Two tiny things this week.
Firstly, the round-robin results are trickling in, and when reviewing Steven Weigand’s data, I noticed a correction I have not come across before: an “exluded volume correction”. As far as I can find (e.g. in this document by Burke & Bothe), this is a correction in the background subtraction, for the loss of background signal due to the presence of scatterer. As you can expect, this would impact the signal significantly at higher volume fractions. I’m sorry to say that I did not know about this correction before, and is therefore omitted in the “Everything SAXS” review paper.
I will figure out more about this correction, its application and impact (and hopefully figure out if it is correlated to the other corrections or not), and write a more lengthy report on it.
The second item for today is that we discovered a bug with the desmearing implementation in the Anton Paar data reduction software SAXSQuant. In our version, it appears that the desmearing procedure silently drops all input datapoints with negative intensity values. We apply this desmearing correction after background subtraction of our unbinned datasets, and therefore our data contains quite some datapoints at high Q with negative intensity values (resulting from subtraction of noisy data from noisy data). Simply discarding all the negative datapoints will skew the profile at higher angles. Desmearing implementations by others will carry negative values through the correction, and so this is a bug.
Next week I will be on holiday, so there will not be a new post. However, contributed articles are always welcome!
Leave a Reply