| Author | Message | ||
     Howard Mark (hlmark) Senior Member Username: hlmark Post Number: 485 Registered: 9-2001 |
Marta - unless both sides of the sample are completely plain (i.e., without any engraving) something you can do besides turning the sample over is to measure the sample in different orientations, and average the readings. If the sample is long and narrow, you should at least turn it end to end. If the sample is square or round, you can use more than two orientations. Since the engraving is generally asymmetric, this helps averaging out the differences in scattering almost as much as turning the sample over. You should still turn the sample over and average the readings from both sides, and if both sides have engraving and you should also rotate it when either side is up. \o/ /_\ | ||
| Sheelagh Halsey (shalsey) Advanced Member Username: shalsey Post Number: 25 Registered: 12-2006 |
Marta I am presuming this is a reflectance measurement from the description in your question. If so, you have done the right thing to include both surfaces of your tablet in your method. Different engravings, break marks etc, will affect the way the light is scattered from your tablets. This will manifest itself in baseline shifts between different surfaces of the same tablet. You may well have used spectral pre-treatments to alleviate these effects when making your model. In addition, multivariate models will also help compensate for known variations. Hopefully your validation will have incorporated data showing that you get the same results whichever side of the tablet is analysed. I would suggest that you discuss what the issue is to show your understanding and discuss how you have compensated for it. Regards Sheelagh | ||
| Marta Lichtig (marta) Junior Member Username: marta Post Number: 8 Registered: 8-2009 |
As response to one of our submissions, regarding NIR testing of individual tablet assay,we were asked to comment on the effect of surface morphology on the NIR reading. It should be mentioned that the tablets have different engravings on the two sides, and we took this fact into account when building the model (the model is based on the average of reading the two sides.Can someone help out with some ideas about what the answer should be? |