Energy & Fuels 2006, 20, 1281-1286
1281
On a Probable Catalytic Interaction between Magnetite (Fe3O4) and Petroleum A. Cosultchi,*,† J. A. Ascencio-Gutie´rrez,† E. Reguera,‡,§ B. Zeifert,| and H. Yee-Madeira⊥ Instituto Mexicano del Petro´ leo, 152 Eje Central “L. Cardenas”, 07730 Mexico City, Mexico, Institute of Materials and Reagents, UniVersity of La Habana, Cuba, Center of Applied Science and AdVanced Technologies of IPN, Mexico, Department of Metallurgical Engineering, ESIQIE-IPN Mexico City, Mexico, and ESFM-IPN, Mexico City, Mexico ReceiVed September 12, 2005. ReVised Manuscript ReceiVed February 13, 2006
Magnetic and nonmagnetic iron compounds were detected as part of asphaltene deposits formed on tubing wall surface. To shed light on the probable role of the iron compounds in the formation of such deposits, magnetite (Fe3O4), one of the intrinsic components of the iron oxide multilayer scale of any carbon steel surface, was contacted with crude oil at 170 °C, a temperature similar to that of the bottom well, and subsequently aged at room temperature. Characterization of the samples was made by using XRD, Mo¨ssbauer, IR, TGA, EDS, and microscopic (SEM and TEM) techniques. Small amounts of new iron phases, magnetic (oxidized magnetite) and nonmagnetic (iron oxyhydroxides), an increase in the content of CdC and C-O bonds of the organic phase, and an increase of the thermal stability of the organic phase indicated the formation of iron complexes of Fe ions and FeOOH with the oxygen functionalities.
Introduction The C-Mn steel tubing is normally covered by a firmly adhered scale of iron oxides rich in magnetite (Fe3O4),1 as a consequence of the manufacturing process. This iron oxide contains both Fe2+ and Fe3+ ions within a compact bulk, and in the presence of the oxidizing agents, it has a low surface reactivity given by surface sites where ferrous ions are accessible.2 Magnetite is also susceptible to corrosion in wetted environments, forming iron (3+) oxyhydroxides, a process known as atmospheric corrosion.3 At the bottom-well conditions, such tubing contacts crude oil, and consequently, chemical changes other than atmospheric corrosion occur depending on the temperature, pressure, and fluid composition, as shown in previous studies.4,5 As crude oil flows up through the tubing well, the fluid pressure progressively decreases until gas phase and water, if any, separate from the liquid one. Reactive gases such as H2S, CO2, and methane, once separated from the liquid phase, also contribute to change the composition of the initial iron oxide layer of the tubing surface.1,6,7 Thus, on a tubing surface exposed to crude oil * Corresponding author. E-mail:
[email protected]. † Instituto Mexicano del Petro ´ leo. ‡ University of La Habana. § Center of Applied Science and Advanced Technologies of IPN. | ESIQIE-IPN. ⊥ ESFM-IPN. (1) Birks, N.; Meier, G. H. Introduction to High-Temperature Oxidation of Metals; Edward Arnold Publishers: London, 1983. (2) Kung, H. H. Transition Metal Oxides: Surface Chemistry and Catalysis; Elsevier Scientific: New York, 1989. (3) Craig, D. C. Practical Oilfield Metallurgy and Corrosion, 2nd ed.; PennWell Books: Tulsa, OK, 1993. (4) Cosultchi, A.; Garcı´a-Borquez, A.; Aguilar-Hernandez. J.; Reguera, E.; Yee, H.; Lara, V. H.; Bosch, P. Surf. Interface Anal. 2002, 34, 384388. (5) Cosultchi, A.; Rossbach, P.; Hernandez-Calderon, I. Surf. Interface Anal. 2003, 35, 239-245. (6) Ramanarayanan, T. A.; Smith, S. N. Corrosion 1998, 46, 66.
contact, compounds such as nonstoichiometric iron sulfides (Fe1-xS), siderite (FeCO3), maghemite, and iron oxyhydroxides (akaganeite, β-FeOOH; lepidocrocite, γ-FeOOH; and goethite, R-FeOOH) can normally be detected.6 In addition, all these compounds are concurring with organic material; organic (asphaltene) deposits have been reported as related to most of the mentioned iron phases as well as to other minerals.4,8,9 However, interaction between the oxide layer and the organic phase and the mutual effect that they can have on their respective composition and morphology remain poorly documented.10,11 It is highly important that the reactivity of iron compounds and their selectivity toward some of the crude oil polar compounds are studied in order to understand the formation of asphaltene deposits in petroleum wells and the role of tubing steel composition in such process. Thus, in this contribution, this subject is studied from samples of magnetite that have been contacted with crude oil following two steps: (a) at high temperature (170 °C) and for a short period of time (170 h) and, (b) at room temperature and for a long period of time (two years). The phase transformation of the magnetite was followed by XRD and Mo¨ssbauer spectroscopy, while the organic phase was observed by FTIR. In addition, electron microscopy techniques were used to evaluate the mutual effect of the interaction between organic and inorganic phases. (7) H2S Corrosion in Oil & Gas Production: A Compilation of Classic Papers; Tuttle, R. N., Kane, R. D., Eds.; National Association of Corrosion Engineers: Houston, 1981. (8) Thawer, R.; Nicoll, D. C. A.; Dick, G. SPE Prod. Eng. 1990, 5, 475480. (9) Branco, V. A. M.; Mansoori, G. A.; de Almeida Xavier, L. C.; Park, S. J.; Manafi, H. J. Pet. Sci. Eng. 2001, 32, 217-230. (10) Nalwaya, V.; Tantayakom, V.; Piumsomboon, P.; Folger, S. Ind. Eng. Chem. Res. 1999, 38, 964-972. (11) Gu, B.; Schmitt, J.; Chen, Z.; Liang, L.; McCarthy, J. EnViron. Sci. Technol. 1994, 28, 38-46.
10.1021/ef0502958 CCC: $33.50 © 2006 American Chemical Society Published on Web 03/16/2006
1282 Energy & Fuels, Vol. 20, No. 3, 2006
Cosultchi et al.
Experimental Procedures A crude oil sample from a Mexican well (P1), free of sediments, was contacted with magnetite powder of analytical grade (from Sigma Aldrich) with a nominal purity of 98%. The mixture was placed within a sealed cylindrical steel cell, pressurized with nitrogen under100 psi, and rolled for 170 h (one week) at 170 °C in an OFITE roller oven. Such a system is normally used in cement and drilling fluids laboratory to evaluate their dynamic behavior. The temperature used in the oven corresponds to the highest temperature estimated for a Mexican petroleum reservoir. Afterward, the solid fraction was separated from the warm oil being filtered at high pressure and under nitrogen atmosphere. Part of this solid fraction was used in the characterization process, and another part was aged in a sealed amber flask and preserved at room temperature for 20 months. At the end of the aging process, the sample was also characterized following the same procedures. These two solid samples will be identified as high-temperature process (HTP) and aging process (AP) samples, respectively. The content of carbon, hydrogen, and nitrogen in both crude oil and sediment (as the solid phase separated by centrifugation from crude oil) was determined by using an Elemental Vario EL analyzer; the sulfur content was measured in a Leco SC-444 apparatus, and the oxygen was estimated in mass difference. The metals, normally concentrated in the crude oil asphaltene fraction, were determined by Atomic Adsorption in a Perkin-Elmer AA 5000 spectrometer. A high-pressure liquid chromatograph, HP series 1100, attached with a diode array detector, was used to identify the following fractions of crude oil: paraffin, aromatic, and polar. The separation process is based on ASTM D-2007-98 standard.12,13 The crude oil, free of asphaltene, was eluted by using a sequence of cyclohexane, dichloromethane, and isopropylic alcohol as mobile phases to drag each fraction along the stationary phase consisting of two packed columns: PAC cyano-amino and dinitroaniline propyl. Several techniques were used to characterize the solid sample (magnetite before and after contact with the organic phase). The X-ray diffraction patterns were collected in a D8 Advance diffractometer (from Bruker) and monochromatic Cu KR radiation. The Mo¨ssbauer spectra were run at room temperature by using a constant acceleration spectrometer (from Wissel) operated in the transmission mode and a 57Co/Rh source. The spectra were fitted by using an iterative least-squares minimization algorithm and pseudo-Lorentzian line shapes to obtain the isomer shifts (δ), quadrupole splitting (∆), line width (Γ), hyperfine magnetic field (Hf), and relative area (A) values. The isomer shift values are reported relative to R-Fe at 298 K. IR spectra were recorded in a FTIR spectrophotometer (EQUINOX 55, from Bruker), and the examined samples were pressed in KBr disks. The thermal behavior of the HTP and AP samples was determined according to the ASTM E-1131 standard method and by using a Perkin-Elmer TGA-7 apparatus. SEM and TEM images were obtained by using JEOL JSM 6300 and JEOL JEM 100CX-II electron microscopes, respectively. For SEM analysis the samples were coated with gold, while for the TEM analysis they were dispersed in ethanol to have small enough powders to produce bright and dark field images. The EDS spectra were recorded with a NORAN Voyager spectrometer attached to the SEM microscope. This microanalysis technique was preferred in the obtainment of the local elemental analyses of both samples, HTP and AP. It is important to indicate that, normally, the atomic % of the light elements obtained from such analysis is not directly proportional to the real content of the studied samples because such elements are in the lowest range of X-ray beam excitation. (12) Altgelt, K. H.; Jewell, D. M.; Latham, D. R.; Seluky, M. L. Separation Schemes. Chromatography in Petroleum Analysis; Marcel Dekker: New York, 1979; pp 185-214. (13) Aske, N.; Kallevik, H.; Sjo¨blom, J. Energy Fuels 2001, 15, 13041312.
Figure 1. XRD powder pattern of: (a) nontreated magnetite sample, (b) HTP sample, (c) AP sample. Symbols: (O) hematite, (*) sodium chloride, (×) silicon oxide.
Results (A) XRD Results. Figure 1a shows the XRD powder pattern of the magnetite sample before contact with crude oil, which also shows minor reflections due to hematite (R-Fe2O3). The pattern of Figure 1b corresponds to an HTP fraction, where small peaks of sodium chloride (JCPDS 72-1668) and silicon oxide (JCPDS 27-605) were identified in addition to the previous iron phases. In addition, as shown in the inset of Figure 1b, the magnetite(111) peak becomes broader compared to that in the inset of Figure 1a. This deformation indicates the possibility of crystal distortion due to organic species adsorption. The same crystalline species were observed accompanying the iron phases in the AP sample XRD pattern (Figure 1c). This last sample also contains a slightly higher contribution of an amorphous species which is observed as a relatively high background around 20° (2θ) (Figure 1c) corresponding either to amorphous iron hydroxide or to the organic phase shown elsewhere.14 The appearance of NaCl and SiO2 as minor phases agrees with a previous study where Na, Cl, and Si have been found at the metal-organic deposit interface of tubing well.14 (B) Mo1 ssbauer Spectroscopy Results. Mo¨ssbauer spectra of all the studied samples are shown in Figure 2, and the calculated parameters were collected in Table 1. As expected, from Mo¨ssbauer spectra, the amount of hematite (R-Fe2O3) is approximately 7 wt %, according to its peaks area (intensity of its subspectrum). Hematite is the impurity fraction of magnetite, and its amount does not change along the experiments, suggesting that this phase remains stable or behaves as inert iron phase when contacted with crude oil. An iron oxyhydroxide, a new iron phase, was identified in spectra of both samples, HTP and AP. This assignment was carried out based on the estimated Mo¨ssbauer parameters (Table 2). A conclusive differentiation between R-FeOOH, β-FeOOH, and γ-FeOOH is not possible (14) Cosultchi, A.; Zeifert, B.; Cordova, I.; Valenzuela, M. A.; Bosch, P.; Lara, V. H. Fuel 2002, 81, 413-421.
Catalytic Interaction between Magnetite and Petroleum
Energy & Fuels, Vol. 20, No. 3, 2006 1283
Figure 2. Mo¨ssbauer spectra at room temperature of: (a) nontreated magnetite sample, (b) HTP sample, (c) AP sample. Table 1. Mo1 ssbauer Parameters at Room Temperature of Fe3O4 Sample before and after Contacted Crude Oil sample
δa (mm/s)
Bhf (T)
∆ (mm/s)
A (%)
assignment
pure magnetite
0.25 0.64 0.53 0.24 0.67 0.52 0.35 0.24 0.67 0.52 0.35
48.9 46.2 51.3 48.9 46.2 51.2 48.9 46.2 51.2 -
-0.16 -0.01 -0.8 -0.12 -0.01 -0.8 0.68 -0.12 -0.01 -0.8 0.68
32 61 7 33 59 7 1 32 58 7 3
magnetite (A) magnetite (B) hematite magnetite (A) magnetite (B) hematite γ-FeOOH magnetite (A) magnetite (B) hematite γ-FeOOH
HTP sample AP sample
a Isomer shift (δ) values relatives to R-Fe. The errors for δ and ∆ are less than 0.01 mm/s and for Bhf are less than 0.1 T.
Table 2. Elemental Composition of P1 Crude Oil, Sediment, and Asphaltene Fraction, wt % element
crude oil
sediment
asphaltene
carbon hydrogen nitrogen oxygen sulfur
84.66 12.98 0.05 1.13 1.17
76.92 12.17 0.12 9.24 1.55
81.5 7.55 0.64 5.03 5.28
from Mo¨ssbauer spectra recorded at room temperature. However, on the basis of Mo¨ssbauer parameters, it is possible to differentiate between oxyhydroxides and any other iron phases such as iron sulfide.15 The amount of the new iron phase is low, and it increases up to 3 wt % at the end of the aging process (AP sample). The intensity ratio from iron in octahedral sites (Fe(III/II)) to tetrahedral ones (Fe(III)) initially is 1.9, corresponding to magnetite with a small degree of oxidation, but the ratio falls to 1.79 in the HTP sample, and then on the aging process a small additional decrease is also observed. This is conclusive evidence of the occurrence of an oxidation process in the magnetite crystals as a consequence of their contact with crude oil. (C) Crude Oil Composition. The crude oil sample has 32.1 degree API and an H/C atomic ratio of 1.84, and it contains 1.93 wt % of asphaltene separated with n-pentane. The amount (15) Greenwood, N. N.; Gibb T. C. Mo¨ssbauer Spectroscopy; Chapman and Hall: London, 1971.
Figure 3. FTIR spectrum of pure magnetite, HTP, AP, and asphaltene samples. Table 3. HPLC Analysis of the P1 Crude Oil, in wt % hydrocarbon type
as aromatic
as alkyl chain
total wt %
relative to crude oil, wt %
saturate one-ring aromatics two-ring aromatics three-ring aromatics four-ring aromatics polar compounds total wt % recovered
0.1 2.8 4.1 1.8 2.0 3.4 11.7 5.4
30.1 20.3 7.7 6.8 7.7 13.2 85.8 32.6
30.2 23.1 11.8 8.6 10.7 16.7 100
11.5 8.8 4.5 3.3 3.7 6.3 38
of sediment separated by centrifugation is 1:30 volume of sediment versus volume of crude oil. The elemental composition of crude oil, sediment, and asphaltene is shown in Table 2. The sediment shows an amount of oxygen higher than that of the asphaltene fraction related undoubtedly to water traces. Asphaltene, the heaviest crude oil fraction, concentrates a small amount of metals, and the crude oil sample also contains Si and Al (commonly related to clays and quartz minerals) as well as other metal traces (Ba, Fe and Na); Na element refers to NaCl compound. The composition of the deasphalted crude oil sample obtained by HPLC is shown in Table 3. The soluble fraction of crude oil contains 11.5 wt % of saturates, 20.2 wt % of aromatic with 1-4 rings in their structures, and 6.3 wt % of polar compounds. It is important to note that most aromatic rings contain mainly alkyl side chains and that most polar compounds are linear. However, following this technique the composition of about 68 wt % of this sample was not elucidated. (D) FTIR Results. Figure 3 shows the IR spectra of magnetite, HTP, AP, and asphaltene samples. The vibration bands identified as unperturbed CH2 and CH3 groups dominate the IR spectra of the HTP, AP, and asphaltene samples. The stretching vibrations of methyl and methylene bands at 2950, 2924, and 2850 cm-1 show variation of their intensities, indicating the variation of the amount of these functional groups or the immobilization of the alkyl chains on mineral surface. The spectrum of the HTP sample has a lower ratio of νCH2 (out-
1284 Energy & Fuels, Vol. 20, No. 3, 2006
Cosultchi et al.
Table 4. Thermal Gravimetric Analysis (TGA) Results in Air magnetite after contact with crude oil nontreated sample
magnetite
temp, in °C
weight gain, %
temp, in °C
HTP sample (1) weight loss, %
temp, in °C
weight loss, %
199.7-378.30
1.088
378.3-1011.1
1.685
29-107 107-311 311-402 402-535
9.06 22.91 5.30 1.40
25-114 114-249 249-362 362-500
0.756 8.918 19.52 9.48
of-phase)/νCH2 (in-phase) vibration bands compared to that of AP and asphaltene samples, indicating that only the in-phase or symmetric motion of hydrogen atoms is allowed. The wide band observed between 3700 and 3200 cm-1 and centered at 3425 cm-1 is observed in the spectra of HTP, AP, and asphaltene. This band corresponds to the hydroxyl group stretching vibration from water molecules probably adsorbed on the mineral surface.17-19 Farmer17 has indicated that lepidocrocite (γ-FeOOH) presents a stretching vibration band (νOH) at 3390 cm-1, an in-plane bending band (δOH) at 1160 cm-1, and an out-of-plane bending vibration band (γOH) at 753 cm-1. In both samples, the in-plane bending band at 1150 cm-1 may be related to lepidocrocite. For the AP spectrum, the intensity of this band is stronger than that for HTP. The IR spectrum of the AP sample shows a main band at 1626 cm-1 and bands at 1660, 1640, and 1580 cm-1 as shoulders, indicating the presence of CdC bonds conjugated to aromatic rings, and also the presence of metallic complex compounds.16 The band at 1150 cm-1 together with the bands at 1110 and 1100 cm-1 observed in the AP sample can also be attributed to C-O groups, while the bands at 1055, 1040, and 1018 cm-1 correspond to CdC and C-O vibration bands.16 The sharp band at 724 cm-1 observed only in the AP sample is normally attributed to the rocking mode of (CH2)n groups in large alkyl straight chains, although disubstituted alkenes and metallic chelate compounds also present strong stretching vibration bands of C-O and CdC groups in this region (775420 cm-1).16 (E) Thermogravimetric Results. The TGA results are presented in Table 4. The mass lost up to 107 °C and 114 °C in the HTP and AP samples, respectively, corresponds to a dehydration process. AP has a lower hydration degree, while for the HTP sample, the missing water was probably integrated within the iron hydroxide structure. The weight loss continues in both samples up to 311 °C, being higher for the HTP sample (22.91 wt %). However, up to 362 °C, the higher amount of mass loss (19.52 wt %) corresponds to that of the AP sample, which suggests that the aging process is related either to a higher adherence of the organic phase to the inorganic surface or to the formation of a new and more thermally stable organic phase such as high molecular weight paraffin or iron complex compounds, for instance. (F) Electron Microscopy Results. Figure 4 shows the micrographs of the samples obtained by TEM (Figure 4a,b) and SEM (Figure 4c,d), which provide a better understanding on the crude oil-mineral interaction. By TEM, the nontreated magnetite sample exhibits flat and massive particles with sizes of about 1 µm (right side of Figure 4a). After contact with (16) Bellamy, L. J. The Infrared Spectra of Complex Molecules, 3rd ed.; Chapman and Hall: London, 1975; Vol. 1. (17) The infrared spectra of minerals; Farmer, V. C., Ed.; Mineralogical Society; London, 1974. (18) Misawa, T.; Asami, K.; Hashimoto, K.; Shimodaira, S. Corros. Sci. 1974, 14, 279-289. (19) Raman, A.; Kuban, B.; Razvan, A. Corros. Sci. 1991, 32, 12951306.
AP sample (2)
petroleum, particles of different sizes are observed in the TEM micrograph in Figure 4b. The magnetite particles are the largest ones, and they preserve only certain features of their regular forms, which suggests that interaction with the organic phases affected the magnetite surface. These particles are surrounded by an amorphous mass with small particles embedded within it. Small acicular particles (