Synthetically Generated Calibration Spectra for Combustion Analysis by FTIR

Jorge E. Perez, CIC Photonics

David Griffith, University of Wollongong

PDF Version

Challenges for FTIR Analysis in Combustion Applications

• Combustion applications generate high amount of water vapor and CO2
• Their absorption spectra strongly overlap with other species of real interest for the combustion analyses
• The difference of concentration level of water vapor and CO2 with respect other combustion products can range from 1 to 5 orders of magnitude
• The high intensity spectral features of water and CO2 at high concentration work vapor like narrow band pass filters to block all the available light in the spectral regions they occupy
• The less intense features of water and CO2 that overlap with other combustion species will normally show a non-linear behavior in their absorption spectrum

Overcoming High Concentrations of H2O and CO2

• Reducing gas cell pathlength
• Pros: Absorption intensity can be reduce to more manageable levels
• Cons: The system sensitivity of the system gets reduced directly proportional to the decrease in path length
• Use of a sample gas conditioning system
• Pros: Moisture is reduce to manageable levels and instrument sensitivity is maintain
• Cons: Some absorption species of real interest may get absorb or reduce by the sampling conditioning syste
• Use of H2O and CO2 spectra at about the same concentration present in the sample gas as part of the quantification analysis
• Pros: Maintain instrument sensitivity
• Cons: Hard to have a calibrated spectral database for water and CO2 big enough to have optimal water quantification analysis
• NL-MALT Software - Synthetically Generated Calibration Spectra
• Pros: Same as above but it overcomes its cons
• Cons: Limited to the compounds present on the HITRAN Database

MALT Features

• Applicable when sample components are known and do not interact
• cf. factor analysis methods
• Require HITRAN-type line parameters or quantitative library spectrum for each component spectrum
• Wide dynamic range
• No Beer’s Law restrictions
• Spectra are fitted in transmittance scale, regions of strong absorption can be included
• All spectral data points have approx. equal noise weighting
• No calibration standards or training sets required

MALT – ILS Parameters

• Following parameters and effect that are included in the instrument function and can be fixed or fitted:
• Wavenumber shift > spectrum shift
• Resolution > line width / shape
• Apodizing > function line shape
• Field of view > line width / shape
• Effective apodization > line width / shape
• Phase error > line asymmetry
• Asymmetry parameter+ > line asymmetry
• Misalignment > line asymmetry

Conclusion

• NLM is an exceptional tool to overcome
• heavily overlapped spectra
• It correctly compensate for standard non -linear behavior of high concentration absorption spectra
• NLM is limited to use those gases present in the HITRAN database

Future Work

• Investigate the effectiveness of MALT while using highly accurate PNNL Vapor Phase Infrared library

References

• The HITRAN 2004 molecular spectroscopic database, L.S. Rothman, et al., Journal of Quantitative Spectroscopy and Radiative Transfer, vol. 96, pp. 139-204 (2005)
• Synthetic Calibration and Quantification Analysis of Gas Phase FT-IR Analysis, David W. T. Griffith, Department of IR Analysis, David W. T. Griffith, Department of Chemistry, University of Wollongong. Applied Chemistry, Applied Spectroscopy, Volume 50, Number 1, 1996, pp 59 70
• PNNL Vapor Phase Infrared Library: nwir.pnl.gov