Development of Pulsed TEM Equipped with Nitride Semiconductor Photocathode for High-Speed Observation and Material Nanofabrication

Round 1

Reviewer 1 Report

The paper entitled “Development of pulse TEM equipped with nitride semiconductor photocathode for high speed observation and materials nanofabrication” deals with the sophisticated pulsed electron sources. The experimental data are clear, and the manuscript is well written and is interesting for the reader of Quantum Beam Science. Some minor corrections/additions are needed before acceptance for publication.

(1) As mentioned in Abstract, the pulsed source has been developed for materials nanofabrication (electron lithography?) as well as high-speed TEM observations. The authors have presented its performance only for TEM observations. There are no comments on its performance for materials nanofabrication in the present manuscript. It is better to add some comments if possible.

(2) A blue solid line (fitted by log-normal function?) is drawn in fig. 4, but no explanation in the caption.

(3) A Greek character “μ” may be missing in line 119. 800 s  should be 800 μs.

(4) A careless mistake in line 79. O² should be replaced with O₂.

Author Response

Thank you very much for the helpful suggestions from the reviewer. I am sending herewith our revised manuscript. Below please find our consideration and modification to the reviewer’s suggestions.

(1) As mentioned in Abstract, the pulsed source has been developed for materials nanofabrication (electron lithography?) as well as high-speed TEM observations. The authors have presented its performance only for TEM observations. There are no comments on its performance for materials nanofabrication in the present manuscript. It is better to add some comments if possible.

We added the following description in the session of Performance characterization in the manuscript. 

The performance of pulsed TEM mentioned above can also be applied to materials nanofabrication. When a semiconductor photocathode is utilized as the electron gun of an electron beam lithography and a semiconductor inspection device, the stability of the electron beam is extremely important. The required performance is a small current fluctuation, which is a standard deviation from the average current during a constant current generation for a long time. In particular, if the current fluctuation over a long time is large, it affects the performance, and the same accuracy as making the TEM image contrast negligibly small is required. It is considered that the fact that sufficient performance can be obtained by applying a semiconductor photocathode to a pulsed TEM is consistent with satisfying the necessary and sufficient conditions for utilizing a semiconductor photocathode in an electron beam lithography and a semiconductor inspection device.

Recently, in the semiconductor manufacturing process, in order to solve the problems of low inspection throughput and drawing throughput, a new demand for higher brightness and a multi-electron beam source in which individual source has high brightness and uniformity is searched. In order to obtain high brightness, an electron source that generates an electron beam with a large current and low emittance characteristics is required. It will be suggested that semiconductor photocathodes that enable generation of high current, low emittance, multi, and high speed pulse electron beam are also useful as next generation electron beam sources which respond to new demands of materials nanofabrication.

(2) A blue solid line (fitted by log-normal function?) is drawn in fig. 4, but no explanation in the caption.

We added the following description in the manuscript and caption.

In order to evaluate the performance of the apparatus, the exposure time required for pulse imaging was measured. Fig. 4 shows a bright-field image which was taken by conventional TEM with thermal emitter, and size dispersion of Au nanoparticles on an amorphous carbon film used for pulse imaging as a standard sample. The size distribution was fitted by lognormal distribution function, and an average particle size is approximately 13 nm. 

(3) A Greek character “μ” may be missing in line 119. 800 s  should be 800 μs.

We have corrected it as follows. 

800 μs

(4) A careless mistake in line 79. O² should be replaced with O₂.

We have corrected it as follows. 

O₂

Reviewer 2 Report

There are no changes or modification made from previous version of the manuscript. 

Author Response

Thank you so much for your comment. 

Reviewer 3 Report

41-43: I recommend reformulating the period which is too long.

79: 0^2, are you sure? Fig. 4 is unclear whether it is acquired by pulsed laser imaging.

In Figure 5 what would be the temporal distance between the laser pulses and how the laser frequency, in addition to the number, could influence the image quality?

line 119: "800 s", is it correct? Better indicate the exposure time dependence.

I strongly suggest to explain better in the text the motivation for which the NEA process is realized and what is the role of the Cs vapor deposition source for NEA surface processing.

Author Response

Thank you very much for the helpful suggestions from the reviewer. I am sending herewith our revised manuscript. Below please find our consideration and modification to the reviewer’s suggestions.

(1) 41-43: I recommend reformulating the period which is too long.

We have separated the sentences by year as follows.　

In 2007 after that, at Lawrence Livermore National Laboratory in the United States, spatial resolution of ~10 nm and temporal resolution of ~1 ns were achieved. [3] In 2008, at California Institute of Technology in the United States, temporal resolution was improved to ~10 ps in the same spatial resolution of ~10 nm. [4] Recently in 2016, at CNRS National Laboratory in France, atomic resolution of ~0.2 nm and temporal resolution of ~370 fs were succeeded to achieve. [5]

(2) 79: 0^2, are you sure? Fig. 4 is unclear whether it is acquired by pulsed laser imaging. 

We have corrected it as follows. 

O₂

In order to evaluate the performance of the apparatus, the exposure time required for pulse imaging was measured. Fig. 4 shows a bright-field image which was taken by conventional TEM with thermal emitter, and size dispersion of Au nanoparticles on an amorphous carbon film used for pulse imaging as a standard sample. The size distribution was fitted by lognormal distribution function, and an average particle size is approximately 13 nm.

(3) In Figure 5 what would be the temporal distance between the laser pulses and how the laser frequency, in addition to the number, could influence the image quality?

As described in the experimental procedure, The imaging conditions were a laser pulse width of 100 μs and a repetition frequency of 5 kHz. Therefore, the interval of laser pulses is 100μs. 

We have added the following description. 

A factor in which laser frequency affects image quality is amount of the specimen drift within the total exposure time. If the total exposure time is shortened, the effect of the specimen drift will be reduced, which will lead to an improvement in image quality. In Fig. 5, even in total 2048 pulses, the total exposure time is about 0.2 sec, and even if the interval of 100 μs is included, it is 0.4 sec, and the effect of the specimen drift is very small. 

(4) line 119: "800 s", is it correct? Better indicate the exposure time dependence.

We have corrected it as follows. 

800 μs

(5) I strongly suggest to explain better in the text the motivation for which the NEA process is realized and what is the role of the Cs vapor deposition source for NEA surface processing. 

We have added the following description to the introduction. 

The advantage of using the NEA semiconductor surface is to reduce  the work function. The role of Cs in NEA surface preparation has not been clarified in detail. The NEA treatment method is to form a dipole layer of Cs with a positive charge on the semiconductor surface by repeating Cs vapor deposition and oxygen adsorption at a substrate temperature of about 550 ° C. This dipole layer is thought to reduce the work function. Lowering of the work function to the energy near the CBM on the semiconductor surface leads that the electrons with an extremely narrow energy width excited just above the CBM are emitted from the surface by the tunnel effect. As a result, it is possible to obtain an electron beam in which the energy dispersion is extremely small. Even if the work function is lowered on the NEA metal surface, the energy dispersion of electrons excited to the unoccupied level cannot be as small as that of a semiconductor with a bandgap. NEA treatment on the semiconductor surface is extremely effective from the viewpoint of electron energy dispersion. 

Round 2

Reviewer 2 Report

Thank you for modifying the manuscript. I don't have any additional comments/questions to authors.

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

The authors propose a new photocathode material for the pulsed electron-source and charaterize the source with some imaging experiments.

The readability of the paper in it's current form is extremely low. I would urge the authors to put some more effort on improving the readability of the article. In it's present form, the significance and scientific merits cannot be judged. Also, the paper lacks, proper comparison of the present work with the previous similar works. 

The paper also lacks the quantitative improvement the new photocathode material brings. I suggest the authors to add this quantitative description.

Reviewer 2 Report

1. Please provide some numerical findings in the abstract of the claimed improvement.
2. shows numerical annotation (1 to 10). Please explain or remove those. It would be best to redraw the figure with proper reference.
3. Page 2; line 57-58: “However, there are not many examples of …” : The claim is not proper. I would suggest removing the claim.
4. Page 2; Section 2: Are there any changes in operational specification to the equipment (Hitachi HF-2000 TEM) due to the modification? Please, describe if such changes exist.  
5. Page 4; Fig 4: The counts are statistically very low and not significant to fit a curve. Please improve the counts.
6. Page 5; line 108-109: “It is evident that ..” Au nano-particles are not visible after 4- 8 pulse. It looks like, the flux is too low to have a meaningful observation in just 4 -8 pulse.
7. Page 56; Fig 6: presented S/N doesn’t translate to the quality of the image in terms of spatial resolution and contrast. Modular transfer function (MTF) would be a standard demonstration to back the claim. Again, I suspect that, the equipment can achieve high timing resolution on the cost of low flux which may not be suitable for high spatial resolution measurements and MTF would be a good measure for that. In addition, please compare the performance measurement (S/N, MTF etc.,) with a standard TEM.