1 Introduction

Additive Manufacturing (AM) is a type of production to ease complex-shaped pieces production and makes fast prototype production, especially in space, aviation, and biomedical (Ref 1, 2). Ti-6Al-4V alloy is one of the common materials used in these aforementioned industries. Moreover, it is a well-acknowledged fact that this material is difficult-to-machine material. Considering all these points, fabricating components made of Ti-6Al-4V alloy utilizing AM method is a reasonable and thus preferred approach. On the other hand, components manufactured through the AM process, including laser powder bed fusion, do not meet the surface requirements of the aforementioned industries. For such reason, post-processes operations such as turning (Ref 3), milling (Ref 4), tribofinishing (Ref 5) (drag finishing, vibratory surface finish etc.), chemical treatment (Ref 6) and blasting (Ref 7) are widely used to enhance surface aspects of AM components

As a mass finishing process, drag finishing operations have been used in industry and academia for a long time for various purposes such as enhancing edge radius of cutting tools (Ref 8, 9), deburring operations of conventionally machined metal parts (Ref 10), surface polishing for conventionally fabricated parts, etc. (Ref 11). Few researchers also utilized it as post-processing operations to enhance the surface quality of additively manufactured stainless steel (Ref 12), maraging steel (Ref 13) and Inconel 718 (Ref 14). These studies reported that the drag finishing process improves surface aspects and topography of additively manufactured metal parts. These works also underlined that drag finishing process is much more efficient to improve surface properties than other tested post-processing operations including vibratory surface treatment. However, all these studies kept abrasive media constant and did not report the role of abrasive media on surface and subsurface aspects of additively manufactured parts. Abrasive media plays a major role in the effectiveness of the drag finishing process, as noted in previous studies that utilized the drag finishing process to improve surface aspects of conventionally manufactured metal components. Moreover, to the authors’ best knowledge, no study investigates the influence of drag finishing on surface aspects of additively fabricated Ti-6Al-4V alloy.

Considering these points, the gap is evident. Furthermore, surface aspects are not limited to roughness and topography, and researchers also put efforts to modify and enhance the wear resistance of additively fabricated Ti-6Al-4V alloy (Ref 15) considering biomedical applications of this material (Ref 16). As drag finish is mass finishing operation and post-processing, its effect on wear resistance should also be investigated and reported. This present work aims to fill these gaps by presenting systematic experimental results focusing on the role of drag finish process parameters and various abrasive media on surface properties of Ti-6Al-4V alloy produced by laser powder bed fusion additive manufacturing. Measured outputs are surface roughness, topography, microhardness and resulting wear resistance.

2 Materials and Methods

The specimens were composed of Ti-6Al-4V alloy powder and produced by Laser Powder Bed Fusion using EOS M 290 under Ar atmosphere. Specimens with equilateral triangle geometry with edge length 6 mm and height of 40 mm are fabricated. The particle size of powder used in this experiment was found to be in the range of 14-45 µm in diameter. Table 1 shows laser powder bed fusion (LPBF) process parameters for fabricating specimens. These laser parameters are recommended by EOS and commonly used for Ti-6Al-4V.

Table 1 LPBF process parameters for fabricating Ti-6Al-4Vspecimens
Full size table

The drag finish post-processing technique is used to enhance the surface characteristics of LPBF processed specimens. Three different processing times (duration), two different rotational speeds and two different abrasive media were used for the DF process, as shown in Table 2. These drag finish process parameters and abrasive media are recommended by surface finishing machine manufacturer company considering Ti-6Al-4V work material and its properties. Future-Tech FM310, Keyence Digital Optical Microscope and Philips Xl 30 Sfeg were used to measure microhardness, surface roughness, and wear marks, respectively. The microhardness of each sample was determined by an average of 4 measurements with a test load of 50 g using the Future-Tech FM310e model instrument with 15 s dwelling time.

Table 2 DF process variables used in this study
Full size table

The sliding wear tests were conducted using a reciprocating tribometer per the ASTM G133 standard. The tests were carried out for 40 min under a contact load of 10 N and a stroke length of 6 mm at room temperature in an air atmosphere without any cooling or lubricant. An Al2O3 abrasive ball was used for the reciprocating wear test. Since the hardness of the ball (1310 ± 10 HV) is much higher than that of the specimens, no wear of the carbide ball was observed during the test, and therefore it did not affect the wear behavior of the material.

3 Results and Discussion

3.1 Surface Roughness

The surface features of the components subjected to dynamic loads or contact forces are important, especially for the aviation, space, and biomedical industry qualifications (Ref 14). Figure 1 shows the surface topographies of as-built specimens for both building and scanning directions. Partially melted powders attached to the surface layer of as-built specimens leads to undesired surface topographies on building direction. This is a commonly observed surface of specimens fabricated by the LPBF process. Much consistent surface is obtained on scanning direction that is also well acknowledged.

Fig. 1
figure 1

Topography images of the building direction (a) and scanning direction (b) surfaces of Ti-6Al-4V samples

Full size image

Figure 2 shows a 3-D examination of the Ti-6Al-4Vspecimen. As can be seen from the figure, the surface roughness value varies in a certain region. The average surface roughness value in building direction is 5 ± 0.1 μm. The overlapping layers and the dust adhering to the surface create a stair-step effect. Therefore, irregular and distorted profiles were formed (Ref 17).

Fig. 2
figure 2

Roughness measurement of as-built Ti-6Al-4V

Full size image

The porous structure on the sample surface negatively affects fatigue under dynamic loads (Ref 18). The DF process decreases surface roughness depending on the operational parameters’ adjustments.

In addition to experimentally measured surface roughness, theoretical calculation of surface roughness is also performed and presented in Fig. 3 by using the following equations proposed by Hashimoto (Ref 21).

$$R\left( t \right) = \left( {{\text{Ir}} - {\text{Dr}}} \right)e^{{ - \frac{t}{T}}} + {\text{Dr}}$$
(1)
Fig. 3
figure 3

(a) Surface roughness values after DF in different conditions, (b) Surface profile lines of samples surface treated using 60 rotational speed, (c) Surface profile lines of samples surface treated using 120 rotational speeds

Full size image

The time constant T can be measured by the process time t when the surface roughness becomes the value R(T) given by Eq 2.

$$R\left( T \right) = \left( {{\text{Ir}} - {\text{Dr}}} \right)e^{ - 1} + {\text{Dr}}$$
(2)

The roughness of finished components can be predicted by using Eq 1 (time constant—T, roughness limitation (Dr, Ra (µm)), the best roughness value obtained during the processing period. Initial roughness (Ir, Ra (µm)), initial roughness value measured from the as-built surface.

To estimate the surface roughness in this model, the system parameters such as the time constant T, the roughness limitation Dr can be measured from the experimental results shown in Fig. 3. System parameters measurements are presented in Table 3.

Table 3 Constant values to be used in the surface roughness estimation model
Full size table

Considering experimental data shown in Fig. 3, it is obvious that post-processing time has notable effect on controlling surface roughness of additively fabricated components. Increasing post-processing time results in reduced surface roughness and thus improve surface quality for all tested abrasive media. The best surface quality was obtained after Al2O3 + plastic media/120 rpm/60 min DF treatment. This value is 0.30 µm and is approximately 94.1% less than as-built parts. With the increase of the operation time and fast rotational speed, large-sized plastic medias tend to decrease the surface roughness of the workpieces. The size of the abrasive media (plastic or Al2O3) and the shape of the abrasive media (conical, powder or nutshell) in the DF effect the impact energy on the sample and the higher energy decreasing the surface roughness (Ref 11, 19).

Different media used in abrasive environments and rotational speed also affect the surface roughness (Fig. 3a). With the increase of the cycle (Al2O3 media), the surface roughness value decreases at a rate of approximately 63.7%. In the 60 rpm cycle, Al2O3 + plastic media decreased the roughness value 74.5% compared to the Al2O3 media. However, at 120 rpm this value has appeared as 86%.

In Fig. 3(b) and (c), the surface profile lines are compared using the Al2O3 and Al2O3 + Plastic medias with different wear durations and cycles. With the cycle and process duration increase, the surface surges decreased and tended to be a straight line.

Calculated surface roughness values show reasonable agreement with the experimentally measured roughness values. The difference is varied between 4 and 24% that is found acceptable. But it should be noted that calculated surface roughness shows remarkable variation considering the first 15 min that is because of the poor surface quality of as-built samples resulting from partially melted powders stick to surface (Ref 20). However, as duration of post-processing time is increased to 30 and 60 min, the calculation offers much better agreement with experimentally measured data. For instance, considering 30 min post-processing duration, the difference between experimental data and calculated data is approximately 5%.

Surface images after the DF process of samples are shown in Fig. 4. Outer layer of samples include partially melted powders and cavities due to incomplete fusion (Ref 22). As can be seen, the DF parameters and abrasive media have had a serious impact on the surface quality. It has been observed that the use of plastic media has a positive effect on surface topography. While Al2O3 removes the surface layer effectively (Ref 5), large-sized plastic abrasive helps finishing the surface topography make the surface much smooth. This eventually helps to enhance surface quality by reducing surface roughness of samples. However, it is apparent that 15 min processing time by either using Al2O3 alone or Al2O3 and plastic media is not capable of removing the outer layer of samples. For this reason, surface topography does not seem to be perfect as yet. Increasing processing time substantially helps to improve the topography of samples as shown in Fig. 4. But at this point, the role of plastic media is very critical. Utilizing plastic media along with Al2O3 at 120 rpm offers the best surface quality among all tested conditions. In this test, outer surface layer of samples is completely removed by Al2O3 as 60 min processing time is given and high rpm is implemented that generated larger kinetic energy (Ref 5, 19). Plastic media polished the generated new surface. But the remaining issue is scratches on the surface. Although it is not tested in this study, further processing by using only plastic media should be effective to remove these scratches from the surface.

Fig. 4
figure 4

Optical microscope images (2D) of DF processed workpieces

Full size image

3-D surface images of samples treated at 120 rpm and after 60 min operation time are shown in Fig. 5. Another interesting result is the scanning direction of the surfaces are not enhanced enough. The reason is dipping the workpieces to the DF receptacle vertically. This situation is about the effects of the abrasive medias on the side surfaces effectively, but it cannot give the intended friction results in the bottom surface.

Fig. 5
figure 5

Surface images formed after the DF process were applied for 60 min at 120 rpm

Full size image

3.2 Microhardness

The surface and subsurface microhardness of as-built and treated samples at 60 and 120 rpm rotational speed are shown in Fig. 6. The hardness of as-built samples is approximately 370 HV. Results indicates that post-processing operations depending on processing time, rotational speed and media influences the surface and subsurface hardness of samples. While at lower rotational speed, process effected layer is approximately 50 µm beneath from the surface, it is approximately 100 µm when 120 rpm rotational speed is used. This is notable finding. Increase in percentage of hardness with rotational speed also support this finding. For instance, maximum increase in hardness with using 60 rpm rotational speed is approximately 2.9%, it is almost 6% when 120 rpm rotational speed is used.

Fig. 6
figure 6

Microhardness values; (a) at 60 rpm rotational speed, (b) at 120 rpm rotational speed

Full size image

In addition to the rotational speed, abrasive media and processing time are also playing role on forming surface and subsurface hardness. It is an obvious that increasing processing time leads to generating harder surface and subsurface. For instance, maximum hardness measured at surface layer is approximately 378 HV when sample is treated Al2O3 + plastic with 15 min processing time, it is 392 HV when processing time increased to 60 min as shown in Fig. 6(b). Considering the role of abrasive on hardness variation, adding plastic media has some limited effect on increasing hardness. For instance, maximum hardness measured at surface layer is approximately 392 HV when sample is treated using Al2O3 + plastic with 60 min processing time, it is 383 HV when it is treated using only Al2O3 abrasive as shown in Fig. 6(b). Adding plastic media to abrasive polished the surface and likely increased residual stresses occurring on the surface (Ref 23). The same has been observed with the increase in the duration of the process (Ref 24).

As it is known, increase in hardness is the results of the formation of refined grains, and dislocation density increased due to plastic deformation (Ref 13). Our results show that major changes in hardness takes place at the surface region or just beneath the surface with longer processing time. Within such a long processing time and high rotational speed, much higher plastic deformation occurs and hence hardness becomes much higher than as-built sample.

4 Wear Behavior

4.1 Friction Coefficient

The wear process occurs in four zones, as described in Fig. 7. The first contact between the specimen and the abrasive ball is the running-in period (Ref 24), which occurs before wear reaches a stable state. This process is usually accompanied by a change in the coefficient of friction and wear rate (Ref 25). Since there is point contact between the specimen and the Al2O3 abrasive ball, high pressure occurs and eventually causes rapid wear. It comes into direct contact with the oxide layer, and a rapid increase in the friction coefficient occurs (I. 0-20 s). It is thought that there is a decrease in the friction coefficient, and the oxide layer between the abrasive ball and the specimen prevents this reduction (II. between 20 and 60 s). As the contact area becomes larger, new contact surfaces are formed between the abraded material and the abrasive material during the wear process. Due to repetitive stress and load cycles, almost all the oxide layer is disintegrated and therefore, there is a significant increase in the friction coefficient. As a result of the disappearance of almost all the oxide layer, the Al2O3 ball and the sample are in full contact, and the increase in the friction coefficient continues (III. 60-150 s). Due to the surface contact area increase, the pressure between the sample and the Al2O3 wear ball decreases with time, and the friction coefficient becomes stable after about 100-150 s till the analysis end (Region IV). This stable character is also clearly shown in the wear test graphs of the study (Fig. 8).

Fig. 7
figure 7

Display of wear stages on the as-built workpiece

Full size image
Fig. 8
figure 8

The COF graph obtained because of the wear process (a) 60 rpm rotational speed, (b) 120 rpm rotational speed

Full size image

The friction coefficient value of Ti-6Al-4V specimens obtained after the wear operation showed that it increased with the first contact at the start of the wear test and took a more stable form in the following test moments. The oscillations in the wear coefficient graphics can be explained as the activation of the wear mechanism of the Three-Body Effect caused by the oxide layer that occurred on the surface in the first place (Ref 26).

All the Coefficient of Friction (COF) graphs obtained are similar and can be seen in Fig. 8(a)-(b). The as-built specimens' COF graph showed more oscillations due to the surface quality that directly affects the contact between the wear ball and the worn (workpiece). If the contact established during the wear process is compatible, the flow of the formed wear residues on the worn surface will also be difficult (Ref 27). Due to the difficulty of flow on the contact surface during wear, abrasion residues sticking to the surface occur. These wear residues adhere to the wear surface and form a protective layer. This protective layer creates a harmonious contact and abrasion resistance between the abrasive and the wearer (Ref 28).

As a result, the strong effect of the contacts established between the geometries in contact on both the wear mechanism and the friction coefficient can be seen in Figure 8. The average friction coefficient values shown in Fig. 8 support this situation. The lowest COF value belongs to the Al2O3 + P medias/120 rpm/60 min DF process. Compared to the as-built specimen, the COF value decreased by 44.18%. The highest COF value belongs to the DF process that Al2O3 media/60 rpm/15 min parameters. Compared to the as-built specimen, the COF value decreased by 13.95% and at the same time, there was a 35.13% change between the conditions of Al2O3 + P medias/120 rpm/60 min. The irregular oscillations in the graphs can be explained in terms of third-body particles trapped between the abraded and abrasive sample during the wearing process (Ref 29). The friction coefficient value usually depends on the oxide layer on the surface, microstructure of the materials, microhardness and surface roughness, and it can develop depending on wear performance (Ref 30).

4.2 Wear Rate

Wear rates are calculated using Eq 3 and 4. The wear rates obtained from the calculations are shown in Fig. 9. Wear volume is calculated using Eq 3 as defined below (Ref 26).

$$V = L\left[ {r^{2} {\text{sin}}^{ - 1} \left( \frac{w}{2r} \right) - \frac{w}{2}\left( {r^{2} - \frac{{w^{2} }}{4}} \right)^{\frac{1}{2}} } \right] + \frac{\pi }{3}\left[ {2r^{3} - 2r^{2} \left( {r^{2} - \frac{{w^{2} }}{4}} \right)^{\frac{1}{2}} - \frac{{w^{2} }}{4}\left( {r^{2} - \frac{{w^{2} }}{4}} \right)^{1/2} } \right]$$
(3)

where V is the wear volume (mm3), r is the wear ball diameter, L is the stroke length, w is wear track width. The wear rate, k (mm3 /Nm), is given by:

$$k = \frac{V}{Nxl}$$
(4)
Fig. 9
figure 9

The wear rates of the as-built and post-treated specimens

Full size image

N indicates the applied force in Newton, l is the total distance traveled during wear as meter. The demonstration of the measurement of the wear width after the wear tests is presented in Fig. 10.

Fig. 10
figure 10

Wear profile chart of workpieces subjected to DF process (a) 60 rpm rotational speed, (b) 120 rpm rotational speed

Full size image

The wear rate calculation results are shown in Fig. 9. The wear rate obtained from the abrasion test performed on the as-built sample was calculated as 3.3 × 10−2 mm3/Nm. Three parameters increase the wear resistance of the surface in the DF process; use of Al2O3 + Plastic media, increase in process time, increase in rpm value. After the long duration of operation and resultant with combined surface enhancement stones, the thin-pieced structure occurred on the surface, and the subsurface caused the material to harden (Fig. 6b). The higher the hardness value, the shallower the depth of the material crushed in the friction and wear process and the less the wear volume (Ref 31). Hardness measures resistance to plastic deformation without fracture under a hard indenter (Ref 32). The high microhardness (Al2O3 + Plastic media/120 rpm/60 min process) value resulted in the alloy’s resistance to wear (Ref 30). However, more plastic deformation occurred because of the abrasion test performed on the as-built sample and the sample treated with Al2O3 media/60 rpm/15 min process. According to Archard's linear (Ref 33) law, the wear volume loss is inversely proportional to the hardness values of the worn material. The microhardness values obtained because of the secondary processing processes applied in this study (Fig. 6) are fully compatible with the wear rates obtained (Fig. 9). As a result, using combined surface enhancement stones and a long duration DF process gave fewer wear rates.

4.3 Wear Profile

The wear trace profiles are presented in Fig. 10. Wear trace & deepness showed differences with the DF process parameters. The as-built sample has the widest (0.967 mm) and deepest (35.13 μm) wear profile. The closest wear scar depth (20.32 μm) and width (886.11 μm) to the as-built sample belong to the sample applied Al2O3 media/60 rpm/15 min process. Depth of wear decreased by 42.15%, and wear width decreased by 8.3%. The lowest width (669.45 µm) and deepest (7.29 µm) belong to the sample applied Al2O3 + plastic media/120 rpm/60 min DF operation. Depth of wear decreased by 79.23% and wear width by 30.74%.

The increase in microhardness with the effect of post-processing plays a serious role in the wear rate case. Rough surfaces cause high friction force, and mechanical interaction increases with an increasing friction force. As a result, plastic deformation increases (Ref 34). As a result of the surface improvement processes, the different media used improved the surface quality depending on the rotation speed and processing time of the DF device. It also increased the microhardness value and wear resistance of surfaces.

SEM images of some wear marks are shown in Fig. 11(a)-(c). SEM images show differences in the tribological properties of the samples. Cracks and cavities cause delamination under and above the surface. The delamination wear mechanism occurs due to the shear deformation on the surface and the combination of cracks (Ref 35). Also, chip removal traces caused by wear are shown in Fig. 11(a). Crack propagation was observed in the transfer layer with the formation of delamination after wear. The delamination wear, which occurs because of the wear process applied to the as-built sample (Fig. 11a), is the porous structure on the surface and under the surface (Ref 36). In this case, it accelerates the formation of delamination, which causes a decrease in wear resistance.

Fig. 11
figure 11

Wear marks in SEM images (a) as-built (b) Al2O3 Media/60 rpm /15 min, (c) Al2O3 + P Media/120 rpm/60 min

Full size image

Numerous fine scratches were observed on the inner surfaces of the groove formed in the wear track. The fact that the scratches are aligned along the sliding direction indicates abrasive wear (Ref 37). The relative extent of the grooves observed in the wear paths depends on the hardness of the abrasive body (Ref 26). When the SEM image of the sample, which was treated with Al2O3 + plastic medium/120 rpm/60 min, is examined, the wear groove is seen in Fig. 11(c). Due to these fine scratches aligned along the said slide direction, Abrasive wear is seen in Fig. 11(c). After the Al2O3 + plastic media/120 rpm/60 min DF process, no significant delamination was observed in the abrasion treated sample. However, in Al2O3/60 rpm/15 min (the process with the worst surface), delamination occurred as shown in Fig. 11(b).

The delamination of the as-built sample obtained due to the wear test is quite high compared to other results. Vertical cracks were formed in the transfer layer because of the wear load. Random clumps of material debris have formed on weathered surfaces because of the severe adhesive and abrasive wear (Ref 38). Concurrently during sliding, metal debris is crushed, mixed, and compacted into the worn surface, forming a tribolayer (Ref 39). A layer of tribo oxide is often found on the corroded surface of Ti-6Al-4V. During the wear process, metal residues combine with oxygen, and the oxygen permeates the tribolayer. At the same time, delamination takes place resulting from the initiation and propagation of micro-scratches. In the wear process, micro-scratches are critically propagated, hence shear delamination occurs parallel with the worn surface and material debris from the worn surface (Ref 39).

5 Conclusions

In this study, the effect of drag finish post-processing parameters and abrasives media on the surface properties and resulting wear behavior of Ti-6Al-4V specimens fabricated by laser powder bed fusion was presented. This work demonstrates that the processing time and the abrasive media have substantial influence on measured outputs. Using Al2O3 + plastic media along with 120 rpm during 60 min processing time contributed to surface quality by reducing arithmetical average surface roughness from 5.1 to 0.30 µm. As a result of the DF operation performed under the same condition, the microhardness value increased by 6% (370 to 396 HV).

The calculated surface roughness by using the model borrowed from the literature for various abrasive media, process duration and rotational speed showed acceptable agreement with the experimentally measured average surface roughness value. It should be noted that while the model shows much better agreement with experiment when Al2O3 media is used, the difference between experimental result and calculated results becomes larger when Al2O3 + plastic media is used.

Wear resistance of specimens subjected to drag finish post-processing showed notable increase as compared to the wear resistance of as-built sample. Moreover, rotational speed and abrasive media play key role on further enhancing wear resistance of Ti-6Al-4V alloy fabricated by laser powder bed fusion. It should be also noted that post-processing operations along with process parameters and abrasive media substantially later wear mechanism of this Ti-6Al-4V alloy. For instance, while larger wear trace and formation of delamination observed on as-built specimen, abrasive and adhesive wears are observed on post-treated specimens. This work demonstrated that drag finish post-processing play remarkable role on enhancing wear resistance and controlling wear rate of additively manufactured Ti-6Al-4V alloy.