Tuning the Morphology and Properties of Nanostructured Cu-ZnO Thin Films Using a Two-Step Sputtering Technique

Round 1

Reviewer 1 Report

The authors reported the Cu-ZnO thin film by using the modified two-step sputtering technique. A change in conductivity and growth mechanism was proposed based on their results. It should be of general interest for people interested in thin-film technology. The reviewer thinks that the work is potentially suitable for publication in Metals. Prior to publication, the following comments should be fully addressed to further improve the manuscript.

1. The author should provide the carrier concentration and mobility of Cu-ZnO film with different Cu contents.
2. please give the comment on Figure 5 for the electrical properties of Cu embedded ZnO film, for example, role of the interface between Cu and ZnO. [ex. ACS Appl. Mater. Interfaces 2020, 12, 4715−4721]
3. please modify the English uses for all the text.

 

Author Response

Reviewer1

The authors reported the Cu-ZnO thin film by using the modified two-step sputtering technique. A change in conductivity and growth mechanism was proposed based on their results. It should be of general interest for people interested in thin-film technology. The reviewer thinks that the work is potentially suitable for publication in Metals. Prior to publication, the following comments should be fully addressed to further improve the manuscript.

1. The author should provide the carrier concentration and mobility of Cu-ZnO film with different Cu contents.
2. please give the comment on Figure 5 for the electrical properties of Cu embedded ZnO film, for example, role of the interface between Cu and ZnO. [ex. ACS Appl. Mater. Interfaces 2020, 12, 4715−4721]
3. please modify the English uses for all the text.

Answers and rebuttals)

We deeply appreciate the comments.

1. We showed the conductivities of Cu-doped ZnO films. Unfortunately, we did not perform Hall measurements because the change of morphologies and related properties were focused in this manuscript.
2. Thank you for the nice comment. However, Fig. 5 just shows the formation mechanism of the broccoli-like structure only and is fully explained in the manuscript.

We believe excess Cu might be existed as atomic clusters (maybe Cu nanoparticles but not certain). We didn’t consider the interface effect during the research so don’t have any direct evidence. So we referenced that article as an assumption in the revised manuscript. (ref. 46)

3. We have checked English again.

Reviewer 2 Report

My biggest concern about this article is in the authors calling the samples a "cu-doped zinc oxide" film. From the deposition method, it is more likely that the film is actualy a bilayer of Cu:ZnO and ZnO. The authors have not decisively shown that the film is definitely homogenous.

Co-sputtering with turning on/off one target is not entirely novel. However, I agree with the authors that there seem to be no study on this bilayer film specifically.

The authors also clain "ultra-high condictivity" but the data (and the authors) suggests that this is because of metallic Cu particles that precipitate out of the film. But this is nothing special? It's not more conductive than metallic Cu, for example?

P2L66-69: description of the gas composition into the chamber during pre-sputtering and 1st stage of sputtering is not found
P2L77: description of the chamber is very important for reproducibility but missing. E.g. target-substrate distance, angles between substrate and target(s), direction of gas flow, gas flow rate, pumping condition (e.g. turbo / cryo pump, where?)
P2L78: It would be most interesting to have C40 as another extreme point comparison (other than C0)
P3L92: clarification: do you mean hexagonal ZnO? i.e. is the lattice constant the same with ZnO too?
P3L95: "Cu-related peaks might be formed from metallic Cu, CuO, and Cu2O" -- this doesn't make sense! Cu peak is simply an indication of the presence of Cu metal, most likely in micro-crystalline form
P3L99: Figures X-axis are in degree. If so, then need to specify X-ray source wavelength or type (e.g. Cu-k-alpha). Alternatively, we can report the X-axis in q.
P3L109: "brocolli-like crystal structure" : calling it a "crystal structure" (hexagonal, diamond, square, triclinic, etc.) is misleading. Calling it "nanostructure" is also misleading since the structure is clearly larger than 200 nm (see the scale in the figures). Morphology" is probably a better terminology. Check for examples https://doi.org/10.1016/j.tsf.2013.03.041.
P4L123: one of the key assumption of W-H analysis is that the film is at least qausi-homogenous within the sampling volume of the X-ray. However, because of the two-step deposition process, I would expect the film to be a Cu:ZnO / ZnO bilayer. This breaks the assumption of WH and therefore this analysis is invalid.

Author Response

Reviewer2

My biggest concern about this article is in the authors calling the samples a "cu-doped zinc oxide" film. From the deposition method, it is more likely that the film is actually a bilayer of Cu:ZnO and ZnO. The authors have not decisively shown that the film is definitely homogenous. Co-sputtering with turning on/off one target is not entirely novel. However, I agree with the authors that there seem to be no study on this bilayer film specifically. The authors also claim "ultra-high conductivity" but the data (and the authors) suggests that this is because of metallic Cu particles that precipitate out of the film. But this is nothing special? It's not more conductive than metallic Cu, for example?

Answers and rebuttals)

We deeply appreciate the comments. We partially agree with the reviewer’s concern. For the structure of ZnO films, we admit that films can be a bi-layer. Getting definite evidence is very difficult because film compositions are the same. In this study, all ZnO films were annealed at 400°C for 1 h. Thus, the lightly Cu-doped ZnO films might be homogeneous after high-temperature annealing.

P2L66-69: description of the gas composition into the chamber during pre-sputtering and 1st stage of sputtering is not found.

Ans) We expressed in the manuscript, “The targets were pre-sputtered with argon plasma for 10 min”. The Ar flow was maintained at 4.55 sccm, and the substrate temperature was set to 150°C.

P2L77: description of the chamber is very important for reproducibility but missing. E.g. target-substrate distance, angles between substrate and target(s), direction of gas flow, gas flow rate, pumping condition (e.g. turbo / cryo pump, where?)

Ans) Vacuum chamber: 50 cm in diameter, target-substrate distance: less than 10 cm, tiled (~15 °) RF/magnetron gun, pumping: turbo-molecular pump. This information is newly added in the supplementary information.

P2L78: It would be most interesting to have C40 as another extreme point comparison (other than C0)

Ans) C40 (fully co-sputtered ZnO) indicated inhomogeneous ZnO films with much metallic Cu clusters. Thus, we set the control sample to be C0.

P3L92: clarification: do you mean hexagonal ZnO? i.e. is the lattice constant the same with ZnO too?

Ans) The lightly doped ZnO maintains a hexagonal crystal structure (Fig. 1). We modified the manuscript. Because peaks were shifted, the lattice constants changed depending on x.

P3L95: "Cu-related peaks might be formed from metallic Cu, CuO, and Cu2O" -- this doesn't make sense! Cu peak is simply an indication of the presence of Cu metal, most likely in micro-crystalline form

Ans) Following this comment, we deleted “Cu-related peaks might be formed from metallic Cu, CuO, and Cu₂O.”

P3L99: Figures X-axis are in degree. If so, then need to specify X-ray source wavelength or type (e.g. Cu-k-alpha). Alternatively, we can report the X-axis in q.

Ans) We used the CuKa radiation. This was added in the Materials and Methods section.

P3L109: "broccoli-like crystal structure" : calling it a "crystal structure" (hexagonal, diamond, square, triclinic, etc.) is misleading. Calling it "nanostructure" is also misleading since the structure is clearly larger than 200 nm (see the scale in the figures). Morphology" is probably a better terminology. Check for examples https://doi.org/10.1016/j.tsf.2013.03.041.

Ans) We understand this concern. In P3L109, we expressed “broccoli-like clusters”.

P4L123: one of the key assumption of W-H analysis is that the film is at least quasi-homogenous within the sampling volume of the X-ray. However, because of the two-step deposition process, I would expect the film to be a Cu:ZnO / ZnO bilayer. This breaks the assumption of WH and therefore this analysis is invalid.

Answers and rebuttals)

We understand the concern of the reviewer and that’s why we used a uniform deformation model (UDM) for W-H analysis to obtain the strain and the size of the ZnO lattice instead of uniform stress-deformation model (UDSM). USDM requires exact material parameters such as Young’s Modulus, but UDM analysis can be accepted with the much simpler conditions if all XRD peaks are independent without overlapping (like the present analysis). We only choose ZnO peaks so the W-H uniform deformation could be accepted if background Cu signal is trivial in the peak area. As this reviewer commented, there could be depth-gradient in these films, but we believe those would become negligible after the thermal annealing process. Also, the W-H strain is valid as the ‘overall’ strain of all X-ray sampling volume, including the strain caused by the internal gradient. For this author, WH analysis is applicable to the system with an inhomogeneously strained crystalline case. In addition, please consider the WH formulae are approximate approach (for simple information, see http://pd.chem.ucl.ac.uk/pdnn/peaks/size.htm).

Reviewer 3 Report

The manuscript is poorly written in language and needs to be significantly improved. No clear story is given and the novelty is not clearly demonstrated. It spents quite a bit efforts to investigate the found broccoli-like crystal cluster and proposes an explanation of its formation. However, the author failed to demonstrate the scientific significance of this crystal.

Author Response

Reviewer3

The manuscript is poorly written in language and needs to be significantly improved. No clear story is given and the novelty is not clearly demonstrated. It spents quite a bit efforts to investigate the found broccoli-like crystal cluster and proposes an explanation of its formation. However, the author failed to demonstrate the scientific significance of this crystal.

Answers)

I am sorry that this paper did not get the attention of the reviewer. However, I consider this manuscript is meaningful to present the variety of sputtering techniques that have not been reported to date. In the manuscript, the formation of broccoli-like clusters is not the main but an outcome for Cu nanocluster formation. In addition, the correlation between defects and Raman peak shifts is meaningful The manuscript is checked again by a native scientist.

Reviewer 4 Report

Line 61: 5 cm x 5 cm (squared-cm)

Line 70: the usual unit is “sccm” (standard cubic centimetre per min).

Materials: the first paragraph should be more logically arranged. Start with the substrate including its distance to the sputter sources. Was there a shutter? Next: all information regarding both magnetrons: size, targets, deposition time, RF frequency (missing), … How large are the sputtering yields/deposition rates of ZnO and Cu?

Line 78: what is the purpose of annealing and what difference does it make?

Line 81: wavelength of x-ray source? employed scattering geometry?

Line 95: the observed peak evidently originates from a Cu lattice, so, to mention Cu2O and CuO is not really helpful here.

Line 98: I do not see how is it possible that the lattice constant can become smaller?

Line 107: how large are the grains of each figure?

Fig. 2: Add scale to each figure. Add electron energy to caption.

Refs. 9 and 32: check/correct author names?

Line 133: Where is figure S2? No supplement attached!

Fig. 3b: what is shown here? Strain and crystal (?) size?

Line 156: delete “rate”

Line 169: …slightly different “value” ?

Line 188: Tauc plots are shown, not band gaps.

Lines 202/203: XRD shows that there are small Cu crystallites in the ZnO lattice. A brief discussion of this is required.

Author Response

Answers and rebuttals) We deeply appreciate the comments.

Line 61: 5 cm x 5 cm (squared-cm)

Answers) We corrected in the revised manuscript.  

Line 70: the usual unit is “sccm” (standard cubic centimetre per min).

Answers) We corrected in the revised manuscript.  

Materials: the first paragraph should be more logically arranged. Start with the substrate including its distance to the sputter sources. Was there a shutter? Next: all information regarding both magnetrons: size, targets, deposition time, RF frequency (missing), … How large are the sputtering yields/deposition rates of ZnO and Cu?

Answers) We added detail information about the sputtering system in the supporting information.  

Line 78: what is the purpose of annealing and what difference does it make?

Answers) We performed high-temperature annealing to make Cu-ZnO films homogeneous. Also, the thermal annealing process densifies the film, resulting in more suitable optoelectronic properties.

Line 81: wavelength of x-ray source? employed scattering geometry?

Answers) We used the CuKa radiation.

Line 95: the observed peak evidently originates from a Cu lattice, so, to mention Cu2O and CuO is not really helpful here.

Answers) We deleted the expression “Cu-related peaks might be formed from metallic Cu, CuO, and Cu₂O”.

Line 98: I do not see how is it possible that the lattice constant can become smaller?

Answers) Lattice constants are fixed value in an ideal crystal. However, internal impurities could deform the crystal lattice so measured lattice constant in XRD data could be different. In this manuscript, 2theta of two XRD peaks, (100H) and (002H) is increased, which indicates shrinkage of Miller distance. d(100H) is a, and d(002H) is the half of c, considering the crystal structure. In conclusion, those lattice constants became smaller.

Line 107: how large are the grains of each figure?

Answers) We newly added the scale bar in the SEM images.

Fig. 2: Add scale to each figure. Add electron energy to caption.’

Answers) Electron acceleration voltage (15kv) is indicated in each SEM image.

Refs. 9 and 32: check/correct author names?

Answers) We corrected the author's names.

Line 133: Where is figure S2? No supplement attached!

Answers) We have uploaded the supplementary information. Please find the uploaded file.

Fig. 3b: what is shown here? Strain and crystal (?) size?

Answers) In Fig. 3b, the y-axis shows strain calculated from WH formula.

Line 156: delete “rate”

Answers) We deleted in the revised manuscript.

Line 169: …slightly different “value” ?

Answers) There is an error. “has a slightly different” changed to “was slightly different”

Line 188: Tauc plots are shown, not band gaps.

Answers) From Tauc plots, we drew a straight line and determined an optical bandgap.

Lines 202/203: XRD shows that there are small Cu crystallites in the ZnO lattice. A brief discussion of this is required.

Answers) In lightly Cu-doped films, there were no Cu. As Cu sputtering time increases, we found metallic Cu in XRD patterns. We discussed in the manuscript that the formation of the broccoli-like structure was attributed to the presence of Cu nanoclusters, which is similar to the previous report.

Round 2

Reviewer 2 Report

OPEN QUESTIONS

Fig S2: Assuming that the dotty area is from the soda glass, I agree that there's no evidence of bi-layerness. However, the film thicknesses, esp. C10, is very different. I believe the authors haven't put this in the main body, although they claim that the thicknesses have been measured with a profilometer (and is needed for 4-point probe). I think this additional data will go well with Fig. 5.

My biggest concern about this article is in the authors calling the samples a "cu-doped zinc oxide" film. From the deposition method, it is more likely that the film is actually a bilayer of Cu:ZnO and ZnO. The authors have not decisively shown that the film is definitely homogenous. Co-sputtering with turning on/off one target is not entirely novel. However, I agree with the authors that there seem to be no study on this bilayer film specifically. The authors also claim "ultra-high conductivity" but the data (and the authors) suggests that this is because of metallic Cu particles that precipitate out of the film. But this is nothing special? It's not more conductive than metallic Cu, for example?

Answers and rebuttals)

We deeply appreciate the comments. We partially agree with the reviewer’s concern. For the structure of ZnO films, we admit that films can be a bi-layer. Getting definite evidence is very difficult because film compositions are the same. In this study, all ZnO films were annealed at 400°C for 1 h. Thus, the lightly Cu-doped ZnO films might be homogeneous after high-temperature annealing.

REPLY)
While the authors have answered the concern about bi-layer (assuming the dotty region is the glass), I would still suggest to refrain the word "doped" here:
1. Doping typically refers to the case where the film is perfectly homogenous in nano-scale, with atomic substitution or addition of the dopant into the host crystal's structure
2. Once we have nano clusters of metals, as evidenced in the X-ray data (fig. 1), it becomes rather misleading to call this doping rather a mixture of Cu and ZnO_x (note that there have been studies that ZnO sputtering with Ar may also lead to oxygen-deficient ZnO film, hence the _x).

What I would suggest is to call this a ZnO-Cu film, and then go into the length of discussing how the Cu starts to precipitate out of the film in C2-C10 as evidenced by X-ray. The author can also claim that most likely the Cu acts as a dopant in C1-C2 because of the lack of XRD peak.

Also, please note that the figure numbers S1 and S2 are reffered to wrongly in the main body of the manuscript (P3L119, P5L141)

 

 

 

CLOSED

P2L66-69: description of the gas composition into the chamber during pre-sputtering and 1st stage of sputtering is not found.

Ans) We expressed in the manuscript, “The targets were pre-sputtered with argon plasma for 10 min”. The Ar flow was maintained at 4.55 sccm, and the substrate temperature was set to 150°C.

P2L77: description of the chamber is very important for reproducibility but missing. E.g. target-substrate distance, angles between substrate and target(s), direction of gas flow, gas flow rate, pumping condition (e.g. turbo / cryo pump, where?)

Ans) Vacuum chamber: 50 cm in diameter, target-substrate distance: less than 10 cm, tiled (~15 °) RF/magnetron gun, pumping: turbo-molecular pump. This information is newly added in the supplementary information.

P2L78: It would be most interesting to have C40 as another extreme point comparison (other than C0)

Ans) C40 (fully co-sputtered ZnO) indicated inhomogeneous ZnO films with much metallic Cu clusters. Thus, we set the control sample to be C0.

P3L92: clarification: do you mean hexagonal ZnO? i.e. is the lattice constant the same with ZnO too?

Ans) The lightly doped ZnO maintains a hexagonal crystal structure (Fig. 1). We modified the manuscript. Because peaks were shifted, the lattice constants changed depending on x.

P3L95: "Cu-related peaks might be formed from metallic Cu, CuO, and Cu2O" -- this doesn't make sense! Cu peak is simply an indication of the presence of Cu metal, most likely in micro-crystalline form

Ans) Following this comment, we deleted “Cu-related peaks might be formed from metallic Cu, CuO, and Cu2O.”

P3L99: Figures X-axis are in degree. If so, then need to specify X-ray source wavelength or type (e.g. Cu-k-alpha). Alternatively, we can report the X-axis in q.

Ans) We used the CuKa radiation. This was added in the Materials and Methods section.

P3L109: "broccoli-like crystal structure" : calling it a "crystal structure" (hexagonal, diamond, square, triclinic, etc.) is misleading. Calling it "nanostructure" is also misleading since the structure is clearly larger than 200 nm (see the scale in the figures). Morphology" is probably a better terminology. Check for examples https://doi.org/10.1016/j.tsf.2013.03.041.

Ans) We understand this concern. In P3L109, we expressed “broccoli-like clusters”.

P4L123: one of the key assumption of W-H analysis is that the film is at least quasi-homogenous within the sampling volume of the X-ray. However, because of the two-step deposition process, I would expect the film to be a Cu:ZnO / ZnO bilayer. This breaks the assumption of WH and therefore this analysis is invalid.

Answers and rebuttals)

We understand the concern of the reviewer and that’s why we used a uniform deformation model (UDM) for W-H analysis to obtain the strain and the size of the ZnO lattice instead of uniform stress-deformation model (UDSM). USDM requires exact material parameters such as Young’s Modulus, but UDM analysis can be accepted with the much simpler conditions if all XRD peaks are independent without overlapping (like the present analysis). We only choose ZnO peaks so the W-H uniform deformation could be accepted if background Cu signal is trivial in the peak area. As this reviewer commented, there could be depth-gradient in these films, but we believe those would become negligible after the thermal annealing process. Also, the W-H strain is valid as the ‘overall’ strain of all X-ray sampling volume, including the strain caused by the internal gradient. For this author, WH analysis is applicable to the system with an inhomogeneously strained crystalline case. In addition, please consider the WH formulae are approximate approach (for simple information, see http://pd.chem.ucl.ac.uk/pdnn/peaks/size.htm).

Author Response

Reviewer 1

Fig S2: Assuming that the dotty area is from the soda glass, I agree that there's no evidence of bi-layerness. However, the film thicknesses, esp. C10, is very different. I believe the authors haven't put this in the main body, although they claim that the thicknesses have been measured with a profilometer (and is needed for 4-point probe). I think this additional data will go well with Fig. 5.

Answer) We appreciate the comments.

The measured thicknesses by SEM and profilometer were different. We used data from the surface profilometer. In the case of C10, the difference was severe. This is because the thickness of the broccoli-like structure using SEM shows a maximum thickness of the irregular structure, whereas the profilometer uses mechanical force between the stylus and the surface. We newly added the thickness profiles of each sample in SI (Table S1). We added in L192 the following statement, “The absorption coefficient was calculated using the measured thickness of each sample shown in Table S1.”

REPLY)
While the authors have answered the concern about bi-layer (assuming the dotty region is the glass), I would still suggest to refrain the word "doped" here:

1. Doping typically refers to the case where the film is perfectly homogenous in nano-scale, with atomic substitution or addition of the dopant into the host crystal's structure
   2. Once we have nano clusters of metals, as evidenced in the X-ray data (fig. 1), it becomes rather misleading to call this doping rather a mixture of Cu and ZnO_x (note that there have been studies that ZnO sputtering with Ar may also lead to oxygen-deficient ZnO film, hence the _x).

What I would suggest is to call this a ZnO-Cu film, and then go into the length of discussing how the Cu starts to precipitate out of the film in C2-C10 as evidenced by X-ray. The author can also claim that most likely the Cu acts as a dopant in C1-C2 because of the lack of XRD peak.

Answer) We appreciate the comments and followed this advice in the revised manuscript.

We changed a previous comment as (L98):

“This means Cu atoms in the high-x films (C5 and C10) not only act as dopants but also form embedded Cu clusters,”

In addition, in L108, L112, L196,

we changed: “lightly doped films” to low-x films, and “highly doped ZnO films” to high-x films

Furthermore, in L194: “Undoped ZnO” to “Pristine ZnO”, and in L206: “Cu-doped ZnO” to “co-sputtered ZnO”

Also, please note that the figure numbers S1 and S2 are reffered to wrongly in the main body of the manuscript (P3L119, P5L141)

Answer) We’ve corrected those in revised manuscripts (L121, L144)

We appreciate again this reviewer for valuable comments and advice.

Reviewer 4 Report

The authors have complied with all points raised by this referee. The paper is improved and now ready for publication

Author Response

We appreciate valuable comments.