February 2022

Bio-based Processing

Impact of biofeed retrofits, co-processing on refinery amine units, SWSs and SRUs—Part 2

Producing diesel with a portion of biologically sourced carbon is being done at an increasing number of conventional crude oil refineries.

Le Grange, P., Tekebayev, K., Goettler, L., Kiebert, J., Sulphur Experts; Sheilan, M., Amine Experts

Producing diesel with a portion of biologically sourced carbon is being done at an increasing number of conventional crude oil refineries. This is commonly achieved by co-processing biofeedstocks in refinery hydrotreaters and FCCUs or through the installation of a dedicated biofeedstock hydrotreater to produce commercial biodiesel products. Generally, refineries are looking to technologies that allow them to easily incorporate biofeedstocks into their existing infrastructure. Part 1 of this article, published in the January issue of Hydrocarbon Processing, described the impacts that this incorporation has on a refinery’s existing amine and SWS units. Part 2 will detail the impacts on the SRU, operational and design options available to manage those impacts, and a specific case study.

Impacts on the SRU

The amine acid gas (AAG) from an amine unit and sour water acid gas from an SWS unit are fed into one or more SRUs. FIG. 8 shows the typical process flow diagram for a two-stage SRU. In the reaction furnace, a portion of the H2S is partially oxidized to SO2, which then reacts with the remaining H2S to form elemental sulfur over the thermal and catalytic stages of the process. A detailed description of the process and chemistry is provided in literature.18 The co-processing of biofeedstock will change the amount and gas composition to the SRU. This will be mostly an increase of CO2 and a reduction of H2S. Processing the changed feed is unlikely to be a problem since most SRUs can handle changes in composition and load to the unit. While this means that the existing equipment may be adequate, it does require careful review before committing to biofeedstock processing.

FIG. 8. Process flow diagram of a two-stage modified Claus unit.

The impact of the extra CO2 on the SRU is limited to the thermal stage (reaction furnace). The thermal stage not only oxidizes H2S and thermally converts it into elemental sulfur, but it also helps to destroy feed contaminants such as hydrocarbons, benzene, toluene, ethylbenzene and xylene (BTEX) and ammonia. If not destroyed, these contaminants can cause operational challenges such as catalyst damage and plugging in the downstream equipment due to salt formation (FIG. 9). The additional CO2 will cool the bulk gas temperature in the reaction furnace, which could reduce contaminant destruction. With only hydrocarbons and BTEX in the feed, the minimum target temperature is 1,050°C. When sour water acid gas (SWAG) is one of the feeds, the key requirement is to ensure good destruction of ammonia, meaning that the process gas temperature in the furnace is recommended to be at least 1,250°C.19 

FIG. 9. Catalyst deactivation by contaminants (left) and by plugging in equipment (right).

The ability to achieve the target temperature will be reviewed for the two most common thermal stage configurations (FIGS. 10A and 10B): the straight-through process (single burner, single chamber) and the split-flow process (single burner, two chambers). The advantages and disadvantages of both configurations have been well documented in literature.19 

FIG. 10. Views of a straight-through furnace configuration (A) and a split-flow furnace configuration (B).

Straight-through furnace. In the straight-through (one burner, one chamber) configuration, the reaction furnace temperature is strictly a function of the process chemistry. The extra CO2 from the biofeedstock only cools the furnace. When processing SWAG, if the resulting furnace temperature is expected to be less than the 1,250°C target value, then more ammonia will leave the furnace and increase the risk of ammonia salt formation and plugging (FIG. 9) in the downstream waste heat boiler and subsequent equipment.

Although increasing the furnace temperature is possible by warming the feed gas(es) and combustion air with preheaters, there is a considerable capital investment in addition to the extra operating cost and maintenance complications. Alternatively, the temperature can be increased by co-firing with some type of fuel gas or by oxygen enrichment of the combustion air stream.

Fuel gas is also used in the SRU for startup, hot standby and shutdown operations, as a reliable source of energy to manage the furnace temperature. The fuel may be natural gas, refinery fuel gas, hydrogen, etc. Whichever gas is used, it is critical that it is of constant composition to maintain good process control. Large fluctuations in the gas composition, and, thus, combustion air demand, will interfere with control of the H2S:SO2 ratio target, meaning lost recovery efficiency and other possible dangers. If continuous co-firing is selected as the heating source, the design of the burner must first be reviewed. Next, ensure the SRU’s maximum throughput or hydraulic load can handle the extra load from co-firings, as co-firing will add a large amount of inert gas to the SRU. The added dilution also reduces the sulfur recovery efficiency, which may be unacceptable. Finally, the CO2 emissions of the facility will increase, which may be a concern in some locations.

A cleaner and better way to improve the furnace temperature is to increase the oxygen concentration of the combustion air. Oxygen enrichment has been used successfully in SRUs that are processing lean (i.e., low concentration H2S and more CO2) acid gas to achieve the desired furnace temperature. However, this requires some modification to the system. The target oxygen concentration will determine the extent of the changes required—low-level oxygen is quite a simple addition, while greater concentrations (up to 100% oxygen) require a special burner and materials with a dedicated oxygen lance.20 The source and cost of oxygen must also be considered. It is important to check the turndown capability of the burner, since the addition of oxygen to the process increases not only the flame temperature, but also the speed of combustion. Usually, systems equipped with oxygen enrichment require more throughput to ensure that the burner operates in a safe regime to protect the metal and refractory components.

Split-flow furnace. In the split-flow process, the SWAG containing the ammonia is processed directly through the main burner with a portion of the AAG. The balance of the AAG is sent to the rear zone of the furnace and is used to control the temperature in the front zone (FIG. 11). The CO2 content of the SWS, especially with biofeedstock processing, may reduce the effectiveness of the bypass design since this increases the gas volume without adding heat of combustion. If the target temperature cannot be achieved with the split-flow design alone, then co-firing or oxygen enrichment can be used.

FIG. 11. Acid gas bypass rate vs. front-zone reaction furnace temperature in a SWAG-processing SRU.

The rear zone of the furnace must also be hot enough to destroy the BTEX in the bypassed acid gas. This requires a  minimum burn temperature of 1,050°C,21 but this threshold was determined using straight-through reaction furnaces only with sufficient residence time, which is another serious process consideration. Conservatively, hotter operation of the second zone is recommended, but BTEX destruction can only be safely confirmed by field analytical testing.

The impact of biofeed co-processing on SRU reaction furnace temperature, and the effect of mitigation methods (such as natural gas co-firing, oxygen enrichment and operation with a split-flow furnace) are discussed further in this article’s case study.

Although more CO2 in the feed(s) will form more carbonyl sulfide (COS) in the reaction furnace, losses associated with this species can be managed in the first Claus converter with temperature control, catalyst choice and catalyst activity. Similarly, more CS2 is formed when co-firing, which is also managed in the first converter. For the many facilities that have an amine-based TGTU after the Claus section, virtually all losses from COS and CS2 will be eliminated, as these species will hydrogenate to H2S and be recycled to the front of the SRU.

Impacts on the TGTU

Today’s strict SO2 emissions requirements are beyond the capabilities of the Claus process. The tail gas cleanup unit is a common addition to adhere to the new guidelines. Tail gas units include processes such as sub-dewpoint, direct oxidation, hydrogenation with amine absorption and stack SO2 recovery. The Claus tail gas hydrogenation process followed by amine absorption—called a TGTU—is the most popular choice when more than 99.5% sulfur recovery efficiency is required.22 The impact of biofeed co-processing on TGTUs is important not only due to its prevalence in refineries, but also because the acid gas from the regenerated amine is recycled to the reaction furnace. Alternate tail gas processes, such as sub-dewpoint or direct oxidation, are not discussed in this article because the changes brought about from biofeed co-processing are minimal, apart from those that have been discussed in the SRU section.

In a TGTU, the Claus tail gas is first reheated and processed in the hydrogenation reactor in which all oxidized sulfur species (SO2, COS and CS2) and the remaining elemental sulfur are hydrogenated or hydrolyzed to H2S on the surface of cobalt-molybdenum catalyst. The hydrogenated process gas is then cooled through an exchanger and water quench column and then treated in the amine absorber. The rich amine is pumped to the regenerator, where the acid gas species are stripped from the solvent and recycled back to the SRU reaction furnace. The TGTU amine is typically a selective amine (most commonly MDEA) that is optimized to minimize CO2 absorption, which minimizes the CO2 recycled to the SRU.

With biofeed co-processing, the composition of the tail gas of a well-operated Claus unit is expected to be like the original case, except for an increase in CO2. While the extra CO2 will pass through the hydrogenation section of the TGTU, the increased CO2 in the feed will increase its pickup by the amine, and the rich amine loading and CO2 recycle will, therefore, also increase. To maintain optimum performance, the amine circulation rates or amine strengths will need to be adjusted or a switch to a more selective amine chemistry may be required.

Installing acid gas enrichment (AGE)

To maintain hydrotreater catalyst activity, a certain amount of sulfur is needed in the feedstock. This is primarily of concern in the HVO production route, where there is not always sufficient sulfur naturally in the hydrotreater feed to do this. This can be overcome by the addition of a sulfiding agent, such as dimethyl disulfide (DMDS), to the hydrotreater feed; however, continuous addition can be expensive, and sulfiding agents are often difficult (and unpleasant) to store and handle. One solution is to install an AGE unit that can concentrate the H2S from the amine process and recycle it to the hydrotreater, keeping the catalyst active. The installation of an AGE unit design is a decision that can be advantageous for some facilities. More facility details on where AGE systems have been installed, as well as the rationale for their installation, are provided in literature.23,24


The fictional 100,000-bpd Experts Refinery has been built for maximum diesel yield, and generally processes high-sulfur crudes. The process conditions are loosely based on the field test results of several operating refineries optimized for diesel yield. The SWS and amine plants were simulated on a proprietary gas treating simulation toola, while the SRU and TGTU were simulated on a process simulation softwareb. The authors have found that these two programs yield the best accuracy of the respective simulated units.

The Experts refinery has two diesel hydrotreaters (cracked and straight-run) and co-processes 10% biofeedstock in both. This has an impact on the treating units in several ways. A comparison of normal and co-processing operation is provided in TABLES 4–7.

The key parameters for the SWS unit are shown in TABLE 4. In this case, a stripped water specification of 50 ppmw ammonium ions is desired. To meet this specification with biofeed co-processing, stripping steam was increased by 23%. Much of this steam was required to process the additional water volume. The additional steam also required more reflux condenser capacity with the associated condenser duty to reduce the SWAG temperature to 85°C, which had increased by 25%. While many facilities will not have this ability, this duty can be offset to an extent by allowing the overhead temperature to increase, provided there is hydraulic capacity in the downstream SRU to handle the additional water vapor in the SWAG stream. In practice, capacity in the water feed-effluent exchanger should also be verified prior to co-processing.

With less residence time in the upstream separator vessels on the diesel hydrotreaters (due to significantly increased water being generated in these units), there will be more hydrocarbons in their sour wastewater. A liquid-liquid coalescer may be required on this stream. The impact of additional organic acids in the sour water due to co-processing (e.g., phosphates) was not considered. In practice, if the level of additional organic acids is significant, caustic can be injected to neutralize them.

The refinery’s main amine system uses DEA, which is common in many refineries. TABLE 5 details the impact of co-processing on key parameters. The table includes a case where the DEA solvent is replaced with an MDEA solvent, which preferentially absorbs H2S over CO2.

Typically, in amine units, the limits on the exchanger duty design are reached before system hydraulics. The extra CO2 absorbed by the DEA, with co-processing, requires more reboiler duty (50 psig stripping steam in this study) to regenerate the amine and, hence, additional condenser duty in the top of the stripper column. In this case, 4%–5% more reboiler and condenser duty were needed. This should not be a problem for systems not near capacity; however, many older amine systems are already operating at capacity—often because utilities have not been expanded at the same rate as refinery capacity—and, for these systems, some equipment capital investment may be required. Alternatively, the solvent could be changed to MDEA.

Switching to an MDEA solvent has advantages in terms of energy consumption and improved H2S:CO2 ratio in the AAG, which limits the need for downstream modifications of the SRU. However, this is not the only consequence of changing to MDEA, and the effects on other units in the refinery need to be considered before a system-wide conversion.

The refinery’s SRU, which processes both SWAG and AAG, was modeled under two different cases: a two-stage Claus only and a two-stage Claus with an amine-based TGTU. The results are provided in TABLE 6. All cases used DEA in the upstream main amine system. The most obvious change in the SRU is the increased CO2 concentration in the feed acid gas when co-processing. The volumetric flow through the SRU increased by 2%–3%, which is minor. The Claus section’s sulfur recovery efficiency declined by 0.1%, mainly due to the decrease of H2S in the acid gas and the cooler reaction furnace.

As expected, an increase of CO2 in the feed reduced the reaction furnace temperature. It should be noted that the temperature presented in TABLE 6 is an adiabatic temperature and does not consider heat losses, which means that the actual temperature of the furnace during co-processing is expected to be less than 1,250°C. As discussed in the SRU section, maintaining the reaction furnace front zone temperature at, or hotter than, 1,250°C is critical for complete destruction of feed contaminants.

If the reaction furnace is equipped with a two-zone chamber, then the acid gas bypass fraction to the rear zone can be increased to maintain the front-zone hotter than 1,250°C. In this case study, the bypass fraction must be between 10%–50% of the inlet acid gas flow for maximum temperature to destroy the contaminants (FIG. 11). The upper limit is based on typical refractory thermal limitations.

As discussed in the SRU section, if the furnace has a straight-through design, the furnace temperature can be managed with oxygen enrichment or co-firing. FIG. 12 shows the reaction furnace temperature profile for a combined acid gas feed (main amine system and TGTU recycle streams), which contains 22.4% CO2 because of co-processing. In this case, the model shows that the minimum required temperature of 1,250°C in the reaction furnace can be easily achieved with low-level oxygen enrichment without other changes.

FIG. 12. Oxygen enrichment concentrations vs. front-zone reaction furnace temperature.

In the case of co-firing, it was assumed that the fuel source is a pure methane natural gas. Simulation shows that the introduction of the co-processed acid gas increased the volumetric flow in the SRU by 2.3%, which should be accommodated in the existing plant. However, every 5 kmol/hr increase of natural gas, which is approximately 3% of the feed acid gas, resulted in an 8% increase in combustion air demand (or air flow requirement). This is equivalent to a decline in SRU sulfur inlet capacity of approximately 6%. For this case study, the required temperature in the reaction furnace was achieved by introducing natural gas at 8% of the feed acid gas—this caused an 18% reduction in SRU capacity. This value will vary depending on the actual feed composition, operating parameters and capacity. FIG. 13 shows the impact of the natural gas co-firing on SRU capacity and air demand.

FIG. 13. Natural gas co-firing vs. the impact on air demand and SRU capacity.

The modeled TGTU amine system uses 40 wt% MDEA with a standalone regenerator. Although the optimized CO2 can reach 85% slip, factors such as solvent over-circulation or the formation of primary and secondary amines from MDEA degradation will increase CO2 pickup in the amine absorber. Therefore, the TGTU is modeled assuming a more conservative 70% CO2 slip. The effect of the TGTU amine unit performance on the reaction furnace temperature is demonstrated in FIG. 14. As expected, increasing CO2 slip correlates to less CO2 in the combined AAG (and total feed gas) and a hotter reaction furnace.

FIG. 14. The effect of CO2 slip on reaction furnace temperature and inlet CO2 concentration.

The TGTU amine absorber inlet contained 70% more CO2 in the co-processing case than it did in the crude oil-only case. This resulted in greater rich loadings and more H2S leaving in the absorber vent gas. The rich loading could be reduced by increasing the amine circulation or by increasing the amine strength, or both. Increasing the circulation requires more energy consumption to circulate and regenerate the amine. Increasing the strength requires more amine inventory and additional energy to strip the amine. TABLE 7 shows the optimized case with increased amine concentration, which required 17% more steam to regenerate the amine. It should also be noted that, when operating with high amine concentrations, the amine losses will also increase. Therefore, when optimizing, operators should consider their requirements and available resources (such as pump capacity, amine inventory and steam capacity).


The impact of biofeed co-processing on downstream treatment units can be quantified with a well-designed test run, appropriate gas and liquid analytical capabilities, and reliable process simulation. Based on this, a fit-for-purpose treating scheme that minimizes capital and operational expenditures and prevents unanticipated reliability issues can be designed and implemented.

A major component in the changes associated with biofeed co-processing is the additional water produced. This will usually require modifications or increases to the SWS system. If these are intelligently implemented, refinery freshwater requirements can often be significantly reduced. The additional water volume increases the unit’s stripping steam requirement and will likely reduce the residence time of the water-liquid hydrocarbon separation vessels, which potentially will leave more liquid hydrocarbon in the wastewater. The total hydrocarbon content—or worse, a fluctuating content—can have detrimental effects and must either be minimized or mitigated.

Another major process change associated with biofeed co-processing is an increase in total system CO2. The main amine system is generally able to handle the additional CO2, although an amine system already at its limits may require additional (usually heat exchanger) equipment installation to manage the extra CO2 picked up by the amine solvent. In some instances, changing the solvent can provide the additional capacity needed without capital investment, and can offset the impact of additional CO2 on the downstream SRU by slipping more CO2 into the product hydrocarbon streams. Depending on the co-processing methodology, foaming and emulsification issues can be severe, and normal management techniques such as anti-foam injection and carbon bed filtering may be insufficient.

The increased CO2 concentration in the combined sour feed gas to the SRU should not be a problem for already-rich feeds (> 90% H2S). However, for leaner feeds, care should be taken to ensure that the reaction furnace temperature remains at a minimum of 1,250°C to avoid contaminant breakthroughs. This can be done by either co-firing with a fuel gas of stable composition, installing an oxygen enrichment system where burner turndown capability allows, or by increasing the bypass fraction in the case of a split-flow furnace design.

As was the case with the main amine system, the absorber of an amine-based TGTU will encounter more CO2. Amine circulation rates, amine strength and solvent selection can be adjusted to increase the CO2 slip in the absorber, thereby preventing it from diluting the SRU feed stream, while holding the H2S concentration in the absorber vent gas to specification. Nonetheless, the extra CO2 will also require more duty from the TGTU regenerator reboiler and reflux condenser.

Understanding how biofeedstocks interplay with existing refinery infrastructure is critical for achieving smooth, easy and cost-effective co-processing. The effects and solutions listed in this article should serve as a guideline for determining the key parameters to monitor at a facility, along with ways to minimize operational issues as the world moves toward more green fuel sources. HP


a Optimized Gas Treating’s ProTreat™
b Aspen Technology’s Aspen-HYSYS™
This work was first presented at the 28th European Biomass Conference & Exhibition (EUBCE 2020).


The authors would like to thank MPR Services for sharing information on FCCU breakdown products, Jeroen Engels for his assistance with graphics and formatting, and Peter Seville for the proofing of this paper.


18. Paskall, H. G., “Capability of the modified-Claus process: A final report to the Department of Energy and Natural Resources of the Province of Alberta,”          Western Research and Development, 1979.
19. Klint, B. W. and P. R. Dale, “Ammonia destruction in Claus sulfur recovery units,” Laurance Reid Gas Conditioning Conference, 1999.
20. Goar, B. G., W. P. Hegarty and T. W. Thew, “How to cope with your sulfide problems (COPE process; use of oxygen enriched air to increase capacity),”            U.S. Department of Energy Office of Scientific and Technical Information, 1986.
21. Klint, B. W., “Hydrocarbon destruction in the Claus SRU reaction furnace,” Laurance Reid Gas Conditioning Conference, Norman, Oklahoma, 2020.
22. Keller, A., E. Nasato, B. Welch and G. Bohme, “Fundamentals of sulfur recovery,” Laurance Reid Gas Conditioning Conference, Norman, Oklahoma, 2021.
23. Lambrichts, J., “Hydrogen sulfide recovery at the Eni Porto Marghera Green refinery using UCARSOL solvent acid gas enrichment (AGE) technology,”              International Refining and Petrochemical Conference (IRPC), 2014.
24. Van Son, M., “Biofuels integration with refinery sulfur complex,” REFCOMM Conference, 2021.

The Authors

Related Articles

From the Archive



{{ error }}
{{ comment.comment.Name }} • {{ comment.timeAgo }}
{{ comment.comment.Text }}