- Email: [email protected]
Effect of precipitation time and solvent power on asphaltene characteristics
Fuel 208 (2017) 271–280
Contents lists available at ScienceDirect
Fuel journal homepage: www.elsevier.com/locate/fuel
Full length article
Effect of precipitation time and solvent power on asphaltene characteristics Estrella Rogel ⇑, Michael Moir Chevron Energy Technology Company, Richmond, CA 94597, United States
h i g h l i g h t s Composition of precipitated asphaltenes changes widely as a function of time. Asphaltene precipitation involves reorganization of the aggregates. Reorganization of aggregates includes expulsion and incorporation of molecules. 24 h is not enough time to reach a steady composition of the asphaltenes.
a r t i c l e
i n f o
Article history: Received 20 March 2017 Received in revised form 5 June 2017 Accepted 24 June 2017
Keywords: Precipitation kinetics Asphaltenes Maltenes Precipitated material Solubility Aggregation model
a b s t r a c t The kinetics of asphaltene precipitation is studied regarding the characterization of the precipitated material. The association of different components of the crude oil depends on the solvent power of the blend heptane/oil and produces aggregates with a variable composition that change as a function of time. It was found that the precipitate contains a significant amount of maltenes whose relative content decreases as a function of time. In agreement with this behavior, the precipitate becomes more hydrogen deficient. These changes also produced a significant increase in the apparent average molecular weight of the precipitate as measured using size exclusion chromatography. This increase correlated with the amount of asphaltenes in the precipitated material independently of time or heptane/oil ratio. Solubility profile measurements showed that the average characteristics of the asphaltenes found in the precipitate varied as a function of time. As more asphaltene molecules precipitated, the distributions of these molecules changed decreasing its average solubility parameter. Considerable differences in the amount collected at low and high heptane/oil ratios as a function of time were also found. Based on the measurements, a new model was developed that describes the complex aggregation behavior involved in asphaltene precipitation. Ó 2017 Elsevier Ltd. All rights reserved.
1. Introduction Asphaltenes are undesirable compounds that precipitate during different stages of petroleum production and refining. As they are not defined as a compound class but as a solubility class, characterization of asphaltenes is challenging since variations in the procedure to obtain them produce not only different amounts but also distributions of compounds with different characteristics [1–3]. It is well known that asphaltene precipitation is affected by several factors, including type of solvent [1,3,4], temperature [5,6], precipitant/sample ratio [3,7], contact time [7], and washing of the filter cake [2,7]. In practical terms, this means that comparison of asphaltenes obtained using even slightly different methodolo⇑ Corresponding author. E-mail address: [email protected] (E. Rogel). http://dx.doi.org/10.1016/j.fuel.2017.06.116 0016-2361/Ó 2017 Elsevier Ltd. All rights reserved.
gies must be made carefully: the asphaltene yields and the chemical composition are going to be different. The main reason for this behavior is the high polydispersity regarding composition and molecular weight of petroleum components. This combination of factors results that under slightly different conditions, a different set of molecules precipitate forming a distinct asphaltene. Although it is expected that the molecules more prone to precipitate are going to be those with the higher aromaticity (poor in hydrogen), higher molecular weight and/or higher polar nature, significant variations in the amount and distribution of the molecules can occur as shown by previous studies [3,5]. Particularly important is the understanding of kinetic aspects related to asphaltene precipitation as this is a variable frequently neglected in asphaltene studies [8] and can have important practical consequences during production and processing of crude oils. Recently, a series of experiments using optical microscopy
272
E. Rogel, M. Moir / Fuel 208 (2017) 271–280
Blend
+
Precipitaon
Filtraon
Nominal pore size: 0.22 µ
Filtered Cake
Crude Oil or Model Oil
n-Heptane
Characterizaon
Shake vigorously for 10 min Fig. 1. Scheme of the experiments.
2. Experimental section 2.1. Materials Crude oil and n-pentane asphaltenes were used in the precipitation experiments. Crude Oil A is a medium crude oil (30.4 API) with an asphaltene content 2.43 wt% using the standard method ASTM D-6560 [16]. n-Pentane asphaltenes were obtained following the standard ASTM D-4055 [17] from a heavy Crude Oil B (7.7oAPI) with an n-pentane asphaltene content of 18.7 wt%. Toluene, nheptane, n-pentane, methylene chloride and methanol are HPLC grade (Fisher Chemical)
2.2. Separation of precipitated material Studied samples included Model Oil and Crude Oil A. The Model Oil was prepared by dissolving the n-pentane asphaltenes in toluene (2 wt%) n-Heptane was used to induce precipitation at room temperature of the Model Oil, and Crude Oil A. Different n-heptane/ samples ratios from 1.5 to 100 were used. Flocculation onset experiments using n-heptane as precipitant agent indicated that precipitation starts at a n-heptane/Crude Oil ratio of 1.5. Blends prepared with a n-heptane/Crude Oil ratio of 1 did not show precipitation during the first 24 h. Solutions are prepared by mixing 10 mL of sample with the corresponding volume of n-heptane and shaking vigorously for 10 min in a shaker. After blending, the solutions are kept in static conditions. Blends were filtered using a Teflon membrane filter with an average pore size of 0.2 mm so that particles could be recovered and analyzed. Fig. 1 shows a scheme of the procedure to obtain the filtered cakes. Since the goal is to replicate the precipitation behavior as close as possible to a real situation, the filtered cake was not washed with additional n-heptane after filtration. Additionally, it has been reported that precipitation procedures where asphaltenes are not washed seem to be more repeatable [2]. Once collected, the filtered cake was dried under nitrogen flow, weighed and analyzed using the characterization methods listed below. Samples were run in duplicates which on average differ by 10%.
30
25 Collected Amount (wt%)
observations and centrifugation based separations [8,9], has shown that kinetic factors are underestimated during conventional flocculation experiments aimed to determine chances of asphaltene precipitation. In fact, it was found that these test over predict solubility as it was observed that the beginning of precipitation could take from minutes to several months depending on the precipitant concentration used [8,9]. From these experiments, we also know that the precipitated amount varies as a function of time and that those variations depend on the precipitant concentration. When asphaltenes precipitate in the reservoir or during refinery operations, factors such as solvent power and time have also a significant influence. In particular, aging can have a definitive impact on deposit characteristics [10,11] or emulsion formation [12]. Additionally, kinetic effects and solvent power play a relevant role in understanding how recently developed solvent injection processes (i.e. Vapex or N-solv) [13,14] might affect the reservoir and crude oil production. In these processes, hydrocarbons in the range C3 to C5 are injected downhole to make the crude oil mobile. Recent physical and numerical simulations [15] of propane injection into a heavy oil reservoir indicated that the kinetics of asphaltene precipitation plays a significant role in the production as well as in the quality of the produced crude oil. From many perspectives, the understanding of how time and solvent power affect asphaltene composition can provide useful information to design new strategies in diverse areas including analytical characterization, precipitation prevention, deposit removal, and new production methods. In the present work, the effect that time and solvent power have on asphaltene properties is investigated using a combination of analytical techniques. We evaluate several physicochemical properties that are strongly linked to precipitation behavior such as hydrogen deficiency, solubility distribution, and molecular size. The changes in the amount and composition of the precipitated material as a function of time are also examined.
20
15
10 1.5 5
2
0 0
50
100
150
200
Time (h)
Fig. 2. Collected amount as a function of time and n-heptane/Model Oil ratio. Trace lines added for visualization purposes.
273
E. Rogel, M. Moir / Fuel 208 (2017) 271–280
changed gradually from pure n-heptane to 90/10 methylene chloride/methanol and then to 100% methanol. This procedure gradually re-dissolves the asphaltenes from the easy to dissolve (low solubility parameter) to the hard to dissolve (high solubility parameter) producing a distribution that represents the range of asphaltenic species with different solubilities present in the sample. In a similar way to the determination of asphaltene content, asphaltenes are quantified using an Evaporative Light Scattering Detector (ELSD). Rather recently, this method has been calibrated to produce a distribution of asphaltenes based on their solubility parameters. A description of the calibration is published elsewhere [20]. Elemental analysis of the samples was carried out by a standard combustion method, using a Carlo Erba model 1108 analyzer. Aggregation studies were carried out by size exclusion chromatography (SEC) using a 30 cm 7.5 mm PLgel ‘‘mixed E” column. Sample solutions (0.01 g in 10 mL of methylene chloride)
2.3. Characterization methods Asphaltene concentrations and asphaltene solubility profiles were determined using on-column filtration techniques [18,19]. In these techniques, a solution of the sample (0.01 g in 10 mL of methylene chloride) is injected into a column packed with an inert material using n-heptane as the mobile phase. This solvent induces the precipitation of asphaltenes and, therefore, their retention in the column. To determine the asphaltene content, the mobile phase is switched to a blend methylene chloride/methanol 90/10 v/v that redissolves the asphaltenes completely. Asphaltenes are quantified using an Evaporative Light Scattering Detector (ELSD). This technique produces results that correlate with the results obtained using the conventional gravimetric technique (ASTM D6560) [16]. In the asphaltene solubility profile method, after the initial injection and precipitation of asphaltenes, the mobile phase is
21.2
a)
2
9.0E-03 1.5
Normalized Response
8.0E-03
5
Pentane Asphaltenes
7.0E-03 6.0E-03
Average Solubility Parameter (MPa0.5)
1.0E-02
21.0 20.8 20.6 20.4 20.2 20.0
Pentane Asphaltenes
19.8 0
2
4 6 8 n-heptane/Model Oil Rao
10
12
5.0E-03 4.0E-03
10
3.0E-03 2.0E-03 1.0E-03 0.0E+00 13
14
15
16
17
18
19
20
Time (min)
1.0E-02
Normalized Response
b) 9.0E-03
1h
8.0E-03
2h
7.0E-03
3h
6.0E-03
24 h
5.0E-03
168 h
4.0E-03 3.0E-03 2.0E-03 1.0E-03 0.0E+00 13
14
15
16
17
18
19
20
Time (min) Fig. 3. Comparison of asphaltene solubility profiles of the filtered cakes. a) Filtered cakes collected at 3 h for different n-heptane/Model Oil ratios (1.5, 2, 5, 10) and original npentane asphaltenes. b) Filtered cakes collected at different times for low (1.5) ratio.
274
E. Rogel, M. Moir / Fuel 208 (2017) 271–280
Fig. 4. Collected amount as a function of time for low (1.5) and high (10) n-heptane/Crude Oil ratios. Insert shows behavior at short times for low (1.5) ratio. Trace lines added for visualization purposes.
3.1. Model Oil Initial experiments were performed using a Model Oil prepared using n-pentane asphaltene in toluene (2 wt%). Fig. 2 shows the amounts of material collected at increasing times for two different n-heptane/Model Oil ratios (1.5 and 2). In this plot, the collected material is calculated as a percentage of the initial mass of npentane asphaltenes. As expected, the amount collected increases as a function of time as well as when the n-heptane/sample ratio increases. Fig. 3a shows the solubility profile for several of the filtered cakes separated using heptane/sample ratios of 1.5–10 and collected at similar times (3 h). The solubility profiles of these materials varied significantly showing that different ratios produced materials that exhibit different solubility characteristics. In particular, the results indicated that lower ratios at this short time (3 h) produced filtered cakes enriched in more soluble material (eluted at shorter times) than larger ratios. In this figure, the solubility profile of n-pentane asphaltenes is also shown for comparison. The more material is collected (Fig. 2), the solubility profile of the asphaltenes becomes closer to the one obtained for the npentane asphaltenes as can be expected. This tendency can be observed in the insert in Fig. 3a showing the average solubility parameter of the asphaltenes as calculated using the solubility profile technique. These values indicate that as more material is precipitated, the average solubility decreases. As the solvent power
a) Content based on crude oil (wt%)
3. Results and discussion
1.40 1.20 1.00 0.80 0.60 0.40 Asphaltenes 0.20
Maltenes
0.00 0
100
200
300
400
500
600
700
800
Time (h) 2.50
b) Content based on crude oil (wt%)
were prepared in methylene chloride. Elution was carried out with 90/10 methylene chloride/methanol at a flow rate of 1.0 mL/min, and temperature kept constant at 25 °C. The molecular weights (Mw) were calculated based on a calibration that uses porphyrins, dyes, and polyaromatics as standards, as published before [21]. Molecular weights were reported as Mw and Mn values [22] Mw have a relative error of 5%.
2.00
1.50
1.00
0.50
Asphaltenes Maltenes
0.00 0
100
200
300
400
500
600
700
800
Time (h) Fig. 5. Asphaltene and maltene content in the filtered cake for low (1.5) and high (10) n-heptane/Crude Oil ratios as function of time. Trace lines added for visualization purposes.
275
E. Rogel, M. Moir / Fuel 208 (2017) 271–280
decreases and more molecules (with high solubility in heptane) move to the precipitate, the solubility parameter of the asphaltenes decreases in average. In contrast with the effect of n-heptane/Model Oil ratio, comparison of asphaltene solubility profiles for the same n-heptane/ Model Oil ratios at different times up to 1 week did not show significant differences in the profile as shown in the example presented in Fig. 3b. 3.2. Crude Oil Fig. 4 shows the amount of solids collected in the filter as a using heptane/Crude Oil ratios of 1.5 and 10. In this figure, changes of the amount collected as a function of time up to 1 month (720 h) indicate different behaviors depending on the ratio. At the low ratio, an initial decrease of the amount is followed by an increase, while for the larger ratio the amount collected increases from the beginning of the experiment. These significant differences at short times are apparently related to the composition of the blend. It has been reported [2] that at low n-alkane concentrations, n-alkanes penetrate asphaltene particles causing swelling and dissolving a fraction of the asphaltenes. An alternative explanation for this contrasting behavior might be found in the diffusion of molecules from the aggregates to the fluid is limited by a higher viscosity. Asphaltene content in the filtered cake was evaluated using the on-column methodology. This test is a measurement that correlates with the asphaltene content measured using ASTM D-6560 [16], the gravimetric standard method that employs heptane as
precipitant. Fig. 5 shows a comparison of how asphaltene and maltene content as measured by the on-column technique changes as a function of time for low (1.5) and high (10) n-heptane/Crude Oil ratios. For the lower ratio, asphaltene content increases in the filtered cake as a function of time, while the maltene amount initially decreases and then starts to increase. In the case of high ratio, both asphaltene and maltene amounts increases, although asphaltene content seems to increase faster and dominate the filtered cake (60–70 wt%). Previous studies of the kinetics of aggregation of asphaltenes have not revealed these differences, mainly because of the washing of the filtered cakes with additional heptane to eliminate trapped crude oil [8,9,23]. In the present work, as mentioned before, the filtered cakes were not washed based on our goal of replicate real precipitation conditions and analyze how the aggregates change with time. We think that these changes are essential to understand the evolution of deposits [11]. Based on previous studies [24], the kinetics of aggregation depends on the viscosity of the blend as well as in the solubility of the heavy hydrocarbons in it. In principle, it can be assumed that once the solubility limit for different components is crossed, diffusion-limited aggregation will occur spontaneously due to the low solubility of these components in the blend. Apparently, at low ratios and short times, an exchange of components takes place where some of the components trapped in the filtered cake go back into solution while new components are incorporated in the aggregates. This exchange indicates that the activation energy of this process is relatively low and the exchange proceeds relatively rapidly. It is important to mention that the redissolution of trapped
a)
b) 168 h 168 h
98 h 24 h
24 h
Response
Response
8h
7h
6h
5h
5h
3h
2h
2h
1h
12
14
16
18
Time (min)
20
1h
22
12
14
16
18
20
22
Time (min)
Fig. 6. Asphaltene solubility profiles for the filtered cakes collected as a function of time at a) 1.5 n-heptane/Crude Oil ratio and b) 10 n-heptane/Crude Oil ratio. For comparison purposes, the profiles were normalized.
276
E. Rogel, M. Moir / Fuel 208 (2017) 271–280
material could also happen at high ratios, but it was not observed because it might happen so fast that the measurements made cannot capture it. Changes in the characteristics of the asphaltenes were also evaluated as a function of time. Fig. 6 shows these changes in the solubility distribution of the heptane asphaltenes in the filtered cake for low and high ratios. For the low ratio, changes are clear: there is an apparent decrease in the more soluble components (1st peak) while the second peak increases (less soluble components). For the large ratios, changes are less significant, but still consistent with a relative decrease in the first peak. In contrast to these results, filtered cakes obtained from the Model Oil did not show major differences in composition as a function of time (Fig. 3b). These different behaviors might suggest that when the precipitation occurs from the whole crude oil, lighter components are involved that are not present when the precipitation occurs in n-pentane asphaltene solutions (Model Oil). Also, the Model Oil blends have lower viscosities than the equivalent Crude Oil blends and, consequently, they reach equilibrium faster.
1200 50 1000
Response
800
100
600 10 400
5
200
2 1.5 0 12
14
16
18
20
22
Time (min) Fig. 7. Asphaltene solubility profile of the filtered cake as a function of n-heptane/ Crude Oil ratio at 24 h.
A significant body of research has pointed out to the presence of components in asphaltenes whose characteristics do not correspond to the expected characteristics of an asphaltene molecule even after significant soxhlet extraction procedures. For instance, a study of experimental data demonstrated the presence of a reversible adsorption phenomenon in the solvent precipitation of asphaltene from the parent crude oil involving the coprecipitation of precipitant-soluble resins and asphaltene -like substances [1]. In fact, the authors of this study estimated that around 70% of heptane asphaltenes in a bitumen were associated resinous material [1]. It is important to point out that the filtered cake was not washed with solvents before analysis. After extensive washing with heptane, a common practice, plots of the filtered cake become more and more like those shown in Fig. 3b for the Model Oil. Concerning changes in the asphaltene characteristics as a function of the heptane/Crude Oil ratio, Fig. 7 shows a comparison of the solubility profiles at 24 h. As shown, there is an increase in the response as the ratio increases consistent with a larger amount of asphaltenes in the filtered cake. As it can be seen, there is also a relative increase in the size of the second peak with respect to the first peak, indicating the enrichment of the filtered cake with less soluble species (longer times: larger solubility parameters) as the ratio increases. However, it is also noticeable that the second peak moves to the left (shorter times: smaller solubility parameters) as the ratio increases, an indication that these new species are in average more soluble. These changes indicate that as more heptane is added to crude oil as happens in a titration experiment, more soluble species precipitate. Also, it shows that most the species that become part of the aggregates at low concentrations of heptane are not the least soluble. Fig. 8 shows changes in the molar H/C ratio for the filtered cakes at 24 h as a function of the composition of the blend (wt% of heptane). There is a decrease in the molar H/C ratio indicating that the filtered cake becomes more and more aromatic as the amount of heptane increases. This corresponds with the enrichment of the filtered cake in asphaltenes as insert in Fig. 8 shows. Similar results were found at different times. Although it was not possible to calculate the average solubility parameter of the filtered cake, it was possible to evaluate the contribution of the asphaltenes to this parameter, based on its weight
Fig. 8. H/C ratio of the filtered cake as a function of the n-heptane concentration in the blend at 24 h. Insert shows increase of asphaltene content as a function of the nheptane/Crude Oil ratio.
277
E. Rogel, M. Moir / Fuel 208 (2017) 271–280
Weight Contribuon to the Solubility Parameter
contribution. The average solubility parameter of the asphaltenes present in the cake was obtained following a known procedure [20] using the asphaltene solubility profiles of the filtered cakes. Fig. 9 shows the correlation between the weight contribution of the asphaltenes to the solubility parameter of the filtered cakes as a function of the molar H/C ratio for different heptane/oil ratios at different times (24 h, 1 and 4 weeks). Even though this correlation lacks the contribution of the maltenes, it is indicative that the solubility parameter is related to the hydrogen deficiency of the
18.00 16.00 14.00 12.00 10.00 8.00 6.00 24 h 4.00 168 h 2.00
720 h
0.00 1.00
1.10
1.20
1.30
1.40
1.50
1.60
1.70
H/C rao Fig. 9. Weight contribution of the solubility parameter of asphaltenes to the solubility parameter of the filtered cake as a function of the H/C molar ratio of the cake.
molecules that precipitate, suggesting a minor role for the presence of heteroatomic functionalities. This result agrees with a previous study that indicates a linear relationship between hydrogen content and the solubility parameter of a group of asphaltenes from diverse origins [20]. However, it is important to point out that heteroatomic functionalities can be important players in the interactions between asphaltenes and surfaces and, therefore, they can influence deposition. In an attempt to evaluate the size of the species recovered in the filtered cakes, size exclusion chromatography was used to determine average molecular weights as well as the relative size of different peaks obtained for filtered cake solutions. Fig. 10 shows a comparison of the chromatograms for different ratios (1.5 and 10) as a function of time. There is an increase in the relative size of the first peak (shorter times) with respect to the second peak as time increases. Since the larger species eluted first, then the first peak correspond either directly to larger species or to the formation of aggregates. These changes as a function of time reveal a significant shift in the characteristics of the filtered cake. Fig. 11 present the average molecular weights obtained from the chromatograms shown in Fig. 10. Fig. 11 shows an increase in the apparent molecular weight for the lowest ratio as the deposit ages. For the largest ratio, there is an initial increase followed by a decrease at longer times. This decrease was also observed for other high ratios and it is probably related to the reorganization of the aggregates that decreases the volume to let to a more compact size. A relative increase in the aggregation peak was found when the n-heptane/Crude Oil ratio increases as shown in Fig. 12 at 24 h. In the insert in this figure, a plot of the average molecular weight as a function of the ratio reveals how the molecular weight of the material in the filtered cake increases as the ratio increases.
a)
b)
98 h
168 h
24 h
24 h
Response
Response
168 h
7h
8h
5h
6h
3h
5h
2h
2h
1h
0
2
4
6
Time (min)
8
10
1h
12
0
2
4
6
8
10
12
Time (min)
Fig. 10. Size Exclusion Chromatograms for the filtered cakes collected as a function of time at a) 1.5 n-heptane/Crude Oil ratio and b) 10 n-heptane/Crude Oil ratio.
278
E. Rogel, M. Moir / Fuel 208 (2017) 271–280
a vast array of different molecules. Initially, in the crude oil, components of the so-called ‘‘asphaltene fraction” are in a relatively good solvent medium. In the original crude oil, before the addition of a non-solvent, a fraction of asphaltenes is forming small aggregates or nano-aggregates. The formation of nano-aggregates is driven by pi-pi non-covalent interactions, and their growth is limited by the presence of alkyl chains that surround the polyaromatic rings of asphaltenes as shown by calculations using molecular thermodynamic modeling [25]. The strength of the pi-pi interaction drives the aggregation as they are the main reason for the low solubility of polyaromatic rings in different solvents. These nano-aggregates are in a pseudo-equilibrium state. Change is triggered by the addition of a non-solvent such as n-heptane. The data indicates that the modifications in the solvent power caused by the addition of a non-solvent initiate a complex aggregation process that, as it has been shown in this work, includes the exchange of molecules in and out of aggregates in addition to the growth of aggregates that has been reported numerous times in the literature [23,24,26]. Fig. 14 shows a proposed model for the aggregation from the molecular point of view. As stated before, it is assumed that the crude oil contains nanoaggregates (Fig. 14a). When the solvent power decreases with the addition of a non-solvent, the formation of disordered larger aggregates is induced. In these aggregates, formed in local environments, a variety of molecules of different kinds are present (Fig. 14 b). Based on an arbitrary definition [27], some are called asphaltenes, and some are called maltenes. They represent a distribution with a continuum of solubilities as we have shown in previous work [19,21]. The process of formation of more organized domains in the aggregates happens next (Fig. 14c). It is driven by pi-pi interactions and, possibly by hydrogen bonding to a lesser extent. This process requires the expelling of soluble molecules with a lower capability to engage in pi-pi interactions as well as the incorporation of less soluble molecules, some of them already forming nanoaggregates. This complex process is driven by the limited solubility of some components in the blend crude oil-non-solvent. There are energetic barriers to this reorganization as non-covalent bonds are
A plot of the area of the first peaks of the chromatograms as a function of the composition of the filtered cake (see Fig. 13) reveals that this peak is proportional to the asphaltene content. This relationship suggests that the increase in molecular weight is the result of more asphaltenes in the filtered cake. Another interesting trend is the decrease of the molar H/C ratio as the n-heptane/Crude Oil ratio increases (Fig. 8). This decrease is an indication that the asphaltenes become more dominant in the filtered cake and therefore, average aromaticity increases as well. However, according to solubility profile evaluations, asphaltene per se became more soluble as the n-heptane/Crude Oil ratio increases. This trend agrees with the decrease of the solvent power of the blend that increases the amount of asphaltenes that precipitate. It is well known that the phenomenon that lead to asphaltene precipitation is the result of multiple noncovalent interactions of
3500 3000
Mn(Da)
2500 2000 1500 1.5 1000
10
500 0
0
20
40
60
80
100
120
140
160
180
Time (min) Fig. 11. Apparent molecular weights calculated as a function of time for low (1.5) and high (10) n-heptane/Crude Oil ratios as function of time. Trace lines added for visualization purposes.
100
3500 3000 2500
Mn (Da)
Response
50
10
2000 1500 1000
5
500 0
2
0
20
40 60 80 Heptane/Oil Rao
100
120
1.5 0
2
4
6 8 Time (min)
10
12
Fig. 12. Comparison of Size exclusion chromatograms for filtered cakes collected at 24 h using different n-heptane/Crude Oil ratios and their respective apparent molecular weight (Mn). Trace lines added for visualization purposes.
E. Rogel, M. Moir / Fuel 208 (2017) 271–280
different processes should vary depending on the composition of the initial blend of n-heptane and crude oil. However, it is possible to assume that the formation of the initial aggregates happens relatively fast (k1), while reorganization and further aggregation (k2 and k3) take longer times. Although at long times, the incorporation of molecules is dominant as shown in Fig. 4, we cannot exclude the idea that molecules are still being expelled from the aggregates considering the evidence provided by sequential asphaltene extraction [28].
1200
1000
Area of first peak
279
800
600
400
4. Conclusions 200
0 0
10
20
30
40
50
60
70
80
90
Asphaltene Content (wt%) Fig. 13. Area of the first peak from the size exclusion chromatograms as a function of the asphaltene content for all the filtered cakes.
broken to expel and incorporate new molecules. The need of long extraction processes [1,28] to somehow ‘‘purify” asphaltenes shows clearly how high these energetic barriers are and also indicates the large amount of components in asphaltenes that apparently are occluded. This reorganization process goes into parallel with the fusion of and fission of aggregates in the blend (Fig. 14d). According to the results of this work, significant expelling of molecules takes place at short times, and it was only observable at low heptane concentrations when the viscosity is high indicating a process limited by diffusion. After some time, the incorporation of molecules becomes prevalent and aggregation starts to slow down. Rates for these
Based on the results obtained using crude oil, it can be hypothesized that kinetics plays a large role in the composition of deposits and 24 h does not seem to be enough time to reach a steady stage at the conditions of these experiments. The study of the composition of the filtered cakes indicates that it changes dramatically overall for low ratios heptane/oil as a function time. Precipitated material is composed mainly by maltenes at low ratios, while at higher ratios, it becomes dominated by asphaltenes or higher aromaticity components of the crude oil. Changes in H/C ratios support this conclusion indicating an increase in the aromatic character of the filtered cake because of its enrichment in asphaltenes as time passes. Increases in asphaltene content were found to be related to the relative increase of high molecular weight species or aggregates in the filtered cake under the conditions of size exclusion chromatography experiments. Experimental evidence shows that the precipitation of asphaltenes is a complex process that involves not only aggregation but also reorganization of the aggregates. The complex behaviors described in this work are usually not taking into consideration in experiments or the modeling of
Fig. 14. The proposed model for the aggregation of asphaltenes under poor solvent conditions.
280
E. Rogel, M. Moir / Fuel 208 (2017) 271–280
asphaltene solubility. From a practical viewpoint, the present study provides new information about the mechanism that leads to asphaltene deposition that might be helpful in the development of new strategies to study this phenomenon. Acknowledgement The authors are grateful to Chevron ETC for providing funding and for allowing the publication of this work. References [1] Strausz OP, Torres M, Lown EM, Safarik I, Murgich J. Energy Fuels 2006;20:2013–21. [2] Nikooyeh K, Reza Bagheri S, Shaw JM. Energy Fuels 2012;26:1756–66. [3] Carbognani Ortega L, Rogel E, Vien J, Ovalles C, Guzman H, Lopez-Linares F, et al. Energy Fuels 2015;29:3664–74. [4] Speight JG, Long RB, Trowbridge D. Fuel 1984;63:616–20. [5] Andersen SI. Fuel Sci Technol Int 1994;12:51–74. [6] Calles JA, Dufour J, Marugan J, Pena JL, Gimenez-Aguirre R, Merino-Garcia D. Energy Fuels 2008;22:763–9. [7] Alboudwarej H, Beck J, Svrcek WY, Yarranton HW, Akbarzadeh K. Energy Fuels 2002;16:462–9. [8] Taqbool T, Balgoa AT, Fogler HS. Energy Fuels 2009;23:3681–6. [9] Haji-Akbari N, Teeraphapkul P, Balgoa AT, Fogler HS. Energy Fuels 2015;29:2190–6. [10] Rogel E, Ovalles C, Vien J, Moir M. Fuel 2016;178:71–6.
[11] Rogel, E, Ovalles, C, Moir, M, Vien, J, Morazan, H. SPE International Symposium on Oilfield Chemistry, The Woodlands, Texas, USA, 13–15 April 2015. SPE 173757. [12] Angle CW, Long Y, Hamza H, Lue L. Fuel 2006;85:492–506. [13] Das SK. SPE J 1998;3:232–7. [14] Nenninger, J, Nenninger, E. Method and apparatus for stimulating heavy oil production US Patent 6,883,607; 2005. [15] Ovalles, C.; Rogel, E.; Alboudwarej, H.; Inouye, A. T.; Benson, I. P.; Vaca, P. Soc. Pet. Eng. – SPE Canada Heavy Oil Tech. Conf., CHOC, 2015. [16] American Society for Testing and Materials. Standard test method for determination of asphaltenes (heptane insolubles) in crude petroleum and petroleum products – IP143/01 ASTM D6560; 2005. [17] American society for testing and materials standard test method for pentane insolubles by membrane filtration ASTM D-4055, 2013. [18] Rogel E, Ovalles C, Moir ME, Schabron JF. Energy Fuels 2009;23:4515–21. [19] Rogel E, Ovalles C, Moir M. Energy Fuels 2010;24:4369–74. [20] Rogel E, Roye M, Vien J, Miao T. Energy Fuels 2015;29:2143–52. [21] Rogel E, Ovalles C, Moir M. Energy Fuels 2012;26:2655–62. [22] Yau WW, Kirkland JJ, Bly DD. Modern size exclusion chromatography. NY, USA: Wiley-Interscience; 1979. [23] Chaisoontornyotin W, Haji-Akbari N, Fogler HS, Hoepfner MP. Energy Fuels 2016;30:1979–86. [24] Haji-Akbari N, Masirisuk P, Hoepfner MP, Fogler HS. Energy Fuels 2013;27:2497–505. [25] Rogel E. Langmuir 2004;20:1003–12. [26] Mohammadi S, Rashidi F, Mousavi-Dehghani SA, Ghazanfari M-H. Can J Chem Eng 2016;94:1820–9. [27] Richardson C. The modern asphalt pavement. New York: John Wiley and Sons; 1905. [28] Chacón-Patiño ML, Vesga-Martínez SJ, Blanco-Tirado C, Orrego-Ruiz JA, Gómez-Escudero A, Combariza MY. Energy Fuels 2016;30:4550–61.
Contents lists available at ScienceDirect
Fuel journal homepage: www.elsevier.com/locate/fuel
Full length article
Effect of precipitation time and solvent power on asphaltene characteristics Estrella Rogel ⇑, Michael Moir Chevron Energy Technology Company, Richmond, CA 94597, United States
h i g h l i g h t s Composition of precipitated asphaltenes changes widely as a function of time. Asphaltene precipitation involves reorganization of the aggregates. Reorganization of aggregates includes expulsion and incorporation of molecules. 24 h is not enough time to reach a steady composition of the asphaltenes.
a r t i c l e
i n f o
Article history: Received 20 March 2017 Received in revised form 5 June 2017 Accepted 24 June 2017
Keywords: Precipitation kinetics Asphaltenes Maltenes Precipitated material Solubility Aggregation model
a b s t r a c t The kinetics of asphaltene precipitation is studied regarding the characterization of the precipitated material. The association of different components of the crude oil depends on the solvent power of the blend heptane/oil and produces aggregates with a variable composition that change as a function of time. It was found that the precipitate contains a significant amount of maltenes whose relative content decreases as a function of time. In agreement with this behavior, the precipitate becomes more hydrogen deficient. These changes also produced a significant increase in the apparent average molecular weight of the precipitate as measured using size exclusion chromatography. This increase correlated with the amount of asphaltenes in the precipitated material independently of time or heptane/oil ratio. Solubility profile measurements showed that the average characteristics of the asphaltenes found in the precipitate varied as a function of time. As more asphaltene molecules precipitated, the distributions of these molecules changed decreasing its average solubility parameter. Considerable differences in the amount collected at low and high heptane/oil ratios as a function of time were also found. Based on the measurements, a new model was developed that describes the complex aggregation behavior involved in asphaltene precipitation. Ó 2017 Elsevier Ltd. All rights reserved.
1. Introduction Asphaltenes are undesirable compounds that precipitate during different stages of petroleum production and refining. As they are not defined as a compound class but as a solubility class, characterization of asphaltenes is challenging since variations in the procedure to obtain them produce not only different amounts but also distributions of compounds with different characteristics [1–3]. It is well known that asphaltene precipitation is affected by several factors, including type of solvent [1,3,4], temperature [5,6], precipitant/sample ratio [3,7], contact time [7], and washing of the filter cake [2,7]. In practical terms, this means that comparison of asphaltenes obtained using even slightly different methodolo⇑ Corresponding author. E-mail address: [email protected] (E. Rogel). http://dx.doi.org/10.1016/j.fuel.2017.06.116 0016-2361/Ó 2017 Elsevier Ltd. All rights reserved.
gies must be made carefully: the asphaltene yields and the chemical composition are going to be different. The main reason for this behavior is the high polydispersity regarding composition and molecular weight of petroleum components. This combination of factors results that under slightly different conditions, a different set of molecules precipitate forming a distinct asphaltene. Although it is expected that the molecules more prone to precipitate are going to be those with the higher aromaticity (poor in hydrogen), higher molecular weight and/or higher polar nature, significant variations in the amount and distribution of the molecules can occur as shown by previous studies [3,5]. Particularly important is the understanding of kinetic aspects related to asphaltene precipitation as this is a variable frequently neglected in asphaltene studies [8] and can have important practical consequences during production and processing of crude oils. Recently, a series of experiments using optical microscopy
272
E. Rogel, M. Moir / Fuel 208 (2017) 271–280
Blend
+
Precipitaon
Filtraon
Nominal pore size: 0.22 µ
Filtered Cake
Crude Oil or Model Oil
n-Heptane
Characterizaon
Shake vigorously for 10 min Fig. 1. Scheme of the experiments.
2. Experimental section 2.1. Materials Crude oil and n-pentane asphaltenes were used in the precipitation experiments. Crude Oil A is a medium crude oil (30.4 API) with an asphaltene content 2.43 wt% using the standard method ASTM D-6560 [16]. n-Pentane asphaltenes were obtained following the standard ASTM D-4055 [17] from a heavy Crude Oil B (7.7oAPI) with an n-pentane asphaltene content of 18.7 wt%. Toluene, nheptane, n-pentane, methylene chloride and methanol are HPLC grade (Fisher Chemical)
2.2. Separation of precipitated material Studied samples included Model Oil and Crude Oil A. The Model Oil was prepared by dissolving the n-pentane asphaltenes in toluene (2 wt%) n-Heptane was used to induce precipitation at room temperature of the Model Oil, and Crude Oil A. Different n-heptane/ samples ratios from 1.5 to 100 were used. Flocculation onset experiments using n-heptane as precipitant agent indicated that precipitation starts at a n-heptane/Crude Oil ratio of 1.5. Blends prepared with a n-heptane/Crude Oil ratio of 1 did not show precipitation during the first 24 h. Solutions are prepared by mixing 10 mL of sample with the corresponding volume of n-heptane and shaking vigorously for 10 min in a shaker. After blending, the solutions are kept in static conditions. Blends were filtered using a Teflon membrane filter with an average pore size of 0.2 mm so that particles could be recovered and analyzed. Fig. 1 shows a scheme of the procedure to obtain the filtered cakes. Since the goal is to replicate the precipitation behavior as close as possible to a real situation, the filtered cake was not washed with additional n-heptane after filtration. Additionally, it has been reported that precipitation procedures where asphaltenes are not washed seem to be more repeatable [2]. Once collected, the filtered cake was dried under nitrogen flow, weighed and analyzed using the characterization methods listed below. Samples were run in duplicates which on average differ by 10%.
30
25 Collected Amount (wt%)
observations and centrifugation based separations [8,9], has shown that kinetic factors are underestimated during conventional flocculation experiments aimed to determine chances of asphaltene precipitation. In fact, it was found that these test over predict solubility as it was observed that the beginning of precipitation could take from minutes to several months depending on the precipitant concentration used [8,9]. From these experiments, we also know that the precipitated amount varies as a function of time and that those variations depend on the precipitant concentration. When asphaltenes precipitate in the reservoir or during refinery operations, factors such as solvent power and time have also a significant influence. In particular, aging can have a definitive impact on deposit characteristics [10,11] or emulsion formation [12]. Additionally, kinetic effects and solvent power play a relevant role in understanding how recently developed solvent injection processes (i.e. Vapex or N-solv) [13,14] might affect the reservoir and crude oil production. In these processes, hydrocarbons in the range C3 to C5 are injected downhole to make the crude oil mobile. Recent physical and numerical simulations [15] of propane injection into a heavy oil reservoir indicated that the kinetics of asphaltene precipitation plays a significant role in the production as well as in the quality of the produced crude oil. From many perspectives, the understanding of how time and solvent power affect asphaltene composition can provide useful information to design new strategies in diverse areas including analytical characterization, precipitation prevention, deposit removal, and new production methods. In the present work, the effect that time and solvent power have on asphaltene properties is investigated using a combination of analytical techniques. We evaluate several physicochemical properties that are strongly linked to precipitation behavior such as hydrogen deficiency, solubility distribution, and molecular size. The changes in the amount and composition of the precipitated material as a function of time are also examined.
20
15
10 1.5 5
2
0 0
50
100
150
200
Time (h)
Fig. 2. Collected amount as a function of time and n-heptane/Model Oil ratio. Trace lines added for visualization purposes.
273
E. Rogel, M. Moir / Fuel 208 (2017) 271–280
changed gradually from pure n-heptane to 90/10 methylene chloride/methanol and then to 100% methanol. This procedure gradually re-dissolves the asphaltenes from the easy to dissolve (low solubility parameter) to the hard to dissolve (high solubility parameter) producing a distribution that represents the range of asphaltenic species with different solubilities present in the sample. In a similar way to the determination of asphaltene content, asphaltenes are quantified using an Evaporative Light Scattering Detector (ELSD). Rather recently, this method has been calibrated to produce a distribution of asphaltenes based on their solubility parameters. A description of the calibration is published elsewhere [20]. Elemental analysis of the samples was carried out by a standard combustion method, using a Carlo Erba model 1108 analyzer. Aggregation studies were carried out by size exclusion chromatography (SEC) using a 30 cm 7.5 mm PLgel ‘‘mixed E” column. Sample solutions (0.01 g in 10 mL of methylene chloride)
2.3. Characterization methods Asphaltene concentrations and asphaltene solubility profiles were determined using on-column filtration techniques [18,19]. In these techniques, a solution of the sample (0.01 g in 10 mL of methylene chloride) is injected into a column packed with an inert material using n-heptane as the mobile phase. This solvent induces the precipitation of asphaltenes and, therefore, their retention in the column. To determine the asphaltene content, the mobile phase is switched to a blend methylene chloride/methanol 90/10 v/v that redissolves the asphaltenes completely. Asphaltenes are quantified using an Evaporative Light Scattering Detector (ELSD). This technique produces results that correlate with the results obtained using the conventional gravimetric technique (ASTM D6560) [16]. In the asphaltene solubility profile method, after the initial injection and precipitation of asphaltenes, the mobile phase is
21.2
a)
2
9.0E-03 1.5
Normalized Response
8.0E-03
5
Pentane Asphaltenes
7.0E-03 6.0E-03
Average Solubility Parameter (MPa0.5)
1.0E-02
21.0 20.8 20.6 20.4 20.2 20.0
Pentane Asphaltenes
19.8 0
2
4 6 8 n-heptane/Model Oil Rao
10
12
5.0E-03 4.0E-03
10
3.0E-03 2.0E-03 1.0E-03 0.0E+00 13
14
15
16
17
18
19
20
Time (min)
1.0E-02
Normalized Response
b) 9.0E-03
1h
8.0E-03
2h
7.0E-03
3h
6.0E-03
24 h
5.0E-03
168 h
4.0E-03 3.0E-03 2.0E-03 1.0E-03 0.0E+00 13
14
15
16
17
18
19
20
Time (min) Fig. 3. Comparison of asphaltene solubility profiles of the filtered cakes. a) Filtered cakes collected at 3 h for different n-heptane/Model Oil ratios (1.5, 2, 5, 10) and original npentane asphaltenes. b) Filtered cakes collected at different times for low (1.5) ratio.
274
E. Rogel, M. Moir / Fuel 208 (2017) 271–280
Fig. 4. Collected amount as a function of time for low (1.5) and high (10) n-heptane/Crude Oil ratios. Insert shows behavior at short times for low (1.5) ratio. Trace lines added for visualization purposes.
3.1. Model Oil Initial experiments were performed using a Model Oil prepared using n-pentane asphaltene in toluene (2 wt%). Fig. 2 shows the amounts of material collected at increasing times for two different n-heptane/Model Oil ratios (1.5 and 2). In this plot, the collected material is calculated as a percentage of the initial mass of npentane asphaltenes. As expected, the amount collected increases as a function of time as well as when the n-heptane/sample ratio increases. Fig. 3a shows the solubility profile for several of the filtered cakes separated using heptane/sample ratios of 1.5–10 and collected at similar times (3 h). The solubility profiles of these materials varied significantly showing that different ratios produced materials that exhibit different solubility characteristics. In particular, the results indicated that lower ratios at this short time (3 h) produced filtered cakes enriched in more soluble material (eluted at shorter times) than larger ratios. In this figure, the solubility profile of n-pentane asphaltenes is also shown for comparison. The more material is collected (Fig. 2), the solubility profile of the asphaltenes becomes closer to the one obtained for the npentane asphaltenes as can be expected. This tendency can be observed in the insert in Fig. 3a showing the average solubility parameter of the asphaltenes as calculated using the solubility profile technique. These values indicate that as more material is precipitated, the average solubility decreases. As the solvent power
a) Content based on crude oil (wt%)
3. Results and discussion
1.40 1.20 1.00 0.80 0.60 0.40 Asphaltenes 0.20
Maltenes
0.00 0
100
200
300
400
500
600
700
800
Time (h) 2.50
b) Content based on crude oil (wt%)
were prepared in methylene chloride. Elution was carried out with 90/10 methylene chloride/methanol at a flow rate of 1.0 mL/min, and temperature kept constant at 25 °C. The molecular weights (Mw) were calculated based on a calibration that uses porphyrins, dyes, and polyaromatics as standards, as published before [21]. Molecular weights were reported as Mw and Mn values [22] Mw have a relative error of 5%.
2.00
1.50
1.00
0.50
Asphaltenes Maltenes
0.00 0
100
200
300
400
500
600
700
800
Time (h) Fig. 5. Asphaltene and maltene content in the filtered cake for low (1.5) and high (10) n-heptane/Crude Oil ratios as function of time. Trace lines added for visualization purposes.
275
E. Rogel, M. Moir / Fuel 208 (2017) 271–280
decreases and more molecules (with high solubility in heptane) move to the precipitate, the solubility parameter of the asphaltenes decreases in average. In contrast with the effect of n-heptane/Model Oil ratio, comparison of asphaltene solubility profiles for the same n-heptane/ Model Oil ratios at different times up to 1 week did not show significant differences in the profile as shown in the example presented in Fig. 3b. 3.2. Crude Oil Fig. 4 shows the amount of solids collected in the filter as a using heptane/Crude Oil ratios of 1.5 and 10. In this figure, changes of the amount collected as a function of time up to 1 month (720 h) indicate different behaviors depending on the ratio. At the low ratio, an initial decrease of the amount is followed by an increase, while for the larger ratio the amount collected increases from the beginning of the experiment. These significant differences at short times are apparently related to the composition of the blend. It has been reported [2] that at low n-alkane concentrations, n-alkanes penetrate asphaltene particles causing swelling and dissolving a fraction of the asphaltenes. An alternative explanation for this contrasting behavior might be found in the diffusion of molecules from the aggregates to the fluid is limited by a higher viscosity. Asphaltene content in the filtered cake was evaluated using the on-column methodology. This test is a measurement that correlates with the asphaltene content measured using ASTM D-6560 [16], the gravimetric standard method that employs heptane as
precipitant. Fig. 5 shows a comparison of how asphaltene and maltene content as measured by the on-column technique changes as a function of time for low (1.5) and high (10) n-heptane/Crude Oil ratios. For the lower ratio, asphaltene content increases in the filtered cake as a function of time, while the maltene amount initially decreases and then starts to increase. In the case of high ratio, both asphaltene and maltene amounts increases, although asphaltene content seems to increase faster and dominate the filtered cake (60–70 wt%). Previous studies of the kinetics of aggregation of asphaltenes have not revealed these differences, mainly because of the washing of the filtered cakes with additional heptane to eliminate trapped crude oil [8,9,23]. In the present work, as mentioned before, the filtered cakes were not washed based on our goal of replicate real precipitation conditions and analyze how the aggregates change with time. We think that these changes are essential to understand the evolution of deposits [11]. Based on previous studies [24], the kinetics of aggregation depends on the viscosity of the blend as well as in the solubility of the heavy hydrocarbons in it. In principle, it can be assumed that once the solubility limit for different components is crossed, diffusion-limited aggregation will occur spontaneously due to the low solubility of these components in the blend. Apparently, at low ratios and short times, an exchange of components takes place where some of the components trapped in the filtered cake go back into solution while new components are incorporated in the aggregates. This exchange indicates that the activation energy of this process is relatively low and the exchange proceeds relatively rapidly. It is important to mention that the redissolution of trapped
a)
b) 168 h 168 h
98 h 24 h
24 h
Response
Response
8h
7h
6h
5h
5h
3h
2h
2h
1h
12
14
16
18
Time (min)
20
1h
22
12
14
16
18
20
22
Time (min)
Fig. 6. Asphaltene solubility profiles for the filtered cakes collected as a function of time at a) 1.5 n-heptane/Crude Oil ratio and b) 10 n-heptane/Crude Oil ratio. For comparison purposes, the profiles were normalized.
276
E. Rogel, M. Moir / Fuel 208 (2017) 271–280
material could also happen at high ratios, but it was not observed because it might happen so fast that the measurements made cannot capture it. Changes in the characteristics of the asphaltenes were also evaluated as a function of time. Fig. 6 shows these changes in the solubility distribution of the heptane asphaltenes in the filtered cake for low and high ratios. For the low ratio, changes are clear: there is an apparent decrease in the more soluble components (1st peak) while the second peak increases (less soluble components). For the large ratios, changes are less significant, but still consistent with a relative decrease in the first peak. In contrast to these results, filtered cakes obtained from the Model Oil did not show major differences in composition as a function of time (Fig. 3b). These different behaviors might suggest that when the precipitation occurs from the whole crude oil, lighter components are involved that are not present when the precipitation occurs in n-pentane asphaltene solutions (Model Oil). Also, the Model Oil blends have lower viscosities than the equivalent Crude Oil blends and, consequently, they reach equilibrium faster.
1200 50 1000
Response
800
100
600 10 400
5
200
2 1.5 0 12
14
16
18
20
22
Time (min) Fig. 7. Asphaltene solubility profile of the filtered cake as a function of n-heptane/ Crude Oil ratio at 24 h.
A significant body of research has pointed out to the presence of components in asphaltenes whose characteristics do not correspond to the expected characteristics of an asphaltene molecule even after significant soxhlet extraction procedures. For instance, a study of experimental data demonstrated the presence of a reversible adsorption phenomenon in the solvent precipitation of asphaltene from the parent crude oil involving the coprecipitation of precipitant-soluble resins and asphaltene -like substances [1]. In fact, the authors of this study estimated that around 70% of heptane asphaltenes in a bitumen were associated resinous material [1]. It is important to point out that the filtered cake was not washed with solvents before analysis. After extensive washing with heptane, a common practice, plots of the filtered cake become more and more like those shown in Fig. 3b for the Model Oil. Concerning changes in the asphaltene characteristics as a function of the heptane/Crude Oil ratio, Fig. 7 shows a comparison of the solubility profiles at 24 h. As shown, there is an increase in the response as the ratio increases consistent with a larger amount of asphaltenes in the filtered cake. As it can be seen, there is also a relative increase in the size of the second peak with respect to the first peak, indicating the enrichment of the filtered cake with less soluble species (longer times: larger solubility parameters) as the ratio increases. However, it is also noticeable that the second peak moves to the left (shorter times: smaller solubility parameters) as the ratio increases, an indication that these new species are in average more soluble. These changes indicate that as more heptane is added to crude oil as happens in a titration experiment, more soluble species precipitate. Also, it shows that most the species that become part of the aggregates at low concentrations of heptane are not the least soluble. Fig. 8 shows changes in the molar H/C ratio for the filtered cakes at 24 h as a function of the composition of the blend (wt% of heptane). There is a decrease in the molar H/C ratio indicating that the filtered cake becomes more and more aromatic as the amount of heptane increases. This corresponds with the enrichment of the filtered cake in asphaltenes as insert in Fig. 8 shows. Similar results were found at different times. Although it was not possible to calculate the average solubility parameter of the filtered cake, it was possible to evaluate the contribution of the asphaltenes to this parameter, based on its weight
Fig. 8. H/C ratio of the filtered cake as a function of the n-heptane concentration in the blend at 24 h. Insert shows increase of asphaltene content as a function of the nheptane/Crude Oil ratio.
277
E. Rogel, M. Moir / Fuel 208 (2017) 271–280
Weight Contribuon to the Solubility Parameter
contribution. The average solubility parameter of the asphaltenes present in the cake was obtained following a known procedure [20] using the asphaltene solubility profiles of the filtered cakes. Fig. 9 shows the correlation between the weight contribution of the asphaltenes to the solubility parameter of the filtered cakes as a function of the molar H/C ratio for different heptane/oil ratios at different times (24 h, 1 and 4 weeks). Even though this correlation lacks the contribution of the maltenes, it is indicative that the solubility parameter is related to the hydrogen deficiency of the
18.00 16.00 14.00 12.00 10.00 8.00 6.00 24 h 4.00 168 h 2.00
720 h
0.00 1.00
1.10
1.20
1.30
1.40
1.50
1.60
1.70
H/C rao Fig. 9. Weight contribution of the solubility parameter of asphaltenes to the solubility parameter of the filtered cake as a function of the H/C molar ratio of the cake.
molecules that precipitate, suggesting a minor role for the presence of heteroatomic functionalities. This result agrees with a previous study that indicates a linear relationship between hydrogen content and the solubility parameter of a group of asphaltenes from diverse origins [20]. However, it is important to point out that heteroatomic functionalities can be important players in the interactions between asphaltenes and surfaces and, therefore, they can influence deposition. In an attempt to evaluate the size of the species recovered in the filtered cakes, size exclusion chromatography was used to determine average molecular weights as well as the relative size of different peaks obtained for filtered cake solutions. Fig. 10 shows a comparison of the chromatograms for different ratios (1.5 and 10) as a function of time. There is an increase in the relative size of the first peak (shorter times) with respect to the second peak as time increases. Since the larger species eluted first, then the first peak correspond either directly to larger species or to the formation of aggregates. These changes as a function of time reveal a significant shift in the characteristics of the filtered cake. Fig. 11 present the average molecular weights obtained from the chromatograms shown in Fig. 10. Fig. 11 shows an increase in the apparent molecular weight for the lowest ratio as the deposit ages. For the largest ratio, there is an initial increase followed by a decrease at longer times. This decrease was also observed for other high ratios and it is probably related to the reorganization of the aggregates that decreases the volume to let to a more compact size. A relative increase in the aggregation peak was found when the n-heptane/Crude Oil ratio increases as shown in Fig. 12 at 24 h. In the insert in this figure, a plot of the average molecular weight as a function of the ratio reveals how the molecular weight of the material in the filtered cake increases as the ratio increases.
a)
b)
98 h
168 h
24 h
24 h
Response
Response
168 h
7h
8h
5h
6h
3h
5h
2h
2h
1h
0
2
4
6
Time (min)
8
10
1h
12
0
2
4
6
8
10
12
Time (min)
Fig. 10. Size Exclusion Chromatograms for the filtered cakes collected as a function of time at a) 1.5 n-heptane/Crude Oil ratio and b) 10 n-heptane/Crude Oil ratio.
278
E. Rogel, M. Moir / Fuel 208 (2017) 271–280
a vast array of different molecules. Initially, in the crude oil, components of the so-called ‘‘asphaltene fraction” are in a relatively good solvent medium. In the original crude oil, before the addition of a non-solvent, a fraction of asphaltenes is forming small aggregates or nano-aggregates. The formation of nano-aggregates is driven by pi-pi non-covalent interactions, and their growth is limited by the presence of alkyl chains that surround the polyaromatic rings of asphaltenes as shown by calculations using molecular thermodynamic modeling [25]. The strength of the pi-pi interaction drives the aggregation as they are the main reason for the low solubility of polyaromatic rings in different solvents. These nano-aggregates are in a pseudo-equilibrium state. Change is triggered by the addition of a non-solvent such as n-heptane. The data indicates that the modifications in the solvent power caused by the addition of a non-solvent initiate a complex aggregation process that, as it has been shown in this work, includes the exchange of molecules in and out of aggregates in addition to the growth of aggregates that has been reported numerous times in the literature [23,24,26]. Fig. 14 shows a proposed model for the aggregation from the molecular point of view. As stated before, it is assumed that the crude oil contains nanoaggregates (Fig. 14a). When the solvent power decreases with the addition of a non-solvent, the formation of disordered larger aggregates is induced. In these aggregates, formed in local environments, a variety of molecules of different kinds are present (Fig. 14 b). Based on an arbitrary definition [27], some are called asphaltenes, and some are called maltenes. They represent a distribution with a continuum of solubilities as we have shown in previous work [19,21]. The process of formation of more organized domains in the aggregates happens next (Fig. 14c). It is driven by pi-pi interactions and, possibly by hydrogen bonding to a lesser extent. This process requires the expelling of soluble molecules with a lower capability to engage in pi-pi interactions as well as the incorporation of less soluble molecules, some of them already forming nanoaggregates. This complex process is driven by the limited solubility of some components in the blend crude oil-non-solvent. There are energetic barriers to this reorganization as non-covalent bonds are
A plot of the area of the first peaks of the chromatograms as a function of the composition of the filtered cake (see Fig. 13) reveals that this peak is proportional to the asphaltene content. This relationship suggests that the increase in molecular weight is the result of more asphaltenes in the filtered cake. Another interesting trend is the decrease of the molar H/C ratio as the n-heptane/Crude Oil ratio increases (Fig. 8). This decrease is an indication that the asphaltenes become more dominant in the filtered cake and therefore, average aromaticity increases as well. However, according to solubility profile evaluations, asphaltene per se became more soluble as the n-heptane/Crude Oil ratio increases. This trend agrees with the decrease of the solvent power of the blend that increases the amount of asphaltenes that precipitate. It is well known that the phenomenon that lead to asphaltene precipitation is the result of multiple noncovalent interactions of
3500 3000
Mn(Da)
2500 2000 1500 1.5 1000
10
500 0
0
20
40
60
80
100
120
140
160
180
Time (min) Fig. 11. Apparent molecular weights calculated as a function of time for low (1.5) and high (10) n-heptane/Crude Oil ratios as function of time. Trace lines added for visualization purposes.
100
3500 3000 2500
Mn (Da)
Response
50
10
2000 1500 1000
5
500 0
2
0
20
40 60 80 Heptane/Oil Rao
100
120
1.5 0
2
4
6 8 Time (min)
10
12
Fig. 12. Comparison of Size exclusion chromatograms for filtered cakes collected at 24 h using different n-heptane/Crude Oil ratios and their respective apparent molecular weight (Mn). Trace lines added for visualization purposes.
E. Rogel, M. Moir / Fuel 208 (2017) 271–280
different processes should vary depending on the composition of the initial blend of n-heptane and crude oil. However, it is possible to assume that the formation of the initial aggregates happens relatively fast (k1), while reorganization and further aggregation (k2 and k3) take longer times. Although at long times, the incorporation of molecules is dominant as shown in Fig. 4, we cannot exclude the idea that molecules are still being expelled from the aggregates considering the evidence provided by sequential asphaltene extraction [28].
1200
1000
Area of first peak
279
800
600
400
4. Conclusions 200
0 0
10
20
30
40
50
60
70
80
90
Asphaltene Content (wt%) Fig. 13. Area of the first peak from the size exclusion chromatograms as a function of the asphaltene content for all the filtered cakes.
broken to expel and incorporate new molecules. The need of long extraction processes [1,28] to somehow ‘‘purify” asphaltenes shows clearly how high these energetic barriers are and also indicates the large amount of components in asphaltenes that apparently are occluded. This reorganization process goes into parallel with the fusion of and fission of aggregates in the blend (Fig. 14d). According to the results of this work, significant expelling of molecules takes place at short times, and it was only observable at low heptane concentrations when the viscosity is high indicating a process limited by diffusion. After some time, the incorporation of molecules becomes prevalent and aggregation starts to slow down. Rates for these
Based on the results obtained using crude oil, it can be hypothesized that kinetics plays a large role in the composition of deposits and 24 h does not seem to be enough time to reach a steady stage at the conditions of these experiments. The study of the composition of the filtered cakes indicates that it changes dramatically overall for low ratios heptane/oil as a function time. Precipitated material is composed mainly by maltenes at low ratios, while at higher ratios, it becomes dominated by asphaltenes or higher aromaticity components of the crude oil. Changes in H/C ratios support this conclusion indicating an increase in the aromatic character of the filtered cake because of its enrichment in asphaltenes as time passes. Increases in asphaltene content were found to be related to the relative increase of high molecular weight species or aggregates in the filtered cake under the conditions of size exclusion chromatography experiments. Experimental evidence shows that the precipitation of asphaltenes is a complex process that involves not only aggregation but also reorganization of the aggregates. The complex behaviors described in this work are usually not taking into consideration in experiments or the modeling of
Fig. 14. The proposed model for the aggregation of asphaltenes under poor solvent conditions.
280
E. Rogel, M. Moir / Fuel 208 (2017) 271–280
asphaltene solubility. From a practical viewpoint, the present study provides new information about the mechanism that leads to asphaltene deposition that might be helpful in the development of new strategies to study this phenomenon. Acknowledgement The authors are grateful to Chevron ETC for providing funding and for allowing the publication of this work. References [1] Strausz OP, Torres M, Lown EM, Safarik I, Murgich J. Energy Fuels 2006;20:2013–21. [2] Nikooyeh K, Reza Bagheri S, Shaw JM. Energy Fuels 2012;26:1756–66. [3] Carbognani Ortega L, Rogel E, Vien J, Ovalles C, Guzman H, Lopez-Linares F, et al. Energy Fuels 2015;29:3664–74. [4] Speight JG, Long RB, Trowbridge D. Fuel 1984;63:616–20. [5] Andersen SI. Fuel Sci Technol Int 1994;12:51–74. [6] Calles JA, Dufour J, Marugan J, Pena JL, Gimenez-Aguirre R, Merino-Garcia D. Energy Fuels 2008;22:763–9. [7] Alboudwarej H, Beck J, Svrcek WY, Yarranton HW, Akbarzadeh K. Energy Fuels 2002;16:462–9. [8] Taqbool T, Balgoa AT, Fogler HS. Energy Fuels 2009;23:3681–6. [9] Haji-Akbari N, Teeraphapkul P, Balgoa AT, Fogler HS. Energy Fuels 2015;29:2190–6. [10] Rogel E, Ovalles C, Vien J, Moir M. Fuel 2016;178:71–6.
[11] Rogel, E, Ovalles, C, Moir, M, Vien, J, Morazan, H. SPE International Symposium on Oilfield Chemistry, The Woodlands, Texas, USA, 13–15 April 2015. SPE 173757. [12] Angle CW, Long Y, Hamza H, Lue L. Fuel 2006;85:492–506. [13] Das SK. SPE J 1998;3:232–7. [14] Nenninger, J, Nenninger, E. Method and apparatus for stimulating heavy oil production US Patent 6,883,607; 2005. [15] Ovalles, C.; Rogel, E.; Alboudwarej, H.; Inouye, A. T.; Benson, I. P.; Vaca, P. Soc. Pet. Eng. – SPE Canada Heavy Oil Tech. Conf., CHOC, 2015. [16] American Society for Testing and Materials. Standard test method for determination of asphaltenes (heptane insolubles) in crude petroleum and petroleum products – IP143/01 ASTM D6560; 2005. [17] American society for testing and materials standard test method for pentane insolubles by membrane filtration ASTM D-4055, 2013. [18] Rogel E, Ovalles C, Moir ME, Schabron JF. Energy Fuels 2009;23:4515–21. [19] Rogel E, Ovalles C, Moir M. Energy Fuels 2010;24:4369–74. [20] Rogel E, Roye M, Vien J, Miao T. Energy Fuels 2015;29:2143–52. [21] Rogel E, Ovalles C, Moir M. Energy Fuels 2012;26:2655–62. [22] Yau WW, Kirkland JJ, Bly DD. Modern size exclusion chromatography. NY, USA: Wiley-Interscience; 1979. [23] Chaisoontornyotin W, Haji-Akbari N, Fogler HS, Hoepfner MP. Energy Fuels 2016;30:1979–86. [24] Haji-Akbari N, Masirisuk P, Hoepfner MP, Fogler HS. Energy Fuels 2013;27:2497–505. [25] Rogel E. Langmuir 2004;20:1003–12. [26] Mohammadi S, Rashidi F, Mousavi-Dehghani SA, Ghazanfari M-H. Can J Chem Eng 2016;94:1820–9. [27] Richardson C. The modern asphalt pavement. New York: John Wiley and Sons; 1905. [28] Chacón-Patiño ML, Vesga-Martínez SJ, Blanco-Tirado C, Orrego-Ruiz JA, Gómez-Escudero A, Combariza MY. Energy Fuels 2016;30:4550–61.