Fourier transform infrared (FT-IR) spectroscopy has been used to identify unknown materials, determine the quality or consistency of a sample, and determine the amount of components in a mixture. Gary Small, of the Department of Chemistry and Optical Science and Technology Center at the University of Iowa, spoke to Spectroscopy about his work using passive FT-IR remote sensing measurements.
A recent paper of yours described the determination of methanol and ethanol using synthetic calibration spectra in passive FT-IR remote sensing measurements. Can you discuss some of the challenges you faced in that study?
Two challenges were most significant in implementing a successful quantitative measurement with this technique: working around some of the inherent limitations of the approach itself and the general difficulty of collecting good data in the outdoor environment where there are many uncontrollable variables.
A passive FT-IR measurement is one in which you use an emission spectrometer to collect naturally occurring infrared radiance within the field-of-view (FOV) of the instrument. The FOV is typically restricted with a telescope so that you can look at a specific target. Your measured signal contains radiance from whatever sources are present. This could be direct emission from the atmosphere or any objects that are present. The amount of radiance you obtain from a given source depends on its temperature and on an inherent property of the emitting material, termed its emissivity. Planck’s radiation law dictates that the total power of the radiance varies as the fourth power of the temperature, so temperature plays a tremendous role in determining how much signal you receive. This also means that temperature variation leads to significant fluctuation in the acquired signal. A passive measurement of naturally occurring infrared sources will always be challenging because of these temperature effects — weak signals that are subject to fluctuation. Contrast this with a conventional laboratory infrared measurement in which the background source is extremely hot (~1500 K) and held at a constant temperature.
If, as in the work described in the paper, your goal is to detect a volatile organic compound (VOC) present in the atmosphere, the signature of that compound can be present in the acquired data in one of two ways. If the VOC is hot relative to the background, you will see direct light emission at the characteristic vibrational frequencies of the molecule. If the VOC is colder than the background, you will see attenuation of the background radiance at the characteristic absorption features of the molecule. So, the temperature of the VOC controls how its spectral signature is manifest in the collected data — another complication related to temperature.
Now, if you want to make a quantitative measurement of a VOC, you have another temperature-related issue. For a quantitative spectral measurement, whether the signal arises from an emission or absorption phenomenon, you normally key the method on the simple principle that if more molecules of the analyte are present in the FOV, the signal arising from the VOC will increase by some smoothly varying function (often a linear one). However, in the passive infrared measurement, because the radiance increases with temperature, you cannot differentiate between a larger number of analyte molecules or the same number of molecules at a higher temperature.
The second challenge in this work was how to collect good quantitative data in the outdoor environment. To mimic an industrial stack monitoring application, it was important to release VOCs into the environment under realistic conditions. We tried to control as many variables as possible, but the bottom line is that we never really knew exactly what was in the FOV at a given time. You may know how much material you introduced into the bottom of the stack, but you can’t say for sure what the concentration of that material was when you actually viewed it with the spectrometer upon its exit from the stack.
The natural follow-on question is why do this when there are so many complicating issues? The simple answer is that, if you can get it to work, the passive FT-IR method offers the ability to do remote monitoring of VOCs continuously, in real time, and with a single-ended instrument. You don’t have to perform a physical sample collection and you don’t have to have a light source on the other side of the sample. This makes the technique amenable to implementation on a variety of platforms, such as moving vehicles or aircraft.
How did you become interested in combining spectroscopy methods with computer-based data analysis?
I’m old enough to have been just coming out of college when dedicated computers were first being put onto laboratory instruments for data acquisition and control. We were also just starting to get dedicated computers in the laboratory for use in implementing sophisticated data analysis methods. So, it was a time when combining analytical chemistry and computer science methodologies was very appealing. I did undergraduate research with Tom Isenhour when he was at North Carolina and graduate research with Peter Jurs at Penn State, two of the pioneers in applying computer-based data analysis methods to spectroscopy data. So, part of it was just being in the right environments where I could learn the methodology and see its utility.
What are the advantages of your method compared with previous methods used for this analysis?
I touched on this briefly in my answer above. The other methods for quantitative monitoring of VOCs in the atmosphere are either cumbersome to implement or have no ability to produce results rapidly. Physical sampling of the atmosphere requires you to go to the measurement site to obtain a sample, followed by a subsequent analysis. While in principle, this could be done rapidly, it is not really amenable to continuous monitoring. The passive infrared measurement can be implemented from a distance, does not require physical collection of the sample, and requires only a single instrumental package that is amenable to mounting in both static and mobile configurations.
What are some other possible applications in complex sampling environments for this methodology?
We work with U.S. EPA’s ASPECT program (http://www.epa.gov/osweroe1/content/partners/aspect.htm). This is an emergency response unit that deploys to natural or manmade disasters where there are potential releases of chemicals into the atmosphere. They operate an aircraft with passive infrared measurement capabilities. The instruments are mounted in a downward-looking orientation and use the natural infrared radiance emanating from the Earth as the background light source. The methodology we are developing is targeted to their applications in which rapid, remote detection and quantification of VOCs is desired for situations in which a chemical release has occurred (for example, chemical plant accidents resulting from natural or manmade causes, train derailments, and terrorist attacks).
What are the next steps in your research with FT–IR remote sensing?
We are interested in continuing to refine the quantitative analysis methodology to overcome some of the limitations in the work described in the paper, especially for application to the airborne monitoring scenario relevant to EPA’s ASPECT program. Central to our approach is the need to estimate the temperatures of the background and the analyte so that we can calibrate the quantitative response to those conditions. In the work described in the paper, we estimated the temperature of the background by fitting background spectra to Planck’s function and estimated the sample temperature by analysis of one data sample released at a known concentration. Our next goal is to modify these elements of the procedure to be consistent with the airborne application.
Related content:FT-IR Technology For Today's Spectroscopists