In that case, if I buy a 35 ppm, 17 L CH4 gas cylinder and then start passing the gas into the chamber, the volume will start reducing. So, if it becomes 12 L (considering 5L of CH4 being passed), the ppm will still be 35 ?

Correct.

I think I am mixing up in between the 'Number of Molecules' and 'Parts Per Million'. I thought, number of PPM is representing the concentration of a gas indicating the number of molecules present in that gas.

Not quite. ppm is proportional to number of molecule per unit volume.

Now, if the ppm doesn't change with the volume of the gas, then what exactly is changing with the volume, the number of molecules present in that gas?

Number of molecules present in the volume of gas you have used or have remaining.

What is the way to determine the amount of gas (for this experiment it's CH4) present in air and how to determine whether this amount is increasing or decreasing?
Thanks

That is what your sensor is for. Volume X ppm is proportional to the number of molecules.

Let's say, you have 100 red beans and 100 black beans mixed together. The mixture is 50% red beans = 500,000 ppm red beans. If you take an ounce of beans from the mixture, say 1 ounce = 20 beans, you will likely have 10 read beans and 10 black beans in the volume you take. The mixture you took is still 50% red beans. The remaining 180 beans is also 50% red beans ( i.e., 90 red beans and 90 black beans). The composition of the mixture doesn't change, but the actual number of beans you have changes. Of course, for such small numbers of beans, probability will pay a role and the numbers won't be exactly what is calculated. For gas molecules, probability still plays a role, but the numbers are so large you will never notice a difference.

 

Let's say, you have 100 red beans and 100 black beans mixed together. The mixture is 50% red beans = 500,000 ppm red beans. If you take an ounce of beans from the mixture, say 1 ounce = 20 beans, you will likely have 10 read beans and 10 black beans in the volume you take. The mixture you took is still 50% red beans. The remaining 180 beans is also 50% red beans ( i.e., 90 red beans and 90 black beans). The composition of the mixture doesn't change, but the actual number of beans you have changes. Of course, for such small numbers of beans, probability will pay a role and the numbers won't be exactly what is calculated. For gas molecules, probability still plays a role, but the numbers are so large you will never notice a difference.

Excellent example! Thanks a lot for making things easier for me. Below is what I have understood (please correct me again if I'm wrong):
PPM is indicating the proportion/ratio of a particular gas particles w.r.t. air/any other solution.
So, a 2.5% CH4 means that 25,000 CH4 molecules are present in a mixture of total 1,000,000 (air+CH4) molecules.
So, this ratio will be same for 1L to 17L of CH4. Now, the amount of CH4 present in any amount of gas (for example 5L) can be calculated as:

Amount of CH4 (number of molecules) = 25000 x 5 = 125,000 molecules

So, if the volume of gas mixture (considering CH4 mixed with Air as the sensor will be monitoring the air) remains constant, the increase or decrease in CH4 concentration can be represented in terms of ppm as for that fixed volume, the number of CH4 molecules is changing.

Thanks

 

Excellent example! Thanks a lot for making things easier for me. Below is what I have understood (please correct me again if I'm wrong):
PPM is indicating the proportion/ratio of a particular gas particles w.r.t. air/any other solution.

So, a 2.5% CH4 means that 25,000 CH4 molecules are present in a mixture of total 1,000,000 (air+CH4) molecules.
So, this ratio will be same for 1L to 17L of CH4. Now, the amount of CH4 present in any amount of gas (for example 5L) can be calculated as:

Amount of CH4 (number of molecules) = 25000 x 5 = 125,000 molecules

Correct, except the ACTUAL number of gas molecules is MUCH LARGER.

Earlier on, I referred to a mole (gram molecular weight) of methane. The molecular weight of methane is 16 (12 for carbon + 4 hydrogens). Therefore, 16 grams of methane contain Avogadro's number of molecules or approximately 6.022 X 10^23 molecules (some of the standard values have changed slightly with adoption of SI units). Therefore, at standard temperature and pressure (pressure = 1 atmosphere) , 22.4 L of gas = 1 mole = 6.022 x 10^23 molecules.

Using those numbers for the volume of gas in that cylinder, your 17 L will contain (17L/22.4L/mole ) x 6.022x10^23 molecules/mole = 4.57 x 10^23 molecules. Note how the units work out: L/(L/mole) = mole; mole x molecules/mole = molecules

EDIT: My calculation just above could have been less ambiguous. That 4.57 X 10^23 molecules is the total number of molecules in 17 L of mixed gases at STP, of which only 35 ppm are methane.[/edit]

In SI, I believe the standard of pressure is 100 kPa, which is slightly lower than the old "atmosphere" and so the volume for 1 mole is slightly larger or 22.7 L. (Described in detail in Wikipedia under "standard pressure" and Avogadro's law.)

Aside from that "small" correction,* I think you've got the right idea.

John

* 6.022 x 10^23 is 4.82 x 10^18 times more than 1.25 x10^5 (i.e., 125,000).

 

Once again Thanks for such elaborate explanation and correcting me.

Now please correct me again if I'm mistaken:

According to the theory, 1L of CH4 is containing C= (4.57 x 10^23) /17 molecules.
For 35 ppm, the ratio of CH4 to air is:
35/(1M - 35)
Again, for 'C' number of CH4 molecules, the total number of Air molecules (A) present in the mixture is:
35/(1M - 35) = C/A
=> A = C/ (35/(1M-35))
Hence, in 1L of CH4+Air mix, total number of particles present is (C+A).
Now, 1M particles has 35 CH4 molecules.
So (C+A) particles has:
X = (35 x (C+A))/1M CH4 molecules.

So for (C+A) amount of particles, the change of ppm is: Z = X/35 times.

 

Aside from that "small" correction,* I think you've got the right idea.

The previous reply may seem quite confusing and not clear what I'm trying to achieve.

I'm actually trying to calculate how I would be able to convert my sensor's output (current) to approximate change in PPM.
For example, for 35 ppm of CH4, If the sensor's output is 2 mA, then 4 mA output would indicate around 70 ppm of CH4.

 

The previous reply may seem quite confusing and not clear what I'm trying to achieve.

I'm actually trying to calculate how I would be able to convert my sensor's output (current) to approximate change in PPM.
For example, for 35 ppm of CH4, If the sensor's output is 2 mA, then 4 mA output would indicate around 70 ppm of CH4.

That is correct, if the response is linear and increases with concentration.

Background:
Beer's law (Beer-Lambert's) states that absorption of light at a particular wavelength is proportional to the concentration of the substance (methane in your case) , the light path (usually in cm), and a constant called the extinction coefficient at that wavelength. The relationship can be written as this:



Source: https://en.wikipedia.org/wiki/Beer–Lambert_law

Practice:
In practice, one does not usually know the absolute intensity of the incident light being absorbed by the compound. A way around that is to use a ratio measurement. That is, the ratio of the incident intensity to the transmitted intensity of light. Since it is a ratio, the absolute intensity does not need to be known. There are two ways to express absorption. The most common today is in units of "absorbance" (written as "A" in the equation shown above). An equivalent term is "optical density" (written as "OD" or "O.D."). They are the same, just different names. The relationship of A to the light absorbed is simple:

A = log(incident intensity/transmitted intensity)

Absorbance (A) can be substituted in the above equation (A= εlc) to calculate concentration.*

What needs to be done:
I could not find in the datasheet for the device I think you are using (post #19) nor in the datasheet for the break-out board that is available the exact relationship of the data with the absorbance. For example, is the extinction coefficient in molar units or PPM? What is the conversion from absorbance to a digital value? There should be a calibration chart showing concentration of methane vs. number transmitted over the interface. In the absence of that, you could make your own calibration chart. Even if you find a chart, I strongly recommend that you do your own calibration. You may well find it is not linear.

There are reasons to expect non-linearity, for example, you are measuring a mixture, not just methane in air. The light being used is not a single wavelength (i.e., not monochromatic). Water vapor also shows absorption in the band of wavelengths being used.

John

*Another expression is % transmitted ("%T"). One can do the algebra and show that A = -log(%T). In quantitative work such as this, use of absorbance (A) is far more common.

 

That is correct, if the response is linear and increases with concentration.

I could not find in the datasheet for the device I think you are using (post #19) nor in the datasheet for the break-out board that is available the exact relationship of the data with the absorbance. For example, is the extinction coefficient in molar units or PPM? What is the conversion from absorbance to a digital value? There should be a calibration chart showing concentration of methane vs. number transmitted over the interface. In the absence of that, you could make your own calibration chart. Even if you find a chart, I strongly recommend that you do your own calibration. You may well find it is not linear.

You are right. Their datasheet doesn't help much. Even their full form datasheet of the sensor doesn't mention anything about the calibrated value of the sensor unlike the SPEC sensors. I contacted them but they said to follow their Calibration and Determination of the linear coefficient guides (I have attached all three docs here for your reference). Unfortunately, I do not have access to such equipment and facilities where I would deal with pure nitrogen for calibration purpose (as instructed in their guideline).
It seems that the relationship isn't linear. In that case, if my calculations (in post #24) is correct, then I will try to take several data points and form a curve (o/p of the sensor Vs concentration of the gas) to get an idea about the changing characteristic of CH4 concentration. From that, may be I would be able to calculate a coefficient which would provide approximate relationship.

The above image is the availability of CH4 from GASCO website. If I use the one which is balanced with Air, can I expect the same characteristics of CH4 in the open air (where the sensor will be operating in real time)?

Regards

 

Hopefully, you have the right sensor, namely something with a dash number like -150A-xxx where xxx (10 to 100) is in PPM. The data from ezPyro is all for full percentages and well above the range we have been discussing. You can change your project to include flammable mixtures, if you have not bought the sensor yet. That might work better for your project. Table 2 in the second attachment (Determination of linear...) shows data in absorbance units.

Unless you are pretty good at soldering (i.e., using reflow), you will be better off getting a break-out board instead of the bare detector.

I agree completely about the need to do your own calibration curve. One way to get various mixtures is with a water trap as mentioned early one. Namely, get a graduated cylinder (500 ml to 2 l), fill with water and invert in a pan of water. Add your gas from the bottom and then dilute with air or nitrogen to a final volume. The problem is that methane is reasonably soluble in water. That will be hard to get around perfectly. You could take the water you intend to use, shake it with your methane mixture to get it to the same concentration, then use that in your mixing apparatus. Avoid bubbling the gas mixture through it and work fast. If you just used ordinary water and worked fast, that might be good enough too.

I assume this is for a school project. If so, determining methane concentration is a gas mixture is quite easily done with gas chromatography at room temperature -- if anyone has such an instrument.

 

Hopefully, you have the right sensor, namely something with a dash number like -150A-xxx where xxx (10 to 100) is in PPM. The data from ezPyro is all for full percentages and well above the range we have been discussing. You can change your project to include flammable mixtures, if you have not bought the sensor yet. That might work better for your project. Table 2 in the second attachment (Determination of linear...) shows data in absorbance units.

Unless you are pretty good at soldering (i.e., using reflow), you will be better off getting a break-out board instead of the bare detector.

I agree completely about the need to do your own calibration curve. One way to get various mixtures is with a water trap as mentioned early one. Namely, get a graduated cylinder (500 ml to 2 l), fill with water and invert in a pan of water. Add your gas from the bottom and then dilute with air or nitrogen to a final volume. The problem is that methane is reasonably soluble in water. That will be hard to get around perfectly. You could take the water you intend to use, shake it with your methane mixture to get it to the same concentration, then use that in your mixing apparatus. Avoid bubbling the gas mixture through it and work fast. If you just used ordinary water and worked fast, that might be good enough too.

I assume this is for a school project. If so, determining methane concentration is a gas mixture is quite easily done with gas chromatography at room temperature -- if anyone has such an instrument.

Thanks for the suggestion. I have already bought the sensor with a breakout board (https://au.mouser.com/ProductDetail/Pyreos/EPY12221-B1?qs=qSfuJ%2Bfl/d5fViuM9%2BpmbQ==).
So for the time being, I will have to proceed with this sensor.
If I proceed with calibrated CH4 gas balanced with Air for a certain PPM (the one you suggested earlier), will it be alright or I'll have to mix it with N2 or water for testing?

Another concern is, so far I have seen most of the valve regulators measure the pressure of the cylinder. If I can't find any such regulator which also measures the volume, then would it be wise to convert the pressure into volume (as you explained earlier) to get the approximate idea that how much gas is being passed?

Thanks

 

Thanks for the suggestion. I have already bought the sensor with a breakout board (https://au.mouser.com/ProductDetail/Pyreos/EPY12221-B1?qs=qSfuJ%2Bfl/d5fViuM9%2BpmbQ==).
So for the time being, I will have to proceed with this sensor.

Glad you got the BOB. Those pinless packages with exposed pads can be very hard to solder by hand.

I misunderstood the data you posted for gas mixtures. We had been talking so much about ppm that I incorrectly thought the dash number applied to sensors with different ranges. On re-reading, I see that is not the case. It appears your sensor is for full percentages (e.g., 0% to 100%) and the dash number applies to gas mixture. Be sure to get one that is within the range of the sensor.

If I proceed with calibrated CH4 gas balanced with Air for a certain PPM (the one you suggested earlier), will it be alright or I'll have to mix it with N2 or water for testing?

Since the measurement is based on IR absorption, not thermal conductivity or other measures, I don't think it much matters whether the makeup gas is nitrogen or air. Neither absorbs IR in the 3.3 um IR region. Nitrogen might be a little safer.

Another concern is, so far I have seen most of the valve regulators measure the pressure of the cylinder. If I can't find any such regulator which also measures the volume, then would it be wise to convert the pressure into volume (as you explained earlier) to get the approximate idea that how much gas is being passed?

I think any needle valve controller (like is used for propane torches) will work. With the small volume of standard you will have, doing mixtures by flow may waste a lot of gas. That is why I suggest making the mixtures my volume. The main problem with volume done by water displacement will be water vapor (it absorbs in the same IR region) and solubility in water. Using higher concentrations of methane will lessen the effect of solubility.

After a little more thought, rather than water displacement, you can use a solid plug (like an old-fashioned glass syringe). That will solve both problems created by water. You might be able to get a plastic syringe of 50 or 100 ml that will be tight enough, or make your own with a plug out of some plastic foam that will slide in a calibrated cylinder/pipe. Plastic plumbing pipe (say 2 " PVC) should be OK. If you can get something that is translucent or clear, that would be easier to read, but is not necessary. I lean toward doing it by volume with some sort of homemade displacement cylinder.

If you want to do it by flow rate, a floating ball gauge is all you need. You do not need a pressure controller. The needle valve will work. Here's what they look like. Notice the one on the right just uses a needle valve for flow control.


I have several collected over the years. A local gas supply house (welding or medical gases) may have something they can loan or just give you.

Once you get over the mechanical part, you are going to have fun with the digital part. I would probably attack that part first. For a cheap test gas, propane or butane from a cigarette lighter will probably give you a signal. That IR band at 3.3 um (3000 cm-1) is from C--H bond stretching. Any hydrocarbon will have absorption in that band. That's why O--H in water interferes. The O--H band is at a higher wavenumber/frequency, but the filter used is relatively broad band. Here's an example of O--H and C--H in that region of the IR spectrum:

 

Thanks a lot John for all the insights and suggestions. I will get back to you once I receive the gas and start testing it.

Regards