The Size Effect on Forming Quality of Ti–6Al–4V Solid Struts Fabricated via Laser Powder Bed Fusion

Round 1

Reviewer 1 Report

The subject of the research presented in the manuscript can be assessed positively. The work concerns application of advanced SLM technology to manufacture of parts with complex shapes, including porous structures. The authors focused on the effect of 2 factors: scanning speed and size of square cross-section of constructional bars manufactured in the SLM technology.

However, in the present form, the manuscript is not suitable for publication because of substantial errors and unclear presentation of the results.

Comments:

1.        The term “porous structure” used in the title and in the text is misleading. For the tests, flat strut structures, frame type, were manufactured.

2.        Employing the volumetric energy density as a variable parameter does not make sense, if the authors changed scanning speed only at constant laser power, hatch distance and layer thickness, and the examinations were of a comparative nature. Identical values of energy density can be obtained at various sets of parameters and results would be different, especially with use of a high power of laser.

3.        Line 104: „Inside to outside scanning strategy with meander hatch style was selected.”. To eliminate uncertainty, the scanning strategy should be shown schematically.

4.        Figure 3a: In the caption, there is no information about dimensions L_s of the struts for that the surface morphology is shown.

5.        It is necessary to show exemplary images of size deviation on cross-sections of overhanging struts with the angle of 0°C.

6.        Lines 186, 187: …”the refined 2θ locations of the strongest peak of struts with Ls of 1.4 mm 186 generally shifted to higher 2θ with the decrease of energy density, implying that the increased 187 energy density would enlarge the grain size.” Shift of the peaks is caused by a change of chemical composition of the phase α or α', not by a change of grain size. Interpretation of diffractograms should be supported by SEM examination.

7.        Lines 202 - 208. Figure 7. It is impossible to determine and compare sizes of columnar grains on the xy plane. It is possible on the plane xz or yz. In order to compare thickness values of plates of the phase α or α', SEM examinations are necessary.

8.        Figure 7d. Microstructure is shown on the area larger than cross-section area of the bar with its side length L_s = 0.4 mm.

9.        Figure 10a. More details of porous structures (surfaces of the struts, geometry of cells on the planes xy and xz) should be shown at larger magnifications.

10.    Lines 259-262. It would be good to identify, on the grounds of their shapes, types of the pores (gaseous, build defects) dominating in the manufactured specimens depending on the scanning speed.

11.    There is no final conclusion to assess poor quality of the porous structures manufactured in the verification experiment.

Summarizing, the manuscript in its present form is not suited for publication and requires a major revision.

Author Response

Point 1:The term “porous structure” used in the title and in the text is misleading. For the tests, flat strut structures, frame type, were manufactured.

Response 1: Thanks for your thoughtful comment and question. According to the comment, we have restructured the article’s structure. And the article only focused on the effect of the geometric characteristic size on the forming quality of the SLM-produced solid struts. Therefore, the title of  this article was changed into “The Size Effect in Fabrication of Ti-6Al-4V Solid Struts via Selective Laser Melting” and the related content has been revised in the text.

 Point 2: Employing the volumetric energy density as a variable parameter does not make sense, if the authors changed scanning speed only at constant laser power, hatch distance and layer thickness, and the examinations were of a comparative nature. Identical values of energy density can be obtained at various sets of parameters and results would be different, especially with use of a high power of laser.

Response 2: Thanks for your careful reading and professional question. As point 1 said, we  gave up making volumetric energy density as factor in the experiment of this article. Only  the geometric characteristic size effect on the forming quality of the solid struts produced by SLM with different scanning speed was discussed. The “volumetric energy density” was introduced to represent the thermal input during the SLM process, which was needed in section of discussion. And related content has been revised in the text.

 Point 3: Line 104: „Inside to outside scanning strategy with meander hatch style was selected.”. To eliminate uncertainty, the scanning strategy should be shown schematically.

Response 3: Thanks for your careful reading and professional suggestion. According to the comment, We have shown the scanning strategy schematically in Figure 2(c).

 Point 4:Figure 3a: In the caption, there is no information about dimensions Ls of the struts for that the surface morphology is shown.

Response 4: Thanks for your careful reading and professional question. In order to ensure the integrity of the article, we have deleted the related content about surface morphology. 

Point 5:It is necessary to show exemplary images of size deviation on cross-sections of overhanging struts with the angle of 0°.

Response 5: Thanks for your professional suggestion. According to the comment, the exemplary images of size deviation on cross-sections of overhanging struts with the angle of 0° have been shown in Figure 5. And the average value of measuring sizes of the solid struts with Ls of 0.4 and 1.4mm was also summarized in Table 4.

Point 6: Lines 186, 187: …”the refined 2θ locations of the strongest peak of struts with Ls of 1.4 mm 186 generally shifted to higher 2θ with the decrease of energy density, implying that the increased 187 energy density would enlarge the grain size.” Shift of the peaks is caused by a change of chemical composition of the phase α or α', not by a change of grain size. Interpretation of diffractograms should be supported by SEM examination.

Response 6: Thanks for your careful reading and professional question. We are sorry for the mistake in expression. The related expression has been changed into  “For the struts with Ls of 1.4 mm, when the scanning speed increased, the diffraction peaks became considerably broadened and the intensity decreased, which implied the formation of considerable refined crystal” and “In addition, under the process condition of v of 700 mm/s, the peak of the strut with Ls of 0.4 mm became thinner than the strut with Ls of 1.4 mm.”. Under the same condition in v of 700 mm/s, the  difference in grain size of martensites α’  between the solid struts with Ls of 0.4 and 1.4 mm has been shown in SEM images in Figure 8.

Point 7: Lines 202 - 208. Figure 7. It is impossible to determine and compare sizes of columnar grains on the xy plane. It is possible on the plane xz or yz. In order to compare thickness values of plates of the phase α or α', SEM examinations are necessary.

Response 7: Thanks for your careful reading and professional question. During the SLM process of TC4 part,  the  columnar grain of X-Y plane is closely affected by the scanning strategy, like checkerboard scanning style. However, the scanning strategy used in this article, as shown in Figure 2(c), had less impact on the shape and size of primary columnar grains of X-Y plane, because the hatch direction kept changing with the increasing layer. So I thought, to a certain degree, the observed shape and size of   primary columnar grains of X-Y plane could reflect the grain size change during SLM process. In addition, the  difference in grain size of martensites α’  between the solid struts with Ls of 0.4 and 1.4 mm has been observed using SEM and shown in Figure 8. 

Point 8:Figure 7d. Microstructure is shown on the area larger than cross-section area of the bar with its side length Ls = 0.4 mm.

Response 8: Thanks for your careful reading and professional question. According to the comment, in order to more clearly show the difference in microstructure of struts with Ls of 0.4 and 1.4 mm, the optical image of Figure 7(d) has been replaced by SEM images (shown in Figure 8)  

Point 9:Figure 10a. More details of porous structures (surfaces of the struts, geometry of cells on the planes xy and xz) should be shown at larger magnifications.

Response 9: Thanks for your careful reading and professional suggestion. After restructuring the article, the focus point was about the forming quality of the solid struts. The experiment and discussion  about porous structures were merely used to verify the above analysis. And the porosity of porous structure is associated with struts size, the relative density is associated with struts’ pore defects. I thought the tow parameters were most important for the verification experiment in this article.   

Point 10:Lines 259-262. It would be good to identify, on the grounds of their shapes, types of the pores (gaseous, build defects) dominating in the manufactured specimens depending on the scanning speed.

Response 10: Thanks for your careful reading and professional suggestion. According to the comment, the related expression has been revised into “Too high temperature of the molten pool with too high energy density will cause evaporation of elements, generating pores with rounded shapes. Too high cooling rate of molten pool with too low energy density will cause the incomplete fusion of powder, also generating pores with sharp shapes and incompletely melted particles. ”.

Point 11: There is no final conclusion to assess poor quality of the porous structures manufactured in the verification experiment.

Response 11: Thanks for your careful reading and professional suggestion. According the comment, the related content (“This finding indirectly indicated that the struts of porous sample with Ls of 1.4 mm achieved more pores and larger positive size error under the v of 700 mm/s. In other words, porous structure with Ls of 0.6 mm and v of 1900 mm/s gained a better forming quality than that with v of 700 mm/s.”) has been added in text.

Author Response File: Author Response.pdf

Reviewer 2 Report

The paper might contain some new and significant scientific information adequate to justify publication. It should be more clearly stated what is new in this research.

However the reporting of the study need to be improved. The reporting of the experimental methods and results should be more complete and accurate.

In AM I recommend to use ISO / ASTM52900 – 15 Standard Terminology for Additive Manufacturing – General Principles – Terminology. For the sake of clarity and for future understandability and indexing when standard name overrules other.

Literature part there should acknowledge earliest research in making implant from Ti-6Al-4V using power bed fusion (DMLS) process and also there is lot of research about making porous structures. Example references to read:

Salmi, Mika, Jukka Tuomi, Kaija-Stiina Paloheimo, Markku Paloheimo, Roy Björkstrand, Antti A. Mäkitie, Karri Mesimäki, and Risto Kontio. "Digital design and rapid manufacturing in orbital wall reconstruction." In Innovative Developments in Design and Manufacturing, pp. 357-360. CRC Press, 2009.

Dinda, G. P., L. Song, and J. Mazumder. "Fabrication of Ti-6Al-4V scaffolds by direct metal deposition." Metallurgical and Materials Transactions A 39, no. 12 (2008): 2914-2922.

Ponader, Sabine, Cornelius Von Wilmowsky, Martin Widenmayer, Rainer Lutz, Peter Heinl, Carolin Körner, Robert F. Singer, Emeka Nkenke, Friedrich W. Neukam, and Karl A. Schlegel. "In vivo performance of selective electron beam‐melted Ti‐6Al‐4V structures." Journal of Biomedical Materials Research Part A: An Official Journal of The Society for Biomaterials, The Japanese Society for Biomaterials, and The Australian Society for Biomaterials and the Korean Society for Biomaterials 92, no. 1 (2010): 56-62.

The test parts made with parameters and measurements should be gathered in table format. Now text is quite hard to read.

What were all the measuring equipment and procedures? How many samples? How many measurements? How much variation there was between measurements? What was the repeatability? Error limits in all measurements.

Example is am sceptimistic that with 17um resolution could give a total result error of 5um. Measuring equipment should be at least 10 more accurate than the values claiming.

Explain in detail how the experiments of the samples have been planned and analyzed (factors, levels, type of experimental plan, replications, analysis of variance and related statistical tests. Was there predefined experimental design, e.g. a factorial plan? Explain how the process variables(if tested) influence the responses (individual effects and possible interactions).

Density (actual) for the material achieved in this experiment is really low from 98 to 95%. With EOS you should be much closer to 100%. Did you used somehow wrong parameters?

What was the Laser size/shape at powder bed?

I recommend Major revision for the paper to see more about the details and novelty of the submission.

Author Response

Point 1: In AM I recommend to use ISO / ASTM52900 – 15 Standard Terminology for Additive Manufacturing – General Principles – Terminology. For the sake of clarity and for future understandability and indexing when standard name overrules other.

Response 1: Thanks for your careful reading and professional suggestion. According the comment, we have corrected the related content in text using ISO / ASTM52900-15 Standard Terminology for Additive Manufacturing-General Principles-Terminology. And this standard is very helpful for our future research work and paper.

Point 2: Literature part there should acknowledge earliest research in making implant from Ti-6Al-4V using power bed fusion (DMLS) process and also there is lot of research about making porous structures. Example references to read:

Salmi, Mika, Jukka Tuomi, Kaija-Stiina Paloheimo, Markku Paloheimo, Roy Björkstrand, Antti A. Mäkitie, Karri Mesimäki, and Risto Kontio. "Digital design and rapid manufacturing in orbital wall reconstruction." In Innovative Developments in Design and Manufacturing, pp. 357-360. CRC Press, 2009.

Dinda, G. P., L. Song, and J. Mazumder. "Fabrication of Ti-6Al-4V scaffolds by direct metal deposition." Metallurgical and Materials Transactions A 39, no. 12 (2008): 2914-2922.

 Ponader, Sabine, Cornelius Von Wilmowsky, Martin Widenmayer, Rainer Lutz, Peter Heinl, Carolin Körner, Robert F. Singer, Emeka Nkenke, Friedrich W. Neukam, and Karl A. Schlegel. "In vivo performance of selective electron beam‐melted Ti‐6Al‐4V structures." Journal of Biomedical Materials Research Part A: An Official Journal of The Society for Biomaterials, The Japanese Society for Biomaterials, and The Australian Society for Biomaterials and the Korean Society for Biomaterials 92, no. 1 (2010): 56-62.

Response 2: Thanks for your careful reading and professional suggestion. According to the comment, we have added the related references into the introduction section and reference section of this article.

“....

13.  Dinda, G.P.; Song, L.; Mazumder, J. Fabrication of Ti-6Al-4V scaffolds by direct metal deposition. Metall. Mater. Trans. A Phys. Metall. Mater. Sci. 2008,12,2914-2922.

14.  Ponader, S.; Von Wilmowsky, C.; Widenmayer, M.; Lutz, R.; Heinl, P.; Körner, C.; Singer, R.F.; Nkenke, E.; Neukam, F.W.; Schlegel, K.A. In vivo performance of selective electron beam-melted Ti-6Al-4V structures. J. Biomed. Mater. Res. - Part A 2010, 92, 56-62.

15.  Salmi, M.; Tuomi, J.; Paloheimo, K.; Paloheimo, M.; Björkstrand, R.; Mäkitie, A.A.; Mesimäki, K.; Kontio, R. Digital design and rapid manufacturing in orbital wall reconstruction. In Innovative Developments in Design and Manufacturing, 1st ed.; Reddy, J.N., Eds.; CRC Press: London, UK, 2009; pp. 357-360.

....”

Point 3: The test parts made with parameters and measurements should be gathered in table format. Now text is quite hard to read.

Response 3: Thanks for your professional suggestion. According the comment, the main measuring objectives and methods in this study was summarized in Table 3.

Table 3. Measuring objectives and methods in this study.

Point 4: What were all the measuring equipment and procedures? How many samples? How many measurements? How much variation there was between measurements? What was the repeatability? Error limits in all measurements. Example is am sceptimistic that with 17um resolution could give a total result error of 5um. Measuring equipment should be at least 10 more accurate than the values claiming.

Response 4: Thanks for your careful reading and professional question. According the comment, the related content has been corrected and added.  And we are sorry for the mistake in expression about the μCT. The parameters of the μCT were re-verified and corrected in the text.

 “....

The 3D models of the fabricated samples were reconstructed through slice images data using commercially available software (VGStudio MAX 3.0; Volume Graphics GmbH, Germany). The same software was also used to detect the the size deviation distribution of as-built samples comparing to the 3D model with an error of 5 μm (n=3).....

....

The relative density of struts with the same square section were measured using the Archimedes method on each set of samples. Archimedes test results are calculated based on a combination of dry weighing and weighing in pure ethanol and on the theoretical density of 4.43 g/cm3 for Ti-6Al-4V . Due to the too small weight of struts, ten samples of each set were measured together, and the arithmetic mean value of relative density was calculated (n=3) ....

....

nanoindentation tests on the polished sections of as-built struts with an angle of 90º were performed using a nano indenter (G200, Agilent, Ltd., USA) to evaluate the mechanical properties, including elastic modulus and nanohardness. A loading-unloading test mode was used with a maximum indentation depth of 2000 nm, a loading speed of 10 nm/s, and a hold time of 10 s (n=5). The Oliver-Pharr method [27] was then applied to calculate the nanohardness and elastic modulus.....

”

 Point 5: Explain in detail how the experiments of the samples have been planned and analyzed (factors, levels, type of experimental plan, replications, analysis of variance and related statistical tests. Was there predefined experimental design, e.g. a factorial plan? Explain how the process variables(if tested) influence the responses (individual effects and possible interactions).

Response 5: Thanks for your professional suggestion. Based on all the existed comments, we have restructured the article’s structure. And the article only focused on the effect of the geometric characteristic size on the forming quality of the SLM-produced solid struts. Therefore, the title of  this article was changed into “The Size Effect in Fabrication of Ti-6Al-4V Solid Struts via Selective Laser Melting” and the related content has been revised in the text. Specifically, the geometric characteristic size (L_s) was the unique factor, and the experiment was conducted to present its effect on forming quality under different SLM process conditions (too low scanning speed, optimized scanning speed, too high scanning speed)

“

....

Porous structure is generally made up of struts with relatively small size and various angles. In this work, the sample model consisted of struts with angles of 0º, 45º, and 90º, which had a square section with side length (Ls) of 0.4 to 1.4 mm (Figure 2(d)). Three sets of parameters were used in this fabrication process, resulting in different energy densities of 95.24, 51.28, and 35.09 J/mm³ (Table 2 and Figure 2(e)).  To verify the correctness of the analysis, two kinds of porous structures with as-designed Ls of 0.6 mm and 1.4 mm were fabricated via SLM process with v of 700 and 1900 mm/s separately (Figure 2(f)).

Table 2. The selected process parameters for SLM fabrication.

....

”

Point 6: Density (actual) for the material achieved in this experiment is really low from 98 to 95%. With EOS you should be much closer to 100%. Did you used somehow wrong parameters?

Response 6: Thanks for your careful reading and professional question. The density of the SLM-produced part was associated with the the process parameters, powder status, protective atmosphere, etc. Then universal optimized combined process parameters of EOS equipment for TC4 are “P_200W, v_1300 mm/s, h_0.1 mm, d_30 μm, D_0.1 mm”. Too low or too high  scanning speed both make a lowing density.

Point 7: What was the Laser size/shape at powder bed?

Response 6: Thanks for your careful reading and professional question. The EOS SLM machine (M290) was equipped with a Yb fiber laser of 400 W with a wavelength of 1064 nm and a Gaussian spot. The laser spot is circular and its size is about 100 μm (shown in Table 2) In addition, the actual size of the laser spot at powder bed is associated with the process parameters. High laser powder and low scanning speed both make a lager laser spot at powder bed. 

Author Response File: Author Response.pdf

Reviewer 3 Report

This article concerns Ti6Al4V porous structures fabricated by SLM method and reports among others their phase composition, microstructure and mechanical properties. The Authors used few methods such as light and scanning microscopy, micro-tomography, X-ray diffraction and nanoidentation test  for comprehensive characterization of obtained samples. The paper is written in logical way, good in English and presented topic can be useful for AM industry community. However it requires deeply correction, included addition of research results, before publishing:

1.       The title should be more comparable with presented work (only one factor influencing on solid struts was analyzed (volumetric energy density)) and more understandable for readers.

2.       The keywords must be more precisely selected, please specify only the most important ones.

3.       Line 61  - “Apart from the scanning speed, the other main process parameters of SLM, including laser power, hatching space, and layer thickness, are rarely discussed.”  - formally, in this work only influence of scanning speed on metallurgical  quality was discussed. Other parameters as laser power, hatch distance etc. were constant.

4.       Fig.1 – please change the Y axis description – should be Frequency [%].

5.       Line 93 – The wavelength 1064 mm is characteristic for YAG laser. You use the Yb fiber laser. Please check its wavelength and put into text.

6.       Fig.2c – What was an overlapping of laser paths? Are you sure that hatch spacing vs. layer thickness was properly chosen?

7.       Archimedes method must be described more precisely – how was open porosity and high developed surface of samples  taken into consideration during density measurements?

8.       Line 131,132 – please use Light Microscopy term – check throughout article.   

9.       Fig.3a – the scale bars are invisible, please improved it. It is clearly seen the surface texture is extremely different for the top and the side of samples. The roughness parameters must be calculated for each E_v variant.      

10.   Fig.3b  - the maximum shape deviation must be precisely given for each variant.

11.   Fig.4 – Does the graph present average density of measured samples? The standard deviation must be added. Please use “densities” in singular form - check throughout article.

12.   Fig.5 – The shape deviation for presented variants is high, especially for interior areas – please comment it.

13.   Line 194 – The figure number is wrong – please correct it.

14.   Fig.5a - In my opinion peak from beta phase is accidentally observed. You should remember that samples after SLM process are characterized by strong texture. Due to beta phase peaks can be invisible. The beta phase must be identified by more precisely methods, e.g. EBSD.

15.   Fig.7 – The presented microstructure pictures for different volumetric energy density are very similar. In my opinion beta phase exists in all technological variants.

16.   Fig.8b,c – Why were only minimum and maximum values for 35.09 J/mm³ and 95.24 J/mm³ were showed? Why high energy variants has the lowest plastic modulus and nanohardness at higher density than samples obtained with the lowest E_v? The hardness value of SLM samples should be evaluated trough micro or macro hardness measurement.

17.   Fig.10b – It is hard to understanding – relative density is almost the same as porosity value. The presented variants - design, 35.09 J/mm³ and 95.24 J/mm³ - should be denoted more clearly.

Author Response

Point 1: The title should be more comparable with presented work (only one factor influencing on solid struts was analyzed (volumetric energy density)) and more understandable for readers.

Response 1: Thanks for your careful reading and professional suggestion. According to the comment, we have restructured the article’s structure. And the article only focused on the effect of the geometric characteristic size on the forming quality of the SLM-produced solid struts under different conditions in scanning speed. Therefore, the title of this article was changed into “The Size Effect in Fabrication of Ti-6Al-4V Solid Struts via Selective Laser Melting” and the related content has been revised in the text.

Point 2: The keywords must be more precisely selected, please specify only the most important ones.

Response 2: Thanks for your careful reading and professional suggestion. According to the comment, we have refined the keywords in the manuscript.

Point 3: Line 61 - “Apart from the scanning speed, the other main process parameters of SLM, including laser power, hatching space, and layer thickness, are rarely discussed.”  - formally, in this work only influence of scanning speed on metallurgical quality was discussed. Other parameters as laser power, hatch distance etc. were constant.

Response 3: Thanks for your careful reading and professional suggestion. As the response of point 1 said, we gave up making volumetric energy density as factor in the experiment of this article. Only the geometric characteristic size effect on the forming quality of the solid struts produced by SLM with different scanning speed was discussed. The “volumetric energy density” was introduced to represent the thermal input during the SLM process, which was needed in section of discussion. And related content has been revised in the manuscript.

Point 4: Fig.1 – please change the Y axis description – should be Frequency [%].

Response 4: Thanks for your careful reading and professional suggestion. According to the comment, we had corrected the Fig 1 in the manuscript.

Point 5: Line 93 – The wavelength 1064 nm is characteristic for YAG laser. You use the Yb fiber laser. Please check its wavelength and put into text.

Response 5: Thanks for your careful reading. We are sorry for the mistake about the laser wavelength. By checking in the equipment specification of EOS M290, the wavelength of the used laser is range of 1000~1100 nm. And related content has been revised in the manuscript.

Point 6: Fig.2c – What was an overlapping of laser paths? Are you sure that hatch spacing vs. layer thickness was properly chosen?

Response 6: Thanks for your careful reading and professional question. According to the comment, we have shown the scanning strategy schematically in Figure 2(c). The selected parameters of hatching space and the layer thickness were the optimized parameters in fabrication TC4 part for EOS M290. And the actual overlapping of the laser path is associated with the laser power and scanning speed. Both high laser power and low scanning speed can result in relatively large overlapping. As a general rule in SLM process experiment, the laser powder and scanning speed are changeable parameters, and the hatching space and the layer thickness are fixed parameters.

Point 7: Archimedes method must be described more precisely – how was open porosity and high developed surface of samples taken into consideration during density measurements?

Response 7: Thanks for your careful reading and professional suggestion. The Archimedes method was used to measure the sample’s relative density. Taking open porosity and high developed surface of samples into consideration, the samples were coated with max after first dry weighing. And the related content has been revised in the manuscript.

“

Taking open porosity and high developed surface of samples into consideration, the samples were coated with max after first dry weighing. The relative density (ρrelative)of the samples was calculated using the Eq(2):

ρrelative=m1/{(m2-m3)/ρethanol-(m2-m1)/ρmax}                  (2)

where m1 was the mass of the sample without max coating in air, m2 was the mass of the sample with max coating in air, and m3 was the mass of the sample with max coating in pure ethanol.

”

Point 8: Line 131,132 – please use Light Microscopy term – check throughout article.  

Response 8: Thanks for your careful reading. We are sorry for the mistake. According to the comment, the related content has been revised in the manuscript.

Point 9: Fig.3a – the scale bars are invisible, please improved it. It is clearly seen the surface texture is extremely different for the top and the side of samples. The roughness parameters must be calculated for each Ev variant.     

Response 9: Thanks for your careful reading and professional suggestion. In order to ensure the integrity of the article, we have deleted the related content about surface morphology.

Point 10: Fig.3b - the maximum shape deviation must be precisely given for each variant.

Response 10: Thanks for your careful reading and professional suggestion. After restructuring this article, Figure 3 was only used for presenting the the geometric profile error distribution among solid struts with different Ls. And the average value of measuring sizes of the solid struts with Ls of 0.4 and 1.4 mm was also summarized in newly added Table 4.

Point 11: Fig.4 – Does the graph present average density of measured samples? The standard deviation must be added. Please use “densities” in singular form - check throughout article.

Response 11: Thanks for your careful reading and professional suggestion. Due to the too small weight of single solid strut, ten struts of each set were measured together, and the arithmetic mean value of relative density was calculated and shown in Figure 4 (n=10). Therefore there was no standard deviation in it. We are sorry for the mistake in “densities”, and related content has been corrected.

Point 12: Fig.5 – The shape deviation for presented variants is high, especially for interior areas – please comment it.

Response 12: Thanks for your careful reading and professional suggestion. According to the comment, the interior areas should mean the overhanging solid strut. The high shape deviation was attributed to the permeation effect on powder bed of molten pool. And the related comment has been added and can be found in the text.

“….

For the overhanging struts with angle of 0º, the actual size was larger than the vertical struts. This was because the molten pool would usually generate many tumor forms in the bottom surface, owning to the permeation effect, causing uneven geometric profile and over dimension as a condition of low scanning speed (Figure 5).

….”

Point 13: Line 194 – The figure number is wrong – please correct it.

Response 13: Thanks for your careful reading. We are sorry for the mistake. According to the comment, the related content has been revised in the manuscript.

Point 14: Fig.6a - In my opinion peak from beta phase is accidentally observed. You should remember that samples after SLM process are characterized by strong texture. Due to beta phase peaks can be invisible. The beta phase must be identified by more precisely methods, e.g. EBSD.

Response 14: Thanks for your careful reading and professional suggestion. In this paper, the XRD was merely conducted for the qualitative analysis of grain size. And the rule of the size effect on mechanical properties could be explained through the XRD result and SEM/LM images of microstructure. Therefore, the content about the beta phase was unnecessary in this paper. According to the comment, the related content has been deleted in the manuscript.

Point 15: Fig.7 – The presented microstructure pictures for different volumetric energy density are very similar. In my opinion beta phase exists in all technological variants.

Response 15: Thanks for your careful reading and professional question. We totally agree with you about the above opinion. In order to show the difference in microstructure of the solid struts with different scanning speed, the Figure 7 has been changed, showing the change of the primary columnar β grain more clearly. And a new figure (Figure 8) has been added to present the difference in martensites α’ of solid strut with different Ls.

Point 16: Fig.8b,c – Why were only minimum and maximum values for 35.09 J/mm3 and 95.24 J/mm3 were showed? Why high energy variants has the lowest plastic modulus and nanohardness at higher density than samples obtained with the lowest Ev? The hardness value of SLM samples should be evaluated trough micro or macro hardness measurement.

Response 16: Thanks for your careful reading and professional question. For the first question, it is because that after restructuring the article, the main topic has been refined, which was about the effect of geometric characteristic size on the forming quality of solid strut with different SLM-process condition in scanning speed. The solid struts with two classical geometric characteristic size (0.4 and 1.4 mm) was regarded as the major research object. Therefore, under different process condition in scanning speed (700, 1300, and 1900 mm/s), the mechanical properties of the solid struts with Ls of 0.4 and 1.4 mm were measured and shown. For the second question, it is because that the hardness is closely related to microstructure and elastic modulus is associated with the residual stress level to a certain extent. Solid strut with highest Ev (lowest scanning speed) had the coarsest grain microstructure and lowest residual stress level, which resulted in lowest plastic modulus and nanohardness. For the third question, it is because that the solid strut has too small size and homogeneous microstructure, so the mechanical performance could be effectively characterized to a certain degree through nanoindentation test.

Point 17: Fig.10b – It is hard to understanding – relative density is almost the same as porosity value. The presented variants - design, 35.09 J/mm3 and 95.24 J/mm3 - should be denoted more clearly.

Response 17: Thanks for your careful reading and professional question. The relative density is not associated with porosity. Having similar value was merely because that the porous structure with Ls of 0.6 had a very high design porosity. And according to the comment, the figure (Figure 11) has been adjusted in the manuscript.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

No further comments.

Author Response

Thanks very much.

Reviewer 2 Report

The terminology is still not correct. Please use example powder bed fusion not SLM.

Author Response

Point 1: The terminology is still not correct. Please use example powder bed fusion not SLM.

Response1: Thanks for your careful reading and professional suggestion. According to the standards of ISO / ASTM52900 , AMS7003, and IOS 17296-2, we have corrected the related content in manuscript, changing “selective laser melting” into “laser powder bed fusion”, changing “scanning speed” into “scan speed”, changing “hatching space” into “hatch spacing”, etc.

Reviewer 3 Report

The article has been significantly improved. The most of Author’s explanations is satisfactory, too. However some corrections are still needed:

1.       The title is still sounding strange – The size effect of what???

2.       The keywords must be more precisely selected – please specify three, maximum four keywords the best describing this work.

3.       Table 1 – please use “(wt. %)” instead of “(mass %).

4.       Fig.2c – Are you sure the strut core is built firstly??  

5.       Table 3 - The apparatus and research methodology information are already given in Materials and Methods section. This table is no needed.

6.       Line 157, 158, 159 – What do the percent values mean? In my opinion absolute dimensional deviation must be given for all technological variants.  

7.       Table 4 – Please order the scan speeds from min. to max. Please marked the place of measuring as 1, 2, 3 and put into the table instead of Angle values. Please change of “Measuring positions” to “Measuring place”. If the size is given as average value the standard deviation must be added.

8.       Fig.6 – What does “Wide 2q” mean? Please comment why only for E_v = 95.24 J/mm³ the peak from beta phase is visible (Fig.7 – the micrographs suggest that beta phase is existed in all analyzed variants).  

9.       Fig.7 – What does “Optical microstructure…” mean. Please define it.

10.   Fig.11 – Where are the Fig.11a and  the Fig.11b? Please precisely explained why e.g. relative density of Ls =  0.6 mm, Ev = 35.09 J/mm³ is ~96% at ~89% of porosity – it is impossible!!

The English, especially highlighted text, must be strongly improved. I recommends native speaker support.

Author Response

Point 1: The title is still sounding strange – The size effect of what???

Response 1: Thanks for your careful reading and professional suggestion. We have change the title into “The Size Effect on Forming Quality Fabrication of Ti-6Al-4V Solid Struts Fabricated via Laser Powder Bed Fusion”.

Point 2: The keywords must be more precisely selected – please specify three, maximum four keywords the best describing this work.

Response 2: Thanks for your careful reading and professional suggestion. According to the comment, we kept four keywords in the manuscript.

Point 3: Table 1 – please use “(wt. %)” instead of “(mass %).

Response 3: Thanks for your careful reading and professional suggestion. According to the comment, we have corrected the related mistake in the manuscript.

Point 4: Fig.2c – Are you sure the strut core is built firstly?? 

Response 4: Thanks for your careful reading and professional suggestion. For the part with small geometric characteristic size or porous structure to be fabricated using EOS M290, the scanning strategy shown in Figure 2c is usually selected through the EOSprint software.

Point 5: Table 3 - The apparatus and research methodology information are already given in Materials and Methods section. This table is no needed.

Response 5: Thanks for your careful reading and professional suggestion. Table 3 was added according to another reviewer’s comment for the purpose of easy reading.

Point 6: Line 157, 158, 159 – What do the percent values mean? In my opinion absolute dimensional deviation must be given for all technological variants. 

Response 6: Thanks for your careful reading and professional suggestion. We are sorry for the unclear expression. The percent values were used the present variation degree of the as-built size relative to the design size. According to the comment, we have corrected the related mistake in Line 157, 158, 159.

Point 7: Table 4 – Please order the scan speeds from min. to max. Please marked the place of measuring as 1, 2, 3 and put into the table instead of Angle values. Please change of “Measuring positions” to “Measuring place”. If the size is given as average value the standard deviation must be added.

Response 7: Thanks for your careful reading and professional suggestion. According to the comment, we have corrected the related content in Table 4.

Point 8:  Fig.6 – What does “Wide 2θ” mean? Please comment why only for Ev = 95.24 J/mm3 the peak from beta phase is visible (Fig.7 – the micrographs suggest that beta phase is existed in all analyzed variants). 

Response 8: Thanks for your careful reading and professional suggestion. We are sorry for the unclear expression. The wide 2θ is the diffraction angle of XRD. And the phase of TC4 alloy can be detected in the 2θ of 30-80°. Actually, the β phase existed in all the SLM-built samples. On one hand, generally, because of too little content, it is difficult to be reflected on the XRD result. And the primary β grain mainly consisted of acicular martensites α’ throughout the entire grain. On the other hand, the content of β phase is associated with the cooling condition of SLM process. The strut with Ls of 0.6 mm and Ev of 95.24 J/mm3 had the worst cooling condition, which mean the melton pool would keep longer time in relatively high temperature. Although occasionally, the beta phase peak of strut with Ls of 0.6 mm and Ev of 95.24 J/mm3 was detected.

Point 9: Fig.7 – What does “Optical microstructure…” mean. Please define it.

Response 9: Thanks for your careful reading. We are sorry for this mistake in expression. It should be “Optical images of microstructure…….”

Point 10: Fig.11 – Where are the Fig.11a and  the Fig.11b? Please precisely explained why e.g. relative density of Ls =  0.6 mm, Ev = 35.09 J/mm3 is ~96% at ~89% of porosity – it is impossible!!

Response 10: Thanks for your careful reading. We are sorry for this mistake about the caption of Figure 11. It should be noted that the porosity is a design parameter only for the porous structure. And the relative density is the classical parameter for the part using powder metallurgy method like SLM. There is no relationship between the two parameters. And the relative density is associated with the pore deficiency, and the porosity is associated with the the size of as-built struts of porous structure. According to the comment, we added a related content “As know, the relative density is associated with the pore deficiency, and the porosity is associated with the size of as-built struts of porous structure….” for explaining the Figure 11.

Author Response File: Author Response.pdf

Round 3

Reviewer 3 Report

Dear Authors,

I hope the following correction list will be the last one:

1.       I suggest use more properly keywords, e.g.: selective laser melting, Ti6Al4V alloy, metallurgical quality, mechanical properties …

2.       Please use “properties” instead of “property” – check throughout article.

3.       Table 3 – Please put space between Measuring and objectives. “Equipment Model and Parameters” – I suggest use “Apparatus”. Please give information about apparatus for density measurements by Archimedes method. Please use “Microstructure observations” as the last objective.

4.        Line 155, 156, 157 – The average value ±SD are enough. The percent values are not needed.

5.        Table 4  - L_s must be given in millimeters (column 2). Please use “Measuring place”
 (column 6).

6.       Fig.4 – If you use average values of relative density the error bars must be added for each measuring point.

Please improved English of highlighted text. I recommend  support of native speaker.

Author Response

Point 1: I suggest use more properly keywords, e.g.: selective laser melting, Ti6Al4V alloy, metallurgical quality, mechanical properties …

Response 1: Thanks for your professional suggestion. According to the comment, we have changed the keywords in the manuscript.

Point 2: Please use “properties” instead of “property” – check throughout article.

Response 2: Thanks for your careful reading and professional suggestion. According to the comment, we have corrected the related content throughout the manuscript.

Point 3: Table 3 – Please put space between Measuring and objectives. “Equipment Model and Parameters” – I suggest use “Apparatus”. Please give information about apparatus for density measurements by Archimedes method. Please use “Microstructure observations” as the last objective.

Response 3: Thanks for your careful reading and professional suggestion. According to the comment, we have corrected the related content in Table 3.

Point 4: Line 155, 156, 157 – The average value ±SD are enough. The percent values are not needed.

Response 4: Thanks for your professional suggestion. According to the comment, we have corrected the related content in Line 155, 156, 157.

Point 5: Table 4  - Ls must be given in millimeters (column 2). Please use “Measuring place”

 (column 6).

Response 5: Thanks for your professional suggestion. According to the comment, we have corrected the related content in the Table 4.

Point 6: Fig.4 – If you use average values of relative density the error bars must be added for each measuring point.

Response 6: Thanks for your careful reading and professional suggestion. We have added the error bars into Figure 4 based the original measuring data.

Point 7: Please improved English of highlighted text. I recommend support of native speaker.

Response 7: Thanks for your careful reading and suggestion. We have asked MDPI a help in language editing.

Author Response File: Author Response.pdf