| Author | Message | ||
     Tony Davies (td) Moderator Username: td Post Number: 263 Registered: 1-2001 |
Hi John, Don, Dave et al., John: You have the answer,; Rule 1: Karl is always right! But, do you have a transmission capability on your instrument? Don: Sorry about spelling your name incorrectly! Dvae: 1) I am dependent on spell checkers but there is none in this program! 2): I agreed with you when you said " a diffusely reflecting reference sample that also reflects about the same intensity of light as is being reflected by your samples." If a blacked-out slide with sample was used, then the reference would have to be a completely blacked-out slide. Both sets of readings would be very small and dominated by noise so there is a danger that the reported spectrum would go negative and I think this might cause an error when the computer attempts to convert it to log 1/R (I'm half remembering an experience with the GrainSpec which is a long while ago!). If you use a clear slide + sample and a standard instrument diffuse reflector for reference then you are measuring two large signals but at least the noise component will be relatively low and (as Gabi said) you do have double the sample thickness. Of course I'm not disagreeing with you Karl, transmission spectroscopy should always be the first choice! You would need to have a clear slide in the reference reading and, Don, would we have to start worrying about the surface reflection component? Isn't wonderful how these "newcomers" come up with difficult and interesting questions? Thanks John! Best wishes, Tony | ||
| Donald J Dahm (djdahm) Senior Member Username: djdahm Post Number: 69 Registered: 2-2007 |
There�s another way of looking at the background issue, which winds up pretty much at the same place as has been discussed. Those of us who had �classical� spectroscopy training were lead to think about the �reference� as a �background correction�. To do an effective background correction, the reference scan must be as similar as possible to the sample scan, except of course, for the presence of the �analyte�. For example, in a transmission experiment though a clear solution, we were encouraged to insert an empty cell in the reference scan, and we believed we had thus corrected for the absorption of the cell itself. The purists then wisely carried the correction one step further and suggested that we include the pure solvent in the reference scan, thereby correcting for any absorption of the solvent as well. It was years later that I learned that there was an additional reason for including the solvent. In an empty cell the refractive index change in the light path is rather dramatic without the solvent, which leads to scatter back and forth between the cell walls. Generally, departing from these guidelines for measurements on solutions does not result in a straight line relationship being lost; the line just moves. When scattering samples were first examined, classically trained spectroscopists assumed that they should use similar principles. This results in some good advice, because the nonlinear nature of the response curve (using essentially any metric we know) alters the response curve in far more complex ways than for solutions. Our defense has been to try to keep everything as �constant� as possible within an analytical system, including the reference. As a theorist, I support the idea of always trying to make measurements for scattering samples on an absolute scale, and then if background corrections are desirable, do them subsequent to the intensity measurements. However, I doubt I could convince you to change your ways by citing great benefits. Adding to Art�s geometry point: If you want to use a mirror, it needs to be tilted at half the angle of the detector. If detection is optimized for 45 degrees, the mirror would be tilted at 22.5 degrees. This winds up using a reflection geometry for what is largely a transmission experiment. I�m not recommending that, it�s just one more possibility that could be used if there was a good reason. | ||
| Jose Miguel Hernandez Hierro (jmhhierro) Intermediate Member Username: jmhhierro Post Number: 19 Registered: 4-2008 |
Dear John, I think that you can use a circular quartz cell and perform your mesurements in transflectance mode. Please, see this reference:FEMS MICROBIOLOGY LETTERS (2008), 284, 135-141. Best Regards Jose | ||
| mathews m john (mathews) Junior Member Username: mathews Post Number: 8 Registered: 6-2011 |
Thanks a lot for the clarification. Thank you :D | ||
| Art Springsteen (artspring) Senior Member Username: artspring Post Number: 41 Registered: 2-2003 |
One last thing. I think Karl hit the nail squarely on the head. If these were my samples, I would definitely be looking at them in diffuse transmittance mode. | ||
| Art Springsteen (artspring) Senior Member Username: artspring Post Number: 40 Registered: 2-2003 |
Mattherw Johns wrote: "can u please elaborate 'The rule of thumb is to measure diffuse against diffuse, specular against specular'?" The reason is that matching scattering characteristics of sample and reference gives better and more reproducible results. Typically, reflectance instruments are designed to measure diffuse reflectance and, from an engineering standpoint, you base a design around what is called the 'perfect reflecting diffuser', a material that is totally lambertian. Now, there are no 'real' materials that are perfectly lambertian, but a lot of stuff is pretty close (most white powder- salt, barium sulfate, packed PTFE, freshly fallen light snow- are close). And most materials we measure are probably closer to lambertian than perfectly specular. You then commented"Let me be a little more specific with the previous question. I don't see why i can't use specular reflectance as a background for diffusely reflecting sample. Light just bounces off from the background, hits the sample and then goes to the detector. So why can't i use the mirror itself?" Geometry, sir. If the instrument is bi-directional (as most NIR instruments are), they exclude the specular component. Most non-fiber optics based instruments are 45:0 or 0:45 geometry, which are designed to ONLY catch scattered light. So, if you measure a mirror, you will (should) see zero signal at the detector. That's why. | ||
| mathews m john (mathews) Junior Member Username: mathews Post Number: 7 Registered: 6-2011 |
Let me be a little more specific with the previous question. I don't see why i can't use specular reflectance as a background for diffusely reflecting sample. Light just bounces off from the background, hits the sample and then goes to the detector. So why can't i use the mirror itself? | ||
| mathews m john (mathews) Junior Member Username: mathews Post Number: 6 Registered: 6-2011 |
@artspringsteen.. can u please elaborate 'The rule of thumb is to measure diffuse against diffuse, specular against specular'? | ||
| Karl Norris (knnirs) Senior Member Username: knnirs Post Number: 53 Registered: 8-2009 |
I agree with the comments against using a mirror for the reference in your measurement, but I would choose to do a diffuse transmission measurement. I would use transmission through the clean slide as the reference, and measure the transmission with the added sample. You can vary the thickness of your sample to obtain the best signal to noise. | ||
| Gabi Levin (gabiruth) Senior Member Username: gabiruth Post Number: 61 Registered: 5-2009 |
Hi guys, I personally would not go for a black back ground - assuming that the smeared layer is thin, I would prefer to have the light go through the layer twice,once on its way through, and once the diffusely reflected light from the diffuse reflector will go through it again, on its way to the detecting system whatever you use. This way your sensitivity will increase. Thanks, Gabi Levin | ||
| mathews m john (mathews) New member Username: mathews Post Number: 5 Registered: 6-2011 |
Thanks a lot for all the inputs. I got a couple more doubts (i am really an amateur in this). 1. How close should the probe be to my sample? 2. If i paint the back of the slide black... i could as well just use that instead of using a diffuse reflecting background right? 3. The NIR system itself has a reference inside. So should i take the reference again with the background I use? | ||
| Gabi Levin (gabiruth) Senior Member Username: gabiruth Post Number: 60 Registered: 5-2009 |
Hi guys, In addition to all other points about mirror reflector - the issue with mirrors is that the reflected intensity that reaches the sensors becomes angle sensitive - it may be difficult to guarantee that the angle is perfect all the time. If by any chance the reflectivity at different wavelengths and different angles is varying you are in trouble. I would of course prefer a diffuse background - but may I suggest - it is possible to get fine ground slides that are almost perfect diffuse reflectors by nature of the roughness. These will be angle insensitive, cheap, cleanable, will not suffer from chemical deterioration and will be quite easy to handle. The other alternative is simply to go for transmittance - through the slide - if the set up permits. Thanks, Gabi Levin | ||
| Howard Mark (hlmark) Senior Member Username: hlmark Post Number: 441 Registered: 9-2001 |
Dave - It's not all that clear-cut, I think. It's the difference between what's likely to be a small signal versus what's about equally likely to be a small change in a large signal. Might be worthwhile trying both ways, though. I believe you're correct about being able to get ceramic disks from Coors. I've bought some from them, although not for some while, though. \o/ /_\ | ||
| Jim Burger (jburger) Junior Member Username: jburger Post Number: 6 Registered: 11-2010 |
For NIR work, black silicon carbide 'sandpaper' makes a great black background material. I use it routinely for NIR imaging work. You can get the material with different particle sizes, but in general: 1) it's very black; 2) the texture helps to hold samples in place - probably not an issue in this case; 3) its CHEAP! However... the diffuse properties make it black when imaging pixels are 100 x 100 microns. I have no idea what it looks like under a microscope - but certainly worth trying. Jim | ||
| David W. Hopkins (dhopkins) Senior Member Username: dhopkins Post Number: 197 Registered: 10-2002 |
Hi Howard, Good observation. That makes me think, I suspect a good alternative would be to use a ceramic disk as a standard, and then make the bacterial smear on the ceramic disk. That might result in a transflectance measurement that would be somewhat more sensitive than the pure reflectance measurement obtained with an absorbing background. I believe that ceramic disks are still available from Coors. Best wishes, Dave | ||
| Art Springsteen (artspring) Senior Member Username: artspring Post Number: 39 Registered: 2-2003 |
Hello again, Howard is 100% correct- you do need to get rid of any back reflection from what is behind the slide. A simple solution is to paint the back of the slides with a diffuse black paint like Krylon UltraFlat Black or Stove Black. | ||
| Howard Mark (hlmark) Senior Member Username: hlmark Post Number: 440 Registered: 9-2001 |
I agree with those commenters who recommend that you use a diffusely reflecting reference, and I also have another recommendation to make. Since your bacterial smear is the sample that is presumably what you want to be the reflector, but is mounted on a glass slide, that slide should be backed with something highly absorbing, to prevent light that misses the sample from being reflected back from any other objects in the instrument that might be present underneath the sample (e.g., screw heads, etc). and reflect light back. Howard \o/ /_\ | ||
| Art Springsteen (artspring) Senior Member Username: artspring Post Number: 38 Registered: 2-2003 |
Hi all, The Foss instruments are typically set up with a diffuse standard because the geometry is 45:0 (45 degree incidence, near normal collection). A mirror would give essentially zero as a measured reflectance. So typically, one uses a diffuse reference, like Fluorilon/Spectralon/sintered PTFE, packed PTFE, or a ceramic. A bacterial smear on a glass slide is sort of diffuse, I would think. The rule of thumb is to measure diffuse against diffuse, specular against specular. But NEVER diffuse against a specular reference. Roughening a mirror with carborundum paper, David? OY!!!! You just totally ruined an expensive reference mirror!!! | ||
| David W. Hopkins (dhopkins) Senior Member Username: dhopkins Post Number: 196 Registered: 10-2002 |
Hi John and Tony, Tony is right, and I expect that a mirror would send most of the light right back into the monochromator, which would possibly introduce some problems with stray light and actually give relatively little light actually falling on the detectors, and particularly, no hot spots, as I suggested earlier. Best regards, Dave | ||
| David W. Hopkins (dhopkins) Senior Member Username: dhopkins Post Number: 195 Registered: 10-2002 |
Hi Tony, Glad to see we are in agreement! No one had replied when I started my reply.... I have to laugh at your spelling of Don's name. As former editor of the Journal, I would think you would have caught that typo. It could have been worse! ; - ) Best wishes, Dave | ||
| David W. Hopkins (dhopkins) Senior Member Username: dhopkins Post Number: 194 Registered: 10-2002 |
Hi John, That is an interesting question. I can understand that for your application, a mirror might be a convenient reference target. Furthermore, it may function reasonably well. However, your target samples are diffusely reflecting, and I suspect that a diffusely reflecting reference target might give somewhat more reproducible results than a mirror. I don't know whether the response of the detectors is possibly influenced by the presence of hot spots that might be present from the imaging of the lamp filament on the surface of the detectors, and any effect on the linearity of the response. I wonder if there is anyone with more intimate knowledge of the Foss instrument who could respond to that issue? I could recommend some experiments to help you decide whether a diffusely reflecting reference would improve your results. My intuition is that you would be best advised to use a diffusely reflecting reference sample that also reflects about the same intensity of light as is being reflected by your samples. Your mirror may be limiting the range of samples you can measure. 1. Compare scans of the energy coming from your mirror reference to a diffusely reflecting reference. Look at the ratio of the two to see if there are any anomalous effects. 2. Evaluate the reproducibility of such reference scans by looking at the WL by WL Stdev of a series of 10 readings of each type of reference. 3. I wonder if you could make your mirror more diffusely reflecting by roughing the first surface using some carborundum paper. I hope that you are just in the early phase of your work, so that you can take the time to understand your measurements more, and optimize your spectroscopic measurements. Best wishes, Dave | ||
| Tony Davies (td) Moderator Username: td Post Number: 261 Registered: 1-2001 |
Hello John, Normally the detector in NIR instruments are arranged to collect diffuse reflection. If you have this arrangement then I would recommend that you use a diffuse reflector and not a mirror. Don Damm might have a different view! Best wishes, Tony | ||
| mathews m john (mathews) New member Username: mathews Post Number: 4 Registered: 6-2011 |
In most of the papers it's mentioned that a reflecting background is needed while taking the reflectance spectra.In our experimental setup we used a mirror as the background. My specimen is a bacterial smear on a transparent glass slide. Will the mirror serve as a good background or should i use a white reflecting background? |