| Author | Message | ||
     venkatarman (venkynir) Senior Member Username: venkynir Post Number: 50 Registered: 3-2004 |
First chemometrics tutorials (in INDIA) wiil be held at Trivandrum Feb 23 2008 in India as pre-conference (International conference of Vibrational spectroscopy ICOVPS2008 ). Kindly visit site for more infromaion http//www:ICOVPS2008.org | ||
| Dr.K.Balasubramanian (drkbala) New member Username: drkbala Post Number: 1 Registered: 9-2006 |
Resolution is not a problem if you have more scans amd the average spectra would give hand. | ||
| MSRajpoot (msarajpoot) New member Username: msarajpoot Post Number: 5 Registered: 7-2006 |
I think the high resolution may required mainly for the calibration transfer from one instrument to another. | ||
| Gavriel Levin (levin) Member Username: levin Post Number: 11 Registered: 1-2006 |
Hi, I am glad that someone with experience like Howard tells the story about the resolution - resolution comes at the expense of noise - and for quantitative work - noise is much more critical than resolution. One more point - the NIR peaks are not so sharp anyway - so resolution contributes in a limited way to quantitative work - up to a point where you start loosing information because of low resolution - but this is not happening even with 6 nm resolution - whatever this translates in cm-1 which is typical of dispersive instruements. Gabi Levin | ||
| Jonathan Hoggatt (jhoggatt) Member Username: jhoggatt Post Number: 12 Registered: 6-2006 |
Thanks Howard for the response. That is partly what I mean by increased time... not only an increase in time per scan, but also an increase in the number of scans per sample to get the same S/N ratio. You say only use high resolution when you need it... when would that be? I am thinking perhaps a pseudo-polymorphism change in a drug substance (crystallinity) detection might require a slightly higher resolution than normal. Any references discussing this topic? Thanks again, Jon | ||
| Howard Mark (hlmark) Senior Member Username: hlmark Post Number: 43 Registered: 9-2001 |
Jonathan - don't forget that along with the extra resolution comes a decrease in S/N level. This is a fundamental phenomenon and you have to average more than proportional number of data points to bring the S/N level up to the level of the lower-resolution spectrum. For instance, if you change resolution from 8 cm-1 to 2 cm-1, you have to have to average together groups of more than four spectral points to restore the S/N level to what you had with the 8 cm-1 spectrum. This is true whether you scan for the same amount of time or collect data from the same number of scans. For this reason you should use the higher resolution setting only when you really need it. \o/ /_\ | ||
| Jonathan Hoggatt (jhoggatt) Member Username: jhoggatt Post Number: 11 Registered: 6-2006 |
Is anyone aware of what effect (if any) the resolution of your instrument has on your calibration model? For instance, let's say you have an FTNIR instrument and you can collect as high as 2cm resolution. Besides the increase time in scanning, is their any inherent disadvantage to collecting spectra at the maximum resolution? What is the current rule of thumb... scan at the highest resolution possible (2 cm in this case) or scan at the lowest resolution possible and still see your analyte of interest? Does a high resolution affect derivatization... it would seem that a GREAT deal of smoothing would be required in order to make any sense out of a derivative? Any thoughts on this subject would be most appreciated. Jon |