| Author | Message | ||
     Ian Goodyer (zinir) New member Username: zinir Post Number: 5 Registered: 5-2006 |
Hi Howard, Yes samples like coal were what I was thinking of. I have experience of imaging soil samples and very black dirty rubber and in these cases I had some success but felt that a brighter source may have increased signal to noise. I don't think that people used to imaging white tablets would find any benefit in a brighter source. Ian | ||
| Ian Goodyer (zinir) New member Username: zinir Post Number: 4 Registered: 5-2006 |
Hi Ralf, Thank you for your post, I obviously have to provide a bit more information. I do understand that a bulb running at 20 Watts cannot give off more than 20 Watts of energy (heat + light) and that all 20 W QTH bulbs will give a similar level of light to each other. What does vary, quite considerably with different commercially available sources is the fraction of this light that falls on the surface of the optical fibre and the proportion that is therefore coupled into the fibre. Even without any lenses or mirrors if you double the distance between the bulb filament and the end of the fibre cable, you would expect the amount of light to fall on the fibre to decrease by a factor of 4. From what you say, it sounds like calibrating light sources would not be required, which is clearly not the case. We are, in essence, a very practical group of people. This may help: take a fibre optic cable and clamp one end so that it points at a light meter, screw the other end into a series of light sources from Ocean Optics, Hamamatsu, AlphaBright, Avantes, Bentham, ThorLabs or whoever and compare the amount of light that falls on the light meter. When we did this experiment we found although the bulb powers only varied from 7 Watts to 20 Watts there was a hundred fold difference in light level hitting the light meter. Ian | ||
| Howard Mark (hlmark) Senior Member Username: hlmark Post Number: 234 Registered: 9-2001 |
Tony - yes, that's probably true, at least for any "real" tablets. It also occurs to me that there are people who try to measure spectra of coal, coke and other very "black" samples, who might benefit from having greater energy available. \o/ /_\ | ||
| Tony Ainscough (tony_ainscough) New member Username: tony_ainscough Post Number: 4 Registered: 5-2009 |
Yes Howard you are right to point that out, it strengthens my argument about there being more than sufficient energy output for NIR regardless of monochromator/interferometer design. Just as an example it is possible to detect transmitted NIR energy through a tablet of 8mm thickness even with a grating spectrometer. Sufficient to accurately assay the API content. What might be interesting though would be if the Assay accuracy could be improved by even illumination of the tablet. The NIR energy dispersion beyond the surface would be so great i suspect it would make little difference. Tony | ||
| Howard Mark (hlmark) Senior Member Username: hlmark Post Number: 233 Registered: 9-2001 |
Ralf - what you say is correct. It is not possible to increase the brightness (in the rigorous sense used by the physicists) of a source at a given temperature. However, there are two caveats to that: First, tungsten is not a particularly good emitter; its emissivity is about 0.4 to 0.5 in the visible region, decreasing to about 0.2 to 0.3 in the NIR, over a fairly broad range of wavelengths and temperatures. Some of this can be compensated for by imaging the filament on itself, so that the reflected light adds to the emitted light; this results in higher energy output from the lamp, and the technique was used in the design of the Technicon InfraAlyzer instruments. Secondly, an image of a coiled filament typically is not a solid block of light; the coils leave gaps in the image, where no light is present. A mirror can take the light from the 'back" side of the filement and use it to "fill in the gaps" so that again, more energy is made available, if not greater brightness. Kohler illumnination is another way to redistribute the light so that it is more evenly spread out over the image. To make a comment on Tony's posting: to some extent that depends on the design of the spectrometer. A grating spectrometer that disperses the light before the sample, only allows a small fraction of the energy to impinge on the sample at any one time, and therefore heating of the sample from that cause is negligible. \o/ /_\ | ||
| Tony Ainscough (tony_ainscough) New member Username: tony_ainscough Post Number: 3 Registered: 5-2009 |
Hi Ian, In most modern NIR spectrometer designs there is usually sufficient light for most applications. In fact with an FT-NIR design there can be too much and occasionaly filters have to be used to avoid detector saturation. There is also the distinct disadvantage of the intense light source heating samples(not good in quantitative NIR)or even worse burning them. That said the NIR world is a weird and wonderful place, i wouldnt bet against there being an application out there that could benefit from this. regards Tony | ||
| Ralf Marbach (ralf) Junior Member Username: ralf Post Number: 8 Registered: 9-2007 |
Ian, The brightness of a thermal source is determined by its temperature and emissivity. These are more or less constant for all tungsten filaments, including yours. Mirrors and lenses can not increase the brightness a.k.a. radiance of the output beam, they can only decrease it, if lossy. It's thermodynamics. Once you fill the optical etendue of your fiber [mm^2 steradian] with your available radiance [Watt/(mm^2 steradian)], that's it, that's the maximum energy the fiber can carry. I don't know what you measured in your comparison studies, but your source can not be brighter than other tungsten sources heated to 2850 degK. Ralf | ||
| Howard Mark (hlmark) Senior Member Username: hlmark Post Number: 232 Registered: 9-2001 |
From my examination of the diagram on the website, this seems to be a fairly straightforward aplication of Kohler illumination. Kohler illumination has been known at least since the 1950s (and likely much earlier), since that's when I read about it in books on microscopy, an application where it was already considered the "standard" method of illumination. The purpose of Kohler illumination was to ensure the evenness of illumination of the sample across the microscope's visual field, not necessarily to increase the brighness. So it's not anything "new" although the application as a generalized illuminator is interesting. \o/ /_\ | ||
| Jerry Jin (jcg2000) Member Username: jcg2000 Post Number: 13 Registered: 1-2009 |
Hi, Ian, In my applications of biomedical optics, the light source needs to be strong enough to penetrate through dark skin and thick fat layer to interact with the deep human tissue we are interested in. But it can't be too intense, otherwise it will burn the patient. So there is compromise. We would like to adjust the source intensity to a low level and use a sensitive detector. It is already good to have a adjustable light source. I am sure there is market for your tungsten halogen light source. The uniform distribution of the wide wavelength range is irreplaceable. This merit is appreciated from the very beginning of spectroscopy and is going to benefit for sure. What you need to do is to make it small and last longer, perhaps. Cheers! Jerry Jin | ||
| Ian Goodyer (zinir) New member Username: zinir Post Number: 3 Registered: 5-2006 |
i am wondering whether people on this forum thought that there would be a need in the NIR community for a brighter QTH based optical fibre coupled light source. I think that there is and my company has developed one, www.alphabright.com, but I would like to test its appeal to the NIR reflectance market. I have had some experience using one of the new detector array-based 'mini-spectrometers'. The unit's detector has a full response from 0 to 65000 but I was only obtaining readings of ~4000 in my application. All I could do was to increase my signal was to increase integration time. With a light source 10 times brighter, the detectors would be working closer to saturation increasing signal to noise ratios and hence the quality of the results without having to increase integration time. I am sure that there are some applications where more light would not be advantageous and others where it will be really useful. I guess my question is what applications would benefit from increased light throughput? For example would you see this particularly beneficial for dark / poorly reflective samples or for imaging from a distance. Anybody have any experience of this? Thanks, Ian |