Asphaltene Adsorption, a Literature Review - Energy & Fuels (ACS

Mar 27, 2014 - Thermal Analysis Coupled to Ultrahigh Resolution Mass Spectrometry with Collision Induced Dissociation for Complex Petroleum Samples: H...
25 downloads 3 Views 3MB Size
Review pubs.acs.org/EF

Asphaltene Adsorption, a Literature Review Jeramie J. Adams* Western Research Institute, 3474 North Third Street, Laramie, Wyoming 82072, United States ABSTRACT: Asphaltene adsorption at solid surfaces is a ubiquitous phenomenon that begins within the production well and continues through the entire production chain. It is generally an undesirable phenomenon that causes damage within reservoirs, fouling of pipelines and transportation equipment, and fouling of refining and upgrading equipment. However, by virtue of this phenomenon, problematic asphaltenes may also be selectively removed from petroleum streams to produce partially upgraded petroleum, which has significantly improved handling, processing, and upgrading qualities. This review covers many basic aspects regarding the chemical and physical nature of asphaltenes and sorbents related to asphaltene adsorption.

1. INTRODUCTION Asphaltenes are the largest, densest, most polar, and surfaceactive fraction of oil generally defined by a solubility regime: insoluble in alkanes such as n-pentane (C5) or n-heptane (C7) but soluble in aromatic solvents such as toluene, benzene, or pyridine. The asphaltene fraction is the major contributor to several issues that pester the petroleum industry, especially as lighter conventional crudes are becoming depleted and the vast reserves of heavy, extra heavy, and other unconventional crude oils are becoming major refining feedstocks.1,2 Complications related to asphaltene stability within the supporting oil matrix affect the entire production chain, beginning at the well where they can reduce oil recovery through changes in reservoir wettability, plugging of rock pores,3−6 and asphaltene deposition within wells.7 Destabilized asphaltenes can cause arterial clogging within pipelines and wellbores,8,9 formation and strengthening of oil and water emulsions,10 adsorption onto refining equipment,11 sedimentation and plugging during crude oil storage, corrosion and fouling of production equipment and catalysts,1,11−13 and coke formation.11,14−16 A more detailed review of asphaltene precipitation-related problems encountered at the well on up to surface facilities can be found in Oilf ield Review.8 To lessen steep economic, environmental, and political penalties that can arise from asphaltene flow-assurance issues, one of two common paradigms can be exercised: asphaltene conservation and asphaltene rejection. Asphaltene conservation mitigates the inherent asphaltene content of the incoming feedstock by adding diluents,8,17 deposition inhibiting chemicals or solubilizing agents,18−25 or by blending with stabilizing oils.26 Properly treating oil by these methods requires extensive knowledge and testing of the incoming feedstocks to minimize costs and to prevent exacerbating the problem. In some cases, to meet pipeline density and viscosity specifications significant amounts of diluents are added so that the oil can be pumped to refineries. The diluent can then be recovered from the oil which is often piped back for reuse.27 Alternatively, asphaltene rejection can be achieved through solvent deasphalting,1,28,29 solvent deasphalting combined with emulsion extraction,30,31 ultrafiltration,32 fluid carbon rejection,33 selective adsorption,25,34−50 and selective oxidation of heteroatom-containing species followed by their selective separation.51 These methods © 2014 American Chemical Society

are generally less specific to the characteristics of the incoming oil and view the asphaltenes as a “waste” byproduct that can be converted into coke or other fuel sources to return some of the energy input needed for refinery operations. In the case of heteroatom oxidation, follow up work shows that selectively oxidized heteroatom-rich asphaltenes can be utilized to produce useful high end chemicals through selective reduction.52 Solvent deasphalting can be costly and impractical because of the large volumes of paraffinic or naphthenic solvents needed to precipitate acceptable amounts of asphaltenes and asphaltenecoated particles. These solvents are later vented or distilled during the refining processes and recycled back into the system, often under supercritical conditions. Solvent methods can also inadvertently reject too much polar-type material which are valuable and not asphaltenic. Methods that employ supercritical conditions utilize much less solvent, use light hydrocarbons such as propane, and can be combined with CO2.28,29 Treating oil by these methods inadvertently rejects some of the valuable oil material as it becomes entrapped within, and coprecipitates with, the undesirable asphaltenes. In the case of oil sands, hot water extraction followed by treating the bitumen froths with paraffinic or naphthenic solvents can remove small amounts of asphalteneswhen removing sedimentswhich produces large volumes of undesirable tailings consisting of caustic water, carcinogenic polyaromatics, and organic/asphaltene coated clays and silicas,30 which are often stored in vast reservoirs.53 Solvent deasphalting combined with an emulsion extraction can lead to nearly quantitative asphaltene rejection and oil recovery by using smaller ratios of solvent (pentane) to oil (depending on the feedstock and temperature 0.2−10:1, with an optimum ratio of 2.5:1). This process however requires more characterization of the incoming feedstock and is further complicated by the additional steps needed to treat the oil− water emulsions.54,55 Carbon rejection by filtration is problematic because ultrafiltration membranes can easily become fouled by adsorbed asphaltenes.32 Fluidized carbon rejection is another viable method for deasphalting. It is carried out by spraying a preheated residue into a fluidized-bed reactor at Received: January 29, 2014 Revised: March 26, 2014 Published: March 27, 2014 2831

dx.doi.org/10.1021/ef500282p | Energy Fuels 2014, 28, 2831−2856

Energy & Fuels

Review

Figure 1. Yen−Mullins model.70,71,78,79 A typical asphaltene molecule with a polyaromatic core is shown at the left, which can then associate into nanoaggregates that contain less than 10 asphaltene molecules, shown in the middle. Groups of nanoaggregates ( Ca2+ > Na+ > K+.110 No quantification was given based on surface area. In a later study, it was shown that kaolin had a capacity near 1.4 mg/m2, which was slightly greater than alumina (1.3 mg/m2) and silica (1.1 mg/m2) and significantly more than the metal oxides MgO, ZnO, and TiO2.139 In a study aimed at reservoir rock minerals, the following trend was established (mg/m2): chlorite (1.85) > Fisher kaolinite (1.55) > Berea sandstone (1.43) > Ward’s kaolinite (1.11) > illite (1.08).143 Others have reported that other sources of chlorite have a much smaller adsorption capacity around 0.24 mg/m2. The difference in chlorite adsorption is likely because chlorite can have several different metals present, and the amount of iron in particular can have a dramatic effect on the minerals adsorption capacity. For example, thuringite, the iron containing chlorite (0.89 mg/ m2) adsorbs 3.7 times as much asphaltenes than a non-ironcontaining chlorite (0.24 mg/m2).112 Nevertheless, the adsorption capacity of thuringite is much lower than other mineral types surveyed (mg/m2): calcite (3.4, Dover) > Ottowa sand and kaolin (2.2) > dickite (2.1) > kaolinite (2.0) > calcite (1.7, Omay Carb UF) > alumina (1.5) > illite (1.1) > Berea sandstone (1.0) > thuringite > chlorite. In similar studies of reservoir materials, a likewise trend was established (mg/m2): hematite (5) > fluorite and calcite (2) > kaolin (1.56) > quartz (1.0) > feldspar (0.94) > smectite (0.29); it was found that the iron-containing mineral hematite had a significantly different adsorption isotherm relative to the other minerals and clay minerals, which was the reason it exhibited such a relatively large adsorption capacity (see section 11).24,144 As illustrated above by calcite and kaolinite, the source from which the mineral based sorbents are derived can vary their adsorption capacities significantly. Nevertheless, similar rankings have been found (mg/m2): calcite > quartz > feldspar;145 calcite (2.74− 3.89) > quartz (1.64−1.69) = dolomite (1.65) under flow conditions;116 kaolinite (3.36) > Indiana limestone (1.82) > Baker Dolomite (1.46) > Berea clays (1.4) > Berea sandstone (0.9);134 quartz > kaolinite > illite = montmorillonite;141 and kaolin > calcite > quartz > dolomite.146 From the above, it can be seen that kaolinite and calcite are highly active minerals for asphaltene adsorption and that kaolinite consistently shows a greater adsorption capacity than illite (0.3−2.7).140,141,147,148 From one study surveying five different sources of asphaltenes over a series of mineral -based sorbents, adsorption showed a general trend despite varying sources of asphaltenes. The adsorption capacities varied between 3.78 and 0.26 mg/m2, which was correlated with the polar interactions of the surface in the following order: hydrophilic silica > TiO2 > kaolin > Fe3O4 ≈ FeS > CaCO3 > hydrophobic silica. Within the series of sorbents, no correlation was found with surface acidity, but it was assumed that H-bonding sites of surface hydroxyls determined the surfaces affinity for asphaltene adsorption.119 A study focused on asphaltene adsorption from asphalts onto different minerals using four different sources of asphalt showed that the adsorption capacity trend for the different minerals was the same regardless of the source of asphalt: those minerals that had lower adsorption capacities had a relatively low adsorption

Figure 4. A plot showing the amount of adsorbed asphaltenes as a function of asphaltene concentration for kaolinite and Athabasca asphaltenes; data were collected using an ultraviolet−visible spectrophotometer to determine the concentration.136

nanoaggregate regime137 for asphaltenes or oil dissolved in aromatic solutions (mainly toluene) at ambient temperature. Within this review, the adsorption capacity will be quantified by the amount of asphaltenes adsorbed per surface area of sorbent (mg/m2) since the adsorption capacity is limited to the number of nucleation sites which is a function of the surface area.117,138 For cases in which the surface area is not reported, then the adsorption capacity will be reported as the weight of asphaltenes per weight of sorbent (mg/g), or only the trends between different sorbents will be given from within a specific study.139 Temperature and solvent effects are discussed separately (sections 7 and 9, respectively). Also, be aware that the mg/m2 comparisons for clay minerals may be more affected by which crystalline faces are most exposed at the surface edges. Minerals. Clay minerals have been extensively studied as asphaltene sorbents because they are known to be strongly associated with asphaltene deposits found in refining equipment. They are also a major component of fines dispersed within crudes since they originate from reservoir surfaces or from heavy oil sandsthey are often regarded as the most reactive components for retaining asphaltenes.140 They are abundant, naturally occurring, and industrially mined and utilized for a variety of commercial products and industrial applications. Their specific chemical composition and purity depends largely on their source. Clay minerals are hydrous aluminum phyllosilicates (aluminosilicates) that contain parallel sheets, or lamellae, of silicate tetrahedra with varying amounts of alumina. Clay minerals and minerals contain cations within their structure which can be exchanged to affect the adsorption behavior of the material. Clay minerals can swell in the presence of water and solvents which have been determined to have a negligible effect on asphaltene adsorption. This is because asphaltenes generally do not penetrate into the interlamellar sites and are adsorbed only at the surfaces of clay minerals.141,142 Common clay minerals used in adsorption studies are kaolinite, illite, montmorillonite (from the smectite group), and to a much lesser extent the polymorph dickite, magnesium containing attapulgite, mica, and chloritewhich depending on its form contains significant amounts of magnesium, manganese, nickel, or iron. The common nonclay minerals used in adsorption studies are quartz (SiO2), calcite Ca(CO)3, feldspar (aluminum silicon oxide), dolomite (CaMg2836

dx.doi.org/10.1021/ef500282p | Energy Fuels 2014, 28, 2831−2856

Energy & Fuels

Review

did poorly crystallized kaolin (2.78 mg/m2).161 A similar adsorption capacity for a raw Chinese kaolinite was also reported in a more recent study (2.75 mg/m2).136 The high surface charge of kaolinite was correlated with its higher adsorption capacity relative to the lower surface charge and adsorption capacity of montmorillonite (∼1 mg/m2).141 As with other sorbents, the surface hydroxyls and adsorption capacity of montmorillonite may be increased through acid washing. The acid washed product, Filtrol, has been demonstrated in a patent to be affective for removing paraffin insolubles from gas oil and condensate.38 The adsorption of asphaltenes onto clay minerals has a tremendous effect on stabilizing emulsions and so-called rag layers which will not be discussed in detail; suffice it to say that the charge heterogeneity observed for aluminosilicates, especially kaolinite, makes them susceptible to adsorbing either partially positively charged asphaltenes, such as those containing amine groups, or negatively charged dissociated carboxylic acid containing asphaltenes or naphthenic acids.154,157 The ability of clay minerals to adsorb asphaltenes from the oil changes them from hydrophilic (water-wet) to more hydrophobic (oil-wet) with some particles having a mixed oil-wet and water-wet character which allows them to effectively build up at the oil and water interface since then they can effectively straddle this hydrophilic/hydrophobic interface.154,162,163 Wettability changes due to adsorbed asphaltenes also play a significant role in reservoir rock permeability and reservoir damage. In a chromatographic separation of oil into saturates, aromatics, and resins, the phylosillicate attapulgus clay has a low adsorption capacity for asphaltenes of 0.36 mg/m2, but depending on the solvent it was highly selective for separating oils similar to alumina.45 It has also been shown to be selective for removing paraffin insolubles from a gas oil and condensate.38 Mica, because it can be cleaved to provide uniform surfaces, has been used to study asphaltene adsorption.4 Mica was reported to have a strong affinity for asphaltenes and to have a higher adsorption capacity than silica,164 and it was reported to have a higher adsorption energy than quartz or dolomite.6 Silica and Alumina. Silica, alumina, silica−alumina, and crystalline zeolites have probably been the most thoroughly studied sorbent substrates largely due to their wide availability, well characterized surfaces, and easily engineered morphologies, and because their surfaces are easily modified. Silica gel, alumina, silica−alumina, and zeolites are widely used as catalysts or catalysts supports which may be impregnated with other metals for petroleum refining and upgrading. Adsorption on hydrophilic silica and alumina is largely mediated by surface hydroxyl silanol (Si-OH) and aluminol (Al-OH) groups, respectively. Silica hydroxyls are especially amenable to modification by solvents or silylating reagents to make them hydrophobic. The number of surface hydroxyls can be reduced through calcining or they can be increased through hydration or acid treatments. Alumina on the other hand is industrially manufactured to be acid, basic, or neutral. Many studies using silica surfaces show that hydrophilic silica, containing the highest number of Si-OH sites, adsorbs the most asphaltenes relative to hydrophobic modified silica.104,106,119−121,165−167 In one dramatic example, it was demonstrated that the adsorption capacity changed from 3.78 mg/m2 to 0.26 mg/m2 when going from hydrophilic Aerosil150 to hydrophobic octamethylcyclotetrasiloxane treated Aerosil

capacity for all asphalts, and minerals with higher adsorption capacities gave higher adsorption over all the asphalts. Also, it is clear from the data that the magnitude of the adsorption capacity was different for each asphalt and consistent across the different minerals: the highest adsorption capacity asphalt gave the highest adsorption on all minerals, and the rank remained the same for the lower adsorption capacity asphalts over all the minerals. This study was carried out differently from other studies mentioned above since they did not use an asphaltene model solution. Instead, the minerals were preheated to 150 °C and mixed with the oil at 150 °C for 4.5 h. The excess oil and weekly bound asphaltenes were removed with a benzene extraction, and the strongly adsorbed material was desorbed with pyridine. The adsorption capacities for the minerals were similar to those reported in other studies (average calculated values for four different asphalts): granite 1.2 mg /m2, quartzite 1.1 mg/m2, Hol limestone 0.81 mg/m2, granite P6 0.68 mg/m2, Riverton limestone 0.63 mg/m2.133 In a separate study, the adsorption of asphalt onto model rock aggregate surfaces showed that minerals comprised of limestone (2.77 mg/m2) and sandstone (2.06 mg/m2) adsorbed significantly more asphaltenes than silica or alumina (0.638 and 0.538 mg/m2, respectively).149 Other studies on reservoir rocks have shown very high adsorption capacities above 100 mg/m2 at concentrations up to 30 000 mg/L in the following the order: Berea sandstone > Berea limestone > dolomite.150 Adsorption onto minerals at various well depth and locations shows that adsorption capacity is highly regulated by the amount of hydroxyl functional groups associated with clay minerals dispersed throughout the quartz matrix.151 Clay minerals have polar surfaces containing tetrahedral SiOH and octahedral Al-OH groups situated at the broken edges and at the exposed lamellae planes terminated by hydroxyls.152 These polar groups are considered to be the active sites for asphaltene adsorption and are considered slightly acidic for kaolin,139,140,153 illite,140 quartz,24 and montmorillonite154 whereas calcite and fluorite are considered to have weakly basic Ca-OH groups.24,139,155 The mineral surface calcite is generally considered to have a positive charge; however, in one study it was reported to have a negative charge with strong interactions between asphaltene functional groups and calcium hydroxyls.156 The surfaces of the clay minerals kaolinite and montmorillonite have a permanent negative charge, and in aqueous suspensions they have localized patches of negative or positive charges at active Al-OH sites depending on the pH and electrolyte.157 Similar studies utilizing photoacoustic Fourier transform infrared (PA-FTIR) spectroscopy148 and Fourier transform infrared (FT-IR) spectroscopy augmented with X-ray adsorption fine structure spectroscopy have shown that the AlOH groups of kaolinite are affected by asphaltene adsorption, and no modification of the silicon environment takes place.153,158,159 A similar modification of the Al-OH environment of montmorillonite upon asphaltene adsorption has also been reported.154 Modifying the kaolin through silylation of the hydroxyl groups reduced the amount adsorbed below the limits of detection.139 By modifying clays in a different way, through slightly reducing the acidity of kaolin by partial cation exchange of Ca2+ with Ba2+, a slightly increased adsorption capacity was observed on a mg/g basis; however, on a mg/g basis they are probably within experimental error (0.012 vs 0.014 mg/m2).160 In a study emphasizing the lattice organization of the clay mineral kaolin, researchers found that well crystallized kaolin had a significantly higher adsorption capacity (4.9 mg/m2) than 2837

dx.doi.org/10.1021/ef500282p | Energy Fuels 2014, 28, 2831−2856

Energy & Fuels

Review

104.119 Even treating silica with methanol is enough to render the surface slightly more hydrophobic, thus reducing its adsorption capacity.130 In another study, the distance (height) between the polar silica surface was varied by adding molecular spacers which significantly affected the adsorption capacity of asphaltenes.166 Increasing the number of carbon spacers for the self-assembled monolayers of trichlorosilanes effectively screens the polar silica surface charge and blocks access to surface hydroxyl groups, which in turn decreases the amount of adsorbed asphaltenes.166 Similarly, it was determined that the distance (height) from the silica surface had a greater affect on the amount of asphaltenes adsorbed than the nature of the resulting surface terminal end (alkyl vs phenyl) formed by the monolayers.168 In two nonconforming cases, it has been reported that hydrophobic silica had a slightly greater adsorption capacity than hydrophilic silica. In one instance, the hydrophobic silane coated CPC-silica-carrier adsorbed slightly more asphaltenes than non-silane coated silica vs CPCsilica-carrier (0.94 mg/m2 and 0.79 mg/m2, respectively).130 In another study, it was reported that hydrophobic silica gave a slightly higher saturation limit (3.68 mg/m2 vs 3.18 mg/m2), but that the hydrophilic silica had a stronger interaction with asphaltenes.169 In any case, the silanol groups of silica are very important for adsorption. Thermal reduction of silica Si-OH groups by calcining at various temperatures has been shown to directly decrease the amount of asphaltenes adsorbed (calcining temperature): 1.11 mg/m2 (400 °C), 1.09 mg/m2 (600 °C), 0.98 mg/m2 (800 °C), 0.90 mg/m2 (1000 °C).139 In a recent patent using the silica Britesorb (a silica used to treat food based oils and beverages), significant amounts of asphaltenes were removed from Arabian Heavy Crude.35 Britesorb has also been shown to remove vanadium impurities from a partially deasphalted crude at ambient temperature which had been diluted with petroleum ether in a flow through chromatographic setup.34 Silica, mainly originating from quartz, has been found to be a major component (97%) of scale deposits in pipelines. Silica isolated from these deposits showed a relatively low adsorption capacity for asphaltenes from a stable oil at 0.5 mg/m2, but for asphaltenes from an unstable highly aromatic oil the adsorption capacity was nearly 10 times greater (4.6 mg/m2), presumably due to multilayer adsorption.92 Alumina, on the other hand, seems to have a much lower adsorption capacity than silica. For a porous alumina catalyst support, the adsorption capacity was reported to be 1.79 mg/ m2 even at high asphaltene concentrations.108 Among different acidity alumina, acidic alumina was reported to have the highest adsorption capacity for asphaltenes, whereas basic alumina adsorbed slightly more than neutral alumina.170 Adsorption onto alumina nanoparticles has recently been investigated. Nanoparticles have several factors that make them attractive as asphaltene sorbents: high surface area/volume ratios, tailorable surface areas, high mobility in porous media, and in situ preparation within the reservoir.77 Nanoparticles of alumina were reported to have enhanced adsorption, 2.13 mg/m2, relative to other porous alumina.77 It is important to note that adsorption data using nanoparticles may be skewed because the dispersed nature of the sorbent can give rise to colloidal effects that can significantly affect diffusion rates.118 Additionally, for nanoalumina particles and porous microalumina particles with similar acidities, it was demonstrated that despite the significantly smaller surface area of the nanoalumina particles

they adsorbed significantly more asphaltenes due to dispersion effects (ca. 1.6 mg/m2 vs 0.4 mg/m2).124 Silica−alumina sorbents exhibit both Brønsted and Lewis acidity, which can selectivity remove basic nitrogen asphaltenic material from oil, such as Cold Lake vacuum residuum and Tia Juana crude.47,49 The amount of nitrogen-containing asphaltenes adsorbed was found to be dependent on the amount of alumina in the sorbent: 13% alumina (more acidic) showed more selectivity than 25% alumina. The amount of asphaltenes adsorbed for the 13% alumina silica−alumina was significant: 20% total asphaltenes by weight, of which 32% were basic nitrogen-containing asphaltenes. This reduced the overall basic nitrogen-containing asphaltene content by 90% at ambient or elevated temperatures (less than 300 °C).48 Likewise, for neutral silica it was reported that no basic asphaltenes were adsorbed.46,47 In another application, a silica−alumina sorbent with 20−25% alumina content that was wetted with 20−30% water, or doped with 3% fluoride (NH4+F−), produced a treated oil that was quantitatively devoid of basic nitrogen content rendering it more oxidatively stable.41 The acidity of silica−alumina, by virtue of the alumina content, has been systematically studied for the adsorption of polycyclic aromatic hydrocarbons (PAH). It was demonstrated that silica−alumina with 10% alumina was more acidic than with 30% alumina and that both are significantly more acidic than the parent mesoporous silica. The adsorption of PAHs were shown to increase with increasing acidity of the sorbent.171 Silica−alumina and zeolites (crystalline silica− alumina) have also been shown to be highly effective for removing polyaromatics from raw reformate40 and for upgrading vacuum residua.39 Glass. At Western Research Institute, it has been observed that a brown layer accumulates on glassware from heptane maltenes solutions over time; this layer has been termed the “maltenes varnish”. The composition of this varnish has been shown to be similar to the 10 μm asphaltenes precipitated with heptane.172 Similarly, others observed that while carrying out adsorption experiments over 72 h at concentrations between 0.02 and 0.1 mg/L, approximately 2.7% of asphaltenes became adsorbed onto the walls of glass test tubes that were used to conduct the experiments.117 Asphaltene adsorption onto glassware is well-known since asphaltenes can adsorb onto glass through surface Si-OH groups.173 Glass slides have also been used to show the stepwise formation of asphaltene adsorption at high concentrations using photothermal displacement spectroscopy114,174 and ellipsometry.175 Metals. Hydrophilic gold surfaces, although not practical in commercial applications, have provided valuable insight into the characteristics of adsorbed asphaltene layers, the affect of resins on asphaltenes, the behavior of asphaltenes in whole oils, solvent effects, and desorption phenomena using highly sensitive quartz crystals microbalance technique with dissipation (QCM-D).106,176−178 In a study using toluene, the adsorption did not appear to level off for asphaltene concentrations between 50 and 10 000 mg/L, which gave a final tightly adsorbed asphaltene layer equal to 7.1 mg/m2.176 In a more recent study, multilayer formation was found to be more pronounced (13.09 mg/m2) for the adsorption of Cold Lake vacuum residua asphaltenes after 2 h. The difference in adsorption capacity observed between the two studies using toluene solutions was attributed to the different sources of asphaltenes (see section 6).106 2838

dx.doi.org/10.1021/ef500282p | Energy Fuels 2014, 28, 2831−2856

Energy & Fuels

Review

further decreased by blocking acidic sites by Na+ deposition.14 Similar to the positive correlations between acid strength and adsorption capacity, for metal oxides impregnated on catalyst substrates, metal oxides and sulfides were reported to have a strong relationship between surface acidity (FeS > Fe2O3 > NiO) and adsorption capacity (1.5, 1.1, 0.6 mg/m 2 , respectively). The primary surface active sites for these substrates were reported to be free hydroxyl and oxide groups.120 Acidic surface sites are key to catalytic upgrading, which are the sites where fouling tends to occur upon coke formation. These active sites have been shown in some cases to become neutralized by basic nitrogen-containing asphaltenes.14,15 Interest in metal oxide nanoparticles has been growing because they can impart catalytic activity for selectively upgrading adsorbed asphaltenes. Asphaltenes adsorbed onto metal oxide nanoparticles have been demonstrated to undergo relatively mild catalytic steam gasification/cracking. The activity of the catalysts has been found to correspond to the adsorption affinity of the nanoparticles. The temperatures necessary to initiate and sustain catalytic activity correlate to adsorption affinity in the following order: NiO > Co3O4 > Fe2O3.186,187 However, the adsorption affinities do not correlate with the adsorption capacity, Co3O4 (1.54 mg/m2) > Fe2O3 (1.44 mg/ m2) > NiO (0.56 mg/m2).187 Other work on metal oxide nanoparticles showed that the adsorption capacities roughly corresponds to the acidic nature of the solid, where basic and amphoteric oxides adsorb more than acidic oxides, from 2.5 to 0.5 mg/m2 (CaO-basic > Co3O4-amphoteric > Fe3O4amphoteric > MgO-basic > NiO-acidic > TiO2-acidic). However, the affinity between the surfaces and asphaltenes showed no relationship to the acidic strength of the oxides.188 Similar to other metal oxide nanoparticles, Fe2O3 has been shown to have very high kinetic rates of adsorption due to its high degree of dispersity. It has also been shown to be an active catalyst for assisting the catalytic oxidation of thermally cracked asphaltenes.189 Increasing the concentration of NiO impregnated nanoparticles on nanoparticulated silica gel successively increased the adsorption capacity of the composite nanoparticles. For nanoparticulated silica gel, the adsorption capacity was 1.4 mg/ m2, while adding some NiO increased the adsorption to ca. 3.8 mg/m2, and increasing the nickel oxide content even more increased the adsorption capacity further to 5.8 mg/m2.104 The unusually large adsorption capacity in the last case was attributed to multilayer formation. Nickel oxide nanoparticles have also been impregnated onto alumina which showed the same trend as silica but to a greater degree: 0.5 mg/m2 for Al2O3 and 7.0 mg/m2 for the largest amount of NiO on Al2O3.190 The ability of these nanoparticles to disperse and stabilize asphaltene aggregates through adsorption coupled with their small size has given rise to the concept of using them to prevent well damage. Nanoparticles can adsorb asphaltenes quickly, which prevents them from further agglomerating which can restore production and lead to better recovery.105 Other nanoparticles that are NiO particles doped onto a barium kaolin substrate held together with a sucrose binder were demonstrated to significantly increase the adsorption capacity relative to the substrate or the nanoparticles (3.1 mg/m2 for barium−kaolin−sucrose−NiO, 2.8 mg/m2 for barium−kaolin− sucrose, 0.5 mg/m2 for NiO). The resulting barium−kaolin− sucrose−NiO substrate was shown to be a highly active catalyst for steam gasification of adsorbed asphaltenes.122 The

Asphaltene adsorption capacities on stainless steel, iron, and aluminum metal surfaces (likely containing significant amounts of surface oxides) have been reported to be similar to minerals. Stainless steel, a major component in petroleum pipelines, was found to have the highest capacity (2.7 mg/m2) followed by iron (1.35 mg/m2) with aluminum adsorbing the least (0.25 mg/m2). The difference in adsorption capacity was attributed to the differences in the surface morphologies of the sorbents as assayed by scanning electron microscopy (SEM).117 A more recent study on adsorption at iron surfaces reported a maximum adsorption capacity of 4.9 mg/m2.118 Between the two studies using iron sorbents, the discrepancy in adsorption capacity was likely due to the diffusion difference between the sorbents types used and the solvents used: foil pieces vs 40 μm powder, and benzene vs toluene, respectively. Metal Oxides and Sulfides. Of metal oxides, iron oxides are of particular interest because they are often associated with asphaltenes and they can become concentrated as a result of pipeline corrosion9,179 or from corrosion of other refining surfaces. Adding various iron compounds to asphaltenes shows a dramatic affect on their resulting polarity,88 and they can enhance asphaltene precipitation and sludge formation.180 The clay mineral thuringite, an iron-containing chlorite, has been shown to adsorb three times as many asphaltenes than noniron-containing chlorite (see Minerals).112 Similarly, hematite was shown to have more than twice the adsorption capacity, up to 5 mg/m2, than clay minerals and minerals.144 A more recent study has shown that iron oxide adsorbed between 3.5 and 4 mg/m2 of Hamaca asphaltenes at near ambient temperatures but at high concentrations (50 mg/L to 30 g/L).181 Impregnating Fe2O3 on kaolinite, montmorillonite, and SiO2 surfaces greatly enhanced adsorption relative to the unmodified sorbent bases.182,183 Impregnating Fe2O3 on attapulgite on the other hand did not increase its adsorption because it already contained a significant amount of iron oxide. A computational study, using hematite as a sorbent, reported that Fe2O3 has a specific attractive interaction for the aromatic moieties of asphaltenes.184 Similar to iron oxide, impregnating other metal oxides on supports, such as alumina or kaolin, can significantly increase the adsorption capacity relative to the individual components of the support or the metal oxide. Alumina impregnated with metal oxides showed a significant increase in adsorption of vanadium-containing asphaltenes.125 Likewise, cobalt and molybdenum supported on alumina, CoMoAl2O3, adsorbed more porphyrins than any of its individual constituents which adsorbed porphyrins in the following order: Al2O3 > MoO3 > Co3O4.16,185 Here it was shown that the increased acidity correlated to an increase in adsorption because Lewis acid centers were mainly responsible for the interactions between the surface and the asphaltenic porphyrins. Similarly, it was established that acidic metal oxides supported on alumina are mainly Lewis acids, but when supported on pyrogenic alumina they become Brønsted acids.46,48 WO3 and Ta2O5 supported on alumina showed a high affinity for basic asphaltenes removing up to 20% by weight.46,47 For the sulfide metal alumina catalysts CoMo-Al2O3, Mo-Al2O3, and Co-Al2O3, it was shown that asphaltene adsorption blocked catalytically active acidic centers when heated in a fixed bed reactor between 200 and 300 °C.16 Likewise for the NiMo/γAl2O3 system, when the acidity was decreased the adsorption capacity also steadily decreased. Doping NiMo/γAl2O3 with tin made the catalyst more basic decreasing the adsorption capacity, which was 2839

dx.doi.org/10.1021/ef500282p | Energy Fuels 2014, 28, 2831−2856

Energy & Fuels

Review

Results f rom QCM with dissipation (QMC-D) gave consistently higher adsorption capacities than UV−visible depletion results for several sources of asphaltenes in toluene at hydrophilic silica, alumina, and TiO2 surfaces. The QCM-D adsorption capacities were 2−9 mg/m2, whereas UV−visible depletion adsorption capacities were 0.26−3.78 mg/m2.119,121 For QCM studies using metal sorbents, electrodeposition has been postulated as the reason for such high adsorption capacities.117 Others have also noted that QCM studies in static solutions also cause a combination of adsorption and deposition, which results in higher than expected values for adsorption.136 Additionally, QCM-D may generate higher adsorption capacity values than other methods due to solvent impurities or crystal contamination which can increase the relative roughness of the crystal leading to an increase in apparent mass. In summary, for ambient temperature adsorption of asphaltenes from static toluene solutions, the adsorption capacity for most surfaces falls under 4 mg/m2 with hydrophilic silica gel most consistently reaching this upper boundary (higher values consistently reported in QMC studies). Higher values due to multilayer formation were obtained for hematite (5 mg/m2), NiO on nanoparticulated silica gel (5.8 mg/m2), and NiO on Al2O3 (7.0−9.0 mg/m2), and SiO2 (4.6 mg/m2) from pipe scale. Hematite may be a special case where the iron oxide surface is able to induce multilayer adsorption behavior by specific interactions with the aromatic moiety of asphaltenes. Nanoparticulated particles impregnated with metal oxide nanoparticles are affected greatly by their enhanced diffusion rates and the characteristics of the metals. For SiO2 from pipeline scale, the multilayer formation was due to using a source of asphaltenes from an oil that strongly associates to form large aggregates in dilute solutions and is known to cause precipitation problems in the field. Regarding asphaltene adsorption capacity, in one view, saturation is rarely reached since most reported values for adsorption from good solvents are less than 10 mg/m2which is an estimated value for total monolayer saturation at an oil/ water interface.117 From this viewpoint, monolayer coverage is limited by the number of available nucleation sites which limits the amount of adsorption. However from another view point, if asphaltene nanoaggregates are assumed to have a diameter of about 3 nm and a density of 1200 kg/m3,106 and if it is assumed that a cubic asphaltene nanoaggregate takes up about 27nm3, then it will have a weight of 3.24 × 10−17 mg, and if the nanoaggregate takes up a surface area of 9 nm2, then coverage for a 1 m2 surface would be 3.6 mg/m2. This should be considered an upper boundary, since if we assume that the nanoaggregates are spherical then the approximate coverage would be 1.88 mg/m2. It can be seen then that the adsorption capacity will vary broadly depending on the asphaltene molecule size which is partially responsible for the size of the resulting asphaltene aggregates, the sizes and homogeneity of the aggregate, and homogeneity of surface coverage. If we look at hydrophilic silica gels having a surface silanol coverage greater than about 3.0 SiOH per nm2,119,196 then this sorbent clearly has an adequate number of nucleation sites to accommodate 3 nm nanoaggregates in a close packed monolayerthis may be the reason why adsorption capacities for hydrophilic silica gel are commonly observed around 3 mg/m2. In a general sense, “effective” monolayer nanoaggregate saturation should be limited to around 3.6 mg/m2, and sorbents that have an adsorption capacity near this threshold should be considered near the physical adsorption limit of the surface. Depending on

importance of surface acidity for various calcium− and barium− kaolin−sucrose substrates was determined from a model study using quinoline-65 as the asphaltene adsorbate analogue. From this study, it was shown that increasing surface acidity corresponded to increased uptake.122 Carbon Based. Carbon-based sorbents have been highly utilized in industry to treat and purify many different gaseous hydrocarbon streams or to treat wastewater effluent. Much less work has been done to adsorb asphaltenes from liquid hydrocarbons. For example, refinery wastewater contaminated with heavy hydrocarbons was reported to be successfully treated using activated carbon.44 As far as using carbon-based sorbents to treat petroleum streams, it has been shown that polyaromatic compounds can be removed from a hydrocracker feed at 288 °C with activated charcoal,43 which after treatment and upgrading the feed produced more liquid and less coke.42 Active carbon and carbon black have been demonstrated to be effective for removing asphaltenes from heavy oil vacuum residue at elevated temperatures using light cycle oil as the solvent. Carbon black in particular was found to be more selective, on a weight basis, for removing the most polar, metal-containing, THF-insoluble asphaltenes.191 Active carbon is also effective for removing paraffin insolubles from gas oil and condensates38 and Ni and V asphaltenic material from heavy oil or Valero Coker gas oil and condensate.37 A variety of other carbon sorbents were reported to be effective for removing asphaltenic material from condensate feeds.36 The carbon CG6 was reported to be more effective at removing asphaltenes than attapulgus clay > Filtrol clay > and SG6 Carbon; however, it is unclear if this was on a mg/g or mg/m2 basis.36 A novel carbon-based sorbent, mimicking inherent asphaltene−asphaltene interactions, was prepared by heating asphaltenes to 350 °C. The asphaltene sorbent was shown to adsorb significantly more asphaltenes from toluene than SiO2 on a mg/g basis (almost 10 times as much for highly associating Ceuta asphaltenes).92 Another later study claims a higher adsorption affinity for activated carbon than silica; however, surface area and pore structure data were not provided.192 Polymers. Depending on the source of asphaltenes, a considerable amount of naphthneic acids can report to the asphaltene subfraction. Carbohydrate copolymers of B-cyclodextrins have been demonstrated to adsorb toxic naphthenic acids from processed water for the treatment of oil sands through a selective guest−host interaction with B-cyclodextrin active sites. Various monomer diisocyanate cross-linkers were used to optimize the textural properties of the resulting copolymer to determine if naphthenic acid selectivity can be achieved.193 The adsorption of these naphthenic acids has mainly been applied to treating wastewater from tailings produced by hot water bitumen extraction of Canadian oil sands.193,194 Note regarding quartz crystal microbalance (QCM) studies: For hydrophilic silica, alumina, clay minerals, and minerals, the reported adsorption capacity using QCM are often signif icantly higher than the capacities reported f rom other studies.121 Using a gravitational f low system, the capacity for hydrophilic silica was reported to be 5.3 ± 0.8 mg/m2.167 Another group using a f low setup to study the displacement of preadsorbed asphaltenes on hydrophilic silica and alumina by ethyl cellulose to alter its wettablity reported that the preadsorbed amount of asphaltenes were ∼7 mg/m2 for silica and 6.2 mg/m2 for alumina. These values are signif icantly higher than reported in other studies mentioned.195 2840

dx.doi.org/10.1021/ef500282p | Energy Fuels 2014, 28, 2831−2856

Energy & Fuels

Review

asphaltenes for a particular sorbent. The attraction of functional groups onto rock minerals has been a significant area of study for paving asphalts (similar to vacuum residuum) since these interactions are of paramount importance to the long-term performance of roads. Strong interactions ensure that the asphalt does not separate from the rock aggregate, weakening the pavement. Asphaltenes adsorbed onto five different rock aggregates (quartzite, two limestones, and two granites) at 150 °C were rinsed with benzene and then desorbed with pyridine to give the strongest adsorbing asphaltenes in the pyridine solution. Infrared analysis of these asphaltenes showed that they were significantly concentrated, by several fold, in oxygen and nitrogen species relative to the bulk asphalt, but sulfur was not found to vary significantly. The strength of the functional group interactions was determined as follows: carboxylic acids > dicarboxylic anhydrides > 2-quinolone types > sulfoxides > nitrogen (pyridinic) > ketones. Affinity for the selective adsorption of the different functional groups was not uniform between all the rock minerals surveyed. Confirmation for the relative strength of these interactions was gained by using model compounds in cyclohexane solutions measured by UV absorbance with decreasing strength in the following order: benzoic acid, quinoline, phenyl sulfoxide, phenyl sulfone, valarophenone, phenol, benxyl benzoate, 1,2,3,4-dibenzantrhracene, and naphthalene.133 In a separate report, model asphaltene compounds were studied to understand how functional groups participated in adsorption on dry silica, and the following affinities were reported: phenylsulfoxide > quinoline > phenol > benzoic acid > benzophenone > benzylbenzoate > pyrene.149 In a similar approach using the sorbent goethite (FeOOH), a different source of asphaltenes, and heptol as the solvent, a different ranking was established: indole < diphenylsulfoxide < benzoic acid.179 Studies using model molecules to mimic Athabasca asphaltenes concluded that the heteroatom content was more important for modeling asphaltene adsorption interactions than the overall chemical composition of the molecule. In particular, N and especially O heteroatoms were important for mimicking Athabasca adsorption behavior, while S was not as important because it imparts a significantly weaker dipole.206,207 When the relatively higher interaction potential for adsorption onto hydrophilic silica compared to hydrophobic silica was examined, it was reported that the mechanism of adsorption is dominated by the polar entities of asphaltenes,169 and the polar entities were delineated as functional moieties consisting of pyridinic, pyrrolic, phenolic, carboxylic, and quinolic groups.168 FT-IR spectroscopy of asphaltenes adsorbed onto montmorillonite,154 and kaolinite or illite showed that carboxylate and carboxylic asphaltene groups participate in surface adsorption.140 Heteroatoms in the bulk matrix can have different affinities for different substrates leading researchers to propose them as markers for predicting the degree of adsorption that can be expected for a particular oil, or as a metric for the selectivity for a particular type of asphaltene at a particular sorbent interface.120 In some studies, adsorbed asphaltenes were reported to be enriched in S and N heteroatoms relative to the original samples indicating selective removal of asphaltenes.98,101 Asphaltenes adsorbed onto kaolinite showed a 6fold increase in the N/C and S/C ratios compared to the oil.159 By varying the crudes, a linear correlation was found between the sum of nitrogen and sulfur content and adsorption.98 In one

the asphaltenes, values significantly less than this may be considered limited by surface adsorption sites, and values above this could be considered an “effective” multilayer. It should be considered, especially in cases where the surfaces have significantly less surface active sites than hydrophilic silica (∼0.1 sites/nm2), that values around 3.6 mg/m2 may actually be due to formation of multilayers. Conclusions about monolayer and multilayer coverage should be taken cautiously, and preferably the term “effective”118especially in reference to monolayersshould be used because of the inherently heterogeneous distribution of asphaltene aggregates in solution and at surfaces (see sections 10 and 11). Perhaps “effective nanoaggregate monolayer” would be a better descriptor for adsorption capacities when experiments are conducted in good solvents below the concentration range 3000−5000 mg/L for well behaved asphaltenes.

4. CHEMICAL CHARACTERISTICS OF ASPHALTENES IMPORTANT FOR SURFACE ADSORPTION Asphaltenes containing mixtures of acids and bases having ionizable groups in aqueous solutions.24,134,197−200 Aggregates in nonaqueous mixtures have also been shown to have a net charge of +1.84 Asphaltenes are known to contain acid−base pairs have been found to be significantly adsorbed at oil and water emulsion interfaces201,202 and are also thought to be important for asphaltene−adsorbent interactions.4,5,24,119,167 Asphaltene polar functional groups203 as well as the aromatic backbone give rise to significant electrostatic interactions.171 Characterization of Athabasca bitumen charge carriers in toluene, using electrodeposition quantified by QCM and functional groups analysis by IR and XPS, showed that the acid−base pairs are unbalanced combinations of high mass Lewis bases (:N) and low mass (or greater multiplicity) Lewis acids (RCO2H).204 It is also important to keep in mind that most oils can contain considerable amounts of water, and this native water can have a dramatic effect on adsorbent−adsorbate interactions (see section 8). It is clear that in aqueous and high dielectric organic solvents ion exchange mechanisms will dominate, and in less polar organic solvents these attractive forces are minimized and van der Waals, π-π, acid−base, coordination, and H-bonding interactions become dominant. Heteroatoms such as nitrogen were initially found to be important regarding asphaltene adsorption. When studying the negatively charged montmorillonite surface, it was determined that long-range interactions between the surface and positively charged (partial) nitrogen groups of asphaltenes are the driving force for the initial attractive interaction leading to adsorption. This was concluded from data that showed that the basic nitrogen concentration increased in the adsorbed layer and became depleted in the bulk solution. It was also postulated that as the distance between the asphaltene aggregates and surface became sufficiently short the initial attractive force becomes small relative to van der Waals and/or π-stacking interactions.110 An earlier study for the retention of crude material in sandstone containing at least 10% clay minerals showed a preferential binding of basic nitrogen compounds and as the basic nitrogen to total nitrogen ratio increased the more strongly the asphaltenes were bound.205 Since initial studies in the 60s and 70s, several publications have related the importance of nitrogen and oxygen heteroatoms to adsorption. Heteroatom and functional groups analyses have been used as a way of predicting the affinity of 2841

dx.doi.org/10.1021/ef500282p | Energy Fuels 2014, 28, 2831−2856

Energy & Fuels

Review

increases closer to the interface.178 In contrast, in a recent study of asphaltene adsorption onto kaolin clay, XPS was unable to detect an increase in N and O upon sputtering time; this however may have been influenced by the significant heterogeneity of the aggregates adsorbed at the surface as determined by XPS depth profiling, contact angle, and time-offlight secondary ion mass spectroscopy.136 Intramolecular π−π stacking and H-bonding forces have been shown to be significant forces in directing asphaltene aggregation.67,213 These forces are also likely to participate in asphaltene surface interactions leading to adsorption. An early adsorption study concluded that since the hydrogen acceptor/ donor solvent mixture of chloroform/methanol was efficient at extracting adsorbed asphaltenes then the primary forces occurring at the asphaltene−surface interface were Hbonding.112 Others have also pointed to the importance of polar interactions between the polar groups of oil components and the surface alluding to the importance of H-bonding.119 Likewise, for asphaltenes on Fe2O3-modified surfaces IR showed significant changes in H-bonding character. Efficient desorption of asphaltenes by water, the prototypical hydrogenbonding molecule, supports the assertion that H-bonding is dominant for Fe2O3 asphaltene interactions.183 With regard to aromatic π−π stacking interactions at the interface, significant aromatic π−π interactions are present between asphaltenes in a highly concentrated solid state; however, it has yet to be directly observed that these interactions are occurring significantly at surfaces.67 Recently, high level computations have shown that π−π interactions are dominant between aromatic cores of asphaltene molecules to form smaller asphaltene moieties such as dimers. The interaction becomes less likely for larger associated asphaltene structures due to the steric blocking imparted by the combination of the two asphaltene molecules’ aliphatic side chains, alkyl linkages, and cyclic naphthenes.82 In one study, it was demonstrated that asphaltenes did not have a preferential affinity for phenyl terminated self-assembled monolayers relative to alkyl-terminated monolayers,168 suggesting that π interactions between aromatic portions of asphaltenes and the surface are not crucial for surface adsorption. On the other hand, the aromaticity of asphaltenes, a well defined parameter of asphaltenes, has been demonstrated to have a significant influence on asphaltene adsorption. Asphaltenes with a higher degree of aromaticity are reported to cause multilayer adsorption making them more prone to precipitate within pipelines.92 In a computational investigation, asphaltenes with increased aromaticity and minimal aliphatic side chains were demonstrated to have a specifically attractive interaction for Fe2O3.184 Adsorption onto kaolin showed that the number of aromatic rings and aromaticity was greater for adsorbed asphaltenes relative to whole asphaltenes from the original oils.100,148 Likewise, after asphaltene adsorption onto illite the asphaltenes from the supernanant oil showed a reduction in aromaticity.101 Visbreaking a vacuum residue decreases the number of aliphatic groups, decreases the average molecular size, and increases the aromaticity of the resulting asphaltenes.214,215 The resulting visbroken asphaltenes showed enhanced adsorption relative to the virgin sample.128,151,160,189 Similarly, hydroconverted asphaltenes with higher aromaticity were found to be more polar and to interact more strongly with a NiMo alumina supported catalysts sorbent.14 In a study using model compounds adsorbed to mineral surfaces, it was observed that the larger 5-aromatic ring structure of 1,2,3,4-

patent, the amount of sulfur present in the asphaltenes seems to correlate with the aggregate size observed on the surface.208 Similar to the first observations by Clementz concerning nitrogen content, a later study showed a positive correlation between the amount of nitrogen and the amount adsorbed for five crudes over several inorganic sorbents.119 In ultrafiltration membrane fouling, the adsorbate−sorbent interaction has been postulated to occur between basic asphaltene nitrogen and acidic surface sites leading to the initial adsorption of a monolayer. Fouling can occur when the adsorbed asphaltenes interact further with other asphaltene aggregates through basic nitrogen and/or acidic oxygen groups to form multilayers.32 Asphaltenes from mildly visbroken vacuum residue at 28.5% conversion showed that an increase in the N content of the resulting oil correlated with increased adsorption of the asphaltenes.128,151,189 In a similar study for visbroken residue and a basic modified kaolin sorbent, increasing N and O content was reported to positively correlate with adsorption (S content decreased due to alkyl sulfide cleavage).160 They also reported that when the N- and O-containing asphaltenes were removed by adsorption the resulting deasphalted oil exhibited an improvement in its stability as gauged by an increased P value.160 In a different study it was reported that the sulfur concentration is higher for asphaltenes adsorbed onto basic surfaces, but that the overall amount of asphaltenes adsorption was lower relative to acidic sorbents which bind more quickly or strongly to lower sulfur-containing species.120 It is apparent that the acid−base character of nitrogen and oxygen functional groups plays a significant role in adsorption41,46−49,170,188,209 and is a key parameter when modeling kinetic and thermodynamic behaviors of asphaltene adsorption and affinity.210 X-ray photoelectron spectroscopy (XPS) is a valuable tool for elucidating the types of functional groups and heteroatoms present in adsorbed asphaltenes. Sputtering petroleum deposits from well tubing by XPS removed small amounts of the adsorbed layer revealing that N, Fe, Al, and especially S content increased with time. This suggest that the most polar asphaltenes containing functional groups are partially responsible for the initial adsorption interaction leading to pipeline deposition.9 Similarly, XPS on clay surfaces has shown that the N and S content are enriched at the surface after adsorption of C5 asphaltenes from toluene. The adsorption capacity of the asphaltenes onto illite correlated linearly with increasing carbon and S.101 For asphaltenes adsorbed onto stainless steel, the functional groups detected by XPS were carboxylic, pyrrolic, pyridinic, thiophenic, and sulfite,211 and on gold they were carboxylic, thiophenic, sulfide, pyridinic, and pyrrolic.178 Sputtering an adsorbed layer on a stainless steel surface gave results similar to asphaltene pipeline depositsthe heteroatoms N and S increase when approaching the sorbent interface suggesting that polar interactions from the associated functional groups of these heteroatoms are responsible for initiating adsorption.212 However, unlike in the case for pipeline deposits, it was also found that oxygen increased significantly upon sputtering, which may also suggest the importance of oxygen functional groups. It should be noted that the oxygen concentration was very high relative to the original sample eluding to oxidation of the surface which may have contributed to such high values.212 Others have also reported similar trends; the concentration of N, O, and to a lesser degree S is enhanced in adsorption layers and is responsible for the polar interactions prompting adsorption,113,120 and the heteroatom concentration 2842

dx.doi.org/10.1021/ef500282p | Energy Fuels 2014, 28, 2831−2856

Energy & Fuels

Review

Figure 5. At the left is a generic representation showing the adsorption interaction of an asphaltene aggregate at a sorbent surface through respective active sites. Major sorbent surface active functional groups (middle) and asphaltene surface active groups (right) are shown (π−π stacking interactions are not shown). It should be noted that asphaltene aggregates can have several different surface active sites and can orient at the surface in a variety of ways and can attach at multiple points. These aggregates can also interact with other asphaltene molecules and/or other oil components.

dibenzanthracene adsorbed appreciably, whereas the smaller 2aromatic ring naphthalene did not.133 When applying a polymer model, it has been shown that asphaltene aggregates initially exhibit long-range steric-entropic repulsive forces, but as aggregates approach one another shortrange attractive forces dominate which are attributed to van der Waals forces and asphaltene chain interdiffusion.4 For surfaceasphaltene interactions, van der Waals interactions have generally been invoked due to unusually small changes in surface IR hydroxyl stretching frequencies, 159 or from computational data.81 Although van der Waals forces are more important for interactions between asphaltene aggregates,82,83,213 they are a rather small portion of all the different forces present within asphaltenes that build up the supramolecular assembly of an asphaltene aggregate;80 therefore, they are not likely to be the dominate force for asphaltene− sorbent interactions. One last important chemical observation should be noted; oxidation of asphaltenes changes the molecular structure of some of the asphaltenes, and this can have an impact on adsorption values. Upon oxidizing asphalts significantly more ketone functional groups become present, and since ketones have a very low affinity for adsorption, adsorption of oxidized ketone type asphaltenes becomes less with increasing oxidation severity.149 The other major byproducts of oxidation often observed with asphalts are carboxylic acids and sulfoxides which are known to strongly adsorb to surfaces.133 In summary, several reports have shown that heteroatoms are important for asphaltene−sorbent interactionsespecially N and O and to a lesser degree S. Figure 5 shows a generic representation for the interaction between sorbent surface active sites and asphaltene surface active sites and the chemical nature of the respective active sites. Despite several observed correlations, a few studies have shown no global conclusions regarding heteroatom content and adsorption. One explanation for this may be the different structures of asphaltenes between different sources which may change in the polyfunctionality or steric hindrance of heteroatoms within the asphaltene molecules. Insights into the later phenomena can be gleaned

from studies using chromatographic separations of pyrrolic nitrogen compounds which show that adsorption and elution are governed by the size of the molecule, steric blocking of the interacting site, and the degree of shielding of the nitrogen group.216 This may be the reason why researchers have had difficulty finding universal correlations between adsorption behavior, asphaltene elemental composition, and acid−base data since these techniques account for an average of the bulk properties and not the local environment of the asphaltene molecule active sites. For example, at low concentrations less than 50 mg/mL, in a good solvent, and assuming that the functional groups are not buried within an aggregate or bound up with other associated agglomerates, asphaltenes with a high nitrogen content and total base number may not adsorb strongly at an acidic surface if the majority of nitrogencontaining species have large sterically blocking aliphatic chains positioned near the nitrogen, if the nitrogen is buried within the pericondensed aromatic core, or near alkyl linkages or naphthenic moieties. Correlations with sulfur can be expected to be even more elusive since sulfur has a significantly smaller dipole than nitrogen, and in the case of sulfides, they are often associated with nonridged sterically blocking alkyl groups.

5. EFFECT OF SATURATES, AROMATICS, AND RESINS ON ASPHALTENE ADSORPTION The various fractions of oil, for example, asphaltenes, maltenes, saturates, aromatics, and resins, have inherently different chemical and physical properties that can affect the degree of asphaltene aggregation in solution and consequently adsorption at the surface. In general, for asphaltene adsorption studies, asphaltenes are isolated from residua, crude, processed or thermally cracked oil, or shale oil using either pentane or heptane by various methods. Variations in the isolation methods changes the amount of resins associated with the asphaltenes. Early work on asphaltene adsorption showed that the adsorption of isolated asphaltenes was similar to the adsorption characteristics of the corresponding whole oil,110 since then several other authors have reported likewise that the adsorption behavior of an oil is determined mostly by its 2843

dx.doi.org/10.1021/ef500282p | Energy Fuels 2014, 28, 2831−2856

Energy & Fuels

Review

compete for adsorption active sites106 or simply reduce the size of the aggregates. Similar observations were found for metal substrates. Mixtures of resins and asphaltenes slightly decreased the adsorption capacity relative to only asphaltene solutions by either competing for binding sites or decreasing the size of the aggregates.117 The effect was also demonstrated through the decreased adsorption of C5 (more resinous) asphaltenes relative to C7 asphaltenes on NiO nanoparticles.188 The same trend was observed in another study that used increasing carbon length n-alkanes (C5, C7, and C8) to precipitate asphaltenes. As the carbon length increased, the amount of associated resins decreased and adsorption increased.73 In another QCM-D study, researchers reported that the adsorption of asphaltenes in the presence of resins onto different types of silica gel had very little effect on the overall adsorbed amount or that the resulting adsorbed amount was decreased slightly.165 In a study using small-angle X-ray scattering (SAXS) to probe the size of asphaltene aggregates, it was determined that as the amount of resins increased the radius of gyration of the asphaltene aggregates decreased leading to decreased adsorption.73 It is apparent that the interactions of resins at surfaces are minimal in good solvents and that resin interactions with preadsorbed asphaltene aggregates are also minimal. The most important effect appears to be how the resins change the overall structure of the aggregated asphaltenes in solution and how the resulting structure is manifested at the sorbent interface. From the literature it is observed that resins influence asphaltene adsorption by at least four mechanisms: competition for sorbent adsorption sites, termination of aggregate growth in solution leading to smaller adsorbed aggregates at the surface, termination of further asphaltene adsorption onto the initially adsorbed layer, and/or intercalation or incorporation of resins within the asphaltene aggregate structure to make them slightly larger. The data seem to suggest that the second mechanism is a dominant feature for asphaltene−resin adsorption interactions at solid surfaces. This appears to support the supramolecular model of asphaltene aggregates put forward by Gray et al.80 where resins become terminators of supramolecular growth.

asphaltenes (adsorption of whole oils are usually slightly less).92,164,165,175,176 From one of the first studies on asphaltene adsorption, it was reported that, when using montmorillonite, resins adsorbed slightly more than asphaltenes because of their smaller size which allowed them to pack in a more condensed arrangement on the surface and/or allowed them to penetrate the interlamellar sites.110 However, in a later study no penetration into interlamellar sites was observed for montmorillonite.111 Several recent studies have shown that resins adsorb very little relative to asphaltenes in aromatic solvents (for toluene 5 wt% for resins and 20 wt% for asphaltenes)92,100,139,156,164,176,188,217 and that when asphaltenes are adsorbed onto a surface the asphaltene−asphaltene interactions are more significant than asphaltene−resin interactions.92,106,176 The isotherms and adsorption capacities of resins adsorbed onto a solid sample of heated asphaltenes were significantly more than on SiO2 from pipeline scale showing some importance of asphaltene− resin interactions.92 As a result for weaker resin adsorption at sorbent surfaces, adsorbed resins can be quantitatively rinsed off from surfaces using aromatic solvents.218 By investigating the adsorption behavior of the partitioned oil fractions of saturates, aromatics, resins, and asphaltenes (SARA) it was shown that saturates and aromatics do not adsorb,100 but resins are more “sticky”;164 hence, they have limited adsorption and asphaltenes exhibit facile adsorption.100 Likewise, several studies have shown that maltenes adsorb to a significantly lesser degree than either isolated asphaltenes or the whole oil from which they were isolated because maltenes are concentrated with smaller less polar molecules (saturates, aromatics, and resins) than asphaltenes.77,111,139,219 QMC-D studies show that resins do not remove or add to already adsorbed asphaltene layers. However, when combining asphaltenes and resins prior to adsorption the resulting adsorption layer was slightly larger.176 The same phenomena was also observed by ellipsometry, and it was reasoned that resins were able to terminate the ability of the aromatic asphaltene surface to organize into a more compact layer.175 In another study, asphaltenes continued to adsorb on glass slides over time, but when asphaltenes and resins were adsorbed surface saturation was reached which did not increase further with time.114,174 The reason the adsorbed layer may sometimes become slightly larger when asphaltenes and resins are combined prior to adsorption176 may be attributed to the swelling of the aggregates from intercalation of resins into the aggregates.94 Adsorption studies from crude oils have also demonstrated that waxes and paraffins have a slight effect on the final aggregate structure which is unrelated to the solvent power of the oil.176 Most researchers on the other hand have found that resins significantly decrease the size of asphaltene aggregates in solution, which translates to smaller aggregates adsorbed at the surface. It was demonstrated that when asphaltenes are added to a crude solution the resulting adsorbed amount is larger relative to adsorption from only the crude, and conversely adding maltenes (resin rich) to the crude decreased the overall adsorption.106,176 The same trend was found for asphaltene aggregates at a model oil solution and water interfaces.177 In a QCM-D study, asphaltenes from a model solution were found to adsorb to a greater degree than their native pitch and much more than their maltenes: 13.09, 7.95, and 4.86 mg/m2, respectively. The decrease in adsorption between the pitch sample relative to isolated asphaltenes shows that resins may

6. PHYSICAL SIZE OF ASPHALTENES AND THEIR AGGREGATES PERTAINING TO ADSORPTION The physical size characteristics of asphaltenes can greatly affect the ability of the asphaltene’s surface active sites to favorably interact with the active sites of a sorbent. Two main physical properties of asphaltenes related to adsorption are the size of the primary asphaltene molecules14 and the resulting size of the asphaltene aggregate. Asphaltene aggregate size has a direct effect on the observed isotherms (see section 11), and they can indicate how strongly aggregates will interact at interfaces. Aggregate size also plays a primarily role in diffusion rates in solution and at porous interfaces,107 and it also affects pore plugging. Clearly, increasing the size of aggregates will decrease the rate of diffusion of the aggregates in solution causing adsorption to be more time dependent.6,121,210 The correlation between aggregation size and rates were reported to be stepwise, and the rates were highly dependent on concentration.107 It has been argued that smaller aggregates are less likely to have buried functional groups and are more likely to interact at sorbent binding sites.113 However, increasing aggregate size often leads to increased adsorption due to the larger aggregates that appear 2844

dx.doi.org/10.1021/ef500282p | Energy Fuels 2014, 28, 2831−2856

Energy & Fuels

Review

on the surface as observed in solvent dependent studies (see section 9). In an interesting approach, a density gradientalso an aggregation gradientof asphaltenes was created upon ultracentrifugation of an asphaltene solution in toluene and adsorption was studied using the different density fractions. The adsorbed amount increased (from 1.79 to 4.9 mg/m2) with increasing density (increased apparent molecular weight and aggregation), but the estimated amount per mole remained about constant.108 When studying adsorption onto silica, alumina, and titanium oxide, a rough correlation was established between the size of aggregates in toluene and the amount adsorbed on the surfaces.121 Depletion of larger asphaltene aggregates from solution was observed after adsorption onto illite since the resulting solution contained lower molecular weight aggregates than the original solution.101 Recently, it was determined by size exclusion chromatography and fluorescence spectroscopy that larger asphaltene aggregates are depleted upon adsorption and that more smaller molecules are released.181 In a study aimed at correlating the bulk asphaltene volume properties of molecular mass and radius of gyration to adsorption, as determined by SAXS, it was clear that for a single source of asphaltenes that as the aggregates became larger the total amount adsorbed onto a surface increased linearly. Nevertheless, when trying to find a global correlation between different sources of asphaltenes, no direct relationship was found between the total amount adsorbed and the molecular masses or radii of gyration. It was also reported that there was significant modification of aggregates upon adsorption relative to solution aggregates. The adsorbed aggregates at the surface were more compact and dense (densification) than the more diffuse fractal aggregates in solution.73 Similarly, from neutron reflectivity it was observed that upon adsorption there is a significant densification of the aggregates. It was also concluded that the adsorbed aggregates are essentially of the same dimensions as the solution aggregates and that no specific reorganization of the asphaltene aggregates occurred when they move from the solvent to the surface interface.169 Visbreaking of vacuum residue results in thermally cleaving aliphatic side chains from asphaltenes producing smaller-sized asphaltene molecules.160,214,215 The reduced steric repulsions of the smaller asphaltene molecules result in better exposure of the asphaltene aromatic core and heteroatoms to sorbent surfaces resulting in higher adsorption relative to virgin vacuum residuum.121,151,160 A similar trend was found for hydroconverted asphaltenes where the smaller molecules bound more strongly to the surface and diffused more into the internal structure of a porous NiMo catalyst supported on alumina.14 Likewise, higher adsorption of thermally cracked Cold Lake vacuum residuum from toluene onto a gold surface was reported than for asphaltenes from North Sea crude.106

corresponds to the size of aggregates adsorbed onto surfaces: increasing the temperature produces decreased adsorption.77,104,117 Several other studies have reported that adsorption decreases with increasing temperature or was unaffected (adsorption is exothermic,77,190 adsorption onto iron oxide described as endothermic181).4,77,104,181,190,192 Likewise, some have claimed that temperature does not have an effect on overall adsorption and is only important at short reaction times.120 However, in real world applications aimed at treating whole oil feedstocksnot just asphaltene solutionsseveral patent applications claim enhanced selectivity of adsorption at elevated temperatures. When using silica, alumina, or silica−alumina as a sorbent it was found that temperatures above about 200 °C were optimal for selective removal of asphaltenic material.39,41,42,48,50 Selectively removing the most polar metalcontaining asphaltenes from a heavy oil vacuum residue, using light cycle oil as the solvent, was enhanced at temperatures up to 150 °C.4 Removing asphaltenes from condensate-containing asphaltenes using CG6 Carbon was enhanced at higher temperatures (170−230 °C), which also showed a fortuitous increase in the adsorbent life.36 Even higher temperatures (288 °C) were reported to be favorable for removing asphaltenes from a vacuum gas oil feed using activated charcoal.43 For uptake of asphaltene aggregates onto mica, elevated temperatures were reported to increase the kinetics of adsorption.164 For the sulfided CoMoAl2O3, MoAl2O3, and CoAl2O3 catalysts in a fixed bed reactor, at temperatures between 200 and 300 °C, asphaltene adsorption from xylene solutions were quantitative at about 250 °C.16 To study asphalt paving performance, it is necessary to understand the adsorption interactions of asphaltenes with rock aggregate under applicable paving conditions which vary but are usually around 160 °C. Asphaltenes were demonstrated to be strongly adsorbed onto several different rock aggregates after contacting asphalt with the rock at 150 °C for 4.5 h (about 1% of an asphalt/vacuum residuum was adsorbed).133 Model solutions comprised of asphaltenes dissolved in solvents attempt to study asphaltene behavior in a controlled manner, but these solutions and routine laboratory methods are not readily amenable to higher temperature adsorption experiments. However, industrial asphaltene adsorption is often studied at higher temperatures as evidenced from the patent literature. For viscous oils, increasing the temperature decreases the viscosity and increases diffusion rates. For whole oils, increasing temperature can enhance adsorption by breaking up the associations that hold together asphaltene aggregates, thus exposing surface active sites that are buried or bound up within aggregate structures. Disrupting asphaltene aggregates makes smaller asphaltene agglomerates, which in turn have faster rates of diffusion which also helps to enhance adsorption.

7. TEMPERATURE EFFECTS ON ADSORPTION It is generally accepted that the size of asphaltene aggregates decreases with increased temperature of the supporting medium.86 There are a few exceptions; one notable recent example showed that for a particular source of asphaltenes at 250 °C the aggregates distributed into two size regimes: a lower aggregate diameter distribution, as expected, and an equally populated larger diameter distribution relative to the average diameter observed at 150 °C.220 Accordingly, for asphaltene solutions the temperature dependence on aggregate size

8. EFFECTS OF WATER ON ADSORPTION AND ADSORBED ASPHALTENES Because of its prototypical hydrogen-bonding character− water−in oil has a significant effect on the aggregation structure of asphaltenes.213,221−223 H-bonding interactions are important for asphaltene−sorbent and asphaltene−asphaltene interactions which can be significantly altered by the presence of water.177,224 Subjecting sorbent surfaces to moisture prior to adsorption showed that adsorption decreases because water strongly competes for surface adsorption sites,77,111,179 but it 2845

dx.doi.org/10.1021/ef500282p | Energy Fuels 2014, 28, 2831−2856

Energy & Fuels

Review

does not completely inhibit adsorption.143 In some cases, water was found to preferentially adsorb onto an adsorbed asphaltene layer,173 while in the case of Fe2O3 modified surfaces water completely desorbs asphaltenes.183 Time-dependent studies showed that when water is contacted with an adsorbent surface for longer periods of time, prior to asphaltene adsorption, it becomes more organized and the amount of asphaltenes adsorbed decreases.161 For mica-coated with asphaltenes, water was shown to increase the sort-range attractions across the oil by discrete charge−charge acid−base interactions. The water was able to penetrate the aggregate layers causing them to swell, and it continued to penetrate building up a water layer at the surface which eventually displaced the asphaltene aggregates.4 Moisture can also decrease the selectivity of the sorbent: silica−alumina sorbents lose their selectivity for basic asphaltenes.46−49 However, in one patent the opposite was shown: by intentionally adding between 20 and 30% of water to an amorphous silica−alumina sorbent, the selectivity for basic nitrogen-containing asphaltenes was increased.41 The ionic strength of water is also important for determining how asphaltenes will interact at surfaces and is very important for reservoir rock studies. For asphaltenes adsorbed onto silica, as the ionic strength of water increased the charges on the asphaltenes also increased thus enhancing desorption.167 For clay minerals exposed to moisture, the degree of hydrophilicity and cation exchange capacity correlated with adsorption capacity.147 As the hydrophilicity of the surface increases, the water becomes more competitive with asphaltenes for adsorption sites.140 Likewise, the higher the cation exchange capacity the higher the surface charge resulting in preferential water adsorption lowering the amount of adsorbed asphaltenes.154 Infrared functional group analysis was used to study which types of asphaltene functional groups are most affected by water. When moisture disrupts the adhesion of asphalt to rock aggregate, by desorbing adsorbed asphaltenes, stripping occurs, which is an important issue for pavement performance and preservation. Asphaltenes adsorbed from asphalt were treated with water and frozen for 21 h at -5 °C followed by heating at 60 °C for 24 h. The functional groups were analyzed by comparing the strongly adsorbed asphaltenes (removed by Soxhlet extraction using pyridine) to the asphaltenes desorbed by water. This study showed that the oxygen-containing functional groups, most especially carboxylic types, were most readily displaced by water due to their affinity to hydrogen bond with water. The process is further driven by complementary hydrogen bonding of water to the respective sorbent active surface sites that had an original affinity for the carboxylic functional groups. The following functional groups were found to be least affected by water desorption: pyridinic < 2-quinolone < ketone.133 In a follow up study, model compounds were used to determine the interplay between the carboxylic acid and basic amine functional groups. Since it was established that nitrogen groups are significantly less affected by moisture than oxygencontaining functional groups, the nitrogen content was used to quantify the amount of material retained using a pyroreactor coupled to a nitrogen detector. After adsorption from model compound solutions of benzene onto different rock aggregates and rinsing with benzene to removed nonadsorbed material, the moisture susceptibility was checked by soaking the adsorbate-rock aggregate in water and benzene (15 mL to10

mL, respectively) solutions for 6 h. The model compounds surveyed were benzoic acid, pyridine, indole, idoleproploic acid, and basic fractions of shale oil, and combinations of these. In general, nitrogen functionalities were strongly retained, but if the surface was treated first with an acidic functional group and then the base, then less basic compounds were retained relative to adsorption of only the base. Mixing acids and bases together prior to adsorption also gave similar results. However, if the surface was first treated with the base and then with the acid, then the bases were still strongly retained, and in some instances there was a symbiotic affect causing significantly more retentions of basic nitrogen-containing material compared to adsorption of only the base. This study also showed that indole functional groups and bases extracted from shale oil were significantly adsorbed and also strongly retained by surfaces after water treatment.224

9. SOLVENT AND CONCENTRATION DEPENDENCE It has been known for some time that the degree of asphaltene aggregation is highly dependent on the strength of the solvent.225 The charge state of the asphaltenes and their aggregates can also be modulated by the solvent, such as in the extreme case where asphaltenes are ionized in strong solvents such as nitrobenzene110−112 or nitromethane.134 The source, concentration, and solvent strength of the solution determines the degree of asphaltene−asphaltene association. In good asphaltene solvents such as benzene, toluene, xylenes, chloroform,110 and CS2,139 asphaltenes are present mainly as primary molecules or dimers at very low concentrations (>50−100 mg/ L), nanoaggregates at relatively low concentrations (as low as 100 mg/L up to 2000 and as high as 5000 mg/L), or more associated clusters of nanoaggregates at higher concentrations (as low as 1500 mg/L but generally between 5000 and 50 000 mg/L).67,70,71,78,137 It is important to point out that within these concentration ranges there is not an absolute cutoff point in aggregation size; rather there remains some population distribution of varying agglomerate sizes. In poor solvents such as heptane-diluted toluene (heptol), the degree of asphaltene aggregation increases as the relative amount of heptane is increased and the solubility parameter of the solvent decreases.226 From these solutions, the larger aggregates manifest themselves as larger amounts of adsorbed asphaltenes relative to stronger toluene solutions which have smaller aggregates and produce smaller adsorbed amounts.77,106,113,117,121,141,166,210 Concentration is known to affect the rates of adsorption at gold and model oil/water surfaces. Kinetics studies of adsorption onto gold have shown for low concentrations (139−278 ppm of asphaltenes) of crude oil in toluene that primary asphaltenes are adsorbed and at higher concentrations (835 ppm of asphaltenes) nanoaggregates are primarily adsorbed.113 For asphaltenes dissolved in solvents, and asphaltenes in bitumen diluted with heptol, it was found that at high concentrations a steady-state saturation plateau is reached, and at low concentrations a saturation plateau is not observed due to non-steady-state behavior.177 In another study, for adsorption onto an iron based sorbent, it was demonstrated from thermodynamic data that the adsorption mass density peaked at concentrations between 250 and 350 mg/L and did not vary up to 1500 mg/L. It was suggested that for concentrations above 250−350 mg/L the asphaltene−asphaltene rearrangements leading to aggregation are more thermodynamically favorable than asphaltene−surface inter2846

dx.doi.org/10.1021/ef500282p | Energy Fuels 2014, 28, 2831−2856

Energy & Fuels

Review

significantly time dependent.169 This phenomena may have more to do with the deposition of asphaltenes at the surface rather than adsorption. Again by QMC-D, adsorption of crude dissolved in various n-alkanes (C7, C10, C12) showed that increasing the number of carbons slightly increases the absorbed amount, but by AFM the size of aggregates slightly decreased with increasing carbon number. It was reasoned that a narrower range of molecular weights are sustainable by longer alkanes producing smaller aggregates that form more compact layers at the surface. In toluene, the aggregates were reported to be more ordered and hydrophilic, but in heptane they were more amorphous, which may cause their polar functional groups to become caged within the aggregate structure. But for a smaller more selective range of aggregates, such as those produced by increasing the n-alkane length to C12, the aggregates are more polar and their functional groups are less blocked. This was confirmed by the increase in the O and S content detected by XPS. Adsorption from toluene gave a rigid monolayer with a thickness of about 3−4 nm that was insensitive to concentration and that remained rigid over time. The adsorbed layers produced from solutions of n-alkanes or heptol gave loser viscoelastic films that increased in viscoelasticity with time because the solvents were able to penetrate the aggregates. The viscoelasticity was found to be maximized near the flocculation point for heptol solutions.113 Swelling of asphaltene aggregates due to n-alkanes and the adsorption of n-alkanes on asphaltene aggregates are known to occur for aggregates suspended in solution,228 so it is not unreasonable that the same affects should be present for asphaltene aggregates adsorbed to a surface.106

action leading to adsorption. The increased energy of aggregation efficiently outcompetes asphaltene adsorption at the surface.118 Most studies report that increasing the solvent strength effectively decreases aggregate size and/or reduces adsorption by competing for surface active sites. For asphaltenes containing metal porphyrins adsorption increased in the order of pyridine and acetonitrile (both showed no adsorption), followed by benzene, and then by cyclohexane which showed significant adsorption.185 Adsorption onto glass plates showed a greater buildup of asphaltenes with decreasing solubility parameter of the solvent in the following order: chloroform < THF < toluene < heptol, and further increases were reported with increasing proportions of heptane for the heptol solutions.173 The correlation between decreasing Hildebrand solubility parameter and increased asphaltene adsorption onto silica surfaces modified with self-assembled monolayers was also observed in the following order: 1-methylnaphthalene < tetralin < toluene < decalin < heptol (20:80 heptane/toluene).166 Elsewhere, a similar trend was reported for asphaltene adsorption on hydrophilic silica particles: 1-methylnaphthalene < xylene ≤ toluene < heptane/toluene (10:90) < heptane/ toluene (25:75).73 The reverse trend for increased adsorption capacity with increasing solvent strength has rarely been observed. For the adsorption of vanadium-containing asphaltenes onto alumina, it was reported that adsorption increased in the following order: pyridine < cyclohexane < toluene.125 Similarly, in an initial study using QCM-D to quantify adsorption onto gold from a 1:1 heptol solution, it was reported that adsorption of compact asphaltene layers had occurred, leveling off at around 1000 mg/ L, but for toluene no plateau was observed and adsorption continued to increase up to 10 000 mg/L. It was proposed that in the case of toluene, multilayers were formed which were removed by rinsing to a greater extent (10%) than layers produced from heptol. The fact that larger multilayers could be rinsed off to a greater degree suggested that the asphaltene− asphaltene interactions are weaker than the surface−asphaltene interactions.176 Data from a similar study using QCM-D showed that the adsorption capacity of asphaltenes on hydrophilic silica, alumina, and TiO2 were sometimes greater from toluene solutions compared to 1:1 heptol solutions.121 Follow up studies using QMC-D and crude in 1:1 heptol solutions do indeed show continuous adsorption of multilayers as was initially reported for toluene.106,177 It was also confirmed that as the size of aggregates increased the aggregates become more loosely packed, 121 and as the colloidal stability approached flocculation, the largest most voluminous aggregates were observed adsorbing to the surface.106,113,227 In a similar study, increasing the amount of pentane precipitant sequentially increased the amount of asphaltenes adsorbed up to the point of flocculation. Above the flocculation threshold, adsorption rapidly increased yielding a very large amount of asphaltenes adsorbed at the surfaces (160 mg/m2 for gold, 18.79 for hydrophilic silica, and 11.37 for hydrophobic silica).106 Additionally, atomic force microscopy (AFM) was used to confirm the growth of larger voluminous aggregates near the onset of precipitation, which also showed that these aggregates were more heterogeneous.227 Others have observed by neutron reflectivity the same phenomenonthat near the flocculation threshold significantly more material is adsorbed and multilayer organization dominates. Beyond the flocculation threshold, asphaltene adsorption occurs faster and becomes

10. MODIFICATION OF THE SORBENT PARTICLES UPON ADSORPTION, AND PROPERTIES OF ADSORBED ASPHALTENE AGGREGATES There are a variety studies that are helping to clarify the complex changes that occur at surfaces upon asphaltene adsorption, which are not only important regarding adsorption onto stationary surfaces but are important in understanding how indigenous particles suspended in oil modified with asphaltenes affect refining, emulsions, and reservoir quality. It is well documented that as asphaltene adsorption increases the hydrophobicity of sorbents also increase, a key issue in reservoir wettability. This is usually demonstrated as an increase in the contact angle and zeta potential,99,139,140,165,229,230 or by photoelectron spectroscopy and time-of-flight secondary ion mass spectroscopy,136 as the adsorbed layer effectively shields the surface charge energy. As adsorption increases particles become more hydrophobic, their immersion behavior changes accordingly.99,140,141,165,210,230 However, their electrophoretic mobility may not be sufficiently changed due to partial surface coverage by asphaltenes.134,156,217 When studying the variables of time, temperature, concentration, and their combinations on suspension stability of a variety of surfaces, it was found that coating hydrophilic silica or kaolin stabilizes the suspension of these particles in water due to steric stabilization (repulsive interactions of the asphaltene shells on the particles keep the particles from collecting together). However for FeS, favorable hydrophobic interactions between the asphaltenes of different particles caused particle destabilization. This was explained by the preferential organization of the asphaltene layers on the particles. This phenomena may be guided by, and specific to, the sorbent surface.230 2847

dx.doi.org/10.1021/ef500282p | Energy Fuels 2014, 28, 2831−2856

Energy & Fuels

Review

produced using an upstroke Langmuir−Blodgett technique. The technique applies a force to order asphaltenes at an oil/ water interface while pulling a silica wafer through the interface, which causes the polar groups of asphaltenes to interact with the polar silica surface leaving behind a well ordered compact hydrophobic layer of asphaltenes.232−234 Others however have reported that the polar orientations of the adsorbed layers are not well organized. Adsorption onto hydrophilic and hydrophobic silica did not show any preferential orientation of asphaltenes: polar groups toward the surface producing a highly hydrophobic shell or the reverse orientation giving a hydrophilic shell, respectively.119 This is probably due to the source and type of asphaltenes adsorbed, since more aromatic asphaltene molecules are not like amphiphilic molecules comprised of well defined polar headgroups and aliphatic tails (most asphaltenes are not like surfactant type molecules, but rather some of the naphthenic acids which can report to this fraction can be surfactant-like).97 As judged by AFM, adsorption of the asphaltene aggregates on a surface of mica was shown to be highly heterogeneous118 in appearance and coverage which varied with time.208 From toluene solutions, adsorbed asphaltene aggregates were observed to grow in size with time, and they continued to organize at the surface.208 The preparation of the asphaltene solutions can also affect the aggregates found on the surface: unfiltered samples were found to give relatively larger fractal aggregates.208 For glass and silica surfaces, adsorption was found to increase as a function of time for a given concentration, leading the authors to conclude that there was a continuous increase in the size of asphaltene aggregates which leads to multilayer formation.68 Along these same lines, SEM images of adsorbed asphaltenes showed that increased adsorption correlated with increased size of aggregates and not an increasing number of aggregates, suggesting a limited number of nucleation sites necessary for specific asphaltene sorbent interactions.117 SEM,117 time-of-flight secondary ion mass spectroscopy (ToF-SIMS),101,148,212 and photoelectron spectroscopy in conjunction with ToF-SIMS have also shown that the surface coverage after adsorption is limited to a number of nucleation sites and that adsorbed aggregates are highly heterogeneous.136 On the other hand, studies on metal surfaces using AFM have shown that surface saturation is achieved by increasing the number of adsorbed aggregatesnot the size of aggregatesfor the concentration range investigated (up to 1500 mg/L).118 Asphaltene aggregates adsorbed onto a surface are generally accepted to be patchy and heterogeneous.148,227 Adsorbed aggregates are considerably denser and associated with less solvent than those found in the solution.169 Despite this densification, the adsorbed nanoaggregates were found to be permeable containing porosities from 0.85 to 0.90, which may also be consistent with a polymer brush structure.106 While studying asphaltene adsorption onto iron surfaces at concentrations within the nanoaggregate range, it was determined by AFM that the surface aggregates were as large as ∼50 nm high and a couple hundred nanometers wide.118 In a similar study, slightly smaller aggregates were observed for aggregates on mica (10−20 nm high).6 From a study using an ellipsometry method measuring asphaltenes adsorbed onto glass, it was reported that the thickness of adsorbed asphaltenes spanned 20−298 nm.175 These values are far beyond the typical values given for solution nanoaggregates or clusters of nanoaggregates outlined by the modified Yen−Mullins

Asphaltene adsorption from water-saturated toluene solutions onto montmorillonite gave results similar to the asphaltene-coated FeS particles. SEM showed that the overall size of small particles of montmorillonites increased upon adsorption. For this system, the negatively charged carboxylate asphaltenes, which were more aliphatic by IR, were preferentially adsorbed at the positive surfaces of the montmorillonite, and the stability of the water suspensions of the asphaltene-coated particles was decreased due to favorable asphaltene−asphaltene interactions between different coated particles.154 For clay minerals, the adsorption of asphaltenes blocks ion exchange sites reducing their cation exchange capacity, expandability, surface area,154,231 and surface charge as observed by microelectrophoresis.154 The adsorption of asphaltenes causes a buildup of polar functional groups, which result in interparticle electrostatic repulsive forces. In organic solvents, when these electrostatic repulsive forces combine with the repulsive steric forces of the asphaltene alkyl side chains, the result is an increase in the colloidal stability of coated particles relative to a noncoated hydrophilic particles.219 Similar results were found for small hydrophilic silica particles, which had a diameter of about 12 μm: they were weakly flocculated in toluene but well dispersed after the adsorption of asphaltenes.73 The source of asphaltenes, adsorption conditions, and charge behavior and hydrophilicity of the particles are important for determining which types of asphaltenes adsorb, the properties of the adsorbed asphaltene aggregates, and how the resulting surfaces become modified. A mechanism for how the stability of asphaltenes affects particle behavior was recently proposed. It states that as the stability of the asphaltenes, within the oil matrix, approaches or goes beyond the point of flocculation, then the resulting asphaltene-coated particles also become significantly destabilized. Using AFM, it was determined that the size of solution aggregates is relative to the size of adsorbed aggregates, which is also relative to the resulting energy of adhesion for the adsorbed aggregates. In other words, it was demonstrated that adding precipitating solvent to the point of asphaltene flocculation caused larger aggregates to form in solution, which caused larger aggregates to be adsorbed. These larger adsorbed aggregates in turn overcome steric repulsions between particles and, due to their enhanced ability to stretch, facilitate bridging between the particles resulting in a driving force that enhances the flocculation of asphaltene coated particlesan important conclusion for understanding how particles or emulsions behave when adding different concentrations of precipitant.227 At surface solution interfaces, it has been reported that preferential ordering takes place for the hydrophilic and hydrophobic portions of asphaltenes upon adsorption. Over several hydrophilic clay minerals and reservoir rocks surfaces, asphaltenes initially adsorb with their more polar surfaces pointing toward the surface causing a hydrophobic shell of asphaltenes. The adsorption of a second layer of asphaltenes onto the first hydrophobic shell can result in a layer that is more hydrophilic due to the buildup of outward facing polar functional groups. This suggests that a layer of reverse orientation relative to the first layer is adsorbed.141 For this study, it is important to point out that the authors had concluded, based upon IR, that the adsorbed asphaltenes were mostly aliphatic and amphiphilic and thus coated the particles similar to a surfactant. The initial ordering and compacting of amphiphilic asphaltenes is similar to asphaltene aggregate films 2848

dx.doi.org/10.1021/ef500282p | Energy Fuels 2014, 28, 2831−2856

Energy & Fuels

Review

model: 2.0 and 5.0 nm, respectively.70 The large size of aggregates observed by AFM on micarelative to the monomers (0.84 relative fraction), dimers (0.07 relative fraction), and trimers observed in the 100 mg/L solution by polarizing fluorescencewas attributed to capillary aggregation. In this study, it was observed that the concentration of asphaltenes was significantly higher within the pores of the sorbent than in the solution. This led to the following mechanistic conclusion: adsorption initially occurs at a rough spot within the micropore of the surface, and as asphaltene molecules begin saturating the micropore the contact points between the neighboring asphaltene molecules have the highest surface curvature making them preferential centers for adsorption leading to accelerated monomer aggregation.6 A similar argument was made for multilayer formation on silica by strongly associating asphaltenes: adsorption of aggregates can cause additional sites for adsorption which are similar to the process of growing asphaltene aggregates in solution.68 Most studies on the other hand have reported values of adsorbed aggregate sizes that are similar to those described by the Yen−Mullins model. In a more recent AFM study, asphaltene aggregates adsorbed from varying pentane to toluene ratios onto gold gave aggregate sizes between 3 and 8 nm high. Within this range, the magnitude of elasticity and heterogeneity were found to increase with the amount of pentane added until the onset of precipitation. Below the threshold of precipitation, the height of the aggregates were between 3 and 5 nm, and above this threshold they were 8 nm high, highly heterogeneous, and elastic.227 Surface aggregate size determined by asymptotic analysis of early adsorption kinetics using QCM-D gave similar values: 0.5−1.6 nm for primary adsorbing species at low concentrations (139−278 ppm asphaltenes) and 2.6−5.6 nm at higher concentrations (835 ppm asphaltenes) for toluene diluted crude sample.113 Another study using a combined QMC XPS approach for adsorbed aggregates on a gold substrate gave similar thicknesses ranging from 6 to 8 nm,178 which were similar to those adsorbed at oil/water interfaces of emulsions (2−9 nm)235 and also similar to those reported by a XPS ToF-SIMS study for asphaltenes on kaolinite and silica, which give a maximum asphaltene aggregate height of 11 nm and mean aggregate thickness of 3 nm.136 Other similar results were obtained by AFM from the adsorption from model toluene solutions (46 mg/L) onto gold surfaces, which showed the majority of aggregates to be between 2 and 6 nm. Some larger aggregates up to 40 nm were observed, but they were attributed to interparticle aggregation at the surface or from additional deposition from the evaporation of the residual asphaltene solution. It should be noted that of the 5 asphaltenes in this last study, Chevron C did not follow the observed trend: when adsorbing 0.5 wt% solutions of Tensleep oil and Chevron C onto gold, AFM showed the Tensleep-treated surface to be covered with typical monolayer asphaltenes, whereas Chevron C gave significant amounts of large aggregates as a multilayer that was highly heterogeneous.236 This study shows the importance of considering the source of asphaltenes.

correlations among the data can be inconsistent. Some show a positive correlation between adsorption capacity and affinity, while others show a negative correlation, and these correlations are highly dependent on which type of isotherm is chosen.124,170,188 In general, most of the isotherms reported are either of the Langmuir or Freundlich type. The Langmuir type I isotherm represents homogeneous monolayer coverage;237 although asphaltene aggregates are nonhomogenous,6 it is a reasonably “effective” model for a compact layer of asphaltenes aggregates adsorbed at low concentrations.118 On the other hand, the Freundlich isotherm models multilayer formation of a heterogeneous surface.238 In addition to the Freundlich170 model and Freundlich with an association Kiselev term,135,150,239,240 several other modifications and other multilayer models have been reported to fit the complex behavior observed for multilayer formation.68,92,135,150,239,241 Clementz first observed a simple Langmuir Type I isotherm for asphaltene adsorption when studying adsorption onto montmorillonite.110 Since then, several other researchers have also observed monolayer Langmuir type isotherms (refs 6, 73, 77, 101, 107, 112, 117−119, 134, 139, 141, 143−145, 148, 149, 159, 178, 188−190, 217, 219, and 229). In some reported cases, the data fit equally well to both Langmuir and Freundlich isotherms.105,179,241 The type of sorbent can have a significant effect on the isotherms, even at low concentrations.177 For the mineral-based sorbents, kaolinite, quartz, and calcite Langmuir type I isotherms gave a good fit, but for dolomite multilayer behavior was fit to a modified Langmuir isotherm by adding a concentration-dependent factor.146 For the metals gold, stainless steel, and aluminum, QCM-D data showed possible multilayer formation.210 Similarly, for gold at early times steady monolayer Langmuir isotherms were reported, but at much longer times multilayer adsorption was observed.177 It should be noted that some researchers have pointed out that in some cases the apparent observed monolayer Langmuir isotherms are fortuitous conclusions.119 It should also be noted that QMC data often gives multilayer formation relative to other methods for the same sorbents.121 In one study using silica gel, it was shown that the resulting monolayer Langmuir isotherms were affected by the source of the asphaltenes, the amount of resins precipitated with the asphaltenes, and the strength of the solvent used to dissolve the asphaltenes. In line with others’ observations, an increased adsorption capacity was observed for weaker solvents and for asphaltenes containing less resins.73 Isotherms can also be affected by sample preparation and reaction conditions. Shale oil inherently contains large amounts of fine material, and filtering does not remove enough of these contaminants, which results in multilayer Freundlich isotherms, but when the oil is extracted with a solvent, excluding most of the fines, a monolayer Langmuir isotherm is observed.24 The nature of the solvent also plays an important role in the type of isotherm observed. In one study, nonpolar solvents resulted in Langmuir Type I isotherms, whereas for the same asphaltenes in nitrobenzene partial ionization resulted in multilayer Langmuir Type II adsorption.112 In strong solvents that do not cause ionization, Langmuir monolayer behavior is typically observed, but when the solvent mixture becomes poor, through addition of alkanes, multilayer behavior is observed, especially when approaching flocculation.169 Isotherms have also been reported to be sensitive to particle size. Nanoalumina was reported to give a monolayer Langmuir isotherm, but for

11. ASPHALTENE ADSORPTION ISOTHERMS A detailed discussion about isotherms is beyond the scope of this review, but it is important to point out some general observations. Isotherms give important information about the adsorption affinity and adsorption capacity of the sorbent− asphaltene systems. As with other adsorption data, general 2849

dx.doi.org/10.1021/ef500282p | Energy Fuels 2014, 28, 2831−2856

Energy & Fuels

Review

Figure 6. At the left is a general representation for a simple adsorption isotherm most commonly reported. At the right is a general representation of a multicomponent isotherm attributed strong stepwise multilayer adsorption.

microalumina a Freundlich isotherm was observed.124 Isotherm data from NiO nanoparticles supported on nanoparticulated silica were also fitted to a Freundlich multilayer isotherm.104 Static and dynamic adsorption conditions can also affect the isotherm. In a study on adsorption on dolomite sand, the static experiments showed typical monolayer Langmuir isotherms, but under flow conditions multilayer behavior was observed.115 At high concentrations (above around 3000−5000 mg/L), and depending on the oil’s source and the sorbent, multiple adsorption regimes for an isotherm can be observed, which correspond to the different adsorption rates of different aggregation states.109,114,135,150,174,175,239,240 The adsorption behavior of these multicomponent isotherms are often described as multilayers, and in some cases the adsorption capacity can be above 100 mg/m2.135,150 Figure 6 shows a generic representation of a simple isotherm most often observed relative to a multicomponent isotherm. For asphaltenes constrained within the porous sorbents of silica, calcite, and dolomite, the adsorption was determined to be monolayer coverage between 0 and 2500 mg/L, bilayer between 2500 and 5000 (and in one case 8000) mg/L, transition monolayers between 5000 (8000)−15 000 mg/L, and larger multilayers between 15 000 and 30 000 mg/L.109 Interestingly, it has also been reported that hematite exhibits a multistep isotherm resulting in an adsorption capacity of 5 mg/ m2 at low concentrations (up to 600 mg/L).144 Likewise, for inherently unstable oilsknown to cause precipitation within pipelinesat concentrations below 3000 mg/L multicomponent adsorption was observed.68,92 For these highly aromatic oils with H/C ratios less than 1, strong multilayer isotherms were observed. From this work, it was suggested that isotherm data might be used to predict whether an oil would be inherently unstable: a multicomponent isotherm would deem the oil unstable, or if it gave a Langmuir Type I isotherm it would be classified as stable.92 For other well behaved sources of asphaltenes the concentration does not appear to be the primary driving force for multicomponent behavior since some studies have shown single-component isotherms up to 40 000 mg/L.108

physical adsorption (deposition).6,138,151 In a recent AFM study using gold surfaces, it was reported that after initial adsorption the surface became irreversibly contaminated by a strongly adsorbed layer of asphaltenes that could not be rinsed off with toluene or chloroform.227 In a different study using pure toluene as a desorption solvent, a less than 3% increase in absorbance of the solution was reported by UV after 50 h.68 Clementz first showed that adsorbed asphaltenes on montmorillonite were unable to be displaced by any detectible amount when using benzene or CS2 and that the strong azeotropic solvent mixture of the chloroform/acetone was able to extract some of the material (up to 55%).111110 Chloroform/ acetone was also found to be a much better solvent than toluene for extracting the most binding basic nitrogen material from sandstone containing clay.205 Other solvents such as CCl4,139 toluene,6,101,107,148,165,173 benzene,133 or CHCl3 were ineffective for removing adsorbed asphaltenes,227 while solvents such as CH2Cl2, CHCl3, and THF were reported to be partially effective.115,173,227 For asphaltenes on carbon sorbents, THF was shown to be selective for desorbing nonmetal containing asphaltenes.191 In the case of adsorbed asphaltenes on mica, some desorption occurred with toluene, less with decalin, and even less with cyclohexane.164 Mixtures containing alcohols or ethyl acetate were found to partially or quantitatively remove adsorbed asphaltenes: CHCl3/MeOH azeotrope (78.7:21.3),112 toluene/EtOH (20:80), toluene/MeOH (85:15),39 toluene/ MeOH (25−50% MeOH by volume),39 cyclohexane/ethyl acetate (70−30:30−70), cyclohexane/EtOH (10:90 or 25:75), cyclohexane/2-propanol (67:33).35 It has also been demonstrated that petroleum fractions produced from catalytic cracking of crudes, containing 10−80 wt % aromatics, can be isolated and reused as an asphaltene desorption medium.36,38 Very sensitive QCM techniques have demonstrated that when using the same solvent as was used for adsorption only a small amount of the weakest adsorbing material can be rinsed off from silica or alumina surfaces,232 and the same was shown for stainless steel, aluminum, and gold.210,232 In a different study using QCM-D, it was shown that the asphaltenes adsorbed to a gold surface from toluene were bound tightly, and about 10−20% of asphaltenes could be desorbed. The degree of desorption was reported to be dependent on how compact the adsorbed aggregates were.121,176 It has been demonstrated that for larger more voluminous aggregates adsorption was partially reversible, but for smaller aggregates adsorption was irreversible.166,176,177 As in other studies,

12. DESORPTION OF ADSORBED ASPHALTENES BY SOLVENTS AND AMPHIPHILES Adsorption of asphaltenes from solution is largely irreversible from the medium in which the adsorption took place and is therefore considered more within the realm of chemical adsorption, similar to a chromatographic process, rather than 2850

dx.doi.org/10.1021/ef500282p | Energy Fuels 2014, 28, 2831−2856

Energy & Fuels

Review

for enhanced oil recovery efforts, which will not be discussed further in this review.

desorbing asphaltenes from silica or alumina using pure toluene gave minimal desorption as detected by QCM-D.195 To study the organic material adsorbed onto deposits of solid asphaltenes formed at Venezuelan facilities, a two-stage extraction process was developed to remove most of the organic material since pure CH2Cl2 extraction attempts left significant organic materials trapped within the deposits. First a (90:10) CH2Cl2/MeOH solution was used, which was then followed by a (90:10) toluene/2-propanol. The alcohol solutions were deemed necessary to remove very polar adsorbed asphaltenic material, while toluene, due to its higher operation temperature, was selected to remove very large alkanes (waxes).182 When desorbing asphaltenic metal porphyrins from the Lewis acid catalyst CoMoAl2O3, cyclohexane was completely ineffective, benzene showed slight removal, and complete desorption was accomplished using pyridine or acetonitrile.185 From acidic silica−alumina or WO3 supported on alumina, THF was able to remove nonbasic asphaltenes, and a second desorption using pyridine quantitatively removed the basic asphaltenes.49 Others have also observed that pyridine is a good solvent for desorption.112,133 The peptizing solvent N-methyl-2pyrrolidone (NMP) under flowing conditions at 80 °C for 2 h was demonstrated to completely restore the original adsorption capacity of silica−alumina by removing basic nitrogencontaining asphaltenes.41 Elsewhere, addition of up to 10% of NMP in toluene significantly increased asphaltene desorption, and an inverse relationship between desorption and pressure was reported.130 Desorbing solvent media must have two key properties: strong chromatographic interaction with sorbent active sites to displace the adsorbed asphaltenes and be able to readily dissolve the desorbed asphaltenes. Therefore, water by itself is generally not a good solvent for desorbing asphaltenes since asphaltenes are poorly soluble in it, but solvent mixtures with water can be good for desorbing asphaltenes. Mixtures of 10% H2O in THF or 5% in pyridine were reported to be potent mixtures for removing the heaviest material from alumina after a chromatographic separation of coal pyrolysis liquid.45 The effect of pure water for desorbing asphaltenes is dependent on the properties of the surface. For Fe2O3 modified surfaces, with dominant H-bonding characteristics, water was reported to be the most efficient at displacing asphaltenes even when compared to several solvent mixtures and commercial additives.183 For industrial applications, hot water44 and particularly steam are employed to regenerate sorbents, which are sometimes coupled with a supplemental desorbing stream such as carbon dioxide.36 With regard to using amphiphiles, often used as asphaltene peptizing agents to prevent precipitation or as competitors for asphaltene adsorption sites, they were found to be ineffective for removing already adsorbed asphaltenes from quartz.23 For asphaltenes adsorbed onto glass, the surfactant sodium dodecylsulfate and water did not desorb any asphaltenes but rather adsorbed onto the already adsorbed asphaltene aggregates.173 However, when treating asphaltenes adsorbed on silica or alumina surfaces with the amphiphile ethyl cellulose, it was found to compete for surface active sites and effectively push away and desorb adsorbed asphaltenes.232 It should be noted that there is a significant body of research using amphiphiles to alter the wettability of reservoir rock material aimed at reducing plugging or to unplug porous reservoir surfaces due to asphaltene adsorption.242 This is very important

13. CONCLUSION Asphaltene adsorption onto surfaces is a ubiquitous phenomenon for petroleum occurring within reservoirs and throughout the entire production and processing chain. This phenomenon can be tuned to optimize asphaltene sequestration from petroleum sources within the production well or at junctures further downstreamwhich is essential to efficiently utilize the vast global unconventional oil resources and to assist with enhanced oil recovery efforts. Asphaltene adsorption is a viable method for removing the most corrosive, fouling, coke forming, and viscous building components in heavy oil and other unconventional sources. Adsorption is practical because sorbents can be regenerated through nonconsumptive desorption processes restoring the activity of the sorbent for many subsequent cycles. Understanding and optimizing sorbent-asphaltene interactions will have a tremendous economic impact for commercializing asphaltene adsorption processes. Congruencies between several different studies show that the key elements of adsorption, with regard to asphaltene selectivity, are the amount of heteroatoms, types of functional groups present, degree of aromaticity, the size of the aromatic core, degree of processing (amount of cleaved alkyl side chains), acid−base characteristics, amount of resinous material, and molecular weight, all of which contribute to the polarity and steric interactions of the individual molecules and their resulting aggregates. These factors ultimately dictate how asphaltenes interact at a particular surface. Increasing polarity, through heteroatom content or increased aromaticity, and decreasing steric hindrance allow the surface active portions of asphaltenes to more favorably interact at surfaces. Several studies have shown that the functional groups consisting of particularly N and O heteroatoms, and to a lesser degree S, are concentrated in the adsorbed asphaltenes and may be significant for initiating surface-specific adsorption interactions. Regarding sorbents contacted with low concentrations of asphaltenes in aromatic solvents, the sorbents surface area, pore characteristics, surface acidity, surface charge, cation exchange capacity (clay minerals), hydrophilicity, water content, number of active sites, size, acid−base character, and chemical composition all affect the overall amount adsorbed. To selectively remove the most polar and problematic apshaltenes from oil or model solutions a likewise highly charged polar sorbent is needed with high acid or base activity and preferably a high number of hydroxyl or other active binding sites. More generally, adsorption is enhanced with decreasing particle size, increasing surface area, and increasing sorbent porosity and pore diameter which prevents diffusion problems and pore plugging. Asphaltene adsorption is irreversible with respect to the oil or solvent system in which adsorption occurred, and desorption is effective only with solvents and mixtures that readily dissolve asphaltenes and are chromatographically strong enough to efficiently compete for surface adsorption sites. For well behaved asphaltenes, laboratory results have shown that the adsorption capacity over most sorbents was limited to under 4 mg/m2, which is near the maximum of 3.6 mg/m2 calculated for continuous converge assuming 3 nm aggregates with a density of 1200 kg/m3. Several studies have shown that adsorbed asphaltene aggregates are similar in size to solution 2851

dx.doi.org/10.1021/ef500282p | Energy Fuels 2014, 28, 2831−2856

Energy & Fuels



aggregates. For asphaltene solutions within the nanoaggregate concentration regime, many studies have shown adsorbed aggregates to be roughly the same size as those in the Yen− Mullins model: 2 and 5 nm for nanoaggregates and clusters, respectively. In the absence of other surface characterization data, for asphaltenes in good solvents within this concentration regime, values under about 3.6 mg/m2 should be considered an “effective” monolayer. Under similar conditions, adsorption capacities above the effective monolayer adsorption capacity can be considered to be an “effective” multilayer; however, it should be kept in mind that “effective” multilayer behavior may only be a manifestation of larger sized aggregates. For asphaltenes that are not prone to precipitation problems, due to unusually strong associations, several situations have been reported under which asphaltenes can be coaxed to undergo effective multilayer adsorption in the nanoaggregate regime. Decreasing the solvent power of the matrix in which the asphaltenes are dissolved causes larger aggregates to form in solution which are consequently adsorbed; for solutions near or at the point of flocculation adsorption capacities are diagnostic of effective multilayers. The highly disperse nature of nanoparticles, especially when coupled with metal oxide selectivity, combine with favorable diffusion rates to give high adsorption capacities. Iron oxide has been shown to adversely affect asphaltene stability in general, and asphaltene adsorption onto hematite has been demonstrated to give effective multilayers and multistage isotherms. On the other hand, for oils which are known to cause precipitation problems in the field, such as highly aromatic oils and oils with highly associating asphaltenes, effective multilayer behavior and multicomponent isotherms are more often observed. Asphaltenes themselves are often regarded as a byproduct and nuisance that must be necessarily mitigated to optimize refining and upgrading. Asphaltene byproducts are often converted into coke and burned to return some of the energy necessary for the refining process. This method of disposing of asphaltenes continues to become costly due to tightening regulations that limit pollutants emitted from these combustion processes. Growing consciousness about environmental impacts, the need to utilize finite natural resources judiciously, and profit driven ingenuity have been pushing the industry to find innovative ways to further upgrade asphaltenes and/or convert them into valuable feedstocks and chemical commodities. Selective removal and fractionation of asphaltenes will be necessary to produce homogeneous feedstocks making them easier to upgrade or to find niches in other chemical industries.



REFERENCES

(1) Rana, M. S.; Samano, V.; Ancheyta, J.; Diaz, J. A. I. Fuel 2007, 86, 1216−1231. (2) Chopra, S.; Lines, L. Soc. Explor. Geophys. 2008, 27, 1104−1106. (3) Cruz, J. L. M. d. l.; Arguelles-Vivas, F. J.; Matias-Perez, V.; DuranValencia, C. d. l. A.; Lopez-Ramirez, S. Energy Fuels 2009, 23, 5611− 5626. (4) Drummond, C.; Israelachvili, J. J. Pet. Sci. Eng. 2004, 45, 61−81. (5) Buckley, J. S.; Liu, Y. J. Pet. Sci. Eng. 1998, 20, 155−160. (6) Syunyaev, R. Z.; Balabin, R. M.; Akhatov, I. S.; Safieva, J. O. Energy Fuels 2009, 23, 1230−1236. (7) Vafaie-Sefti, M.; Mousavi-Dehghani, S. A. Fluid Phase Equilib. 2006, 247, 182−189. (8) Akbarzadeh, K.; Hammami, A.; Kharrat, A.; Zhang, D.; Allenson, S.; Creek, J.; Kabir, S.; Jamaluddin, A.; Marshall, A. G.; Rodgers, R. P.; Mullins, O. C.; Solbakken, T. Oilfield Rev. 2007, 22−43. (9) Cosultchi, A.; Rossbach, R.; Hernandez-Calderon, I. Surf. Interface Anal. 2003, 35, 239−245. (10) Kokal, S. SPE Intl. 2005, SPE 77497, 1-13. (11) Gawel, I.; Bociarska, D.; Biskupski, P. Appl. Catal., A 2005, 295, 89−94. (12) Bartholomew, C. H. Appl. Catal., A 2001, 212, 17−60. (13) O’Connor, P.; Takatsuka, T.; Woolery, G. L., Deactivation of Zeolite Catalysts by Coke. In Deactivation and Testing of HydrocarbonProcessing Catalysts; Masuda, T.; Hashimoto, K., Eds.; American Chemical Society: Washington, DC, 1996; Vol. 634, pp 62−76. (14) Marchal, C.; Abdessalem, E.; Tayakout-Fayolle, M.; Uzio, D. Energy Fuels 2010, 24, 4290−4300. (15) Absi-Halabi, M.; Stanislaus, A.; Trimm, D. L. Appl. Catal. 1991, 72, 193−215. (16) Faus, F. M.; Frange, P.; Delmon, B. Appl. Catal. 1984, 11, 281− 293. (17) Argillier, J.-F.; Henaut, I.; Gateau, P.; Heraud, J.-P. SPE International Thermal Operations and Heavy Oil Symposium; November 1−3, Calgary, Alberta, Canada, 2005, SPE 97763; Society of Petroleum Engineers: Richardson, Texas, 2005; pp 1−7. (18) Barcenas, M.; Orea, P.; Buenrostro-Gonzalez, E.; ZamudioRivera, L. S.; Duda, Y. Energy Fuels 2008, 22, 1917−1922. (19) Chang, C.-L.; Fogler, H. S. Fuel Sci. Technol. Int. 1996, 14, 75− 100. (20) Leon, O.; Rogel, E.; Urbina, A.; Andujar, A.; Lucas, A. Langmuir 1999, 15, 7653−7657. (21) Barcenas, M.; Orea, P.; Buenorostro-Gonzalez, E.; ZamudioRivera, L. S.; Duda, Y. Energy Fuels 2008, 22, 1917−1922. (22) Ostlund, J.-A.; Nyden, M.; Fogler, J. S.; Holmberg, K. Colloids Surf., A 2004, 243, 95−102. (23) Gonzalez, G.; Middea, A. Colloids Surf. 1991, 52, 207−217. (24) Gonzalez, G.; Moreira, M. B. C. Aspahltenes Asphalts 1994, 40, 207−231. (25) Bilden, D. M.; Jones, V. E. Aspahltene Adsorption Inhibition Treatment. U.S. Patent 6051535, April 18, 2000. (26) Hasan, S. W.; Ghannam, M. T.; Esmail, N. Fuel 2010, 89, 1095− 1100. (27) Hawkins, D. J.; Perry, G. F. Oil Gas J. 2002, Oct. 28, 64−69. (28) Yeung, C. K. System for the Decontamination of Asphaltic Heavy Oil and Bitumen. U.S. Patent 7625466, December 1, 2009. (29) Wolk, R. H.; Nongbri, G.; Alpert, S. B. Pentane Insoluble Asphaltene Removal. U.S. Patent 3948756, April 6, 1976. (30) Romanova, U. G.; Yarranton, H. W.; Schramm, L. L.; Shelfantook, W. E. Can. J. Chem. Eng. 2004, 82, 710−721. (31) Zhao, Y.; Wei, F. Fuel Process. Technol. 2008, 89, 941−948. (32) Duong, A.; Smith, K. J. Can. J. Chem. Eng. 1997, 75, 1122−1129. (33) Wang, H.-l.; Wang, G.; Shen, W.-j.; Xu, C.-m.; Gao, J.-s. Ind. Eng. Chem. Res. 2011, 50, 12501−12511. (34) Osaheni, J. A.; Fyvie, T. J.; O’Neil, G. A.; Matis, H. Methods and System for Removing Impurities from Heavy Fuel. U.S. Patent 8088277, May 29, 2012.

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.



Review

ACKNOWLEDGMENTS

The author would like to thank Shell International Exploration and Production for funding this literature review. The author would also like to thank Dr. Frans G. A. van den Berg and Dr. John Schabron for fruitful discussions and guidance with editing this paper. 2852

dx.doi.org/10.1021/ef500282p | Energy Fuels 2014, 28, 2831−2856

Energy & Fuels

Review

(35) Osaheni, J. A.; Fyvie, T. J.; Haitko, D. A.; O’Neil, G. A.; Glaser, P. B. Methods for Regeneration of Adsorbent Material. U.S. Patent 8187991, May 29, 2012. (36) Janssen, M. J.; Ou, J. D. Y.; Heeter, G. A.; Oorshot, C. W. M. V. Removal of Asphaltene Contaminants from Hydrocarbon Streams Using Carbon Based Adsorbents. U.S. Patent 2012/0132566, 2012. (37) Osaheni, J. A.; Bablin, J. M.; Haitko, D. A.; Soloveichik, G. L. Methods and Systems for Removing Metals from Low Grade Fuel. U.S. Patent 7947167, May 24, 2011. (38) Ou, J. D. Y.; Strak, R. D. Apparatus and Process for Cracking Hydrocarbonaceous Feed Treated to Adsob Parafin-Insoluble Compound. U.S. Patent 2009/0156876, June 18, 2009. (39) Gorbaty, M. L.; Errughelli, T. T.; Olmstead, W. N.; Miseo, S.; Soled, S. L.; Robbins, W. K. Selective Adsorption Process for Resid Upgrading. U.S. Patent 6245223, June 12, 2001. (40) Kuehl, G. H. Removal of Large Molecules from a Fuild. U.S. Patent 5583277, December 10, 1996. (41) Yao, K. C. Removal of Basic Nitrogen Compounds from Extracted Oils by Use of Acidic Polar Adsorbents and the Regeneration of Said Adsorbents. U.S. Patent 4846962, July 11, 1989. (42) Ngan, D. Y. Process for Removal of Polynuclear Aromatics from Hydrocarbon in an Endothermic Reformer Reaction System. U.S. 4804457, February 14, 1989, 1989. (43) Reno, M. E. Hydrocracking Process with Feed Pretreatment. U.S. Patent 4775460, October 4, 1988. (44) Johnson, R. W.; Hribik, W. f.; Hilfman, L. Process for Hydrotreating Hydrocarbonaceous Charge Stock. U.S. Patent 4719007, January 12, 1988. (45) Long, R. B.; Griffel, J. Selective Removal of Coke Precursor from Hydrocarbon Feedstock. U.S. Patent 4624776, November 25, 1986. (46) Gould, K. A.; Clark, G. D.; Long, R. B.; Murrell, L. L.; Pieters, W. J. M. Separating Basic Asphaltenes Using Metal Oxide Acid Catalysts. EP Patent 0076090, July 24, 1985. (47) Murrell, L. L.; Grenoble, D. C.; Long, R. B. Separating Basic Asphaltenes Using Transition Metal Oxide Acid Catalysts. U.S. Patent 4424114, January 3, 1984. (48) Long, R. B.; Caruso, F. A. Adsorption of Basic Asphaltenes on Solid Acid Catalysts. U.S. Patent 4419219, December 6, 1983. (49) Gould, K. A.; Grenoble, D. C.; Murrell, L. L.; Pieters, W. J. M. Separating Basic Asphaltenes using Bronstead Acid Transition Metal Oxide Acid Catalysts. U.S. Patent 4422926, December 27, 1983. (50) Woodle, R. A. Adsorption-Desorption Process for Removing Unwanted Comonent from a Reaction Charge Mixture. U.S. Patent 3767563, October 23, 1973. (51) Rankin, J. P.; Vreeland, J. L.; Litz, K. E.; Jordan, T. M.; Rossetti, M. N.; Burnett, E. H. Methods for Upgrading of Contaminated Hydrocarbon Streams. U.S. Patent 8197671, June 12, 2012. (52) Litz, K. E.; Vreeland, J. L.; Rankin, J. P.; DeLancey, T. W.; Thompson, T. A. Reaction System and Products Therefrom. U.S. Patent 2012/0067777, March 22, 2012. (53) BGC Engineering Inc., Oil Sands Tailings Technology Review; Oil Sands Research and Information Network, University of Alberta, School of Energy and the Environment: Edmonton, Alberta, OSRIN Report No. TR-1, 2010; 136 pp. (54) Yeung, C. K. Decontamination of Asphaltic Heavy Oil and Bitumen. U.S. Patent 2010/0116716, May 13, 2010. (55) Yeung, C. K. System for the Decontamination of Asphaltic Heavy Oil and Bitumen. U.S. Patent 7,625,466, December 1, 2009. (56) Mullins, O. C.; Sheu, E. Y.; Hammami, A.; Marshall, A. G. Asphaltenes, Heavy Oils, and Petroleomics; Springer: New York, 2007. (57) Sabbah, H.; Marrow, A. L.; Pomerantz, A. E.; Zare, R. N. Energy Fuels 2011, 25, 1597−1604. (58) Liu, P.; Shi, Q.; Chung, K. H.; Zhang, Y.; Pan, N.; Zhao, S.; Xu, C. Energy Fuels 2010, 24, 5089−5096. (59) Shi, Q.; Hou, D.; Chung, K. H.; Xu, C.; Zhao, S.; Zhang, Y. Energy Fuels 2010, 24, 2545−2553. (60) Acevedo, S.; Guzmán, K.; Labrador, H.; Carrier, H.; Bouyssiere, B.; Lobinski, R. Energy Fuels 2012, 26, 4968−4977.

(61) Schabron, J. F.; R., J. F., Jr.; Sanderson, M. M.; Loveridge, J. L.; Nyadong, L.; McKenna, A. M.; Marshall, A. G. Energy Fuels 2012, 26, 2256−2268. (62) Mullins, O. C.; Sheu, E. Y. Structures and Dynamics of Asphaltenes; Plenum Press: New York, 1998. (63) Mullins, O. C.; Betancourt, S. S.; Cribbs, M. E.; Dubost, F. X.; Creek, J. L.; Andrews, A. B.; Venkataramanan, L. Energy Fuels 2007, 21, 2785−2794. (64) Erdman, J. G.; Harju, P. H. J. Chem. Eng. Data 1963, 8, 252− 258. (65) McKenna, A. M.; Purcell, J. M.; Rodgers, R. P.; Marshall, A. G. Energy Fuels 2009, 23, 2122−2128. (66) Dechaine, G. P.; Gray, M. R. Energy Fuels 2010, 24, 2795−2808. (67) Andersen, S. I. Association of Petroleum Asphaltenes and the Effect on Solution Properties. In Handbook of Surface and Colloid Chemistry, 3d ed.; Birdi, K. S., Ed.; CRC Press, Taylor and Francis Group: Boca Raton, FL, 2009; pp 703−718. (68) Acevedo, S.; Ranaudo, M. A.; Garcia, C.; Castillo, J.; Fernandez, A.; Caetano, M.; Goncalvez, S. Colloids Surf., A 2000, 166, 145−152. (69) Mostowfi, F.; Indo, K.; Mullins, O. C.; McFarlane, R. Energy Fuels 2009, 23, 1194−1200. (70) Mullins, O. C.; Sabbah, H.; Eyssautier, J.; Pomarantz, A. E.; Barre, L.; Andrews, A. B.; Ruiz-Morales, Y.; Mostowfi, F.; McFarlane, R.; Goual, L.; Lepkowicz, R.; Cooper, T.; Orbulescu, J.; Leblanc, R. M.; Edwards, J.; Zare, R. N. Energy Fuels 2012, 26, 3986−4003. (71) Mullins, O. C. Annu. Rev. Anal. Chem. 2011, 4, 393−418. (72) Wang, S.; Liu, J.; Zhang, L.; Masliyah, J.; Xu, Z. Langmuir 2010, 26, 183−190. (73) Simon, S.; Jestin, J.; Palermo, T.; Barre, L. Energy Fuels 2009, 23, 306−313. (74) Spiecker, P. M.; Gawrys, K. L.; Trail, C. B.; Kilpatrick, P. K. Colloids Surf., A 2003, 220, 9−27. (75) Yarranton, H. W.; Alboudwarej, H.; Jakher, R. Ind. Eng. Chem. Res. 2000, 39, 2916−2924. (76) Tanaka, R.; Sato, E.; Hunt, J. E.; Winans, R. E.; Sato, S.; Takanohashi, T. Energy Fuels 2004, 18, 1118−1125. (77) Faus, F. M.; Grang, P.; Delmon, B. Appl. Catal. 1984, 11, 281− 293. (78) Mullins, O. C.; Seifert, D. J.; Zuo, J. Y.; Zeybek, M. Energy Fuels 2013, 27, 1752−1761. (79) Mullins, O. C. Energy Fuels 2010, 24, 2179−2207. (80) Gray, M. R.; Tykwinski, R. R.; Stryker, J. M.; Tan, X. Energy Fuels 2011, 25, 3125−3134. (81) Murgich, J. Pet. Sci. Technol. 2002, 20, 983−997. (82) Sedghi, M.; Goual, L.; Welch, W.; Kubelka, J. J. Phys. Chem. B 2013, 117, 5765−5776. (83) Porte, G.; Shou, H.; Lazzeri, V. Langmuir 2003, 19, 40−47. (84) Kok, W. T.; Tudos, A. J.; Grutters, M.; Shephard, A. G. Energy Fuels 2011, 25, 208−214. (85) Altgelt, K. H.; Harle, O. L. Ind. Eng. Chem. Prod. Res. Dev. 1975, 14, 240−246. (86) Pierre, C.; Barre, L.; Pina, A.; Moan, M. Oil Gas Sci. Technol. 2004, 59, 489−501. (87) Kharrat, A. M. Energy Fuels 2009, 23, 828−834. (88) Nalwaya, V.; Tantayakom, V.; Piumsomboon, P.; Fogler, S. Ind. Eng. Chem. Res. 1999, 38, 964−972. (89) Joshi, N. B.; Mullins, O. C.; Jamaluddin, A.; Creek, J.; McFadden, J. Energy Fuels 2001, 15, 979−986. (90) Cosultchi, A.; Garciafigueroa, e.; Mar, B.; Garcia-Borquez, A.; Lara, V. H.; Bosch, P. Fuel 2002, 81, 413−421. (91) Wattana, P.; Fogler, H. S. Energy Fuels 2005, 19, 101−110. (92) Acevedo, S.; Ranaudo, M. A.; Escobar, G.; Gutierrez, L.; Ortega, P. Fuel 1995, 74, 595−598. (93) Koots, J. A.; Speight, J. G. Fuel 1975, 54, 179−184. (94) Leon, O.; Contreras, E.; Rogel, E.; Dambakli, G.; Acevedo, S.; Carbognani-Arambarri, L.; Espidel, J. Langmuir 2002, 18, 5106−5112. (95) Pfeiffer, J. P.; Saal, R. N. J. J. Phys. Chem. 1940, 44, 139−149. (96) Rogel, E. Energy Fuels 2008, 22, 3922−3929. 2853

dx.doi.org/10.1021/ef500282p | Energy Fuels 2014, 28, 2831−2856

Energy & Fuels

Review

(97) Czarnecki, J.; Tchoukov, P.; Dabros, T. Energy Fuels 2012, 26, 5782−5786. (98) Tu, Y.; Woods, J.; McCraken, T.; Kotlyar, L.; Sparks, B.; Chung, K. Clay Sci. 2006, 12, 188−193. (99) Zahabi, A.; Gray, M. R.; Czarnecki, J.; Dabros, T. Energy Fuels 2010, 24, 3616−3623. (100) Tu, Y.; Woods, J.; Kung, J.; McCraken, T.; Kotlyar, L.; Sparks, B.; Mingzhe, D. Clay Sci. 2006, 12, 183−187. (101) Tu, Y.; Kung, J.; McCraken, T.; Kotlyar, L.; Kingston, D.; Sparks, B. Clay Sci. 2006, 12, 194−198. (102) Pourabdollah, K.; Moghaddam, A. Z.; Kharrat, R.; Mokhtari, B. Oil Gas Sci. Technol. 2011, 66, 1005−1016. (103) Nassar, N. N.; Hassan, A.; Pereira-Almao, P. Energy Fuels 2011, 25, 1566−1570. (104) Bortes, F. B.; Ejia, J. M.; Ruiz, M. A.; Benjumea, P.; Riffel, D. B. Energy Fuels 2012, 26, 1725−1730. (105) Franco, C. A.; Nassar, N. N.; Ruiz, M. A.; Pereira-Almao, P.; Cortes, F. B. Energy Fuels 2013, 27, 2899−2907. (106) Zahabi, A.; Gray, M. R. Energy Fuels 2012, 26, 1009−1018. (107) Acevedo, S.; Ranaudo, M. A.; Garcia, C.; Castillo, J.; Fernandez, A. Energy Fuels 2003, 17, 257−261. (108) Tayakout, M.; Ferreira, C.; Espinat, D.; Picon, S. A.; Sorbier, L.; Guillaume, D.; Guibard, I. Chem. Eng. Sci. 2010, 65, 1571−1583. (109) Castro, M.; Cruz, J. L. M. d. l.; Buenrostro-Gonzalez, E.; Lopez-Ramirz, S.; Gil-Villegas, A. Fluid Phase Equilib. 2009, 286, 113− 119. (110) Clementz, D. M. Clays Clay Minerals 1976, 74, 312−319. (111) Czarnecka, E.; Gillott, J. E. Clays Clay Minerals 1980, 28, 197− 203. (112) Dubey, S. T.; Waxman, M. H. SPE Res. Eng. 1991, 389−395. (113) Adubu, A.; Goual, L. Energy Fuels 2009, 23, 1237−1248. (114) Acevedo, S.; Castillo, J.; Fernandez, A.; Goncalvez, S.; Ranaudo, M. A. Energy Fuels 1998, 12, 386−390. (115) Piro, G.; Galbariggi, G.; Bertero, L.; Carniani, C. SPE Prod. Facil. 1996, 30109, 156−160. (116) Saraji, S.; Goual, L.; Piri, M. Energy Fuels 2010, 24, 6009−6017. (117) Alboudwarej, H.; Pole, D.; Svrek, W. Y.; Yarranton, H. W. Ind. Eng. Chem. Res. 2005, 44, 5585−5592. (118) Balabin, R. M.; Syunyaev, R. Z.; Schmid, T.; Stadler, J.; Lomakina, E. I.; Zenobi, R. Energy Fuels 2011, 25, 189−196. (119) Dudasova, D.; Simon, S.; Hemmingsen, P. V.; Sjoblom, J. Colloids Surf., A 2008, 317, 1−9. (120) Xing, C.; Hilts, R. W.; Shaw, J. M. Energy Fuels 2010, 24, 2500−2513. (121) Dudasova, D.; Silset, A.; Sjoblom, J. J. Dispersion Sci. Technol. 2008, 29, 139−146. (122) Hassan, A.; Lopez-Linares, F.; Nassar, N. N.; CarbognaniArambarri, L.; Pereira-Almao, P. Catal. Today 2013, 207, 112−118. (123) Deng, S. Sorbent Technology. In Encyclopedia of Chemical Processing; Taylor and Francis: U.K., 2006; pp 2825−2845. (124) Nassar, N. N.; Hassan, A.; Pereira-Almao, P. Energy Fuels 2011, 25, 3961−3965. (125) Saint-Just, J. Ind. Eng. Chem. Prod. Res. Dev. 1980, 19, 71−75. (126) Shedid, S. A.; Abbas, E. A. A. Pet. Sci. Technol. 2006, 24, 1457− 1467. (127) Cruz, J. L. M. d. l.; Arguelles-Vivas, F. J.; Matias-Perez, V.; Duran-Valencia, C. d. l. A.; Lpez-Ramirez, S. Energy Fuels 2009, 23, 5611−5625. (128) Lopez-Linares, F.; Carbognani, L.; Sosa-Stull, C.; PereiraAlmao, P.; Spencer, R. J. Energy Fuels 2009, 23, 1901−1908. (129) Drushel, H. V. Prepr. Pap.-Am. Chem. Soc., Div. Pet. Chem. 1972, 17, 4. (130) Pernyeszi, T.; Dekany, I. Colloids Surf., A 2001, 194, 25−39. (131) Carbognani, L.; Gonzalez, M. F.; Lopez-Linares, F.; Sosa-Stull, C.; Pereira-Almao, P. Energy Fuels 2008, 22, 1739−1746. (132) Ching, M.-J. T. M.; Pomerantz, A. E.; Andrews, A. B.; Dryden, P.; Schroeder, R.; Mullins, O. C.; Harrison, C. Energy Fuels 2010, 24, 5028−5037.

(133) Plancher, H.; Dorrence, S. M.; Petersen, J. C. Proc. Assoc. Asphalt Paving Technol. 1977, 46, 151−175. (134) Kokal, S.; Tang, T.; Schramm, L.; Sayegh, S. Colloids Surf., A 1995, 94, 235−265. (135) Marczwski, A. W.; Szymula, M. Colloids Surf., A 2002, 2008, 259−266. (136) Wang, S.; Liu, Q.; Tan, X.; Xu, C.; Gray, M. R. Energy Fuels 2013, 27, 2465−2473. (137) Akbarzadeh, K.; Hammami, A.; Kharrat, A.; Zhang, D.; Allenson, S.; Creek, J.; Kabir, S.; Jamaluddin, A.; Marshall, A. G.; Rodgers, R. P.; Mullins, O. C.; Solbakken, T. Oilfield Rev. 2007, Summer, 22−43. (138) Somasundaran, P.; Krishnakumar, S. Colloids Surf., A 1997, 123−124, 491−513. (139) Fritschy, G.; Papirer, E. Fuel 1978, 57, 701−704. (140) Jada, A.; Debih, H. Composit Interfaces 2009, 16, 219−235. (141) Pernyeszi, T.; Patzko, A.; Berkesi, O.; Dekany, I. Colloids Surf., A 1998, 137, 373−384. (142) Cosultchi, A.; Cordova, I.; Valenzuela, M. A.; Acosta, D. R.; Bosch, P.; Lara, V. H. Energy Fuels 2005, 19, 1417−1424. (143) Collins, S. H.; Melrose, J. C. Proceedings of the SPE International Symposium on Oilfield and Geothermal Chemistry; Denver, CO, June 1−3, 1983; SPE 11800. (144) Gonzalez, G.; Moreira, M. B. C. Colloids Surf. 1991, 58, 293− 302. (145) Gonzalez, G.; Travalloni-Louvisse, A. M. SPE Prod. Facil. 1993, 91−96. (146) Mohammadi, M.; Mousavi-Dehghani, S. A. ACS National Meeting, 2012. (147) Saada, A.; Siffert, B.; Papirer, E. J. Colloid Interface Sci. 1995, 174, 185−190. (148) Tu, Y.; Kingston, D.; Kung, J.; Kotlyar, L. S.; Sparks, B. D.; Chung, K. H. Pet. Sci. Technol. 2006, 24, 327−338. (149) Curtis, C. W.; Joen, Y. W.; Clapp, D. J. Fuel Sci. Technol. Intl. 1989, 7 (9), 1225−1268. (150) Cruz, J. L. M. d. l.; Castellanos-Ramirez, I. V.; Ortiz-Tapia, A.; Buenrostro-Gonzalez, D.; Duran-Valencia, C. d. l. A.; Lopez-Ramirez, S. Colloids Surf., A 2009, 340, 149−154. (151) Lopez-Linares, F.; Carbognani, L.; Spencer, R. J.; PereiraAlmao, P. Energy Fuels 2011, 25, 3657−3662. (152) Tombacz, E.; Szekeres, M. App. Clay Sci. 2006, 34, 105−124. (153) Bantignies, J.-L.; Moulin, C. C. d.; Dexpert, H. J. Pet. Sci. Eng. 1998, 20, 233−237. (154) Jada, A.; Debih, H.; Khodja, M. J. Pet. Sci. Eng. 2006, 52, 305− 316. (155) Kuiyavar, S. I.; Vetrivel, R.; Hegde, S. G.; Ramaswamy, A. V.; Chakrabarty, D.; Mahapatra, S. J. Mater. Chem. 2000, 10, 1835−1840. (156) Gonzalez, G.; Middea, A. Colloids Surf. 1988, 33, 217−229. (157) Jiang, T.; Hirasaki, G. J.; Miller, C. A.; Ng, S. Energy Fuels 2011, 25, 545−554. (158) Bantignies, J.-L.; Moulin, C. C. d.; Dexpert, H.; Flank, A.-M.; Williams, G. C. R. Acad. Sci. Paris 1995, 320, 699−706. (159) Baboriau, H.; Saada, A. Chemosphere 2001, 44, 1633−1639. (160) Lopez-Linares, F.; Carbognani, L.; Hassan, A.; Pereira-Almao, P.; Rogel, E.; Ovalles, C.; Pradhan, A.; Zintsmaster, J. Energy Fuels 2011, 25, 4049−4054. (161) Dean, K. R.; McAtte, J. L. Appl. Clay Sci. 1986, 1, 313−319. (162) Kotlyar, L. S.; Sparks, B. D.; Woods, J. R. Energy Fuels 1999, 13, 346−350. (163) Jiang, T.; Hirasaki, G. J.; Miller, C. A.; Ng, S. Energy Fuels 2011, 25, 2551−2558. (164) Kumar, K.; Dao, E.; Mohanty, K. K. J. Colloid Interface Sci. 2005, 289, 206−217. (165) Hannisdal, A.; Ese, M.-H.; Hemmingsen, P. V.; Sjoblom, J. Colloids Surf., A 2006, 276, 45−58. (166) Turgman-Cohen, S.; Fisher, D. A.; Kilpatrick, P. K.; Genzer, J. App. Mater. Interfaces 2009, 1, 1347−1357. (167) Farooq, U.; Sjoblom, J.; Oye, G. J. Dispersion Sci. Technol. 2011, 32, 1388−1395. 2854

dx.doi.org/10.1021/ef500282p | Energy Fuels 2014, 28, 2831−2856

Energy & Fuels

Review

(168) Turgman-Cohen, S.; Smith, M. B.; Fisher, D. A.; Kilpatrick, P. K.; Genzer, J. Langmuir 2009, 25, 6260−6269. (169) Jouault, N.; Corvis, Y.; Cousin, F.; Jestin, J.; Barre, L. Langmuir 2009, 25, 3991−3998. (170) Nassar, N. N.; Hassan, A.; Pereira-Almao, P. J. Colloid Interface Sci. 2011, 360, 233−238. (171) Araujo, R. S.; Azevedo, D. C. S.; Cavalcente, C. L.; JimenezLopez, A.; Rodriquez-Castellon, E. Microporous Mesoporous Mater. 2008, 108, 213−222. (172) Schabron, J. F.; Boysen, R. B.; Kalberer, E. W.; Rovani, J. F. Hydrocarbon Separation and Analysis Apparatus and Method. U.S. Patent 20130067991, March 21, 2013. (173) Akhalaq, M. S.; Gotze, P.; Kessel, D.; Dornow, W. Colloids Surf., A 1997, 126, 25−32. (174) Castillo, J.; Goncalvez, S.; Fernandez, A.; Mujica, V. Opt. Commun. 1998, 145, 69−75. (175) Labrador, H.; Fernandez, Y.; Torvar, J.; Munoz, R.; Pereira, J. C. Energy Fuels 2007, 21, 1226−1230. (176) Ekholm, P.; Blomberg, E.; Claesoon, P.; Auflem, I. H.; Sjoblom, J.; Kornfeldt, A. J. Colloid Interface Sci. 2002, 247, 342−350. (177) Goual, L.; Horvath-Szabo, G.; Masliyah, J. H.; Xu, Z. Langmuir 2005, 21, 8278−8289. (178) Rudrake, A.; Karan, K.; Horton, J. H. J. Colloid Interface Sci. 2009, 332, 22−31. (179) Jeon, Y. W.; Yi, S. H.; Choi, S. J. Fuel Sci. Technol. Intl. 1995, 13, 195−214. (180) Garrouch, A. A.; Malallah, A. H.; Al-Enizy, M. M. Proceedings of the SPE Annual Technical Conference and Exhibition; San Antonio, TX, September 24−27, 2006; SPE 102129. (181) Rogel, E.; Roye, M. Prepr. Pap.-Am. Chem. Soc., Div. Energy Fules 2013, 58, 509−511. (182) Carbognani, L.; Orea, M.; Fonseca, M. Energy Fuels 1999, 13, 351−358. (183) Carbognani, L. Pet. Sci. Technol. 2000, 18, 335−360. (184) Alvarez-Ramirez, F.; Garcia-Cruz, I.; Travizon, G.; MartinezMagadan, J. M. Pet. Sci. Technol. 2004, 22, 915−926. (185) Morales, A.; Galiasso, R. Fuel 1982, 61, 13−17. (186) Nassar, N. N.; Hassan, A.; Pereira-Almao, P. Energy Fuels 2011, 25, 1566−1570. (187) Nassar, N. N.; Hassan, A.; Pereira-Almao, P. Colloids Surf., A 2011, 384, 145−149. (188) Nassar, N. N.; Hassan, A.; Pereira-Almao, P. Energy Fuels 2011, 25, 1017−1023. (189) Nassar, N. N.; Hassan, A.; Carbognani, L.; Lopez-Linares, F.; Pereira-Almao, P. Fuel 2012, 95, 257−262. (190) Franco, C.; Patino, E.; Benjumea, P.; Ruiz, M. A.; Cortes, F. B. Fuel 2013, 105, 408−414. (191) Dakanishi, K.; Saitio, I.; Watanable, I.; Mochida, I. Fuel 2004, 83, 1889−1893. (192) Ramos, A. C. d. S.; Haraguchi, L.; Notrispe, F. R.; Loh, W.; Mohamed, R. S. J. Pet. Sci. Eng. 2001, 32, 201−216. (193) Headley, J. V.; Peru, K. M.; Mohamed, M. H.; Wilson, L.; McMarticn, D. W.; Mapolelo, M. M.; Lobodin, V. V.; Rodgers, R. P.; Marshall, A. G. Energy Fuels 2013, 27, 1772−1778. (194) Mohamed, M. H.; Wilson, L. D.; Headley, J. V.; Peru, K. M. Process Saf. Environ. Prot. 2008, 86, 237−243. (195) Wang, S.; Segin, N.; Wang, K.; Masliyah, J. H.; Xu, Z. J. Phys. Chem. C 2011, 115, 10576−10587. (196) Zhuravlev, L. T. Colloids Surf., A 2000, 173, 1−38. (197) Hashmi, S. M.; Firoozabadi, A. Soft Matter 2011, 7, 8384− 8391. (198) Salou, M.; Siffert, B.; Jada, A. Fuel 1998, 77, 343−346. (199) Salou, M.; Siffert, B.; Jada, A. Fuel 1998, 77, 339−341. (200) Rayes, B. H.; Pernyeszi, T.; Lakatos, I.; Toth, J.; Proceedings of the SPE International Symposium on Oilfield Chemistry; Houston, TX, February 5−8, 2003; SPE 80265. (201) Stanford, L. A.; Rodgers, R. P.; Marshall, A. G.; Czarnecki, J.; Wu, X. A.; Taylor, S. Energy Fuels 2007, 21, 973−981.

(202) Stanford, L. A.; Rodgers, R. P.; Marshall, A. G.; Czarnecki, J.; Wu, X. A. Energy Fuels 2007, 21, 963−972. (203) Hashimi, S. M.; Firoozabadi, A. Soft Matter 2011, 7, 8384− 8391. (204) Goual, L.; Horvath-Szabo, G.; Masliyah, J. H.; Xu, Z. Energy Fuels 2006, 20, 2099−2108. (205) Reed, M. G. Clays Clay Miner. 1968, 16, 173−178. (206) Gonzalez, M. F.; Sosa-Stull, C.; Lopez-Linares, F.; PereiraAlmao, P. Energy Fuels 2007, 21, 234−241. (207) Lopez-Linares, F.; Carbognani, L.; Gonzalez, M. F.; Sosa-Stull, C.; Figueras, M.; Pereira-Almao, P. Energy Fuels 2006, 20, 2748−2750. (208) Toulhoat, H.; Prayer, C.; Rouquet, G. Colloids Surf., A 1994, 91, 267−283. (209) Strausz, O. P.; Peng, P.; Murgich, J. Prepr. Pap.-Am. Chem. Soc., Div. Fuel Chem. 1999, 44, 743−745. (210) Xie, K.; Karan, K. Energy Fuels 2005, 19, 1252−1260. (211) Abdallah, W. A.; Taylor, S. D. Nucl. Instrum. Methods Phys. Res., Sect. B 2007, 258, 213−217. (212) Abdallah, W. A.; Taylor, S. D. J. Phys. Chem. C 2008, 112, 18963−18972. (213) Costa, L. M. d.; Stoyanov, S. R.; Gusarov, S.; Tan, X.; Gray, M. R.; Stryker, Y. M.; Tykwinski, R.; Carneiro, J. W. d. M.; Seidl, P. R.; Kovalenko, A. Energy Fuels 2012, 26, 2727−2735. (214) Schabron, J. F.; Rovani, J. F.; Sanderson, M. M. Energy Fuels 2010, 24, 5984−5996. (215) Labadidi, H. M. S.; Sabti, H. M.; AlHumaidan, F. S. Fuel 2014, 117, 59−67. (216) Li, M.; Larter, S. R.; Frolov, Y. B. J. High Resol. Chromatogr. 1994, 17, 230−236. (217) Gonzalez, G.; Middea, A. J. Dispersion Sci. Technol. 1987, 8, 525−548. (218) Balabin, R. M.; Syunyaev, R. Z. J. Colloid Interface Sci. 2008, 318, 167−174. (219) Marlow, B. J.; Sresty, G. C.; Hughes, T. D.; Mahajan, O. P. Colloids Surf. 1987, 24, 283−297. (220) Acevedo, S.; Garcia, L. A. Energy Fuels 2012, 26, 1814−1819. (221) Andersen, S. I.; Rio, J. M. d.; Khvostitchenko, D.; Shakir, S.; Lira-Galeana, C. Langmuir 2001, 17, 307−313. (222) Murgich, J. Langmuir 2002, 18, 9080−9086. (223) Horvath-Szabo, G.; Masliyah, J.; Czarnecki, J. Can. J. Chem. Eng. 2004, 82, 1089−1095. (224) Petersen, J. C.; Plancher, H. Pet. Sci. Technol. 1998, 16, 89− 131. (225) Moschopedis, S. E.; Fryer, J. F.; Speight, J. G. Fuel 1976, 55, 227−232. (226) Spiecker, P. M.; Gawrys, K. L.; Kilpatrick, P. K. J. Colloid Interface Sci. 2003, 267, 178−193. (227) Zahabi, A.; Gray, M. R. Energy Fuels 2012, 26, 2891−2898. (228) Nikooyeh, K.; Bagheri, S. R.; Shaw, J. M. Energy Fuels 2012, 26, 1756−1766. (229) Menon, V. B.; Wasan, D. T. Colloids Surf. 1987, 23, 353−362. (230) Dudasova, D.; Flaten, G. R.; Sjoblom, J.; Oye, G. Colloids Surf., A 2009, 335, 62−72. (231) Clementz, D. M. SPE/DOE 1982, 10683, 131−138. (232) Wang, S. University of Alberta, Thesis, 2011. (233) Wang, S.; Liu, J.; Zhang, L.; Masliyah, J.; Xu, Z. Langmuir 2010, 26, 183−190. (234) Wang, S.; Liu, J.; Zhang, L.; Xu, Z.; Masliyah, J. Energy Fuels 2009, 23, 862−869. (235) Sztukowski, D. M.; Jafari, M.; Alboudwarej, H.; Yarranton, H. W. J. Colloid Interface Sci. 2003, 265, 179−186. (236) Goual, L.; Abudu, A. Energy Fuels 2010, 24, 469−474. (237) Langmuir, I. J. Am. Chem. Soc. 1916, 38, 2221−2295. (238) Rouquerol, F.; Rouquerol, J.; Sing, K. Adsorption by Powders and Porous Solids: Principles, Methodology and Applications; Academic Press: London (UK), 1999. (239) Szymula, M.; Marczweski, A. W. Appl. Surf. Sci. 2002, 196, 301−311. 2855

dx.doi.org/10.1021/ef500282p | Energy Fuels 2014, 28, 2831−2856

Energy & Fuels

Review

(240) Marczewski, A. W.; Szymula, M. Ann. Univ. Mariae CurieSklodowska Lublin-Polonia 2003, 4, 69−79. (241) Giraldo, J.; Nassar, N. N.; Benjumea, P.; Pereira-Almao, P.; Cortes, F. B. Energy Fuels 2013, 27, 2908−2914. (242) Bilden, D. M.; Jones, E. Asphaltene Adsorption Inhibition Treatment. U.S. Patent 6051535, April 18, 2000.



NOTE ADDED AFTER ASAP PUBLICATION This article published April 16, 2014 incorrectly citing an author surname in references 128, 131, 151, 160, 189, and 207. The correct version published April 24, 2014.

2856

dx.doi.org/10.1021/ef500282p | Energy Fuels 2014, 28, 2831−2856