 BMG as a function of the distance from the shot-peened surface: (a) hardness H, (b) reduced Young's modulus E r , (c) normalized nanoindentation plastic energy U pl /U tot , and (d) elastic yield strain r y /E, estimated as described in the text.