Refiners will find that their choices of process technologies and catalysts will determine their ability to produce ultra-low sulfur (30 ppm) gasoline and diesel.
In particular, producing ultra-low sulfur diesel will be more of a challenge than producing 30-ppm-sulfur gasoline.
Ultra-low sulfur diesel production will require larger investments in high-pressure hydrotreating than will gasoline. Although it will not solve all diesel quality and demand problems, retrofiting existing reactors and choosing the better catalysts will help ease the need for additional diesel hydrotreating capacity.
The prospect of making less-than-30-ppm sulfur streams poses other challenges:
- The possibility of contamination with higher sulfur streams.
- The optimization of the H2S stripper operation so that there is no leakage.
- The contamination by feed and effluent exchanger leaks.
- Finding reliable analytical techniques to measure sulfur.
The sulfur compounds present in the naphtha destined for the gasoline pool are less refractory and easier to remove than the sulfur compounds in diesel.
Naphthas contain, for the most part, sulfur compounds up to and including benzothiophene. About 90% of the total sulfur is thiophenes or lighter compounds. Although only 10% of the total sulfur is benzothiophene or heavier compounds, this could present a problem, depending on the total sulfur level in the stream and the contribution to the gasoline pool sulfur.
Refiners should choose the best sulfur-reduction process from a number of options. Options to treat the gasoline include hydrotreating of the whole cut, caustic washing, adsorption processes, fractional distillation followed by selective hydrotreating of the heavy cut, catalytic distillation, and blending.
Likely, refiners will use one or more of these approaches to reach the low-sulfur gasoline targets. The best route for a particular refiner will depend on the refinery configuration, process units, current pool sulfur, and economics.
One other consideration is octane barrels. Severe hydrotreating to meet low-sulfur levels will result in a considerable loss in octane number in some of the process options.
As an alternative to treating the gasoline product, refiners can choose to apply severe hydrotreating in the fluid-catalytic cracking unit (FCCU) feed pretreater.
Pretreating the feed to the FCCU can reduce the sulfur level of the FCCU naphtha to an acceptable limit. It also improves naphtha yield and quality, reduces SOx emissions, and lowers the sulfur content of light cycle oil.
While many refiners are focusing on large projects to help manage low-sulfur production, a properly designed distributor tray in the hydrotreating reactor can greatly increase the refinery-desulfurization capacity. In fact, a properly designed distribution tray is required to meet high-severity hydrotreater specifications for diesel.
In several retrofitted hydrotreating units, Haldor Topsoe Inc. has seen a 45° F. lower start-of-run (SOR) temperature advantage by replacing the distributor tray. This is equal in performance improvement to adding a new, same-size reactor or halving the liquid hourly space velocity (LHSV). It is considerably less expensive, however, and thus, has a rapid payback.
An efficient distribution tray maximizes the available catalyst activity and thereby optimizes run length, whatever the product sulfur goals may be.
Some hydrotreaters do not have a distributor tray at all. Some older units have a tray, but it was designed for a different feedstock, flow rate, or product requirement than what the operation sees today.
An inefficient tray or just 1% bypassing of the catalyst in the reactor will make it very difficult, if not impossible, to meet diesel-sulfur specifications.
Although catalyst activity in the hydrotreater is the primary factor that determines overall performance, the liquid-distribution tray is also important because it evenly distributes reactants across the catalyst bed. Even dispersion of liquid feed and treat gas across the cross section of the reactor optimizes catalyst utilization, minimizes maldistribution, and prevents the formation of hot spots.
A properly designed tray should have close spacing, low pressure drop, low fouling potential, and provide some vapor and liquid mixing. The tray should have a low sensitivity to levelness and be able to handle varying vapor and liquid ratios, as the degree of vaporization changes from SOR to end-of-run (EOR).
Given today's refining environment with variability in feedstock quality, feedstock blend properties, sulfur content, and operating conditions, refiners will require assistance from catalyst vendors to determine the optimum catalyst load for a given operation to make ultra-low sulfur diesel.
In many cases, this will be easy and will require only a simple evaluation. In other cases, pilot plant evaluations will determine the optimum catalyst fill for that particular operation. These types of evaluations assume ideal distribution in the reactor and zero H2S in the treat gas, unless otherwise specified.
Haldor Topsoe conducted a number of tests that showed that nickel molybdenum (NiMo) catalysts removed sulfur compounds from sterically-hindered compounds better than cobalt molybdenum (CoMo) catalysts. CoMo catalysts are preferred for higher liquid hourly space velocity (LHSV) reactor operations.
- NiMo better at removal of dibenzothiophenes with side chains. Dibenzothiophenes with side chains are among the most difficult sulfur compounds to remove.
Fig. 1 shows the different reactivities of the sulfur species. As the relative reaction rate decreases, the sulfur compounds become more difficult to remove.
The well-established rule of thumb for hydrotreating catalyst selection says that sulfur-removal applications prefer CoMo catalyst types and nitrogen-removal applications prefer NiMo types. While this is basically correct, it is not always true for feeds that require deep desulfurization
In using deep desulfurization to make ultra-low-sulfur diesel, sterically hindered sulfur compounds like 4-methyl dibenzothiophene and 4,6 dimethyl dibenzothiophene are less reactive than dibenzothiophenes without side chains by several orders of magnitude. These sterically hindered dibenzothiophene isomers must be desulfurized to reduce the sulfur content to 30 ppm or less.
Fig. 2 shows two possible reaction pathways for sulfur removal. The first is a direct route. In the direct route, extraction of the sulfur atom from the molecule does not first require hydrogenation of the aromatic ring.
The second route is indirect. That is, extraction of the sulfur atom follows hydrogenation of the aromatic ring. Not surprisingly, the direct route is better catalyzed by a CoMo type catalyst, and the indirect route by a NiMo type.
It should be noted, however, that the direct route (CoMo) is inhibited both by H2S and nitrogen compounds. The indirect route (NiMo), on the other hand, is an intrinsically faster reaction. In this case, the main inhibitors are the nitrogen compounds.
This work has also shown that for HDS of dibenzothiophene (DBT), the direct route is fastest. For the 4-methyl dibenzothiophene, the reaction proceeds at an equal rate along both routes, but for the 4,6 dimethyl dibenzothiophene, the preferred route is the indirect hydrogenation one.
On this basis, it is an advantage to have a catalyst with a hydrogenation function (NiMo type) to remove sulfur from the sterically-hindered compounds.
Fig. 3 compares the effectiveness of the two catalysts in removing sterically hindered compounds. These plots are actual chromatograms on desulfurized diesel using a gas chromatograph with a Sulfur Chemiluminescence Detector (GC-SCD). Use of this analytical technique provides valuable insight as to the type and concentration of the specific sulfur compounds in the feed to and product from the hydrotreater.
In this example, the feedstock was a blend of 50% straight-run gas oil and 50% LCO; the total sulfur content was 1.5 wt %. The HDS severity was 99%, corresponding to a diesel product sulfur content of 150 ppm (wt).
Fig. 3a shows the results using CoMo-type catalyst. At 99% HDS, all the dibenzothiophene and all the 4 methyl dibenzothiophenes were removed. The compound remaining with the highest concentration is 4,6 dimethyl dibenzothiophene. All of the other unidentified components are sterically-hindered dibenzothiophenes with alkyl substituents in both 4 and 6 positions.
Fig. 3b shows the results using a NiMo-type catalyst on the same feedstock and at the same severity of 99% HDS to produce a product sulfur of 150 ppm (wt). Even though there is no dibenzothiophene left, there is some residual 4 methyl dibenzothiophene.
Again, the compound with the highest concentration is 4,6 dimethyl dibenzothiophene, with some additional sterically-hindered dibenzothiophene with alkyl substituents in the 4 and 6 position.
In the case of the NiMo catalyst, the concentration of these sterically-hindered species is significantly lower. This experiment shows that although a CoMo-type catalyst is better in removing dibenzothiophene and 4 methyl dibenzothiophene, it is inferior to a NiMo-type in removing sterically-hindered sulfur types like 4,6-dimethyl dibenzothiophene and heavier sulfur species.
Although these pilot plant tests were carried out at industrially relevant conditions in terms of LHSV, operating pressure, and treat-gas rate, the severity was not high enough to make the proposed 30-ppm sulfur diesel.
Haldor Topsoe deliberately chose 99% HDS, equal to 150 ppm (wt) product sulfur to have sufficient total sulfur in the product to enable analysis by the GC-SCD sulfur speciation with a reasonable degree of accuracy.
Routine analysis of diesel sulfur content in the less than 50 ppm (wt) range will present additional challenges to the industry in terms of analytical method, sample containers, and techniques.
- NiMo better for higher pressures, lower LHSVs. In hydrotreater operating conditions, LHSV and hydrogen partial pressure are extremely important in determining the reactor SOR temperature and catalyst deactivation rate. These parameters will dictate the cycle length.
To transition from 500-ppm to 30-ppm sulfur diesel, the obvious answer is to build a new high-pressure hydrotreater designed for that sole purpose. There are other options for existing units, however:
- Adding additional catalyst volume, a new reactor, or lower LHSV.
- Increasing hydrogen partial pressure or total pressure.
- Installing a distributor designed for the relevant operation.
- Building an H2S removal system if there is not one.
All these measures will lower the SOR temperature and the catalyst deactivation rate, which will maximize the operating cycle. Even with other measures, it will be necessary to operate with an LHSV lower than that currently used, to make ultra-low sulfur diesel.
At the higher LHSVs, like 2.5 hr-1, CoMo catalyst is better than the NiMo catalyst at all operating pressures. At lower LHSVs, like those less than 1 hr-1, the ranking of the two catalysts changes with respect to HDS activity with increasing pressure.
The hydrogen partial-pressure dependency is different for the two different reaction routes. The direct extraction route (for which the CoMo catalyst has the highest activity) has a relatively low hydrogen-pressure dependency, whereas the hydrogenation route (for which the NiMo catalyst has the highest activity) exhibits a high hydrogen partial pressure dependency.
If the hydrogen partial pressure is high (greater than 600 psi), the thermodynamic equilibrium constraint for removal of sulfur via the hydrogenation route is moved to a higher temperature, and the apparent activation energy of HDS via the hydrogenation route is pressure-dependent.
Therefore, a NiMo catalyst is preferred over a CoMo one for a diesel hydrotreater operating at high pressure and low LHSV for production of ultra-low sulfur diesel.
Brian M. Moyse is the hydroprocessing technology manager in the hydroprocessing sales group at Haldor Topsoe Inc. He is an international expert in the application of grading technology for fixed-bed hydrotreating and has been working on the development and application of this technology for 20 years.
Moyse has worked for Shell Oil and Topsoe. His 40 years of experience include unit start-up, troubleshooting, technical service, sales, and catalyst research and development assignments.