Interactions between Asphaltenes and Water in Solutions in Toluene

3096

Energy & Fuels 2008, 22, 3096–3103

Interactions between Asphaltenes and Water in Solutions in Toluene† Daria S. Khvostichenko‡ and Simon I. Andersen*,§ Department of Chemistry, St. Petersburg State UniVersity, 199034 St. Petersburg, Russia, and IVC-SEP, Department of Chemical Engineering, Technical UniVersity of Denmark, DK-2800 Lyngby, Denmark ReceiVed December 13, 2007. ReVised Manuscript ReceiVed April 15, 2008

Binding of water by asphaltenes dissolved in toluene was investigated for two asphaltene samples, OMV1 and OMV2, from the same reservoir deposit. Solubility of water in asphaltene solutions in toluene was found to increase with an increasing asphaltene concentration, indicative of solubilization of water by asphaltenes. Fourier transform infrared (FTIR) spectroscopy of stretching modes of OH groups in the region of 3800-3100 cm-1 was used to obtain insight into the state of water in water-unsaturated asphaltene solutions in toluene. The number of water molecules bound to one asphaltene molecule was determined for water-saturated solutions (OMV1 and OMV2) and for water-unsaturated solutions (OMV1 only). These numbers were found to decrease from several water molecules per asphaltene molecule to below unity upon an increase of the asphaltene concentration in toluene from 0.1 to 20 g/L, suggesting that binding of water by asphaltenes dramatically depends upon the aggregated state of asphaltenes and that the mechanism of water solubilization by asphaltenes differs significantly from water solubilization in reverse micelles formed by conventional surfactants.

Introduction Asphaltenes belong to the heavy end of petroleum distillates and are separated as solids by the addition of an excess amount of n-alkanes to crude oil.1 Among constituents of petroleum that cause problems in recovery and refinery operations, asphaltenes stand out as a fraction having a wide range of undesirable properties, from their tendency to form a precipitate in recovery operations2 to their ability to stabilize water-in-oil emulsions3 and poison catalysts used in refineries.2 Characteristic structural features of asphaltenic molecules are a fused aromatic system decorated with saturated side chains and naphthenic fragments and the presence of heteroatoms (sulfur, oxygen, and nitrogen) that account for polarity of asphaltene molecules.1 The formation of molecular aggregates by asphaltenes in solvents and crude oils is a well-recognized phenomenon confirmed by measurements of small-angle neutron scattering,4,5 surface tension,5–7 and calorimetric titration,8–10 but the mechanism remains a matter of debate. Understanding the behavior of † Presented at the Third International Conference on Petroleum Phase Behavior and Fouling. * To whom correspondence should be addressed. Telephone: +45-4527-2113. Fax: +45-47-20-2999. E-mail: [email protected]. ‡ St. Petersburg State University. Current address: Department of Chemistry, University of Illinois, 600 S. Mathews Ave., Urbana, IL 61801. § Technical University of Denmark. Current address: R&D Refinery, Haldor Topsøe A/S, Nymøllevej 55, DK-2800 Kgs. Lyngby, Denmark. (1) Speight, J. G. The Chemistry and Technology of Petroleum, 3rd ed.; Marcel Dekker: New York, 1999. (2) Sheu, E. Y. Energy Fuels 2002, 16, 74. (3) Sjoblom, J.; Aske, N.; Auflem, I. H.; Brandal, O.; Havre, T. E.; Saether, O.; Westvik, A.; Johnsen, E. E.; Kallevik, H. AdV. Colloid Interface Sci. 2003, 399, 100–102. (4) Gawrys, K. L.; Kilpatrick, P. K. J. Colloid Interface Sci. 2005, 288, 325. (5) Storm, D. A.; Sheu, E. Y. Fuel 1995, 74, 1140. (6) Sheu, E. Y.; De Tar, M. M.; Storm, D. A.; DeCanio, S. J. Fuel 1992, 71, 299. (7) Rogel, E.; Leon, O.; Torres, G.; Espidel, J. Fuel 2000, 79, 1389. (8) Murgich, J.; Merino-Garcia, D.; Andersen, S. I.; del Rio, J. M.; Lira Galeana, C. Langmuir 2002, 18, 9080.

asphaltenes at the molecular level is a challenging task because of the high polydispersity of the structure and size of the constituents of this fraction.11 Water at trace levels is present in most technical and natural systems, including oil reservoirs, where the contact time between water and oil is of the magnitude of geological time, allowing the system to reach equilibrium conditions. Trace amounts of water in apolar solvents can significantly alter the aggregative behavior of surfactants,12 yet the interaction between asphaltenes and water dissolVed in an apolar solvent has received relatively little attention (stabilization of water-in-oil emulsions by asphaltenes, however, has been studied extensively).3 One study found that the solubility of water in asphaltene solutions in toluene increased with an increasing asphaltene concentration, suggesting a solubilization-type mechanism of interaction.13 It was also found in that work that the presence of dissolved water in toluene was required for observation of a CMC-like transition in asphaltene solutions in toluene in calorimetric titration experiments13 (CMC is the critical micellization concentration or, more accurately, a narrow range of concentrations of a surfactant in solution, defined as “separating the limit below which virtually no micelles are detected and the limit above which virtually all additional surfactant molecules form micelles”).14 Although the latter result was not confirmed in later studies with a more sensitive instrument,8,9 thorough control of humidity was required to ensure reproducibility of calorimetric (9) Merino-Garcia, D.; Andersen, S. I. J. Pet. Sci. Technol. 2003, 21, 507. (10) Merino-Garcia, D.; Andersen, S. I. J. Dispersion Sci. Technol. 2005, 26, 217. (11) Klein, G. C.; Kim, S.; Rodgers, R. P.; Marshall, A. G.; Yen, A.; Asomaning, S. Energy Fuels 2006, 20, 1965. (12) Eicke, H.-F. In Interfacial Phenomena in Apolar Media; Eicke, H.-F., Parfitt, G. D., Eds.; Marcel Dekker: New York, 1997. (13) Andersen, S. I.; del Rio, J. M.; Khvostitchenko, D.; Shakir, S.; Lira-Galeana, C. Langmuir 2001, 17, 307. (14) Muller, P. Pure Appl. Chem. 1994, 66, 1077.

10.1021/ef700757h CCC: $40.75  2008 American Chemical Society Published on Web 07/29/2008

Asphaltene-Water Interactions in Toluene

Energy & Fuels, Vol. 22, No. 5, 2008 3097

titrations,9 thus indirectly indicating the importance of interactions between water and asphaltenes. The scope of the present work is to gain an understanding of the mechanism of the interaction between asphaltenes and water in solutions in toluene. Two different approaches have been used: (a) measurements of solubilization of water by asphaltenes dissolved in toluene and (b) investigation of the state of water in asphaltene solutions in toluene at different stages of water saturation based on the analysis of IR adsorption bands of hydroxyl groups in the region of 3800-3100 cm-1. Because the spectra reflect a multitude of hydrogen-bonding features, the intention of the present work is to assess trends rather than to calculate exact numbers. In combination with the ambiguity regarding the distribution of molecular types within the asphaltene fraction, interpretation of data on the observed interactions between water and asphaltenes in solutions in toluene can only provide a simplified picture.

IR spectra of all samples were taken on a Perkin-Elmer PARAGON 1000 Fourier transform infrared (FTIR) spectrometer at room temperature. Spectra of all OMV1 asphaltene solutions and of the OMV2 solution in CS2 were taken using a Specac cell with a path length of 0.5 mm with NaCl windows. Spectra of the OMV2 asphaltene solutions in toluene were taken using a Specac cell with a path length of 0.2 mm with NaCl windows. Every sample was given 256 scans at a resolution of 4 cm-1. To eliminate the contribution from the absorption by the solvent, the spectrum of dried toluene was used as the background. The Microcal Origin software package was used for curve-fitting of the hydroxyl stretching region (3800-3100 cm-1) with three Gaussian peaks. All experiments, including equilibration of asphaltene solutions in toluene with the aqueous phase, measurement of water content, and recording of IR spectra were performed at room temperature (23 ( 1 °C).

Experimental Section

Differences between the OMV1 and OMV2 samples can be assessed by analyzing their IR spectra. Because of the large variations of absorbance values in spectral regions corresponding to various structural fragments (Figure 1a), which is typical of asphaltene spectra,16 ratios of absorbances (Figure 1b) rather than the absolute values in Figure 1a appeared most instructive. IR band assignments were made on the basis of literature data.16–18 The main change in the OMV2 asphaltene compared to the OMV1 is the increase of the relative content of oxygenated functional groups: the ratio of absorbances in the spectral regions corresponding to CdO and sulfoxide groups was higher than that in the aliphatic and aromatic regions. Oxidation of the OMV2 sample was probably a consequence of storage of the unpurified raw material in the air atmosphere for an additional year after the purification of the OMV1 batch. At the same time, similar levels of the curve in the aliphatic and aromatic absorbance regions (Figure 1b) showed that the ratio of saturated to aromatic fragments within the sample had not changed appreciably. Petersen et al. reported that oxidation of asphalts and asphaltenes resulted in the formation of carbonyl groups (mainly ketones) generated from benzylic carbons,19,20 and in the formation of sulfoxide upon oxidation of aliphatic or naphthenic sulfides.20 The latter was formed even at mild conditions, and model studies indicated that these species were 2 orders of magnitude more reactive compared to other abundant sulfur species in petroleum and asphalts. Analyses of the unpurified OMV deposit upon its arrival showed that its content of SdO moieties was rather low, leaving the potential for its further increase by conversion of non-oxidized thio compounds in the deposit to the SdO form during storage. As can be seen in Figure 1a, absorbance of asphaltene samples above 3100 cm-1 is rather low and the ratio in that region is likely not a reliable indicator of the relative composition of species (OH and NH groups) that absorb in that region. In addition, this ratio can be biased by the contribution from light scattering by asphaltene aggregates to absorbance, which is different for the two samples, and can be traced to wavenumbers as low as 3600 cm-1 (see Figures S1 and S2 in the Supporting Information).

The OMV asphaltene is a reservoir deposit that originates from separator vessels in a gas condensate reservoir treatment plant and is known to contain ∼81% of n-heptane insolubles.15 Earlier data on structure and elemental composition of the deposit and its heptanewashed asphaltene (see Table S1 in the Supporting Information) indicate that properties of these samples are similar to those of crude oil asphaltenes reported in the literature.1 An original unpurified deposit was stored under air in a closed container. The deposit was purified in two batches: the second batch, OMV2, was prepared approximately 1 year later than the first batch, OMV1. Purified samples were obtained by 72 h Soxhlet extraction with heptane with subsequent vacuum drying at 60 °C for 12 h. Each batch was used for experiments within 2 months following purification. Heptane (HPLC grade) and toluene (glass-distilled grade) were obtained from Rathburn; both solvents were used as received, except for drying the toluene when required for experiments. Watersaturated toluene was obtained by mixing toluene with excess water and allowing the mixture to stand for at least 3 days. Unsaturated solutions of water in toluene were obtained by mixing dried and water-saturated toluene in varying proportions. Molecular sieves (diameter of 4 Å) were used to obtain dry toluene with the water content of approximately 0.001% wt. Deionized water was used for preparation of the samples. Asphaltene solutions for water solubilization measurements were prepared as follows: a series of asphaltene solutions in toluene were obtained by diluting a concentrated stock solution, which was prepared by ultrasonication. To saturate an asphaltene solution in toluene with water for water solubility measurements, the asphaltene solution was slowly poured onto a water phase, avoiding mixing of the organic and aqueous phases. The two-phase systems were kept in sealed vials. After equilibration in the dark for at least 72 h, samples of the organic phase were taken and analyzed for water content (according to Andersen et al.,13 72 h is sufficient time to achieve saturation in systems in question). The water content was determined by Karl Fischer titration on a 756 KF coulometer using Hydranal coulomat AG (RH 34836) with 30% toluene as a reagent; the relative error of measured water contents was within 5% for each sample. The KF system was calibrated with solutions of known water content. To obtain a series of asphaltene solutions having the same water content in unsaturated toluene, a concentrated stock solution of an asphaltene was prepared by ultrasonicating the asphaltene in unsaturated toluene with a given water content. Solutions with lower asphaltene concentrations were obtained by diluting the stock solution with toluene of the same water content as used for preparation of the stock solution. After preparation, all solutions were kept in sealed vials. (15) Andersen, S. I.; Potsch, K. Paper presented at the 1st International Conference on Petroleum Phase Behavior and Fouling, AIChE Spring Meeting, Houston, TX, 1999.

Results and Discussion

(16) Yen, T. F.; Erdman, G. Prepr. Pap.sAm. Chem. Soc., DiV. Pet. Chem. 1962, 7, 5. (17) Moschopedis, S. E.; Speight, J. G. Fuel 1976, 55, 187. (18) Andersen, S. I. Ph.D. Thesis, Technical University of Denmark, Lyngby, Denmark, 1990. (19) Petersen, J. C. Transportation Research Report. National Research Council, Washington, D.C., 1984. (20) Petersen, J. C.; Dorrence, S. M.; Nazir, M.; Plancher, H.; Barbour, F. S. Prepr. Pap.sAm. Chem. Soc., DiV. Pet. Chem. 1981, 26, 898.

3098 Energy & Fuels, Vol. 22, No. 5, 2008

KhVostichenko and Andersen

Figure 1. IR spectra of OMV1 and OMV2 asphaltenes (a) and their ratio (b). Asphaltene solutions with a 10 g/L concentration in CS2. Spectrum of OMV2 in a is offset by 0.5 absorbance units. Because of noisiness, data in b are not shown in wavelength regions corresponding to absorbance by the solvent.

Figure 2. Saturation concentration of water in solutions of OMV1 and OMV2 asphaltenes in toluene: OMV1, [; OMV2, 0. Table 1. Measured Water Content in Mixtures of Dried and Water-Saturated Toluene Used as Solvents for the OMV1 Asphaltene dried/water-saturated toluene ratio (vol/vol)

water content (% wt)

3:1 1:1 1:3 0:1

0.013 0.019 0.027 0.047

The saturation concentration of water cwsat in solutions of the OMV1 and OMV2 in toluene increased with an increasing asphaltenes concentration ca (Figure 2), in agreement with earlier findings for a different asphaltenes sample.13 A similar trend is known to hold for solutions of surfactants in apolar media, where the dependence of cwsat on the surfactant concentration is attributed to the accumulation of water in the interior of reverse micelles (so-called solubilization).21 This suggests that changes of water solubility with the asphaltene concentration resulted mainly from the interaction between asphaltenes and water, but the picture appeared more complex than in reverse micellar solutions, as discussed later. Although both OMV1 and OMV2 (21) Rabie, H. R.; Vera, J. H. Fluid Phase Equilib. 1996, 122, 169.

asphaltenes originated from the same reservoir deposit, their properties with respect to binding of water were different, as follows from different solubilities of water at the same asphaltene concentration for those two samples (Figure 2). The measured solubility of water in toluene (Table 1) agreed well with available literature data.22 To obtain insight into the state of water in the solutions, IR spectra of solutions of asphaltenes in toluene were obtained for water-saturated asphaltene solutions and solutions of asphaltenes in toluene with a constant water content below the saturation level. Three series of water-unsaturated solutions were prepared using mixtures of dry and water-saturated toluene with 1:3, 1:1, and 3:1 volume ratios, respectively, as solvents for asphaltenes (see the Experimental Section for details). Figures 3a and 4a show examples of spectra of asphaltene solutions in toluene with different water contents in the region of 3800-3100 cm-1. Characteristic of all spectra were two narrow peaks centered at approximately 3680-3678 cm-1 (peak 1) and 3592 cm-1 (peak 2) and a broad band in the region of 3800-3100 cm-1. Positions of the two narrow peaks were close to those observed in solutions of water in benzene (3684 and 3595 cm-1),23 suggesting that the peaks corresponded to vibrations of free water molecules dispersed in the solvent and not incorporated in hydrogen bonding with any polar moieties. According to literature data,23,24 these bands arise from symmetric and asymmetric O-H stretching modes of water molecules. On the basis of available spectral data for water/ surfactant/apolar solvent mixtures,24–27 the broad band was assigned to water solubilized within the asphaltene aggregates (22) Sorensen, J. M.; Arlt, W. Liquid-Liquid Equilibrium Data Collection, Chemistry Data Series; Behrens, D., Eckermann, R., Eds.; Dechema: Frankfurt, Germany, 1979; Vol. 5, Part 1. (23) Tassaing, T. Vib. Spectrosc. 2000, 24, 15. (24) D’Angelo, M.; Onori, G.; Santucci, A. J. Phys. Chem. 1994, 98, 3189. (25) Onori, G.; Santucci, A. J. Phys. Chem. 1993, 97, 5430. (26) Jain, T. K.; Varshney, M.; Maitra, A. J. Phys. Chem. 1989, 93, 7409. (27) Bey Temsamani, M.; Maeck, M.; El Hassani, I.; Hurwitz, H. D. J. Phys. Chem. B 1998, 102, 3335.



Figure 3. (a) Spectra of asphaltene solutions in dried and waterunsaturated toluene. Solid line 1, asphaltene/toluene/water (water content of 0.027 wt %); solid line 2, asphaltene/dried toluene; dashed line, subtraction (lines 1 - 2). OMV1, concentration of 5 g/L. (b) Subtracted spectrum fitted with three Gaussian peaks: dashed line, original spectrum (lines 1 - 2) of a; solid line, fitted spectrum; gray lines, individual fitted Gaussian peaks.

and involved in hydrogen-bonded networks of various sizes and degrees of distortion (“bulk-like” water). In water/surfactant/ apolar solvent mixtures, however, the two narrow bands of monomeric water molecules can only be observed if the concentrations of water and surfactant are low.24 Typically, only the broad band of the bulk-like water is apparent in these systems because of the high concentration of water solubilized in the micelles relative to that of water dispersed in the solvent.24,25 Lowering of the absorbance of the peaks in asphaltene solutions in dried toluene (Figures 3a and 4a) relative to absorbance by asphaltene solutions in toluene with water confirmed that the observed bands in asphaltene/toluene/water systems were indeed a result of the water present in the system. A comparison of spectra of dried toluene with and without dissolved asphaltenes (Figure 5) clearly shows that asphaltenes contributed to absorbance in the spectral interval of interest; weak absorbance by dried toluene could result from approximately 0.001 wt % of residual water remaining in toluene after drying. A variety of bands exhibited by asphaltenes in that region in addition to those of water is attributed to vibrations of phenol O-H and pyrrole- and indole-type N-H groups,16,17 with a broad band at 3450-3200 cm-1 arising because of hydrogen bonding between asphaltene molecules. Semi-quantitatively partition of water between the free (dispersed in toluene) and asphaltene-bound states in the system could be assessed from a dependence of monomeric water peak areas (peak 1 or peak 2) on the asphaltene concentration in a series of solutions with the same total water content. To remove the contribution of asphaltenes to the OH vibration region, spectra for analysis were obtained by subtraction of the spectra

Figure 4. (a) Spectra of asphaltene solutions in dried toluene and after saturation with water-toluene with and without water. Solid line 1, water-saturated solution of asphaltene in toluene (water content of 0.067 wt %); solid line 2, asphaltene/dried toluene; dashed line, subtraction (lines 1 - 2). OMV1, concentration of 5 g/L. (b and c) Subtracted spectrum fitted with four Gaussian peaks: dashed line, original spectrum (lines 1 - 2) of a; solid line, fitted spectrum; gray lines, individual fitted Gaussian peaks.

Figure 5. Spectra of dried toluene and OMV1 (5 g/L) in dried toluene.

of asphaltenes dissolved in dry toluene from the spectra of asphaltene/toluene/water systems with the same asphaltene concentration (Figure 3b). Peak areas for water-unsaturated solutions were obtained by fitting these subtracted spectra with tree Gaussian peaks (Figure 3b). An example of a fitted spectrum for a solution of water in toluene in the absence of asphaltenes is shown in Figure 6; the low wide band still observed below 3500 cm-1 in this solution might indicate the presence of some impurities in toluene, because it also appeared in the spectrum



titrator did not allow us to measure the concentrations of water in toluene used for preparation of the OMV2 asphaltene solutions, and because deviations of the actual water content from that in OMV1 solutions were unknown, quantitative analysis was not carried out for this sample. On the basis of the linear fit obtained for solutions of water in toluene (Figure 8), one could calculate the concentration of free water cwfree in asphaltene solutions in toluene from areas asph P1 . Because the total water content cw was constant throughout each series, it is possible to estimate the number of water molecules bound to one asphaltene molecule, nb nb ) Figure 6. Spectrum of water dissolved in toluene (0.013 wt %) fitted with three Gaussian peaks: dashed line, original spectrum; solid black line, fitted spectrum; gray lines, individual fitted Gaussian peaks.

of dried toluene (Figure 5). An attempt to analyze spectra of water-saturated solutions was unsuccessful: the spectra required at least four Gaussian peaks for fitting, and fits of comparable accuracy could be obtained with multiple combinations of peak parameters (parts b and c of Figure 4). Figure 7 shows the ratio of the peak 1 area in asphaltene/ asph water/toluene solutions, P1 , to that in the solution of water in tol toluene with the same water concentration, P1 , as a function of the asphaltene concentration for each series of waterasph tol unsaturated solutions. P1 /P1 corresponds to the ratio of the concentration of water not bound to asphaltenes, cwfree, to the total concentration of water in the system, cw (To simplify the estimations, all water in solutions in toluene in the absence of asphaltenes was assumed to be in the monomeric state. This may be an oversimplification, especially in the saturated case, but the focus on the sharp peaks merits the approach). The original intention was to obtain more detailed information about the region of low asphaltene concentrations, and solutions of the OMV2 sample only spanned the concentration range of 0.01-2.0 g/L, while OMV1 concentrations varied from 0.1 to 20 g/L. However, the properties of the two samples turned out to be substantially different, and the features of the partition of water between the bound and free states in the OMV1 solutions were not retained in solutions of the OMV2 asphaltene. Although the solubilization capacity of the OMV1 asphaltene was higher than that of the OMV2 asphaltene, data for waterunsaturated solutions showed the opposite trend: at low asphaltene concentrations, there was less free water in OMV2 solutions than in OMV1 solutions. In the case of the OMV1 sample, all three independently obtained curves exhibited a drop in the free water content around the asphaltene concentration of 0.1 g/L, a plateau in the concentration range of 0.5-2.0 g/L, and a gradual decrease at higher asphaltene concentrations; such nonmonotonous variation in the free water content suggested that binding of water by asphaltenes depends upon the state of asphaltenes in solution. Quantitative analysis of spectroscopic data was carried out for solutions of the OMV1 asphaltene. Concentrations of water in mixtures of dry and water-saturated toluene used for the preparation of asphaltene solutions were measured (Table 1), and it was found that the dependence of the peak area on the water concentration in the range of 0.013-0.047 wt % could be fitted with a straight line in accordance with Lambert-Beer’s law (Figure 8). Some degree of numerical dispersion was observed, causing the linear fitted regression lines not to pass through the origin of the plot. Unfortunately, failure of the KF

(cw - cwfree) MWa · ca MWw

(1)

where cfree w , cw, and the asphaltene concentration ca are expressed in g/kg or g/L and MWw and MWa are molecular weights of water and asphaltene, respectively. In addition, the number of water molecules bound per asphaltene molecule in watersaturated asphaltene solutions in toluene, nsat b , was calculated for water-saturated solutions of asphaltenes in toluene. To avoid uncertainty related to calculation of the free water content from fitted values of peak areas for these solutions, molar water solubilization capacity nsat b was estimated from results of water solubilization measurements as follows: nsat b )

(cwsat - cwtol) MWa · ca MWw

(2)

where csat w is the concentration of water in the solution at a given asphaltene concentration obtained in water solubilization measurements. The concentration of water in the bulk of the solvent, cwtol, was assumed constant in all solutions and equal to the saturation concentration of water in pure toluene. The values of nsat b were calculated for both OMV1 and OMV2 samples. The molecular weight of the OMV asphaltene was not known; however, available literature data indicate that molecular weights of most asphaltenes likely lie in the lower part of the 500-2000 g/mol interval,11,28–30 and the value of 1000 g/mol was selected for calculating nb and nsat b . As shown in Figure 9, the number of water molecules bound to an asphaltene molecule was relatively large at low asphaltene concentrations and decreased at higher asphaltene concentrations, unlike in solutions of conventional surfactants, where solubilization capacity increases with an increasing surfactant concentration. At a given OMV1 asphaltene concentration, the values of nsat b were higher than those of nb, with the exception of solutions with an asphaltene concentration of 0.1 g/L. Discrepancy in the latter case could result from the calculation of nb and nsat b from the difference of two close values (cw cwfree) and (cwasph - cwtol), respectively. Calorimetric data on the interaction of asphaltenes with p-nonylphenol31 indicate that the number of hydrogen-bonding sites per asphaltene molecule is in the range of 2-6 (also selecting 1000 g/mol as the average asphaltene molecular weight). Low values of nb calculated for the OMV1 asphaltene at higher asphaltene concentrations (10-20 g/L) were not unexpected because the amount of water available for binding in water-unsaturated solutions was ultimately limited by the total water concentration, and the total water/asphaltene molar ratios in solutions with 0.013, 0.019, (28) Groenzin, H.; Mullins, O. C. Energy Fuels 2000, 14, 677. (29) Groenzin, H.; Mullins, O. C. J. Phys. Chem. A 1999, 103, 11237. (30) Yarranton, H. W.; Alboudwarej, H.; Jakher, R. Ind. Eng. Chem. Res. 2000, 39, 2916. (31) Merino-Garcia, D.; Andersen, S. I. Langmuir 2004, 20, 1473.



Figure 7. Free water content in solutions of OMV1 (b, 2, and 9) and OMV2 (O, 4, and 0) in series with a constant total water content.

Figure 8. Linear fit of the peak area versus the water concentration for solutions of water in toluene: peak 1, b; peak 2, 2.

and 0.027 wt % water in toluene at the asphaltene concentration of 20 g/L were as low as 0.32, 0.46, and 0.65, respectively. However, the values of nsat b reflected the largest possible number of water molecules that could be bound by asphaltenes, yet they steadily decreased with an increasing asphaltene concentration and, at ca of 20 g/L, fell just below unity for the OMV1 asphaltene and were even lower for the OMV2 sample (Figure 9). Low values of nsat b did not support the existence of clusters of water in the interior of asphaltene aggregates and suggested that the broad hydrogen-bonding envelope in the IR spectrum (3100-3400 cm-1) was related mostly to water-asphaltene interactions rather than water-water interactions, unlike in solubilization by conventional surfactants.24–27,32 Recent studies suggest that the distribution of monomeric asphaltene molecular weights peaks at ∼500 g/mol,11,33 with the average molecular weight value of ∼900 g/mol33 and, hence, the selected value of 1000 g/mol might be slightly too high. Ambiguity in the choice of the MWa value inevitably results in (32) Li, Q.; Weng, S.; Wu, J.; Zhou, N. J. Phys. Chem. B 1998, 102, 3168. (33) Martı´nez-Haya, B.; Hortal, A. R.; Hurtado, P.; Lobato, M. D.; Pedrosa, J. M. J. Mass Spectrom. 2007, 42, 701.

some uncertainty of calculated values of the number of water molecules bound per asphaltene molecule. However, because these numbers are directly proportional to the asphaltene molecular weight (eqs 1 and 2), recalculation for any value of MWa other than 1000 g/mol work is straightforward. For example, using MWa ) 500 g/mol shifts the range of nsat b to 1.9-0.4 and 1.9-0.05 for the OMV1 and OMV2 samples, respectively. At the same time, the choice of the asphaltene molecular weight will not affect overall trends in Figure 6 since all values of nb and nsat b will be scaled uniformly. Because the aim is to give trends rather than exact numbers, we find that this merits the use of an average unit molecular weight. Observed trends in the variation of nb and nsat b with the asphaltene concentration are markedly different from those observed for conventional surfactants. First, the solubilization capacity of surfactants, such as Aerosol OT and NaDEHP (10-40 mol of water/mol of surfactant)21,26,32 greatly exceeds that of asphaltenes. Second, depending upon the composition of the aqueous phase used in solubilization experiments, nsat b for Aerosol OT is either independent of or grows with an increasing surfactant concentration,21 whereas for asphaltenes, a decrease in water solubilization capacity with an increasing asphaltene concentration is observed. A high solubilization capacity of conventional surfactants is the consequence of their structural properties: the well-defined inner hydrophilic region of reverse micelles that contains the polar head groups21,26,32 of surfactant molecules can accumulate clusters of up to several hundred of water molecules.34 The structure of asphaltene aggregates and the mechanism of asphaltene-asphaltene interactions are not known in detail, and elucidation of those properties is impeded by polydispersity of the asphaltene fraction. X-ray diffraction studies35,36 suggest that asphaltene aggregates are formed through π stacking of polyaromatic fragments of asphaltene molecules, with some disorder introduced by saturated side chains. On the contrary, molecular (34) Eicke, H.-F.; Rehak, J. HelV. Chim. Acta 1976, 59, 2883.



Figure 9. Values of nb in solutions of OMV1 in toluene with a constant total water content: 0.013 wt %, 0; 0.019 wt %, O; 0.027 wt %, 2. The inset shows values of nsat b in water-saturated solutions of OMV1 (1) and OMV2 (3).

simulations predict a large variation in preferred relative orientation of neighboring asphaltene molecules, from parallel face-to-face to perpendicular edge-on, depending upon the structures of model asphaltenes and environmental conditions (temperature and presence or absence of solvent molecules) selected for simulations.37,38 In either configuration, the presence of well-defined hydrophilic cavities that can accommodate appreciable amounts of water within asphaltene aggregates is dubious, which could explain relatively low values of nsat b . It is, however, possible that in some cases water molecules could actually facilitate otherwise hindered interactions by serving as bridges between polar groups of asphaltene molecules, as was found in molecular dynamics simulations.8,9 In addition to π stacking, other types of interactions, such as the formation of charge-transfer complexes, acid-base interactions, and hydrogen bonding, can take place as a result of various heteroatom-containing functional groups in asphaltene molecules. Relative contributions of these interactions to the total intermolecular interaction energy vary for different asphaltene samples.39,40 Interactions that explicitly involve localized polar functional groups (such as hydrogen bonding) are likely the most important ones when the binding water to asphaltenes is considered.41 Each asphaltene molecule has only a few polar groups,28,31 and a decreasing number of bound water molecules per one asphaltene molecule with an increasing asphaltene concentration could result from competition between asphaltene-asphaltene and asphaltene-water binding as well as from the changes in the aggregate structure. The formation of asphaltene aggregates, presumably dimers, is evident in fluorescence measurements at concentrations as low as 0.05-0.06 g/L;9,29 hence, most of the solutions studied in our work contained aggregated asphaltenes (possibly with the exception of the OMV2 solutions at concentrations of 0.01-0.05 g/L). However, the decrease in the heat of dissociation with increasing asphaltene concentrations observed in calorimetric measurements8,9 suggests that the strength of asphaltene-asphaltene interactions (35) Yen, T. F.; Erdman, J. G.; Pollack, S. S. Anal. Chem. 1961, 33, 1587. (36) Tanaka, R.; Sato, E.; Hunt, J. E.; Winans, R. E.; Sato, S.; Takanohashi, T. Energy Fuels 2004, 18, 1118. (37) Alvarez-Ramirez, F.; Ramirez-Jaramillo, E.; Ruiz-Morales, Y. Energy Fuels 2006, 20, 195. (38) Zhang, L.; Greenfield, M. L. Energy Fuels 2007, 21, 1102. (39) Gawrys, K. L.; Blankenship, G. A.; Kilpatrick, P. K. Energy Fuels 2006, 20, 705. (40) Strausz, O.; Torres, M.; Lown, E. M.; Safarik, I.; Murgich, J. Energy Fuels 2006, 20, 1023. (41) Xu, Y.; Dabros, T.; Hamza, H.; Shelfantook, W. E. J. Pet. Sci. Technol. 1999, 17, 1051.

is weaker at low concentrations, possibly indicating that these aggregates have a more open structure that is easily penetrated by water molecules. Under this assumption, the following picture emerges: polar groups of asphaltene molecules in small weakly bound aggregates at lower asphaltene concentrations were easily accessible to water molecules. Stronger interactions between asphaltenes in and a growing size of aggregates with increasing asphaltene concentrations effectively reduced the accessability of polar asphatene groups for binding with water. First, asphaltene–asphaltene bonds were harder to break for asphaltene– water bonds to form. Second, the number of polar groups in the interior of the aggregates inaccessible for asphaltene-water binding increased with an increasing aggregate size. The nonmonotonous dependence of the free water content on the asphaltene concentration in OMV1 solutions with a constant total water content could result from the interplay between the changes in the aggregated state of asphaltenes and the decrease of the total water/asphaltene ratio with an increasing asphaltene concentration. Dependence of asphaltene-water binding on the state of asphaltenes in solution might also account for the counterintuitive behavior of the OMV2 asphaltene: (1) solubilization capacity of the OMV2 sample, despite its higher relative content of oxidized moieties, was lower than that of the OMV1; (2) although the OMV2 could solubilize less water than the OMV1, it bound more water in water-unsaturated solutions than the OMV1 asphaltene. In water saturation experiments, a solution of asphaltenes (presumably already aggregated) was brought in contact with the aqueous phase and no mechanical mixing, such as ultrasonication or stirring, was applied to avoid emulsion formation. Hence, only the sites on the exterior of asphaltene aggregates were initially available for water-asphaltene interaction, and the formation of additional asphaltene-water bonds would have required breaking of asphaltene-asphaltene bonds. Because interactions between polar sites play an important role in asphaltene-asphaltene binding, it is possible that the aggregates formed by the OMV2 asphaltenes were bound more tightly and contained more asphaltene molecules per aggregate than those of the OMV1. Both features appear unfavorable for asphaltene-water binding, as described above: surface/mass ratio decreases with an increasing aggregate size, resulting in fewer exterior sites readily available for binding, while stronger asphaltene-asphaltene bonding would impede breaking of the aggregates, effectively making the binding sites in the interior of asphaltene aggregates less accessible to water. Upon dissolution of asphaltenes in water-unsaturated toluene solutions, the stock solutions of asphaltenes, toluene, and water were



subjected to ultrasonication. The latter method is known to have a drastic effect on the structure of aggregated systems and is commonly used to release drug molecules encapsulated in micelles.42 Thus, ultrasonication might have allowed water to occupy asphaltene polar sites that would otherwise have remained locked in asphaltene aggregates. Although our argument is speculative, trapping of asphaltenic systems as well as reverse micellar solutions in a certain state that depends upon the preparation procedure is not unprecedented. For example, it was observed in precipitation/dissolution studies of asphaltenes, where it was rationalized in terms of a high kinetic barrier for transition from one state to another.43 Water solubilization capacity of Aerosol OT/cosurfactant systems was also shown to depend upon the method used for preparation of watersaturated solutions.44 Relatively poor water-binding properties of asphaltenes are not in disagreement with a well-known fact that asphaltenes show interfacial activity and stabilize water-in-oil emulsions:3,45–47 our asphaltene solutions containing solubilized water were rather different from water-in-oil emulsions. Emulsion formation typically requires mechanical mixing,3,45,46,48 while in the experiments described in this paper this was avoided. Spiecker et al.48 reported that solutions of asphaltenes in heptane/toluene mixtures (3.7 g of asphaltene/L solvent) retained from 1.5 to 5.4 g emulsified water per 4 mL of solution. Thus, emulsified water content in these systems was at least 300 g/kg, exceeding the amount of water solubilized by asphaltene solutions in toluene at similar asphaltene concentrations of 2 and 5 g/L in our experiments (Figure 2) over 500-fold. Stabilization of significant amounts of water in emulsions was shown to result from rigid skins assembled from asphaltene monomers or whole aggregates at the water droplet/oil interface.3,45–47 Such a mechanism in fact does not rely on the penetration of water into preformed asphaltene aggregates and disrupting the asphaltene-asphaltene interaction in the aggregates to facilitate

asphaltene-water binding. The curvature radius of the micrometer size water droplets in emulsions is orders of magnitude bigger than the size of individual water molecules or asphaltene monomers and aggregates, 2-5 nm,49 and water in the droplets must be in essentially the bulk state, with only a surface layer, i.e., only a small fraction of water, being in direct contact with the asphaltenes.

(42) William, G. P.; Ghaleb, A. H.; Bryant, J. S. Expert Opin. Drug. DeliVery 2004, 1, 37. (43) Beck, J.; Svrcek, W. Y.; Yarranton, H. W. Energy Fuels 2005, 19, 944. (44) Rabie, H. R.; Helou, D.; Weber, M. E.; Vera, J. H. J. Colloid Interface Sci. 1997, 189, 208. (45) Czarnecki, J.; Moran, K. Energy Fuels 2005, 19, 2074. (46) Wu, X. Energy Fuels 2003, 17, 179. (47) McLean, J. D.; Kilpatrick, P. K. J. Colloid Interface Sci. 1997, 189, 242. (48) Spiecker, P. M.; Gawrys, K. L.; Trail, C. B.; Kilpatrick, P. K. Colloids Surf., A 2003, 220, 9.

Supporting Information Available: Elemental composition and and 13C NMR structural data for the OMV unpurified deposit and one of its earlier purified asphaltenes (Table S1), spectra of OMV1 and OMV2 asphaltenes in the range of 7000-600 cm-1, with a 10 g/L concentration in CS2 (Figures S1 and S2). This material is available free of charge via the Internet at http://pubs.acs.org.

Conclusions Our findings demonstrated that solubilization of water by asphaltenes and water-asphaltene binding strongly depend upon the state of aggregation of asphaltenes. Decreasing water-binding capacity of asphaltenes with an increasing asphaltene concentration suggests that asphaltene-asphaltene binding competes with asphaltene-water binding. Solubilization of non-emulsified water by asphaltenes appears to occur because of the binding of water by polar sites on the exterior of asphaltene aggregates rather than through the formation of clusters of water inside the asphaltene aggregates. Reversal of water-binding properties of the two investigated asphaltene samples in water-saturated and water-unsaturated asphaltene solutions shows that structural characteristics of asphaltene molecules alone, such as the content of polar groups, are not the only factor that determines the water-binding capacity of a given asphaltene sample. Taking into account different preparation procedures for water-saturated and water-unsaturated samples, these findings may indicate that one of the factors that affects asphaltene-water binding is the state of asphaltenes before they are brought in contact with water. However, further investigations on the size and properties of asphaltene aggregates in these systems will be required to fully elucidate the mechanisms behind the observed trends. Acknowledgment. The authors thank Nordic Energy Research for financial support. S.I.A. also thanks the STVF for a grant under the Talent Project. 1H

EF700757H (49) Rahmani, N. H. G.; Dabros, T.; Masliyah, J. H. J. Colloid Interface Sci. 2005, 285, 599.

Interactions between Asphaltenes and Water in Solutions in Toluene

Recommend Documents