Author |
Message |
Gabi Levin (gabiruth)
Member Username: gabiruth
Post Number: 15 Registered: 5-2009
| Posted on Tuesday, September 15, 2009 - 4:57 pm: | |
Hi Jerry, OK - first it is not true that tungsten or any regular incadescent source is featurless. The output energy at different wavelengths follows a distributiion law that has been formulated years ago by physicists and it is temperature sensitive. Second, in theory, if you have a true real time dual beam spectrometer that measures the inetnsity of light from the source at each wavelength and divides the returning light at each wavelength to the launched light than the "features"a as you call them shall not interfere with the measurement. The problem may arise when you have such "strong" features that they can cause "stauration" of the signal in those wavelengths and leave other areas "weak" signal. Another aspect is how stable the "features" are - if they are very stable than their influence can be regarded as marginal. If the features are unstable, then there will be significant problems using it without real time dual beam arrangements I hope this helps you. Gabi Levin Brimrose |
Ian Goodyer (zinir)
Junior Member Username: zinir
Post Number: 7 Registered: 5-2006
| Posted on Tuesday, September 15, 2009 - 2:40 pm: | |
Jerry, can you tell us a bit more about the light source you are using. Do you have its spectrum? Are there any public websites available that we could look at that describe the source in more detail or is this an experimental source?? Ian |
Howard Mark (hlmark)
Senior Member Username: hlmark
Post Number: 266 Registered: 9-2001
| Posted on Tuesday, September 15, 2009 - 1:44 pm: | |
OK, OK, I won't ask why. I will ask, though, how large the peak is, compared to the rest of the spectral output? When you use tungsten-halogen lamp, while the spectrum can be measured over the entire range of the detector's sensitivity curve, most of the information in the active region is around the blackbody peak emission of the lamp, and most of the useful spectral information is where the lamp emission is more than, (very) roughly, 1/3 of the peak emission. Part of the reason for that is the residual detector (and system) non-linearity. Referencing works properly only when the actual instrument conditions conform to the theoretical expectations, one of which is that the detector output is exactly proportional to the optical energy input, i.e., perfectly linear. So any nonlinearity in the detector will prevent the referencing from working properly, and the ratio of sample/reference reading will not be indpendent of the signal level. The larger the peak emission is, compared to the rest of the emission spectrum, the more likely it is that you'll run into trouble from these non-linearities. The extreme case is where the detector "flat-tops" (also called "saturates") and the output doesn't change at all with further increases in the optical energy on it. Other sources of system non-linearity can occur, too; one possiblity is stray light. Instrument design is a complicated, tricky affair. A modern NIR instrument does not necessarily use the best possible individual components, it uses the ones that work together the best, which can be difficult to determine. The other side of that coin is that changing parts cannot be done lightly. You'd better have a REALLY good reason to do something like change the type of lamp. And if you do, then you'd better be prepared to spend the time and effort needed to make the new parts work together properly. \o/ /_\ |
Jerry Jin (jcg2000)
Intermediate Member Username: jcg2000
Post Number: 18 Registered: 1-2009
| Posted on Tuesday, September 15, 2009 - 12:45 pm: | |
The common light source for NIR is usually tungsten lamp which is "featureless" in the NIR range. I am talking about another NIR light source, some light emitting diode (LED) source. There is feature/peak in the spectral profile of the light output. I was wondering if these features can be carried over to the spectrum of samples. I guess the use of reference will cancel out the feature from light source. But I am not sure of that since LED is not a traditional NIR source. Thanks for your input. Jerry Jin |
Howard Mark (hlmark)
Senior Member Username: hlmark
Post Number: 265 Registered: 9-2001
| Posted on Tuesday, September 15, 2009 - 12:26 pm: | |
Jerry - there you go again, asking questions without telling us what you're doing or what you're using. I'm not aware that any of the sources typically used (i.e., tungsten-halogen lamps) for NIR measurements have sharp emission features. If you're using a non-typical source, then you surely have to tell us all about it, so we can put your question into context. \o/ /_\ |
Jerry Jin (jcg2000)
Intermediate Member Username: jcg2000
Post Number: 17 Registered: 1-2009
| Posted on Tuesday, September 15, 2009 - 11:06 am: | |
Dear all, Can anyone explain the influence of spectral feature in light source on the measurement? For instance, there are several sharp peaks in the emission spectrum of the light source, does that affect the accurate measurement of light absorbance using that light source? Thanks. Jerry Jin |
|