Light/Matter Interactions: Physics of Diffuse Reflectance (Beer's Law "Challenge")
Session Chair: Don Dahm, Rowan University
Summary: Although many have taken advantage of the utility of mid-IR and NIR diffuse reflectance spectroscopy, many are not fully aware of the underlying physics and theories that govern light-matter interactions. This has relevance in many phases of a diffuse reflectance project: including the design of the sampling interface and light collection optics, and spectral data pre-processing.
NOTE: The unique format of this session allows participants to contribute their own data in order to either illustrate unique light-matter interactions, or to poll conference participants to offer explanations for "difficult" data. To contribute or participate in this forum, please contact the session chair, Don Dahm at DJDahm@aol.com
Plans for “Physics” session
It’s the Volume, Folks! - Howard Mark
Milk Coagulation: As the size of the floating particles increases, what happens to the NIR spectra? - Karl Norris and Christian Hansen
Effect of Particle Size on Log(1/R) when Contaminant is a Particle - Karl Norris
Beer’s Law, Scattering Samples, and Time of Flight Spectroscopy -
Kevin Dahm
Perspectives on Remote Sensing - Peter Nastasiak
Contributed Discussions -Attendees
At the time of registration, and throughout the week, we will post questions about theory to test our knowledge base. The idea is to get people talking about the answers to seemingly simple questions, many of which involve poorly understood concepts.
At some convenient spot, we will post data that illustrates poorly understood concepts. I have a few in hand, and am looking for others. In the afternoon, some of us will gather to discuss issues raised by the data with interested parties.
At the Friday session, there will be a chance to discuss these matters in an open forum. However, it is my hope that the above activities will have gone a long way to having educated the attendees on the issues.
There seems to be a fairly good appreciation of the fact that scatter in a sample “increases pathlength”, and that as a consequence absorbance measured in transmission goes up (compared to the same material in the absence of scatter). In Time of Flight or Time Resolved Spectroscopy, illumination is by a pulse of light, and a profile of the intensity of light reaching the detector is recorded as a function of time. Given the constant speed of light, it seems “obvious” that all light reaching the detector in the same time will have traveled the same distance. Then the incorrect reasoning goes, for a sample of a single material, Absorbance at any time will be given by the absorptivity of the analyte (a constant) times the concentration of the analyte (a constant) times the pathlength (determined from the lapse between the time of the pulse and the time of detection). Our session will have:
- A short talk (Kevin Dahm) showing the fallacy of this reasoning;
- A description of an approach (Marc Leger) to determining the absorption coefficient of a material in the face of these problems;
- (Possible: A talk by a representative of a group in Sweden that is at the forefront of applying TOF to particulate materials)
- A discussion of these issues and the potential of time resolved spectroscopy.
Practitioners of NIR tend to shun theory in favor of Chemometrics. While I admire the power of the tools, I believe that future advances will be more likely if we have a better theoretical understanding of diffuse reflectance, and complementary data from other techniques. Both of these add a dimension of “reality” to conclusions and reduce the risk of the Chemometrics running amok.
It is my hope that Remote Sensing activities will be able to shed like on how we might proceed if we are to succeed in an activity like Spectroscopic Imaging of the interior of the intact human body. To this end we have Neil Gallagher available for discussion of the importance of using theoretical framework in calibration; and Pete Nastasiak who has spend a career wrestling “intelligence” out of remotely obtained data and in a sort talk will give an example or two of how this is done in settings quite foreign to some of us.
This topic may come up in discussion, because it is certainly a way of simplifying a complex system, but because of time constraints, will not be an advertised topic for the session.
D. J. Dahm