Compact Optical System Based on Scatterometry for Off-Line and Real-Time Monitoring of Surface Micropatterning Processes

Round 1

Reviewer 1 Report

The authors analyze the diffraction pattern obtained by the reflection of the micropatterning surface. I am concerned about associating two-dimensional distribution failures with only following order separation. I believe an extra measurement process needs to be done and not only use the standard grating equation and observe orders intensity on a CCD camera. I have more concerns that could help explain their proposal.

1(1)      Please add a diagram showing the explanation in the first and second paragraphs of section 2.

2(2)      Figure 1 and table 1 explanations need to be addressed appropriately. Maybe separation into different sections could help with their description.

3(3)      Please add a monitoring system diagram on page 4, second paragraph.

Regards

The Reviewer

Author Response

We would like to acknowledge the Reviewer for her/his comments that allowed us to improve the quality of our manuscript. Here below, we include a point by point response of the review.

Query 1: I am concerned about associating two-dimensional distribution failures with only following order separation. I believe an extra measurement process needs to be done and not only use the standard grating equation and observe orders intensity on a CCD camera.

Answer 1: We agree that an exhaustive characterization of the surface morphology cannot be accomplished by using methods based solely on scatterometry. However, we prove in this study that this optical technique can be used in industrial environments for monitoring laser texturing processes by just measuring the positions of the diffraction peaks and their relative intensities. To that end, we suggest in the final part of the section 3.2 that according to the real process it is important to identify which process variables (e.g., period, focus position, patterning strategy, etc.) would be subject to fluctuations and should be monitored, while assuming that the rest of the variables can be considered fixed. Then, from the scatterometry measurements, the oscillations of the process variables could be quantified and the overall surface texture quality could be assessed. We also agree that adding sensors to complement our proposed monitoring method could enhance the accuracy of the surface characterization and could allow the user to retrieve more information from the process. As discussed in section 3.2 including a powermeter to monitor the laser fluence might be beneficial, as our method was not successful to detect fine variations in the fluence. Analyzing the impact and synergy of adding extra sensors (e.g., acoustic, spectroscopic, microscopy, etc.) could be interesting for further research towards implementation in real laser texturing stations, however, this great deal of work is beyond the scope of our study.

Q2: I have more concerns that could help explain their proposal. Please add a diagram showing the explanation in the first and second paragraphs of section 2.

A2: In the new version, schematics of both the DLIP setup and optical system are added for the reader´s convenience.

Q3: Figure 1 and table 1 explanations need to be addressed appropriately. Maybe separation into different sections could help with their description.

A3: In the revised version of the manuscript, section 2 now contains bullet points describing each set of experiments for a simpler understanding of the proposed methodology.

Q4: Please add a monitoring system diagram on page 4, second paragraph.

A4: As explained in Q2, a diagram showing the main components of the monitoring system is now included in the manuscript.

Reviewer 2 Report

The authors have demonstrated a compact optical system for real-time and offline monitoring of surface micropatterning processes. The technology provided them with precise estimates of the spatial period, focus shift identification, and texturing. The manuscript is very well written. Please find my comments below,

1.      The authors were trying to provide detection limits for their detection system, and they have provided exciting results. However, please give an experimental setup of the real-time and offline monitoring system for the reader's convenience.

2.      Does your optical system differentiate between samples B and C? If yes, how do we know which parameter (focus or fluence) to change to get the required pattern?

3.      Is there an optimum position for your CCD in the detecting system to accurately anticipate the spatial period with minimum error?

4.      Figure 2f is -300 um according to line 206.

5.      Does your real-time monitoring system impose any restrictions on the repetition rate of your laser?

6.      Please check the sentence in line 329

7.      In Figure 4, topographical spatial period and monitoring system spatial period error. Can you describe the discrepancies in the error values estimated from your monitoring system with the spatial period obtained using the topographical image?

Comments for author File: Comments.pdf

Author Response

We would like to acknowledge the Reviewer for her/his comments that allowed us to improve the quality of our manuscript. Here below, we include a point by point response of the review.

Reviewer´s comment:

The authors have demonstrated a compact optical system for real-time and offline monitoring of surface micropatterning processes. The technology provided them with precise estimates of the spatial period, focus shift identification, and texturing. The manuscript is very well written. Please find my comments below,

Query 1: The authors were trying to provide detection limits for their detection system, and they have provided exciting results. However, please give an experimental setup of the real-time and offline monitoring system for the reader's convenience.

Answer 1: A CAD drawing of the optical system, including a short description of the main components, is now shown in Figure 1. As explained in section 2, the same experimental setup was used for the offline and real-time measurements, whereas the code that converts the CCD image into the different signals (period and order intensities) was adapted to allow the real-time mode.

Q2: Does your optical system differentiate between samples B and C? If yes, how do we know which parameter (focus or fluence) to change to get the required pattern?

A2: As described in the Results and Discussion section, the optical system was not successful to detect fine variations of the topography that were produced with different fluence levels. On the contrary, by monitoring the spatial period, i.e. tracking the position of the diffraction peaks, it was possible to detect fluctuations in the focus position due to the intrinsic shift of the period within the interference volume, when the sample is placed outside the optimum focus position (cf. Equation 2 in the original manuscript). If a proper control of the overlapping angle between the beams can be guaranteed, then the measured variations of the spatial period can be directly ascribed to shifts in the focus positions. Therefore, the proposed method is able to differentiate between samples B and C. These above-mentioned concepts were discussed in the last part of section 3.2 where different hypothetical scenarios are proposed. Furthermore, this paragraph was modified for a better explanation.

Q3: Is there an optimum position for your CCD in the detecting system to accurately anticipate the spatial period with minimum error?

Yes, the parameters and positions of all the optical components were first simulated using ZEMAX software to optimize the formation in the CCD sensor of the image corresponding to the diffraction pattern. Then these optical components were finely adjusted and aligned in the experimental setup. The description of the design and simulation model of this system can be found in the Reference 31.

Q4: Figure 2f is -300 um according to line 206.

A4: The caption of Figure 2 was corrected accordingly.

Q5: Does your real-time monitoring system impose any restrictions on the repetition rate of your laser?

A5: To avoid confusions, we would like to clarify that the real-time measurements were taken after the laser texturing process. In other words, so far, the monitoring system was not tested in on-line mode or during the laser texturing. Therefore, there are in principle no restrictions on the repetition rate of the laser for an accurate monitoring of the produced topographies. It is well-known that changing the repetition rate of the laser source influences the resulting surface morphology on the sample surface not only using the DLIP method but also other laser-texturing processes. Further studies could be conducted to assess whether the proposed scatterometry-based method is useful to monitor fluctuations in the repetition rate. On the other hand, the laser diode used in the monitoring system works in continuous mode and thus only its wavelength and intensity (which is controlled with the polarization optics) have an influence on the captured images.

Q6: Please check the sentence in line 329

A6: We modified the sentence accordingly. Now it reads:

As described in Table 1, samples D and E were fabricated using a patterning strategy that intentionally leaves line- and square-shaped untreated areas on the surface.

Q7: In Figure 4, topographical spatial period and monitoring system spatial period error. Can you describe the discrepancies in the error values estimated from your monitoring system with the spatial period obtained using the topographical image?

A7: As observed in the original Figures 2 and 3, the diffraction peaks captured in the CCD images are not well-defined Gaussian peaks, but rather a diffuse cloud of illuminated pixels. Furthermore, these clouds are not round nor symmetric, due to the optical aberrations from the lenses and the projected orders onto the flat CCD sensor. Stray light, internal reflections and electrical noise can be further sources of errors that distort the shape of the measured peaks. As a consequence, the captured peaks do not have a uniform and systematic shape, but they are rather random. As the position of their centroid determines the texture spatial period, there is in all cases an associated error which according to our studies was below 2.5% relative to the spatial period measured with microscopy methods. The main text (section 3.2, lines 273-277) was modified to include this comment.