September 2024
Special Focus: Refining Technologies
Thermal decomposition of particulate mercury sulfides in petroleum—Part 1
Traces of particulate mercury sulfide (HgS) in stabilized crude oils transform to elemental mercury below 400°C in refinery distillation units. The authors evaluated this transformation and measured reaction rates and activation energies in crude oil and Hg-spiked mineral oil, using glass vessels at atmospheric pressure and a microunit at 1,000 psig.
Traces of particulate mercury sulfide (HgS) in stabilized crude oils transform to elemental mercury below 400°C in refinery distillation units. The authors evaluated this transformation and measured reaction rates and activation energies in crude oil and Hg-spiked mineral oil, using glass vessels at atmospheric pressure and a microunit at 1,000 psig. The results show that HgS decomposes irreversibly above ˜150°C. The data indicated the presence of two Hg species. Approximately 88% of the HgS decomposed with an activation energy of 56±7 kJ mol-1. The remaining 12% decomposed with an activation energy of 130±10 kJ mol-1. Near-identical kinetic parameters were found with differential measurements at atmospheric pressure and with integral measurements in the liquid phase at > 1,000 psig, indicating that mild pressures do not change the reaction kinetics. The results show that > 99% of HgS in crude oil decompose in a typical crude distillation unit (FIG. 1). This article covers the background, methods and experimental studies at atmospheric pressure. Part 2—to be published in the October issue—will include the results at 1,000 psig, the model of kinetics and further discussion.
FIG. 1. Initial evolution of mercury Hg species during refinery distillation.
Background. A previous study of Hg in crude oils and other petroleum processing streams found that metacinnabar (β-HgS) was the primary form of Hg with minor amounts of chemically related Hg-thiol.1 Collectively, these species are referred to as HgS in this article. Transmission electron microscopy (TEM) evaluation of particles recovered from a crude sediment showed nano-sized (< 100 nm) Hg-containing particles coating larger aggregates containing carbon (C), oxygen (O), aluminum (Al), silicon (Si), iron (Fe) and arsenic (As); and confirmed the colocation of Hg and sulfur, suggesting that Hg in these crude oils may be nano-sized HgS associated with other solids, which are taken to be a formation of sand and iron-corrosion debris. Hg in the aggregates was found to be non-volatile and could be removed by laboratory filtration and centrifugation.2 Filtration found the size distribution of these aggregates ranging from < 0.2 microns to > 20 microns, which varied significantly with the source. Other studies showed that Hg in crude can be extracted into an aqueous phase by solutions of ammonium sulfide or sodium sulfide.3 Since these reagents are not expected to dissolve formation material or iron-corrosion debris, the majority of the nano-sized HgS are believed to be located on the exterior surface of the micron-sized solids.
While Hg in crude oil is primarily non-volatile HgS, during refining, elemental Hg is often observed in light products from the distillation of crude [e.g., liquid petroleum gas (LPG) and naphtha],4 indicating that at least a portion of the solid HgS transforms to elemental Hg. However, older references to the thermal decomposition of HgS indicated it decomposed at 580°C, as detailed in literature.5 This value is significantly higher than the maximum temperature during crude distillation (400°C) and appears to contradict observations at crude oil refineries. Literature values for the thermal decomposition temperatures of metacinnabar (β-HgS) and related cinnabar (α-HgS) range between 150°C and 450°C, which are all significantly lower than the common reference values (TABLE 1).
Two studies determined the kinetics of HgS (cinnabar) thermal decomposition. Leckey and Nulf12 determined the decomposition of cinnabar powder to be first order with an activation energy of 230 kJ mol-1 for temperatures between 265°C and 345°C.
Town and Rao16 also determined the decomposition of cinnabar powder to be first order, stating, “Activation energies for cinnabar were found to be about 105 kJ mol-1 in nitrogen, 125 kJ mol-1 in air and 155 kJ mol-1 in vacuum, with little shift in the rate of reaction with respect to temperature under the three different atmospheres.”
These studies were done on dry powders and may not represent the diluted nano-scale particles of metacinnabar found in crude oil.
Materials and methods. The thermal decomposition of Hg in commercial crudes was studied (TABLE 2).
NA-1 is the crude used to derive Sample 1 in the previous publication, and SEA-3 is the crude used to derive Sample 2.1 Several of these crudes are sold as gas condensates, even though they are produced at the wellhead as liquids, along with liquid water, and are not condensed from a vapor. Therefore, the term “crude oil” is more applicable and is used in this article. All samples used in this study have been commercially stabilized—i.e., light volatile hydrocarbon species such as propane and isobutane have been thermally removed to conform to vapor pressure specifications for crude oil shipping.
Hg-containing crudes also have a unique distribution of sulfur species. They typically contain low but detectable concentrations of sulfur. However, they do not appear to contain appreciable amounts of mercaptans, hydrogen sulfide (H2S) or disulfides (the latter two species could not be detected in the above crudes). Their sulfur is almost exclusively in the form of aromatic thiophenes, which may reflect their geochemical origin. This unique sulfur species distribution has a practical use, as it can be used to identify likely high-Hg crudes from initial downhole samples.17 This is a more reliable method of identifying potentially Hg-containing crudes than a direct analysis of Hg from downhole samples.
In addition to these crudes, powder samples of metacinnabar and cinnabar were obtained from Alfa Aesar and blended with Superla™ mineral oil to 3,000-parts-per-billion (3,000-ppb) Hg to study the decomposition of these reagents in a simulated crude. To compare the desorption rate of elemental Hg with the decomposition rate of particulate Hg in crude, a sample of Superla mineral oil was sparged with elemental mercury vapor.3
The thermal decomposition was studied in two reactor configurations: batch glass vessels at atmospheric pressure, and a continuous-flow, stainless-steel microunit at 1,000 psig. The pressure in the latter was sufficient to keep the crude as a liquid.
The batch glass atmospheric pressure reactor configuration is shown in FIGS. 2 and 3, where 100 ml of crude was used with nitrogen sparging. The volatile low-molecular-weight portion condensed in a knockout pot. Nitrogen, with the evolved elemental Hg, flowed to a polysulfide adsorber, which captured the Hg. Tests on the gas exiting the absorber did not detect Hg. The Hg contents of the crude, knockout pot liquid and polysulfide solution were monitored over time. The Hg content of the knockout pot liquid was typically below detection. Analyses of the liquids and gases were performed by use of a proprietary analyzera. The gases were analyzed by use of a clean 1-cc gas syringe with injection through a septum into the air supply line to the analyzer. When liquid samples reported Hg values below 100 ppb, samples were analyzed by the U.S. Environmental Protection Agency’s (EPA’s) modified 1631E method, which has a limit of detection < 1 ppb.
FIG. 2. Schematic of the glass atmospheric pressure apparatus.
FIG. 3. Photo of the glass atmospheric pressure apparatus.
Material balances for total mass and Hg were calculated for each sample interval. The mass of Hg in the three-neck flask was based on the measured concentration and the initial mass of crude but was adjusted for the loss of crude due to vaporization. Mass balances averaged 89%. Hg mass balances at the end of the experiment ranged from 80% to 120%, with an average of 96%. Since Hg in the crude oil is particulate and tends to settle, obtaining a consistent sample from the storage container was difficult, even with vigorous shaking. Analyses were based on the measured Hg in the flask, rather than on the Hg content of the sample in the storage container.
The data from the atmospheric pressure studies were processed as follows:
- Measurements of the Hg in the three-neck flask were made over time, along with measurements in condensate, gas phase above the crude, and the polysulfide solution.
- Since a portion of the crude boils during this operation (primarily during the heating period), the measured concentration was adjusted (decreased) for the lost light hydrocarbons. This loss was assumed to occur during the 30-min heating period.
- A differential first-order kinetic analysis plot of the ratio of natural log of the initial Hg concentration in the crude oil to the Hg concentration in the crude oil measured over time vs. time was created—i.e., [Hg(0) / [Hg(t)] vs. time.
- The slope of this line provides the first-order rate. In some cases, the data fell on two straight lines. In these cases, two rates were recorded.
- The natural log of the first-order rates was then plotted vs. the reciprocal of the absolute temperature. The slope of this line gives the activation energy, and the intercept gives the pre-exponential factor.
The batch atmospheric pressure measurements provided measurements of the rate over time. In contrast, the continuous microunit studies only report an integral of the rate, but both measurements have utility.
Results. The Hg percentage (left ordinate) and temperature measurements (right ordinate) over a typical run are shown in FIG. 4. The Hg measurements show the percentage of Hg in gas, crude oil, the polysulfide adsorbent, the sample aliquots and the overall loss. The percentages of Hg in the gas and sample aliquots were very small proportions.
FIG. 4. Distribution of Hg vs. time for SEA-3 at 175°C.
Hg mass balance identified a fraction of Hg that was not accounted for in the gas, crude oil, sample aliquots or the polysulfide trap. This Hg was most likely due to elemental Hg adsorbed on the glass surfaces of the tubing and condenser. Initially, losses were high and then declined. Presumably, this was due to Hg first adsorbing on the surfaces and then desorbing once most of the Hg in the crude had been converted and swept into the polysulfide solution.
FIG. 5 depicts the progression of normalized Hg concentration in the crude oil sample with time, adjusted for volume losses. Consistent with previously reported studies, the thermal decomposition of the particulate sulfidic Hg in crude oil appeared to follow first-order kinetics.12,16
FIG. 5. First-order kinetics plot of Hg decomposition in SEA-3 at 175°C.
FIG. 6 shows an Arrhenius plot for data from a single crude (SEA-3). A linear regression of the Arrhenius plot was conducted to determine the slope and intercept. The slope gives an activation energy of 56±7 kJ mol-1.
FIG. 6. Arrhenius plot of reactive HgS from SEA-3.
FIG. 7 shows an Arrhenius plot of the rate data for all crudes. A linear regression of the Arrhenius plot was conducted to determine the slope and intercept. The activation energy appears to be consistent with the value determined from the single crude (SEA-3), but individual crudes are somewhat offset.
FIG. 7. First-order kinetic rate constants vs. 1,000/T(°K) for reactive Hg.
In approximately half of the experiments, the Hg content declined logarithmically following first-order kinetics. However, when approximately 90% had been reacted, the decline slowed and followed a different linear relationship, as shown in FIG. 8.
FIG. 8. First-order kinetic plot for HgS decomposition in SEA-3 at 200°C.
For convenience, the two kinetic species were called reactive and refractory, where the reactive species has the lower activation energy. Considering all tests that demonstrated this behavior, approximately 12% of the Hg was characterized as refractory by way of the kinetic analysis. The refractory rate was found to have an activation energy of 130±10 kJ mol-1, as shown in FIG. 9. The refractory rate cannot always be measured due to its higher activation energy, which may explain the absence of refractory Hg in 18 out of 28 experiments. At temperatures above 250°C, the rate of refractory Hg decomposition is so rapid that it cannot be measured accurately. At temperatures below 175°C, the rate of reaction of the refractory Hg is too slow (several hours) to allow measurements during the time of the experiments. The cause of this different kinetic behavior is unknown.
FIG. 9. First-order kinetic rate constants vs. 1,000/T(°K) for refractive Hg.
Two additional experiments were performed. The decomposition of reagent particulate HgS was studied to compare with the HgS in crude oils. The stripping of elemental Hg from Superla mineral oil was also studied to determine if the disappearance of Hg from the crude samples was limited by decomposition or by the stripping itself.
FIG. 10 shows changes in first-order rate constants with temperatures for particulate metacinnabar (β-HgS), particulate cinnabar (α-HgS) and elemental Hg blended with Superla mineral oil. The rate constant for decomposition of metacinnabar powder was found to be approximately 60% lower than the average value for crude oils, but the activation energy was not significantly different. The rate of decomposition of cinnabar was approximately 50% lower than that of metacinnabar. The difference between cinnabar and metacinnabar might be explained by the heat of formation, which is -53.35 kJ mol-1 and -49.17 kJ mol-1 for cinnabar and metacinnabar, respectively.19 Therefore, cinnabar is more stable than metacinnabar, and it decomposes at a slower rate.
FIG. 10. First-order kinetic rate constants for α- and β-HgS and Hg0 in Superla mineral oil.
In contrast, the rate constant for stripping elemental Hg was approximately 20 times higher than the constant for Hg decomposition in crude oils, indicating that the decomposition was rate limiting vs. the stripping.
FIG. 11 depicts the apparent influence of particle size on the reactive rate of HgS decomposition at 200°C. The reactive rate constant for the different crudes increased as the percent of Hg in the smallest-size fraction increased.
FIG. 11. Apparent influence of aggregate particle size on the refractory rate of HgS decomposition at 200°C.
This is consistent with the lower rates for the reagent powders that had particle sizes that were significantly larger than those found in crude oils. However, it must be noted that the particle size reported here is the composite of the nano-meter-scale metacinnabar on the micron-scale formation material, and not a measurement of the metacinnabar particles themselves. The cause of this different kinetic behavior is unknown.
Part 2. Part 2 of this article will be featured in the October issue.
NOTE
a Lumex RA-915+ analyzer
SUPPLEMENTAL MATERIALS
Supplemental materials are available from the corresponding author(s) upon request.
LITERATURE CITED
1 Avellan, A., J. P. Stegemeier, K. Gai, J. Dale, H. Hsu-Kim, C. Levard, D. O’Rear, T. P. Hoelen and G. V. Lowry, “Speciation of mercury in selected areas of the petroleum value chain,” Environmental Science and Engineering, 2018.
2 Khan, R. M., et al., “Sweetening of sour crudes,” U.S. Patent No. EP1532227A1, 2011.
3 O’Rear, D., R. E. Cooper and S. Husain, “Method for removing mercury from crude oil,” U.S. Patent No. US10179879B2, 2019.
4 Hadden, R. and T. Moss, “Dealing with mercury in refinery processes,” Digital Refining, July 2010, online: https://www.digitalrefining.com/article/1000496/dealing-with-mercury-in-refinery-processes#.XSEx1-hKjIU
W
6 Golas, P. J., Science and Civilization in China: Volume 5, Chemistry and Chemical Technology, Cambridge University Press, 2020.
7 Montgomery, H., “Mercury facts,” University of Texas, 1999, online: http://www.utdallas.edu/~mont/C_MAP/CMAP%20version%20of%20Mercury%20Fact%20Sheet.pdf
8 Tersic, T., M. Gosar and H. Biester, “Influences of ore roasting sites from the first period of mining on the mercury dispersion in the surroundings of Idrija (Slovenia),” Berichte der Geologischen Bundesanstalt, November 2008.
9 Erspamer, E. G. and R. R. Wells, “Selective extraction of mercury and antimony from cinnabar-stibnite ore,” Bureau of Mines, September 1956, online: https://www.govinfo.gov/content/pkg/GOVPUB-I-PURL-gpo89700/pdf/GOVPUB-I-PURL-gpo89700.pdf
10 Bell, D. A., B. F. Towler and M. Fan, Coal Gasification and its Applications, Elsevier, Burlington, Massachusetts, 2011.
11 Balaz, P. and E. Godocikova, “Thermal reduction of mechanically activated cinnabar (HgS) and stibnite (Sb2S3), Journal of Thermal Analysis and Calorimetry, 2001.
12 Leckey, J. H. and L. E. Nulf, “Thermal decomposition of mercuric sulfide,” U.S. Department of Energy, October 28, 1994, online: https://www.osti.gov/servlets/purl/41313
13 Wu, S., M. A. Uddin, S. Nagano, M. Ozaki and E. Sasaoka, “Fundamental study of decomposition characteristics of mercury compounds over solid power by temperature-programmed decomposition desorption mass spectrometry,” Energy & Fuels, December 2010.
14 Taimni, I. K. and S. N. Tandon, “Thermolysis of sulphides precipitated by sodium sulphide: Sulphides of arsenic, mercury, rhenium, antimony and molybdenum,” Analytica Chimica Acta, 1960.
15 Wilhelm, S. M. and M. Nelson, “Interaction of elemental mercury with steel surfaces,” The Journal of Corrosion Science and Engineering, September 2010.
16 Town, J. W. and P. D. Rao, “Thermogravimetric and distillation studies on mercury, antimony and arsenic sulfides,” Mineral Industry Research Laboratory, University of Alaska, December 1975, online: https://core.ac.uk/download/pdf/162573744.pdf
17 O’Rear, D. J., C. D. Wehunt and G. A. Winslow, “Process, method and system for removing heavy metals from fluids,” U.S. Patent No. US20150322764, 2011.
18 Liang, L., “CEBAM Analytical,” online: http://cebam.net/
19 Sharma, R. C., Y. J. L. Chang and C. Guminski, “The Hg-S (mercury-sulfur) system,” Journal of Phase Equilibria, 1993.
Comments