Effect of Different Sandblasting Parameters on the Properties of Additively Manufactured and Machined Titanium Surfaces

Abstract

Background/Aim

In dentistry, the surfaces of titanium implants are often sandblasted and acid-etched in order to support successful osseointegration. The aim of this study was to investigate the impact of various sandblasting parameters on the surface roughness, contact angle and surface energy of additively manufactured (TiAl6V4) and machined commercially pure titanium (cpTi) surfaces.

Materials and Methods

A total of 56 disc-shaped samples were produced using either laser powder bed fusion (TiAl6V4) or using precision cutting (cpTi). The samples were then sandblasted with different angles, distances, and pressures using an automated sandblasting machine. Afterwards, surface roughness and contact angle for water and diiodomethane were measured, and scanning electron microscopy images were taken.

Results

The results showed that the initially rough TiAl6V4 samples became smoother after sandblasting, while the smooth cpTi surfaces became rougher. Sandblasting pressure had the most significant influence on surface roughness. The surface energy of sandblasted TiAl6V4 samples showed no significant change compared to the as-built state (26.6±1.3 to 26.3±1.8 mJ/m²). In contrast, cpTi samples showed a reduction in surface energy after sandblasting (32.3±1.6 to 26.8±1.2 mJ/m²). Scanning electron microscopy revealed irregular surfaces with grooves and ridges for both types of samples. The roughness of TiAl6V4 decreased at higher sandblasting pressures, whereas cpTi surfaces became rougher.

Conclusion

Surface roughness after sandblasting is strongly influenced by the initial surface, which differs in additively manufactured TiAl6V4 samples compared to machined cpTi surfaces.

Introduction

Additive manufacturing (AM) has seen notable technological progress in recent years. Unlike conventional machining processes, AM enables the layer-by-layer production of components directly from digital models (1,2). This technology opens up new possibilities in designing and producing parts with complex geometries and individualised designs that are challenging to achieve with conventional manufacturing processes (3). Medical technology benefits from AM, as it enables the production of customised implants that can be optimally adapted to the anatomical conditions of the patient (4-6).

For medical applications, titanium and titanium alloys play an important role and are widely accepted (7). Titanium alloys, particularly TiAl6V4, are characterised by high strength, superior corrosion resistance and biocompatibility, making them especially suitable for medical implants (8). Laser powder bed fusion is a common method for processing such alloys and facilitates the production of patient-specific implants (9,10). Despite advanced manufacturing techniques, additive manufactured implants often require subsequent surface treatment to achieve optimal surface properties. For machined surfaces, sandblasting is a common process used for such surface treatments (11-13).

In dentistry, the process of sandblasting is commonly employed to roughen smooth surfaces. For example, the abutment contact area is roughened in some cases using this method to enhance the surface for cementing the abutment to a crown (14,15). A recent review by Zimmer et al. found that sandblasting improves bond strength in resin composite repair (16). For dental titanium-based implants, a combination of sandblasting and acid etching is frequently employed to create sandblasted and acid-etched surfaces, which enhances the healing process (known as osseointegration) of the implant with the surrounding biological environment (17-20). However, the exact parameters of the sandblasting process, such as the distance used, are often insufficiently detailed in literature (21), as sandblasting is usually carried out manually in scientific studies, which makes it difficult to determine the precise sandblasting parameters.

The aim of the current study was to investigate the influence of sandblasting with different parameters on the surface properties of conventionally machined titanium (cpTi) and additively manufactured TiAl6V4 samples.

Materials and Methods

In this study, an automated sandblasting machine was used to treat conventionally machined titanium (cpTi, Grade 4) and additively manufactured TiAl6V4 (Grade 5 Eli) samples varying blasting pressure, blasting angle and blasting distance. After blasting, the contact angles for water and diiodomethane were measured following EN ISO 19403-2 (22). Furthermore, the surface roughness was analysed using optical profilometry and scanning electron microscopy (SEM) images were taken using a secondary electron detector. The contact angle results and surface roughness were statistically analysed. Details are given in the following sections.

Sample preparation. Disc-shaped samples with a diameter of 12 mm and a thickness of 2 mm were manufactured from cpTi (Grade 4) and TiAl6V4 (Grade 5 Eli, which is also called Grade 23). The cpTi samples (Figure 1A) were cut with a linear precision saw (Brillant 220; ATM Technologies GmbH, Nienhagen, Germany) from 12 mm titanium rods (L. Klein SA, Biel, Switzerland) at 5,000 rpm and a cutting speed of 1.5 mm/min. For cutting, a diamond cutting disc (Buehler GmbH, Braunschweig, Germany) with a diameter of 203 mm and a thickness of 0.9 mm was used.

Figure 1.

(A) Machined titanium rod (cpTi) specimens, stored in well plates. (B) Built plate with the additively manufactured (TiAl6V4) specimens.

The TiAl6V4 samples were produced using a Lasertec 12 SLM (DMG Mori AG, Bielefeld, Germany) with a 400 W fiber laser (1,070 nm, continuous wave, minimum spot diameter of 35 μm), oriented as shown in Figure 1. The process parameters included a laser power of 120 W and a laser speed of 860 mm/s for the contour, while a laser power of 175 W and a laser speed of 1,050 mm/s were applied for the hatch.

The powder utilized was Ti-6Al-4V Grade 5 Eli (ECKART GmbH, Hartenstein, Germany) characterized by its predominantly spherical morphology and particle sizes between 20.0 μm and 53.0 μm. According to ISO 20160 (23), medical implants must undergo heat treatment to reduce internal stresses and refine the microstructure. The samples were heated to 1,050˚C for 4 h, followed by cooling in an oven.

Sandblasting machine. The automated sandblasting machine used in this study was based on previous work (24) and was customised for this study. The machine was based on a on a 3D-printer construction kit (Reptile, Locxess, Germany), with three axes for controlled translational movements (Figure 2). A sandblasting unit with pressure regulator (IP Mikro-Sandy; IP Division Technische Produkte GmbH, Haimhausen, Germany) was mounted instead of the printing head. The sandblasting nozzle (IP Division Technische Produkte GmbH) had a diameter of 1.3 mm. The axes were controlled via G-Codes, allowing parameter-controlled movement of the nozzle. Additionally, the sandblasting angle was set between 45˚ and 90˚ in 5˚ increments using a locking plate (Figure 2B). To enhance its containment of sandblasting particles and prevent their uncontrolled spread, a blasting chamber was employed. The chamber consisted of a perforated plate that formed the base and a wall around the base on two inverted U-profiles. The U-profiles had openings and were placed orthogonally on top of each other as shown in Figure 2B. Both U-profiles sat loosely on top of each other, allowing them to move freely in the horizontal X and Y directions. An extraction system was installed beneath the perforated plate generating a vacuum in the sandblasting chamber, which helped to reduce the spread of particles. A sample holder was attached to the perforated plate using a plug-in system. The samples were placed at predefined positions on the sample holder and secured with double-sided adhesive tape, ensuring a defined position and fixation of the samples.

Figure 2.

(A) Sandblasting machine used in the study. (B) Schematic cross-sectional view of the sandblasting chamber.

Sandblasting parameters and post-processing. The samples were sandblasted using Corundum particles (Al₂O₃) (SHERA Werkstoff-Technologie GmbH, Lemförde, Germany) with a nominal grain size of 110 μm. A full factorial experimental design was applied, varying pressure, angle and pressure and material as follows:

• three different sandblasting distances: 5 mm, 10 mm and 15 mm; • three different sandblasting angles: 45˚, 60˚ and 90˚; • three different sandblasting pressures: 2 bar, 4 bar and 6 bar; • two different materials: cpTi and TiAl6V4.

This resulted in 54 sandblasted groups in total. Additionally, two control groups were established, one for TiAl6V4 and one for cpTi, which remained untreated. For each of the resulting 56 groups, two samples were prepared, resulting in a total of 112 samples. The allocation of the samples is shown in Table I.

Table I. Overview of the parameter group numbers (PGN) and the corresponding sandblasting parameters.

To ensure precise sandblasting distances, a spacer was made for each possible combination of sandblasting distance and angle, resulting in nine spacers in total (see Figure 3A). The sandblasting path was meander-shaped (see Figure 3B) with a path offset of 0.3 mm and a forward speed of 1 mm/s, resulting in an effective sandblasting time of ~8 min for each sample. These values were determined through preliminary tests. Path offsets between 0.15 mm and 0.50 mm and forward speeds of 0.5 mm/s to 1.5 mm/s were tested (data not shown here) to achieve a uniformly sandblasted surface within a reasonable sandblasting time.

Figure 3.

Schematic representation of (A) a spacer with a blasting distance of 15 mm at 45° (trajectory indicated with a green line), (B) sandblasting path (indicated with red lines) and profilometry path (indicated with blue lines), and (C) positions of the drops for the contact angle measurement.

After sandblasting, the double-sided adhesive tape was removed. To ensure the removal of adhesive residues and loose sandblasting particles, the surface was cleaned in an ultrasonic bath (Emmi-H30; emmi Emag AG, Mörfelden-Walldorf, Germany) with acetone (J. T. Baker, Phillipsburg, NJ, USA) for 15 min. Three further cleaning stages followed in an ultrasonic bath with distilled water for 10 min each.

Surface roughness characterisation. The arithmetic mean value of the roughness (Sa) was determined using an optical profilometer (MicroProf 100; FRT GmbH, Bergisch-Gladbach, Germany). The analysis included a measurement field of 3.2 mm×3.2 mm (10.24 mm²) at the centre of the sandblasted surface, including 800×800 (640,000) data points within this area. The software Gwyddion (version 2.63; Petr Klapetek, Brno, Czech Republic) as used to determine the mean roughness. In a first step, a plane level correction was applied to the raw data. In order to reduce the distortion of the measurement results due to measurement artifacts, the three standard deviations (3σ) rule was applied (25). The 3σ rule is a statistical method for identifying outliers in a data set. It is based on the assumption that the data follows a normal distribution. The principle is that approximately 99.7% of the data lie within 3σ of the mean. Values outside the 3σ range around the mean were considered outliers and were excluded.

SEM analysis. The samples were imaged using an SEM (EVO MA10; Carl Zeiss AG, Oberkochen, Germany). A 500× magnification over an area of 450×600 μm and a 1000× magnification for a 225×300 μm area was used for the imaging, allowing a detailed inspection of the surface topography and checking for microstructural changes. This analysis was performed at a working distance of 10.5 mm, utilizing a spot size of 400 nm and an acceleration voltage of 15 kV.

Contact angle measurement. Contact angle measurements were used to determine the surface energy according to EN ISO 19403-2. For this purpose, an OCA 40 contact angle measuring device (DataPhysics Instruments GmbH, Filderstadt, Germany) equipped with SCA20 software was employed. Prior to measurement, samples underwent a two-stage cleaning process in an ultrasonic bath, initially with 2-propanol (J. T. Baker), and subsequently with ultrapure water type 1 (Milli-Q Advantage A10; Merck Chemicals GmbH, Darmstadt, Germany) for 15 min each. Following this, the samples were left to dry and acclimate in the ambient air of the measuring room for at least 16 h. The temperature during the measurements was 24˚C with a relative humidity of 57%, which is slightly outside of the range of ISO 19403-2 (50±5%). Contact angles were measured at three points per sample, as depicted schematically in Figure 3C. The static method (26) was applied to these measurements. Two liquids were applied, ultrapure water type 1 with a drop volume of 2 μl and diiodomethane (Thermo Fisher Scientific, Waltham, MA, USA) with a drop volume of 1 μl. The dosing rate was 0.5 μl/s. After depositing the drop, an image was captured for measurement after 10 s. The camera for measuring the contact angle was tilted by 2˚ and a twofold magnification was used. Hot-pixel correction served as the method for image correction. The manual polynomial method was used to define the outline of the drop to determine the contact angle. The surface energy was determined based on the linear equation (Eq. 1) developed by Owens, Wendt, Rabel and Kaelble (OWRK) (27,28). In this equation, the slope and the intercept were obtained from the square root of the polar as well as the disperse components of the surface energy.

$$\frac{{\sigma }_1·(1+\mathrm{cos\; \theta })}{2·\surd {\sigma }_1^d}=\surd {\sigma }_s^d+\surd {\sigma }_s^p\surd \frac{{\sigma }_1^p}{{\sigma }_1^d}$$

(Eq. 1)

where θ: the mean value of the measured contact angles for the respective liquid;

σ₁: the surface tension of the respective liquid;

${\sigma }_1^p$: the polar component of the surface tension;

${\sigma }_1^d$: the dispersive component of the surface tension;

${\sigma }_s^p$: the polar component of the free surface energy;

${\sigma }_s^d$: the dispersive component of the free surface energy.

The values from the DIN EN ISO 19403-2 standard for water and diiodomethane were used to calculate the points for the linear regression (see Table II).

Table II. Calculated surface tension parameters for water and diiodomethane used in the linear equation for surface energy calculation.

The abscissa described the ratio of $\surd {\sigma }_1^p/{\sigma }_1^d$ (1.5295 for water, 0 for diiodomethane). The values for the ordinate were determined by multiplying the term (1+cos θ) with 7.7960 for water and 3.5637 for diiodomethane.

Statistical analysis. The influence of the sandblasting parameters on roughness, contact angle, and surface energy was analysed using a three-way analysis of variance (MATLAB function ANOVAN) in MATLAB (Version R2022a; The MathWorks, Natick, MA, USA). The results were considered statistically significant for values of p<0.05.

Additionally, MATLAB’s boxplot command was used with notches. If notches in a plot do not overlap, it can be concluded with 95% confidence that the true medians do differ, according to the MATLAB documentation. The notches are calculated as notch_hi/low=p50±1.57*(p75-p25)/√(n), where n is the number of elements and p25, p50, p75 are the percentiles [further details can be found in (29,30)]. In certain cases, the notches may extend beyond the range of the box itself, indicating a wider confidence interval. This phenomenon arises due to the calculation method and does not compromise the validity of a plot or its interpretation.

Results

TiAl6V4. In the as-built state, a mean roughness Sa of 15 μm was measured for the TiAl6V4 samples with parameter group number (PGN) 28 (see Table I). After sandblasting, a decrease in the mean roughness was observed (PGN 1 to 27). The roughness ranged between 5 μm to 11 μm, depending on the combination of sandblasting parameters used (Figure 4). The median roughness for blasting angle ranged from 7.8 μm (45˚) to 7.2 μm (90˚), the median roughness for blasting distance was between 7.0 μm (5 mm) and 7.9 μm (15 mm), and for blasting pressure the mean ranged between 9.5 μm (2 bar) and 6.1 μm (6 bar).

Figure 4.

Boxplots illustrating the main effects of sandblasting on the roughness of additively manufactured TiAl6V4 samples (PGN 1-27). (A) Results according to sandblasting angle. (B) Results according to sandblasting distance. (C) Results according to sandblasting pressure. The red line represents the median. The boxes represent the 25^th to 75^th percentile. The black whiskers extend to the minimum and maximum values.

The ANOVA (Table III) shows that the sandblasting pressure had a significant influence (F=15.3, p<0.05), whereas no significant effect was observed for the blasting angle and blasting distance.

Table III. Analysis of variance of surface roughness (Sa) of sandblasted TiAl6V4 samples.

Figure 5A shows the microstructure of the additively manufactured TiAl6V4 surface in the as-built state (PGN 28), imaged with SEM at a magnification of 500×. A large number of spherical particles of different sizes characterise the surface. These led to an irregular texture with numerous elevations. Areas with particle clusters were also visible. After sandblasting, the surface (Figure 5B, PGN 1) had a different texture compared to the as-built state. No spherical particles were visible. Instead, a rough, uneven structure with many grooves and ridges created by sandblasting dominates. However, an uneven structure with irregular areas remained apparent, indicating that the spherical particles had not been completely blasted down.

Figure 5.

Scanning electron microscopy images of the TiAl6V4 surface in (A) the as-built state (PGN 28) and (B) after sandblasting (PGN 1).

Figure 6 shows SEM images for different combinations of sandblasting parameters. In general, it can be seen that all sandblasted samples had a rough, uneven structure with many grooves and ridges. The surfaces that were sandblasted with higher pressures exhibited a more homogeneous distribution of grooves and ridges in general. In addition, these surfaces also appeared smoother, which also correlates with the roughness measurements. The surfaces with a lower sandblasting pressure showed a ‘hilly’ structure. In addition, small areas that appeared smooth as in the as-built state were visible, which indicates that these sandblasting particles had not blasted the surface completely homogeneously. This effect was slightly more pronounced with a sandblasting angle of 45˚. However, this effect was no longer apparent at higher sandblasting pressures of 4 or 6 bar.

Figure 6.

Scanning electron microscopy images of the TiAl6V4 surface after treatment with different combinations of sandblasting parameters. To reduce the number of images, combinations with a distance of 10 mm and an angle of 60° are not shown.

The mean contact angle on TiAl6V4 in the as-built state (PGN 28) was 132.0˚±0.9˚ for water, and 67.6˚±2.1˚ for diiodomethane. This resulted in a surface energy of 26.6 mJ/m² in the as-built state. For the sandblasted samples (pooling data for all blasting parameters), the mean contact angles, surface energies and ranges are given in Table IV.

Table IV. Contact angle measurements for sandblasted TiAl6V4 samples.

Figure 7 shows the main effects of the contact angle experiment in the form of box plots. Figure 7A shows the contact angle for water. With increasing sandblasting angle, an increasing contact angle is noticeable. To a lesser extent, this trend is also visible for the sandblasting distance. The opposite trend can be seen for the sandblasting pressure. Here the contact angle becomes smaller with increasing sandblasting pressure. Only the effect seen for the pressure was statistically significant (F=4.5, p<0.05, see Table IV).

Figure 7.

Boxplots illustrating the main effects on the contact angle and surface energy of additively manufactured TiAl6V4 samples (PGN 1-27). (A) Contact angles for water. (B) Contact angles for diiodomethane. (C) Surface energy. The red line represents the median. The boxes represent the 25^th to 75^th percentile. The black whiskers extend to the minimum and maximum values.

Figure 7B shows the results of the contact angle measurement with diiodomethane. For the sandblasting angle and the sandblasting distance, the observed trend is similar to the results for water. However, an upward trend in the contact angle for diiodomethane can be observed for increasing blasting pressure. However, none of these trends was statistically significant (see Table IV).

Figure 7C shows the surface energy for the main effects. A downward trend is observed for increasing blasting angle, distance and pressure. Again, only the effect seen for pressure was statistically significant (F=5.4, p<0.05, see Table IV).

Figure 8 shows the linear regression lines based on the contact angle measurements using the OWRK method. The value at the intercept with the ordinate axis for the as-built state (green line, 4.9 mN/m) lies within those of the entire sandblasted group (blue marked area, 4.5 to 5.12 mN/m; mean value=4.8 mNm). The dispersed part of the surface energy for the as-built state is slightly higher than the mean value of the sandblasted samples (PGN 1 to 27). The slope in the as-built state is negative and averages −1.5 mN/m. The sandblasted groups show a slope between −2.5 mN/m and −1.4 mN/m, whereas the mean slope is −1.8 mN/m. The polar component of the surface energy in the as-built state also lies within the group of sandblasted samples. Here, the polar component of the surface energy was lower than the mean value of the sandblasted samples (red line). Generally, the disperse component was several times higher than the polar component in all additively manufactured TiAl6V4 samples.

Figure 8.

Owens, Wendt, Rabel and Kaelble plot for the additively manufactured TiAl6V4 samples. The green line represents the regression line for the surface in the as-built state (PGN 28). The red line is the regression line for the mean value of all sandblasted samples (PGN 1-27). The blue area is where all sandblasted regression lines are located.

cpTi. In the as-cut state, a mean roughness Sa of 0.5 μm was measured for the cpTi samples (PGN 56). An increase in the mean roughness can be observed after sandblasting (PGN 29 to 55). The roughness ranged between 1.6 μm to 2.7 μm, depending on the sandblasting parameter combination. Figure 9 shows box plots for Sa roughness values for the main effects. The roughness appears to increase with increasing sandblasting angle (from 45˚ to 60˚) and then to decrease (from 60˚ to 90˚). There is no clear trend for this main effect. An upward trend can be recognised for the sandblasting distance: as the sandblasting distance increases, the roughness also increases. The same effect can also be seen in a more pronounced form for the sandblasting pressure.

Figure 9.

Boxplots illustrating the main effects on the roughness of machined titanium rod (cpTi) samples (PGN 29-55). (A) Results for varied sandblasting angle. (B) Results for varied sandblasting distance. (C) Results for varied sandblasting pressure. The red line represents the median. The boxes represent the 25^th to 75^th percentile. The black whiskers extend to the minimum and maximum values.

The ANOVA shows (Table V) that sandblasting angle, distance and pressure had a statistically significant influence on the surface roughness (p<0.05). The sandblasting angle had the weakest main effect (F=4.9), followed by the sandblasting distance (F=5.7). The sandblasting pressure had the greatest influence (F=29.8).

Table V. ANOVA of surface roughness (Sa) for machined cpTi samples.

Figure 10A shows the microstructure of the machined cpTi surface at a magnification of 500×. The surface had a uniform, linear structure characterised by parallel line patterns. These lines ran mainly horizontally across the entire surface and resulted from the cutting process. The surface appeared smooth overall, with very few irregularities and indentations visible as small notches or ridges along the lines. The line patterns indicate a controlled removal of material, which is typical of a precision cutting machine. Figure 10B shows the microstructure of a cpTi surface after sandblasting. The surface had a different texture compared to the original cut surface. A rough and irregular structure with numerous grooves and ridges created by sandblasting is visible. The even, parallel line patterns of the cut surface were no longer visible and had been replaced by a complex, jagged texture. The original smooth structure was broken up by sandblasting, which led to an increase in surface roughness, corresponding with the roughness measurements.

Figure 10.

Scanning electron microscopy images of the cpTi surface in (A) the as-cut state (PGN 56) and (B) after sandblasting (PGN 37).

Figure 11 shows SEM images for cpTi samples after different combinations of sandblasting parameters. The sandblasted surfaces appeared homogeneously rough with many small, sharp-edged, irregularly distributed depressions and elevations. In general, surfaces with a sandblasting distance of 15 mm had deeper surface craters than those with 5 mm. The same effect can be seen with higher sandblasting pressure. This result corresponds with the roughness values (Figure 9). These deeper indentations led to increased roughness.

Figure 11.

Scanning electron microscopy images of the cpTi surface with different combinations of sandblasting parameters. To reduce the number of images, combinations with a distance of 10 mm and an angle of 60° are not shown.

The mean contact angle on as-cut cpTi (PGN 56) was 91.1˚±3.0˚ for water and 57.0˚±2.3˚ for diiodomethane. This resulted in a surface energy of 32.3 mJ/m² in the as-built state. For the sandblasted samples (pooling data for all blasting parameters), the mean contact angles, surface energies and ranges are given in Table VI.

Table VI. Results of contact angle measurements for sandblasted cpTi samples.

Figure 12 shows the main effects of the contact angle for the sandblasted cpTi samples (PGN 29 to 55) in box plots. Figure 12A shows the contact angle for water. A noticeable upward trend can be seen as the sandblasting angle increases, although the variations within the sandblasting angle of 45˚ are relatively large compared to the other sandblasting angles. No clear trend is recognisable for sandblasting distance and sandblasting pressure. Furthermore, the data for a pressure of 2 bar show a notably larger variation than the data for other pressures. The ANOVA (Table VI) showed significant effects for the sandblasting angle (F=63.2, p<0.05) and sandblasting pressure (F=12.8, p<0.05).

Figure 12.

Boxplots illustrating the main effects on the contact angle and surface energy of machined titanium rod (cpTi) samples (PGN 29-55). (A) contact angles for water (B) contact angles for diiodomethane (C) Surface energy. The red line represents the median. The boxes represent the 25^th to 75^th percentile. The black whiskers extend to the minimum and maximum values.

Figure 12B shows the contact angle for diiodomethane. The contact angle increases with increasing sandblasting angle. A slightly lower contact angle is observed for increasing sandblasting distance. There is no clear trend in the sandblasting pressure. The ANOVA (Table VI) found a significant effect only for the sandblasting angle (F=4.9, p<0.05).

Figure 12C shows the surface energy. There seems to be a downward trend for high sandblasting angles and a slight upward trend for increasing sandblasting distances. No clear trend is visible for the sandblasting pressure. The ANOVA (Table VI) found no significant effects for the surface energy.

Figure 13 shows the regression lines based on the contact angle measurements using the OWRK method. It can be seen that the intercept at the ordinate before sandblasting (green line, 5.5 mN/m) lies above the sandblasted group (blue marked area, 4.8 mN/m to 5.2 mN/m, mean value=4.9 mNm, red line). The dispersive part of the surface energy was therefore higher compared to the sandblasted samples (PGN 29 to 55). The slope is positive before sandblasting and equates to 1.4 mN/m. The sandblasted groups show slopes of −1.8 to −1.2 mN/m, the mean value here was −1.6 mN/m. The polar part of the surface energy was, therefore, within the sandblasted group but slightly lower than the mean value (red line). In general, as was the case with the TiAl6V4 samples (PGN 1 to 28), the dispersive component was many times higher than the polar component.

Figure 13.

Owens, Wendt, Rabel and Kaelble plot for the machined titanium rod (cpTi) samples. The green line represents the regression line for the surface in the as-cut state (PGN 56). The red line is the regression line for the mean value of all sandblasted samples (PGN 29-55). The blue area is where all sandblasted regression lines are located.

Discussion

The aim of this study was to investigate the influence of sandblasting with different parameters on surface properties of conventionally machined titanium (cpTi, Grade 4) and additively manufactured TiAl6V4 (Grade 5 Eli) samples.

Due to the manufacturing process cpTi initially had a much lower mean roughness (Sa=0.5 μm) than the additively manufactured TiAl6V4 (Sa=15 μm). The observed mean roughness is close to the roughness observed by Greitemeier et al. (Ra=13.0±0.7 for direct metal laser sintering of TiAl6V4) (31). Overall, sandblasting led to rougher surfaces in the case of cpTi and less rough surfaces in the case of TiAl6V4. In the case of TiAl6V4, the sandblasting process was able to remove the spherical features that originated from the presence of partially molten particles. The strongest effect on the roughness was observed using higher pressure, which can be explained due to the higher kinetic energy per particle. For a similar setup in previous work, the particle speed was 51 m/s for sandblasting pressures of 2 bar and 75 m/s for 6 bar (32). As the machine was slightly modified (different nozzle and a longer blasting tube), those speeds may differ in this work.

In the SEM images, similar features were observed for cpTi and TiAl6V4. For both surfaces, irregular surface with grooves and ridges were observed, whereas in the case of TiAl6V4, some hilly features were observed, especially using a low sandblasting pressure, which likely originated from the spherical particles observed in the as-built state. With higher pressures, these structures disappeared and both surfaces looked more similar in SEM imaging. The sandblasted images also look similar to the titanium grade 2 sandblasted surface in the work by Rupp et al. (33).

The contact angle measurement results for the cpTi and TiAl6V4 titanium surfaces seem to correlate with the surface structure and roughness factors. The TiAl6V4 surface was very hydrophobic in the as-built state and had a low surface energy. This hydrophobicity remained after sandblasting.

The high contact angle that was observed for the additively manufactured TiAl6V4 surfaces in the as-built state indicates that the Cassie-Baxter state dominates. In this state, the drop sits on the peaks of the rough structure, with air pockets trapped under the drop, resulting in reduced wetting and therefore higher contact angles (34). After sandblasting numerous small craters are formed on the surface, which probably increases the number of air pockets on which the drop sits and further increase the Cassie-Baxter state. This results in further increased contact angles and lower surface energies.

In contrast, the as-cut cpTi surface showed a moderate contact angle, indicating the dominance of the Wenzel state. In this state, the drop follows the contours of the surface completely, resulting in a larger actual contact area and thus lower contact angles (35). After sandblasting, the contact angle also increases on these surfaces. This also suggests that small air pockets form under the drop.

Similarities to the study by Yoshimitsu et al. are also apparent, which showed that the Wenzel state dominates at low roughness factors (1.00<Rw<1.10), while the Cassie-Baxter state dominates at higher roughness factors (Rw >1.23) (36). The results of our study show similar trends: TiAl6V4 in the as-built state had a roughness factor of Rw=3.02, those for sandblasted TiAl6V4 samples those for ranged from 1.20 to 1.66, while as-cut cpTi had a Rw=1.02 and sandblasted cpTi samples ranged from 1.14 to 1.27. These parallels indicate that both the additively manufactured titanium surfaces in the as-built state and the sandblasted surfaces are dominated by the Cassie-Baxter state.

A more hydrophilic surface is advantageous for biological applications such as dental implants, as it promotes improved cell adhesion and tissue integration (37). This can be achieved through an additional acid etching process of the sandblasted surfaces, as demonstrated with the sandblasted and acid-etched active surface.

Numerous studies in the literature have measured the contact angle (including water) for titanium (38-41). Elias et al. analysed machined cpTi, among others, and found water contact angles of 85.2˚ at a mean roughness (Ra) of 0.65 μm (38). Ferguson et al. also reported a water contact angle of 91.3˚ for cpTi (39). Similarly, Hotchkiss et al. reported a mean roughness (Sa) of 0.59 μm and a water contact angle of 93.6˚ for smooth cpTi surfaces (40). Taborelli et al. also found a water contact angle of around 90˚, but in their study, the cpTi samples were polished with a roughness (Sa) of 0.06 μm (41). The results of Rupp et al. show that sandblasting with larger nominal particle sizes (250-500 μm) resulted in a mean roughness (Sa) of 2.48 μm and a dynamic water contact angle of 106.6˚ (33). Although the sandblasting particles in our study were smaller (110 μm) and we measured static contact angles, their study showed that the contact angle increased after sandblasting. However, the dynamic contact angle was lower than our static contact angles, which might be due to differences in measurement methods and particle sizes. Chen et al. compared polished and sandblasted TiAl6V4 samples and concluded that the water contact angle increased after sandblasting (42). Strnad et al. (43) analysed CNC-manufactured and sandblasted (250 μm to 300 μm silica particles) TiAl6V4 samples and found a water contact angle of 82.23˚ and a roughness (Ra) of 2.694.

Studies have shown that an increase in the polar and disperse components of surface energy leads to improved cell adhesion and proliferation in osteoblasts and fibroblasts (44-46). For the additively manufactured TiAl6V4 surfaces in this study, the polar component of the surface energy after sandblasting was slightly greater on average compared to the as-built state (−1.5 mN/m in the as-built state and −1.8 mN/m after sandblasting). For cpTi, a clear decrease in the disperse component was seen (5.5 mN/m after cutting and 4.9 mN/m on average after sandblasting). This leads to the assumption that sandblasting alone is not particularly favourable for cell adhesion and proliferation from the point of view of surface energy.

Our study’s findings may be affected by certain limitations. In the static contact angle measurements, standard deviations were observed that in some cases exceeded the limit of 3˚ specified in the EN ISO 19403-2 standard. Regarding the water contact angle, the 3˚ limit was exceeded for one sample each for cpTi (PGN 1, 3.13˚) and TiAl6V4 (PGN 52, 3.5˚). For diiodomethane, the limit was affected more often. For cpTi, one sample exceeded the limit (PGN 37, 3.3˚) and for TiAl6V4, 11 samples exceeded the limit (highest standard deviation was 6.63˚). These deviations likely originate from the irregular surface topography of the sandblasted samples; this notably affected the TiAl6V4 surface in the strongest way when diiodomethane was used. The relative humidity during the contact angle measurements (57% in this study) was also slightly above the recommended range specified in EN ISO 19403-2 standard (50%±5%). For comparison with other studies, another limitation is that sandblasting pressure, together with the unknown size distribution of the sandblasting particles (nominal size 110 μm; size distribution not analysed in this study), may lead to different kinetic energies of the particles when another study repeats the experiment with same pressure but using a different sandblasting machine and/or particle distribution.

In future research within the research unit FOR 5250 (47), the sandblasted additively manufactured surface will undergo acid etching, a common practice for cpTi surfaces in the dental implant sector. It is planned to analyse whether these surfaces can be meaningfully combined with polyelectrolyte multilayer coatings (48) for an improved in vivo response of (dental) implants.

Conclusion

Sandblasting significantly reduced the surface roughness of additive manufactured TiAl6V4 and increased the roughness of machined cpTi.

The surface energy of additively manufactured TiAl6V4 samples did not seem to be significantly affected by sandblasting in the tested parameter range (as-built: 26.6 mJ/m², sandblasted: 26.3 mJ/m²). On the contrary, initially smooth-cut cpTi-surfaces showed a significant reduction in the surface energy after sandblasting (as-cut: 32.3 mJ/m², sandblasted: 26.8 mJ/m²).

Of all the sandblasting parameters investigated, sandblasting pressure appears to have the greatest influence in the tested parameter range. The study showed that a higher sandblasting pressure led to greater material removal from the surface, which plays a decisive role in precise control of the sandblasting process to achieve the desired surface properties.

Funding

This work was supported by the Deutsche Forschungsge-meinschaft (DFG) under project number 449916462.

Conflicts of Interest

The Authors declare no conflicts of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.

Authors’ Contributions

Osman Akbas: Writing – review and editing, writing – original draft, visualization, methodology, investigation, formal analysis, data curation, conceptualization. Leif Reck: Writing – review and editing, investigation, formal analysis, data curation. Anne Jahn: Writing – review and editing, visualization. Jörg Hermsdorf: Writing – review and editing, supervision. Meike Stiesch: Writing – review and editing, supervision, project administration, funding acquisition, conceptualization. Andreas Greuling: Writing – review and editing, supervision, project administration, funding acquisition, conceptualization.