March 2023

Special Focus: Petrochemicals Technology

An advanced extractive distillation process for aromatics recovery

This article focuses on a novel extractive distillation (ED) process for recovering benzene, toluene and xylene (BTX) aromatics from a full-boiling-range pyrolysis gasoline or reformate feedstock—a process that has been commercially demonstrated and is ready for licensing or a JV.

This article focuses on a novel extractive distillation (ED) process for recovering benzene, toluene and xylene (BTX) aromatics from a full-boiling-range pyrolysis gasoline or reformate feedstock—a process that has been commercially demonstrated and is ready for licensing or a JV. The process’s solvent regeneration system, operated at low temperatures and pressures with no energy requirement, completely prevents the accumulation of heavy hydrocarbons and polymeric sludge in the lean solvent, thus maintaining a 99.9% purity with essentially no solid particles (not even the color bodies).

The modified ED column (EDC) can maintain benzene specifications of the EDC (overhead) raffinate and simultaneously keep C₈ naphthenes away from the EDC (bottom) extract to ensure mixed-xylenes product purity and toluene product recovery. Through proprietary operational adjustments and an optimized tray design, the modified EDC provides more stable and efficient operation in the upper two-liquid-phase region. This ED technology is also suitable for revamping conventional liquid-liquid extraction (LLE) or ED processes. Converting a conventional LLE process into the novel ED process can save 20%–30% in energy and increase throughput by up to 20%.

The LLE process has dominated the hydrocarbon processing industry in purifying BTX from petroleum streams for more 50 yr.^1–3 However, the LLE process is complicated—typically including steps such as LLE, extractive stripping (ES), raffinate washing and solvent recovery (SR), as well as being energy intensive. An alternative process is ED, which was first developed in the late 1970s.^4,5

The ED process has the following advantages over LLE:

• Less major equipment and smaller plot plan
• Lower energy consumption
• No need to purge the light products in the recycle stream to the LLE column
• Light products can be fed to the EDC (there is no requirement for a depentanizer for the reformate feed)
• Ability to handle a wider range of aromatic concentrations in the feed
• More flexible than the LLE process
• EDCs can operate like conventional distillation columns
• Startup of the ED process takes only hours vs. days or weeks for the LLE process.

Conventional ED process

A conventional ED process for BTX aromatics recovery using sulfolane as the extractive solvent is shown in FIG. 1. The hydrocarbon (HC) feed is fed to the EDC through Line 1. The principal of ED is to significantly increase the relative volatility between the close-boiling aromatic and non-aromatic compounds in the HC feed by introducing a polar compound selective solvent (via Line 2) into the upper portion of the EDC. The solvent preferentially extracts the more polar compounds (mainly aromatics) in the rising vapor mixtures, thus allowing fewer polar compounds (mainly non-aromatics) to rise to the top of the EDC to produce the raffinate product through Line 3. The EDC bottom rich solvent stream containing the solvent and aromatics is transferred directly to a solvent recovery column (SRC) via Line 4. The aromatics are stripped by steam from the rich solvent stream in the SRC and withdrawn from Line 5 as the extract product from the overhead of the SRC. The stripped lean solvent—exiting from the bottom of the SRC—is recycled to the upper portion of the EDC, via Line 2. For solvent regeneration, a small slipstream of the lean solvent from Line 2 is fed via Line 6 to a thermal solvent regenerator. Additional details on the characteristics, design and operation of a conventional ED process have been thoroughly discussed in literature.^6–9 

FIG. 1. A conventional ED process.

Challenges of the conventional ED process

Despite the advantages over the LLE process, the ED process, since its first commercial practice more than 50 yr ago, has never been a significant competitor against the LLE process for BTX aromatics recovery from pyrolysis gasoline or reformate. Although numerous improvements have been made in recent years, the developers are still unable to solve some of the basic challenges of the ED process for BTX aromatics recovery. The major challenges of a conventional ED process, using sulfolane as the extractive solvent, are:

1. Preventing excessive accumulation of heavy HCs and polymeric sludge in the lean solvent. Extractive solvent in both ED and LLE processes for aromatics recovery is internally circulated in the process system in a closed loop. To remove heavy HCs and sludge from the lean solvent, the conventional commercial LLE or ED process employs a thermal solvent regenerator. Only a small slipstream of lean solvent (approximately 1%–2%) is withdrawn from the solvent loop and heated with or without stripping steam in the regenerator to recover the solvent and any heavy components with boiling points lower than that of the solvent. The basic concepts of this thermal solvent regeneration scheme have been described in literature^10,11 in relation to an LLE process for aromatics recovery.
   
   This means that up to 99% of heavy HCs stay in the closed solvent loop as the possible species to be dehydrogenated into active olefins/diolefins and then polymerized into sludge and catalyzed under iron oxides/iron sulfides in the lean solvent. These catalytically active iron oxides/iron sulfides are generated through corrosion of the carbon-steel process equipment at high temperatures by acids in the lean solvent. These acids are produced through the oxidation of the sulfolane solvent by air—most likely leaking through the SRC unit, which is operated under negative pressure and high temperatures.
   
   Nearly all commercial ED processes for aromatics recovery have adapted the thermal regeneration design from the LLE process. It should be noted that approximately 67% of heavy HCs in the lean solvent in an LLE process are removed with the raffinate stream in the LLE column. The thermal solvent regeneration design has been commercially successful in keeping the relatively minor amounts of heavy HCs and sludge in the lean solvent at a tolerable level for the LLE process. However, this thermal regenerator is unsuitable for ED processes, since its lean solvent loop is loaded with approximately three times more heavy HCs and sludge than in the LLE process, and these can only be removed by the thermal regenerator. This is a potentially unwanted situation, since the presence of excessive heavy HCs and sludge in the lean solvent not only significantly changes the solvent properties (selectivity and solvency), but also plugs process equipment (such as pumps, valves, column internals and lines) to render the ED process inoperable.

2. Maintaining benzene specifications of the EDC (overhead) raffinate and keeping C₈ naphthenes away from the EDC (bottom) extract. A commonly acceptable benzene specification of an EDC overhead raffinate stream is 1 vol% or less to be qualified for gasoline blending. Meanwhile, C₈ naphthenes in the EDC bottom stream must be controlled at a certain level, since they form minimum-boiling azeotropes with toluene, causing purity or recovery problems of toluene and xylenes. Possible azeotropes of C₈ naphthenes and toluene are listed in TABLE 1.
   
   
   With a conventional EDC design and configuration for processing a full-range (C₆–C₈) pyrolysis gasoline or reformate feed, it is very difficult to maintain benzene specifications of the EDC overhead raffinate stream while keeping C₈ naphthenes away from the EDC bottom extract stream.

3. Maintaining stable EDC operation in the upper two-liquid-phase region. Traditionally, it is believed that the successful operation of an EDC requires a high concentration of solvent to avoid forming two immiscible liquid phases within the column. When immiscibility occurs, the raffinate liquid phase (second liquid phase) is considered to reduce the relative volatility of the key components to be separated. In addition, when using the ED process with furfural as the solvent for C₄ HC separation, flooding was reported at a relatively low HC feed rate if an immiscible liquid phase was allowed to separate on trays below the HC feed tray.¹² Davies, et al.¹³ found that foaming occurred in trays with a liquid composition near that of the one-liquid phase/two-liquid phase transition. While sulfolane is a highly selective solvent for aromatics recovery, the solubility of paraffins and naphthenes in sulfolane is quite low—therefore, the paraffin/naphthene-rich top section of an EDC tends to operate under two liquid phases, even under a high solvent-to-HC feed ratio.

An advanced, proprietary ED process^a for aromatics recovery

Recognizing the defects of the conventional ED process for BTX aromatics recovery, one of the co-authors (F. M. Lee) has spent more than 40 yr serving at major U.S. energy companies, as well as engineering and construction companies, to develop technologies that not only eliminate or minimize the basic challenges of the ED process, but that also further improve the process performance through commercial demonstrations.^14–28 A schematic diagram of this commercially ready ED process^a is shown in FIG. 2, with the implemented proprietary technologies shown in the yellow-shaded steps.

FIG. 2. Schematic of the advanced, proprietary ED process^a.

Proprietary solvent regeneration system (SRS). A different proprietary approach for removing any heavy HCs, sludge and/or metal-oxide particles from the extractive solvent is implemented in FIG. 2.  

As shown schematically in FIG. 2, the entire stream of lean solvent from the bottom of the SRC is fed directly (via Line 18) to the proprietary SRS without cooling to conserve process energy. Any solid particles—including polymeric sludge, metal oxides and sulfides—are removed. Nearly-solids-free (99.9% purity) lean solvent is generated from the SRS unit and transferred through Line 12. Note: The SRS can remove small particles with sizes in the nanometer range. Furthermore, the SRS is operated without any energy consumption for heating or cooling. The SRS can be operated with automated operating and regenerating cycles without risking hazardous exposure to the process fluids. The environmental costs for sludge disposal are also substantially reduced.

Proprietary EDC overhead system (SEOS). The conventional EDC overhead system is replaced by the proprietary SEOS, which is operated without energy consumption. Operations of the modified EDC are adjusted accordingly to be more effective. These steps include:

1. Removing heavy non-aromatics—especially the C₈ naphthenic compounds from the EDC bottom stream—to increase the purity of mixed xylenes and the recovery of toluene to match those produced from other LLE processes
2. Removing the dissolved heavy HCs from the lean solvent by feeding a slip solvent stream to the SEOS
3. Controlling and recovering benzene from the raffinate product to maintain its quality as a gasoline blendstock and thus increase the recovery of benzene in the aromatic products.

Two-liquid phases in the upper portion of the EDC. As previously mentioned, the non-aromatic components in the feed mixture cause the formation of a second liquid phase in the upper portion of the EDC due to their significantly lower solubility in the sulfolane solvent than the aromatic components. In a distillation column with two-liquid phases, if not well mixed, the two-liquid phases tend to separate on trays and cause poor separation efficiencies and unstable operations. 

In the proprietary ED process^a, the upper portion of the EDC is operated to significantly reduce the formation of the second liquid phase as one way to improve the solvent performance toward that of the single-liquid phase.

In addition, trays with a design to promote phase mixing are installed in the upper portion of the EDC, causing the two-liquid phases to behave as homogeneous pseudo-single-liquid phases. The improved agitation will minimize stagnant liquid holdup on the tray, while allowing proper vapor/liquid disengagement from tray to tray.

Commercial demonstration of the proprietary ED process^a for the recovery of BTX aromatics from a full-boiling-range pyrolysis gasoline

According to the proprietary ED process configuration presented in FIG. 2, an 8,000–10,000-bpd LLE plant was revamped into an ED plant for recovering BTX aromatics from a full-boiling-range (C₆–C₈) pyrolysis gasoline. The HC feed—with a composition shown in TABLE 2—was fed to the middle portion of the EDC via Line 11 at a flowrate of 56 m³/hr (approximately 8,000 bpd). Lean solvent—with a composition shown in TABLE 2—was fed to the upper portion of the EDC, while a slipstream of the lean solvent was fed to the SEOS unit. The raffinate stream from the top of the EDC was also introduced to the SEOS unit.

The upper portion of the modified EDC was designed and operated to reduce the detrimental effects of the two-liquid phases by minimizing the formation of a second liquid phase and improving the tray design with enhanced mixing capabilities. Controlling the C₈ naphthene content in the bottom extract stream was the primary focus of this modified EDC operation. The Extract Product (Line 17) section of TABLE 2 (extract product) shows that the C₈ naphthene content in the extract stream (Line 17) was 0.233 wt%; however, it can be further reduced, if necessary, by delivering more benzene into the EDC overhead raffinate stream. The excess benzene is then recovered by the SEOS unit and returned to the EDC.   

The SEOS successfully serves two purposes:

1. Recovering excess benzene from the EDC raffinate stream to qualify it for gasoline blending (with less than 1 vol% benzene). As shown in the Raffinate Entering the SEOS Unit (Line 13) section of TABLE 2, the raffinate stream entering the SEOS contained 2.596 wt% benzene (Line 13), and the raffinate stream leaving the SEOS contained only 0.831 wt% benzene (a 68% reduction) (Line 14).
2. Removing and recovering dissolved heavy HCs from the lean solvent. Line 12 of TABLE 3 shows that circulating lean solvent entering the SEOS was already very clean (containing only 0.022 wt% C₉–C₁₁ aromatics and 0.0732 wt% C­₁₂₊ heavy HCs), but that lean solvent leaving the SEOS was even cleaner with essentially no detectable C₉₊ HCs.  
   
   

Rich solvent (containing the extract and the solvent) was transferred from the bottom of the EDC to the middle portion of the SRC for steam stripping to recover the extract product from the overhead of the SRC (through Line 17). The Extract Product (Line 17) section of TABLE 2 shows that C₈ naphthenes in the extract stream (Line 17) were 0.233 wt%, but operation of the EDC could be adjusted to reduce it further, if required.

Lean solvent exiting from the bottom of the SRC was passed through the heating coils of a water stripper as the heat source. Then, the entire lean solvent stream was fed to the SRS to remove any detectable sludges. As shown in the Lean Solvent (Line 12) section of TABLE 2, lean solvent leaving the SRS had a 99.9 wt% purity with only trace amounts of heavy HCs (0.0013 wt%). The solvent showed a transparent straw color (FIG. 3), meaning that no heavy color bodies with nanometer particle sizes existed in the solvent.

Takeaways

A proprietary ED process^a has been commercially demonstrated for recovering BTX aromatics from a full-boiling-range pyrolysis gasoline feedstock. This development summarizes the non-interrupted research and development efforts of the past 40 yr. The successful commercial test demonstrated:

• This technology completely prevented the accumulation of heavy HCs and polymeric sludge in the lean solvent to maintain a purity of 99.9% with essentially no solid particles (not even the color bodies). The proprietary SRS is operated at low temperatures and pressures with no energy requirement.
• The new process maintained benzene specifications of the EDC (overhead) raffinate and simultaneously kept C₈ naphthenes away from the EDC (bottom) extract to ensure mixed-xylenes product purity and toluene product recovery. In other words, the operation of the modified EDC can be adjusted to meet any C₈ naphthenes requirement in the extract stream, while still satisfying the benzene requirement of the raffinate product.
• Through proprietary operational adjustments and an optimized tray design, this process maintained stable EDC operation in the upper two-liquid-phase region.

It is conceivable that customers can easily adapt the proprietary SRS to replace the low-efficiency and high-cost thermal solvent regenerator. The proprietary SRS comprises a low-cost, standalone unit that is easy to operate. This system requires no energy to maintain the circulating lean solvent with a purity up to 99.9% and with only trace amounts of sludges for LLE or ED processes. The proprietary ED process^a is suitable for revamping conventional LLE or ED processes. Converting a conventional LLE process into the proprietary ED process^a can save 20%–30% in energy costs and increase throughput up to 20%. HP  

NOTE

^a Shin-Chuang Technology’s SCT Extractive Distillation process

LITERATURE CITED

1. Deal, C. H., H. D. Evans, E. D. Oliver and M. N. Papadopoulus, “Extraction of aromatics with sulfolane,” 5th World Petroleum Congress, New York, New York, 1959.
2. Voetter, H. and W. C. G. Kosters, “Industrial experiences and new possibilities in the field of solvent extraction with sulfolane,” Erdoel Kohle, 1966.
3. Broughton, D. B. and G. F. Asselin, “Production of high-purity aromatics by the Sofolane process,” 7th World Petroleum Congress, Mexico City, Mexico, 1967.
4. Cines, M. R., “Extractive distillation,” U.S. Patent No. 4053369, October 11, 1977.
5. Lee, F. M. and D. M. Coombs, “Two-liquid-phase extractive distillation for aromatics recovery,” Industrial and Engineering Chemistry Research, March 1987.
6. Atking, G. T. and C. M. Boyer, “Application of McCabe -Thiele Method to Extractive Distillation Calculations” Chemical Engineering Progress, September 1949.
7. Chambers, J. M., “Extractive distillation—Design and applications,” Chemical Engineering Progress, September 1951.
8. Hachmuth, K. H., “Industrial viewpoints on separation processes—Part 3,” Chemical Engineering Progress, December 1952.
9. Butler, R. M. and J. A. Bichard, “Separation of aromatics from hydrocarbon streams,” U.S. Patent No. 3,114,783, December 1963, online: https://patentimages.storage.googleapis.com/a4/47/d3/08a2e15dad100b/US3114783.pdf
10. Asselin, G. F., “Aromatic extraction with solvent recovery and regeneration,” U.S. Patent No. 4,048,062, September 1977.
11. Asselin, G. F., “Aromatic hydrocarbon separation via solvent extraction,” U.S. Patent No. 4,046,676, September 1977.
12.  Happel, J., P. W. Cornell, D. Eastman, M. J. Fowle, C. A. Porter and A. H. Schutte, Transactions of the American Institute of Chemcial Engineers, Vol. 42, 1946.
13. Davies, B., A. Zafar and K. E. Porter, “Distillation of systems containing two liquid phases,” AIChE Journal, January 1987.
14. Lee, F. M., “Extractive distillation for aromatics recovery,” U.S. Patent No. 9,221,729, December 2015, online: https://patentimages.storage.googleapis.com/62/1d/16/5f5ebad7a30fd4/US9221729.pdf
15. Lee, F. M., et al., “Methods for removing heavy hydrocarbons from extractive solvents,” U.S. Patent No. 8,680,358, March 2014, online: https://patentimages.storage.googleapis.com/56/01/86/ced545bf03f922/US8680358.pdf
16.  Lee, F. M., et al., “Extraction process with novel solvent regeneration methods,” U.S. Patent No. 8,663,461, March 2014, online: https://patentimages.storage.googleapis.com/ff/bb/47/25cec0dff9360f/US8663461.pdf
17. Lee, F. M., et al., “Low-energy extractive distillation process for dehydration of aqueous ethanol,” U.S. Patent No. 8,172,987, May 2012, online: https://patentimages.storage.googleapis.com/77/51/ed/8b30b0e6c8460f/US8172987.pdf
18.  Lee, F. M., et al., “Low-energy extractive distillation process for dehydration of aqueous ethanol,” U.S. Patent No. 8,002,953, August 2011, online: https://patentimages.storage.googleapis.com/ca/02/de/96c0608e00825b/US8002953.pdf
19.  Lee, F. M., et al., “Energy efficient and throughput enhancing extractive process for aromatic recovery,” U.S. Patent No. 7,879,225, February 2011, online: https://patentimages.storage.googleapis.com/fa/19/dc/9865d2a7861549/US7879225.pdf
20. Lee, F. M., et al., “Extractive distillation processes using water-soluble extractive solvents,” U.S. Patent No. 7,871,514, January 2011, online: https://patentimages.storage.googleapis.com/4b/da/bf/951d91cd87b436/US7871514.pdf
21. Lee, F. M., “Aromatic purification from petroleum streams,” U.S. Patent No. 6,781,026, August 2004, online: https://patentimages.storage.googleapis.com/52/5e/fd/b0dcf14cbf6c08/US6781026.pdf
22. Lee, F. M., et al., “Process of removing sulfur compounds from gasoline,” U.S. Patent No. 6,551,502, April 2003, online: https://patentimages.storage.googleapis.com/09/d6/35/96a12a84a78f87/WO2001059033A1.pdf
23. Lee, F., “Recent development of extractive distillation: A distillation alternative,” Encyclopedia of Chemical Processing and Design,” Vol. 69, 2001.
24. Lee, F., “Extraction/distillation,” Encyclopedia of Separation Science, Summer 2000.
25. Lee, F., et al., “Extractive distillation for aromatics purification,” AIChE National Meeting, Houston, Texas, April 2001.
26. Lee, F., et al., “Extractive distillation: Separating close-boiling-point components,” Chemical Engineering, November 1998.
27. Lee, F., et al., “New extractive distillation process for aromatic recovery,” AIChE Spring National Meeting, Houston, Texas, March 1995.
28. Lee, F., et al., “Two-liquid-phase extractive distillation for aromatics recovery,” Industrial and Engineering Chemistry Research, March 1987.

The Authors

Related Articles

From the Archive

Comments

Comments

{{ error }}

Add Comment

{{ comment.comment.Name }} Editor • {{ comment.timeAgo }}

{{ comment.comment.Text }}