Qualitative Lubricant Analysis Question

[dwiggins]

I need to be able to distinguish between different lubricants and the method and column that was being used before I got the responsibility for this gives nothing but a very broad non descript chromatogram. I realize that I will probably never be able to a really good separation of the different components within the lubricants but I need some help. The method used was:

column: 1% dexsil 300 on 100/120 mesh Chromosorb WHP packed column
Injection temp: 225 C
Det: (FID) temp 320 C
oven: 155 C hold for 4 min. ramp to 300 C at 12 C/min

I then modified the method to

injection temp: 300 C
Det temp: 340 C
Oven: 90 C hold for 2 minutes ramp 20 C/ min to 200 C then ramp to 320 C at 6 C/min

and got some separation (a broad "hump" with peaks on top) which was good enough for me

BUT I could not reproduce this

the samples were 1% v/v lubricant in hexane and 1 ul injections

So basically I am wondering a couple of things:
1. Is the problem most likely an injection problem (I am doing manual injections)
2. Is it a method problem or column problem
and more generally
3. Is there any rule of thumb for what the injection temperature should be with respect to the boiling points of the analytes or solvents?

I will try to get the chromatograms posted

[Paulhurleyuk]

I actually used to do this for a global lubricant company in the UK, we used to analyse competitors products to see what was in them. (for us, it was metalworking lubricants)

We used to seperate a sample into aqueous / organic components, then split the organic components down using SPE, then analyse each fraction. If I remember right we used to split a sample into four or five fractions with differing proportions of Hexane / diethyl ether.

The higher hexane fractions were then analysed using simdist GC. Try doing a literature search on Simulated Distillation or SimDist. You might need some specialist columns as it is generally high temp, but it should enable you to fingerprint different mineral oils used in lubricants (although beware, if your samples contain chliorine based lubes, they can foul the column as they don't come off, even up to 600 degrees)

Hope this helps...

Paul.

[dwiggins]

Here are the chromatrograms (hopefully)

the "good" run of lubricant
http://img314.imageshack.us/my.php?image=goodiu0.jpg

the typical run of lubricant
http://img192.imageshack.us/my.php?image=lubtp3.jpg

[Bruce Hamilton]

Hi,

Couple of suggestions, and some comments.

The flash point of lubricant base grades will be around 225C, which is probably telling us that at the temperature the molecules are starting to thermally crack. Residence time in a GC is usually short, but packed columns can provide more activity.

Lubricant base grades can vary an awful lot ( depending on refinery, crude feed etc, etc. ), so you need to obtain some "standard samples" that you use to ensure what you see is real. ASTM sell simulated distillation standards if you want to buy them ( Restek used to have some as well ). Alternative use some base grades.

In my very limited experience, the hedgehog profile is usually indicative of paraffinic oils, whereas the turtle profile is typical of naphthenic base oils.

I've run base grades through an injector at 320C and seen both sorts of profiles that you show for different base grades, but never that much variation on the same sample. If it really is the same sample, then I'd susspect that part of your system is chemically active - either rubbish in the injector or on the column.

I would suspect that you are partially breaking down the oil - either in the injector or on the column. Using standard mixes well help identify if you have a problem. Your temperatures alone should not have caused such a change, but if you have known standards you can investigate the effects.

I used to have injectors at 340C, and still see profiles the same as at 300C. I've never used <300C for the injector with base grades, but was using hydrogen carrier, split injection, and 5m x 0.1 x 0.1 non-polar column.

I would condition the column overnight at the recommended isothermal maximum temperature, and try and obtain a couple of different base grades and repeatably inject them with solvent blanks between injections as well ( to ensure you have a flat baseline ). that's to give you confidence that the system is consistent.

Look at the size of the peak, if necessary add a internal standard
( n-alkane around C30? ) to check your injection volume repeatibility, and to also get an idea of how much of the lubricant base grade is reaching the detector.

Hope this helps, but please keep having fun,

Bruce Hamilton

[dwiggins]

So I have been doing a little bit of searching:

At least some of the lubricants are synthetic lubricants an example being Mobil 629. From what I have found that does not bode well for getting a very good chromatogram (a "hedgehog").

So does anyone have any sugguestions? I should still be able to do simulated distillation right?

[Bruce Hamilton]

So I have been doing a little bit of searching:

At least some of the lubricants are synthetic lubricants an example being Mobil 629. From what I have found that does not bode well for getting a very good chromatogram (a "hedgehog").

So does anyone have any sugguestions? I should still be able to do simulated distillation right?

First of all. What information do you want? The GC profile may give you an indication of the boiling range of the lubricant, and you can also get distinctive profiles of different base grades ( including synthetic, semi-synthetic, and mineral oil ), depending how high a distillation range you want to go to, however the higher you go, the harder it is to spot differences. But will that information meet your needs?.

It can be useful information - provided each if the base grades ( synthetic or mineral ) has a unique profile. As well as the tortise and hedgehog, there's also the armadillo, which usually indicates a mineral oil base grade that is higher boiling and has both paraffinic and aromatic molecules. However semi-synthetic base grades can be quite diverse and represent many reptiles and dinosaurs.

The chromatograms are fingerprints, so hedgehogs aren't good or bad ( unless you're a snail ), but will be representative of the base grade. Provided that you get similar profiles on repeat injections, the boiling range and shape can be very helpful, especially if you have reference base grades ( synthetic, semi-synthetic, and mineral oil. The profile and retention time can be useful, especially with reference materials.

But always remember that GC and HPLC suffer from the same problem, you only see what manages to reach your detector, unlike TLC, Iatroscan, etc. As the boiling temperature increases, the shape ( eg an armadillo's tail as the temperature approaches maximum ) will tell you that there are probably are heavier components that you're not seeing. If you get an area that's small, very little will have reached the detector, and coated your GC injector and column instead.

If you can obtain sufficient profile differences and build up a library of the base grades likely to be used ( based on the application ), you could identify them. Without that information, you may find your time better spent looking as other properties of the material, which will give you some guidance on what base grades were used anyway.

By looking at properties such as viscosity index, oxidation stability, aromatic content, and price, you'll get an idea of candidates. You're unlikely going to find 12H or 30/100 base grade as a major component of a thick goo ( unless emulsified ), or 650/75 as a component of a fluid, or find a true synthetic in a cheap product.

So, only you can decide what information is useful, but I'd recommend that you get some reference samples and have a play.

Bruce Hamilton

[dwiggins]

So...
If I understand you what you are saying the "spikes" are from the paraffins and it is very possible that other lubricants such as aromatic, synthetic, and naphthenic derived lubricants won't have the spikes unless there are some paraffins in it as well but each should still give a distinctive chromatogram shape. Is there any way to perhaps concentrate the possible paraffin "contaminates" in such lubs or would that vary too much from batch to batch and not provide a good indication of what an unknown lubricant is? (I need to determine what the lubricant is that is contaminating our product during the manufacturing process so we can know where to look for a problem as well as to know if we need to destroy the product or not.)

Are there other good and easy ways to help distinguish between different lubricants besides GC in case the chromatograms are too similar to differentiate? (We use FTIR as well but that usually doesn't help too much)

As you can see I am not too knowledgeable about lubricants and I appreciate the help.

[Bruce Hamilton]

So...
Is there any way to perhaps concentrate the possible paraffin "contaminates" in such lubs or would that vary too much from batch to batch and not provide a good indication of what an unknown lubricant is? (I need to determine what the lubricant is that is contaminating our product during the manufacturing process so we can know where to look for a problem as well as to know if we need to destroy the product or not.)

If that's the information you seek, then GC will only be small part of your analytical arsenal. If you know, and can obtain samples of, all the possible lubricant contaminants, you could run them through the GC and build up areference library, however lubricants vary from batch to batch, and even change signifciantly over a year. That's because the lubricants are made from base grades that are traded on the open market, and a base grade of 30/100 could come from different refineries, different crudes.

The chemical composition can vary a lot, as the critical properties are viscosity ( the first number is the viscosity at 40C, the second indicates how little it changes when heated to 100C, so higher is better ). At room temperature 30/100 is similar to olive oil, 650/75 is similar to treacle. Synthetic polymeric Additives are used that can take the second number over 200, ie the engine oil doesn't turn from treacle to olive as the engine warms up, but thins much less, and multigrade oils tend to have lots of such additives. Many formulated lubricants will have 10 or more added chemicals.

Because most base grades are produced by vacuum distillation and some sort of chemical/solvent refining, characterising lubricant base grades is usually performed by looking at the family properties ( aromatics, olefine, cycloparaffins, paraffins ), as well as the general distillation, viscosity, stability properties, and also the unwanted impurities such as sulphur, nitrogen.

However, for most purposes, lubricant base grades are merely organic solvents, and all the goodies are dissolved in it. For example, a gear oil without additives will chew 1mm/100km off the contact surfaces on a truck's differential - I've seen one, very unhappy owner . So additives are added, usually at concentrations of 1 - 10 %, and formulated lubricants are often identified by examining the additives ( often organics containing sulphur, phosphorus, synthetic polymers, and metals ( eg zinc, magnesiun, calcium, boron ) etc.

The normal strategy is to use column chromatography or solvent extraction to separate out the lubricant additives from the base grade and contaminated product, and sometimes they will have a unique components that can be used to identify the contaminant. I'd suggest looking at a mxture of TLC, SPE, HPLC and GC if you want to characterise your lubricant contaminant, focussing initially on TLC with different spray reagents. Why? - because TLC with different sprays can show almost everything, not just what reaches the end of the column.

There are batch to batch variations, and the habit of oil companies is to continually change the formulations ( the same specalist additives often are used by serveral companies, one major global supplier of additives is Lubrizol ). Formulating lubricants is a very complex field ( very price sensitive ), and it's hard for end users to identify unknowns, that's why pharmaceutical and food industries specify just a few types, so they can quickly identify possible sources.

I'd strongly recommend discussing your needs with the technical section of the suppliers of the lubricants that your facility mainly uses. If you want to identify lubricants, you need to narrow down the choices, and they can help. If you want to play, first look up some of the literature, and start by separating the additives and base grades from your product, then look at which is best to characterise.

I'd suggest that mineral, semi-synthetic, or synthetic base grade will tell you you have a lubricant contaminant, but you'll need to look at the additives to find out the contaminant origin, unless you only have a few choices, and you've built a reference chromatogram library that is regularly updated.

Bruce Hamilton