Effect of Textural Properties and Surface Chemical Nature of Silica Nanoparticles from Different Silicon Sources on the Viscosity Reduction of Heavy Crude Oil.

The main objective of this study is to evaluate the effect of the textural properties and surface chemical nature of silica nanoparticles obtained from different synthesis routes and silicon precursors, on their interactions with asphaltenes and further viscosity reduction of heavy crude oil (HO). Four different SiO2 nanoparticles were used, namely, commercial fumed silica nanoparticles (CSNs) and three in-house-synthesized nanoparticles (named based on the silicon source) modifying the silicon precursor: sodium silicate (SNSS), tetraethylorthosilicate (TEOS) (SNT), and rice husk (SNRH). The nanomaterials were characterized through dynamic light scattering (DLS), transmission electron microscopy (TEM), Fourier transform infrared (FTIR) spectroscopy, N2 physisorption (S BET), atomic force microscopy (AFM), and X-ray photoelectron (XP) spectroscopy (XPS). The adsorption of asphaltenes over the different nanoparticles was evaluated at a concentration of 1000 mg·L-1 in toluene. The asphaltene-nanoparticle interactions are closely related to several textural properties, such as roughness, surface area, and hydrodynamic diameter, as well as the surface chemical nature of the materials. The results in the textural characterization exhibited that the sizes of the nanoparticles from TEM ranged between 6.9 and 11.5 nm. Nevertheless, the standard deviation of the measurements showed that the sizes are statistically similar. Inversely, the hydrodynamic diameter changed, affecting the surface silanol group's availability due to a hindering effect on functional groups as the hydrodynamic size of the material increased. The rheological measurements were performed at a fixed nanoparticle dosage of 1000 mg·L-1 and showed that the trend of the degree of viscosity reduction (DVR) was CSN > SNT > SNSS > SNRH with the highest value yielding at 30%. The results of DVR are in accordance with the nanoparticles' adsorptive capacity as higher values were obtained with the material that leads to a higher amount of adsorbed asphaltenes. Also, the oxygen amount related to silanol groups, estimated by the XPS analysis, showed a direct relation regarding adsorption capacity and further HO viscosity reduction.

Introduction

Nowadays, most of the worldwide petroleum reserves are composed of heavy oil (HO) and extra-heavy oil (EHO).[1−3] Although there is a wide availability of this type of reserves, the vast majority of these remain underexploited due to the costs and technical challenges associated with their production and transportation.[4,5] The issues presented in these operations are related to the HO and EHO physicochemical properties, particularly the viscosity, which hinders the crude oil mobility at surface and subsurface conditions.[6−10] It is well known that this reduced mobility is derived mainly from the asphaltene contents[11,12] and their interactions with the other chemical compounds in the crude oil, which lead to the formation of complex microstructures.[13−19] Thus, different types of nanomaterials have been developed for capturing asphaltenes through an adsorption phenomenon, inhibiting the aggregation of these molecules. Different nanoparticles have been evaluated including alumina,[20−23] Fe3O4,[24−29] silica,[30−35] functionalized materials,[22,36−39] and other oxides,[20,40−43] which can modify the viscoelastic networks of HO and EHO and hence reduce their viscosity.[29,44]

Silica nanoparticles are among the most used nanomaterials for these purposes due to their high affinity toward asphaltene adsorption.[35] Cortés et al.[36] evaluated SiO2 nanoparticles as support for nickel oxide nanocrystals for improving their adsorption. The high affinity of these composite materials for asphaltenes leads to the reduction of asphaltene aggregate sizes, which in EHO and HO causes a reconfiguration of the microstructure, decreasing the viscosity, and, thus, improving the mobility.[44,45] Taborda et al.[44] evaluated the effect of nanoparticles and nanofluids on the HO rheology and mobility, finding that the performance of viscosity reduction is directly related to the nanoparticles’ adsorptive capacity. The authors found that the best results are obtained using acidized silica nanoparticles with a degree of viscosity reduction (DVR) of 90% with dosages between 1000 and 10000 mg·L–1.

In further studies, Taborda et al.[46,47] found that the maximum HO viscosity reduction for the evaluated silica nanoparticles was reached with a concentration near 1000 mg·L–1, with a DVR of roughly 52%. Also, Taborda et al.[45] studied the mechanisms leading to HO viscosity reduction through dynamic rheological tests, establishing that the crude oil microstructure was altered upon nanoparticle addition, which reduced the elastic contributions of the viscoelastic fluid after being subjected to stress, suggesting a disruption in the asphaltene viscoelastic network. The positive effects of silica nanoparticles on the crude oil rheological properties and thus in their mobility have also been evaluated for alternated benefits. Montes et al.[48] evaluated the diluent/dispersant ratio for carrier fluid designing and its synergistic process alongside silica nanoparticles. The authors found that the viscosity reduction perdurability would be enhanced with the nanoparticulated system with an increase in viscosity of just 20% after 30 days compared with an increment of 80% after 10 days in the absence of nanoparticles.

Due to the high performance of silica nanoparticles in capturing asphaltenes on their surface, some authors have studied the effects of different properties on the adsorption phenomenon according to their nature in terms of the textural properties such as surface area, particle size, and surface chemical composition.[30,33,35,36,49−51] The different values of surface areas with specific active site availability could substantially change the effect of the sorbent and its impact on viscosity reduction. For Poole et al.[50] and Doner,[49] the asphaltene surface attractive forces of silica nanoparticles depend on properties such as low particle diameter, high surface area, and available active sites, specifically Si–OH (silanol) groups on the surface of nanoparticles, which have a great selectivity for polar molecules.[30] The presence of this type of structure generates acid–base interactions with the most polar compounds in HO N-containing functional groups such as pyridine, quinolone, and pyrrole species.[52−54]

Furthermore, recent studies[30] aimed at determining the role of particle size and surface acidity in the asphaltene–silica nanoparticle interactions with nanomaterials ranging from 11 to 240 nm. It was found that the lowest nanoparticle size (S11) generated the highest asphaltene adsorption, which is related to a higher surface area availability. Moreover, the surface of the S11 nanoparticles was modified by acidizing (S11A) and basifying (S11B) treatments, finding that the best performance was obtained with the S11A material and is narrowly related to the generation of higher amounts of hydroxyl (silanol) groups. Other authors have also found that the surface availability of silanol groups has a significant role in asphaltene–silica interactions.[22,36] It is worth mentioning that the surface chemical nature and the textural properties are the keys to the asphaltene adsorption phenomenon, which in HO causes changes in its microstructure leading to viscosity reduction.

However, to the best of our knowledge, there are no studies evaluating the effects of different textural properties and surface chemical nature of silica nanoparticles on the viscosity reduction of HO and EHO based on different silicon sources and synthesis routes. Therefore, the main objective of this study is to determine the effect of the textural properties and surface chemical nature of silica nanoparticles obtained from different sources and synthesis routes on their interactions with asphaltenes and HO viscosity reduction. Thereby, distinct synthesis routes and silicon precursors for obtaining silica nanoparticles were used, ensuring a similar particle size and varying the roughness, hydrodynamic diameter, and surface area, resulting in a change of the surface silanol concentration and availability. It is expected that this work will generate a wider landscape about the role of nanotechnology to guarantee a cost-effective and better exploitation of heavy and extra-heavy oils.

Results and Discussion

The results are divided into four sections, namely: (i) nanoparticle characterization including the analysis of textural and surface chemical nature, (ii) rheological behavior and asphaltene adsorption evaluation, (iii) assessment of the effect of textural properties on viscosity reduction, and (iv) effect of surface chemical composition of silica nanoparticles on HO viscosity reduction.

Nanoparticle Characterization

Textural Property Assessment

Figure shows the textural characteristics of the four silica nanoparticles in terms of particle size and visual roundness assessed through the high-resolution transmission electron microscopy (HR-TEM) technique. It can be seen in Figure a,b that the silica nanoparticles synthesized from rice husk (SNRH) and silica nanoparticles synthesized from sodium silicate (SNSS) have an amorphous agglomerate texture as Musić et al.[55] and Ghorbani et al.[56] stated in their respective studies. In contrast, commercial silica nanoparticle (CSN) and silica nanoparticles synthesized from tetraethylorthosilicate (TEOS) (SNT) present a spherical and nonagglomerate shape as Stöber et al.[57] suggested. Additionally, the nanoparticles’ mean particle size was calculated through HR-TEM by analyzing the images using ImageJ—Fiji software.[58] In brief, the images that were considered as a representative bulk sample were selected, and a negative photographic image was generated. The contrast and brightness were adjusted to improve the image quality, and finally, after selecting the color target, the software reproduces the nanoparticles’ size histogram. The obtained results are shown in Table and Figure alongside the standard deviation of the measured sizes. The results indicate that silica nanoparticles are restrained in the range from 6.9 to 11.5 nm approximately, and as observed from the standard deviation, the sample sizes are statistically similar, as it is also shown in the box plot in Figure .[59] The aforementioned fact was also validated with the skewness and kurtosis values, which are shown in Table S1 of the Supporting information. It can be seen that the values for both parameters tend to 0, and it can be established with sufficient confidence that the particle size distribution is Gaussian, and thus the standard deviation alone is adequate for stating that all of the distributions are statistically similar.

Figure 1

High-resolution transmission electron microscopy (HR-TEM) for (a) silica nanoparticles synthesized from rice husk (SNRH), (b) silica nanoparticles synthesized from sodium silicate (SNSS), (c) commercial silica nanoparticles (CSNs), and (d) silica nanoparticles synthesized from TEOS (SNT) at the 100 nm scale.

Table 1

Textural Properties of the Evaluated Silica Nanoparticles Synthesized from Rice Husk (SNRH), Silica Nanoparticles Synthesized from Sodium Silicate (SNSS), Commercial Silica Nanoparticles (CSNs), and Silica Nanoparticles Synthesized from TEOS (SNT), Obtained through High-Resolution Transmission Electron Microscopy (HR-TEM), Dynamic Light Scattering (DLS), Brunauer–Emmett–Teller (BET) Surface Area (SBET), and Atomic Force Microscopy (AFM)

particle	diffusivity in aqueous suspension (m2·s–1)	mean hydrodynamic diameter, DLS (nm)	mean particle diameter, HR-TEM (nm)	SBET (m2·g–1)	surface roughness, AFM (nm)
SNRH	1.0 × 10–11 ± 0.1 × 10–11	45 ± 4	11.5 ± 4	166	1.18 ± 1
SNSS	5 × 10–11 ± 1 × 10–11	12 ± 2	6.9 ± 4	382	3.17 ± 1
SNT	6 × 10–11 ± 1 × 10–11	11 ± 2	10.7 ± 3	210	8.42 ± 1
CSN	7 × 10–11 ± 2 × 10–11	9 ± 2	8.3 ± 2	344	6.86 ± 1

Figure 2

Box plot for nanoparticles measured by TEM. The full boxes represent the interval of 25–75% of the data, while the hollow boxes account for the mean.

Also, in Table are presented the BET surface area (SBET) and mean particle size (dp50) results obtained by N2 physisorption and DLS analysis, respectively. It is observed from Table that SNSS and CSN have the highest SBET values, while the medium and low surface areas represent SNRH and SNT, respectively. The four samples of silica evaluated are nonporous materials based on the SBET values reported in Table . Moreover, the SNRH sample has a higher hydrodynamic diameter compared to the one reported with HR-TEM, which is mainly due to a higher aggregation of the nanoparticles in the aqueous suspension used. This phenomenon affects the hydrodynamic diameter of the nanomaterials according to the Stokes–Einstein equation (eq ), which is used for DLS equipment for the indirect measurement of the hydrodynamic diameter. In this sense, when the hydrodynamic diameter increases for the nanoparticle cluster, the movement of nanoparticles is reduced and then diffusivity also decreases.

Equation is the Stokes–Einstein equation,[60] used for determining the hydrodynamic diameter of nanoparticles through DLS measurements, where KB (1.98 × 10–23 m2·kg·s–2·K–1) is the Boltzmann constant, T (K) is temperature, η (cP) is the viscosity of the medium, and D (m2·s–1) is the nanoparticle diffusion coefficient. It is observed that as the diffusivity of the particles in the suspension decreases, the hydrodynamic size increases (Table )

Moreover, as the HR-TEM images just provide visual insights on the nanoparticles’ roundness or sphericity, the roughness of the material was evaluated through AFM analysis, and its results are also shown in Table . It is observed from these results that the roughness trend of the nanomaterials is SNT > CSN > SNSS > SNRH. An example of the analyzed micrographs is observed in Figure , where the noncontact micrograph is visible, as well as the high profiles and three-dimensional (3D) micrographs. Additional information on the nanoparticles’ textural properties regarding N2 adsorption isotherms and pore volume can be found in Figure S1 and Table S2 of the Supporting Information, respectively.

Figure 3

AFM micrograph example of silica nanoparticles synthesized from rice husk (SNRH). (a) Noncontact AFM micrograph, (b) high profiles extracted from the top micrograph, and (c) 3D AFM micrograph.

Functional Groups and Composition

The surface chemical composition of the silica nanoparticles has an important effect on the nanoparticle–asphaltenes interactions,[30] which lead to viscosity reduction.[44] Specifically, siloxane and silanol groups are of primary importance for governing the adsorption process. A brief scheme of the commonly found silanol and other Si–O functional groups in the surface of the nanoparticles is shown in Figure , where Figure b–d are the structures corresponding to silanol groups, while Figure a represents siloxane groups. To study the effect of the chemical nature of nanoparticles on HO viscosity reduction, the nanomaterial surfaces were properly characterized through Fourier transform infrared (FTIR) and X-ray photoelectron spectroscopy (XPS) for determining composition changes and quantifying the presence of both functional groups’ species.

Figure 4

Surface groups of silica nanoparticles: (a) siloxane, (b) isolated silanol, (c) geminal silanols, and (d) associated silanols. Reproduced with permission from Poole and Riffenburgh.[61]

Figure shows the FTIR spectra obtained for the evaluated nanoparticles. From Figure , the typical vibrations of the siloxane and silanol groups can be observed in all of the nanomaterials. The bands at around 750 and 810 cm–1 represent the Si–O bond flection, and the wide band between 850 and 960 cm–1 corresponds to the O–Si–O stretching vibrations. Moreover, the adjacently encountered band between 1000 and 1300 cm–1 denotes the same bonds’ asymmetric stretching.[62,63] On the other hand, the presence of silanol groups is related to the band centered between 2500 and 3800 cm–1, which is produced by the O–H bond vibrations. The band centered at 1600 cm–1 shows hydroxyl scissoring.[64] It can be also seen from Figure that the bulk intensity of common silanol bands follows the order SNSS > SNRH > CST > SNT. However, as the FTIR spectra correspond to bulk measurements, it is not enough for determining the tested nanomaterial’s surface activity regarding the attractive forces between these and asphaltene aggregates. In this sense, a more refined surface analysis was carried out using the XPS technique.

Figure 5

FTIR spectra for commercial silica nanoparticles (CSNs) and silica nanoparticles synthesized from different sources: rice husk (SNRH), sodium silicate (SNSS), and TEOS (SNT).

Figure shows the XP spectra, which were carried out for identifying the surface chemical nature of the four nanoparticles evaluated. Figure A represents the survey spectra, in which it is observed the presence of distinct elements in each nanoparticle surface. The binding energy (BE) of carbon 1 (C 1s) was taken as a reference value to perform spectra corrections, and the signals were decomposed using Gaussian/Lorentzian combination functions. In Figure B, the O 1s high-resolution spectra are shown, which schematizes the total surface oxygen in the nanomaterials and its different chemical surroundings. The green area corresponds to the silanol-group-related oxygen (SO, at around 533.7 eV of BE), while the blue one denotes the siloxane-group-related oxygen (≈532.7 eV). It is important to clarify that Figure A shows peaks associated with the presence of C 1s carbon, which is generated by adventitious carbon and does not resemble the surface composition of the nanoparticles.

Figure 6

XP spectra for commercial silica nanoparticles (CSNs) and silica nanoparticles synthesized from different sources: rice husk (SNRH), sodium silicate (SNSS), and TEOS (SNT). (A) Survey spectra and (B) decomposed O 1s high-resolution spectra of the different nanomaterials, where the signal in blue denotes the siloxane-related oxygen (≈532.7 eV), while the green one is associated with the silanols (≈533.7 eV).

It is also observed that SNRH has sodium and sulfur presence at the surface, which indicates the presence of sodium sulfate salt from the synthesis route. The SNSS spectrum presents Cls and sodium characteristic peaks, indicating the presence of sodium silicate in the sample and chloride ions from sodium chloride used in its respective synthesis process. It is also shown in Figure B that the curved area of the silanol on the surface of the nanomaterials has the trend CSN > SNT > SNSS > SNRH. Presumably, the asphaltene–nanoparticle interaction forces would have a similar tendency, as in the silanol groups, and the O–H bond formed in this type of structure generates a delocalization of the charge toward oxygen, due to the high difference of electronegativity of these atoms, causing a greater interaction between these groups with the polar components of the crude oil.[65,66]

Table summarizes the quantified amount of the different elements conforming each nanoparticle surface as atomic percentages. The total surface oxygen by percentage has the trend SNSS > SNT > SNRH > CSN. However, it is worth mentioning that these quantities are not related to surface SO enrichment, as different synthesis processes could lead to distinct water chemisorption degrees on the surface of the nanoparticles, which is narrowly related to the presence of SO.[67] A better approximation could be obtained by calculating the O/Si ratios for each type of nanomaterial. Nonetheless, there would be some overestimations related to Si–O–Si structure existence. Hence, further analyses were carried out by quantifying the normalized total oxygen conforming the surface of the nanoparticles and calculating the SO contribution from Figure B.

Table 2

XPS Analysis and Quantification for Commercial Silica Nanoparticles (CSNs) and Silica Nanoparticles Synthesized from Different Sources: Rice Husk (SNRH), Sodium Silicate (SNSS), and TEOS (SNT)

material	atom-electron orbital	surface composition (atom %)	binding energy (eV)	%O/%Si
SNRH	O 1s	54.77	530.70	1.48
	Si 2p	37.11	102.70
	Na 1s	0.72	1070.00
	S 2p	7.41	169.70
SNSS	O 1s	51.58	532.80	1.25
	Si 2p	41.45	103.87
	Na 1s	1.12	1071.87
	Cl 2p	5.85	200.87
CSN	O 1s	57.81	532.50	1.37
	Si 2p	42.19	103.50
SNT	O 1s	53.82	532.93	1.16
	Si 2p	46.18	103.93

The SO amount for each evaluated nanoparticle calculated from the normalized total oxygen content in its surface is shown in Figure . It is seen that unlike the previous analyses, the higher SO contribution was obtained for the CSN, with a value yielding 100% of the total oxygen presented in the nanoparticle surface, while the SNT, SNSS, and SNRH have contributions of 90, 82, and 53%, respectively. This last analysis would suggest that the CSN has a higher surface concentration of asphaltene–nanoparticle interaction points, which may result in higher amounts of adsorbed asphaltenes that generate further viscosity reduction.[45,46] Nonetheless, the silanol active sites available on the surface of the nanoparticles could change according to the nanomaterials’ textural properties such as surface area and hydrodynamic diameter.

Figure 7

Silanol-group-related oxygen (SO) amount for commercial silica nanoparticles (CSNs) and silica nanoparticles synthesized from different sources: rice husk (SNRH), sodium silicate (SNSS), and TEOS (SNT) calculated from the normalized total oxygen content on their surface.

Figure shows the influence of the nanoparticles’ textural properties on the presence of SO on their surface in terms of their (A) surface area and (B) the mean hydrodynamic diameter (dp50). No clear trend is seen when comparing the surface area to the SO, as the SNSS material, which has the highest surface area, does not have a vast majority of surface silanol groups as the CNS does. This means that the availability of this type of active sites is not necessarily related to a higher area and would not lead automatically to higher adsorption. On the contrary, a clear correlation is observed between the nanomaterials’ hydrodynamic diameter and the availability of the discussed active sites, as a decrease in the dp50 value is related to an increase in the SO. This latter fact could be associated with nanoparticle agglomeration, as it suggests an important nanoparticle–nanoparticle interaction, hindering the exposition of active sites for asphaltene–nanoparticle interactions.

Figure 8

Influence of the nanoparticles’ textural properties on the presence of silanol-group-related oxygen (SO) on their surface: (A) surface area and (B) hydrodynamic diameter (dp50). The four different nanoparticles are commercial silica nanoparticles (CSNs) and silica nanoparticles synthesized from different sources: rice husk (SNRH), sodium silicate (SNSS), and TEOS (SNT).

Rheological Behavior and Asphaltene Adsorption Evaluation

The effect of silica nanoparticles on HO rheological behavior has been widely studied in previous works,[44,46,47] where also the mechanisms governing viscosity reduction have been explored.[45] The most important conclusions regarding these studies indicate that the asphaltene adsorption over the surface of the nanoparticles leads to an aggregate size reduction,[33,44] producing a modification of the fluid microstructure, which causes a viscoelastic network disruption.[45] Moreover, although the effect of this phenomenon caused by different nanoparticle surfaces’ chemical nature has been evaluated, a proper analysis regarding the influence of nanoparticle’s identified functional groups is yet to be carried out. In this sense, Figure shows the flow curves for the HO in the absence and presence of the studied nanoparticles at a concentration of 1000 mg·L–1. The mentioned dosage is in agreement with other studies, as it is close to the optimum concentration for obtaining the best viscosity reduction performance.[46−48]Figure shows the typical pseudoplasticity of these types of fluids with a shear-thinning behavior.[11,68] Moreover, a viscosity reduction upon nanoparticle addition to the HO is observed in all of the ranges of the shear rate used for determining the samples’ rheological behavior. The highest reduction was observed when using CSN, followed by SNT, SNSS, and SNRH, which is in agreement with a higher silanol availability on the surface of the nanoparticles.

Figure 9

Rheological behavior (viscosity vs shear rate) for the heavy oil (HO) in the absence and presence of the evaluated nanoparticles at a concentration of 1000 mg·L–1 measured at 25 °C. The four different nanoparticles are commercial silica nanoparticles (CSNs) and silica nanoparticles synthesized from different sources: rice husk (SNRH), sodium silicate (SNSS), and TEOS (SNT). Error bars represent data ± standard deviation.

Further microstructural changes can be assessed by determining the effect of nanoparticles on HO pseudoplastic behavior. Rheological models such as Cross,[69] Sisko,[70] and Herschel–Bulkley,[71] among others, are commonly used for gathering this information.[72,73] These phenomenological-like models are useful for obtaining insights on rheological properties such as thixotropy and elastic responses to strain. Most of them introduce viscosity plateau terms, in which the viscosity of the examined sample has a Newtonian behavior at very low or high deformations, known as the viscosity at zero and infinite shear rates, i.e., viscosity values at which an asymptotic behavior is observed regardless of a shear rate increase or decrease. Nonetheless, these types of models have the restriction that they can be used in cases in which the experimental data is near the mentioned plateaus to increase analysis precision for providing phenomenological approaches instead of mere fitting uses. Since the experimental data does not approach this asymptotic behavior (in a log-scale plot),[74] it was decided to use a simple power-law model (eq ) for estimating the changes in the shear-thinning behavior of the HO upon nanoparticle additionwhere η (Pa·s) is the estimated viscosity in the shear-thinning region; γ ˙(s–1) is the shear rate; k is the consistency index, which has the particular unit of Pa·s; and n is the power-law index, which if it is equal to 1, denotes Newtonian behavior, while if it is <1, it represents a shear-thinning behavior.[75] The estimated values for n have the trend of HO-CSN > HO-SNT > HO-SNSS > HO-SNRH > HO, being 0.38, 0.36, 0.34, 0.29, and 0.26. These results denote that the inclusion of nanoparticles in the HO matrix modifies the shear-thinning behavior, which would be related to a microstructure modification,[39,44,48] indicating a more Newtonian behavior and possibly lesser elastic contributions to the stress response of the system.[45]

The degree of viscosity reduction (DVR) was estimated from the flow curves at a shear rate of 7 s–1, as after identifying the residual wall slip, a shear rate was chosen in which the reported viscosity would be lesser affected by this phenomenon but also not as high for not being close to an initial transportation condition via the pipeline.[9] As it was discussed above, the highest DVR was obtained when adding the CSN with a value of 30%, and the lowest value (10.3%) was observed when using SNRH. It is expected that the nanomaterials with a better performance in viscosity reduction have the highest asphaltene–nanoparticle attractive forces due to the presence of active sites conformed by silanol groups. In this sense, the DVR should be related to the amount of adsorbed asphaltenes in the surface of the nanoparticles as it has also been suggested in previous studies.[44,48]

Thus, nanoparticle adsorption was evaluated for different nanoparticles at 25 °C and a constant asphaltene concentration in toluene model solutions of 1000 mg·L–1. H bonds between Si–O–H and heteroatoms present in asphaltenes, mainly N and S, can explain the affinity between these molecules and the SiO2 surface, as well as the increase of the adsorption capacity with the increase of the silanol surface concentration. Some works have studied the interaction between asphaltenes and various types of sorbents, in which five different mechanisms were identified driving asphaltene adsorption based on the van der Waals forces, π–π stacking, cation−π stacking, acid–base interaction, coordination, and H-bonding.[51] However, the significance of coordination, π–π stacking, and H-bonding is mainly related to the self-association of asphaltene once an asphaltene monolayer has been previously adsorbed onto the sorbent’s surface.[66] Similarly, it has been demonstrated that the effect of van der Waals forces on the adsorption is higher when the asphaltene aggregates are close to the first layer of adsorbed asphaltenes.[76] Nevertheless, a different behavior has been encountered regarding the acid–base interactions as it has been determined that the initial attractive asphaltene–adsorbent force is dominated by the presence of N-containing functional groups,[77] i.e., the monolayers formed by the initially adsorbed asphaltenes contain more polar species, and the additional layers in a multilayer assembly are formed mainly through van der Waals forces, H-bonding, and π–π stacking.[78]

In this regard, sulfur has been established to have a minor role in asphaltene adsorption compared to other species with oxygen and nitrogen in their structure[65,79] and is mainly related to the fact that the nitrogen groups, which are partially positively charged, can produce acid–base interactions with the typically acidized sorbent surfaces. This behavior leads to an increase in asphaltene adsorption as the nitrogen content increases in the oil-heavy fractions.[20,40,80] This increased adsorption is derived from the explained acid–base interactions, which in silica surfaces are enhanced by the presence of hydroxyl Si–OH functional groups (silanol), as the partially positive charged nitrogen species are enabled to interact with the negatively charged silanol groups.[52−54]

Hence, the measured amount of asphaltenes adsorbed (Qads) was compared with the DVR and is shown in Figure . Here, a direct relation between viscosity reduction and the amount of adsorbed asphaltenes is observed, indicating that the highest DVR was obtained with the nanomaterials with the highest amount of silanol active sites. Therefore, it is feasible to establish a direct relationship with silanol-adsorption-viscosity reduction. Nonetheless, further analyses will be carried out in the following section for isolating the distinct effects concerning HO viscosity reduction.

Figure 10

Degree of viscosity reduction (DVR) vs the amount of adsorbed asphaltenes (Qads) obtained for each evaluated nanomaterial in model solutions of asphaltenes at 1000 mg·L–1 in toluene. The four different nanoparticles are commercial silica nanoparticles (CSNs) and silica nanoparticles synthesized from different sources: rice husk (SNRH), sodium silicate (SNSS), and TEOS (SNT). Error bars represent ±data standard deviation.

Effect of Nanoparticle Texture

Some authors have studied the effect of the physical properties of different sorbents in the asphaltene adsorption process, such as surface area,[81] pore volume and sizes,[82] and particle sizes,[83−85] concluding that these textural properties have a significant role in the asphaltene–sorbent interactions and subsequent adsorptive capacity.[26,86] In the particular case of silica nanoparticles, the evaluation of different particles with distinct surface area, size, and surface chemical nature has been reported, concluding that these textural and chemical properties are the key variables in the asphaltene–nanoparticle interactions.[30,33,49,50] Hence, it is of particular importance for the present study to isolate the possible effects of the distinct textural characteristics of the evaluated nanoparticles on asphaltene adsorption and further HO viscosity reduction.

Figure shows the effects of the nanoparticles’ (A) surface roughness and (B) surface area on the degree of viscosity reduction of heavy oil. The roughness was considered as an essential parameter to be evaluated on asphaltene–nanoparticle interactions as the adsorption of the monolayer formed by the most polar-heavy fractions could be affected due to the asphaltene molecules’ adjustment in the sorbent surface, leading to some surface area blockage produced by steric effects.[87] However, from the observed results in Figure A, there is no clear trend between the surface roughness of the nanoparticles and the DVR as the SNT deviates from the other particles. Nonetheless, it can also be said that as the roughness increases, the DVR increases too, except for a single point. However, this trend should be corroborated in further studies with more than four particles with different textural properties. Also, it has been established that the asphaltene adsorption phenomenon is enhanced with an increase in the surface area of materials with the same composition.[30,81]

Figure 11

Effect of the nanoparticles’ (A) roughness and (B) surface area on the degree of viscosity reduction (DVR) of heavy oil. The four different nanoparticles are commercial silica nanoparticles (CSNs) and silica nanoparticles synthesized from different sources: rice husk (SNRH), sodium silicate (SNSS), and TEOS (SNT). Error bars represent ±data standard deviation.

Nonetheless, as observed in Figure B, there is no evidence that a larger surface necessarily will lead to a higher amount of asphaltenes adsorbed for the nanoparticles evaluated. In this sense, it is more precise to conclude that the adsorption is more related to the type of active sites exposed in the sorbent surface. Further, the effect of the hydrodynamic diameter in viscosity reduction will be discussed, as it was found that this property has a significant effect on the amount of exposed silanol on the surface of each nanoparticle.

Although the textural properties of sorbents could play an essential role in the asphaltene–sorbent interactions, the process is primarily driven by the attractive forces between the asphaltenes and the sorbent surface.[88,89] Thereby, Figure shows the effect of the nanoparticle’s hydrodynamic diameter effect on the DVR. It can be seen from Figure that the DVR increases as the hydrodynamic diameter decreases and is related to a higher exposure of the silanol groups on the surface of the evaluated nanoparticles. A similar trend is observed as the one in Figure B, which means that a more significant silanol exposition on the surface of the nanomaterials would lead to a higher viscosity reduction. Hence, the chemical surface effect will be analyzed in the next section to corroborate this hypothesis.

Figure 12

Effect of the mean nanoparticle hydrodynamic diameter (dp50) on the degree of viscosity reduction (DVR) of heavy oil. The four different nanoparticles are commercial silica nanoparticles (CSNs) and silica nanoparticles synthesized from different sources: rice husk (SNRH), sodium silicate (SNSS), and TEOS (SNT). Error bars represent ±data standard deviation.

Effect of Nanoparticle Surface Chemical Nature

According to the results in previous sections, it has been possible to establish a direct relationship between the exposed active sites on the surface of the nanoparticles and their performance on viscosity reduction. To further demonstrate the validity of this affirmation, the effect of nanoparticle surfaces’ chemical nature on the DVR will be analyzed in this section. Figure shows the effect of (A) total surface oxygen (SO and siloxane groups) and (B) solely SO groups on the DVR. From the obtained results, it is seen that there is no correlation between the bulk amount of Si–O structures on the nanoparticle surface and the performance on viscosity reduction. However, a trend between SO presence and its effect on the obtained DVR is observed, which proves the importance of the number of exposed silanol groups on the asphaltene–nanoparticle attractive forces due to their capacity of generating acid–base interactions with the most polar-heavy fractions. These results agree with those reported by Pernyeszi and Dékány[82] and Betancur et al.,[30] where it was identified that the fundamental property that enhances the asphaltene–silica interactions is the chemical nature of the surface, particularly the silanol groups’ presence.

Figure 13

Effect of (A) total surface oxygen and (B) silanol-group-related oxygen (SO) on the degree of viscosity reduction (DVR) of heavy oil. The four different nanoparticles are commercial silica nanoparticles (CSNs) and silica nanoparticles synthesized from different sources: rice husk (SNRH), sodium silicate (SNSS), and TEOS (SNT). Error bars represent ±data standard deviation.

Conclusions

The study of the influence of distinct silica nanoparticle surfaces, namely, CSN, SNRH, SNT, and SNSS, in terms of their textural properties and chemical nature on their interactions with asphaltenes and their effect on viscosity reduction, was done correctly by assessing different characteristics such as roughness, surface area, hydrodynamic diameter, and exposed silanol groups on the surface of the nanomaterials.

It was concluded that some surface textural properties such as area and roughness, which have been conventionally considered to play an essential role in asphaltene adsorption, have minor effects on this phenomenon compared to the availability of surface active sites. The number of silanol groups on the nanomaterial surface was preliminarily considered as the driving parameter of asphaltene adsorption and heavy oil viscosity reduction, and these were quantified by estimating the surface oxygen related to these groups (SO) through the XPS technique. The SO amount was identified to follow the trend of CSN > SNT > SNSS > SNRH, and the hydrodynamic diameter was observed to affect active site availability due to hindering effects. The degree of viscosity reduction (DVR) was observed to follow the same trend, with the highest DVR being 30%, followed by 18, 14, and 10.3%, respectively. This phenomenon was explained by the measured adsorption capacity, which showed that higher availability of silanol active sites enhances the asphaltene–nanoparticle attractive forces and interactions. Finally, it was established that the type of chemical structures presented in the sorbent surfaces had an important role in the process, as no important effects on DVR were observed when analyzing the bulk or total surface oxygen representing both silanol and siloxane groups.

Experimental Section

Materials

Four silica nanoparticles were used for this study: three of them were in-house synthesized, while the fourth one was a commercial nanoparticle from Sigma-Aldrich (St. Louis, MO). The synthesis of the in-house-prepared nanoparticles had three different silicon precursors, namely, tetraethylorthosilicate (TEOS, Sigma-Aldrich, St. Louis, MO), sodium silicate (Sigma-Aldrich, St. Louis, MO), and rice husk from a local Colombian mill. The commercial silica nanoparticles and the synthesized nanoparticles from TEOS, sodium silicate, and rice husk were labeled as CSN, SNT, SNSS, and SNRH, respectively. Additionally, reagents such as ethanol (99% purity), ammonia (25% purity), chloride acid (37% purity), and sodium hydroxide (98% purity) were purchased from Merck KGaA (Darmstadt, Germany), while sodium chloride (99% purity) and sulfuric acid (98% purity) were obtained from PanReac AppliChem (Darmstadt, Germany), and finally, cetrimonium bromide (CTAB) (99% purity) was purchased from Sigma-Aldrich (St. Louis, MO) for carrying out the distinct synthesis processes. A Colombian HO of 11.2° API with saturate, aromatic, resin, and asphaltene (SARA) contents in mass fractions of 12.8, 30.4, 40.1, and 16.7%, respectively, was employed for asphaltene isolation and rheological measurements. API gravity measurement was made according to the ASTM D 1250 standard, while SARA analysis was carried out through a micro deasphalting technique coupled with thin-layer chromatography (TLC) following the IP 469 standard and using a TLC-FID/FPD Iatroscan MK6 (Iatron Labs Inc., Tokyo, Japan).

Methods

Nanoparticle Synthesis

The SNT nanoparticles were synthesized through the sol–gel method.[57] Briefly, TEOS/H2O/NH4OH/ethanol was mixed in a 1:3:0.2:1.1 volumetric proportion; the solution was stirred for 1 h and then heated at 110 °C for 5 h to remove excess water. Calcination is not needed as the polymerization process for obtaining this material is very effective and the nanoparticles precipitate right after this reaction is terminated, which is why it is feasible to obtain a clean material just by dehydration as a final synthesis step.

The synthesis of SNSS nanoparticles was carried out according to the procedure proposed by Zulfiqar et al.[90] with some modifications. Accordingly, a solution of sodium silicate and CTAB in deionized water was aged for 1 h. Then, HCl at a concentration of 2.5% (in mass fraction) was added to the solution until a pH of 3 was obtained. Afterward, the solution was centrifuged, and the precipitate was washed with deionized water to remove impurities. Finally, the obtained solid was dried at 120 °C and further calcined at 600 °C to obtain the silica nanoparticles.

The SNRH synthesis using rice husk was carried out based on the route proposed by Ghorbani et al.[56] with slight modifications. The rice husk was washed with deionized water to remove soil and dirt and further heated at 90 °C for 1 h to remove humidity. Then, a treatment with HCl 1 M was used for 2 h at 80 °C. The rice husk was further rewashed to remove heavy elements, and it was calcined at 600 °C for 6 h. The obtained ashes were treated with a basic solution of NaOH 0.5 N at 90 °C under vigorous stirring for 6 h. The solution was filtered and neutralized with H2SO4 15% to reach a pH of 8. The obtained gel was left to stand for 24 h and further centrifuged at 4500 rpm. The material was filtered and washed with deionized water to remove the sodium sulfate salts and finally dehydrated and calcined under the same temperature conditions of the SNSS to obtain the SNRH material.

Nanoparticle Characterization

The nanoparticles were texturally characterized by their surface area (SBET) through the Brunauer–Emmett–Teller method,[91] by N2 physisorption at −196 °C using an Autosorb-1 Quantachrome. High-resolution transmission electron microscopy (HR-TEM) was used for determining the nanomaterials’ mean particle size (dp50) with a Tecnai F20 Super Twin TMP (FEI Company, Oregon) with a resolution of 0.1 nm, operated at 200 kV as the accelerating voltage. The hydrodynamic diameter was obtained in an aqueous phase using a NanoPlus-3 (Micromeritics, GA). The nanomaterial roughness was assessed through atomic force microscopy (AFM) using AFM 5500 (Agilent Technologies, Chandler, AZ) equipment, and the obtained results were analyzed using Gwyddion software.[92] An aqueous drop of 10 mg·L–1 of the different silica nanoparticles was deposited over a mica surface and dried at 50 °C during one night and then measured into the AFM.

The Fourier transform infrared (FTIR) technique[63] was used to identify the functional groups in the surface of the nanoparticles with an IRAffinity-1 FTIR (Shimadzu, Japan) spectrophotometer. Also, surface groups were determined through X-ray photoelectron spectroscopy (XPS) with a PHOIBOS 150 1D-DLD (SPECS GmbH, Berlin, Germany) photoelectronic X-ray spectrometer (NAP-XPS) analyzer using monochromatic light of Al Kα (1486.7 eV, 13 kV, 100 W) with 100 and 30 eV energy steps for general and high-resolution spectra; the values for general and high-resolution spectra were 1 and 0.1 eV, respectively. The spectra were analyzed using CasaXPS software.[93]

Asphaltene Adsorption

The asphaltenes were isolated from the HO based on their solubility as it has been accounted for in other studies.[32,37] The interactions between silica nanoparticles and the asphaltenes were evaluated in terms of the amount of asphaltenes adsorbed (Qads) over the selected nanoparticles using a UV–vis Genesys 10S spectrophotometer (Thermo Scientific, Waltham, MA) at a wavelength of 295 nm. The tests were carried out following the procedure proposed by Guzmán et al.[33] with solutions at an asphaltene concentration of 1000 mg·L–1 prepared from a stock solution of 5000 mg·L–1. A fixed dosage of 10 g·L–1 of nanoparticles was used in each prepared solution for prompting their deposition upon asphaltene adsorption and providing a better supernatant analysis.

Rheological Measurements

A sample preparation protocol was implemented to guarantee proper dispersibility into HO and avoid experimental errors in the rheological measurements.[44,94] For this, the heavy crude oil was weighed, and the silica nanoparticles were added and further manually stirred for 15 min at 35 °C due to the high viscosity of the samples that make stirring by other mechanical methods unfeasible. The HO in the absence of the nanoparticles was also stirred under the same time and temperature conditions to have a similar disturbance and precondition regarding the samples with nanoparticles.[94] Later, the samples were left to rest until the sample temperature was reduced to 25 °C. A nanoparticle concentration of 1000 mg·L–1 was used as it has been demonstrated in previous studies that it is close to the optimum dosage for obtaining the highest viscosity reduction.[45,46]

The rheological behavior of the samples was evaluated using a Kinexus Pro (Malvern Instruments, Worcestershire, U.K.) rheometer equipped with a solvent trap and with a Peltier cell for controlling the temperature with a precision of 1 × 10–2 °C. The tests were carried out using a parallel-plate geometry of 20 mm with a GAP of 0.3 mm at a temperature of 25 °C with a 2 min shear rate ramp starting from 0.01 to 50 s–1. This small GAP is used to reach high shear values; thus, the misalignment of the instrument and residual wall slip was checked as these might have an impact on the measured viscosity. It was found that the viscosity is independent of GAPs larger than 0.5 mm, while it tends to decrease for lower GAPs. In this sense, it can be considered that the viscosity of the samples is underestimated mainly due to an intrinsic instrument misalignment, which also causes a wall slip phenomenon at low GAPs; however, as the main objective of this study is to compare the effect of different surface physicochemical nature of distinct nanoparticles on their interactions with HO heavy fractions and viscosity reduction, it is not necessary to adjust the viscosity values through a wall slip correction as it is considered that the instrument misalignment affects all of the samples equally.[95−97] Furthermore, 2 min was used for increasing the shear rate from 0.01 to 50 s–1 as it has been reported by other studies for preventing edge fracture and flow instabilities.[10,47] The viscosity reduction obtained by the nanoparticle usage was assessed by the degree of viscosity reduction (DVR). The DVR was calculated according to eq where μHO and μHO–SNP are the base HO viscosity in the absence and presence of the different types of silica nanoparticles, respectively.[29,44] All of the rheological measurements were carried out by triplicate to ensure the accuracy of the experiments.