During the 2018 American Fuel and Petrochemical Manufacturers Operations & Process Technology Summit (formerly Q&A and Technology Forum), Oct. 1-3, 2018, in Atlanta, Ga., US domestic and international refiners addressed questions about hydroprocessing operations selected and answered by industry experts from refining companies and other technology specialists.
The respondents
Robert Steinberg, Motiva Enterprises LLC
John Kulach, Honeywell UOP LLC
Richard Todd, Norton Engineering LLC
Jeff Caton, Axens North America Inc.
Jessica Schlicting, Criterion Catalysts & Technologies LP
Dennis Haynes, Nalco Champion
Brandon Miller, Criterion Catalysts & Technologies LP
Raul Romero, Nalco Champion
This annual meeting addresses real problems and issues refiners face in their plants and provides an opportunity for members to sort through potential solutions in a discussion with panelists and other attendees.
This is the first of three installments based on edited responses in the 2018 official answer book. Part 2 in the series (OGJ, Oct. 7, 2019) will highlight discussion surrounding gasoline processes, while the final installment (OGJ, Nov. 4, 2019) will focus on fluid catalytic cracking (FCC).
The only disclaimer for respondents was that they discuss their own experiences, their own views, and the views of their companies. What has worked for them in their plants or refineries might not be applicable to every situation, but it can provide sound guidelines for what would work to address specific issues.
What factors do you consider when coprocessing jet fuel in a distillate hydrotreater vs. processing the jet fuel separately (including feedstock and unit consideration)?
Steinberg: There are several considerations when deciding if jet fuel and diesel should be coprocessed or hydrotreated separately. The most important consideration is if jet fuel will be blended into the diesel product or if separate products are desired. The decision may depend on if you’re looking at constructing new facilities or making the best use of existing equipment. If a refiner needs to build a new unit to increase distillate hydrotreating, building a jet fuel unit and using existing units for diesel will normally be less expensive than building a new diesel hydrotreater. If more capacity is required than jet fuel supply can satisfy, some straight-run diesel can be blended into the jet fuel hydrotreater. If the existing unit is too low a pressure to handle the diesel effectively, however, it may be preferable to build a new unit for the most-difficult-to-treat streams such as light cycle oil (LCO) or light coker gas oil instead.
If a new refinery is to be built and new jet fuel and distillate hydrotreaters are needed, it would be simpler and cheaper to build a single unit and coprocess the jet fuel with the diesel. Exceptions include if capacity would be too large for a single unit, or separate jet fuel and diesel products were desired. For a refiner looking to optimize existing operations, it’s assumed that a separate jet fuel product isn’t required. If it was required, feeds would need to be processed separately unless there was a fractionator on the back end to make jet fuel and diesel cuts. Having such a fractionator is probably unusual and would require a lot of energy to vaporize the jet fuel to remove it from the diesel.
For a refiner with only a single hydrotreater, this becomes a question of batch processing diesel and jet fuel or blending them and processing together. If the refiner has multiple hydrotreaters, it will generally make sense to send the easiest-to-treat streams (such as jet fuel) to the lowest-pressure or mildest hydrotreater, or blend just enough of the easiest streams into the more-severe unit to let it achieve its desired run length. Alternately, batch processing jet fuel and diesel in the same unit instead of coprocessing wouldn’t generally be recommended unless there were special circumstances. This would be more complicated, as it would require frequent switching of feeds and changing operating conditions. During such changes, lower charge rates may be needed to reduce the risk of making off-spec products or having to pull more naphtha out of the stripper than otherwise required to maintain product flash point. Diesel requires more severe reactor conditions than jet fuel to meet the same sulfur specification.
Mixing jet fuel into the distillate hydrotreater lowers the average boiling point of the feed as well as feed sulfur and nitrogen content. This means lower reactor temperatures and less chemical hydrogen consumption, less treat gas needed to maintain the desired ratio of hydrogen availability to consumption, smaller exotherms, and less quench gas. The lower start-of-run temperature can extend the catalyst run length. Pressure drop for jet fuel and diesel in the same hydrotreater would be similar if the same amount of hydrogen circulation was used: more of the jet fuel would vaporize and increase velocity, but the diesel—which is more viscous—has more mass at the same barrel-per-day charge rate. Less treat gas is needed for jet fuel due to the lower chemical hydrogen consumption, however, and exotherms are generally low enough with jet fuel to not need any quench. This effectively means less hydrogen circulation with jet fuel and lower pressure drop in the reactor, exchangers, and furnace. Blending jet fuel into the diesel hydrotreater will reduce pressure drop, and if the run length is limited by pressure drop, this can extend catalyst life.
Another consideration can be product blending and diesel cetane. If there are multiple units, or jet fuel and diesel are batch processed, some products may not meet sulfur, cloud point, cetane, or other product specs. In some cases, products from different units can be blended to make an on-spec ultralow-sulfur diesel (ULSD). While this complicates operations, adjusting the amount of jet fuel that’s blended into each unit can help keep all products on-spec and reduce the risk of having to deal with an off-spec product tank. That said, jet fuel has a lower boiling point than diesel and thus a lower cetane number. If a refiner makes both a higher and a lower-cetane product, it may be necessary to minimize the jet fuel in the high-cetane product; this is like making a separate jet fuel product in that it can require separate processing.
Motiva has a relatively mild hydrotreater that was revamped for ULSD 15 years ago. Initially, it charged mostly straight-run diesel to make ULSD. Later, when the refinery was expanded, a new diesel hydrotreater was built, and the additional jet fuel was charged to this unit. The unit now charges only jet fuel and operates in ultralow-sulfur kerosine (ULSK) mode; the product has less than 5 ppm sulfur and can be blended into either jet fuel or diesel as desired. The unit has minimal catalyst aging, illustrating both that jet fuel can be the easiest product in the refinery to hydrotreat giving very long catalyst life and that adding jet fuel to a diesel hydrotreater results in milder reactor conditions and longer catalyst life.
Motiva normally produces more kerosine in our crude units than we have capacity for in our jet fuel hydrotreaters. The surplus kerosine is mixed into our diesel hydrotreaters. This is the easiest feed the diesel hydrotreaters process, with more kerosine reactor temperatures able to be lowered while maintaining product sulfur. While this reduces the required reactor severity, there are some drawbacks. For example, the lighter feed reduces the delta API gravity between feed and product, meaning volume swell is reduced. Also, the lighter feed reduces the density of feed going through the charge pump as well as the discharge pressure that the pump can produce. One of the diesel hydrotreaters sometimes needs to lower operating pressure to maintain charge rate.
Motiva normally hydrotreats as much kerosine as there is capacity for in our jet fuel hydrotreaters and blends the remaining kerosine into diesel hydrotreaters. One of the jet fuel hydrotreaters swings between sending its product to jet fuel and diesel but operates to be on-spec for both products. We don’t have any units where we switch feeds or operating conditions to sometimes make jet fuel and sometimes diesel.
Kulach: Considerations for coprocessing jet and distillate in the same unit are the feed rates and feed quality, which go into the selection of operating pressure, space velocity, and catalyst. Distillate hydrotreating for ULSD typically requires higher hydrogen partial pressure, lower liquid hourly space velocity (LHSV), and more active catalysts compared to treating jet because of the need to convert stable sulfur compounds such as benzothiophenes and dibenzothiophenes. If distillate feed includes coker gas oils, LCO, or other difficult-to-process streams such as extracts and condensates, the design would call for a more severe operation with higher reactor temperatures. Inorganic contaminants such as silica and arsenic can be removed from the feed upstream of the active desulfurization catalysts by using filters, particulate traps, and demetalization catalysts in the top of the reactor.
Jet hydrotreating requires enough catalyst and hydrogen partial pressure for mercaptan sulfur and total acid number (TAN) removal. This is usually a low-severity operation compared to distillate hydrotreating, as jet fuel can be very color sensitive, exasperated by high reactor-temperature operation. On the other hand, jet fuel hydrotreating can require more severe operation if the feed contains high aromatics or naphthenes that will require some saturation to meet composition and combustion specifications such as aromatic content and smoke point. In some cases, this might require a noble-metal catalyst or a second-stage operation.
While a typical distillate-hydrotreater fractionation section would consist of a stripper column to remove light ends and meet flash-point specification, a hydrotreater coprocessing distillate with jet feed would require a more complex fractionation section. The fractionation configuration would depend on the relative rates of distillate and jet as well as the need to meet jet-fuel volatility and fluidity specifications such as distillation and freeze point. Whether a coprocessing unit is more economical than individual diesel and kerosine hydrotreating units (DHT, KHT) will depend on the relative feed rates, feed quality, and product specifications. Coprocessing might be favored if the DHT feed is relatively easy (e.g., a straight-run diesel) and if the jet feed is more difficult to treat. The key takeaway is that jet fuel processing is very dependent on feed quality and required specifications.
Todd: Coprocessing jet fuel in a distillate unit may result in poorer unit performance than expected due to higher vaporization of the jet fuel components, which in turn causes lower H2 partial pressure. This varies based on, among others, unit conditions, H2 partial pressure, and LHSV. Heavier distillate desulfurization is normally controlling, so reactor temperatures are set by the distillate requirements. Cracked stocks in the jet-fuel range may have higher olefins increasing H2 consumption, and again in turn, decreasing H2 partial pressure, causing an increase in deactivation.
What considerations do you give to coprocessing or block-mode operations with renewables in an existing hydroprocessing unit?
Caton: Unfortunately, the coprocessing incentives under the Renewable Fuel Standard (RFS) have not been strong enough to lead to mass adoption of coprocessing renewable feedstock by US refiners. As these incentives become more clearly defined, there is little doubt refiners will each adopt a coprocessing vs. block-mode strategy. In all scenarios, a refiner must consider their RFS obligations; potential biodiesel tax incentives; biodiesel merchant market; existing hydrotreating units’ utilization rates, capabilities, and designs; renewable feedstock types; availability, price, and pretreatment requirement; and product specifications and storage constraints.
The primary benefit to full-time or block-mode operation with 100% renewable feedstock is the production of 100% renewable diesel (R100). R100 can be blended with conventional diesel to meet the refiner’s RFS obligations or could be sold to others for their blending requirements. R100 production also can be attractive in terms of a potential biodiesel tax credit. Unfortunately, this biodiesel tax credit is approved on a retroactive basis, and approvals haven’t been timely, consistent, or predictable. As the long-term validity is in question, it does make it difficult to justify economics based on the biodiesel tax credit.
For a refiner who has an unused or underutilized hydrotreating unit, full-time or block-mode operations with 100% renewable feedstock in this asset may be very economical. An inherent benefit of block-mode operations is the ability to switch processing from block mode with 100% renewable feedstock to 100% conventional diesel-kerosine feed mode (or even to a coprocessing mode) depending on the market environment. Note that throughput capacity of the unit in 100% renewable feedstock mode may be limited to 10-30% of the unit’s nameplate capacity when in conventional diesel-kerosine hydrotreating mode, which is attributable to the high exotherm and hydrogen consumption when processing 100% renewable feedstock. To deal with these challenges, a large amount of product will typically be recycled from the back-to-front as liquid quench, thus limiting the fresh-feed capacity to 10-30%. In addition, the need for carbon monoxide (CO), carbon dioxide (CO2), light ends, and water removal must be considered. Finally, protection against carbonic acid corrosion through metallurgy upgrades may be required. Segregated storage of R100 may be required, so storage and handling capabilities should be considered.
When a refiner doesn’t have an unused or underutilized hydrotreating unit that can be dedicated to full-time or block-mode operations with 100% renewable feedstock, coprocessing with conventional diesel or kerosine may be a feasible solution in meeting the refiner’s RFS obligations. It may be possible to coprocess anywhere from a few percent upwards of 30% of renewable feedstock depending on the renewable feedstock type and unit capabilities.
Schlicting: Renewable feedstocks (e.g., vegetable oil, animal fats) can be coprocessed with traditional diesels in existing ULSD units. Coprocessing has some distinct advantages over building a dedicated renewables processing unit because it’s a low-capital expenditure (capex) option that may allow refiners to obtain credits as part of the RFS program.
There are some unique considerations that should be made if renewables are coprocessed in an existing unit. The following is a list of some of the differences between renewables and traditional crude oil-derived feedstock and resulting evaluations that may need to be done before coprocessing this material:
- Oxygen content. The oxygen content of vegetable oil and animal fats is high, making this an important consideration in coprocessing. By-products of the deoxygenation reactions are water, CO2, CO, and methane. The presence of these byproducts in the reactor system can reduce H2 partial pressure and impact recycle-gas compressor operation. The relative number of by-products produced is heavily influenced by hydrogen sulfide (H2S) partial pressure, type of catalyst, and fraction of renewable feed in total feed. Water is generated through the deoxygenation reaction and will be present in the reactor effluent system. Presence of the water will elevate the reactor-effluent dew point and can result in increased chloride corrosion upstream of the water-wash injection point.
- Olefin content. Vegetable oil and animal fats are also very olefinic. This will result in higher hydrogen consumption and heat release, which will increase with increased percentages of renewable feeds processed. As with any olefinic feed processed in a hydroprocessing unit, this feed tends to polymerize, depending on the temperature profile in the unit and reactor, so this should be considered.
- Earth alkaline metals, phosphorus (P). Vegetable oils have a higher content of metals not commonly encountered in petroleum-derived feedstock, such as sodium, potassium, calcium, and magnesium. Vegetable oils and animal fats contain P, a known catalyst poison, which can have a major impact on catalyst activity. The impact of these contaminants on catalyst activity should be discussed with the catalyst provider.
- Chlorides. In addition to other contaminants, vegetable oils and other animal fats can introduce chlorides to the reactor system. This is a potential corrosion concern in the unit which should be evaluated.
- Product properties. Because of the paraffinic nature of the products of coprocessing vegetable oils in hydroprocessing units, the cold-flow properties (pour point, cloud point) of diesel product are typically higher when compared to these properties in diesel product without coprocessing. As the fraction of renewables in feed increases, the resulting product cold-flow properties likely will increase.
- Renewable feedstock storage. Vegetable oil and animal fat feedstock may turn rancid if stored in tankage too long, especially if ambient temperature is high. If a refinery expects to store the feedstock in tanks on site, the feed may need to be deodorized using bleach or something like it to prevent foul smells.
Coprocessing renewable feedstock will likely be most successful in hydroprocessing units loaded with nickel-molybdenum catalyst (vs. cobalt-molybdenum catalyst). Because of the increased hydrogen consumption, heat release, increased corrosion potential, and increased required weighted-average bed temperature to meet product specifications while running renewable feedstock, coprocessing less than 10% renewable fuels will likely be most feasible in an existing hydroprocessing unit. These considerations should be reviewed during a management of change or equivalent process evaluation before choosing to coprocess renewable fuels such as vegetable oils or animal fats in an existing hydroprocessing unit.
What are the sources of silicon that can impact a hydrotreater? How does silicon affect hydrotreater operations? What are your best practices for managing-mitigating silicon poisoning?
Haynes: One source of silicon is antifoam chemistries. These may be introduced into hydrodesulfurization feeds via use in the refinery’s coker process or use in production upstream of the refinery. The ability to minimize silicon content introduced due to antifoam within the refinery depends on antifoam-application practices and controls along with antifoam-product selection.
Miller: There are many potential sources for silicon that ends up in a hydrotreater feed stream. Historically, the principal source of silicon to hydrotreaters was the antifoam addition to delayed cokers at the refinery, which results in polysiloxanes being cracked into the naphtha and light-coker distillate boiling range. Over the years, higher-viscosity antifoams have been introduced, resulting in increased silicon levels in heavier coker-product streams. Silicon can also come into the refinery directly with the crude. Coker-derived synthetic crudes, like those from Canada and Venezuela, contain silicon due to the antifoam used during production. Silicon is also added to many crudes via use of flow improvers-drag reducers for increased production capability. Silicon can also be present in some crudes derived from their source in sand or aluminum silicate clay. Many of these silicon compounds tend to end up in the distillate and vacuum gas oil fractions.
Silicon deposits on the surface of alumina-based catalysts; not on the active metals’ sites, but through adsorption of the silicon-containing molecule onto the surface hydroxyl groups of the alumina support. This results in a thickening layer of silicon formed on the alumina surface. Over time, the pathways into the catalyst pellet become blocked, resulting in effective deactivation of the catalyst.
Once the silicon contamination has been minimized at the source, the next step is to mitigate in the hydroprocessing unit. Regardless of the source, effective mitigation can be accomplished with proper operational management and the appropriate catalyst solution.
We need to make sure, however, that we design the right catalyst system to pull the silica out of whatever type of feed we are processing. For example, a catalyst with a high surface area and a small pore diameter can be very effective with lighter feeds, such as coker naphtha, but will not be effective in heavy feeds due to the diffusion limitations resulting in rapid pore-mouth plugging, or egg shelling.
Generally, the catalyst design to handle silicon employs three types of catalyst:
- Silicon trap catalysts, which are no- or low-activity but high surface area, designed to be the main silicon depository for most silicon in the feed.
- Transition-dual function catalysts, which are active but still very silicon-tolerant catalysts designed to uptake a large amount of silicon over the run while still contributing to the activity of the system.
- Main-bed active catalysts, which are the least silicon tolerant but the highest activity. These catalysts finish off the more difficult treating reactions and are protected by the other layers to ensure good activity throughout the cycle as the upper layers become deactivated.
- Catalyst particle size is also an important consideration for silicon uptake. In many units, particularly naphtha units, silicon uptake is diffusion limited, so smaller-sized catalyst will directionally pick up more silicon in each reactor volume. This benefit is often balanced against reactor pressure-drop limitations.
These catalyst types and relative amounts need to be properly balanced in the unit to achieve the desired performance and cycle life. To find the right balance, the silicon uptake and catalyst activity should be monitored throughout the cycle. Silicon uptake can be measured by taking routine feed composite samples and then validated through spent catalyst analysis.
In addition to the proper catalyst system and monitoring plan, reactor temperature is a key factor. Higher temperature generally leads to more silicon uptake but may also lead to poor silicon diffusion and lay down on the catalyst surface, prematurely blocking sites that otherwise could be used and effectively reducing the total uptake potential of a given catalyst. Operating at too low of a temperature may also limit the total silicon capacity of a catalyst. This is often the case for naphtha hydrotreaters, where typical temperature regimes may limit the uptake capacity of the lead catalyst beds. In distillate and heavier service, however, the typical operating temperature regime is high enough to achieve maximum silicon uptake.
In some cases, space velocity may be another lever available to manage silicon uptake. In general, higher space velocity drives less efficient silicon deposition, resulting in a lower overall silicon-uptake capacity of a given catalyst. Space velocity is often something we can’t control, but it comes into play in select situations where there are multiple units of varying space velocity that could process a silicon-containing feed.
Gas oil hydrotreaters must often contend with additional catalyst poisons, like nickel and vanadium. These metals do require active-metals site functionality on the catalyst. Since silicon deposits on the catalyst surface, these metals aren’t necessarily in direct competition for a deposition site. The deposition of these metals, however, can indirectly affect total uptake capacity through diffusion limitations and steric hinderance of the potential sites. These interactions should be understood and considered when designing, comparing, and optimizing a catalyst system, typically with specific unit performance history, if possible.
Romero: A typical source of silicon on streams coming from a delayed coking unit (DCU) is antifoams applied on its drum during filling operation. A more general balance should be considered, including silicon-based antifoam applied upstream of the DCU, like preflash or main fractionators in crude and vacuum units, and crude feed itself. DCU-feed silicon content should be considered along with best practices for antifoam application on drums. Operating conditions—including drum inlet temperature and drum pressure—also play a key role in optimizing antifoam consumption.
