Characterization of ZrBSiTaN_x Films

Round 1

Reviewer 1 Report

The authors investigated the structural properties of ZrBSiTaNx films grown by sputtering under several conditions, as N-flow ratio as reactive gas, sputter power, and post-annealing. The evaluation of structural properties employs many techniques such as XRD, XPS, TEM, and nanoindentation testing. The manuscript is quite long but well-written. I suggest minor corrections in the following points:

i) It is missing important details about the experimental setup, as radiation wavelength employed in XRD and XPS. In addition, details about the measurement of mechanical properties.

ii) Is the uncertainty of deposition rate really 0,01 nm, which is smaller than an atom? Please, adjust the deposition rate with correct significant algorisms.

Author Response

1. It is missing important details about the experimental setup, as radiation wavelength employed in XRD and XPS. In addition, details about the measurement of mechanical properties.

A: Thanks for the suggestion. Related illustrations were modified as:

“The bonding characteristics of films were analyzed using an X-ray photoelectron spectroscope (XPS, PHI 5000 Versaprobe II, ULVAC–PHI, Kanagawa, Japan) with a monochromatic Al K_α X-ray beam operated at 15 kV. The C 1s line from the carbon contamination on the free surface of a Zr₁₀B₉Ta₁₈N₆₃ sample was 284.33 eV. The XPS spectra of Zr 3d, B 1s, Si 2p, Ta 4f, and N 1s core levels were recorded. The splitting energies were 2.43 and 1.91 eV for Zr 3d and Ta 4f doublets [32], respectively. The intensity ratios of I(3d_5/2):I(3d_3/2) and I(4f_7/2):I(4f_5/2) were set as 3:2 and 4:3 for the Zr and Ta doublets, respectively. An Ar+ ion beam of 3 keV was used to sputter the films for depth profiling; the sputter etching rate was 8.2 nm/min for SiO₂. XPS analyses were conducted at depths of 8.2, 16.4, 24.6, 32.8, 41.0, and 49.2 nm. The backgrounds were corrected by a Shirley function and the peaks were fitted with a mixed Gaussian–Lorentzain function. The phases of the films were verified using X-ray diffraction (XRD; X’Pert PRO MPD, PANalytical, Almelo, The Netherlands) with Cu K_α radiation using a grazing incidence technique at an incidence angle of 1°. The applied accelerating voltage and current of XRD were 45 kV and 40 mA, respectively.” in lines 69–83 and “The mechanical properties of the films were measured using a nanoindentation tester (TI-900 Triboindenter, Hysitron, Minneapolis, MN, USA) equipped with a Berkovich diamond probe tip and calculated using the Oliver–Pharr method [33]. The indentation depth was 80 nm.” in lines 85–88.

2. Is the uncertainty of deposition rate really 0,01 nm, which is smaller than an atom? Please, adjust the deposition rate with correct significant algorisms.

A: Thanks for the suggestion! The resolution of the deposition rate is corrected to 0.1 nm as shown in line 125 and Table 2.

Reviewer 2 Report

The submitted manuscript provides highly detained physical and chemical characterization of ZrBSiTa co-sputtered films with the objective of identifying most stable amorphous films suited for mechanical strong coatings for anti-diffusion barriers, high temperature applications and anti-oxidation protective layers. Characterization involves GIXRD, XPS, SEM, EDS, TEM of as-deposited, rapid annealing under both inert and oxygen atmospheres where on-set of crystallization observed only after high temperature annealing and then only as TM oxides.

The work is similar in concept and goal as performed before by other investigators which have been referenced in the paper. Most notably, Musil et al. (ref 16) with ZrSiN films and Bakhit et al., (ref 26) with ZrTaB films. In the submitted work, ZrBSiTa films are examined with and without nitridation.

Overall, the results of measurements are well described and well organized in a clear systematic manner for this multi-parameter study. The conclusions follow results well for the most part and worth publishing. However two specific points in conclusions I did not follow from results of this study. Namely, in point one, “All the as-fabricated ZrBSiTa and (ZrBSiTa)Nx films exhibited amorphous structures due to incorporating B and Si”. I did not follow how the “due to incorporating B and Si” is to be concluded from the data presented here.  GIXRD show no crystallinity in any of the sputtered samples as deposited -even without Si.(Fig 1) I see more stable amorphous phase at higher temperature with incorporation of increased Si.(Fig 6) But there is no clear variation of B content in this study. The B content seemed to be ~1 to 1.5 that of Zr in all samples.  I suspect this conclusion comes from statements in lines 113-114 which I think was inferred from prior studies with other ceramic materials. There is no mention of preparing SiN of BN films in this study in Materials and Methods section. I didn’t see how this was a “conclusion” but only a “educated suspicion”. Authors should elaborate statement in line 113-114 for better understanding and review point one conclusion.

The other point in conclusions to address is in point 2, “The as-fabricated ZrBSiTa films exhibited hardness values of 14.3–19.1 GPa and Young’s modulus values of 242–264 GPa. The hardness of the (ZrBSiTa)Nx thin films decreased with increasing N content due to the high amount of soft amorphous BN phase”. First, I don’t know what an “amorphous phase of BN” is in these multi-elemental films. I believe authors referring to lack on BN crystallites. Based on Table 5, I see the incorporation of N decreasing hardness somewhat but I didn’t see how BN content is to be interpreted in an amorphous film where only N content varied. Further, hardness varied as 11.0-19.1 GPa and modulus as 181-264 GPa. This need to be corrected here.

Some other points to note.

The introduction is ok although since others have achieved amorphous structures as well authors should state more clearly what the objective of their work is over past work done. Also, the Materials and Methods should provide more detail. More specifically and most importantly details of the reactor used (commercial (vendor & model #) or lab-built with more specific such as target-substrate distance. Same for RTA but more importantly what temperature ramps were used. Lastly, on the XPS, what are operating parameters such as pass energy and xray used? How was data analyzed as to energy calibration, background correction or other fitting software used. How were % composition determined from peak areas and how are the uncertainties reported determined? Were there any surface preparation prior to XPS analysis such as light Ar sputtering? Some results are labelled as “49 nm sputter depth”. (Line 121 for example and elsewhere.) That needs to be defined more clearly. Is it a sputtered film at a 49 nm thickness or a film prepared which is Ar sputtered down 49 nm below initial surface. Again, more method details are needed here.

In description of Fig 1 there is no mention as to what the broad peaks observed are due to. That should be mentioned in text - especially since it varies in different samples. Further in lines 110-112, a high delta would not lead favorably to uniform alloy composition. So in your case it is driven more with relative enthalpy values. May want to elaborate that a bit more. Sounded like high delta is contributing to the amorphous nature of film.

In viewing the composition of the sputtered films I see relative Zr and B near ~1-1.5 yet the target used is ZrB2 and others using ZrB2 target had B content about ~2x or higher that of Zr. (ref 26). If there is know reason it be good to state.

Other more minor comments.

 Line 56, "targets were 50.8" should be "targets which were 50.8".

Lines 91-92, Si2N3 has lowest enthalpy so the preference of bonding does not seem right. And in lines 93-94, authors relate most stable equilibrium of individual nitride stoichiometries to the co-sputtered stoichiometries observed. I think “therefore” is too strong a statement here for the amorphous film stoichiometries observed. For example, both substoichiometric, and over stoichiometric ZrNx films are possible with magnetron sputtering. (Catalysis Today, Volume 89, Issue 3, 30 March 2004, Pages 307-312)

 In XPS analysis of Fig 2 Zr 3d5/2 observed is 178.49-178.70 eV and stated as Zr-B since ZrB2 ref # reported 178.9 eV and another report gave 179.2 eV.( Surf. Sci. Spectra 7, 310–315 (2000)). Other Zr literature values are:  Zr-N is 179.3 eV., Zr-Si 178.8 eV, Zr-Ta 178.8 eV. I don’t think we can conclusively state it is Zr-B. Other designations in seemed speculative as well yet stated as conclusive in text.

Also, in Fig 2, why is XPS missing here but shown in Fig 3?

Fig 3e shows N-M label. What does this label refer to?

In comparing Tables 3 and 4, it’s interesting to see ~0.4 eV shift in all XPS peaks with and without N present but no shift in absence of Si.

In line 163, the range of modulus is not same as in the table.

The rest of the results on annealing with inert and air atmosphere is quite interesting. Any discussions on how results of these studies compare to that previously done, especially references 16 and 26 would add more quality to the manuscript.  

Author Response

Overall, the results of measurements are well described and well organized in a clear systematic manner for this multi-parameter study. The conclusions follow results well for the most part and worth publishing. However two specific points in conclusions I did not follow from results of this study. Namely, in point one, “All the as-fabricated ZrBSiTa and (ZrBSiTa)Nx films exhibited amorphous structures due to incorporating B and Si”. I did not follow how the “due to incorporating B and Si” is to be concluded from the data presented here.  GIXRD show no crystallinity in any of the sputtered samples as deposited -even without Si.(Fig 1) I see more stable amorphous phase at higher temperature with incorporation of increased Si.(Fig 6) But there is no clear variation of B content in this study. The B content seemed to be ~1 to 1.5 that of Zr in all samples.  I suspect this conclusion comes from statements in lines 113-114 which I think was inferred from prior studies with other ceramic materials. There is no mention of preparing SiN of BN films in this study in Materials and Methods section. I didn’t see how this was a “conclusion” but only a “educated suspicion”. Authors should elaborate statement in line 113-114 for better understanding and review point one conclusion.

A: New sentences, “In our previous study [40], TaZrN films crystallized into a face-centered cubic phase and revealed a columnar structure. The addition of B and Si into TaZrN films affected the phase structures.”, were added in lines 129–131.

Point 1 in Conclusions was corrected as “ … The over-stoichiometric ratio (x > 1) was obtained for the (ZrBSiTa)N_x films.”.

The other point in conclusions to address is in point 2, “The as-fabricated ZrBSiTa films exhibited hardness values of 14.3–19.1 GPa and Young’s modulus values of 242–264 GPa. The hardness of the (ZrBSiTa)Nx thin films decreased with increasing N content due to the high amount of soft amorphous BN phase”. First, I don’t know what an “amorphous phase of BN” is in these multi-elemental films. I believe authors referring to lack on BN crystallites. Based on Table 5, I see the incorporation of N decreasing hardness somewhat but I didn’t see how BN content is to be interpreted in an amorphous film where only N content varied. Further, hardness varied as 11.0-19.1 GPa and modulus as 181-264 GPa. This need to be corrected here.

A: The sentence “Residual stress was recognized as a minor effect on the mechanical properties of crystalline films, in which crystalline size and phase structure were the major factors.” was mentioned in lines 204–205. A new sentence, “In contrast, residual stress dominated the mechanical properties of amorphous films.”, was added in lines 205 and 206.

 Point 2 in Conclusions was revised as “The as-fabricated ZrBSiTa films exhibited hardness values of 14.3–19.1 GPa and Young’s modulus values of 242–264 GPa. The hardness and Young’s modulus values of the (ZrBSiTa)N_x thin films decreased to 11.0–15.0 and 181–223 GPa, respectively.”

Some other points to note.

The introduction is ok although since others have achieved amorphous structures as well authors should state more clearly what the objective of their work is over past work done. Also, the Materials and Methods should provide more detail. More specifically and most importantly details of the reactor used (commercial (vendor & model #) or lab-built with more specific such as target-substrate distance. Same for RTA but more importantly what temperature ramps were used. Lastly, on the XPS, what are operating parameters such as pass energy and xray used? How was data analyzed as to energy calibration, background correction or other fitting software used. How were % composition determined from peak areas and how are the uncertainties reported determined? Were there any surface preparation prior to XPS analysis such as light Ar sputtering? Some results are labelled as “49 nm sputter depth”. (Line 121 for example and elsewhere.) That needs to be defined more clearly. Is it a sputtered film at a 49 nm thickness or a film prepared which is Ar sputtered down 49 nm below initial surface. Again, more method details are needed here.

A: The “Materials and Methods” was rewritten.

In description of Fig 1 there is no mention as to what the broad peaks observed are due to. That should be mentioned in text - especially since it varies in different samples. Further in lines 110-112, a high delta would not lead favorably to uniform alloy composition. So in your case it is driven more with relative enthalpy values. May want to elaborate that a bit more. Sounded like high delta is contributing to the amorphous nature of film.

A: Related illustrations were modified as “The broad peaks observed at 2θ of 36°–38° for ZrBSiTa films and at 2θ of 34° for (ZrBSiTa)N_x films indicated that all the as-fabricated ZrBSiTa and (ZrBSiTa)N_x films formed amorphous structures. In our previous study [43], TaZrN films crystallized into a face-centered cubic phase and revealed a columnar structure. The addition of B and Si into TaZrN films affected the phase structures. Multi-component alloys could form distinct structures (solid solution, intermediate phase, and bulk metallic glasses) depending on their atomic size difference (δ), mixing enthalpy (ΔH_mix), and mixing entropy (ΔS_mix) [44]. Multi-component bulk metallic glasses have larger δ (6%–18%) and more negative ΔH_mix (−25–−37 kJ/mol) [44]. The batch A (ZrBSiTa) samples exhibited high δ values of 24.0%, 22.6%, 24.2%, and 19.6%, significant and negative ΔH_mix values of −48, −58, −65, and −66 kJ/mol, and medium mixing entropy values of 8.6, 10.7, 11.3, and 10.7 J/K.mol for Zr₂₁B₃₀Ta₄₉ (A1), Zr₂₀B₂₄Si₁₃Ta₄₃ (A2), Zr₁₈B₂₉Si₂₁Ta₃₂ (A3), and Zr₁₅B₁₅Si₄₂Ta₂₈ (A4), respectively, which resulted in forming amorphous structures.” in lines 127–139.

In viewing the composition of the sputtered films I see relative Zr and B near ~1-1.5 yet the target used is ZrB2 and others using ZrB2 target had B content about ~2x or higher that of Zr. (ref 26). If there is know reason it be good to state.

A: New sentences, “ZrB_y, as well as TiB_y [38], tended to form overstoichiometric diboride thin films through sputtering [26]. In our previous study [39], a ZrB_2.5 (28.4% Zr–70.5% B–1.1% O) film was fabricated using a ZrB₂ target. However, all the ZrBSiTa and (ZrBSiTa)N_x films had understoichiometric B/Zr ratios of 0.9–1.6, which could be attributed to scattering and resputtering of light B atoms during cosputtering [26].”, were added in lines 112–117.

Other more minor comments.

Line 56, "targets were 50.8" should be "targets which were 50.8".

A: This sentence was revised as “All the targets were 50.8 mm in diameter.” In line 60.

Lines 91-92, Si2N3 has lowest enthalpy so the preference of bonding does not seem right. And in lines 93-94, authors relate most stable equilibrium of individual nitride stoichiometries to the co-sputtered stoichiometries observed. I think “therefore” is too strong a statement here for the amorphous film stoichiometries observed. For example, both substoichiometric, and over stoichiometric ZrNx films are possible with magnetron sputtering. (Catalysis Today, Volume 89, Issue 3, 30 March 2004, Pages 307-312)

A: The standard formation enthalpy of Si₃N₄ is −744.8 kJ/mol, which is equal to −186.2 kJ/mol of N. The standard formation enthalpies of ZrN, BN, and TaN are higher than that of 1/4(Si₃N₄) on the base per mole of N. The x values >1 for all the (ZrBSiTa)N_x films were realistically observed.

 In XPS analysis of Fig 2 Zr 3d5/2 observed is 178.49-178.70 eV and stated as Zr-B since ZrB2 ref # reported 178.9 eV and another report gave 179.2 eV.( Surf. Sci. Spectra 7, 310–315 (2000)). Other Zr literature values are:  Zr-N is 179.3 eV., Zr-Si 178.8 eV, Zr-Ta 178.8 eV. I don’t think we can conclusively state it is Zr-B. Other designations in seemed speculative as well yet stated as conclusive in text.

A: Fig. 2 and Table 3 show the XPS analyses of batch A samples without composition N. So, no Zr-N needs to be considered. New sentences, “The Zr 3d_5/2 was also reported to be 178.8 eV in ZrSi₂ [46]. However, the standard formation enthalpy of ZrSi₂ at 298 K is −159.4 kJ/mol [32], which is lower than that of ZrB₂. ZrSi₂ should be not the preferentially formed compound.”, were added in lines 155–157.

Also, in Fig 2, why is XPS missing here but shown in Fig 3?

A: This comment is not clear. The signals of B and Zr are shown in Fig. 2a.

Fig 3e shows N-M label. What does this label refer to?

A: “N-M” was corrected as “N 1s”.

In comparing Tables 3 and 4, it’s interesting to see ~0.4 eV shift in all XPS peaks with and without N present but no shift in absence of Si.

A: The titles of Tables 3 and 4 were corrected. The comment for peak shift was revised as “All the aforementioned binding energies of the Zr₁₀B₉Ta₁₈N₆₃ (D1) sample were lower than those of the Zr₇B₆Si₁₀Ta₁₃N₆₄, Zr₅B₆Si₁₇Ta₉N₆₃, and Zr₄B₄Si₂₁Ta₈N₆₃ samples, which could be attributed to charge effect for the Si-containing films with high resistivity.” in lines 180–182. 

In line 163, the range of modulus is not same as in the table.

A: Thanks for the reminder! The typo in line 193 was corrected.

The rest of the results on annealing with inert and air atmosphere is quite interesting. Any discussions on how results of these studies compare to that previously done, especially references 16 and 26 would add more quality to the manuscript.  

A: New sentences, “In this study, B and Si in (ZrBSiTa)N_x films stabilized the amorphous phase. However, the oxidation resistance of (ZrBSiTa)N_x films was determined by the amorphous Si₃N₄ content, which was comparable with the oxidation behavior of ZrSiN films with a high Si content (≥25 at.%) [16].”, were added in lines 293–296.

Ref. 26 explored the mechanical properties and toughness of ZrTaB films. Oxidation resistance was not the topic.