Combined severe plastic deformation processing of commercial purity titanium enables superior fatigue resistance for next generation implants

for replacing


Introduction
Titanium is one of the most attractive metallic biomaterials for medical applications.It is used as raw material for manufacturing of numerous types of medical implants including artificial hip joints, knee joints, bone plates, screws for fracture fixation, cardiac valve prostheses, pacemakers, and artificial hearts [1].Ti-6Al-4 V ELI (Ti64; grade 23) is one of the most widely used alloys for such implants [2].However, in the human body, corrosion-induced release of elemental ions from grade 23 has been shown to cause toxic reactions, which seems a particular concern in dental applications.Specifically, vanadium ions have been associated with neurotoxicity and inhibition of cellular proliferation and differentiation in vitro as well as other health effects [3][4][5].Moreover, aluminum is considered as a potential cause of Alzheimer's disease among other neurological disorders [6].Corrosion behavior of implant materials strongly influences foreign body reactions in the vicinity of the implantation site and thus, plays a significant role in the biocompatibility of implants.
Commercial purity Ti (cp-Ti) has been reported to have the lowest corrosion rate when tested in physiological saline solution [7] and is considered as one of the most biocompatible metallic materials [1].This has enabled cp-Ti grade 4 to be used in various dental implant applications and to become a promising candidate for replacing potentially toxic grade 23 in orthopedic implant applications.However, low mechanical and fatigue properties of cp-Ti are major limiting factors when it comes to replacing grade 23 and other Ti alloys.Moreover, next generation dental implants manufactured from cp-Ti grade 4 with increasingly narrow diameters (Ø < 3 mm) are a recent focus of research for industry [8,9].Therefore, suitable production methods are needed for improving the mechanical static and foremost, dynamic (fatigue), properties of cp-Ti in order to enable such downsizing of dental implants while maintaining the overall integrity of the implants.Besides what is considered a R1-strategy (refine), the increased interest in substituting Ti alloys with commercial purity material, is a R2-strategy (replace) aiming to optimize existing implant systems, particularly orthopedic, in terms of mechanical properties and fatigue resistance while waiving potentially harmful alloying elements.
In the quest to significantly increase the mechanical performance of cp-Ti and in turn, pursue R1 or R2 strategies, numerous studies have focused on the use of severe plastic deformation (SPD) methods [7,.Such methods have been effective at generating ultrafinegrained (UFG) microstructures in cp-Ti with improved mechanical strength, good ductility and in some instances, enhanced biocompatibility and osseointegration.Besides grain refinement, SPD processing has contributed to enhanced strength in cp-Ti through various other mechanisms including evolution of substructures and strong fiber texture [22,37], suppression of dislocation slip and twinning, and accumulation of dislocation densities [34] among others.Nevertheless, it has been demonstrated that both microstructure and texture of cp-Ti are highly dependent on which SPD process is used, ultimately resulting in variations in mechanical strength, ductility and fatigue properties.
Of the variety of SPD methods, equal channel angular pressing (ECAP) is desirable due to its capability of generating homogeneous and UFG microstructures while maintaining cylindrical or rectangular bar shape.This process involves passing a sample through a die consisting of two channels of equal cross section intersecting at a specific angle [16].During ECAP, a high strain is introduced within the sample and in turn, results in significant grain refinement.Such grain refinement is reported as the dominant strengthening mechanism for cp-Ti processed by ECAP [20,24].Deformation twinning and increased accumulation of dislocations have been reported as underlying mechanisms for the improved strength as well [24,37].However, since the length to diameter ratio of ECAP samples should not be much >10, there are certain limitations to the workpiece lengths that need to be overcome.As such, it is often necessary to combine ECAP with a second thermo-mechanical process to permit the production of long rods suitable for SWISS type turning; a standard machining process for the production of dental implants, orthopedic screws, and even some types of bone plates.Alternatively, a continuous process known as ECAP-Conform (ECAP-C) can be used [44].This process was developed by modifying the Conform technique originally designed for continuous extrusion of wires, rods and tubes [45,46].The ECAP-C process is advantageous in that it can produce long rods with UFG microstructure and improved mechanical properties.Nevertheless, despite the practicality of ECAP-C and its capability for alloy strengthening, anisotropic material properties force repetitive processing even in this case while industrial scalability and process stability have yet to be demonstrated.In addition, it is still necessary to combine it with a second thermo-mechanical process for cp-Ti to effectively exceed the mechanical strength of grade 23 and unlock the potential for cp-Ti to replace grade 23 or other Ti alloys in medical applications [47][48][49][50][51][52].
In this study, an approach combining ECAP with subsequent RS (hereafter referred to as ECAP/RS) was used for cp-Ti grade 4. Rotary swaging is an effective forming method capable of producing highstrength rods, tubes and wires with high processing speed and shape accuracy.To date, only a few studies have investigated the applicability of RS for Ti alloys [53][54][55][56][57][58][59], demonstrating improved mechanical and fatigue performance, but often reporting minimal ductility.Additionally, other works have reported on combined ECAP/RS for cp-Ti [60,61], for which high strength was accompanied by UFG microstructure and significant anisotropy of mechanical properties.However, in contrast to the current study, these works utilized drawing as a final processing step following RS, which would have resulted in distinct microstructures and texture.
The goal of this research was to investigate the suitability of combined ECAP/RS processing of cp-Ti grade 4 for medical implant applications.Incorporating the advantages of ECAP with those of RS can potentially lead to an industrially-suitable technique, capable of manufacturing cp-Ti medical implants with superior strength and fatigue endurance.In this regard, cp-Ti grade 4 samples were evaluated for microhardness, tensile properties, and fatigue endurance, along with extensive microscopy to characterize grain size and morphology, as well as texture properties.Furthermore, corrosion resistance and biocompatibility of cp-Ti grade 4 were characterized.

Equal channel angular pressing and rotary swaging
The cp-Ti grade 4 was received as hot rolled rods with diameter of 22 mm from E. Wagener GmbH in Heimsheim, Germany.The composition of the as-received (AR) material as documented by the supplier's material certificate is presented in Table 1.Prior to ECAP, the rods were sliced to 160 mm lengths.ECAP experiments were carried out at Meotec GmbH in Aachen, Germany, at temperatures of 300 • C and 475 • C, following the B C processing route with a total of four passes at a speed of 0.8 mm/s.The ECAP die consisted of two channels with 22 mm diameters and an angle of intersection of 120 • .For the ECAP experiments carried out at 300 • C (denoted as ECAP-300 samples from hereon), the cp-Ti samples were initially annealed at 680 • C for 1 h and cooled in air.In the case of ECAP experiments carried out at 475 • C (denoted as ECAP-475 samples from hereon), no prior annealing was performed since the temperature was deemed high enough to ensure sufficient ductility during processing.For both ECAP temperatures, the billets were preheated in a furnace for 30 min at temperatures 75 • C higher than that of the processing temperature (i.e., 375 • C and 550 • C for the 300 • C and 475 • C trials, respectively).The billets were continuously reinserted into the furnace and held at preheat temperatures between passes during ECAP.
Following ECAP experiments, the ECAPed rods were examined and any defects generated from ECAP were either completely sectioned or ground away.The ECAP-300 samples were subjected to further annealing at 250 • C for 1 h prior to RS.No annealing was carried out on the ECAP-475 samples prior to RS. Rotary swaging was performed at room temperature for all samples.For simplicity, the samples processed further by RS will be denoted from hereon as Swage-300 and Swage-475 for those first ECAPed at 300 • C and 475 • C, respectively.
The Ø22 mm ECAPed rods were rotary swaged to Ø6 mm, resulting in a total area reduction of 93 % and true strain of 2.6.To attain such reduction, several RS steps were necessary and combined with intermediate stress-relief annealing.The final length of the bars after ECAP/ RS was ~2 m.

Tensile and fatigue testing
Mechanical properties of cp-Ti grade 4 specimens were investigated via microhardness measurements, and tensile and fatigue testing.Microhardness measurements were carried out on both ECAPed and combined ECAP/RS specimens, while tensile testing included cp-Ti grade 4 specimens in the AR condition.Finally, fatigue testing was performed solely on AR samples and combined ECAP/RS specimens, since such specimens are those considered production-suitable.Ti grade 23 samples in the AR condition (i.e., hot rolled) were also subjected to tensile and fatigue testing as reference material for comparative purposes.
Vickers type microhardness measurements were performed at the Austrian Institute of Technology in Wiener Neustadt, Austria, on a fully- automatic hardness tester (EMCO-TEST DuraScan 80) using a load of 9.807 N (HV1) according to the EN ISO 6507-1 standard.Sample preparation for microhardness testing was carried out following metallographic procedures described in Section 2.3.Hardness mappings with up to 700 individual indents were performed on the cross-section perpendicular to the rod axis.Moreover, at least 40 individual indents were taken along the longitudinal axis of the rod.Tensile specimens were machined according to international standard DIN EN ISO 22674, suitable for the fabrication of dental restorations and appliances, with a 3 mm gauge diameter and 15 mm gauge length.Tensile testing was carried out at ambient temperature on a Zwick/Roell Z010 tensile testing machine.The samples were subjected to a constant pulling speed of 2 mm/min.A minimum of five specimens were tested for each condition.
Fatigue testing was performed according to international standard ISO 1143 on specimens with a 3.5 mm gauge diameter and 10 mm gauge length using a rotating bending fatigue testing machine.A minimum of 15 samples per condition were tested at ambient temperature with rotation frequency of 75 Hz.Runout samples were defined as those that exceeded 10 7 cycles.For each load level within the region of finite life fatigue strength, the finite number of cycles was calculated for a failure probability of 50 %.A partial regression line was laid through the calculated points according to the arcsin√P -method of Dengel [62].The fatigue endurance limit (50 % failure probability) was determined using the raw data of load levels which contained both failed and runout samples.The results are presented as S -N curves with corresponding fatigue endurance limits.

Microscopy
Samples for optical microscopy were sectioned from the corresponding bars and subsequently ground using SiC grinding paper up to a fineness of P4000.Following grinding, the samples were ultimately polished using a CHEM-PAD polishing cloth with OPS solution.The samples were then etched by immersing for a minimum of 30 s in a solution consisting of distilled water (100 ml), 30 % solution hydrogen peroxide (5 ml) and 40 % solution hydrofluoric acid (2 ml).Optical microscopy was carried out using a Keyence laser scanning confocal microscope.
Transmission electron microscopy was carried out at the Central Facility for Microscopy in Aachen, Germany, for the Swage-300 condition.Here, electron transparent samples were obtained from the central region of the cross-sectional area by focused ion beam (FIB) machining using a FEI Strata 400 S (FEI Company, Eindhoven, The Netherlands).The samples were investigated using a 200 kV Tecnai F20 transmission electron microscope (FEI Company) equipped with a Veleta S04F camera (EMSIS GmbH, Munster, Germany) and an ADF detector (E.A.: Fischione Instruments, Inc., Export, Pennsylvania, USA) and an ASTAR-System (NanoMEGAS SPRL, Brussels, Belgium).Selected area electron diffraction (SAED) patterns were recorded from areas using a field limiting aperture with a projected diameter of ~3.7 μm.Evaluation of the SAED patterns was carried out using the program ELD (Calidris, Sollentuna, Sweden).Finally, areas of ~4 μm 2 were mapped by scanning electron nanobeam diffraction (SEND) using a lateral step-by-step resolution of 10.5 nm.The thereby obtained datasets were subsequently processed with the software pack ACOM-TEM [63] and further analyzed with OIM Analysis version 8.2 (EDAX-TSL).
Fractography was also performed on selected ruptured samples following tensile and fatigue testing.For this analysis, a Philips XL30 scanning electron microscope (FEI Company) was used at 20 kV in secondary electron imaging mode with working distances ranging from 10 to 12 mm.

Corrosion testing
Electrochemical (EC) tests were performed on AR and Swage-300 specimens.The samples were chemically cleaned prior to measurements in order to eliminate potential influence from combined ECAP/RS (i.e., lubricants, oil).Cleaning was performed by immersing the samples in a solution consisting of 5 % hydrofluoric acid for 1 min.In addition, two surface treatments were also carried out on separate Swage-300 specimens prior to corrosion testing.Specifically, plasma electrolytic oxidation (PEO) surface treatment and sandblasting and acid etching (etched) were performed to alter surface roughness and mimic the treatments commonly applied to commercial Ti-based dental implants [64,65].For PEO, similar processing parameters and composition to those reported in [65] were used.For the etched samples, after sand blasting, the samples were etched for 5 min in a solution consisting of hydrochloric and sulfuric acid.
All EC tests were carried out using a Biologic potentiostat (SP-150 Biologic Science Instruments, Seyssinet-Pariset, France) and a standard three-electrode EC measurement cell including a saturated Ag/AgCl electrode as reference electrode, a graphite rod as counter electrode and the Ti sample as working electrode.The electrodes were immersed in SBF at a constant temperature of 37 ± 1 • C. EC measurements were started by recording the open circuit potential (OCP).Subsequently, potentiodynamic polarization (PDP) measurement was conducted, starting at cathodic region (OCP -1.0 V) scanning with a fixed scan rate of 5.0 mV/s to anodic (OCP + 1.0 V) direction.The resulting current was recorded constantly.Results were plotted and Tafel exploration analysis was performed.

Biocompatibility
L-929 mouse fibroblasts (LGC Standards, Wesel, Germany) were cultured in Minimum Essential Medium supplemented with 10 % fetal bovine serum, penicillin/streptomycin (100 U/ml each) (all from Life Technologies, Carlsbad, USA) and L-glutamine (Sigma-Aldrich, St. Louis, USA) to a final concentration of 4 mM at 37 • C, 5 % CO 2 and 95 % humidity.As a toxic control, RM-A (Hatano Research Institute, Food and Drug Safety Center, Japan) was used for indirect testing.For the livedead staining assay, TC coverslips (Sarstedt, Nürmbrecht, Germany) were used as nontoxic controls and RM-A was utilized as the toxic positive control material.Cytocompatibility experiments were carried out following Jung et al. [66] for indirect viability and cytotoxicity assays, as well as for the direct live-dead staining assays.
Briefly, for indirect assays, the four test specimens, namely the Swage-300 sample in etched, PEO, and ECAP/RS conditions, as well as AR cp-Ti grade 4, were extracted individually along with the toxic control samples using 3 cm 2 /ml of cell culture medium for 72 h under cell culture conditions.Cell culture medium alone was incubated under identical conditions to serve as a negative control extract for indirect testing.After removal of the specimens, the remaining extracts were centrifuged at 14,000 rpm for 10 min.The supernatants were used for the assays.Subsequently, the extracts were used as medium to culture cells for 24 h under cell culture conditions.On the next day, viability and cytotoxicity were measured using XTT-(Cell Proliferation Kit II, Roche Diagnostics, Mannheim, Germany) and LDH (BioVision, Milpitas, USA) assay kits respectively according to the manufacturer's instructions.
For the live dead staining assay, the sample surfaces were seeded directly with cells in cell culture medium at a surface to volume ratio of 5.65 cm 2 /ml in wells of a 12 well plate.After an incubation for one day under cell culture conditions, 60 μl per ml propidium iodide stock solution (50 μg/ml in PBS) and 500 μl per ml medium fresh fluorescein diacetate working solution (20 μg/ml in PBS from 5 mg/ml FDA in acetone stock solution) were added to each well.Following a brief incubation of 3 min at room temperature, the specimens were rinsed in PBS and immediately examined with an upright fluorescence microscope (Nikon ECLIPSE Ti-S/L100, Nikon, Düsseldorf, Germany), equipped with a filter for parallel detection of red and green fluorescence.

Mechanical properties 3.1.1. Microhardness
The results of microhardness tests are shown as color-coded mappings in Fig. 1 and presented as mean values in Fig. 2. A legend depicting the correlation between color and hardness is presented in Fig. 1.Spaceresolved hardness mappings revealed inhomogeneous microhardness distribution after RS, as harder areas are observed along the center regions of the specimens.In contrast, the hardness distribution after ECAP is more uniform.Similar results were achieved in previous studies [26,27,57].In the case of RS, lower hardness is typically observed along the edges due to friction between the dies and the surface of the material [57].A notable difference in microhardness was seen for the different ECAP processing temperatures.Specifically, the 300 • C processing temperature was found to increase microhardness as more green areas are visible in the ECAP-300 map compared to the ECAP-475 map.This is consistent with the results in Fig. 2 for microhardness along the crosssection.The discrepancy in microhardness between processing temperatures was expected and is in agreement with literature, given that higher processing temperatures can increase grain size due to the inhibition of possible recrystallization, and decrease the dislocation density, both leading to a lower strength improvement [67].Moreover, such increased hardness was carried through the RS process, as depicted by the larger red region of the Swage-300 map, and as numerically confirmed in Fig. 2. Finally, differences in microhardness values between the two directions (i.e., longitudinal and cross-section) directions suggest anisotropy of mechanical properties, with increased hardness along the direction of processing (longitudinal).

Tensile
The tensile properties for cp-Ti grade 4 in the AR, ECAPed and ECAP/ RS conditions are presented in Fig. 3 along with those of AR Ti grade 23.The results for cp-Ti grade 4 are in good agreement with microhardness values in Fig. 2. The yield stress and UTS of cp-Ti grade 4 was found to continuously increase both after ECAP processing and combined ECAP/ RS.In the case of the ECAPed samples, the results show that processing at lower temperature (i.e., 300 • C) generated higher strength, as was the case for the microhardness values.Specifically, average values of yield stress and UTS were measured as 916 ± 19 MPa and 947 ± 12 MPa, respectively, for the ECAP-300 sample, and 797 ± 0 MPa and 845 ± 7 MPa, respectively, for the ECAP-475 sample.This improved strength was also maintained throughout room-temperature RS.An average yield stress and UTS of 1383 ± 27 MPa and 1396 ± 12 MPa respectively, was obtained for the Swage-300 condition.To the best of the authors' knowledge, these are the highest values recorded for yield stress and UTS of cp-Ti.In contrast, a slightly lower average yield strength and UTS of 1275 ± 19 MPa and 1334 ± 6 MPa respectively, was measured for the Swage-475 condition.Both conditions exceeded the strength of AR Ti grade 23, as shown in Fig. 3.The increase in strength after ECAP/RS was countered by a decrease in ductility for all conditions, indicative of strain hardening of the material.Nevertheless, moderate ductility was maintained for both the Swage-300 (i.e., 9.7 %) and Swage-475 (i.e., 13.3 %) conditions.
When comparing the tensile properties of cp-Ti processed by ECAP/ RS to those of previous works, it is important to take into account the effect of solid-solution strengthening.In cp-Ti, this mechanism is governed by the total content of interstitial solutes (i.e., oxygen, nitrogen, and carbon), and is, in turn, represented by the oxygen equivalent,  which is determined from the content of such solutes [38].For the cp-Ti grade 4 used in this study, the oxygen equivalent was 0.33 (see Table 1).Although solid-solution strengthening was likely a contributing factor to the improved strength recorded in this study, other microstructural effects were more effective, given that higher values of oxygen equivalent were recorded in works reporting lower mechanical strength for cp-Ti after combined SPD processing [15,60,71].

Fatigue
The S -N (Wöhler) curves for the AR cp-Ti grade 4 and Ti grade 23 samples, along with the Swage-300 and Swage-475 samples are illustrated in Fig. 4, and the corresponding numerical values of fatigue endurance limit are listed in Table 2.With respect to AR cp-Ti grade 4, combined ECAP/RS resulted in a 43 % and 41 % increase in fatigue endurance for the Swage-300 and Swage-475 conditions, respectively.This way, the combined processing was able to almost fully obtain the fatigue endurance limit of grade 23 (i.e., 640 MPa), with a significantly reduced difference (i.e., from 53 % for AR cp-Ti grade 4 to only 6 % for Swage-300).Only a minimal difference in fatigue endurance was measured between the two ECAP processing temperatures, despite the differences observed for microhardness and tensile strength.This is likely attributed to the higher ductility of the Swage-475 sample, since ductility typically promotes higher fatigue endurance for cp-Ti [68].
Numerous authors have reported on the fatigue properties of cp-Ti grade 4 for various loading conditions (e.g., bending, axial) and testing parameters (e.g., frequency, cycle symmetry) [40,[68][69][70].Given the influence of such parameters on fatigue endurance limit, it is somewhat difficult to directly compare results among studies.However, in this work, the values of fatigue endurance limit are consistent with those reported in literature for cp-Ti grade 4, particularly with those generated from combined ECAP and cold forming processes (e.g., drawing) [15,71].When compared to conventional ECAP processing, the fatigue endurance limit of cp-Ti obtained in this study is superior [32], thereby demonstrating the potential for combined ECAP/RS processing.This potential is further illustrated by the significantly higher fatigue endurance limit for cp-Ti compared to that achieved solely by RS [53,55].To the best of the authors' knowledge, only a few studies reported a higher fatigue endurance for cp-Ti grade 4 [71,72].However, this was only attained following a thermal annealing treatment.Indeed, future research should investigate the effect of annealing following ECAP/RS processing, as it was not the focus of this research.A. Kopp et al.

Optical microscopy
Optical micrographs are provided in Fig. 5 for the AR, annealed, ECAP-300 and Swage-300 conditions.Optical microscopy revealed similar microstructures for ECAP-475 and Swage-475 samples as that for ECAP-300 and Swage-300 samples, respectively, thus these micrographs are not illustrated.The images were taken along the longitudinal direction of the samples (i.e., in the direction of processing).The AR sample (Fig. 5(a)) shows typical hot-rolled microstructure with a combination of both equiaxed and elongated α-Ti grains.Following annealing, the microstructure was homogenized to a complete equiaxed grain structure, as shown in Fig. 5(b).The average grain size was ~30 μm.After ECAP, the resulting microstructure was significantly refined, as shown in Fig. 5(c).The grains are also slightly elongated due to the shearing process, and oriented to the billet extrusion direction.Moreover, due to the channel intersecting angle, such elongated grains are also tilted with respect to the direction of feeding.Such microstructure is characteristic of the ECAP B c processing route [30].A further refinement in microstructure is depicted for the Swage-300 sample.It is important to note the 20 μm scale bar depicting higher magnification for this sample in Fig. 1(d).It is therefore evident that grain refinement played a role in improving the strength of cp-Ti after combined ECAP/RS processing.Moreover, a more pronounced banded microstructure is observed for the Swage-300 sample, consistent with those observed in cp-Ti following processing by RS [58].In order to better analyze the microstructure after combined ECAP/RS, transmission electron microscopy was performed.The results are presented in Section 3.2.3.

Fractography
Scanning electron micrographs of fracture surfaces are presented in Figs.6(a) and (b) from ruptured ECAP-300 and Swage-300 tensile samples, respectively.Both fracture surfaces displayed areas of ductile dimple fracture (i.e., microvoid coalescence) along the center of the sample and a more brittle cleavage fracture (i.e., shear lip) along the outer circumference.A greater area of the surface was covered with dimple rupture for the ECAP-300 specimen.In contrast, the fracture surface of the Swage-300 specimen exhibited greater areas of cleavage fracture and a more pronounced shear lip.Nevertheless, the fracture surface in Fig. 7 confirms the presence of dimple rupture indicating a sufficient amount of ductility present within the Swage-300 specimen.
Fracture surfaces of two Swage-300 specimens after rotating bending testing are illustrated in Figs. 8 and 9. Specimen 1 (Fig. 8) was tested at high stress (σ max = 700 MPa) and had low endurance (N = 28,943 cycles).Specimen 2 (Fig. 9) was tested at a stress slightly higher than the fatigue limit (σ max = 625 MPa) and experienced a high endurance (N = 9.4 × 10 6 cycles).Significant differences were observed on the fracture surface of the two specimens.The fracture surface of specimen 1 exhibits a large area of crack initiation (Fig. 8(a)).Moreover, a significant portion of the fracture surface was characterized with dimple rupture (Fig. 8(b)), which indicates static failure due to the limited endurance of the specimen.Crack fatigue striations were virtually absent in the zone of rapid crack propagation for this specimen.
Specimen 2 tested at a load close to the fatigue limit while experiencing a number of cycles nearing that of runout, has a fracture pattern differing from that of Specimen 1. Crack initiation was significantly less  pronounced in comparison to that for Specimen 1.Moreover, the surface of specimen 2 is characterized by fatigue striations in the zone of crack propagation (Fig. 9(a) and (b)) which is an indication of a more ductile mechanism of fracture [73].Finally, in comparison to specimen 1, the fracture surface of specimen 2 demonstrates a narrow zone of transition to static breakage, consisting of radial brittle striations (Fig. 9(c)) that suggest accelerated development of fracture [73,74].

Transmission electron microscopy
3.2.3.1.Grain structure and texture.Transmission electron microscopy was employed for the Swage-300 sample to investigate the microstructure from cross-sectional FIB-Lamellae.This sample was selected since it had the highest fatigue endurance limit.To facilitate understanding of microstructure with respect to RS processing, it is necessary to first establish the directions corresponding to the workpiece; the flow direction (FD) is defined along the workpiece direction of bar feeding during swaging.The transverse direction (TD) is defined perpendicular to FD as is the normal direction (ND) which is orthogonal to TD. TD and FD span the longitudinal plane, ND and FD span the hoop plane (see inlets in Fig. 10), whereas TD and ND span the radial plane.
The STEM bright field micrographs in Fig. 10 show overview images of two orthogonal thin film specimens (FIB-Lamellae) extracted from the center of the Swage-300 bar sample radial plane.Hence, the resulting cross sections project (a) the longitudinal and (b) the hoop plane of the workpiece.Visual examination of bright field images shows that the Swage-300 lamellae exhibit a predominantly banded structure parallel to FD with grains of up to several micrometers in length.The horizontal width of grains inside the banded structure reaches several hundred nm (both in ND and TD).The distribution of bands with varying width appears to be heterogeneous, with arrays of finer banded grains in-between larger grain "fibers".Equiaxed grains are sporadically observable throughout the banded matrix, indicative of substructure evolution.
As-obtained ACOM grain orientation maps for the Swage-300 sample are presented in Fig. 11.The maps were generated for both longitudinal (Fig. 11(a)) and hoop cross sections (Fig. 11 (b)).The color code, red for {0003} (basal plane), blue for {3030} (1st order prism plane) and green for {2110} (2nd order prism plane), shown in the standard stereographic triangle (inset in Fig. 11 (a)), corresponds to the crystallographic orientation of each grain with respect to the viewing direction.Orientation maps on the left-hand side correspond to color code projections of grain orientations observable from the transmission direction perspective (transmission maps).The maps on the right-hand side correspond to the color code projection of grain orientations viewed along the flow direction (flow maps).The color variation within the grains qualitatively represents the difference in internal misorientations.
The maps demonstrate numerous elongated grains along FD that form a predominantly banded style microstructure.Overall, the microstructure is somewhat heterogeneous as distinct areas consisting of coarse (> 1 μm wide) elongated grains, thin banded structures, and ultrafine equiaxed grains are visible.In both cross-sections, individual grain lengths sometimes exceed the vertical dimensions of the scanned areas.Coarse elongated grains in Fig. 11(a) are mostly deformation free, as the color variation within these grains suggests minor misorientation accumulated within the grain.The thin banded regions consist of grains with approximately 50 nm in width and several micrometers in length, sometimes exceeding the vertical dimensions of the scanned areas.A small fraction of ultrafine substructure grains is seen to have nucleated along the boundaries of elongated grains, with the smallest grains ranging from 20 to 30 nm in diameter.Furthermore, the maps of both cross sections showing orientations projected along FD display a distinct and almost uniform blue coloring.This indicates a strong preferred  11a).The 3D models below the pole figures display the preferred orientation according to the PF hot spots.The colored prism faces correspond to the basal ({0001}, red) and 1st order prism planes ({1010}, blue).The unit cell is displayed with dashed lines.SAED projected aperture diameter is approximately 3.7 μm.orientation of 1st order prism planes orthonormal to the bar feeding direction.The color grading in transmission perspective maps is dominated by red-yellow-green (i.e., basal and 2nd order prism orientations) as well as intermediate (of basal and 2nd order prism) orientations.Secondly, the longitudinal cross section shows more intermediate to 2nd order prism orientations, whereas the basal orientations seem to be encountered more frequently in the hoop plane.The flow maps reveal grains with more random orientations in finer, partially equiaxed grain structures.Furthermore, larger, elongated 2nd order prism-oriented grains are sporadically observable in flow maps.The latter grains are clearly distinct in transmission maps as well, showing an intermediate orientation between basal and 1st order prism planes (purple coloring).Grains that are not following the preferred orientation along FD seem to be more frequent in the hoop plane (Fig. 11(b)) (although that can be attributed to a larger fraction of fine-grained regions visible in the scanned areas.) Selected area electron diffraction patterns and complementary pole figures are presented for the Swage-300 sample in Fig. 12  Combinations of d 1010 , d 0002 and d 1011 can be used to construct ideal low indexed zone axis orientations (as shown in Fig. 14) in order to simplify and characterize the majority of observable Bragg reflection intensity.The derived zone axis orientations are {0001}<1010>, {1210}<1010>, and {1213}<1010> which in combination, account for all observed intensity at d 1010 , d 0002 and d 1011 , their equivalent positions and the majority of higher order reflections, whilst sharing the d h0h0 reflection row along FD (with regards to the aforementioned azimuthal spread).However, it is prudent to assume intermediate orientations between {0001}<1010>, {1210}<1010>, and {1213}<1010> contribute to the occurring patterns as well.Some differences are observed in the diffraction patterns, however.Specifically, in the longitudinal plane (Fig. 12), the diffraction pattern depicts a less continuous distribution of spots and stronger contributions from d 0002 and d 1011 positions.This indicates a stronger presence of grains with intermediate and 2nd order prism orientations.On the other hand, the hoop plane is dominated by reflection positions of the basal plane lattice.Consequently, both diffraction patterns indicate a strong preferred orientation of (1010) (i.e., 1st order prism) planes orthonormal to the flow direction as indicated by the d h0h0 Bragg reflection row residing on the vertical axis (i.e., parallel to FD) of the diffraction patterns.The occurrence of more intermediate and 2nd order prism orientations in the longitudinal plane is an expected outcome, considering that both lamellae were extracted orthogonally with respect to each other.Additionally, both diffraction patterns show predominantly a streaking of spots, rather than clear concentric rings, suggesting an accumulation of high dislocation density, lattice distortions, as well as preferred orientation of grains [38,78].
The findings from SAED pattern analysis strongly support the ACOM results as demonstrated by the complementary pole figures both in Figs. 12 and 13, which are based on the orientation maps shown in Fig. 11.Overall, the ACOM and SAED analysis for both longitudinal and hoop plane strongly suggest an alignment of basal planes parallel to FD, whereas 1st order prism planes reside orthonormal to FD and therefore describe a well-developed {1010} fiber texture [79].The hot spot of the longitudinal plane (0001) pole figure (Fig. 12(d)) shows a strong presence of intermediate grain orientations, as confirmed by the SAED analysis.On the other hand, the hoop plane pole figure (Fig. 13(d)) hot spot resides in the center of the pole, indicating that the majority of observed orientations belong to basal plane orientations.This is in good agreement to the high fraction of red colored grains observed in Fig. 10 (b) indicating a stable orientation toward basal.It should be noted, that even though the lamellae were extracted from orthogonal positions respectively, a complete shift of the hot-spot in the (0001) pole figures of 90 • between both lamellae was not observed.This can be attributed to preparational artifacts (e.g., if a lamella bent from releasing internal tensions during the ion thinning process or if the lamella was fixed not completely horizontal to the carrier grid).
The 1st order prism (1010) pole figures (Figs. 12(d) and 13(d)) depict a strong texture with 1st order prism planes oriented parallel to the FD.This {1010} fiber texture is typical for cp-Ti processed by RS and/or drawing [55,61].In contrast to these previously demonstrated textures, however, a stronger basal texture intensity is evident in Fig. 12(d) and 13(d).Strong basal texture has been demonstrated during ECAP of cp-Ti [80].Hence, it is possible that processing by ECAP prior to RS contributed to the formation of basal texture, given that initial texture is known to impact the final texture during SPD processing [81,82].Nevertheless, Fig. 14.Inner reflex positions of simulated zone axis ED patterns (left) with associated unit cell orientations (middle) as well as corresponding hexagonal prisms (right) of benchmark zone axis orientations (a) {0001}<1010>, (b) {1210}<1010> and (c) {1213}<1010>.Reflection positions used for deriving the respective zone axis orientations are marked red.Unit cell models and corresponding coordinate systems are displayed in perspective view.For reference, unit cell boundaries are integrated in hexagonal prisms as thick dashed lines.Hexagonal prism faces associated with equivalent hexagonal lattice planes were colored red ({0001}, basal plane) and blue ({1010}, 1st order prism planes).ED pattern simulations were performed with JEMS [76].Structural drawings were performed by help of the program VESTA3 [77]. in addition to grain refinement, the resulting texture for cp-Ti processed by combined ECAP/RS would influence the material's mechanical properties, albeit through distinct mechanisms.
There are four possible slip systems on three slip planes for cp-Ti.These include <1120> slip on the prismatic plane {1010}, <1120> slip on the basal plane {0001}, as well as <1120> and < 1123> slip on the pyramidal plane {1011} [61,83].Activation of these slip systems is dependent on loading orientation, as governed by the Schmid factor.Although Schmid factors were not measured in this work, a study reporting similar texture from RS [55] calculated highest Schmid factors for prismatic slip {1010}<1120> and pyramidal slip {1011}<1120>, for a loading direction parallel to the rod axis (i.e., same loading as the current study).The authors concluded that the soft orientation of both the prismatic and pyramidal slip systems relative to the tensile axis was instrumental in maintaining moderate ductility for RSed cp-Ti, while the UFG microstructure contributed significantly to the high strength.Given the similar {1010} fiber texture and moderate ductility generated from combined ECAP/RS, it is reasonable to suggest a similar explanation for the results of this research.In addition, the higher tensile strength of cp-Ti grade 4 measured in this study can be attributed to the evolution of strong basal texture.When measuring Schmid factors, Wang et al. [55] determined a zero value for basal slip {0001} <1120>, so given the Fig. 16.Distribution of grain size depicting median grain diameter along the longitudinal plane of the Swage-300 sample and corresponding to a second orientation map with similar dimensions.The particle fraction >700 contributes 1 % to the total number fraction.Fig. 17.Distribution of grain size depicting median grain diameter along the hoop plane of the Swage-300 sample and corresponding to the orientation map shown in Fig. 11 (b).The particle fraction >700 contributes 1 % to the total number fraction.similar RS processing and loading conditions, it is likely that basal slip was hindered during combined ECAP/RS leading to strong texture and enhanced mechanical strength for cp-Ti grade 4.
The strong crystallographic texture also suggests the potential for significant anisotropy of mechanical properties.In this research, tensile testing was only carried out in the direction of processing (i.e., along the swaged rod axis), and thus, cannot be used to confirm this hypothesis.However, the microhardness measurements in Fig. 2 illustrate some anisotropy, with higher microhardness values along the longitudinal direction (rod axis).Nevertheless, despite possible anisotropy, fatigue strength of cp-Ti grade 4 was improved under bending.Texture was shown to influence bending stress for an hcp Mg alloy [84].The authors found that under bending, samples with weak basal texture possessed much lower bending stress that those with strong basal texture.This suggests that basal texture may have also contributed to the improved fatigue strength observed in this study.Additionally, Kitahara et al. [85] observed a strong effect of crystallographic texture on fatigue crack initiation, demonstrating a greater delay in crack initiation for loading orientations consistent with ease of pyramidal slip.Thus, the {1010} fiber texture evolving from combined ECAP/RS processing of cp-Ti grade 4 may have enhanced fatigue strength through a delay in crack initiation.Further research is needed however, to better relate texture to fatigue properties.

Grain size and grain boundary misorientation.
Grain size distribution charts are presented in Figs. 15 to 18.The charts in Figs. 15 and 17 were generated with OIM from the same areas as those for the grain orientation maps in Fig. 11, whereas the charts in Figs.16 and 18 correspond to additional ACOM mapped areas for each TEM lamella, respectively.Grain size is given as the median diameter of grains with a respective misorientation larger than 5 • .Average median grain size was recorded as 125 nm and 135 nm from mappings across the longitudinal plane, as well as 119 nm and 105 nm for the grains across the hoop plane, respectively.A general decreasing trend for grain diameter is seen along both planes.The high fraction of grains with grain size <100 nm confirms the presence of UFG microstructure, thereby demonstrating the Fig. 18.Distribution of grain size depicting median grain diameter along the hoop plane of the Swage-300 sample and corresponding to a second orientation map with similar dimensions.The particle fraction >700 contributes 0.75 % to the total number fraction.potential of combined ECAP/RS processing to form such a microstructure.However, the distribution charts display a broad dispersion of grain sizes (see fraction >700 nm).This is in agreement with the observation of the heterogeneous microstructure in the grain orientation maps (Fig. 11) and STEM-BF images (Fig. 10).Since several grains were clipped by the measurement field of view (technical limitation by maximum scanning field size), complementary measurements were conducted manually for determining the horizontal diameter of grains in   both longitudinal and hoop planes.Fig. 19 shows results from horizontal measurements conducted at ¼, ½ and ¾ scanning field height for the grain orientation maps in Fig. 11, as well as for one additional ACOM mapping per sample plane.The results of such horizontal measurement closely resemble the grain size distribution obtained from OIM for both sample planes, although diameters above 700 nm were not detected in the manual measurements.Average horizontal grain diameter size was recorded as 131 nm (hoop diameter) and 116 nm (longitudinal diameter) for the grains across the longitudinal and hoop planes, respectively.These recorded numbers are in agreement with the results from statistical analysis performed with OIM.
Fig. 20 presents the distribution of grain misorientation angles corresponding to the grain maps in Fig. 11.Once again, the chart in Fig. 20 (a) corresponds to the measurements along the longitudinal plane while that in Fig. 20(b) corresponds to those along the hoop plane.The low angle grain boundaries (LAGBs) are those with misorientation angles between 1.5 and 15 • , while the high angle grain boundaries (HAGBs) are those with misorientation angles >15 • .The charts demonstrate a high fraction of LAGBs with 80 % along the longitudinal plane and 89 % along the hoop plane.These results are in good agreement with the presence of predominantly elongated grains showing a high grade of deformation within the microstructure.
The high fraction of LAGBs observed in this study is typical of cold processing for cp-Ti and is consistent with the heterogeneous grain structure (i.e., banded elongated grains and equiaxed substructure) observed in Figs. 10 and 11.Moreover, a high fraction of LAGBs is indicative of accumulated dislocation densities and lattice distortion, as confirmed by SAED in Figs. 12 and 13.The high fraction of LAGBs suggest that, in addition to grain refinement and solid-solution strengthening, substructure strengthening was also responsible for the significant increase in mechanical strength.This has also been demonstrated during SPD processing of cp-Ti, as reported by Dyakonov et al. [34], and hence, the results of this study are consistent with literature.
An increase in the fraction of HAGBs and subsequent transition from banded microstructure to a more equiaxed one, can likely be implemented through post-processing thermal treatments (e.g., annealing), as demonstrated in previous studies [71,72].These studies also showed improved fatigue properties for samples with a high fraction of HABGs.It is therefore worthwhile to attempt similar post-processing treatments following ECAP/RS processing.Additionally, the number of HAGBs has been shown to be dependent on various ECAP parameters including temperature and channel die angle [86,87].While the die angle is fixed, lower ECAP temperatures could be attempted as a means to potentially enhance fatigue properties of cp-Ti grade 4 during ECAP/RS processing, but challenges attributed to material ductility may hinder the processability.

Corrosion
Potentiodynamic polarization curves are presented in Fig. 21 for AR cp-Ti grade 4, Swage-300, Swage-300_etched, and finally Swage-300 with an applied PEO surface treatment (i.e., Swage-300_PEO).In addition, Table 3 lists the values of electrochemical parameters from the polarization curves for the various cp-Ti grade 4 samples, including Tafel slope of the cathodic and anodic branches, b c and b a , respectively, as well as corrosion density, I corr , and corrosion potential, E corr .Cathodic branches for the samples illustrate similar profiles and only minor differences in slopes, as shown in Fig. 21 and Table 3, respectively.As polarization measurements were started at cathodic potential (-1 V to OCP), all samples showed a positive (anodic) shift in E corr thereby indicating formation of a passive film.Minor differences in E corr were nevertheless recorded for the AR, Swage-300 and Swage-300_etched samples, with the Swage-300_etched sample demonstrating the largest anodic shift.However, in the case of the Swage-300_PEO sample, a less pronounced positive shift was recorded (E corr = − 0.822 V).This was likely due to an already present protective passive layer on the sample, generated by the initial PEO surface treatment.Upon immersion in SBF, the protective PEO surface likely inhibited corrosive attack and hence formation of a passive layer.In contrast, for the Swage-300_etched sample, given that etching was carried out to mimic surfaces typical of Ti alloy dental implants (see Section 2.4), and that such surface treatments typically result in increased surface roughness, it is likely that upon immersion in SBF, the sample was more susceptible to corrosive attack, resulting in higher potential for passive film formation (i.e., greater positive shift).
The anodic branches of the PDP curves for the cp-Ti grade 4 samples showed both similarities and differences among the samples.In the case of the AR and Swage-300 samples, an initial passive behavior was seen followed by a breakdown of the passive layer and finally re-passivation, as illustrated by the blue and gray curves in Fig. 21.This ultimately led to similar I corr values for both samples, thereby suggesting that combined ECAP/RS processing had no significant impact on corrosion resistance of cp-Ti grade 4.This is consistent with some of the literature reporting on corrosion properties of UFG cp-Ti [88,89].In other instances, UFG microstructure generated from SPD processing was found to significantly improve corrosion resistance of cp-Ti.This has been attributed to, e.g., fast generation of passive layers at grain boundaries where passivation films nucleate [80,90].A large number of grain boundaries in UFG Ti enhance the passivation kinetics, leading to improved corrosion resistance.This is difficult to confirm in this research, but given the similarity in passive layer breakdown along the anodic branch of the PDP curves for both the AR and Swage-300 samples, it is unlikely that this was the case.Additionally, Balyanov et al. [90] suggested that the improvement in corrosion resistance of UFG Ti after ECAP was due to increased segregation of impurities to grain boundaries in coarsegrained Ti.Nevertheless, the results of corrosion resistance for cp-Ti have shown a strong dependence on corrosion medium, thereby leading to conflicting results regarding the effect of grain size.
In the case of texture, most studies [80,91,92] reported enhanced corrosion resistance for UFG cp-Ti samples having strong basal texture, in comparison to, e.g., shear texture generated from ECAP processing.The underlying reasons therefor are not clear, however.The effect of texture was not examined in this research and should be further investigated in future studies.
The anodic branches of the surface-treated samples showed specific differences.For the Swage-300_etched sample, a similar profile as that for the AR and Swage-300 samples was observed, but a larger breakdown of the passive layer was recorded, as depicted by the red curve in Fig. 21.This is indicative of increased electrochemical pitting occurring along the surface of the etched sample.The increased surface roughness attributed to such surface treatments typically promotes electrochemical pitting when immersed in corrosive environments, and has been previously demonstrated for SLA-treated Ti alloy specimens [93,94].Finally, distinct differences were recorded along the anodic branch for the Swage-300_PEO sample, as demonstrated by the green curve in Fig. 21 and the values of b a in Table 3. Specifically, the surface of the sample initially demonstrated a strong passive behavior followed by a slight breakdown in potential and almost immediate re-passivation.This suggests good stability of the PEO-treated surface during immersion in SBF.

Biocompatibility
The results of the indirect XTT assay with the respective controls are shown in Fig. 22.Each test group demonstrated a distinct and significant tendency toward the negative, non-toxic control.There were no significant differences among the groups themselves.Within the DIN EN ISO 10993-5:2009 norm, the nontoxic range is defined to lie within a deviation of 30 % above or below the negative control, which was easily accomplished in this case.The results of the indirect LDH assay are shown in Fig. 23.In accordance with the XTT assay, the test groups within the LDH assay demonstrated low absorbance values similar to the negative control, hereby indicating the absence of cytotoxic effects.To assess the direct influence of the material modifications on the vitality of cells and concomitantly investigate cell attachment with respect to the different surface conditions, live-dead staining was performed.The results are shown in Fig. 24.While the nontoxic control material was densely covered with living, spindle-shaped cells, hardly any live cells were found on the toxic control.Overall, no dead cells were found on the surfaces of either of the material groups.Cell density was slightly lower for all groups in comparison to the negative control, although there did not seem to be any differences with respect to cell density among the groups themselves.In general, cells were homogeneously distributed among the surfaces in all cases, the etched group showing only few areas with diminished density.Cell morphology was spindle shaped in almost all cases, hereby indicating cell well-being and sufficient cell attachment.Only few roundly shaped cells were observed within the etched group.Hence, the results confirm no detrimental effect on biocompatibility for cp-Ti grade 4 following combined ECAP/RS processing.

Conclusions
This study demonstrated high potential for improving the mechanical properties of cp-Ti grade 4 through a two-stage processing consisting of ECAP and RS, thereby strengthening the candidacy of cp-Ti grade 4 for replacement of potentially toxic Ti64 in advanced dental and orthopedic implant applications.The following conclusions can be drawn from this study: i.Combined ECAP/RS processing resulted in an average yield strength and UTS of 1383 MPa and 1396 MPa, respectively, while maintaining moderate ductility of 9.7 %.This exceeded the static tensile strength of Ti64 and is, to date, the highest strength recorded for cp-Ti grade 4. As a consequence, the fatigue endurance limit increased up to 600 MPa for cp-Ti grade 4, coming within only 6 % of that for Ti64 (640 MPa).ii.The increase in tensile and fatigue properties for cp-Ti grade 4 was attributed to the formation of an ultrafine-grained microstructure with a minimum of 65 % of grains having diameters <100 nm, and average grain size ranging from 105 to 135 nm, depending on sample orientation.Additional strengthening  mechanisms were attributed to a strong fiber texture and a high fraction (> 80 %) of low angle grain boundaries for cp-Ti processed by combined ECAP/RS.Such low angle grain boundaries were due to accumulation of dislocations, high lattice distortion and formation of substructures within the microstructure of UFG cp-Ti grade 4. iii.Potentiodynamic polarization testing confirmed that the corrosion resistance of cp-Ti grade 4 was virtually unaffected by combined ECAP/RS processing.Although, some differences were observed in anodic behavior for samples with surfaces treated by plasma electrolytic oxidation, and sand blasting and etching.iv.Finally, the results of indirect XTT assay and live dead staining confirmed no detrimental effect on biocompatibility for cp-Ti grade 4 processed by ECAP/RS either, including for samples subjected to surface treatments.
The results indicate that novel ECAP/RS processing enables a new class of ultra-fine grained commercial purity titanium (UFG cp-Ti grade 4) that offers superior static and dynamic mechanical properties, while at the same time waiving any changes in other important measurements for medical implants, such as corrosion resistance or biocompatibility.
Hence, this new class of implant material might offer the ability to advance dental or orthopedic implants with optimized mechanical properties or smaller geometrical features as well as potentially allowing for a waiver in using Ti alloys containing any elements of concern.
Future research will investigate the potential for implementing postprocessing thermal treatments that can aid in promoting less banded, and more equiaxed grains, higher fraction of high angle grain boundaries and further enhancement of fatigue properties for cp-Ti grade 4 rods processed via combined ECAP/RS.Visualization.

Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Fig. 8 .
Fig. 8. Fracture surface of specimen 1 taken from Swage-300 sample after rotating bending fatigue tests (σ max = 700 MPa; N = 28,943 cycles): (a) general view of the fracture surface (the white arrow depicts the direction of crack propagation); (b) zone of crack propagation.

Fig. 9 .
Fig. 9. Fracture surface of specimen 2 taken from Swage-300 sample after rotating bending fatigue tests (σ max = 625 MPa; N = 9.4 × 10 6 cycles): (a) general view of the fracture surface (the white arrow depicts the direction of crack propagation); (b) zone of crack propagation; (c) zone of rapid crack propagation.

Fig. 10 .
Fig. 10.STEM bright field micrographs of thin film specimen (FIB-Lamellae) cross-sections extracted from the central region of the Swage-300 sample along (a) the longitudinal plane and (b) the hoop plane in respect to the bar axis (FD) as illustrated by the inlets (top right).Pointers below the images show the established inplane workpiece directions.

Fig. 11 .
Fig. 11.Grain orientation maps for the Swage-300 condition of the (a) longitudinal cross-section and (b) tangential cross section with respect to the bar axis (FD).Orientation maps on the left represent misorientations with respect to the transmission direction, i.e.(a) along the normal direction of the and (b) the transversal direction of the cross section.Maps on the right represent misorientations with respect to the flow direction.Scale bars correspond to the length of 1 μm.Black lines indicate grain boundaries assigned by ACOM between diffraction patterns showing misorientations >5 • for grains >50 nm (median diameter).

A
.Kopp et al.

Fig. 12 .
Fig. 12.(a) SAED pattern for Swage-300 projecting longitudinal plane grain orientations of the swaged bar.(b) Table showing Bragg reflections indexed according to crystallographic information file (CIF) data.Reflections d 1010 , d 0002 and d 1011 are marked red in (a).(c) Qualitative lattice reflection contribution according to benchmark ZA orientations derived from the Bragg reflections listed in (b).Hexagonal prism models illustrate the crystallographic orientation of unit cell (dashed lines) as well as basal ({0001}, red) and 1st order prism plane ({1010}, blue) orientations associated with the shown ZA orientations.(d) Complementary pole figures depicting 0001 (basal) and 1010 (1st order prism) plane orientations extracted from the corresponding orientation map in Fig.11a).The 3D models below the pole figures display the preferred orientation according to the PF hot spots.The colored prism faces correspond to the basal ({0001}, red) and 1st order prism planes (longitudinal plane) and Fig. 13 (hoop plane).The SAED patterns display Bragg diffraction patterns from grains over an integrated area size comparable to the SEND area dimensions and are representative of the respective lamella orientation.The first three Bragg reflection positions are marked according to the interplanar lattice spacing d 1010 , d 0002 and d 1011 of hexagonal close packed Ti as-indexed in reference to literature data [75] and shown in the tables given in Figs.12(b) and 13(b).The lattice reflections d 1010 , d 0002 and d 1011 allow the identification of low indexed zone axes as benchmark orientations in order to dissect the diffraction data.From visual inspection it is obvious that multiple grain orientations are present in both diffraction patterns shown in Figs. 12 and 13.Furthermore, reflections form arcs, indicating a variation of azimuthal orientation of approximately 15 • in groups of similarly oriented grains.The majority of Bragg reflection intensity distribution concentrates on d 1010 , d 0002 and d 1011 as well as equivalent positions.Finally, strong intensity is observable for d 1010 reflections residing on the vertical radius (therefore || FD) in the SAED patterns of both lamellae.This suggests that the majority of the observed zone axis orientations share this particular reflex position and therefore the majority of observed grains are oriented with 1st order prism planes aligned orthonormal to FD.

Fig. 13 .
Fig. 13.(a) SAED pattern for Swage-300 projecting hoop plane grain orientations of the swaged bar.(b) Table showing Bragg reflections indexed according to crystallographic information file (CIF) data.Reflections d 1010 , d 0002 and d 1011 are marked red in (a).(c) Qualitative lattice reflection contribution according to benchmark ZA orientations derived from the Bragg reflections listed in (b).Hexagonal prism models illustrate the crystallographic orientation of unit cell (dashed lines) as well as basal ({0001}, red) and 1st order prism plane ({1010}, blue) orientations associated with the shown ZA orientations.(d) Complementary pole figures depicting 0001 (basal) and 1010 (1st order prism) plane orientations extracted from the corresponding orientation map in Fig. 11(b).The 3D models below the pole figures display the preferred orientation according to the PF hot spots.The colored prism faces correspond to the basal ({0001}, red) and 1st order prism planes ({1010}, blue).The unit cell is displayed with dashed lines.SAED projected aperture diameter is approximately 3.7 μm.

Fig. 15 .
Fig. 15.Distribution of grain size depicting median grain diameter along the longitudinal plane of the Swage-300 sample and corresponding to the orientation map shown in Fig. 11 (a).The particle fraction >700 contributes 1.5 % to the total number fraction.

Fig. 19 .
Fig. 19.Distribution of grain size depicting horizontal diameter measurements along the (a) longitudinal plane and (b) the hoop plane of the Swage-300 sample.Measurements from horizontal profiles were conducted for two orientation maps per sample plane each at ¼, ½ and ¾ height in vertical scanning field positions.The total number of grains investigated were 102 for the longitudinal and 105 for the hoop plane.

A
.Kopp et al.

Table 1
Chemical composition of as-received cp-Ti grade 4 material (in wt%).
A.Kopp et al.

Table 2
Fatigue endurance limits of as-received cp-Ti grade 4 and grade 23, and Swage-300 and Swage-475 samples.

Table 3
Electrochemical parameters from the polarization curves of various cp-Ti grade 4 samples.