Chemical Polishing of Additively Manufactured, Porous, Nickel–Titanium Skeletal Fixation Plates

Abstract

Nickel–titanium (NiTi) alloys have shown promise for a variety of biomedical applications because of their unique properties of shape memory, superelasticity, and low modulus of elasticity (Young's modulus). Nevertheless, NiTi bulk components cannot be easily machined (e.g., CNC, rolling, grinding, casting, or press molding) due to their thermomechanical sensitivity as well as inherent superelasticity and shape memory. Thus, powder bed fusion (PBF) additive manufacturing has been used to successfully fabricate NiTi medical devices that match the geometric and mechanical needs of a particular patient's condition. However, NiTi PBF fabrication leaves unmelted particles from the source powder adhered to external surfaces, which cause minor dimensional inaccuracy, increase the risk of mechanical failure, and once loose, may irritate or inflame surrounding tissues. Therefore, there is a need to develop a chemical polishing (cleaning) technique to remove unmelted powder from the surfaces of PBF-fabricated implants, especially from inner surfaces that are difficult to access with mechanical polishing tools. This technique is especially useful for highly porous devices printed at high resolution. In this study, a chemical polishing method utilizing HF/HNO₃ solution was used to remove loosely attached (i.e., unmelted) powder particles from surfaces of porous, skeletal fixation plates manufactured by PBF AM. It was observed that 7 min of polishing in an HF/HNO₃ solution comprising 7.5 HF: 50 HNO₃: 42.5 H₂O enabled successful removal of all relatively loose and unmelted powder particles. A microcomputed tomography study examination found that the volumetric accuracy of the polished skeletal fixation plates was ±10% compared with the computer-aided design (CAD) model from which it was rendered. This postprocessing chemical polishing protocol is also likely to be useful for removing loose powder, while maintaining CAD model accuracy and mechanical stability for other complexly shaped, porous, three-dimensional (3D), printed NiTi devices.

Introduction

Nickel–titanium (NiTi) (also known as nitinol, NiTi alloy) is a shape memory alloy that is widely used in aerospace, civil, and biomedical applications.^1,2–7 NiTi alloys have tunable superelasticity and shape memory properties. These two unique properties are mainly due to the fact that NiTi alloys are able to undergo a reversible solid-state phase transformation in response to stress or a change in temperature between the austenite and martensite phases.^8,9 Moreover, this material exhibits high corrosion resistance and is biocompatible when properly alloyed.⁷ However, shape memory and superelasticity properties, as well as its high reactivity and ductility,¹⁰ make machining and processing of NiTi a challenging task. Therefore, the use of conventional manufacturing methods to render complex geometries is imprecise, expensive, time-consuming, and in some situations impossible.^11–13 However, additive manufacturing (AM) methods are relatively new, attractive alternatives for fabrication of complex geometries from a wide range of materials, including NiTi,^12,14 titanium,^15,16 nickel,^17,18 and other alloys,^19,20 directly from a computer-aided design (CAD) file. Thanks to AM techniques for these metals, complexly shaped lattice or pore structures can be achieved that enable the manufacture of lightweight devices with engineered mechanical properties.²¹

In recent years, these methods have been adapted for fabricating patient-specific implants and medical devices.^9,21–24 Due to the high difference in Young's modulus between bone and conventional implants, primarily fabricated from Ti-6Al-4V, stress shielding and subsequent bone resorption may occur. Stress shielding may lead to bone resorption if the medical device leads to insufficient loading of the reconstructed bone.^25,26 Among other metallic medical materials, NiTi has a relatively low Young's modulus and thus low stiffness. NiTi Young's modulus ranges between 55 and 80 GPa, which is closer to compact bone (12–17 GPa) than for most commonly used titanium alloys ∼110 GPa, stainless steel ∼190 GPa, and cobalt-based alloys ∼210 GPa.²⁷ Implant stiffness, which depends on its geometry and Young's modulus, can be reduced by introducing porosity through conventional methods^28,29 or by the use of multiple materials.³⁰ Nevertheless, conventional manufacturing methods may not provide sufficient control over strut and pore geometry and thus the mechanical properties, especially the attempt to surgically reconstruct normal stress–strain trajectories, may fail. Therefore, to solve these problems, AM technologies have recently been extensively explored to produce structures called scaffolds with engineered porosity and mechanical properties from biocompatible metal materials,^7,31–33 polymers,^33–35 ceramics, and composites.^36,37

Among metal AM techniques, we can distinguish powder bed fusion (PBF) (e.g., selective laser melting [SLM] and electron beam melting) and flow-based solutions (e.g., laser engineering net shaping or direct meal deposition).³⁸ SLM PBF has, to date, attracted the most attention and use for creation of complexly shaped NiTi-based parts.^11,13,39 Current SLM PBF devices achieve rendering accuracy of up to 20 μm.^40,41 This high accuracy allows parts to be built with confidence. However, setup parameters must be optimized to allow achieving the necessary or, if needed, full accuracy.

One of the challenges is that PBF-fabricated metal parts have loose, unmelted powder particles fused to all external surfaces. This phenomenon is explained by the nature of the melt pool in PBF fabrication. The high-energy laser produces sufficient heat to create the melt pool at the surface of the powder bed.⁴² The temperature surrounding the melt pool is immediately increased by heat conduction. However, by design, particles in this zone do not melt, but may become lightly sintered to the surface of the part.^43–45 These loosely adhered particles may be released into adjacent tissue and can cause minor differences between the intended dimensions of the CAD model and the dimensions after the fabrication process and can lead to undesirable and unexpected mismatch with the modeled mechanical properties, especially in parts with small features.^16,21,46 Rough surfaces created by partially melted powder particles also increase stress concentrations and can cause early fatigue failure of the components. Thus, for these reasons, the loose powder particles sintered to the surfaces in PBF fabrication are undesirable and postprocessing is necessary to remove them.

Traditional surface treatments for biomaterials include mechanical polishing, sandblasting, and shot peening^47–49; however, in areas where minute porous spaces do not allow for the application of mechanical cleaning devices, these methods are not feasible. In these instances, it may be possible to adequately remove loose powder particles through chemical and/or electrochemical polishing methods. During chemical polishing processes, the part's material and the applied etchant undergo a chemical reaction that leads to removal of loose powder particles. The success of chemical polishing depends on the extent of this reaction, which in turn depends on the applied etchant, temperature of etching, duration of etching, and applied agitation. For titanium and its alloys, chemical polishing in HF/HNO₃ (hydrofluoric acid/nitric acid) solution is well documented and has been extensively utilized.^15,50–52 Usually, this method involves the fabricated part being immersed in the cleaning solution to dissolve the area where powder particles are loosely sintered to the surface of the part. The goal of this activity is to remove loose powder, obtain the dimensions seen in the CAD file, and achieve the predicted short- and long-term mechanical properties (Fig. 1).⁵² Additionally, if needed, the particle removal process can be enhanced with an ultrasonic bath device.⁵⁰ The proper composition of the solution, which allows effective removal of powder particles, and the appropriate polishing time must be determined experimentally. In addition, when the printed parts contain very fine features, the composition of the etchant and the duration of the procedure are critical parameters because excessive polishing will lead to removal of small features. Therefore, the outcome of the chemical polishing procedure must be carefully evaluated to make sure that critical features are not removed during polishing.

FIG. 1.

Chemical polishing postprocessing for removal of (a) unmelted powder particles sintered to the surface of a 3D printed skeletal fixation plate. (b) Powder particle dissolution in an appropriate cleaning solution is enhanced by the addition of ultrasonic wave energy. (c) Smooth part surface free of unmelted sintered particles. (d) Visualization and dimensions of the studied skeletal fixation plate CAD model with the scheme of functional components. 3D, three-dimensional; CAD, computer-aided design.

Chemical etching is commonly utilized to process NiTi and other metallic parts. In the case of NiTi, it serves to remove surface contaminants, defects, and heavy deformation zones after cold drawing.⁵³ Moreover, previous studies have shown that chemical processing of NiTi is known to be efficient for elimination of defective surface layers and surface oxidizing,^54,55 as well as surface structuring.⁴⁹ Additionally, chemical surface modification of NiTi elements for biomedical applications is performed to reduce cytotoxic nickel ion release from the surface. Furthermore, chemical polishing can roughen a smooth surface. It should be noted that two different types of roughness are observed on SLM-fabricated parts. The first type of roughness is created by partially melted powder particles adhering to the external surfaces. As previously noted, this is undesirable because the loose powder fused to the surface of medical devices may be released into adjacent tissues and could lead to toxicity, infection, and allergic or foreign body reactions. The second type of roughness is generated by mechanical⁵⁶ and chemical methods^31,51 or by providing coatings.⁵⁷ This may enhance cell adhesion and proliferation and overall biointegration, which is desirable for some medical applications.^53,58 Chu et al.⁵⁹ used chemical etching of NiTi in HCl-HNO₃ pickling solution to remove struts created during a laser cutting process. Indeed, chemical polishing of NiTi has been mainly used for conventionally manufactured components. Chemical treatment of NiTi elements made by AM methods to remove unmelted powder particles from the surface and, more importantly, to verify the CAD dimensions and predicted mechanical function of a three-dimensional (3D) printed NiTi part using microcomputed tomography (μCT) scanning, to our knowledge, has not been previously done. Thus, the aim of this study was to develop a chemical polishing procedure to remove loosely sintered, unmelted powder particles from the surfaces of NiTi parts fabricated through an SLM PBF process. This protocol was tested to evaluate the possibility of ensuring CAD model dimensions and to obtain a mechanically reliable and biomedically safe skeletal fixation device surface.

Materials and Methods

Fabrication procedure

Skeletal fixation plates with mechanical properties desirable for mandibular graft fixation were achieved, in part, by including custom-designed porosity. An SLM machine (3D Systems PXM, Rock Hill, SC), equipped with a 300 W ytterbium fiber laser, was used to produce Ni-rich, NiTi skeletal fixation plates. A slightly Ni-rich Ni50.8Ti49.2 (at%) ingot was subjected to electrode induction melting gas atomization by TLS Technik GmbH (Bitterfeld, Germany) to produce the NiTi powder. To ensure adequate layer resolution and powder flow, particle size ranging from 25 to 75 μm was used. To minimize the level of impurity content within the resultant part, the oxygen level was decreased to 500 ppm before fabrication. Process parameters used in the fabrication are shown in Table 1 (energy input is defined as E = p/(v.h.t)).^32,33,60 The shape and dimensions of the fixation plate studied are shown in Figure 1d.

Table 1.

Selective Laser Melting Process Parameters for Fabrication of Skeletal Fixation Plates

Laser power (p) [W]	Scanning speed (v) [mm/s]	Hatch spacing (h) [μm]	Layer thickness (t) [μm]	Energy input (E) [J/mm3]
250	1250	80	30	83.33

Polishing procedure

The NiTi fixation plates were ultrasonically cleaned in deionized water before immersion in a beaker with HF/HNO₃ acid solutions. The HF/HNO₃ solutions used to remove attached (i.e., unmelted and sintered) powder particles were as follows: solution 1: 10 HF: 40 HNO₃: 50 H₂O (S1); solution 2: 3.3 HF: 13.9 HNO₃: 82.8 H₂O (S2); and solution 3: 7.5 HF: 50 HNO₃: 42.5 H₂O (S3). The composition of solutions S1 and S2 was selected on the basis of the literature,^31,59,61 while the S3 solution composition was based on characterization of the fixation plates polished in solutions S1 and S2. The assessment of solutions S1 and S2 was done with partial plate sections (Fig. 1d), while full-size skeletal fixation plates were used for the analysis of polishing using solution S3. Solution ratios, polishing times, and reagent concentrations are presented in Table 2. Moreover, additional polishing in solution S3, referred to as repolishing, was applied to samples that were also studied by μCT scanning (Table 3). Repolishing was done to achieve a total polishing time of 7 min. This experiment allowed us to determine whether repolishing could be used successfully with NiTi plates as when the process of polishing is interrupted, an insufficient amount of powder would be removed. All of the polishing and repolishing experiments were carried out in an ultrasonic bath at room temperature. After all chemical polishing procedures, polished samples were rinsed with DI water and cleaned in an ultrasonic bath of DI water. Each sample was ultrasonically cleaned three times for 10 min to remove all remaining acid traces.⁶² The DI water was replaced after each iteration of ultrasonic cleaning.

Table 2.

Chemical Composition of Test Solutions and Polishing Time

Solution	Component volume [mL]			Polishing time [min]
	HF (40%)	HNO3(60%)	H2O
S1	10	40	50	3, 4, 5, 6, 7
S2	3.3	13.9	82.8	3, 6, 12
S3	7.5	50	42.5	3, 4, 5, 6, 7, 8, 9, 10

Table 3.

Volume Measurement of Samples Before and After Polishing in Solution S3

As-made		Polishing			Repolishing (polishing+additional polishing)
Sample No.	Volume [mm3]	Etching time [min]	Volume [mm3]	Volume loss [%]	Etching time [min]	Volume [mm3]	CAD accuracy [%]
CAD model	60.2
1	93.8	3	77.6	17.3	7 (3 + 4)	65.7	109.1
2	93.4	4	66.9	28.4	7 (4 + 3)	54.2	90.1
3	94.8	5	64.7	31.8	7 (5 + 2)	63.4	105.3
4	94.4	6	64.2	32.1	7 (6 + 1)	61.2	101.7
5	93.3	7	58.3	37.5	7	58.3	96.9
6	98.5	8	23.6	76.1
7	93.7	9	23.5	74.9

CAD, computer-aided design.

The polishing times were adapted according to the observations of polishing speed, as determined by the intensity of gas released from the plate's surface during the immersion.

Characterization methods

The mass loss owing to chemical polishing was determined by taking each specimen's weight before and after the chemical polishing procedure. More specifically, each fixation plate's mass was nonconsecutively measured three times before and after the chemical polishing. The samples' surfaces were observed using scanning electron microscopy (SEM) on an Apreo LoVac HR scanning electron microscope (Thermo Fisher, Waltham, MA). μCT scans were performed using a HeliScan (Thermo Fisher Scientific, Waltham, MA) μCT scanner with a voxel size of 7.6450 μm, before and after chemical polishing, to verify the difference in sample volume before and after subsequent stages of chemical treatment in the final solution. Source voltage and source current were set to 100 kV and 45 μA, respectively. A 2-mm aluminum filter material was chosen. The scanning procedure was carried out by performing a rotation of an emitted X-ray by 180°, 1800 projections per resolution, and an exposure time of 0.25 s per projection. The accuracy of the rendered CAD model was evaluated based on a comparison of the standard tessellation language (STL) model volume with the original (used for SLM PBF fabrication) CAD model volume.

Results

Mass loss measurement

Polishing of fixation plate sections was performed in solutions S1 and S2, while full skeletal fixation plates were polished in solution S3. Mass measurements indicated that for all tested solutions, mass loss increased in relation to increasing chemical polishing time. Chemical polishing in solution S1 caused the greatest mass loss, while mass lose due to polishing in solution S2 was significantly less, including after a much longer soak time. The weight of samples polished in solution S3 decreased slightly faster than for samples polished in solution S2 and much more slowly than for solution S1 (Fig. 2). Complete dissolution of the plate was observed after 7 min of polishing in solution S1 and after 10 min of polishing in solution S3.

FIG. 2.

Percentage weight loss results during polishing in three different solutions: S1, S2, and S3.

SEM observations

SEM pictures of unpolished and polished fixation plates are shown in Figure 3. Each HF-HNO₃ solution used for the polishing experiments acted in a different way on the loosely sintered powder particles on NiTi plate surfaces. It was observed that solution S1 caused rapid dissolution of loosely sintered powder particles; nevertheless, linear edges (Fig. 3) were not eliminated. The acid concentration in solution S2 was too low to remove all loosely sintered particles from the surface, even after relatively long polishing times. Solution S3 removed all loosely sintered particles from the surface and left it smooth after 7 min, without causing excessive dissolution of the processed material.

FIG. 3.

SEM images of the as-made sample and samples polished in solutions S1, S2, and S3 for a specified time. SEM, scanning electron microscopy.

μCT validation

μCT scans and volume reconstructions were performed for the plates before and after the chemical polishing process. Volume loss measurements are summarized in Table 3 and presented in Figure 4. Three-dimensional reconstructed models were created as STL files from the original μCT slice data. By comparing the STL files of samples before and after polishing, the influence of the polishing time on volume was determined. The unpolished, μCT-derived surface image files were overlaid on the relevant original CAD files initially to study the rendering accuracy, especially of the most detailed and fine features, for each skeletal fixation plate. The CAD file and three μCT scans of polished plates are shown in Figure 5.

FIG. 4.

Volume change during polishing in solution S3.

FIG. 5.

(a) CAD model of the sample and μCT visualization of (b) the as-made sample. (c) Sample polished for 7 min in solution S3. (d) Overetched sample polished for 9 min in solution S3. μCT, microcomputed tomography.

Discussion

Mass loss measurement

During polishing in solution S1, intense gas release was observed at the surface of each sample throughout the full range of applied immersion times. After 7 min of polishing in solution S1, the sample was damaged and almost completely dissolved and its mass loss was above 95%. Alternatively, there was almost no gas release from the sample surface during polishing in solution S2, even for 12 min of immersion, and there were no changes visible to the unaided eye in the sample appearance during immersion. Mass measurement following soaking in solution S2 showed minimal to modest weight loss, which was never greater than 12%, throughout the study. Polishing in solution S3 did not cause intense gas release and slow and uniform change in volume was observed through the first 7 min of immersion. However, rapid mass loss was observed after 7 min of polishing in solution S3. Furthermore, polishing times longer than 9 min resulted in complete dissolution of the parts in solution S3. Likewise, 6 min of polishing in solution S1 resulted in a 50% weight loss and after 7 min, the sample was almost completely dissolved. This may indicate that for both the S3 and S1 solutions, at the initial stage of polishing, the connection sites of loose and unmelted particles dissolve and this results in an almost constant weight loss. After the particles have been completely removed from the solution, the solid part itself begins to dissolve, which results in rapid weight loss, something that did not occur with specimens soaked in solution S2. The results of the weight loss measurements show that the weight of samples polished in solution S3 decreased much slower than for samples polished in solution S1 (Fig. 2). Furthermore, it was observed that surfaces of the polished samples became shinier as soak time increased for all three solutions studied.

SEM observations

SEM observations of samples polished in solutions S1 and S3 showed that polishing times less than 6 and 7 min, respectively, did not provide removal of all powder particles and sharp edges from the surface. On the other hand, for both solution S1 and solution S3, longer soak times eventually led to strut damage and dissolution. Although much loose powder adhered to all parts after SLM PBF printing, all loosely annealed powder particles were removed after 4 and 7 min of polishing in solution S1 and solution S3, respectively. We also observed that in both solution S1 and solution S3, larger plate surfaces became rough and sharp edges also became visible. The sharp edges are not desirable features and do not meet the requirements of the postprocessed fixation device because they can irritate adjacent tissues and they appear with other negative features (pitting pores and the remaining unmelted powder particles). These sharp edges can lead to stress concentrations that propagate cracks, especially because they occur in solutions with intense hydrogen generation. Due to this risk, it is recommended that sharp edges be removed from the surface of the part. Polishing in solution S1 was severe and too aggressive for small struts. It caused strut cracking and splitting, which was also observed for solution S3 after 8 min of polishing. It can be observed that 7 min of polishing in solution S3 ensured full powder particle removal, while maintaining adequate surface smoothness, and longer polishing times left the part overetched. Polishing in solution S2 for 12 min did not completely remove loose particles from the plate's surface. Therefore, the acid concentration in solution S2 was also deemed to be too low for the devices that were tested.

It has been reported that the HF/HNO₃ ratio has the greatest influence on homogeneous removal of unmelted powder particles from pure titanium scaffolds made with the SLM PBF process.⁶² This phenomenon might be explained by different changes in the surface chemical composition after soaking in HF and HNO₃ acids. It has also been reported that HF works as a dissolvent, while HNO₃ mostly passivates the surface of the processed material.^51,63 Properly selected proportions of both acids provide a balance between etching and passivation of the material so that uniform dissolution is observed over the entire surface, which is then smooth and free of rough patches or sharp edges. The HF/HNO₃ ratio in solutions S1 and S2 was similar, close to 0.25. However, the dissolution of the material in both solutions varied due to the different HF and HNO₃ concentrations. Our analysis showed that the concentration of HF and HNO₃ acids in solution S2 was too low to fully remove loose particles from the surface in a relatively short time, while the concentration of acids in solution S1 was too high and it was difficult to avoid surface damage. The ratio of HF/HNO₃ in solution S3 was decreased to 0.12. The SEM analysis revealed that solution S3 successfully removed the unmelted powder particles from part surfaces. A smooth and powder-free surface was generated after 7 min of polishing. Polishing in solution S3 resulted in slower material dissolution than in solution S1, thereby improving surface smoothness. Based on results obtained from the experiments reported here, solution S3 was selected to be studied in the next step of the polishing evaluation by μCT.

μCT validation

The μCT reconstruction and volume measurements showed that the volume of as-made samples increased by more than half on average compared with the CAD model. That is unacceptable for elements with predesigned mechanical properties and likely poses a risk for the release of large amounts of loosely sintered powder particles into surrounding tissue. The results of volume measurements for polished samples showed that sample volume gradually decreased as the polishing time increased (Fig. 4). During 4–6 min of polishing, the volume of the elements changed very little and was close to the CAD model volume. However, by comparing these results with SEM observations, it is observed that the surface smoothness is not acceptable for biomedical applications that require a reliable feature shape for immediate biomechanical function, long-term fatigue strength, and no possibility of loose powder invading adjacent tissues. After 7 min of solution S3 polishing, the volume reduces and becomes closest (96.9%) to the volume of the CAD model. At the same time, comparing the results with SEM images, we can see that the surface smoothness has significantly improved and this amount of time is optimal for SLM PBF fabricated plates. Furthermore, polishing longer than 7 min caused a significant and detrimental dissolution of the plate material with either solution S1 or solution S3, resulting in unacceptable volume loss of more than 60% (Fig. 4) as well as gaps in some of the struts and formation of excessive surface pitting (Fig. 5). During the first 4 min of polishing, a slight decrease in the plate surface area was observed. After this time, the surface area starts to increase and after 7 min, it is almost 6% larger than for the as-made sample (Table 4). The rapid volume loss after 7 min of polishing may be due to the increased surface area of the bulk sample surrounded by the chemical solution after removing all unmelted powder particles from its surface. We consider it most likely that after 7 min of polishing in S3, the loose powder particles were completely dissolved from the plate inner and outer surfaces (Fig. 5c), and when they had completely disappeared, the solution acted directly on the bulk plate, causing significant dissolution and volume decrease. These results are also confirmed by the SEM observations.

Table 4.

Surface Area Values for Samples Polished in Solution S3

Polishing time	3 Min	4 Min	5 Min	6 Min	7 Min	8 Min	9 Min
	Surface area [mm2]
As-made	597	577	575	587	588	563	574
Polished	574	566	581	598	622	392	416

It is important to note that samples polished in the range of 1–6 min in solution S3 were additionally polished to achieve a total polishing time of 7 min. We refer to this sequential operation as repolishing. As previously noted, this two-part procedure was done to check whether the polishing procedure can be stopped and restarted with the same final result. The repolishing was done to achieve a total polishing time of 7 min because 7 min of polishing resulted, cumulatively, in the optimal loose powder removal and best fixation plate accuracy when compared with the CAD model. With solution S3, it was observed that after 7 min, the accuracy of volume, compared with the CAD model, was ±10% (Table 3). Thus, our study confirmed that using this protocol, the surface of our 3D printed NiTi skeletal fixation plates can be safely and usefully made free of unmelted powder particles (Fig. 5c). Solution S3 polishing times longer than 7-min induced pitting formation and intense dissolution of the material (Fig. 5d).

Conclusions

Chemical polishing in the HF/HNO₃ solution was found to be an effective method to remove unmelted particles from the outer and inner surfaces of skeletal fixation plates (3D) printed from NiTi powder through SLM PBF. The proper ratio of hydrofluoric acid, nitric acid, and water components of the polishing solution, as well as time of polishing, was determined to obtain high-quality surfaces that are free of loosely sintered powder particles. Since the degree of chemical polishing is determined by the time of polishing and type of solution, reducing the amount of HF, responsible for dissolution, and increasing the amount of HNO₃, responsible for passivation, were found to be the best methods to optimize the polishing process kinetics for these fixation plates. The volume and weight of the samples gradually decreased as polishing time increased and the intensity of this phenomenon depended on the solution used. We found that solution S3 was the most appropriate for removing loosely sintered powder particles from the surfaces of these 3D printed NiTi skeletal fixation plates. The weight and volume decrease during the first 6 min of polishing with solution S3 was not significant and increased rapidly after this time. Likewise, SEM observations showed that the best surface quality or smoothness of surfaces was obtained at 7 min of polishing in solution S3. Exceeding this time caused critical features to dissolve. This fact was confirmed by both SEM and μCT analysis. The reason for this may be that at the initial stage of polishing, the connections of the unmelted powder particles to the surface of the skeletal fixation plate dissolve, and after the removal of a significant portion of these powder particles, the surface of the skeletal fixation plate itself is more exposed and undergoes more significant dissolution. In addition, it was observed that additional periods of polishing (i.e., repolishing fixation plates that have previously been polished) for a shorter time can be effectively used. Other studies are being conducted to confirm that finite element method modeled mechanical properties of the polished parts are accurate and that after polishing, potentially dangerous (cytotoxic) Ni ion release has been minimized.

Author Disclosure Statement

D.D. and M.E. are coauthors on patent applications that are licensed to RegenFix, LLC, by the University of Toledo and The Ohio State University. Both D.D. and M.E. have founder's equity in RegenFix, LLC.

Funding Information

This work was supported by a TVSF grant, #TECG20170372, from the Third Frontier (State of Ohio); RegenFix, LLC (Toledo, OH); the Department of Plastic and Reconstructive Surgery at The Ohio State University; and, a grant from the Warsaw University of Technology Faculty of Materials Science and Engineering.