Gas Chromotography

[Danh]

First, I acknowledge that I'm beyond my skill level on this subject, hence the question.

There seem to be two UOA methods for determining fuel dilution of engine oil: 1) inferring a dilution % based on an oil sample's flashpoint (e.g. Blackstone), and 2) gas chromotography (e.g. Polaris). Gas chromotography seems to be regarded as the more precise method and typically shows a higher level of fuel contamination than the flashpoint method.

As I understand gas chromotography, it analyzes and categorizes sample vapors based on molecule length, or basically the number of carbon atoms. Gasoline has fewer carbon atoms than engine oil and is segregated on this basis. But the difference in the number of carbon atoms in gasoline isn't radically different from engine oil and I'm wondering if the GC method is as precise as advertised. For example, doesn't normal mechanical shearing essentially "shear" carbon atoms from engine oil molecules, thus creating the opportunity for these sheared molecules to appearing same general category as some gasoline molecules?

This is relevant to me because, as I've posted here before, my 2015 Honda CRV has serial UOAs from Polaris showing fuel dilution of >5%. One UOA from Blackstone showed a flashpoint of 360 F and an inferred fuel dilution of about 1%. More than 5% dilution seems extreme, and looking through Blackstone UOAs this would translate into a flashpoint of 260 F or so, which seems absurdly low. As Honda can't find anything wrong with my CRV I'm left to wonder if (hope?) the measurement method is to blame. And of course, how much precision can we expect from a $20-$30 analysis?

As noted, I'm far beyond my level of competence here so be gentle.

 

[Tom NJ]

Gas chromatography can separate molecules by various attributes, but the most common for lubricants is separation by molecular weight and/or polarity. Under the conditions used to analyze motor oils, gasoline will flash out of the column very quickly and well before the smallest molecules of the base oil and additives, and below sheared base oil as well. Therefore isolating and measuring the fuel dilution by GC should be more accurate, certainly better than flash point. Flash point by ASTM D92 has a reproducibility of +- 15°F, not nearly sensitive enough to measure a few percent of fuel.

Here is a lay man's description of the GC method I wrote here many years ago:

Gas Chromatography

Tom NJ

 

[kschachn]

Well the average molecular weight of a viscosity index improver is going to be much higher than either gasoline (C8) or motor oil. I don't know specifics about VII polymers but somewhere around a million wouldn't be surprising.

You aren't going to shear short chain length hydrocarbons such as octane or even C20-C50 motor oils like you will a million carbon chain polymer. An example of this is how multi-vis oils may exhibit a viscosity change due to shear whereas a mono-grade does not.

And you don't "shear carbon atoms" from the molecules, if it happens at all it is a cleaving of a very long chain. An automobile engine isn't going to shear apart what are relatively short chain hydrocarbon molecules.

 

[Leo99]

I've run GC methods in pharmaceutical labs to look for residual solvents but not with any petroleum products. GC is extremely sensitive.

GC analysis provides a chromatogram. Here is one for example. What this shows is they mixed 63 separate compounds together and the GC separated them all out after a 90 minute analysis. Basically, you inject a very small aliquot of the sample and the GC column will separate them and then a detector detects them.

The GC doesn't separate into oil and gas. You have to look at the compounds it finds in your sample and then know where they came from. If they detect xyz compound in your sample and xyz comes from gasoline and not oil then they deduce that fuel must have gotten into your oil. The time that the compounds come out in the chromatogram can shift a little and it can lead to errors in identification. You can see some of the peaks are very close together and your sample will not have all the peaks so it's easy to misidentify a peak. For $30, a chemist is not looking at every chromatogram; they are letting the chromatography software identify the peaks.

Did they provide you with any chromatograms?

 

[JHZR2]

I'd recommend reading this thread in full. It was some work I did at the University when doing my PhD.

https://bobistheoilguy.com/forums/ubbthreads.php/topics/456671#Post456671

It is chromatograms of fuels and additives and some discussion that will help.

 

[MolaKule]

Originally Posted By: JHZR2
I'd recommend reading this thread in full. It was some work I did at the University when doing my PhD.

https://bobistheoilguy.com/forums/ubbthreads.php/topics/456671#Post456671

It is chromatograms of fuels and additives and some discussion that will help.

Excellent read!

 

[kschachn]

I don't think the OP is directly asking about the details of gas chromatography, he's asking if oil molecules can be sheared down to the C8 range and "fool" the CG results into thinking more gasoline exists in the sample than is actually there.

Which has been shown to not be the case by the shear stability of mono-grade oils.

 

[Danh]

Originally Posted By: kschachn
I don't think the OP is directly asking about the details of gas chromatography, he's asking if oil molecules can be sheared down to the C8 range and "fool" the CG results into thinking more gasoline exists in the sample than is actually there.

Which has been shown to not be the case by the shear stability of mono-grade oils.


Thanks, that is the question. In my case Mobil1 0w-20 with a new 100C viscosity of 8.7 had a UOA viscosity of 7.0-7.2 after 1,000-2,000 miles. Some of this may be fuel dilution of course, but if mechanical shearing is going on as well, could GC be "fooled" into interpreting some of the oil as gasoline? And if this isn't how mechanical shearing works keep in mind my background is finance...

I appreciate the responses.

 

[JHZR2]

Originally Posted By: kschachn
I don't think the OP is directly asking about the details of gas chromatography, he's asking if oil molecules can be sheared down to the C8 range and "fool" the CG results into thinking more gasoline exists in the sample than is actually there.

Which has been shown to not be the case by the shear stability of mono-grade oils.


This is why I pointed OP to my old post. Gas Chromatography methods can be implemented with different temperature rise parameters in different sequences. For example, start at x degrees, ramp at some rate to some temperature for some time, then when at y degrees, do something else. This can compress or expand the time it takes for various chemical specied to elude the column.

But what I was hoping would be observed from my old thread is that there is a drastic difference in elusion time between the light, low boilers, and the heavier stuff.

Notionally the magnitude of each peak aligns to a concentration or fraction of each species. As you can see, gasoline is a large conglomeration of different low boilers, and in the methods I used for looking even at diesel fuels, the gasoline fractions were a pretty overlaid chunk of peaks. If a method was optimized for even heavier boilers, like motor oil, all the light end stuff would start to look like one massive peak.

Notionally, this is where common sense of the analyst would come into play. If a massive, overlaid peak was observed early on, another technique or techniques would be utilized to see if better detail could be had, such as by doing ASTM D7593 versus a method optimized for looking at the heavier species. To optimize for motor oil, a whole lot of higher molecular weights could be massed into nearly one group...

At some point whether it is fuel or sheared oil that is in the right MWT range is irrelevant, its either there or not. If flagged, other methods may be better to identify what exactly is going on... But one should also note that not all oil would shear to C8 or whatever. That's probably the biggest determination... Sheared oil would have another distribution of heavier boilers.

Therefore fuel dilution would give a bimodal distribution of fuel-range peaks and oil-range peaks. Sheared motor oil would either smear the overall distribution of the high boilers, or be bimodal but with a far broader distribution of the lower boilers that are the sheared species.

Tom, Mola, etc. - please peer review my claim in bold.

 

[Leo99]

Originally Posted By: JHZR2
Originally Posted By: kschachn
I don't think the OP is directly asking about the details of gas chromatography, he's asking if oil molecules can be sheared down to the C8 range and "fool" the CG results into thinking more gasoline exists in the sample than is actually there.

Which has been shown to not be the case by the shear stability of mono-grade oils.


This is why I pointed OP to my old post. Gas Chromatography methods can be implemented with different temperature rise parameters in different sequences. For example, start at x degrees, ramp at some rate to some temperature for some time, then when at y degrees, do something else. This can compress or expand the time it takes for various chemical specied to elude the column.

But what I was hoping would be observed from my old thread is that there is a drastic difference in elusion time between the light, low boilers, and the heavier stuff.

Notionally the magnitude of each peak aligns to a concentration or fraction of each species. As you can see, gasoline is a large conglomeration of different low boilers, and in the methods I used for looking even at diesel fuels, the gasoline fractions were a pretty overlaid chunk of peaks. If a method was optimized for even heavier boilers, like motor oil, all the light end stuff would start to look like one massive peak.

Notionally, this is where common sense of the analyst would come into play. If a massive, overlaid peak was observed early on, another technique or techniques would be utilized to see if better detail could be had, such as by doing ASTM D7593 versus a method optimized for looking at the heavier species. To optimize for motor oil, a whole lot of higher molecular weights could be massed into nearly one group...

At some point whether it is fuel or sheared oil that is in the right MWT range is irrelevant, its either there or not. If flagged, other methods may be better to identify what exactly is going on... But one should also note that not all oil would shear to C8 or whatever. That's probably the biggest determination... Sheared oil would have another distribution of heavier boilers.

Therefore fuel dilution would give a bimodal distribution of fuel-range peaks and oil-range peaks. Sheared motor oil would either smear the overall distribution of the high boilers, or be bimodal but with a far broader distribution of the lower boilers that are the sheared species.

Tom, Mola, etc. - please peer review my claim in bold.


Good post. I'm thinking that a proper GC method can differentiate between sheared oil and gasoline. And I agree that other methods will be better but that will likely require lower temps and more shallow ramps up in temps and longer dwells. That takes more time. Time is money...

 

[bunnspecial]

Something else worth mentioning is that a gas chromotograph is only part of the picture. You need to stick a detector on the end of it to actually see what is coming out.

For general hydrocarbon analysis, a flame ionization detector is common because it's inexpensive, sensitive(to flammables-everything else is invisible) and repeatable.

Increasingly, though, a mass spectrometer is becoming the detector of choice. In fact, we have a bit of a divide in the field-someone who considers themselves a chomotographer(like me) will say that a mass spectrometer is a great GC(or LC) detector. A mass spectroscopist will say that a gas chromatograph makes a great inlet for a mass spectrometer.

In any case, the mass spectrum provides a different dimension to the data. Without actually looking at the relevant spectra, I strongly suspect that a lubricating oil that has "sheered" to gasoline range will fragment differently from a typical gasoline hydrocarbon. Gasoline(and similar light hydrocarbons) certainly have a "signature' in a mass spectrometer or GC-MS that will look different from a much heavier lubricating oil.