| Author | Message | ||
     Gabi Levin (gabiruth) Senior Member Username: gabiruth Post Number: 70 Registered: 5-2009 |
Peter, I think it is basics - the I0 should be the intensity at each wavelength of the signal detected on the same detector and electronics that is used for the sample signal - there is no other way to assign value to the incident light - and this is why it becomes important to make sure the "standard" is always positioned in a very reproducible way - e.g. in some single beam such as rotating grating instruments the standard is built in, so it is supposed to be very reliable - and it has to be extremely reliable - otherwise you would have great errors in predictions. If you have a single beam instruement where you have to manually place the standard to create the I0 signal for the whole spectrum - you become dependent on your capability - I find it hard to believe this is the case. Gabi Levin | ||
| Howard Mark (hlmark) Senior Member Username: hlmark Post Number: 479 Registered: 9-2001 |
Peter - I also forgot to point out that the angular distirbutions of the reflected light from the sample and from the reference reflector may not be the same. This exacerbates the effects of the samples' reflectance not being the same as the reference, and our practical inability to measure all the light coming back. \o/ /_\ | ||
| Howard Mark (hlmark) Senior Member Username: hlmark Post Number: 478 Registered: 9-2001 |
Peter - in the strictest sense, no. Theoretically, Io is the amount of energy impinging on the sample and Id is the total energy coming off (reflected from) the sample. It's difficult enough to measure ALL the energy being reflected back from the sample due to its reflection, and in all situations except specially set up reflectometers, it's impossible, in any practical sense, to measure the energy impinging on the sample. Therefore, to get around that impossibility, the light reflected by a reference reflector is used as a surrogate for the amount of energy impinging on the sample. Ideally, the reflector would reflect 100% of the light, and therefore any sample would always reflect less, and then the measured absorbances would always be positive. In practice, of course, nothing reflects 100% (in addition to the other causes that were described), and therefore for real measurements, you may soemtimes run into cases where the sample reflects more light than the reference reflector does. \o/ /_\ | ||
| Peter (drpratai) New member Username: drpratai Post Number: 3 Registered: 4-2012 |
Howard, I am still in contact with the company, I may yet receive an answer. If the information is indeed considered to be proprietary, then I will respect that. Therefore, I do not wish to disclose this information. Gabi, thanks you for your reply. From your response can I assume (in general with single beam spectrometers): R = Id / I0 , with I0 corresponding to the detected illumination of the standard and not the incident illumination? | ||
| Gabi Levin (gabiruth) Senior Member Username: gabiruth Post Number: 69 Registered: 5-2009 |
Hi guys, I have watched this discussion for some time - but not with great interest until now, mainly because it makes no difference if you get negative values or not. Negative values result when the signal at a given wavelength is stronger than the signal from the reference at same wavelength. For single beam spectrometers, such as most instruments - the reference signal is supposed to be highest possible as it is collected from a very high reflectance "standard" surface. Thus, if a negative value occurs it means that at a given wavelength the signal from the sample is higher than from the reference - resulting in a value >1 and the -log of absorbance becomes negative. The question is - why is the signal from the reference lower - if this is due to some contamination, it will be limited to certain wavelengths, or it is improperly positioned in an angle, or maybe the distance is not the optimal distance so the signal is reduced. I expect the negative values to occur only in those parts of the spectrum where the absorption of the sample is the lowest. So, either leave it be or look for improper positioning, contamination. All this is correct for single beam spectrometers because they use the same complete set, detector, amplifiers, A/D converters etc for both signals. For real time, full dual beam spectrometers like the Brimrose - negative values arise from the fact that the reference channel signal is based on a separate detector that measures the reference beam at each wavelength all the time at same time it scans the sample -in such case we fix the reference channels "amplification" in the factory and there is no need to collect reference from a "standard" sample at any time. However, it could be possible that in some cases, the reflectance from a sample at some wavelengths will be higher than the preset amplification signal of the reference beam and result in negtaive values. None the less, as I said, it has no significance at all - if you take the derivatives - there is no importance to negative values. Gabi Levin brimrose | ||
| Howard Mark (hlmark) Senior Member Username: hlmark Post Number: 477 Registered: 9-2001 |
Peter - unfortunately, it may be that the manufacturer considers the information "proprietary". Another possibility is that someone else on the discussion group knows exactly what that manufacturer does, but since you haven't said who it is, they can't respond. Do you have any reason to not say who it is? If not, why not tell us so that if anybody knows their internal data-handling, they can tell you. \o/ /_\ | ||
| Peter (drpratai) New member Username: drpratai Post Number: 2 Registered: 4-2012 |
Thank you for your helpful replies! Howard, I've gotten in contact with the manufacturer and as of yet have not received any insightful information. David, thank you for your answer! I have a ceramic calibration plate, which is regularly used to calibrate the device. I can now investigate further. I would be very grateful if anyone can supply references to literature concerning this subject as I wish to gain a greater understanding and to add references in my paper. | ||
| David E Honigs (dhonigs) New member Username: dhonigs Post Number: 4 Registered: 4-2012 |
Dear Peter Jenkins, I would like to try to partially answer your question about negative absorbance numbers. Absorbance numbers can be quoted on an absolute or a relative scale. Most instruments measuring in reflection quote absorbance (-log reflectance) as a relative number. The number is relative to some standard like an uncoated ceramic piece. The standards are never perfectly white. It is possible for the relative absorbance to be negative in this case as the sample can reflect more than the standard. If the absorbance is measured and corrected to an absolute scale then the numbers can be negative only to the extent that small fluctuations in the source or electronics make the amount of detected light change between sample and reference. Normally these deviations or negative numbers are very small on an absolute scale. On a relative scale they can be a few hundredths to even a tenth depending on the reference. If they are larger then the reference is a darker material or could be dirty. Another reason negative absorbances exist is the sample geometry. The amount of light returned to the instrument is generally a function of the illumination and detection distance. If the sample is closer than the reference it can return more light just from the geometry change and this can cause negative numbers. This is not an exhaustive list. Those are simply the two most common reasons for negative absorbances that I have seen. Sincerely, David Honigs | ||
| Howard Mark (hlmark) Senior Member Username: hlmark Post Number: 476 Registered: 9-2001 |
Peter - the only one who knows for sure is the instrument manufacturer. While some things are more-or-less constant in the industry, oftentimes the details of how the results are obtained, and what is being displayed, are unique to each company. I recommend you give them a call and ask to speak to someone in their Applications departmant. Howard Mark \o/ /_\ | ||
| Peter Jenkins (drpratai) New member Username: drpratai Post Number: 1 Registered: 4-2012 |
Hello there, I'm a student currently working on an assignment regarding diffuse reflectance spectroscopy. I wish to learn about determining moisture content in different samples using NIRS. The spectral device supplies two different values regarding moisture. One is the content in % (which is inaccurate as it has not yet calibrated) and the other, a value without a unit of measurement, is I suspect absorbance. As I understand light is passes through filter(s) at certain wavelengths (e.g. for moisture detection 1940 nm & 1450 nm) and is directed on the samples. The reflected light is then directed to a detector where the absorbance is determined. The absorbance can be calculated as follows: A = log(1/R), with R (reflectance) = I/I0 I believe reference values are generally used to correct any drift errors etc.. Later I intend to calibrate these absorbance values with the water percentages from laboratory measurements. Before this I would like to gain a greater understanding of this absorbance value that is outputted. Can anyone point me in the direction of good literature that explains exactly what this value is? A further question is that the outputted value is sometimes negative (I think when my sample has a moisture content of ~2%, I have to still confirm this). Is it possible that this value can be negative? Kind regards, Peter |