The big concern about asphaltene is due to the field observation that sometimes some “asphaltenic” deposit is formed. We feel that, from a logical point of view, two approaches were employed in affording this topic: a “top-down” approach and a “bottom-up” one.

The first one starts from the asphaltene definition: the idea is that by knowing the molecular structure of the asphaltene (or asphaltenes, as they are an ensemble of compounds) it would be possible to “dominate” the argument. Here starts the search for an average (what kind of an average?) molecular structure of asphaltene and the need for better standards  for both separation and characterisation.

This approach is focused on asphaltene, while the oil is generally seen as an accident to be eliminated. Even if this search for the “arabic phoenix” would be successful (and we doubt of this), we wonder if it would be fruitful: in our opinion the objective should be to predict (that is,  model) the phenomenon, and it seems that the molecular parameters (molecular weight distribution, etc.) alone won’t be useful to this purpose.

Two fundamental aspects are underevaluated in this approach:        

• the fact that instability comes from interaction between asphaltene and the surrounding medium, whatever it is;
• the availability of well developed thermodynamic (or colloidal) tools to deal with such systems.

For the first point, at present there are different pictures of the asphaltene status in the oil, which correspond to different behaviour description: LLE, SLE, sterically stabilised colloid, etc. Incidentally, note that, if it were a classical phase equilibrium, the asphaltene standard definition (including n-paraffin removal and asphaltene washing steps) would imply that all information is lost about the equilibrium concentrations in the original medium!

The second aspect is a more general one, and concerns the present status of phase equilibrium modelling: to obtain a model able to predict the phase equilibrium, properties of compounds like, e.g., critical (or pseudo-critical) properties are needed. Now, it is hard to believe that such properties can be estimated on the basis of chemical characterisations. Correlations would be necessary in order to make this jump, which would require a very extensive and difficult experimental work.

Alternatively, a molecular dynamic simulation could be performed but, in spite of the extensive use of it in many condensed matter physics, this kind of approach seems to be applicable only to relatively simple systems.

The “bottom-up” approach is phenomenological: it starts from the observation of the phenomenon of interest and tries to gain insight into the physics in order to model and “predict”. As someone may deduce from the previous paragraph, we adopted this one: starting from literature and experimental evidences we tried to make a tool to give practical answers to some field questions.

We agree with dr. Fotland about the misleading power of asphaltene definitions, but we do believe that it’s worth while not to loose the whole previous experience. The peculiar behaviour of asphaltene is that the separation/deposition happens to occur due to the pressure decrease; besides, asphaltene deposits from field result to be analytically similar (though not identical) to the ones obtained by means of standard lab isolation techniques. Of course field deposits may enclose a lot of oil components , but we do believe that this constitute a good starting point.

Due to the previously discussed inadequacies of the “top-down” approach, we believe that informations needed to model the phenomenon should be obtained from onset measurements. We’ve made (till now) measurement by employing different apparatuses and then by measuring different physical properties: UV-VIS absorbance, NIR-absorbance, electrical impedance, electrical resistance, refractive index. Each time, when setting up a new apparatus, we’ve made a “consistency check” between two devices. In spite of this extensive experience (or perhaps due to this experience), we agree with dr. Buckley that a standardisation effort is necessary.

The overall (starting) properties of the oil should be known. In our method, we deduce both the oil and the asphaltene solubility parameters from a series of onset measurements. While we too use a “solubility parameter model”, we believe that choosing to measure this parameter means that a previous choice has been made about the model to use.

So, selecting it could not be enough general for information-exchange purposes. However, we’d like to have comments both on the characterisation parameters and methods.

Any way, the focus should be on the onset conditions determinations. In our opinion, the main objective should be an accurate and unambiguous measurement of the amount of n-paraffin to add in order to achieve onset conditions. This is twice important: from an empirical point of view, it allows a better description of the phenomenon; from a thermodynamic point of view, it could be useful in order to judge between different physical models. Last but not least, of paramount importance is the definition of (standard) criteria to account for the kinetics: a high flow rate of paraffin addition can dramatically change experimental results (and then model predictions).

A few words at the end on the samples. Of course, in order to directly (“in vivo” !) studying the phenomenon, live oil samples would be necessary. But these samples have prohibitive costs and are often of doubtful significance. Besides, in our everyday experience, due to other constraints, it is often impossible to obtain live oil samples. These considerations induced us to use stock tank oil samples to find our model parameters; of course, if asphaltene separation was a classical phase equilibrium, this procedure would not allow to obtain right data. The fact that, instead, our model is able to do reliable estimates seems indicate that the phenomenon is a different one. However, we think that this aspect should be better investigated.