The kinetics of the precipitation and flocculation of asphaltenes can impact the operation of solvent deasphalting processes such as paraffinic froth treatment as well as the potential for deposition from destabilized crude oil. It has also been reported that the solvent content at which asphaltenes precipitate (the onset condition) is a time dependent property and may be far lower than reported because the precipitation kinetics are slow. However, it is possible that the slow kinetics are caused by oxidation or oxygen catalyzed polymerization.
The objectives of this thesis are to: 1) assess the role of oxygen in asphaltene precipitation kinetics; 2) determine the kinetics of asphaltene precipitation from diluted bitumen near the onset of precipitation, and; 3) investigate the kinetics of asphaltene flocculation near the onset of precipitation. The data were collected from a Western Canadian bitumen diluted with n-heptane at 23oC. The n-heptane content at the onset of precipitation was determined using a combination of optical microscopy and gravimetric analysis. Asphaltene precipitate yields were measured over time gravimetrically, and asphaltene flocculation was investigated using a Lasentec particle size analyzer.
The onset of asphaltene precipitation (defined as the first appearance of 0.5 μm asphaltene particles) in n-heptane diluted bitumen occurred at 57.5 wt% heptane. When heptane was added above the onset amount, precipitation yields increased at an exponentially decaying rate over 48 hours. In air, the yields increased and the onset n-heptane content decreased at a slow rate for at least a month. In nitrogen, the yields reached a plateau and there was no change in the onset condition after 48 hours. Once oxygen related artifacts were eliminated, the precipitation kinetics could be described with first order kinetics requiring an ultimate yield and a rate constant. The steady state yield is a function of the solvent content and has been modeled with regular solution theory. A single rate constant was found to apply to all of the data.
The effects of heptane content and shear rate on the kinetics of asphaltene flocculation were investigated and modeled using a population balance based model adopted from Rastegari et al. (2004) and Daneshvar (2005). Unfortunately, the particle size analyzer could not detect the particle size distribution in unadulterated bitumen near the onset. Therefore, data was collected at dilutions well above the onset and the possibility of extrapolating the model to near onset conditions investigated.
The main parameters governing flocculation are the particle concentration, the fractal dimension, and the magnitude and ratio of the flocculation and shattering rates, kf/ks. The fractal dimension of settled flocs was found to be 2.3 ±0.1 (with a slight dependence on the n-heptane content) and this value was assumed to apply for suspended flocs. At the dilutions considered in this thesis, adding n-heptane decreased the individual particle concentration and therefore decreased flocculation. An increase in the shear rate resulted in a decrease in the asphaltene volume mean diameter, while the total number of asphaltene flocs in the mixture remained relatively constant. This observation indicates that asphaltene flocs likely become more compact at higher shear rates.
The flocculation model was fitted the data at different dilutions. A constant value of the shattering rate was sufficient to fit the data. Both the fractal dimension and the reaction rate ratio could be correlated to the mass fraction of n-heptane. The effect of shear rate on the kf/ks ratio at various heptane contents was investigated for two different bitumen samples. The kf/ks ratio appears to decrease with increasing Reynolds number up to 12000, above which the ratio becomes constant. The decrease in the ratio corresponds to the transition zone, while the plateau corresponds to the establishment of a turbulent zone.
While the model parameters could be correlated to n-heptane content, a long extrapolation with a significant a change in parameter values is required to estimate the behavior near the onset condition. Hence, the flocculation rates near the onset conditions cannot be determined from the available data with any confidence. It is recommended that data be collected for toluene-bitumen feedstocks where the onset occurs at higher n-heptane contents so that a more reliable extrapolation can be performed. If successful, both precipitation and flocculation rates can be predicted and used to guide process design and operation.