XRD is a great technique for understanding crystalline structure and even elemental ratios in products, but it has it’s operational limitations. A limiting factor is that in order to be confident in the output it is necessary to achieve a certain “peak intensity” in the analysis. Peak intensity is determined primarily by “dwell time”, the length of time a measurement is taken for any given area of the sample. Long story, short – to achieve a satisfactory peak intensity (say 4000) usually take around 30 minutes per sample.
So, working out the maximum number of samples that can analysed in a working day is pretty easy.
9 samples in a holder take 4.5 hours, allow 30 minutes to take samples off and put new ones on, and in 8 hours you can make two runs, so in a working day that’s just 18 samples. But of course if you put the first 9 on at 8.00am, the next on at 1.00pm and the last on at 5.30 then you can squeeze 27 samples into a “working day” of 9 ½ hours. Which is fine as long as you don’t have clients needing longer than 30 minutes runs, or more than 27 samples in a day! So when we were faced with both those challenges issues, we had to identify how to solve the problem, and we were quite clear this is our problem, not the client’s – who quite rightly just wants the data!
Shift systems can give an initial boost in capacity, but are not really a sustainable approach, and in any event it could work out expensive for us in the long term (and therefore the client).
Increasing the number of samples in the sample holder, so that we could run all night, offered an option. However, that changed the sample size (and therefore the controlled method we use) and the electronics of our cGMP controlled Bruker D8 XRD kit. Neither change was something we would want to leap into, and the nature of such changes is that they are time consuming, so not a sensible option.
The we considered the Bruker “Lynx eye” attachment to our detector. This is an array of 192 detectors rather than just 1. This allowed much more of the data from different parts of the sample to be collected in a single run. In theory a ½ hour run could be reduced to a few minutes and still reach the required peak intensity.
In discussion with Bruker we agreed this was the right approach, and instigated the change over a 6 week period (from first discussion to delivery and installation). On the first use we found a 6 minute run time produced a peak intensity of over 20,000 – well above our target level. We’re now establishing the precise dwell time we need. The bottom line is that we’ve increased our capacity from around 30 samples a day to over 100. That gives us both the greater capacity and the cover we need for routine servicing and other operational issues.
Just goes to show, “it ain’t what you do, it’s the array that you do it (with) .. er sort of!