Absorption: Path Integrated
Interferometric spectrometers such as FTIRs represent a high-resolution spectroscopic solution that could benefit many researchers of in-cylinder properties. When compared to a grating spectrometer at equal spectral resolution, a typical interferometric spectrometer has a throughput at least 1,000 times greater while also having the advantage of broad spectral coverage. FTIRs operate by monitoring the resulting interferogram produced by varying the relative path length between two mirrors divided by a beamsplitter. The Fourier-transform of that interferogram produces a spectrum. However, because FTIR mirror scan speeds are much slower than most in-cylinder combustion events, FTIRs typically are only used for exhaust gas measurements.
We have adapted FTIRs to operate in-cylinder. Although FTIRs do not have sufficient time resolution to gather crank-angle resolved engine spectra for a single engine cycle, by collecting data from many engine cycles, we are able to obtain crank angle resolved data. We have measured gas temperature, fuel and H2O concentrations in a spark-ignition engine using a commercial FTIR. Source light from a heated ceramic globar is first passed through the FTIR’s Michelson interferometer prior to being introduced into the engine. Optical access to the engine is provided by an optical spark plug probe (OSPP). The OSPP contains two fibers used as transmit (Io) and receive (I) engine access. Both the engine and the mirror in the Michelson interferometer are allowed to free-run as piston position, mirror position, and interferogram signal are logged. The engine piston and interferometer mirror move in an uncorrelated fashion, so that after many minutes of repeatable engine operation, the superposition of many engine cycles results in a complete interferogram at each piston position. By taking the Fourier-transform of the interferograms, the results are ultimately cast as absorption spectra versus piston position.
The measured absorption spectra are then compared to spectroscopic databases to produce the gas temperature and mole fraction results shown in the movie. These measurements demonstrate the technique’s ability to measure multiple species simultaneously over a wide range of temperatures and pressures.
For more details see papers such as:
Rein, K.D., Sanders, S.T., Lowry, S.R., Jiang, E.Y., and Workman, Jr, Jerome J., "In-cylinder Fourier-Transform Infrared Spectroscopy," Measurement Science and Technology, 19, 043001, 2008.
Wavelength-agile Temp / H2O measurements in an HCCI Engine
In this study, methods are developed for measuring temperature and water mole fraction in an HCCI engine using absorption spectroscopy. These measurements need to be at a sufficient crank angle resolution (i.e., very fast), and therefore fast measurement techniques need to be developed. This involves improving wavelength-agile light sources. Wavelength-agile light sources (usually lasers) produce light such that the light's wavelength (color) varies in time. For example, if the light source was in the visible color range, the wavelength-agile source would first output blue light, and as time went on - the color would turn to green to yellow to red (from a low wavelength to a high wavelength). If done quickly (typically 5 µs), this allows spectra to be measured quickly over a wide range of wavelengths, in a straightforward way (the relationship between time and wavelength is known).
Light from a wavelength-agile source was pitched through an engine that had windows fixtured so that gas could be viewed through TDC. The transmitted light was collected into a photo detector in order to measure light intensity. The transmitted light intensity (I), along with the initial intensity (Io) could then be converted into an absorption spectrum: a plot of absorption coefficient (‘kl’ in Beer's Law) versus wavelength (see the absorption spectroscopy tutorial).
The measured absorption spectra are then compared to a database of known spectra (the 'HITEMP' database) at a given temperature (T), water mole fraction (X), and pressure (P). It is therefore possible to extract these variables (T, X, or P) from a single absorption spectrum. The top panel of the above movie shows the measured spectra (pink) and the known spectra at a similar condition (blue). The bottom panel shows the results from the measured spectra: the temperature (red, center) and water mole fraction (blue, right). The pressure (blue, left) is measured from a pressure transducer independent of the spectra. These measurements are useful in that they are non-intrusive (i.e., they don't disrupt the fluid dynamics), they are accurate in a wide variety of temperature and pressure conditions that are normally difficult to measure, and they are sufficiently fast which allows for a new level of combustion research and understanding to be achieved.
For more information contact Prof. Sanders
OH Absorption Measurements During HCCI Combustion
The HCCI combustion process produces very low emissions of NOx, but at very light loads the emissions of unburned hydrocarbons and CO can be quite high. It has been suggested that the chemical reactions quench at light loads when the peak temperature gets below a value of ~1500 K because the OH radical concentration becomes too low. OH is known to be the primary species responsible for the oxidation of CO.
In this study the OH radical concentration was measured quantitatively using time-resolved absorption spectroscopy. A broadband Deuterium lamp, which produces light in the ultraviolet portion of the spectrum, was delivered to the engine through a fiber optic cable. The light was collimated and passed through the engine. A second fiber was used to collect the transmitted light and the output of this fiber was put into an imaging spectrometer, which separated the light into its component wavelengths. A fast-kinetics camera that collects rows of data very quickly was used to capture the absorption spectra. The raw transmission data were corrected for beamsteering and broadband absorption, and using Beer's Law the absorbance was calculated as a function of wavelength. In the figure, plots of the absorbance are shown as a function of the crankangle time denoted by the vertical line in the heat release plot. The characteristic features of absorption centered at 283 and 308 nm correspond to the spectral signature of OH uniquely.
Measurements of the OH absorption were acquired for a wide range of operating conditions and engine-out emissions were measured as well. It was confirmed that the observed increase in the engine-out CO emissions coincided with the disappearance of observable OH absorption. The detection limit of the optical system was on the order of 1 part per million.
Acknowledgements: This work was performed by Sean Younger and was supported by Ford Motor Co.
For more information contact Prof. Ghandhi