Sub-Terahertz Computer Generated Hologram with Two Image Planes

Round 1

Reviewer 1 Report

The paper presents experimental demonstration of the possibility of using 3d printing to generate transmissive computer generated hologram masks for Sub-terahertz light. Moreover, the authors developed a variant of the existing "ping-pong" hologram computation algorithm, adapted for optical setups not satisfying the paraxial approximation assumption, and for placement of the hologram in a plane different from the pupil plane of the system.

The paper is definitely interesting and deserving publication due to its novelty and its convincing results. Nonetheless, the presentation of the paper could be definitely improved to guarantee reproducibility of the results and a wider adoption of the reported method. In particular:

- The hologram computation algorithm could be described better. In particular, the two most interesting aspects of the computation (propagation through modified convolution method and off-axis propagation) are reported as simple references to previous papers. While this is sufficient to present the method, it would be very helpful for the reader if the authors added, even just as supplementary material, a more detailed description of the computation required, or even better some commented source code from their software.

- A very important aspect of the paper is the fabrication of the 3d printed phase mask. The quality of the hologram produced can be assumed to be strongly dependent on the quality of the printer itself. It would therefore be very useful if the authors specified the maker and model of the 3d printer they employed, together with more accurate specifications. In particular, while the authors report the nominal resolution of the printer, they did not specify the layer thickness, which is a quite important parameter for holograms fabrication.
Moreover, better details on the data processing between computation of the hologram and printing of the mask should be provided. The authors specify that the resolution of the printer is 117 micrometers, and the size of the hologram is 10 centimeters, which would imply the computed holograms should have a resolution of 855x855 pixels. Instead, the authors computed the mask on 8192x8192 pixels to obtain more accurate results in the simulations. The rescaling process between the two sizes could significantly affect the quality of the hologram, and should be better explained.

- The authors refer multiple times that the size of the produced images are wider than the size of the hologram. I can understand how this is important since it makes the system non-paraxial. However, it would be even more important to compare the image sizes with the theoretical maximum angular field of view of the printed holograms. This would be determined by the maximum gradient of the wavefront achievable with the printer, which i assume will be limited by the lateral resolution.

- Figures 5 and 6 should present scale bars.

Author Response

Dear Reviewer,

Thank you very much for your helpful remarks. We have taken into account all your suggested requests and comments. Below we copy remarks and answer them point by point – showing how they influenced the changes in the new version of the manuscript.

“The hologram computation algorithm could be described better. In particular, the two most interesting aspects of the computation (propagation through modified convolution method and off-axis propagation) are reported as simple references to previous papers. While this is sufficient to present the method, it would be very helpful for the reader if the authors added, even just as supplementary material, a more detailed description of the computation required, or even better some commented source code from their software.”

Response: Modified convolution method and off-axis propagation are thoroughly described in [1]  and [2] correspondingly (both sources were already referenced in article before review). As for the supplementary material we agree that providing the clear code could be helpful for reader. However, code used in LightSword software may be not readable for someone who is not familiar with its syntax and little quirks. This is why we decided to provide pseudocode in form of separate table and reference it properly in article.

“A very important aspect of the paper is the fabrication of the 3D printed phase mask. The quality of the hologram produced can be assumed to be strongly dependent on the quality of the printer itself. It would therefore be very useful if the authors specified the maker and model of the 3d printer they employed, together with more accurate specifications. In particular, while the authors report the nominal resolution of the printer, they did not specify the layer thickness, which is a quite important parameter for holograms fabrication.”

Response: In lines 132-134 we provided information about model and maker of 3D printer. Also we enumerated most important parameters from specification such as layer thickness and nominal accuracy of printer.

“Moreover, better details on the data processing between computation of the hologram and printing of the mask should be provided. The authors specify that the resolution of the printer is 117 micrometers, and the size of the hologram is 10 centimeters, which would imply the computed holograms should have a resolution of 855x855 pixels. Instead, the authors computed the mask on 8192x8192 pixels to obtain more accurate results in the simulations. The rescaling process between the two sizes could significantly affect the quality of the hologram, and should be better explained.”

Response: We haven’t used any rescaling. Size of the hologram on calculation matrix had 855x855 px (we used rectangular mask of this size to remove all unnecessary data). In order not to misguide reader we added comment about that in lines 130-131.

“The authors refer multiple times that the size of the produced images are wider than the size of the hologram. I can understand how this is important since it makes the system non-paraxial. However, it would be even more important to compare the image sizes with the theoretical maximum angular field of view of the printed holograms. This would be determined by the maximum gradient of the wavefront achievable with the printer, which i assume will be limited by the lateral resolution.”

Response: The resolution of 3D printer is 39 micrometers (size of voxel). We assumed the resolution of 117 um (3 times bigger voxel) due to the fact that in printed hologram there are many places that have relatively large changes of height – from 0 to almost 3 mm. Nevertheless, such resolution is dictated by the possibilities of printing machine to assure the best quality of manufacturing. In such case the wavefront gradient can cover the angles up to +/- 45 degrees. The necessity of optical design does not require such resolution. The determination of possible field of view could be related to the size of the structure (100 mm) and the distance from the structure to the image plane (150 mm). In case of described hologram it is hard to relate to the field of view, especially that it has two image planes, that is why we gave all distances and dimensions of structure and images. In our opinion the simplest and most unequivocal way to determine the field of view related to angles at which radiation can be bent is the comparison to diffraction grating.

If we would produce blazed grating with period of 2 mm - the first order of diffraction would have half-angle of about 60 degrees. [Our team during separate experiments used kinoform (blazed) gratings with period 2 mm made with PA12 for the same wavelength. It’s first diffraction order was obtained at around 60 degrees angle from normal to the surface of grating.  Our team also done some experiments with antireflective pyramids with 1 mm base and 2 mm of height – so the manufacturing of such gratings was experimentally verified.]

This can be stated as maximal possible FOV angle obtainable with diffractive element.

“Figures 5 and 6 should present scale bars.”

Response: We provided figures 5 and 6 with bars.

Sources:

[1] Maciej Sypek. Light propagation in the fresnel region. new numerical approach. Optics communications 1995, 116(1-3), 43–48

 [2] M. Sypek, C. Prokopowicz, M. Górecki, "Image multiplying and high frequency oscillations effects in the Fresnel region light propagation simulation", Optical Engineering, vol 42, no 11, p.  3158-3164 , (2003) – IF2014 0,926

Reviewer 2 Report

The authors present the design and simulation of a single computer generated Fresnel hologram for sub-terahertz band. I found the results of the manuscript interesting and would like to recommend the paper for publication upon addressing the following major and minor issues.

1-      The authors do not discuss pros and cons of the proposed system compared to the prior art. For instance, the authors have made the statement that the hologram area is smaller than the operation wavelength in their designed system. However, they do not compare their systems with similar reports to see whether their approach is indeed superior to the current state of the art or not.

2-      The authors do not discuss the available bandwidth of their computing system. Considering the seemingly strange chosen operation frequency (0.17 THz), it looks like to me that the system under study is not very wide band. If it is the case, then discussion should be added into the manuscript in this regard, citing the prior arts about reconfigurable computing terahertz systems (see for instance: Optics letters 42.10 (2017): 1954-1957) which can, in principle, go beyond the bandwidth limitation of narrow-band systems.

3-      While the manuscript is well written, I am still finding some minor typo in it. For instance, a space should be included between the numbers and the units. Like 60 mm in line 59, etc. Furthermore, in the abstract, the authors would better avoid using past tenses (line 7 and 8). Altogether, I recommend the authors proofread everything one more time.

4-      The introduction does not go through the recent advances in the field of terahertz technology well. Not only has this frequency range found to establish a perfect platform for computing applications over the past decade, but also it has been of great importance for a large variety of other technology oriented applications including sensing   (Applied Physics Letters 77.24 (2000): 4049-4051, IEEE Sensors Journal 16.11 (2016): 4338-4344, Nanoscale 7.29 (2015): 12682-12688, Optics Communications 371 (2016): 9-14), spectroscopy (Laser & Photonics Reviews 5.1 (2011): 124-166, Materials today 11.3 (2008): 18-26.), and so forth.     

5-      In the discussion section, the authors should discuss the possible future direction of the work, rather than discussing the methods they have used to generate the results (line 113-116).     

6-      What is the scale bar in Fig. 6? I am guessing that the authors have used normalized values in this figure. As far as I know, the conversion efficiency of most of terahertz sources and detectors are very low. In fact, lots of efforts are being invested to enhance the conversion efficiency of terahertz sources and detectors (for instance: Nature communications 4 (2013): 1622, Journal of Nanophotonics 10.3 (2016): 036005, etc). I suggest the authors to provide the scale bar in the figure and make some statement in this regard.

7-      The authors make the argument that their proposed system can be potentially used in MIMO telecom application. This statement, however, is neither supported by a scientific argument or, at least, some proper references.

8-      The bibliography does not look complete to me. Citing only 25 prior works is not sufficient for a full scientific article.       

Author Response

Dear Reviewer,

Thank you very much for your helpful remarks. We have taken into account all your suggested requests and comments. Below we copy remarks and answer them point by point – showing how they influenced the changes in the new version of the manuscript.

“The authors do not discuss pros and cons of the proposed system compared to the prior art. For instance, the authors have made the statement that the hologram area is smaller than the operation wavelength in their designed system. However, they do not compare their systems with similar reports to see whether their approach is indeed superior to the current state of the art or not.”

Response: Produced hologram is not “smaller than the operation wavelength” but its size is relatively small comparing with traditional hologram (designed for visible spectrum of radiation). We agree that this statement was not clear. For this reason we have made amendments in abstract and introduction  (lines 5-6,  56-60).  Pros and Cons  were also added (lines 39-48).

“The authors do not discuss the available bandwidth of their computing system. Considering the seemingly strange chosen operation frequency (0.17 THz), it looks like to me that the system under study is not very wide band. If it is the case, then discussion should be added into the manuscript in this regard, citing the prior arts about reconfigurable computing terahertz systems (see for instance: Optics letters 42.10 (2017): 1954-1957) which can, in principle, go beyond the bandwidth limitation of narrow-band systems.”

Response: There is nothing strange about selected wavelength/frequency. As it was already pointed in section 3. The design wavelength was chosen to be matched to the source frequency chosen for experimental part, together with detector and good optical properties of hologram material at this frequency. It is true that system is not wide band because it wasn’t meant to be. It is actually much easier to obtain image from hologram for single frequency than for multiple frequencies. In this paper we concentrated on creating multiple images in multiple planes rather than to create images for many frequencies. However, broadband working can be achieved using kinoforms of higher order [Optics Express, 22(3), 3137-3144. (2014)].

“While the manuscript is well written, I am still finding some minor typo in it. For instance, a space should be included between the numbers and the units. Like 60 mm in line 59, etc. Furthermore, in the abstract, the authors would better avoid using past tenses (line 7 and 8). Altogether, I recommend the authors proofread everything one more time.”

Response: We checked the text once more and tried to eliminate all mistakes.

“The introduction does not go through the recent advances in the field of terahertz technology well. Not only has this frequency range found to establish a perfect platform for computing applications over the past decade, but also it has been of great importance for a large variety of other technology oriented applications including sensing   (Applied Physics Letters 77.24 (2000): 4049-4051, IEEE Sensors Journal 16.11 (2016): 4338-4344, Nanoscale 7.29 (2015): 12682-12688, Optics Communications 371 (2016): 9-14), spectroscopy (Laser & Photonics Reviews 5.1 (2011): 124-166, Materials today 11.3 (2008): 18-26.), and so forth.”

Response: We added some review articles that are strictly related to advances in the field of terahertz technology [references number 11,12,13 and 14 in revised article]

 “In the discussion section, the authors should discuss the possible future direction of the work, rather than discussing the methods they have used to generate the results (line 113-116).”

Response: This is valid comment. First paragraph of section 4. Discussion was rephrased a little bit and placed in section 3. Methods in lines 113-117.

“What is the scale bar in Fig. 6? I am guessing that the authors have used normalized values in this figure. As far as I know, the conversion efficiency of most of terahertz sources and detectors are very low. In fact, lots of efforts are being invested to enhance the conversion efficiency of terahertz sources and detectors (for instance: Nature communications 4 (2013): 1622, Journal of Nanophotonics 10.3 (2016): 036005, etc). I suggest the authors to provide the scale bar in the figure and make some statement in this regard.”

Response: We added scale bars in Fig. 6 and the information about using normalized values in all diagrams.

 “The authors make the argument that their proposed system can be potentially used in MIMO telecom application. This statement, however, is neither supported by a scientific argument or, at least, some proper references.”

Response: Proper argument was provided by adding sentences in lines 33-37.

Round 2

Reviewer 1 Report

The authors have done a good job addressing my minor concerns about the paper, so i can recommend it to be published in its current form.

However, in my personal opinion, the considerations the authors made in reply to my last question regarding field of view, are very interesting and well discussed and deserve to be somehow included in the paper itself. Nonetheless, i leave the choice on whether to include them in the paper or not to the authors.

Reviewer 2 Report

I believe the authors have properly addressed my previous concerns, hence I recommend the publication of the work at this stage.