Jin Kwei Koh

I'm not sure I understand your question but you state 'XRD data results show the d spacing, crystallinity size and lattice constant'. I'm not sure I agree with that statement as this information is derived from the raw intensity - (x-ray) scattering angle (2Ɵ) data - it's not 'seen' directly. It's this 2Ɵ (Bragg) information that I'd consider the raw data (unless you're talking of counts/pulses at the detector). The derived information is subject to assumptions (impinging wavelength, phase content, crystalline form, crystallite form, the actual processed peak (e.g. for Scherrer). Other information (e.g. amorphous content) comes from that basic information/plot. You need to explain what you mean by 'Not sure which directions to explain the statistical scattering parameters to support XRD data?'' as the scattering data are exactly the XRD data used to derive other information that will be considered more useful.

Alan F Rawle Thanks for your feedback. Actually I just conduct my material under XRD testing, its data only can derive the information of d spacing, crystallinity size and lattice constant.

I showed this data in my revised article, but reviewer request the statistical scattering of parameters to support those d spacing, lattice constant derived from XRD.

Hence, I am not sure what is related information regarding statistical scattering that can support to those data? I saw you said "the scattering data are exactly the XRD data used to derive other information..." Seek your help advise on that. Appreciate and thanks.

Jin Kwei Koh

I must admit I’m a little confused by the reviewer’s points, but I see 2 possibilities for the questions:

• Counting statistics. See, for example: http://clay.uga.edu/courses/8550/XRD.html This link is also useful: https://www.researchgate.net/post/How_can_I_calculate_total_time_for_an_XRD_measurement

To improve the accuracy by a factor of 2 will require 4 times the total counts

• Standard variable/error balance based on the 3R’s (repeatability, reproducibility, robustness) principle.

- Did you conduct your experiment a minimum of 5 times consecutively (repeatability/intra-assay precision)? Here you can gage the effect of the sample stability and any experimental variations and you may be able to look at something like lattice parameter(s) +/- calculated variation

- Did you take multiple aliquots from your same bulk sample (heterogeneity effects/reproducibility)? In many systems bulk heterogeneity can exert an enormous effect on the end result as only small amounts of sample are being considered. Think of SEM and an ore sample. Here's a webinar example in particle size (registration required):

Keys for Successful Analysis – Representative Sampling & Estimation of Standard Error Calculation

https://www.malvernpanalytical.com/en/learn/events-and-training/webinars/W190613KeysLD

- Did you undertake a standard variable balance budget looking at the effect that each experimental parameter can effect the end result (robustness)? In the latter case we’d know that getting the wavelength (e.g. Cu Kα) incorrect is unlikely and will not have a large effect on the end result. Estimating the peak position (2θ) by eye is likely to incur a large error.

This error/variable balance used to be a feature of standard error estimation in experiments and seems to be little practiced these days (one result is perfect as nothing to compare it against)

Dear Alan F Rawle ,

I agree with you.

I am also confused with respect to the comment of Jin Kwei Koh 's reviewer;

requesting 'the statistical scattering of parameters to support those d spacing, lattice constant derived from XRD'.

This 'statistics' can only be related to the repeateabilty of the results such as d (or lattice constant(s)) and FWHMs,

when

a) the scan is repeated,

b) some scan parameters, such as step width and (local) scan time are changed,

c) Tube voltage is changed a bit,

d) sample removed from the setup and being replaced by another one of the same material (if there is enough of material at all).

I must admit, that in principle there are a lot of scans to do.

But it is, to my opinion, not necessary to do such repetitions for each kind of individual sample, but only for a representative one.

So, Jin Kwei Koh, to my opinion, in your article you should attach some 'measurement error' to the d and or FWHM(s) based on some repeated scans with different input parameter (scan parameters).

Good luck

and

best regards

G.M.

Gerhard Martens

You comments are extremely useful - as always. Your last sentence highlights one absolutely vital point: '...but only for a representative one'. This is key in all metrology and representative sampling will always be (wrongly?) assumed. In XRD we may be looking at a small sample from the whole. We don't know the type/nature of sample being measured by Jin Kwei Koh and this information will be valuable in gaging possible or probable variation(s). The webinar above (registration required) provides some guidelines that would be valuable in ore sampling (but not in colloidal metals):

Keys for Successful Analysis – Representative Sampling & Estimation of Standard Error Calculation

https://www.malvernpanalytical.com/en/learn/events-and-training/webinars/W190613KeysLD

I think that the reviewer could be asked to clarify the comments and perhaps provide an example of what is required.

I apologize in advance. I'm also confused by the requirement to estimate the error on the d values. This would really require a lot of repeated measurements. However, a number of programs for Rietveld refinement (e.g. Topas, BGMN) perform error estimates or estimation of the standard deviation in the calculation of cell parameters and domain sizes. Wouldn't that be enough?

If it is taken exactly, even for d values it is possible to state the standard deviation (compared to the analytical profile function, usually pseudovoight's f. ). But I never did.

Thanks for your comments, Alan F Rawle and Dalibor Matysek.

My sample tested is silica, which is commercially used in lab.

For my reviewer's question, I will try to solve as what you both suggested. Now is look like guess on the reviewer's requirement. Appreciate all your help.

Jin Kwei Koh Ask the reviewer what is required. Don't try to guess him/her...

Please check the fully-ready preprint article DOI: 10.13140/RG.2.2.27720.65287/3, Title: Qualitative Analyses of Thin Film-Based Materials Validating New Structures of Atoms