Adsorption and Subsequent Oxidation of Colombian Asphaltenes onto Nickel and/or Palladium Oxide Supported on Fumed Silica Nanoparticles

Article pubs.acs.org/EF

Adsorption and Subsequent Oxidation of Colombian Asphaltenes onto Nickel and/or Palladium Oxide Supported on Fumed Silica Nanoparticles Camilo A. Franco,† Tatiana Montoya,† Nashaat N. Nassar,*,‡,§ Pedro Pereira-Almao,‡ and Farid B. Cortés*,† †

Grupo de Investigación en Yacimientos de Hidrocarburos, Facultad de Minas, Universidad Nacional de Colombia Sede MedellínKra, 80 No. 65-223, Medellín, Colombia ‡ Department of Chemical and Petroleum Engineering, University of Calgary, 2500 University Drive NW, Calgary, Alberta, Canada § Department of Chemical Engineering, An-Najah National University, Nablus, Palestine ABSTRACT: High asphaltene content in heavy crude oil normally generates adverse rheological properties that affect the flow through the reservoir, preventing optimal hydrocarbon production. It has been demonstrated that using nanoparticles may improve the mobility of oil. Nanoparticles may be used as adsorbents and catalysts in the oil industry for in situ upgrading. The main objective of this study was to investigate the sorption kinetics and the thermodynamic equilibrium for asphaltene sorption onto nickel and/or palladium oxides supported on fumed silica that was nanoparticulated at different times, temperatures, and concentrations. After adsorption, thermally cracked asphaltenes from Colombian crude oil were investigated using catalytic oxidation. The asphaltenes adsorbed onto the selected nanoparticles were subjected to thermal decomposition up to 700 °C in a thermogravimetric analyzer. This study was realized using an experimental design with a measured simplex−centroid of the three components by varying the wt % of the palladium and nickel oxides as well as the fumed silica as the support. The silica nanoparticles were characterized using N2 adsorption at 196 °C and X-ray diffraction. The Langmuir and Freundlich models were used to correlate the experimental sorption equilibrium data. The experimental asphaltene adsorption isotherm data were adequately adjusted using the Freundlich model. The adsorption of asphaltenes on NiO and/or PdO supported on fumed silica was much higher than that over fumed silica over the range of the tested equilibrium concentrations. Pseudo-first- and pseudosecond-order kinetic models were applied to the experimental data obtained at different asphaltene concentrations from 100 to 1500 mg/L for the virgin fumed silica (S) and fumed silica-supported materials (SHSs); better fits were found for the pseudosecond-order model. However, the nanoparticles significantly decreased the asphaltene decomposition temperature and activation energy. The catalyst kinetics was calculated using the Ozawa−Flynn−Wall Model (OFW). All of the nanoparticles demonstrated high catalytic activity toward asphaltene decomposition in the following order at 0.2 mg/m2 asphaltene concentration on nanoparticle surfaces: SNi1 < SNi1Pd1< SNi0.66Pd0.66 < SPd1 < SPd2 < SNi2 < S. Consequently, using nanoparticles significantly enhanced the thermal decomposition of asphaltenes, improving the mobility of heavy oils in situ.

1. INTRODUCTION Conventional crude oil reserves are declining, but the worldwide energy demand is increasing rapidly. Accordingly, developing new alternative energy resources to sustain industrial activities is of paramount importance. Therefore, unconventional materials such as heavy and extra-heavy oil have recently become important sources of fossil fuel. Worldwide oil resources contain 40% heavy and extra-heavy oils,1 and Colombia is a representative case.2 Heavy and extra-heavy crude oils have extremely high asphaltene content, which causes high viscosity and low specific gravity, hindering processing, production, and transportation. The chemical structures of asphaltenes is generally based in one polycyclic aromatic hydrocarbon (PAH) or many cross-linked PAHs, forming island and archipelago architectures, respectively;3 asphaltenes also contain nonmetallic heteroatoms, such as nitrogen, oxygen, and sulfur, as well as metals, including vanadium, iron, and nickel.4,5 This structure facilitates asphaltene nucleation and growth because its amphiphilic behavior affects viscosity; therefore, oil mobility decreases.6−10 In addition, having © XXXX American Chemical Society

asphaltenes present in crude oil causes many problems during oil recovery and upgrading, such as pore and pipeline blockages caused by asphaltene deposition and catalyst deactivation and poisoning; these pitfalls translate into increased economic expense11 and negative environmental impact. Therefore, removing asphaltenes from crude oil will help reduce the costs of crude oil production and transportation, while improving the crude oil quality to meet stringent market specifications with less environmental impact. Nanoparticle technology is a rapidly growing field that may supply an alternative for the currently available techniques for heavy oil upgrading and recovery.12−15 Nanoparticles might be employed as adsorbents/catalysts for heavy and extra-heavy oil during in situ upgrading and recovery processes, such as removing the asphaltene (heavy molecules) and subsequent oxidation, steam gasification, or thermal cracking, to produce Received: September 16, 2013 Revised: November 3, 2013

A

dx.doi.org/10.1021/ef4018543 | Energy Fuels XXXX, XXX, XXX−XXX

Energy & Fuels

Article

light-quality or synthetic gases.16−26 The unique properties of nanoparticles, such as their high surface area to volume ratio, dispersibility, high adsorption affinity, mineral composition, and catalytic activity, make nanoparticles excellent adsorbent/ catalysts for asphaltene capture and subsequent oxidation/ gasification reactions below the natural oxidation temperature of virgin asphaltenes. The adsorption and subsequent oxidation of asphaltenes on metal oxide nanoparticles was introduced by Nassar et al.18,22,24−27 and has since been tested by several investigators.8,9,22−30 Nassar et al. investigated the adsorption/ oxidation of Athabasca asphaltenes using different metal oxide nanoparticles (i.e., NiO, Co3O4, Fe3O4). During the adsorption process, the authors observed a monolayer adsorption isotherm. After the adsorption process, the authors evaluated the catalytic effect of the nanoparticles; the oxidation temperature of asphaltene decreased by 140 °C, 136 °C, and 100 °C relative to the noncatalytic oxidation of virgin asphaltenes in the presence of NiO, Co3O4, and Fe3O4 nanoparticles, respectively. Further, the nanoparticles significantly decreased the activation energy, confirming their catalytic activity toward asphaltene decomposition. The authors confirmed that the asphaltene adsorption/oxidation is metal oxide specific. A correlation between the adsorption affinity and the catalytic activity of the metal oxide nanoparticles was reported, indicating that higher adsorption affinities increase the catalytic activity. In addition, Nassar et al.22 investigated the adsorption and catalytic oxidation of asphaltenes using two different alumina particles (micro- and nanoparticles) that have different particle sizes and surface acidities.22,24 When the surface acidity is constant, nanoalumina has an adsorption capacity for asphaltenes higher than that of microalumina based on the surface area, while the microalumina has a catalytic activity during asphaltene oxidation higher than that of nanoalumina, revealing that the textural properties are more important than the particle size during the catalytic oxidation of asphaltenes. For different surface acidities, acidic alumina has the highest adsorption of asphaltenes, and basic alumina has the highest catalytic activity toward asphaltene oxidation based on its surface area.25 In a recent study, Nassar and coworkers incorporated NiO nanoparticles into mesoporous−macroporous metakaolin intended to be both a catalyst support and an adsorbent.16 The authors demonstrated that incorporating 3 wt % NiO nanoparticles into these catalyst supports enhanced the adsorption and steam gasification of the adsorbed asphaltenes, generating CH4, CO, and H2 as the major products. More recently, our group also has demonstrated that Colombian asphaltenes might be easily removed from the heavy oil matrix via adsorption onto different nanoparticle surfaces with and without a support. We also confirmed that asphaltene adsorption is metal oxide specific.10 Moreover, the transport of alumina nanoparticles suspended in aqueous solution with potential adsorption value is feasible because the nanoparticles allowed the system to flow, demonstrating the inhibition of both precipitation and deposition as well as the enhanced durability against asphaltene damage in porous media.9 This study continues our previous work, aiming to employ nanomaterials based on palladium and/or nickel oxides supported on fumed silica for adsorption and subsequent oxidation of asphaltene molecules extracted from Colombian heavy crude oil. This study focuses on investigating the sorption kinetics and thermodynamic equilibrium of asphaltene

sorption onto virgin fumed silica nanoparticles as well as onto nickel and/or palladium oxide nanoparticles supported on the same silica nanoparticles. Additionally, we evaluated the catalytic effect of these nanoparticles toward asphaltene oxidation. Furthermore, a simplex−centroid mixture design (SCMD) was used to optimize the concentration of the nickel and/or palladium oxides on the silica surfaces to enhance the catalytic activity.

2. EXPERIMENTAL SECTION 2.1. Materials. 2.1.1. Asphaltenes. The “CAPELLA” crude oil is produced from a reservoir located in the south of Colombia (Putumayo, South of Colombian). The produced crude oil has 10.5°API (with a specific gravity of 0.996), viscosity of 475 886 cP at 25 °C, and an approximate content of 9 wt % of asphaltenes. This crude oil was used as the source of asphaltenes. 2.1.2. Solvents and Salt Precursors. n-Heptane (99%, SigmaAldrich, St. Louis, MO) was used for asphaltene extraction from crude oil. Ni(NO3)2(6H2O) (MerkK GaG, Germany), Pd(NO3)2 (MerkK GaG, Germany), and distilled water were used for functionalization of nanoparticles. Nanoparticles of fumed silica were purchased from Sigma-Aldrich. Toluene (99.5%, MerkK GaG, Germany) was used to prepare heavy oil model solutions. 2.2. Methods. 2.2.1. The Asphaltene Extraction Protocol. Asphaltenes were precipitated from CAPELLA crude oil following a standard procedure.26,27 In brief, an excess amount of n-heptane was added to the crude oil in a volume ratio of 40:1. The mixture was then sonicated for 2 h at 25 °C and further stirred at 300 rpm for 20 h. Black precipitates formed at the bottom. The precipitated fraction was filtered using a 8 μm Whatman filter paper and washed with n-heptane at a ratio of 4:1 (g/mL). Asphaltene samples were centrifuged at 5000 rpm for 15 min. Then, the obtained cake was washed with n-heptane several times until the color of the asphaltenes became shiny black. Finally, the obtained asphaltenes were homogenized and fined using a pestle and mortar and left to dry at 25 °C in a vacuum oven for 12 h. The model solutions for the batch adsorption experiments were prepared by dissolving a desired amount of the obtained asphaltenes in toluene. All samples were prepared from a stock solution containing 2000 mg/L asphaltenes diluted to different concentrations by addition of toluene. The initial concentration of asphaltene solutions used in the adsorption experiments ranged from 100 to 1500 mg/L. 2.2.2. In-House-Prepared Nanoparticles. Commercial fumed silica particles purchased from Sigma-Aldrich were used as a support. Fumed silica nanoparticles were dried at 120 °C and posteriorly impregnated with an aqueous solutions of nickel nitrate Ni(NO3)2 and/or Pd(NO3)2 at different mass percentages (based on the SCMD) for 3 h and then further dried at 120 °C for 6 h. The obtained solid was calcined at 450 °C for 6 h to obtain the hybrid nanoparticle support.31The aqueous solution used for impregnation containing an X wt % of Pd(NO3)2 and/or Y wt % of Ni(NO3)2 quantity was used for the synthesis of SNiYPdX nanoparticles on the silica support with the incipient wetness technique.8It should be noted here that the Ni and Pd precursors are hygroscopic salts. These materials become oxides after calcination. Table 1 lists the specifications and surface area characteristics of nanoparticles used in this study. The hybrid nanomaterials obtained in this study are called supported hygroscopic salt (SHS) and denoted by the initial letter of the support followed by the symbol of the cation of the resulting metal oxide after calcination and the weight percentage of the aqueous solutions of nickel nitrate Ni(NO3)2 and/or Pd(NO3)2 used for impregnation. For instance, SNi1Pd1 denotes a SHS synthesized by using fumed nanosilica as support and containing 1 wt % of Ni(NO3)2 and 1 wt % Pd(NO3)2, which would produce, after calcination, a nanosilica with NiO and PdO nanoparticles on its surface. 2.2.3. Surface Area and Crystallite Size Measurements. The surface areas (SBET) of the prepared nanoparticles were measured following the Brunauer−Emmett−Teller (BET) method.32,33 This was achieved by performing nitrogen adsorption−desorption at 196 °C, using an Autosorb-1 from Quantacrome. The samples were degassed B

dx.doi.org/10.1021/ef4018543 | Energy Fuels XXXX, XXX, XXX−XXX

Energy & Fuels

Article

2.4. Modeling. The adsorption isotherms for the asphaltenes on different nanoparticles were modeled using two commonly used adsorption models, the Langmuir and Freundlich models.36,37 The adsorption kinetics were modeled using pseudo-first-order38 and pseudo-second-order models.29 For the kinetic oxidation of the asphaltenes on the nanoparticles, the Ozawa−Flyn−Wall (OFW) method was employed to describe the reaction kinetics and estimate the effective activation energies.39 2.4.1. The Langmuir Model. The Langmuir model has been widely used to correlate experimental data on equilibrium adsorption.36 This model assumes that monolayer adsorption occurs on a homogeneous surface. It was originally derived from the kinetic data of adsorption and desorption, taking into account the fact that equilibrium is obtained when the rates of adsorption and desorption are equivalent. The Langmuir equation can be expressed as follows:

Table 1. Estimated Values of NiO and PdO Nanoparticle Diameter and Surface Area of Selected Nanoparticles material

SBET ± 0.01 m2/g

dp-NiO ± 0.2 nm

dp-PdO ± 0.2 nm

S SNi2 SPd2 SNi1 SPd1 SNi1Pd1 SNi0.66Pd0.66 SNi0.29Pd1.32

389.00 233.63 205.15 290.31 265.56 201.50 220.79 211.01

− 2.9 − 1.9

− − 4.1 − 2.4 2.2 1.2 3.1

1.3 0.9 0.7

⎛ KLC E ⎞ Nads = Nads,max ⎜ ⎟ ⎝ 1 + KLC E ⎠

at 140 °C under N2 flow overnight before analysis. Surface areas were calculated using the BET equation. The size of the nanoparticles was determined using an X́ Pert PRO MPD X-ray diffractometer (PANalytical, Almelo, Netherlands), with Cu Kα radiation operating at 60 kV and 40 mA with a θ/2θ goniometer. The mean crystallite size of the particles (dp‑s: supported nanoparticle diameter) was obtained by applying the Scherrer equation to the main diffraction peak. 2.2.4. Equilibrium Adsorption Isotherms. Batch adsorption experiments were carried out in a set of 10 mL vials by mixing 100 mg nanoparticles with 10 mL of the prepared heavy oil solution containing a known concentration of asphaltenes at 25 °C, 35 °C, 45 °C, and 55 °C. Subsequently, the mixtures were agitated at 200 rpm by placing the vials in a temperature-controlled incubator and were allowed to equilibrate for 24 h because that was the amount of time needed to ensure equilibrium.19,26 The mixture was separated by allowing the nanoparticles containing the adsorbed asphaltenes to precipitate and by decanting the supernatant. The residual asphaltene concentration in the supernatant was measured using UV−vis spectrophotometry34,35 with a Genesys 10S (Thermo Scientific, Waltham, MA) spectrophotometer at 410 nm. A calibration curve of the UV−vis absorbance at 410 nm against the asphaltene concentration was established using standard model solutions with known concentrations prepared inhouse. UV−vis spectra of the asphaltene solutions were selected based on the absorption linearity range (absorbance < 2.0).35 For high asphaltene concentrations (>250 mg/L), the solutions were diluted with toluene to the desired absorbance value, and the actual concentration was estimated by multiplying the obtained concentration by the dilution factor. The amount of adsorbed asphaltenes (mg of asphaltenes/m2 surface area of nanoparticles) was determined using the mass balance in eq 1:

Q=

Co − C E V A

(2)

where Nads is the amount of asphaltenes adsorbed onto the nanoparticles (mg/m2), CE is the equilibrium concentration of asphaltenes in the solution (mg/L), KL is the Langmuir equilibrium adsorption constant related to the affinity of binding sites (L/mg), and Nads,max is defined as the monolayer saturation capacity, representing the maximum amount of asphaltenes adsorbed per unit surface area of nanoparticles for complete monolayer coverage (mg/m2). 2.4.2. The Freundlich Model. Freundlich proposed an empirical expression to represent the isothermal variation of the adsorption of a quantity of mass adsorbed by unit surface area of solid adsorbent at equilibrium concentration:37 Nads = KFC E1/ n

(3)

where KF is the Freundlich constant related to the adsorption capacity ((mg/m2)(L/mg)1/n), and 1/n is the adsorption intensity factor (unitless). It should be noted here that the Langmuir and Freundlich constants were estimated from the slopes and intercepts of the best-fit line of the linear forms of eqs 2 and 3.27 2.4.3. Kinetic Model. To predict the rate at which asphaltenes adsorb onto solid surface, two adsorption kinetic models, pseudo-firstorder and pseudo-second-order, have been used. The pseudo-firstorder kinetic model is expressed as:38 dNads = k1(Nads,eq − Nads) dt 2

(4) 2

where Nads,eq (mg/m ) and Nads (mg/m ) are the amounts of adsorbed asphaltene on solid surface at equilibrium and at time t, respectively, and k1 (min−1) is the kinetic constant of pseudo-first-order adsorption. The model is expressed in a linear form to plot the experimental data, to obtain the slope and intercept and determine Nads,eq and k1. Pseudosecond-order kinetic model is expressed as:31

(1)

where Co is the initial concentration of asphaltenes in solution (mg/L), CE is the equilibrium concentration of asphaltenes in the supernatant (mg/L), V is the solution volume (L), and A is the dry surface area of nanoparticles (m2). 2.3. Thermogravimetric Analysis of Asphaltenes. The thermogravimetric analyses were carried out with a TGA/DSC analyzer (Q500, TA Instruments, Inc., New Castle, DE) by heating the samples (nanoparticles containing adsorbed asphaltenes) in air from 30 °C to 700 °C at four different rates (5, 10, 15, and 20 °C/ min). The sample mass was kept low to circumvent the diffusion limitations (ca. 5 mg). The airflow rate was a constant 100 cm3/min throughout the experiment. Fresh silica nanoparticles were also heated to 700 °C to obtain a complete mass loss and heat change profile. Before the samples were transferred for TGA analysis, they were dried overnight in a vacuum oven. The nanoparticles selected for TGA analysis have the same amount of asphaltenes/surface area (ca. 0.2 mg/m2). Further, before the TGA experiments, the instrument was calibrated for mass and heat changes as well as temperature readings, using nickel as a reference material. Each run was repeated at least twice to confirm the producibility of the experiment, which was achieved successfully.

dNads = k 2(Nads,eq − Nads)2 dt

(5) −1

The second-order kinetic constant k2 (min ) and the amount of asphaltenes adsorbed at the equilibrium Nads,eq (mg/m2) are obtained from the slope and intercept of the experimental data plot, fitted to the linear form of the model. 2.4.4. Thermodynamic Properties. To characterize the adsorption phenomenon and process spontaneity, thermodynamic properties were calculated. They were estimated from the classical expression for the standard Gibbs free energy change ΔG0ads:40 0 ΔGads = − RT ln K

(6)

where R is the universal constant for ideal gases, T is the absolute temperature, and K is the adsorption equilibrium constant. K is related to the energy of adsorption and can be expressed as KLCs,8,26 where Cs is the solvent molar concentration. The changes of the sorption enthalpy (ΔH0ads) and entropy (ΔS0ads) were calculated from the plot of ΔG0ads against temperature. C

dx.doi.org/10.1021/ef4018543 | Energy Fuels XXXX, XXX, XXX−XXX

Energy & Fuels

Article

Figure 1. Simplex−centroid mixture design with fumed silica (S), palladium oxide (Pd), and nickel oxide (Ni). 0 0 0 ΔGads = ΔHads + T ΔSads

(7)

log(β) = δ − 0.4567

The thermodynamic properties were calculated, taking into account the extremes of the typical range of asphaltene molecular weights reported in the literature (750−5000 g/mol).41,42 This is because asphaltenes’ molecular weight property is still an unsolved issue in the literature, and it depends on the chemical nature of the crude oil from which asphaltenes were isolated and the intrinsic complexity of the asphaltene structure.8,27 2.4.5. Ozawa−Flynn−Wall (OFW) Model. The catalytic activity of the nanoparticles toward asphaltene oxidation was confirmed by estimating the effective activation energy using the isoconversional method developed by OFW.39,43 This technique assumes that the reaction rate at a given degree of conversion is only a function of the state and temperature.44 Accordingly, the equation for the rate of the oxidation reaction used in this kinetic study of oxidizing virgin asphaltenes and asphaltenes adsorbed on catalytic nanoparticles can be expressed as follows:

⎛ E ⎞ dα = Kα exp⎜− α ⎟f (α) ⎝ RT ⎠ dt

g (α) =

∫0

dα = f (α)

∫0

T

(10)

⎛K E ⎞ δ = log⎜ α α ⎟ − 2.315 ⎝ Rg (α) ⎠

(11)

At different heating rates and constant degrees of conversion, a linear relationship is observed when plotting log(β) against 1/T. The activation energy is obtained from the slope of the best-fit linear function. 2.5. Simplex−Centroid Mixture Design. Simplex−centroid mixture design (SCMD), performed with the STATGRAPHICS Centurion XVI software (StatPoint Technologies Inc.), was used to determine the optimal mixture of palladium, nickel, and fumed silica nanoparticles to minimize the oxidation temperature of the asphaltenes. Normally, SCMD is used to study the relationships between the proportions of different variables and responses.46 The fractions of each component must meet the following constraints:47 q

∑ xi = x1 + x2 + ... + xq = 1,

(8)

i=1

where Kα is the pre-exponential factor (1/s), Eα is the activation energy (kJ/mol), R is the ideal gas constant (J/mol·K), t is the reaction time (s), T is the reaction temperature (K), and α is the reaction conversion. The reaction conversion is equal to (m0 − mt)/(m0 − mf), where m0, mt, and mf are the initial mass of the sample, the current mass of the sample at temperature and time t, and the final mass of the sample at a temperature where the mass loss remains unchanged, respectively. The values of the α range between 0 and 1.0. Replacing the heating rate β = dT/dt in eq 8 and integrating, yields:39 α

Eα RT

Kα exp(− Eα /RT ) dT β

xi ≥ 1

(12)

where xi and q are the proportions of each component and the number of components in the mixture, respectively. In this study, q was 3 when x1 = S, x2 = Pd, and x3 = Ni. The amount of Pd and/or Ni oxides was restricted up to 2% in the design. Consequently, the limits of each compound are the following:

0.98 ≤ S ≤ 1.00

(13)

0 ≤ Pd ≤ 0.02

(14)

0 ≤ Ni ≤ 0.02

(15)

Using a three-component SCMD evaluates seven points at different concentrations of each component to adjust the response variable for the design and the desired model. In this case, each component was evaluated at the maximum allowable fraction at an intermediate and the centroid points of the mixture, as displayed in Figure 1. The regression model for the oxidation temperature was established using a special cubic regression fitting. The regression model equations were as follows:

(9)

Evaluating the activation energy from eq 9 clearly depends on the approximation used to estimate the integral on the right-hand side of the equation. For the OFW model, the Doyle approximation44,45 was employed: D

dx.doi.org/10.1021/ef4018543 | Energy Fuels XXXX, XXX, XXX−XXX

Energy & Fuels

Article

q

Nj =

∑ βi xi′ + ∑ βijxi′x′j ∑ i=1

xi′ =

xi − Li 1−L

i SNi0.66Pd0.66 > SNi1 > S. The data for the equilibrium adsorption isotherms describing asphaltene behavior with the S and SHS samples at 25 °C were fit to the Langmuir and Freundlich models. Table 2 summarizes the estimated Langmuir and Freundlich parameters for S and SHS. As indicated by the regression coefficients, the experimental data were adequately described by the Freundlich model, suggesting that the asphaltenes were adsorbed onto a heterogeneous surface via multilayer adsorption. This result was expected because the Langmuir model assumes that there is a homogeneous surface, while the synthesized hybrid nanomaterial has no such surface. Additionally, during the sorption experiments, even at the highest concentration tested, the isotherms did not display a plateau (maximum) in the adsorbed uptake. The highest values for the Langmuir equilibrium adsorption constant (KL) and the monolayer capacity (Nads,max) were obtained when the S and SHS samples for the entire range of asphaltene concentration were tested because of the limitations of the model for hybrid materials; for the specific cases studied, the isotherms did not plateau, a behavior untypical for Type I isotherms.8,28 However, for the SNi0.66Pd0.66 nanoparticles, the experimental data fit the Langmuir model better than the Freundlich model, indicating there was monolayer adsorption and indicating that asphaltene adsorption was surface specific.31 3.3. Adsorption Kinetics. Figure 3 displays the results obtained during the kinetics studies for Capella asphaltene adsorption onto fumed silica and SHS samples at 25 °C. Figure 3 reports the amount of adsorbed asphaltenes relative to surface area as a function of time for an initial asphaltene concentration of 500 mg/L. For every nanoparticle sample, the adsorption increased rapidly during the first 10 min and remained unchanged after 50 or 60 min of contact, indicating that equilibrium was nearly reached within 10 min. This fast asphaltene adsorption might be caused by the intermolecular forces between the most polar asphaltene components (mostly functional groups and heteroatoms) and the NiO and/or PdO on the silica surface, as well as the high dispersion of the NiO and/or PdO on the surface of the hybrid nanomaterials. The time required for the samples to reach the adsorption equilibrium occurred in the following order: SPd1 ≈ SNi0.66Pd0.66> S ≈ SNi1 ≈ SNi2> SPd2 ≈ SNi1Pd1. It was observed that SNi1Pd1 and SPd2 nanoparticles have faster adsorption kinetics and higher asphaltene uptake than the other nanoparticles considered in this study. However, for the

(16) (17)

where Nj is the oxidation temperature for a desired amount of asphaltenes adsorbed onto the adsorbent surface, βi are the coefficients of the linear terms of the components, βij are the components of the binary mixtures of the nonadditive components, and βijk is the component of the nonadditive ternary mixture. Equation 17 is a pseudocomponent of xi (fraction) and is used because of the restrictions mentioned in eqs 13−15. In these equations, Li is the lower limit of each component, and L is the sum of the lower boundaries.

3. RESULTS AND DISCUSSION 3.1. Surface Characterization. Table 1 lists the BET surface areas and crystallite sizes of the studied nanoparticles (both commercially available and those prepared in-house). Each type of nanoparticle exhibited a different chemical nature. The fumed silica had a distinct size (7 nm) and BET surface area (389 m2/g). As expected, for the hybrid materials (SHS), the surface area of the samples decreased as the metallic oxide content of the nanoparticles increased. Therefore, depositing nickel and/or palladium oxide nanoparticles blocked some of the pore spaces on the silica support, reducing its original surface area. In addition, SHS materials preserve the characteristic size of the support (fumed silica, 7 nm). 3.2. Asphaltene Adsorption Isotherms. Figure 2 shows the asphaltene adsorption isotherms at 25 °C for the S and SHS

Figure 2. Asphaltene adsorption isotherms onto different surfaces of nanoparticles at 25 °C. Adsorbent dose, 10 g/L; shaking rate, 200 rpm. The symbols are experimental data, and the solid lines are from the Freundlich model (eq 3).

samples. SHS clearly adsorbed more asphaltenes than S across the entire range of asphaltene concentrations; the difference was more noticeable at higher concentrations (CE > 200 mg/ L). The sorption isotherms for every sample revealed an increase in the adsorbed asphaltene content as the asphaltene concentration increased. These results are in excellent agreement with those reported by Cortés et al.,28 Franco et al.,8,9 Giraldo et al.,10 Nassar et al.,18,19,26,27 and Dudásǒ vá et al.48 for the adsorption of different asphaltene types onto different solid surfaces. E

dx.doi.org/10.1021/ef4018543 | Energy Fuels XXXX, XXX, XXX−XXX

Energy & Fuels

Article

Table 2. Estimated Parameters of Freundlich and Langmuir Models Freundlich model

Langmuir model

materials

KF ± 0.001 (mg/m2) (L/mg)1/n

1/n ± 0.001

R2

Nads,max ± 0.004 (mg/m2)

KL ± 0.008 (L/mg)

R2

S SNi2 SPd2 SNi1 SPd1 SNi1Pd1 SNi0.66Pd0.66

0.0061 0.0287 0.0324 0.0227 0.0335 0.1223 0.0105

0.5684 0.4590 0.4735 0.4913 0.4319 0.3020 0.6065

0.99 0.99 0.99 0.99 0.99 0.95 0.98

0.3569 0.5698 0.7724 0.6071 0.4742 0.52495 0.7083

0.0027 0.0077 0.0061 0.0056 0.0134 0.1652 0.0032

0.98 0.96 0.97 0.97 0.97 0.81 0.99

nanoparticles, the adsorption equilibrium was reached at the same time for the support, indicating that there was not any synergic effect between NiO and fumed silica nanoparticles for this case. To investigate the kinetic mechanism that controls the adsorption process, the experimental data presented in Figure 3 were analyzed using Lagergren’s pseudo-first- and pseudosecond-order models, presented in eqs 4 and 5, respectively. Table 3 summarizes the kinetic parameters (Nads,eq, k1, and k2) for both models. According to the values of the regression coefficients, the experimental data for Capella asphaltene adsorption onto S and SHS samples were more adequately fit by a pseudo-second-order model. The kinetic efficiency represented by the adsorption rate value (k2) decreased along with the initial asphaltene concentration, agreeing with the data reported by Nassar,16 Cortés et al.,28 and Franco et al.8 The kinetic efficiency for the S and SHS samples initially at 500 mg/ L occurred in the following order: SNi1 > SPd2 > SNi1Pd1 > SNi2 > SNi0.66Pd0.66 > S ≈ SPd1. However, when the initial concentration was 1000 mg/L, the order was SNi1 >SNi1Pd1 ≈ SPd2> S > SPd1 > SNi0.66Pd0.66 ≈ SNi2. Finally, when the initial concentration was 1500 mg/L, the order was S > SNi1> SNi1Pd1≈ SPd2> SPd1 > SNi0.66Pd0.66 ≈ SNi2. Notably, as

Figure 3. Adsorption kinetics for S and SHS samples for an asphaltene initial concentration of 500 mg/L and temperature of 25 °C. Adsorbent dose, 10 g/L; shaking rate, 200 rpm.

bimetallic system with 0.66 wt % of both Pd and Ni, the adsorption equilibrium was reached slower than for even the support. The same effect was observed with SPd1, indicating that by adding less than 1 wt % of Pd without the presence of the Ni, the time to reach the adsorption equilibrium is increased. By adding 1 and 2 wt % of Ni to the fumed silica

Table 3. Estimated Parameters for the Pseudo-First- and Pseudo-Second-Order Models pseudo-first-order material S

SNi1

SPd1

SNi1Pd1

SNi0.66Pd0.66

SPd2

SNi2

pseudo-second-order

Ci (mg/L)

Nads,eq exptl ± 0.01 (mg/ m2 )

Nads,eq ± 0.01 (mg/ m2)

k1 ± 0.0003 (min‑1)

R2

Nads,eq ± 0.01 (mg/ m2)

k2 ± 0.0003 (min‑1)

R2

500 1000 1500 500 1000 1500 500 1000 1500 500 1000 1500 500 1000 1500 500 1000 1500 500 1000 1500

0.10 0.16 0.23 0.19 0.33 0.45 0.17 0.31 0.43 0.25 0.45 0.65 0.18 0.34 0.44 0.22 0.39 0.54 0.19 0.34 0.46

0.02 0.03 0.04 0.02 0.06 0.07 0.04 0.05 0.09 0.05 0.07 0.13 0.04 0.10 0.18 0.08 0.08 0.12 0.04 0.14 0.23

0.0334 0.0293 0.0284 0.0926 0.0211 0.0115 0.0307 0.0285 0.0238 0.0810 0.0639 0.0348 0.0455 0.0194 0.0370 0.1293 0.0521 0.0370 0.0425 0.0330 0.0295

0.98 0.92 0.98 0.92 0.94 0.83 0.92 0.94 0.83 0.96 0.96 0.98 0.98 0.93 0.92 0.96 0.95 0.94 0.99 0.99 0.96

0.10 0.17 0.23 0.19 0.33 0.46 0.18 0.32 0.44 0.25 0.46 0.67 0.18 0.36 0.56 0.22 0.40 0.56 0.19 0.37 0.50

0.0110 0.0070 0.0063 0.0687 0.0105 0.0049 0.0056 0.0033 0.0020 0.0315 0.0087 0.0027 0.0111 0.0016 0.0012 0.0343 0.0084 0.0026 0.0123 0.0017 0.0011

0.99 0.99 0.99 0.99 0.99 0.99 0.99 0.99 0.99 0.99 0.99 0.99 0.99 0.99 0.99 0.99 0.99 0.99 0.99 0.99 0.99

F

dx.doi.org/10.1021/ef4018543 | Energy Fuels XXXX, XXX, XXX−XXX

Energy & Fuels

Article

Table 4. Calculated Values of ΔG0ads, ΔH0ads, and ΔS0ads for Asphaltene Adsorption on the S and SHS

the initial concentration of asphaltenes increased, the kinetic efficiency of S increased. 3.4. Thermodynamics Studies of Asphaltene Sorption. Thermodynamic studies are important for understanding the temperature effects on the adsorption of asphaltenes on different nanoparticle surfaces. Figure 4a and 4b shows the

material S SNi1 SPd1 SNi1Pd1 SNi0.66Pd0.66 SPd2 SNi2

asphaltene molar mass (g/mol)

−ΔG0ads (kJ/mol)

−ΔH0ads (kJ/mol)

ΔS0ads (J/mol K)

R2

5000 750 5000 750 5000 750 5000 750 5000 750 5000 750 5000 750

29.22 24.51 30.85 26.15 32.68 27.98 36.45 31.74 29.53 24.83 31.84 27.14 31.73 27.03

22.86 22.86 23.58 23.58 24.21 24.21 23.85 23.85 23.49 23.49 25.72 25.72 25.02 25.02

21.28 5.50 24.37 8.60 28.44 12.66 42.26 26.48 42.26 26.48 20.55 4.78 22.45 6.68

0.99 0.99 0.99 0.99 0.99 0.99 0.99 0.99 0.99 0.99 0.99 0.99 0.99 0.99

thermogravimetric conditions. The oxidation tests were carried out in an air atmosphere for a specific asphaltene loading in every sample (ca. 0.20 mg/m2), and each sample had a minimal mass to preclude any mass-transfer limitations.22,25 Figure 5a−c reveals (a) the rate of mass loss as a function of temperature for the virgin Capella asphaltenes and asphaltenes in presence of fumed silica, (b) the plot of heat flow that describes the heat changes that occurred during oxidation, and (c) the conversion (α). Figure 5a reveals that there is no significant mass change in the asphaltenes below 400 °C. However, between 400 °C and 470 °C, the asphaltenes underwent thermal cracking, and above 560 °C, the oxidation to form gaseous products was complete. These results agree with those of Nassar et al.,18 who studied the rate of mass loss in asphaltenes obtained from the Athabasca bitumen. According to Figure 5b, the mass loss is not exothermic until 470 °C, when the complete oxidation to gaseous products by the presence of air occurs, as revealed by the similarity between the curves in Figure 5a and 5b for the virgin asphaltenes in this region. However, for the asphaltene oxidation with fumed silica nanoparticles, the oxidation temperature decreased to 355 °C, indicating that the nanoparticles catalyze asphaltene oxidation. Figure 5b illustrates an exothermic behavior similar to that of the virgin asphaltenes except that it occurs at a temperature near the oxidation temperature of the asphaltenes with S, confirming the catalytic role of S. Figure 5c confirms the catalytic effect of the silica nanoparticles. The oxidation is enhanced by lowering the temperature at which the process starts. Without S, oxidation begins at approximately 400 °C, while the system with the silica nanoparticles initiates oxidation at 300 °C. Clearly, the nanoparticles increased the reaction rate and the conversion. For example, at 400 °C, the oxidation process began for virgin asphaltenes without nanoparticles, with a conversion of approximately 5−6%; however, the conversion is approximately 63% for samples with S nanoparticles run at the same temperature. 3.5.1. Asphaltene Oxidation in the Presence of Monometallic SHS. For practical reasons, the full range of temperatures used for the thermogravimetric analysis will be divided into three regions for all SHSs. The first region corresponds to the low-temperature region (LTR) between 200 °C and 250 °C. The second falls between 251 °C and 450 °C

Figure 4. Asphaltenes adsorption isotherms onto (a) S and (b) SNi1Pd1 at 25 °C, 35 °C, 45 °C, and 55 °C. Adsorbent dose, 10 g/L; shaking rate, 200 rpm.

adsorption isotherms for asphaltenes onto nanoparticles at 25 °C, 35 °C, 45 °C, and 55 °C for S and SNi1Pd1. For both S and SNi1Pd1, the amount of adsorbed asphaltenes decreases as the temperature increases because asphaltene adsorption onto nanoparticle surfaces is an exothermic process, while temperature affects the colloidal state of the asphaltenes.28,38,49 For the remaining samples, the same behavior was observed. Table 4 lists the calculated thermodynamic parameters for the sorption of asphaltenes with “different molecular weights” onto S and SHS samples. The negative ΔG0ads values reveal the feasibility and spontaneity of the adsorption processes; specifically, the adsorption process is spontaneous and thermodynamically favorable. The negative values of the standard enthalpy change (ΔH0ads) confirmed the exothermic nature of the sorption process. The positive values for the standard entropy change (ΔS0ads) were caused by the increased randomness at the solid−liquid interface during the sorption process. These findings agree with those reported by Nassar et al.,27 Cortés et al.,28 and Franco et al.8 for the adsorption of different asphaltenes onto different nanoparticle surfaces. 3.5. Catalytic Oxidation of Asphaltenes. 3.5.1. Virgin Asphaltenes and Asphaltenes in the Presence of S. Catalytic thermal oxidation of asphaltenes was performed to obtain insight into the effects of the selected nanoparticles under G

dx.doi.org/10.1021/ef4018543 | Energy Fuels XXXX, XXX, XXX−XXX

Energy & Fuels

Article

Figure 6. (a) Plot of rate of mass loss as a function of temperature and (b) conversion for asphaltenes in the presence of monometallic SHS, asphaltenes adsorbed, 0.2 mg/m2; heating rate = 10 °C/min; airflow, 100 cm3/min.

approximately 355 °C, there is another significant mass loss caused by the oxidation of larger chains. Finally, in the HTR region, the heaviest compounds are oxidized. The functionalization of silica nanoparticles with 2 wt % Pd leads to a higher mass loss in the LTR region, while materials with 1 wt % Pd exhibit the opposite effect. Figure 4b displays the conversions for SPd1 and SPd2. The asphaltene oxidation is initiated at essentially the same temperature for both samples (220 °C), but when the temperature increases, the conversion is higher with SPd2. For example, at 400 °C, the conversions are 43 and 58 wt % for SPd1 and SPd2, respectively. Further, the presence of Ni on the silica surface does not generate a peak in the LTR, indicating that SPd1 and SPd2 are better catalysts than the nanoparticles functionalized with only nickel oxide. Only two peaks are observed for SNi1 and SNi2. The peaks in the MTR region fall at 316 °C and 347 °C and are the largest mass losses for SNi1 and SNi2, respectively. In the HTR, the final asphaltene oxidation step occurs approximately at 480 °C in both samples. According to Figure 6b, below 300 °C (before oxidation begins), the conversions exhibited by both samples functionalized with Ni are quite similar. However, above 300 °C, the conversion for SNi2 is almost 5% higher than that for SNi1, indicating that, similar to the Pd-loaded nanoparticles, increasing the metal oxide content on the silica surface enhances the catalytic activity. The conversion with SNi2 is higher than the conversion with SPd2 above 300 °C; this result might occur because, between 200 °C and 300 °C, the SPd2 oxidized fraction of the asphaltenes was adsorbed, while SNi2 had not yet begin its oxidation process. 3.5.2. Asphaltene Oxidation in the Presence of Bimetallic SHS. Two bimetallic samples were used to evaluate the effects

Figure 5. (a) Plot of rate of mass loss as a function of temperature; (b) DTA-plot heat flow as a function of temperature; (c) conversion of virgin Capella asphaltenes and asphaltenes in the presence of fumed silica. Amount of asphaltenes adsorbed = 0.2 mg/m2; heating rate = 10 °C/min; air flow = 100 cm3/min.

and is called the midtemperature region (MTR). The last region covers temperatures from 451 °C to 700 °C and is the high-temperature region (HTR). Because asphaltenes are not pure compounds, they exhibit different behaviors during thermal decomposition and oxidation because compounds with lower molecular weight oxidize at lower temperatures. Figure 6a and 6b reveals (a) a plot of the rate of mass loss relative to temperature and (b) the conversion for asphaltenes in the presence of monometallic SHS for a specific amount of adsorbed asphaltene (ca. 0.2 mg/m2). As observed, the samples containing palladium oxide exhibit a peak in the lowtemperature range (Figure 6a), suggesting that there is a significant change in the sample mass due to asphaltene oxidation. Asphaltene oxidation begins at 244 °C and 215 °C for SPd1 and SPd2, respectively. This temperature reduction reveals that increasing the Pd content on the silica surface enhances the catalytic activity. In the MTR region at H

dx.doi.org/10.1021/ef4018543 | Energy Fuels XXXX, XXX, XXX−XXX

Energy & Fuels

Article

SNi0.66Pd0.66 is 4% higher than for SPd2, while the conversion with SNi1Pd1 is 94% and therefore 38% higher than for SPd2. The asphaltenes are nonpure components; therefore, NiO and PdO could have different selectivity for different compounds in the asphaltenes. When the fumed silica nanoparticles are functionalized with just NiO or just PdO, the catalytic reaction will be based on the compound of preference. However, when the fumed silica nanoparticles are functionalized with both PdO and NiO, the combination of individual constituents enhances the catalytic activity. As reported by Wu et al.,50 hybrid structures with noble metals such as palladium and non-noble metals such as nickel allow better surface electronegativity, leading to more efficient conversion in catalytic processes. 3.5.3. Minimization of Oxidation Temperature of Asphaltenes from SCMD. The temperature where the oxidation process begins for each sample was used as a reference when performing the minimization. The statistical analysis of the results is generated using the STATGRAPHICS Centurion XVI software. Table 5 lists the values of β for the cubic special model of the oxidation temperature as a function of the pseudocomponents S′, Pd′, and Ni′. The special cubic model validates the experimental data; R2 = 1.0 and suitably predicts the temperature at which the asphaltene oxidation begins over any amount of S, Ni, and Pd. The optimal concentration of Ni and Pd on the silica surface needed to minimize the asphaltene oxidation temperature is 0.29% Ni and 1.32% Pd (SNi0.29Pd1.32). Figure 2 depicts the sorption isotherm obtained for SNi0.29Pd1.32 (blue line). Below 50 mg/L, the adsorption capacity of the material exceeds that of SNi1, SNi0.66Pd0.66, and S. However, above 50 mg/L, only SPd2 and SNi1Pd1 perform better than the optimized sample. Figure 8 presents the conversion and a plot for the rate of mass

of using bimetallic catalysts during asphaltene oxidation. Figure 7a and 7b presents (a) a plot of the rate of mass loss versus

Figure 7. (a) Plot of rate of mass loss as a function of temperature and (b) conversion for asphaltenes in the presence of bimetallic SHS, asphaltenes adsorbed, 0.2 mg/m2; heating rate = 10 °C/min; airflow, 100 cm3/min.

temperature and (b) the conversion for the asphaltene oxidation with bimetallic SHS materials for the same amount of asphaltenes used with the monometallic SHSs. Figure 7a reveals that both samples exhibited three peaks, similar to SPd1 and SPd2. The difference is that SNi1Pd1 has one peak in the MTR (255 °C) that is very close to the boundary for the LTR. In addition, the peak in the HTR does not represent as significant of a mass loss as the other two. A similar observation can be noted for the SNi0.66Pd0.66 samples. Three peaks are observed, with one in each region. The more significant mass losses in this case are revealed at 219 °C, 347 °C, and 517 °C for the LTR, MTR, and HTR, respectively, with the first being the largest. Clearly, the Pd and Ni exert a synergistic effect during asphaltene oxidation because having both metals present enhanced the asphaltene oxidation in the LTR at the expense of the MTR mass loss. Another indication of the synergistic effects exhibited by the bimetallic SHS is revealed by the conversion analysis. As observed in Figure 7b, the conversion under these samples begins near 200 °C and is higher with SNi1Pd1. When making the same comparison as above, at 400 °C, the conversion with

Figure 8. Conversion and plot of rate of mass loss as a function of temperature (secondary axis) with an asphaltene loading of 0.2 mg/m2 onto SNi0.29Pd1.32; heating rate = 10 °C/min; airflow, 100 cm3/min.

loss versus temperature for SNi0.29Pd1.32 with an asphaltene loading of 0.2 mg/m2. Figure 8 reveals that the optimized

Table 5. Calculated Parameters of the Special Cubic Model for the Capella Asphaltene Temperature of Oxidation in the Presence of S and SHS β1

β2

β3

β12

β13

β23

β123

R2

354.91

213.73

308.37

−275.88

−60.28

−184.20

−423.41

1.0

I

dx.doi.org/10.1021/ef4018543 | Energy Fuels XXXX, XXX, XXX−XXX

Energy & Fuels

Article

Figure 9b displays conversion for asphaltenes with SPd2 at the three tested loadings. As expected, when the amount of asphaltenes decreases, the conversion increases because a larger amount of active sites are available for the reaction.23 3.5.5. Activation Energy for the Oxidation of Asphaltenes in the Presence and Absence of Nanoparticles. The activation energies (Eα) required to oxidize the virgin Capella asphaltenes, as well as those in the presence of the S and SHS nanoparticles, were calculated using the thermal analysis data with the OFW method;43 the results are presented in Figure 10. Different

sample also displays three peaks with one in each region, similar to the other bimetallic samples. However, the peak in the MTR falls between the peaks of the other bimetallic SHSs. With regard to conversion, the shape of the curve is similar to that for SNi0.66Pd0.66. Therefore, when there is a lower NiO and PdO content on the surface of S, results can be obtained that are similar to those for SNi1Pd1. 3.5.4. Effect of Asphaltene Loading on Asphaltene Oxidation. SPd2 displays three peaks for asphaltene oxidation (Figure 6) when the asphaltene loading is 0.2 mg/m2; this sample was used to study the effect of asphaltene loading on the oxidation temperature. Figure 9a depicts a plot of the mass

Figure 10. Activation energies evaluated for thermal cracking of Capella virgin asphaltenes and Capella asphaltenes in the presence of S and SHS by the OFW method.

nanomaterials clearly utilized different mechanisms. Notably, while the Eα values for the virgin asphaltenes decreased as the conversion increased, the opposite was observed for S and the SHS samples. Because the OFW method evaluates the activation energy for different heating rates at a constant degree of conversion, large variations in the Eα force the OFW method to average the values in the profile up to the given conversion. Consequently, as the Eα decreases with α, the OFW method underestimates the activation energy.43 Additionally, the difference in the activation energy trends for the virgin and functionalized asphaltenes might occur because the virgin asphaltenes undergo thermal cracking in one “homogeneous” step; there is no discrimination based on the aggregate size of the asphaltenes. However, the adsorption process onto the nanoparticle surfaces from the oil model solution might first adsorb the smaller asphaltene aggregates that reach the nanoparticles via diffusive processes before the larger aggregates are adsorbed in a new layer. Therefore, the catalysts must react with the asphaltenes in the initial layer (i.e., asphaltenes with lower molecular weight) first before interacting with the asphaltenes in other layers. Figure 10 indicates that the Eα for asphaltene oxidation lowered dramatically in the presence of S and SHS, favoring the catalytic activity. Eα is lower for the SHS samples than for the S materials. For conversion between 0% and 40%, the activation energy decreased in the following order, confirming the synergetic effect of the bimetallic SHS: S > SNi2 > SPd1 > SNi0.66Pd0.66 > SPd2 > SNi1Pd1. However, after α = 40%, the activation energy of SNi2 decreases below that for SPd1 because, before 40% conversion is reached, SPd1 oxidized the smaller asphaltene aggregates but afterward required more energy to react with the larger aggregates, while SNi2 required higher Eα values to start the process; this

Figure 9. (a) Plot of rate of mass loss as a function of temperature with an asphaltene loading of 0.2, 0.05, and 0.03 mg/m2 and plot enthalpy changes (red secondary axis) as a function of temperature for an asphaltene loading of 0.2 and 0.03 mg/m2, and (b) conversion for asphaltenes in the presence of SPd2 with an asphaltene loading of 0.2, 0.05, and 0.03 mg/m2; heating rate = 10 °C/min; airflow, 100 cm3/ min.

loss rate versus temperature for SPd2 with asphaltene loadings of 0.03, 0.05, and 0.20 mg/m2 and a plot of the enthalpy changes for SPd2 with 0.2 mg of asphaltenes per m2. The three samples clearly exhibit three peaks, with one at each temperature range. However, as the asphaltene loading decreases, the peak in the LTR grows larger while the other two tend to form plateaus, indicating that the oxidation process begins at the LTR and the mass loss remains constant until the oxidation is completed. The plots of the enthalpy changes confirm that, for the 0.2 mg/m2 sample, there are three exothermic peaks associated with asphaltene oxidation, while the 0.03 mg/m2 sample displays one defined peak in the LTR and a plateau associated with a constant exothermic character. J

dx.doi.org/10.1021/ef4018543 | Energy Fuels XXXX, XXX, XXX−XXX

Energy & Fuels

Article

(6) Nikooee, E.; Sayyad Amin, J.; Ayatollahi, S.; Alamdari, A.; Jahromi, H.; Escrochi, M. Proceedings of the 1st International Petroleum Conference & Exhibition, Shiraz, Iran, EAGE, 2009. (7) Leontaritis, K. J.; Amaefule, J. O.; Charles, R. E. SPE Prod. Facil. 1994, 8, 157−64. (8) Franco, C.; Patiño, E.; Benjumea, P.; Ruiz, M. A.; Cortes, F. B. Fuel 2012, 105, 408−414. (9) Franco, C.; Nassar, N. N.; Ruiz, M. A.; Pereira, P.; Cortés, F. B. Energy Fuels 2013, 27, 2899−2907. (10) Giraldo, J.; Nassar, N. N.; Benjumea, P.; Pereira, R.; Cortés, F. B. Energy Fuels 2013, 27, 2908−2914. (11) Zahedi, G.; Fazlali, A. R.; Hosseini, S. M.; Pazuki, G. R.; Sheinkhattar, L. J. Petrol. Sci. Eng. 2009, 68, 218−222. (12) Snow, R. H. 2011 SPE Heavy Oil Conference and Exhibition, Kuwait City, Kuwait, Dec 12−14, 2011; SPE: Richardson, TX, 2011. (13) Cavallaro, A. N.; Galliano, G. R.; Moore, R. G.; Mehta, S. A.; Ursenbach, M. G.; Zalewski, E.; Pereira, P. J. Can. Petrol. Technol. 2008, 47, 23−31. (14) Weissman, J. G.; Kessler, R. V.; Sawicki, R. A.; Belgrave, J. D. M.; Laureshen, C. J.; Mehta, S. A.; Moore, R. G.; Ursenbach, M. G. Energy Fuels 1996, 10, 883−889. (15) Pereira, P.; Larter, S. Forum 04: Unconventional petroleum resources. Presented at 19th World Petroleum Congress, Madrid, Spain, June 29−July 3, 2008. (16) Hassan, A.; Lopez, F.; Nassar, N. N.; Carbognani, L.; Pereira, P. Catal. Today 2013, 207, 112−118. (17) Hashemi, R.; Nassar, N. N.; Pereira Almao, P. Energy Fuels 2013, 27, 2194−2201. (18) Nassar, N. N.; Hassan, A.; Carbognani, L.; Lopez-Linares, F.; Pereira-Almao, P. Fuel 2012, 95, 257−262. (19) Nassar, N. N.; Hassan, A.; Pereira-Almao, P. J. Therm. Anal. Calorim. 2012, 110, 1327−1332. (20) Hashemi, R.; Nassar, N. N.; Pereira-Almao, P. Energy Fuels 2012, 26, 1645−1655. (21) Nassar, N. N.; Husein, M.; Pereira-Almao, P. Oil Gas Rev. 2011, 9, 46−48. (22) Nassar, N. N.; Hassan, A.; Pereira-Almao, P. Energy Fuels 2011, 25, 3961−3965. (23) Nassar, N. N.; Hassan, A.; Pereira-Almao, P. Colloids Surf., A 2011, 384, 145−149. (24) Nassar, N. N.; Hassan, A.; Pereira-Almao, P. J. Colloid Interface Sci. 2011, 360, 233−238. (25) Nassar, N. N.; Hassan, A.; Pereira-Almao, P. Energy Fuels 2011, 25, 1566−1570. (26) Nassar, N. N.; Hassan, A.; Pereira-Almao, P. Energy Fuels 2011, 25, 1017−1023. (27) Nassar, N. N. Energy Fuels 2010, 24, 4116−4122. (28) Cortés, F. B.; Mejía, J. M.; Ruiz, M. A.; Benjumea, P.; Riffel, D. B. Energy Fuels 2012, 26, 1725−1730. (29) Mohammadi, M.; Akbari, M.; Fakhroueian, Z.; Bahramian, A.; Azin, R.; Arya, S. Energy Fuels 2011, 25, 3150−3156. (30) Pernyeszi, T.; Patzkó, A.; Berkesi, O.; Dékány, I. Colloids Surf., A 1998, 137, 373−384. (31) Cortes, F. B.; Chejne, F.; Carrasco-Marín, F.; Pérez-Cadenas, A.; Moreno-Castilla, C. Energy Convers. Manage. 2012, 53, 219−223. (32) Rouquerol, F.; Rouquerol, J.; Sing, K. S. W. Adsorption by powders and porous solids: principles, methodology and applications; Academic Press: London, U.K., 1999. (33) Brunauer, S.; Emmett, P. H.; Teller, E. J. Am. Chem. Soc. 1938, 60, 309−19. (34) Mullins, O. C. Optical interrogation of aromatic moieties in crude oil sand asphaltenes. In Structures and dynamics of asphaltenes; Mullins, O. C.; Sheu, E. Y., Eds.; Plenum Press: New York, 1998. (35) Goncalves, S.; Castillo, J.; Fernández, A.; Hung, J. Fuel 2004, 83, 1823−1828. (36) Langmuir, I. J. Am. Chem. Soc. 1916, 38, 2221−2295. (37) Freundlich, H. M. F. Z. Phys. Chem. (Leipzig) 1906, 57, 385− 470.

energy was employed to oxidize both small and medium aggregates. For the optimized sample (SNi0.29Pd1.32), the activation energy exhibited a trend similar to that of SPd1, indicating that the amount of PdO dominates the catalytic activity of the sample. However, after 55% conversion, the activation energy of the sample falls below that of SPd1, indicating that the energy employed earlier cracked a greater number of large chains.

4. CONCLUSIONS Fumed silica nanoparticles were functionalized with nickel and palladium oxides and successfully employed for the adsorption/ oxidation of asphaltenes extracted from a Colombian extraheavy crude oil sample. In addition, the asphaltene sorption isotherms for S and SHS were determined and fitted to the Freundlich and Langmuir models; a better fit was found for the former model. The SNi1Pd1 material demonstrated a better affinity for the asphaltenes than the other samples. Additionally, the adsorption kinetics revealed that SNi1Pd1 and SPd2 reached adsorption equilibrium faster than the other samples. The thermodynamic properties of the asphaltene adsorption onto S and SHS confirmed the spontaneous and exothermic nature of the asphaltene adsorption process. For the TGA experiments, functionalizing the silica nanoparticles lowers the asphaltene oxidation temperature as well as the activation energy; better results were achieved using the bimetallic samples rather than the monometallic ones. A simplex−centroid mixture design of three components, in addition to seven runs, was used to minimize the asphaltene oxidation temperature. Although a 0.29% Ni and 1.32% Pd composition was obtained, its minimized temperature was close to that for the SNi1Pd1, SNi0.66Pd0.66, and SPd2 materials. The estimated activation energies of asphaltene oxidation in the presence and absence of nanoparticles by the OFW method confirmed that different nanoparticles utilized different reaction mechanisms.

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors acknowledge COLCIENCIAS, Petroraza SAS, and Universidad Nacional de Colombia for logistical and financial support.

■

REFERENCES

(1) Energy Information Administration. World consumption of primary energy by energy type and selected country groups, 1980−2006; 2008; http://www.eia.doe.gov/pub/international/iealf/table18.xls (accessed on July 9, 2013). (2) Cámara Colombiana de Servicios Petroleros − CAMPETROL. The Production of Different Types of Crude Oil in Colombia; http:// www.oilproduction.net/cms/files/estadisticaspcp.pdf (accessed on July 10, 2013). (3) Mullins, O. C. Annu. Rev. Anal. Chem. 2011, 4, 393−418. (4) Sayyad Amin, J.; Nikooee, E.; Ghatee, M. H.; Ayatollahi, Sh.; Alamdari, A.; Sedghamiz, T. Appl. Surf. Sci. 2011, 257, 8341−8349. (5) Groenzin, H.; Mullins, O. C. J. Phys. Chem. A 1999, 103, 11237− 45. K

dx.doi.org/10.1021/ef4018543 | Energy Fuels XXXX, XXX, XXX−XXX

Energy & Fuels

Article

(38) Alboudwarej, H.; Pole, D.; Svrcek, W. Y.; Yarranton, H. W. Ind. Eng. Chem. Res. 2005, 44, 5585−5592. (39) Ozawa, T. Bull. Chem. Soc. Jpn. 1965, 38, 1881−1886. (40) Smith, J. M.; VanNess, H. C.; Abbott, M. M. Introduction to Chemical Engineering Thermodynamics; McGraw-Hill: New York, 2005. (41) Rudrake, A.; Karan, K.; Horton, J. H. J. Colloid Interface Sci. 2009, 332, 22−31. (42) Groenzin, H.; Mullins, O. C. Energy Fuels 2000, 14, 677−684. (43) Flynn, J. H.; Wall, L. A. J. Polym. Sci. Pol. Lett. 1966, 4, 323−328. (44) Doyle, C. D. J. Appl. Polym. Sci. 1961, 5, 285−292. (45) Doyle, C. D. Nature 1965, 207, 290−291. (46) Scheffe, H. J. R. Stat. Soc., B 1963, 25, 235−263. (47) Cornell, J. A. Experiments with Mixtures. Designs, Models, and the Analysis of Mixture Data, 3rd ed.; John Wiley & Sons, Inc.: New York, 2002. (48) Dudásǒ vá, D.; Simon, S.; Hemmingsen, P.; Sjöblom, J. Colloids Surf., A 2008, 317, 1−9. (49) Yarranton, H. W.; Alboudwarej, H.; Jakher, R. Ind. Eng. Chem. Res. 2000, 39, 2916−2924. (50) Wu, Y.; Wang, D.; Xhao, P.; Niu, Z.; Peng, Q.; Li, Y. Inorg. Chem. 2011, 50, 2046−2048.

L

dx.doi.org/10.1021/ef4018543 | Energy Fuels XXXX, XXX, XXX−XXX