April 2026

Process Optimization

The use of mercaptide sensors to increase safety, increase unit health and mitigate downtime

IP: 10.1.228.95

Caustic removal of mercaptan species from hydrocarbon streams has been known for centuries. However, the extent of mercaptan removal depends on the solubility of the mercaptan species in the aqueous phase and the equilibrium between the weak acid formed when the mercaptan ionizes and the strong base in aqueous solution. Applying this process to commercial scale required large volumes of caustic used once through to maintain the desired amount of mercaptan removal. During the 1950s, the co-author’s company invented and patented the Merox process, which features a caustic regeneration system, vastly reducing the quantity of fresh caustic consumed and the associated waste caustic disposal.  

The key to the success of the Merox process is its ability to use high-activity catalysts to efficiently oxidize sodium mercaptide produced in the extraction step to easily separated disulfides. This results in regenerated caustic (lean caustic), with a residual sodium mercaptide level low enough to ensure the desired level of mercaptan extraction when contacting the hydrocarbon again. Poor regeneration leads to higher levels of sodium mercaptide in the lean caustic, which results in insufficient mercaptan extraction from the hydrocarbon due to equilibrium limitations. Given the criticality of the lean caustic sodium mercaptide content, a field test was developed to gage the state of the regenerated caustic and give a loose correlation with its mercaptide concentration. This test is called the shake test and has been the leading indicator of regenerated caustic health for the past 70 yrs. 

The catalyst used in the regeneration section is dark olive green in the reduced state. Even though the caustic has been regenerated, it is not fully oxidized and will remain dark green until it is fully oxidized and becomes intense blue. The operator takes advantage of this color change to indicate the degree of regeneration by taking a 250-ml (about 8 oz) sample of hot lean caustic in a 500-ml (about 16 oz) clear glass bottle, then simultaneously starting a timer and shaking the sample to determine if the color change takes place between 30 sec and 120 sec. FIG. 1 shows the color change observed in the lean caustic while performing the shake test used in Merox units.  

FIG. 1. The dull green to bright blue color change between 30 sec and 120 sec. Photo courtesy of a North American refiner. 

While the shake test has remained a quick and useful field test, it is a semi-quantitative test that has inherent limitations and drawbacks. There are some safety concerns with shaking 15-wt% aqueous sodium hydroxide at 49°C (120ºF) in a glass bottle for 120 sec or longer. The manually intensive operation typically results in two datapoints per day being taken and can miss changes in operation or mercaptide load, resulting in poor regeneration performance being missed and incorrect conclusions being acted upon. It is worth noting that many refiners do not perform this test at this frequency—or at all—due to this technique not being passed on with knowledge transfer to new employees. This can lead to improper process variable adjustment negatively affecting regeneration performance.  

Various hydrocarbon streams, such as fluid catalytic cracking or coker liquified petroleum gas, may also have color bodies that extract into the caustic, making the caustic very dark upon oxidation and difficult to see the color change. Additionally, some operators may have visual impairments like color blindness, making it very difficult to observe the color change while performing the shake test. Multiple forms of human-introduced variability can also lead to incorrect observations of the shake test:  

• Taking the shake test at the same time every day  

• Performing the shake test less than twice a day or not at all  

• Operators estimating time when a stopwatch is not available 

• Allowing time to lapse between sampling and performing the shake test, biasing the results by:  

• Smaller samples than recommended being taken, resulting in thermal instability due to small thermal mass  

• Allowing the sample to cool or warm, affecting the reaction rate  

• Mercaptide oxidation occurring with oxygen in the air even though it is not being shaken 
• Exposure to sunlight, thus catalyzing the reactions and generating side reactions. 

These scenarios can result in production of off-specification products that must be rerun or devalued and sold as another product. To minimize these losses, operations will often initiate an unplanned caustic replacement in response to return the product back to specification; however, the off-specification product still affects the downstream units, resulting in increased chemicals consumption or complete catalyst replacement. An off-specification occurrence for a 10,000-bpd extraction unit can result in $250,000 in lost product value and up to $50,000 in caustic purchasing/disposal. Effects to downstream units can result in > $100,000 in chemicals consumption and > $1 MM for catalyst replacement.  

To address these concerns, the co-author’s company developed a novel technology consisting of a sensor and algorithm (mercaptide sensors) to directly measure mercaptide concentration in the rich and lean caustic, resulting in greater insight into the plant’s operation.  

CURRENT VERIFICATION OF CAUSTIC REGENERATION  

The use of mercaptide sensors technology to identify caustic regeneration. By implementing new mercaptide sensors, refiners can realize improved operation via greater process insights from continuous data, resulting in increased on-specification production and less waste caustic generation. The online mercaptide sensor directly measures mercaptide concentration. This enables operators to make rapid corrective actions to the unit process variables (e.g., catalyst adjustments and increased air sent to the unit, thus improving regeneration of spent caustic). The use of these sensors ensures quantitative, consistent and timely results, as well as increased safety by reducing caustic handling to once every two weeks, and no shake test is needed.  

The implementation of mercaptide sensors as a technology was initially installed and tested at the Pembina Redwater Alberta, Canada site, as these sensors have demonstrated a year-minimum lifespan within Merox caustic fluids. FIG. 2 shows the field-testing timeline of these sensors at Pembina’s site and highlights technology improvements. 

FIG. 2. Development of sensor technology over time.  

Integration of mercaptide sensors to refinery technology. To achieve the advanced monitoring described, the sensors were installed on both the rich caustic stream from the extractor and the lean caustic stream from the disulfide separator. A system with two extractors and a common shared regeneration system requires a rich caustic sensor in the rich caustic line from each extractor and one in the combined rich caustic line by the caustic heater for full functionality. The raw data from these sensors were then sent to the co-author’s company through a digital connection, along with other process and lab data, where they were then transformed via a confidential algorithm and reported back to the refiner through the co-author’s company’s digital platform’s^a graphical user interface (GUI). FIG. 3 shows a typical process flow diagram highlighting the approximate installation locations of the mercaptide sensors (in green) and the key process variable (in red) for the Merox regeneration system.  

FIG. 3. Merox extraction process, with the installation of new mercaptide sensors.  

The sensors’ data, now transformed into mercaptide sulfur values, can be used in conjunction with typical key process variables of the Merox unit [these are typically known as COACH variables (catalyst, oxygen, alkalinity, contact and heat)] to determine the root cause for poor regeneration and avoid off-specification product and unnecessary caustic and catalyst replacement. When compared to the variability and frequency of the shake test, the continuous mercaptide data from these sensors deliver vastly improved fidelity and reliability, resulting in earlier detection and rapid response times to unit disruptions. 

The implementation of this technology will grant operational benefits to customers, including increased data frequency of mercaptan concentration. The mercaptide readings from the sensors will allow unit optimization and tighter control of key process variables, including catalyst injection rates, air injection, caustic circulation and water/caustic makeup. Occupational safety is promoted, as there are vastly reduced health, safety or environmental (HSE) concerns from exposure to toxic chemicals.  

Installation at Pembina’s facility. The mercaptide sensors and the associated sampling systems can be installed inline (requiring a shutdown) or while onstream, utilizing a slip stream from the higher-pressure area to the low-pressure area. The sensors were installed utilizing a slip-stream configuration for the testing at Pembina and were implemented while the unit remained online. FIG. 4 exemplifies the use of implementing slip-stream installations, where the rich caustic mercaptide sensors were installed across the inlet and outlet of the heater (on the left side), and the lean caustic mercaptide sensors were installed from the low-point drain on the outlet of the caustic circulation pumps to the bottom “broken ring trap” nozzle of the oxidizer (on the right side). 

FIG. 4. A red-lined process flow diagram showcasing the new installations that were made for a customer while the unit was running.  

The mercaptide sensor’s benefits are further highlighted by the ease and simplicity of their maintenance and replacement, once the analyzer is powered down. Operators can quickly retract the sensor into the upper section of the high-pressure retractor housing, isolate it from the process, drain and flush with water, replace the sensor and reinstall. FIG. 5 shows an example of a sensor replacement. 

FIG. 5. A Pembina operator replacing a second-generation mercaptide sensor. 

Benefits of mercaptide sensors and the digital platform^a in unit troubleshooting. In the wake of higher-than-normal summer temperatures at their Redwater facility in Alberta, Pembina had seen higher than usual temperatures in the feed to its Merox unit. To offset decreases in extraction efficiency resulting from high feed temperatures, operators had to reduce the feed rate in the daytime and increase the feed rate during cooler temperatures at night. Mercaptan load to the regeneration section was shifting due to the temperature and feed rate variations, and residual mercaptide in the lean caustic was building as regeneration was falling behind. However, the low frequency of the shake test, along with variability, did not detect the trend toward off-specification product, ultimately causing the product to go off-spec.  

Further investigation with the digital platform’s^a visualization interface revealed mercaptide sensors showing the oscillation in both the lean and rich caustic streams, along with the upward increasing trend in lean caustic mercaptide concentration. FIG. 6 visualizes the off-spec incident that occurred in July 2023, when the mercaptans reached above the 120 wppm specification mark, based on the co-author’s company’s offline algorithm for mercaptan concentrations.  

FIG. 6. Shown in the digital platform’s^a GUI. For mercaptide concentrations, green signifies optimal parameters, yellow signifies a warning but within expected parameters and red signals that the high/low limit has been exceeded. 

Given that new engineers are typically assigned to the treating units and operations personnel are being cross functionally trained, there is a decrease in experience on these units across the industry. To help with interpretation of the data shown in FIG. 6, the displayed GUI seeks to help by using a stoplight-like color scheme to visually show “good” “caution” and “bad” conditions. Additionally, dialogue boxes provide guidance and recommendations on what the data means, along with some predictive indicators describing where the data is trending. General recommendations are also provided at this time, with reference material links to the Merox knowledge base to help with training and interpretation. 

Two other off-spec events occurred at the customer’s site in January and February of 2024. In a January 2024 plant air trip, the use of mercaptide sensors showed increasing sulfurs which tracked alongside the product sulfur analyzer, showing a peak and then a gradual recovery in the unit (FIG. 7). Interestingly, this is the type of data that would not be shown by a shake test and would likely result in drastic action like caustic and catalyst replacement. In this case, Pembina was able to ride out the “blip” and wait for the regeneration to catch back up.  

FIG. 7. Two separate distributed control system trend images of unit upsets (the top from January 2024, with the bottom from February 2024), showing how the mercaptide sensors aided in reducing any potential downtimes for the customer.  

In February 2024, data had shown that mercaptides were trending up but then corrected back to the optimal levels. A large caustic changeout was being performed on the unit that diluted the catalyst concentration. The large volume of the caustic being replaced resulted in low catalyst concentration, stalling the oxidizer reaction, which the mercaptide sensor showed as an uptick in the lean caustic mercaptide levels. 

Takeaway. As evident from the mercaptide sensor at Pembina’s facility, the mercaptide sensors are now a leading indicator that provide operators and engineers previously unavailable insight to the operation of the Merox unit, allowing greater understanding of off-specification root causes and helping to provide guidance to prevent these events. The enhanced precision and recording of live data provided by these sensors reduce HSE concerns on operations, while saving anywhere from $25,000–$130,000 per off-spec incident. Based on the value added to data presentation and troubleshooting, the customer is adding these sensors to all Merox units at its Alberta facility. 

NOTE  

^a Honeywell Forge 

The Authors

From the Archive

Comments