Combustion Analysis of Nitric Oxide (NO), Acetylene (C2H2), and Hydrogen Cyanide (HCN) in the presence of high concentration moisture based on FTIR spectroscopy.
While the general purpose of using a long path gas cell is to increase the sensitivity of an FTIR Spectrometer, in the case of combustion analysis, it tends to affect the measurements of some components. Up until at least a decade ago, it was common thought, that to achieve detection limits in the low ppm range a long path gas cell was required.
ISO standard 19702 was mainly developed under those premises, where the overwhelming overlapping of moisture’s absorbance spectra with some of the absorption species becomes a hard problem to solve.
The upcoming new revision of ISO19702:2014 states on section 6.7 that:
“6.7 A resolution of 0.5 cm-1 or finer is recommended for minimization of interferences in fire gas analysis, although a coarser resolution (>0.5 cm-1) is allowed as long as spectral interference from overlapping compounds are identified and corrected for by the software.”
Measurement of Nitric Acid(NO)
As an example, Figure 1 shows the absorption intensity of 30% of moisture between 1970 to 1770 cm-1. The moisture absorption is so strong, that 75% of the infrared light has been absorbed in this portion of the spectrum by the moisture itself!. This leaves very few regions that effectively can be used for the measurement of NO. Hence the recommendation of using a resolution of 0.5cm-1 or higher, as a way to mitigate the spectral overlap and to aid the separation of the moisture’s absorption spectra from the absorption spectra of NO and other overlapping species.
It is worth pointing out that most of the absorption peaks of moisture are showing a very strong non-linear behavior, which is hard to compensate for using multivariate quantification analysis algorithms. This problem is independent of the resolution used for the analysis. Even using a finer resolution will not allow the software to separate the spectra if most of the light is absorbed and the absorption peaks behaving extremely non-linear.
Trying to go to longer pathlength gas cells to achieve more sensitivity only makes the problem worst. Is there a better way to solve this problems?
The short answer is: Yes!
The use of FTIR spectrometers with higher sensitivity allow us to explore other alternatives.
Figure 1 Collection Conditions: 0.5cm-1 resolution, 5.1meter pathlength, 1 atm, 120 Deg C, 1, 30% Water, and 1ppm of NO
Over the years, new FTIR spectrometers have been introduced into the market. One of them is the ABB’s MB3000. After many years of research and experience, they were able to develop an FTIR spectrometer with a great increase in sensitivity due to the superior optical and electronic design.
The higher sensitivity of the spectrometer allows, in the combustion analysis arena, the use of a shorter pathlength to obtain the same level of sensitivity as seen in older FTIR spectrometers using long path gas cell.
The main advantage of the use of a short path gas cell is that the absorption intensity of moisture at high concentrations is greatly reduced, and it follows a more linear behavior that is easy to compensate for. (See Figure 2.)
Figure 2 shows that the absorption spectrum of moisture now is more clear, and does not show signs of oversaturation in the same area. In fact, only 12% of the total infrared light in the area has been absorbed by moisture alone. The moisture absorption in the area still strong, some of the stronger peaks do display slightly non-linear behavior, but it can be compensated for by multivariate quantification analysis algorithms. This alone makes the use of lower resolutions absorption spectra perfectly feasible to separate the overlapping of NO from the Moisture, as stipulated by ISO19702:2014 section 6.7 – “… coarser resolution (>0.5 cm-1) is allowed as long as spectral interference from overlapping compounds are identified and corrected for by the software.”
Figure 2 Collection Conditions: 2cm-1 resolution, 0.15 meter pathlength, 1 atm, 120 Deg C, 1, 30% Water, and 1ppm of NO
Another related problem of using a long path gas cell is that it restricts the useful upper range to measure other combustion gases. The absorption intensity of those gases, at high concentrations grow so much, that they may show a strong non-linear behavior.
In the case of NO, the absorption intensity at high concentration (see Figure 3) has grown so much, that it really shows a non-linear behavior that is its difficult to compensate; for even with the use of a non-linear correction curve (See Figure 4). At first sight, the non-linear correction curve seem to go through all the calibration data points, but it can be observed that the curve is being over compensated. This overcompensation of the non-linear compensation curve is more evident when we look at the curve in a logarithmic scale (See Figure 5), and it can be seen that it will increase the reading errors of NO at low concentrations. Again, in the case of NO, a high resolution spectrum is not helping to alleviate this problem.
Figure 3 Collection Conditions: 0.5cm-1 resolution, 5.1 meter pathlength, 1 atm, 120 Deg C, 1, 30% Water, and 0.5% of NO
Figure 4 NO's non-liner correction curve for the range of 0 to 0.5% in a 5.1m pathlegth gas cell and 0.5cm-1 resolution
Figure 5 NO's non-liner correction curve (Logarithmic scale) for the range of 0 to 0.5% in a 5.1m pathlegth gas cell and 0.5cm-1 resolution
Once again, the solution to this problem arrives with the use of a spectrometer with a higher sensitivity, and the use of a short path gas cell. The absorption intensity of NO at 0.5% is greatly reduced (See Figure 6 and 7).
The non-linear correction curve is practically linear, and it will not introduce error in the proper quantification and correction of the concentration of NO.
Figure 6 Collection Conditions: 2cm-1 resolution, 0.15 meter pathlength, 1 atm, 120 Deg C, 1, 30% Water, and 0.5% of NO
Figure 7 NO's non-liner correction curve for the range of 0 to 0.5% in a 0.15m pathlegth gas cell and 2cm-1 resolution
Figure 8 NO's non-liner correction curve (Logarithmic scale) for the range of 0 to 0.5% in a 0.15m pathlegth gas cell and 2cm-1 resolution
Measurement of Acetylene (C2H2) and Hydrogen Cyanide (HCN)
The ISO standard 19702:2014 also states on Annex G.3 that:
“G.3 It is important to acknowledge that FTIR determines the infrared absorption of a gas or mixture of gases over a wide range of wavenumbers (typically 400 cm-1 to 4 500 cm-1 (for KBr windows) and 650 cm-1 to 4500 cm-1 (for ZnSe Windows). The infrared absorption bands for the main fire gases are given in Annex I. It should be noted that many other gases found in fire atmospheres, e.g. hydrocarbons, also absorb infrared radiation over this region. It is therefore rare for a fire gas to absorb infrared within a given wavenumber region without interference from other fire gases in the same region. However using a spectrophotometer resolution of 0.5 cm-1 or higher, provides an acceptable solution for avoiding overlap of absorbances for fire gases such as HCN and acetylene (C2H2), and nitric oxide (NO) and water, that would not be possible at lower resolutions.”
But they also added the following note to the passage:
“NOTE 1: Measurements using spectrophotometer resolutions as low as 4 cm-1 have proved acceptable for measurements of fire gases, as long as spectrally interfering (overlapping) compounds can be addressed in the mathematical evaluation of the spectra .”
As mentioned earlier, the bulk of the ISO19702 document assumes the use of a long path gas cell for the analysis, and focuses on overcoming the strong absorption intensity of moisture. As we have shown, a shorter path length and a more sensitive spectrometer provided a better solution since the absorption spectrum of moisture can be easily compensated by the software since it exhibits a more linear behavior.
This also applies to the overlapping of C2H2 with HCN. Both of the species overlap heavily with each other, and the use of a long path gas cell will make them behave very non-linear at high concentrations. The recommendation of the standard is to only use high resolution spectrometers (See figure 9). With this heavy overlapping, even at high resolution, the whole spectrum cannot be used to measure HCN. The best option is to use the region between 3400 to 3360cm-1, and provide a non-linear compensation curve. To measure C2H2 it is better to use the region between 3240 to 3200cm-1 while using a non-linear correction curve. Both non-linear correction curves should be as closer to linear as possible. I must point out, that the use of a lower resolution can also be use for the same regions while using a long path gas cell.
Figure 9 Overlapping of the absorption spectrum of HCN and C2H2 at high concentration
The magnitude of this problem gets greatly reduced by use of a short path gas cell and a highly sensitive spectrometer.
The absorption intensity of both gases will exhibit almost linear behavior at high concentrations, and a multivariate quantification analysis algorithm should be able to separate them easily. No need for a high resolution spectrum here (See figures 10 and 11)