The molecular weight of heavy aromatics or paraffin's in a crude a key parameters in solubility models. The solubility depends exponentially on this parameter. Actually you should fit these weights to solubility measurements. Actually Asphaltenes are defined as a solubility fraction (obtained by precipitation by adding a certain alkane to the crude followed up by washing with a polar solvent such as toluene). For asphaltenes due to aggregation the molecular weight depends strongly on the "purity" of asphaltenes. Asphaltenes monomers have a typical molecular weight between 300 and 600. These are strongly polar aromatic sheets that tend to stack up. This can lead to molecular weights measured by freezing point depression of up to 30.000. Resins are a quite similar to asphaltenes and aggregate with asphaltenes, but they block the polymerisation process in crude limiting the aggregate size (one asphaltene plus two resins) to molecular weight between 800 and 2000.

When the solubility of asphaltenes in crude is of the order of 0.1% (such as for very light cruces) a difference of a factor 2 in the molecular weight changes the solubility by a factor 1000. This indicates that there is no way to define the molecular weight except by using a thermodynamic model which you should fit to the experimental data on solubility available. So claiming that you can determine the molecular weight of SARA fractions is quite stupid. 

I would be interested to know what you wish to achieve with your molecular weights because I think as input they will be highly suspect because of definitions and measurement techniques. There may be better ways of reaching your objective.

I fully agree that the relevant question is:

"What do you intend to do with the information"?

As in the original question reference was made to a thermodynamic model, I stressed that the molecular weight is an input to a thermodynamic model rather than an answer. Furthermore when considering heavy polar fractions, the molecular weight one measures depends strongly on the procedure followed to extract this fraction. Any thermodynamic model would be a fit of some experimental data. So another way to rephrase your comment is: 

"What kind of experimental data do you want to fit with a model?"

You say that the molecular weight is an input data. However, I was asked to find a way to determine the molecular weight of each SARA fraction. I've been searching many papers and information sources but any of them clearly states how to calculate it based on other input data (i.e composition, Tb, among others). I was wondering if someone with a wider knowledge in this field knew something about a model to do this. I'm sorry if you have misunderstood me.

I understand you have been asked to provide the mol weights of each fraction, but this is difficult to achieve and the answers you get as input to your modelling will probably be useless, so I was asking what the ultimate purpose of the modelling was in order to see if there was a better way of getting there! Molecular weights in themselves are of course only part of the characterisation of a fluid, for EoS work you would presume critical properties (Critical pressure, temperature), acentic factor. There are correlations that relate these inexactly.

Characterisation may be easier tuning to experimental data rather than hunt out molecular weights in isolation.

I would recommend you ask your supervisor what is behind the need for the molecular weights to tackle the issue from a broader front.

I do not know of a method that would separate the fractions and permit molecular weigh measurements on each fraction. Whole oil samples or distillation cuts can be measured by freezing point depression, heavier molecular weights can be separated by gel permeation chromatography, but these techniques split the oil differently from the SARA fraction definitions. This is why you won't find the papers you are searching.

Try looking up research on asphaltenes, in particular try searching on Advances in Asphaltene Science and the Yen–Mullins Model

Brian.

I fully agree with Brian Moffatt. This is what I was trying to say.

When you get an unsolvable question from a supervisor, go back and ask what his final aim is!

There are some methodologies to separate SARA fractions through liquid chromatography. Indeed, asphaltenes may be firstly precipitated with an alkane like heptane or hexane (it depends on the metodology the author use), once they are little soluble in these solvents.

Further steps to determine MW distribution of each fraction may be characterization through Mass Spectrometry or other analytical techniques.

The molecular model of an Asphaltene fraction precipitated by adding an alkane will entirely depend on the extraction procedure followed and the method used to measure the molecular weight.

Molecular weights measured by freezing point depression/vapor pressure can vary between 500 and 50 000 (due to aggregation). So such molecular weight measurements have very little aded value for a prediction of Asphaltene precipitation.