Converting hydrocarbons to methane

Discussions about sample preparation: extraction, cleanup, derivatization, etc.

13 posts Page 1 of 1
I am attempting to use FTIR for measuring non-methane hydrocarbons in gas samples but the NMVC is to be reported as methane or ethane equivalents. I am looking for a converter that I can run gas through that will break down each larger hydrocarbon into methane equivalents. Does anyone know of anything that may be of use?
Are you trying to chemically convert generic hydrocarbons to methane or convert the response on the FTIR to methane equivalents? The latter is easy, the former might be a challenge.
Steve Reimer wrote:
Are you trying to chemically convert generic hydrocarbons to methane or convert the response on the FTIR to methane equivalents? The latter is easy, the former might be a challenge.



I don't care if it converts the hydrocarbons to methane chemically (I know this is the reverse of what is usually done), I just want my FTIR to be able to give me an accurate representation of non-methane hydrocarbons as methane equivalents.

I am attempting to get certified under the NFPA 1989 firefighters grade breathing air using my FTIR equipped with a gas cell. I am able to accurately measure the methane concentration of gases using the unique methane peak at 1305 cm^-1. Larger hydrocarbons are difficult to quantify by FTIR because their unique peaks start to get in the range of water vapor making them extremely inaccurate. To combat this, I was subtracting the methane concentration from the total hydrocarbon region around 3000 cm^-1. The problem is I can't distinguish the difference between say, octane and butane using this method. So if I find there is 1 ppm of "other" hydrocarbons, this differs greatly between propane and octane when I have to report as methane equivalents. If you have any way to help me with this problem it would be ridiculously appreciated.
So what you are not doing is measuring the C-H stretch and reporting the result as if it were methane, instead you are attempting to find a region where you can isolate different hydrocarbon bands and determine the concentration of the true components. Is that correct?
Steve Reimer wrote:
So what you are not doing is measuring the C-H stretch and reporting the result as if it were methane, instead you are attempting to find a region where you can isolate different hydrocarbon bands and determine the concentration of the true components. Is that correct?



I was trying to isolate each of them individually at first but the only one that works well is methane.

My most recent attempt goes as follows:

-Measure methane at 1305. Lets call it 5 ppm.

-Measure the total hydrocarbons using the C-H stretch at 3000. Lets call the result 10 ppm.

-I then subtract the methane, 5 ppm, from the total of 10 ppm, which gives me 5 ppm of other hydrocarbons. For some standards I know only have methane and ethane, I could then assume the 5 ppm remaining was ethane, double it and call the "non-methane hydrocarbons as methane" to be 10 ppm. But if I don't know it is only methane and ethane this falls apart, nor did it work very well when it was just methane and ethane.

I assume you are getting at something along the lines of calibrate the C-H stretch at 3000 to just be total carbon equivalents? I assumed that this wouldn't work. But it seems to me you are saying I can run a 10 ppm octane standard and an 80 ppm methane standard, and the two spectra should have roughly equal area at the 3000 cm^-1 C-H stretch, correct?
The intensity of the C-H stretch signal should calibrate to methane equivalents but not to intact higher hydrocarbons, since moles of C-H is a function of moles of carbon but the slope of moles of carbon to moles of intact molecule varies with MW, the only complication being that the relative C-H content declines with carbon number (because the terminal carbons have three -Hs each). Unsaturation will also be a source in inaccuracy of course.a

Peter
Peter Apps
Peter Apps wrote:
The intensity of the C-H stretch signal should calibrate to methane equivalents but not to intact higher hydrocarbons, since moles of C-H is a function of moles of carbon but the slope of moles of carbon to moles of intact molecule varies with MW, the only complication being that the relative C-H content declines with carbon number (because the terminal carbons have three -Hs each). Unsaturation will also be a source in inaccuracy of course.a

Peter



Thank you for the reply. So with the "The intensity of the C-H stretch signal should calibrate to methane equivalents but not to intact higher hydrocarbons" sentence, you are saying that I should not do the method I mentioned in the previous post, or that the methane equivalents won't work for higher hydrocarbons? I understand the unsaturation and terminal carbon effects will give some source of error, but is this something you have tried or will I just have to test it experimentally?
Also, do you think it would be possible to make it more accurate by calibrating it to the specific number of C-H bonds instead of just to methane? Say I use ethane and methane as my cal gases, I could call each ppm methane as "4" and each ppm ethane as "6"? Doesn't seem to be the intuitive way to do it, but thought it may be worth exploring.
This all rather depends on what "methane equivalent" means. Is it based on moles of carbon, or moles of (C+4H), or how much energy you get if you burn it !?

You can calibrate your signal vs number of C-H bonds as long as you know the identity of the calibrant (because as you observe the number of C-H per molecule differs for different compounds). The problem with using that calibration is that you do not know the identity of the hydrocarbons, so you cannot simply calculate from moles C-H to moles hydrocarbon.

Peter
Peter Apps
Peter Apps wrote:
This all rather depends on what "methane equivalent" means. Is it based on moles of carbon, or moles of (C+4H), or how much energy you get if you burn it !?

You can calibrate your signal vs number of C-H bonds as long as you know the identity of the calibrant (because as you observe the number of C-H per molecule differs for different compounds). The problem with using that calibration is that you do not know the identity of the hydrocarbons, so you cannot simply calculate from moles C-H to moles hydrocarbon.

Peter



Thank you again for all your help. Looking at methane equivalents as number of carbons specifically.

I tried calibrating to the C-H bonds but I was getting less accurate results than when I just calibrated to total carbon atoms. I have a few gas cylinders of methane in balance nitrogen, as well as a few of ethane in balance nitrogen. Using the 1305 cm^-1 peak for methane and varying the pressures slightly to simulate different concentrations I am able to get a very good correlation methane curve (0.999 from around 5 ppm methane up to 40 ppm methane).

Using the exact same scans that give me the 0.999 correlation at 1305 to try to calibrate for total hydrocarbons in the C-H stretching region around 3000 cm^-1 only gives me a correlation of around 0.971. And this is before adding in the ethane which drops it to around 0.940. I have been playing around with different resolutions and apodization functions to try to find the best fit but there is so little literature online about FTIR gas analysis it has mostly just been a trial and error process.

Currently I am running an MCT detector with KBr beam splitter and MIR source. 32 scans, 0.1 resolution with a 0.5 cm^-1 aperture @ 4000 cm^-1. I know that the box-car apodization is traditionally used for gas analysis but I was concerned with the large ripple effect associated with this apodization. I have been trying hapf-ganzel with 8 zero fill the past few days. At first I was making the baseline very far from the actual region of interest but recently I have noticed it works better if the baseline is basically the endpoints of the region of interest.

I definitely feel like I am moving in the right direction but any advice on how to improve my performance would be extremely helpful. Thanks again to both of you for all the advice you have given me.
This;

"Using the exact same scans that give me the 0.999 correlation at 1305 to try to calibrate for total hydrocarbons in the C-H stretching region around 3000 cm^-1 only gives me a correlation of around 0.971. And this is before adding in the ethane which drops it to around 0.940."

is a puzzle. When you calibrate C-H vs methane ppm (I am assuming this means mole fraction, can you confirm) you get a good correlation, but when you calibrate the same readings against total hydrocarbons in the same methane standard the fit is poorer ? How is this possible ? - if methane is the only hydrocarbon then total hydrocarbons = methane. The mole fraction of methane is exactly the same as the mole fraction of total hydrocarbon - you should be putting the same numbers into the calibration and getting the same answer out.

Peter
Peter Apps
Peter Apps wrote:
This;

"Using the exact same scans that give me the 0.999 correlation at 1305 to try to calibrate for total hydrocarbons in the C-H stretching region around 3000 cm^-1 only gives me a correlation of around 0.971. And this is before adding in the ethane which drops it to around 0.940."

is a puzzle. When you calibrate C-H vs methane ppm (I am assuming this means mole fraction, can you confirm) you get a good correlation, but when you calibrate the same readings against total hydrocarbons in the same methane standard the fit is poorer ? How is this possible ? - if methane is the only hydrocarbon then total hydrocarbons = methane. The mole fraction of methane is exactly the same as the mole fraction of total hydrocarbon - you should be putting the same numbers into the calibration and getting the same answer out.

Peter



So my methane standards were prepared gravimetrically, so I assume the 10 ppm and 25 ppm are weight/weight. The ethane standards do not actually have any identification of if it is a mol fraction, v/v, or w/w but I have emailed them to find out. This could cause some of my issues when I add the ethane to the curve but as you said, when I am using the exact same scans and inputting the exact same concentrations, it should give me the same correlation regardless of which methane region I use. I believe a lot of this has to do with finding the best area to use and the best baseline locations. I have noticed changing these can vary the results a lot.

So for the 1305 region that is working well I:

-filled gas cell up to around 3 PSIg with the 10 ppm methane cylinder (prepared gravimetrically). Converted to PSIa by adding 3 to 14.7 (I am on the shore so pretty much as close to 0 altitude as possible) and divided this total by 14.7. This gives me a correction factor of 1.20408 so my 10 ppm cylinder is now acting as a 12.04 ppm standard. I will drop the pressure by about 0.5 PSIg down until about -0.5 PSIg to get a few different data points.

-I then do the same with the 25 ppm methane cylinder (this was also prepared gravimetrically but has 18% oxygen and the rest balance nitrogen instead of the entire balance being nitrogen).

-From this I have about 12 points covering the range I am interested in.

This image is my Methane curve and a scan showing the region I am using.

Image




This last image is one I tried today and it is actually coming out very well but I am still a little concerned with the left part of the ethane region considering I can't get a very good baseline. This image is all unprocessed. I will try to zero-fill and apodize it and test it with some PTs I have and come back with an update.

Image
You absolutely have to find out what "ppm" means - usually in gas work if refers to mole fraction (which is the same as volume fraction near as dammit). The standards having been made gravimetrically does not mean that their compositions are expressed as mass fractions.

To see why this is important, calculate how many moles of hexane are in any given weight and the volume that the hexane vapour will occupy, and then do the same calculation for methane, ethane, butane and pentane.

If you take your 3000 calibration for methane only is the poor linearity because the points fall on a curve, or are they scattered around the straight line ?

Peter
Peter Apps
13 posts Page 1 of 1

Who is online

In total there is 1 user online :: 0 registered, 0 hidden and 1 guest (based on users active over the past 5 minutes)
Most users ever online was 1117 on Mon Jan 31, 2022 2:50 pm

Users browsing this forum: No registered users and 1 guest

Latest Blog Posts from Separation Science

Separation Science offers free learning from the experts covering methods, applications, webinars, eSeminars, videos, tutorials for users of liquid chromatography, gas chromatography, mass spectrometry, sample preparation and related analytical techniques.

Subscribe to our eNewsletter with daily, weekly or monthly updates: Food & Beverage, Environmental, (Bio)Pharmaceutical, Bioclinical, Liquid Chromatography, Gas Chromatography and Mass Spectrometry.

Liquid Chromatography

Gas Chromatography

Mass Spectrometry