Hydrocarbon Gas Streams

Application Note: Hydrocarbon Gas Streams (#3076)

GUIDED WAVE’S product line includes a near-infrared (NIR) dual beam spectrometer (NIR-O™) and a multi-wavelength photometer (ClearView®  db) that are well suited for measuring many hydrocarbon gas streams. Guided Wave’s NIR instrumentation can make both qualitative and quantitative measurements in gas streams using long path gas cells coupled to our analyzer systems using fiber optics. This note will discuss the use of Guided Wave hardware and software tools for the measurement of hydrocarbon gases under various conditions. NIR measurements can be used in real time directly in process monitoring to save both time and money when compared to traditional methods.

Background

The NIR region of the electromagnetic spectrum allows the use of the overtone and combination bands of the C-H, O-H, and N-H fundamentals. By measuring the NIR spectra of a series of gas samples of known concentration, a quantitative model can be developed which will allow the measurement of future samples based only on their NIR spectrum. Guided Wave analyzer systems use fiber optics to allow the sample probe to be located in hazardous, remote locations away from the spectrophotometer itself.

Most NIR spectroscopy is done on liquid samples where the hindered rotation blurs any rotational fine structure resulting in broad peaks. However, gases have well resolved rotational-vibrational (ro-vibrational) fine structure which can be seen in high resolution. Most NIR process spectrometers and photometers are low resolution devices hence do not resolve the fine structure but instead record the spectrum as integrated P, Q, and R bands. When dealing with gases, one must always keep in mind that the real spectrum underneath is a complex series of narrow sharp peaks.

The molecular density of gases and vapors is significantly lower than that of liquids, hence much longer probe pathlengths are required. Good quality NIR spectra of gases can be obtained with probe pathlengths in the range of 25 to 100 cm. Not only is the density lower, but it varies with temperature and pressure according to the Ideal Gas Law. While the spectrometer measures the number of molecules in the beam, it may be necessary to measure temperature and pressure as well to exactly quantify the gas mixture.

Another important aspect of gas phase spectroscopy is that the spectra are pressure and temperature dependent. Increasing pressure will broaden the rovibrational lines due to pressure broadening. For first order effects, the integrated absorption observed with a low resolution spectrometer should remain the same as pressure is increased, however, there may be second order effects that have to be considered and added to the modeling efforts.

Temperature can pose a bigger problem. Many multi-atom molecules will have low lying bending mode bands whose Boltzmann thermal population will increase as the temperature is increased. These combination “hot” bands will appear in the spectra robbing intensity (population) from the fundamental ground state transition bands. It is best to calibrate a gas phase measurement at one temperature and maintain that temperature for the process measurement.

Finally, not all molecules exhibit an IR or NIR spectrum since a change in the dipole moment is required for a ro-vibrational IR spectrum to occur. Homonuclear molecules such as N2, O2, Cl2, and Hdo not have dipole moments hence do not have IR spectra. Fortunately, most hydrocarbons, and many other important gases such as NH3, CO2, H2O, etc. do have strong IR spectra and as a result NIR spectra. Unfortunately many common sulfur compounds have weak IR spectra and correspondingly weaker NIR spectra.

Experimental

The vapor phase NIR spectra of a group of different hydrocarbons were measured between 1550 and 1900 nm using a Guided Wave Model NIR-O process spectrometer. Figure 1 shows the absorbance spectra of several short chain alkanes collected using an on-line process probe with a 25 cm pathlength. Methane has the most unique spectrum, but this is to be expected as it is the simplest and has the smallest moment of inertia. This study concentrates on the first overtone region of the C-H stretching bands near 1700 nm. Figure 2 illustrates the differences between ethylene and propylene. Olefins typically have a strong band at 1625 nm.

This is also seen in Figure 3 in the vapor phase spectra of several other olefins.

Figure 1 - Vapor Phase NIR Spectra - Short Chain Alkanes

Figure 2 - Vapor Phase NIR Spectra - Ethylene, Propylene

Figure 3 - Vapor Phase NIR Spectra - Olefi ns

Hydrocarbon Vapors Common Concern

A common concern of hydrocarbon vapors, especially in ambient air mixtures, is the concentration relative to the Lower Explosion Limit (LEL). Table 1 shows the detection limits for several light hydrocarbons under the following conditions: Instrumental Signal-to-Noise Ratio of 40 μAU, a minimum concentration equal to 10 x the SNR or 400 μAU, a pathlength of 1 m, and a neat sample condition at 20°C. Note that the detection limits are all significantly below the respective LEL.

A further series of tests were conducted with butane in nitrogen to study the self and foreign gas broadening issues. Collisional energy transfer also contributes to changes in absorption coefficients. Neat butane to 1 atm was measured as well as various fixed levels buffered with N2 to 1 atm. The concentration matrix studied is shown in Figure 4.

Table 1 - Vapor Detection Limits

Figure 4 – Butane Concentration Matrix

Analysis

Simple data analysis was used to prepare a model for butane in N2. Prediction of neat butane concentrations up to 100 kPa is straight forward. A four wavelength model produced the regression results shown in Figure 5. Quantification is to about ±0.46 kPa.

Looking at butane concentration at constant total pressure, again in an N2 matrix, butane shows reasonably linear relationship to absorbance for dilute mixtures. This is shown in Figure 6 where mixtures that are more than 80% butane deviate significantly as the broadening and energy transfer mechanisms transition from foreign gas to self-broadening. This nonlinearity was modeled with a quadratic equation. Regressing all of the non-neat butane data using a two wavelength model plus linear and quadratic terms in total pressure produces the regression results seen in Figure 7. Concentrations were predicted to ±3 kPa.

Figure 5 - Butane results

Figure 6 - Butane at constant total pressure

Figure 7 - Butane results, modeling pressure non-linearity

Discussion

The data presented demonstrates that neat vapors and gases can be measured by NIR spectroscopy to better than 500 ppm-m. The data also shows that foreign gas collisional energy transfer significantly alters the absorption coefficients. The foreign gas induced spectral shape changes are small but measurable (and modelable).

Conclusion

We have shown that simple hydrocarbon gases and vapors have unique NIR spectra that can be easily measured with excellent sensitivity. For simple systems, quantitative mixture analysis can be accomplished by picking a few strategic wavelengths and performing a simple linear regression. A cost effective solution for simple systems using discrete wavelengths is the ClearView db.

For more complex mixtures that involve pressure broadening, linear methods can still be used but more calibration data will be required. While not demonstrated here, mixture analysis can be done by using the neat spectra of the individual gases as a basis set to deconvolve the unknown spectrum, thus extracting the individual concentrations. A NIR-O spectrometer would be more suitable for these complex systems.

The vapor phase measurements discussed here are both fast and reliable utilizing Guided Wave analyzer hardware and software tools. Results are available in real-time (seconds) for complex streams. For more detailed information regarding system specifications please contact a Guided Wave sales or technical
specialist.