Multiscale Approach of Investigating the Density of Simulated Fuel for a Zero Power Reactor

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The study investigated potential fuels for Molten Salt Reactors, which is an important research topic given the need of carbon neutral power generation to slow climate change.

The experimental design and the characterization methods are good, but the interpretation of some of the data and the text of the manuscript needs major changes and significant improvement.

Specifically:

-            The evaluation of the XRD data has multiple fundamental flaws and important details are not explained:

o   How exactly were the wa, wb weight fractions determined from the XRD patterns?

o   Eq. 7-10 work only if the material has no crystallographic texture. A texture measurement should be performed to confirm that, or if there is texture, it is needed to discuss the uncertainties in the results caused by its presence.

o   A more detailed explanation is needed here regarding the observed peak shifts. Only vague guesses are mentioned for the possible reasons of the peak shifts, no specific hypothesis is given.

o   Crystallinity cannot be determined using the method described in the paper. The description is not very clear, but if the authors mean that the ratio of the total peak areas with respect to the total area of the pattern (including the background under the peaks/under the baseline) is considered as the crystallinity ratio, that is not correct. A polycrystalline sample even with 100% crystallinity (i.e., having no amorphous volume fraction) can still have a significant background (intensity under the baseline) due to various effects: diffuse scattering from point defects, grain boundaries and dislocation cores, thermal diffuse scattering, background caused by X-ray fluorescence (e.g., if an element in the sample has an absorption edge value in keV just below the X-ray energy and the XRD instrument doesn't filter out the fluorescence properly), etc.

o   The Scherrer equation is only valid for peak at low 2theta values, only for the first 2-3 peaks. For peaks at higher angles strain broadening has a significant contribution which invalidates the Scherrer equation. Thus the grain size values in Table 4 for peaks above 110 (or 101 if one allows more uncertainty) cannot be considered reliable.

o   Before using the Scherrer equation, the measured FWHM values must be corrected for instrumental broadening. It is not mentioned if this crucial correction was done. It should also be added how the instrumental broadening was measured and how exactly the correction was done. Not considering the instrumental broadening invalidates any results from the Scherrer equation even for low-angle reflections.

o   As a suggestion, for the density calculations based on XRD and crystallography, Rietveld refinement of the patterns would provide a much more accurate and more robust result than the calculations presented here.

Comments on the Quality of English Language

The text of the manuscript has to be significantly improved. Many sentences are not clear, some are incomplete, and there are a few typos left in the text as well. Some (not a comprehensive list) of unclear/incomplete sentences can be found in lines 42, 139. At line 112 the values of the inner and outer diameters are not given.

Author Response

Response to Reviewers in ‘Blue’

Comments and Suggestions for Authors:

The study investigated potential fuels for Molten Salt Reactors, which is an important research topic given the need of carbon neutral power generation to slow climate change.

The experimental design and the characterization methods are good, but the interpretation of some of the data and the text of the manuscript needs major changes and significant improvement.

Thank you for the comments. Addressing the comments below.

Specifically:

-            The evaluation of the XRD data has multiple fundamental flaws and important details are not explained:

o   How exactly were the wa, wb weight fractions determined from the XRD patterns?

Table 5 shows the ratio of pellet components according to the mol%. Equation 7 has been used to calculate the one by keeping the other constant.

o   Eq. 7-10 work only if the material has no crystallographic texture. A texture measurement should be performed to confirm that, or if there is texture, it is needed to discuss the uncertainties in the results caused by its presence.

The set of equations were used to quantify the phases and density measurement. The quantification involves the intensity of the peaks which also address the texture of the sample.

o   A more detailed explanation is needed here regarding the observed peak shifts. Only vague guesses are mentioned for the possible reasons of the peak shifts, no specific hypothesis is given.

Position of the peaks in XRD scan are affected by the purity of the sample. Peaks in the scan of the mixture sample has shifted from their positions in the scans of the pure sample as shown in the figure 10.

o   Crystallinity cannot be determined using the method described in the paper. The description is not very clear, but if the authors mean that the ratio of the total peak areas with respect to the total area of the pattern (including the background under the peaks/under the baseline) is considered as the crystallinity ratio, that is not correct. A polycrystalline sample even with 100% crystallinity (i.e., having no amorphous volume fraction) can still have a significant background (intensity under the baseline) due to various effects: diffuse scattering from point defects, grain boundaries and dislocation cores, thermal diffuse scattering, background caused by X-ray fluorescence (e.g., if an element in the sample has an absorption edge value in keV just below the X-ray energy and the XRD instrument doesn't filter out the fluorescence properly), etc.

o   The Scherrer equation is only valid for peak at low 2theta values, only for the first 2-3 peaks. For peaks at higher angles strain broadening has a significant contribution which invalidates the Scherrer equation. Thus the grain size values in Table 4 for peaks above 110 (or 101 if one allows more uncertainty) cannot be considered reliable. Good point … Then Table 4 has reduced for CeCl3.

o   Before using the Scherrer equation, the measured FWHM values must be corrected for instrumental broadening. It is not mentioned if this crucial correction was done. It should also be added how the instrumental broadening was measured and how exactly the correction was done. Not considering the instrumental broadening invalidates any results from the Scherrer equation even for low-angle reflections.

The software has processed the data and FWHM values have obtained to be used in the Scherrer equation. So, the crystallite size measurement involves the peaks and FWHM of the respective component.

o   As a suggestion, for the density calculations based on XRD and crystallography, Rietveld refinement of the patterns would provide a much more accurate and more robust result than the calculations presented here.

Thank you for the suggestion. Rietveld refinement was avoided to reduce the chance of overseeing the information in the obtained scan.

Comments on the Quality of English Language:

The text of the manuscript has to be significantly improved. Many sentences are not clear, some are incomplete, and there are a few typos left in the text as well. Some (not a comprehensive list) of unclear/incomplete sentences can be found in lines

42 - Heat transfer Thermodynamic properties of molten salts allow the MSRs to work at near atmospheric pressure in the reactor primary loop,

139. Calculations of separation (pitch) effects are carried out in Section 3.1

At line 112 the values of the inner and outer diameters are not given - Calculations for 3.5 mm pellets radius (see results in Section 3.2) gives relative density results presented in Fig 3. Clads in LWR typically have 0.1 mm thickness in ZPSR they could be slimmer .

Line 222 - they can form a quadratic or hexagonal array.

Thank you for the comments, all addressed. Native English speaker in the authors team has carefully checked the manuscript again.

Reviewer 2 Report

Comments and Suggestions for Authors

Please see the attached file.

Comments for author File: Comments.pdf

Comments on the Quality of English Language

There are numerous mistakes and omissions in the text. I've flagged some.

Author Response

Response to Reviewers in ‘Blue’

The authors have characterized pellets of CeCl3-NaCl using a variety of techniques. Preparation of a solid-mixture of powders as a surrogate fuel is a challenge. The authors have attempted to describe ways of determining the density/porosity of the fuel. The rationale for the measurement of the solid-state densities is not clear as the argument seems to hinge on fueling a ZPR. However, there is verbiage in the introduction on molten salts, which doesn’t appear to be relevant. Also, the author’s presentation of the different density measurements and the deviations does not present a conclusion as to which is the most useful for a reactor performance analysis. Finally, why was a BET analysis not done, seeing as knowing porosity is crucial for determining/validating measured densities?

Thanks for understanding the challenges. ZPR is type of MSR and will preferably work close to the room temperature so that is why density of solid mixture is very important. Different types of densities were presented to differentiate theoretical and practical measurements. These techniques are used to measure the density by addressing the pores present in the mixture to understand true density. BET analysis was not performed due to lack of facility and did not outsource keeping in view the hygroscopic nature of salts, moisture (during transportation) affects the specific surface area and pore volume. BET concerns sorption isotherms and provide first specific surfaces data and in porous object the open porosity is the only part of the pores that are detected. In this case however, both open and close porosity must be considered for the neutronic, consequently BET is not useful here.

General comments.

The manuscript has many grammatical errors and needs a technical edit.

Typographical concerns:

Page 2 line 5. Reduced from what? Individual salts have higher melting points, but different salts mixture have reduced melting points than their individual self as usually presented in the phase diagrams.

Page 2 line 8. “Considering…” is not a sentence. Addressed

Page 2 line 14. Rewrite the sentence “it acts as a coolant…” because logically the sentence states that the fuel is being removed, which it isnt’t. Addressed

Page 2 line 34. Run on sentence. Addressed

Technical comments:

Page 2. Explain why a ZPR is important when the concept is first introduced. How would the system go from room temperature to a molten state, several hundred degrees, in an accident? Addressed

Page 2 line 33. A reference is needed that explains how physical properties change with burnup, especially in the case of a ZPR. Sorry this may be puzzling, the fuel composition is not significant during ZPR operation.

How does it follow that high purity sample preparation is needed when the samples from a running reactor will have many contaminants? Good point, but the contaminants at beginning of life are different from the fission products generated during burnup.

Page 2. What is the TRL of an MSR?

The Technology Readiness Level e.g. Carmack  et al  (2017) should be discuss in the frame of the renaissance of the MSRs. After success at ORNL in the 60ies (thermal and breeder MSR) there is a certain interest today.

Line 166 – Write out the Scherrer equation and explain the important terms. Addressed

There is a great deal of reliance on XRD as a method of density measurement. How do the authors

account for the depth profile that will be present in pressed material? There is no indication that the

pellets were homogeneous – slicing and analysis along the z-direction were not presented.

The prepared sample is very thin 0.71mm to minimize the residual stresses. At the same time, the techniques will be used for the radioactive materials which will be even lesser in the thickness (to use least possible quantity of radioactive materials). Thinner samples provide more accurate profile and slicing technique cannot be supported as the materials must be fully contained to avoid contact with air (only handled inside the glovebox). Pellets are semi-homogenous due to the different nature and particle sizes of the salts.  

Line 187. Would the authors explain how they get 5 nm resolution from their SEM? How did they sample the surface structure using ImageJ? Did they use a canned algorithm? If so, what was the assumption on the geometry of the pores? Where are the nm-scale data presented? Figure 4 shows 200 micron images and Figure 5 shows pore sizes down to about 0.5 micron. SEM is a surface analysis technique and does not give any information on subsurface structure.

Instrument can achieve resolution upto 5nm. For surface structuring using ImageJ program with following steps; first micrographs were calibrated according to the scale, adjusting threshold, set measurements and then analyse particles. Micrographs at nm scale were not needed as particle size of salts are in um range. Some of the pores measured through ImageJ software are below micron range, Figure 5 showing them. Yes, SEM is surface analysis that is why Figure 8 shows the 3D reconstruction of SEM images in Figure 4 to obtain as much information as possible. 3D laser technique has also used to bring the third dimension into consideration.

Details can be found under  https://www.jeolusa.com/RESOURCES/Electron-Optics/Documents-Downloads/resolution-in-sem (e.g. for a 1 nm beam)

Lines 218-223. Discussion of pitch needs a diagram. Figure 2 is not adequate. What is the thickness of the cladding. What is the axial/radial separation between pellets and how is the spacing maintained?

Calculations for 3.5 mm pellets radius (see results in Section 3.2) gives relative density results presented in Fig 3. Clads in LWR typically have 0.1 mm thickness in ZPSR they could be slimmer e.g. factor 2 to 4.

The discussion of porosity is confusing. One line 223, there is a statement that 5% porosity was assumed (with no justification). Later, the measured porosity values appear to be much higher – line 256. Figure 5 shows that pore diameters ranged from 0.5 to 40 micron – how does that relate to porosity?

As this study is for simulated fuel for the nuclear reactors so used all the assumptions which are suitable for the radioactive materials. Infact, the data shows the theoretical, experimental and densities measured using different techniques for a mixture. The 5% assumption is for theoretical density (this is the classical order of magnitude for a UO₂ pellet, commercially), the higher values of porosity obtained from measured pellet after mixing the specific weights of salts, which is normal as the pellet here is experimental product not a iso-commercial product. The one is the Figure 5 is measured through SEM images, that is why it is different from the rest. Porosity values have changed with the change in type of density.

Colours are discussed in the caption of figure 4, but not the yellow and green colours. It is obvious that the pellets are inhomogeneous. How does that affect density determinations?

Yellow-green colours got visible through shadow formation as usually happen in topography analysis. Pellet is semi-homogeneous due to the different particle sizes of NaCl and CeCl₃, but through mixing was carried out to achieve maximum homogeneity.

Figure 7 shows laser mapping of the surface of the pellet. While this is useful information on the surface roughness, it cannot replace a bulk porosity analysis such as BET.

Yes, you are right. It was tried to keep the analysis as comprehensive as possible keeping in view the constraints of working with radioactive and hygroscopic materials.

Line 332. Pore volume is given, but it needs to be placed in context, e.g., mm3 per gram. Also, the

uncertainty is extremely low given the apparent inhomogeneity of the samples. Addressed

The use of XRD to determine densities via peak shifts/heights needs more detail. How does the analysis account for an inhomogeneous sample and the XRD analysis area?

In case of XRD, RIR and intensities of the individual peaks were considered to quantify the density. The intensities were normalized for the calculations. Below literature has also studied for phase quantification using XRD but not limited to these articles:

1. Robert D. Mariani, DeeEarl Vaden, Modeled salt density for nuclear material estimation in the treatment of spent nuclear fuel, Journal of Nuclear Materials, Volume 404, Issue 1, 2010, https://doi.org/10.1016/j.jnucmat.2010.06.022
2. Toraya, H. (2016). A new method for quantitative phase analysis using X-ray powder diffraction: direct derivation of weight fractions from observed integrated intensities and chemical compositions of individual phasesions of individual phases. J. Appl. Cryst. 49, 1508a1516. https://doi.org/10.1107/S1600576716010451
3. Alfons Berger, Marco Herwegh, Jens-Oliver Schwarz, Benita Putlitz, Quantitative analysis of crystal/grain sizes and their distributions in 2D and 3D, Journal of Structural Geology, Volume 33, Issue 12, 2011. https://doi.org/10.1016/j.jsg.2011.07.002

Line 417 – “has 9.7%” is missing text. Addressed

Line 423 – was irradiation induced swelling studied? Where does this assertion come from?

There is possibility of void swelling due to defect in crystal structure which can be caused by mechanical stress. There is a bunch of evidence that fast neutrons affect materials. The rearrangement of atoms can also contribute to void swelling. However these effects are not expected to be significant because of the very low flux in ZPR. All the investigations are in context of CeCl₃ is surrogate of UCl₃ and other radioactive materials.

Reviewer 3 Report

Comments and Suggestions for Authors

I have reviewed this manuscript and I believe there are a few ways in which the paper should be improved before publication, but overall I believe the data in the manuscript is useful and impactful to the scientific community and the methodology appears to be sound. Please see my recommendations below:

 

--page 2, line 38, it is claimed that molten salts have high thermal conductivity. This is generally not true, molten chlorides especially tend to have thermal conductivities which are 0.6 W/mK or less. This is thus on par or less than water generally. Please see: https://doi.org/10.1016/j.solmat.2014.03.028

-- page 2, line 76, it is claimed that very little property data is available for MSR fuels and coolants, this is perhaps an overstatement. There are gaps in the data, particularly for fuel salts. The authors are encouraged to look into the Molten Salt Thermal Properties Database (https://mstdb.ornl.gov/). 

-- page 7, line 232 it should be justified where the 5% porosity assumption comes from

-- In Tables 1, 2, and 3, there should be a separate row for the mixture, Instead of putting mixture data in the same row as NaCl. 

-- In Figure 5, it does not make sense to connect all the data points with a smooth line. The data points should all be shown separately, perhaps with a least-squares fit of a function that is relevant to the data. 

--Figure 7 f labels are hard to see. Also, What does the green horizontal line represent in the Figure?

--In table 5, it needs to be more clear what the 2D SEM and 3D LM data means, which was not determined with the Mountain Software. 

-- In the conclusion, it is not clear what this sentence means: "Through density comparison it is found out that the actual density has 9.7%, from 2D SEM technique 10.8% and from 3D LM technique."

--It is claimed in the conclusion that "The investigation carried out at grain (nm) level using XRD found the porosity of the grain 10.28% more than the calculated one." Based on Table 5 I'm not seeing where "10.28 %" more is coming from.

--The last sentence of the conclusion states "The investigations have been performed as a feasibility study for a ‘Zero Power Reactor Experiment’." How is this a feasibility study? This seems more like a characterization study. How does this study define whether or not the fuel may or may not be feasible?

 

Comments on the Quality of English Language

No comment

Author Response

Response to Reviewers in ‘Blue’

I have reviewed this manuscript and I believe there are a few ways in which the paper should be improved before publication, but overall I believe the data in the manuscript is useful and impactful to the scientific community and the methodology appears to be sound. Thank you so much for the comments.

 Please see my recommendations below:

--page 2, line 38, it is claimed that molten salts have high thermal conductivity. This is generally not true, molten chlorides especially tend to have thermal conductivities which are 0.6 W/mK or less. This is thus on par or less than water generally. Please see: https://doi.org/10.1016/j.solmat.2014.03.028

Addressed

-- page 2, line 76, it is claimed that very little property data is available for MSR fuels and coolants, this is perhaps an overstatement. There are gaps in the data, particularly for fuel salts. The authors are encouraged to look into the Molten Salt Thermal Properties Database (https://mstdb.ornl.gov/). 

Thank you for the comments. Salt mixture data is quite limited.

-- page 7, line 232 it should be justified where the 5% porosity assumption comes from

As this study is for simulated fuel for the nuclear reactors so used all the assumptions which are suitable for the radioactive materials. Infact, the data shows the theoretical, experimental and densities measured using different techniques for a mixture. The 5% assumption is for theoretical density (this is the classical order of magnitude for a UO₂ pellet, commercially).

-- In Tables 1, 2, and 3, there should be a separate row for the mixture, Instead of putting mixture data in the same row as NaCl. The table is showing the pellet components.

-- In Figure 5, it does not make sense to connect all the data points with a smooth line. The data points should all be shown separately, perhaps with a least-squares fit of a function that is relevant to the data. 

Please see lines 283 onwards, ‘From 0.47 to 38.7 µm the average slope of log (δN/δd) vs log d was -2.3 (Figure 5b) which mean that the contribution of the large pores is more significant than the small ones (Pareto distribution)’. This allow to make a precise analysis of the distribution.

--Figure 7 f labels are hard to see. Also, What does the green horizontal line represent in the Figure?

Green horizontal line represents the laser light filter before penetrating the sample.

--In table 5, it needs to be more clear what the 2D SEM and 3D LM data means, which was not determined with the Mountain Software. This data means, processed using ImageJ for 2D SEM and built-in software of 3D LM. Lines 271 and 315 are describing in detail.

-- In the conclusion, it is not clear what this sentence means: "Through density comparison it is found out that the actual density has 9.7%, from 2D SEM technique 10.8% and from 3D LM technique."

Thanks for the comment. As the whole article is describing the densities at different scales so this sentence is trying to summarise it.

--It is claimed in the conclusion that "The investigation carried out at grain (nm) level using XRD found the porosity of the grain 10.28% more than the calculated one." Based on Table 5 I'm not seeing where "10.28 %" more is coming from.

It is the difference of the percentage of porosity. Column 4 and 7 of the table calculates it. thanks

--The last sentence of the conclusion states "The investigations have been performed as a feasibility study for a ‘Zero Power Reactor Experiment’." How is this a feasibility study? This seems more like a characterization study. How does this study define whether or not the fuel may or may not be feasible?

The experimental and characterisation side of the project is part of the whole plan ‘Defining the draft for zero power reactor’. Thanks for the comment.

Comments on the Quality of English Language

No comment – Thank you.

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

Comment 1. The neutronics/activation will be so different in a ZPR than an operating reactor, the utility of the iMAGINE project as presented is still not clear. The reason for the temperature excursion was not presented and it doesn't seem to be addressed in the text.

Comment 2. If the fuel composition is not important, then how does that jive with the statement that "even small changes in composition cause changes in the physical properties of the salt" - line 68.

Comment 3. The authors continue that they are confident that their 2D techniques, slicing, SEM images etc are as robust as 3D measurements (3D CT, BET). At the least, they need to understand the uncertainty involved in doing a 3D reconstruction from 2D data - even if the data are gathered through slicing (which can affect the substrate). BTW - BET data are useful to understand fission gas migration through fuel - although I guess the argument that will be used by the authors is that for a ZPR there are no fission gases, or at least not enough to be important.

 

Author Response

 Many thanks for the comments the response is in Blue

Comment 1. The neutronics/activation will be so different in a ZPR than an operating reactor,

the ZPR neutron flux is about 106 smaller reducing the energy production then the temperature in core as well as the production rate of fission product and actinide.

the utility of the iMAGINE project as presented is still not clear.

Imagine is an advanced nuclear technology to produce energy without requiring expensive processing technologies. Most of the processing is close to room temperature and can even use spent-nuclear fuel.

The reason for the temperature excursion was not presented and it doesn't seem to be addressed in the text.

Reason of temperature excursion may be due to neutronic effect during energy increase e.g. initiation e.g. due to delayed neutrons. Next articles will be based on the high temperature studies.

Comment 2. If the fuel composition is not important, then how does that jive with the statement that "even small changes in composition cause changes in the physical properties of the salt" - line 68.

During reactor operation e.g. under a neutron flux of 10⁸ cm^-2 s^-1  (LWR flux : 10¹⁴ cm^-2 s^-1), even small changes in the fuel composition may affect the associated physical properties of the salts, therefore new methods must be developed for high purity sample preparation and advanced analytical techniques should be applied for reproducible measurement of each property, see  Degueldre (2017)[29]. Density, melting temperatures, density, specific heat capacity, thermal diffusivity and thermal conductivity are the important salt characteristics that affect the designs of MSRs, for fuel salts.

Degueldre (2017)[29] already given in the list

Comment 3. The authors continue that they are confident that their 2D techniques, slicing, SEM images etc are as robust as 3D measurements (3D CT, BET). At the least, they need to understand the uncertainty involved in doing a 3D reconstruction from 2D data - even if the data are gathered through slicing (which can affect the substrate). BTW - BET data are useful to understand fission gas migration through fuel - although I guess the argument that will be used by the authors is that for a ZPR there are no fission gases, or at least not enough to be important.

To be accurate CT is 3D with 3D image reconstruction of open and close pores, BET is 0D (no image) and concerns open pore only!

for a ZPR there are no quasi fission gases, or at least not enough to be important (in terms of release mainly).

Author Response File: Author Response.pdf