U.S. refiners make complex-model RFG as they prepare for next hurdle

Citgo Petroleum Corp. is adding a 60,000 b/d FCC feed hydrotreater at its 304,000 b/d Lake Charles, La., refinery. The unit, scheduled to start up this month, will put Citgo in a good position to produce Phase 2 federal reformulated gasoline beginning in 2000. Photo courtesy of Citgo.

How a California Refiner Meets Tight Rfg Specs [109,716 bytes]

Anne K. Rhodes
Associate Managing Editor-News

U.S. refiners are facing a third round of reformulated gasoline (RFG) requirements just as the second round of preparations has come to a close.

Since the U.S. Environmental Protection Agency's requirements went into effect 3 years ago, all but a handful of U.S. refiners have been using the simple model to estimate emissions from gasoline combustion. The model calculates emissions based on product qualities such as Reid vapor pressure (Rvp), sulfur content, benzene levels, and distillation characteristics.

As of Jan. 1, 1998, however, all U.S. refiners are now calculating RFG emissions using the complex model.

At the same time, refiners must now begin preparing for the Phase 2 complex model regulation, which takes effect Jan. 1, 2000. This model will impose even stricter emissions limits.

Only a few refiners have announced projects aimed at the Phase 2 regulation. These projects involve primarily desulfurization, either of feed to the fluid catalytic cracking unit (FCCU) or of FCCU gasoline.

Background

RFG was introduced by EPA as a measure to help polluted areas meet National Ambient Air Quality Standards (Naaqs) for atmospheric ozone concentrations. Beginning Jan. 1, 1995, all gasoline sold in areas out of compliance with Naaqs ozone standards (called "nonattainment areas") had to be reformulated.

Refiners were given a choice of methods to calculate gasoline emissions based on product qualities. These choices initially included the simple model and early-use complex model (OGJ, Jan. 17, 1994, p. 16). An alternative simple model was introduced later because of gasoline fungibility issues.

All of the models use what EPA calls a baseline for determining quality limits. The purpose of the baseline is to prevent the "dumping" of lower-quality feedstocks into conventional gasoline, which makes up about 70% of the U.S. gasoline market.

During the past 3 years, refiners could either use their individual 1990 gasoline quality baseline, calculated from detailed blending and operating records, or they could opt to use a stricter, statutory baseline.

Most chose to undertake the arduous task of calculating their individual baselines.

The so-called "antidumping" rule put non-U.S. refiners exporting conventional gasoline to the U.S. at a competitive disadvantage, because EPA required them to use the statutory baseline, as it was unable to verify their 1990 baselines.

Venezuela and Brazil protested to the World Trade Organization about the regulation, and EPA has proposed changes to the rule (OGJ, May 12, 1997, p. 36). The changes involve allowing non-U.S. refiners to apply for individual baselines and monitoring the quality of imported conventional gasoline in order to prevent dumping.

Phase 1 complex model

The Phase 1 complex model is more detailed and stricter than the simple model, says Eric Schneider, a consultant with Ernst & Young Wright Killen (EYWK), Houston.

The complex model calculates emissions using eight gasoline parameters, including aromatics concentration, evaporative qualities, and sulfur content. In order to use the model, the gasoline a refiner produces, distributes, or sells must fall within accepted ranges for all eight parameters (see table [27,400 bytes], p. 24).

"There are tighter complex-model limits on some parameters-for example, sulfur," said Bob Greco, a senior manager at the American Petroleum Institute. RFG made under the complex model will be limited to a maximum of 500 ppm (for sulfur, in order to be eligible) to use the complex model."

With the simple model, the statutory sulfur limit was 1,000 ppm, but most refiners chose to use the baseline calculation method and, as such, were limited to their (1990) baseline sulfur content, said Greco.

"The complex model will give refiners more flexibility, with respect to emissions reductions, than the simple model," Greco explained. "The simple model was limited to just a few variables-oxygen, benzene, and Rvp-whereas, in the complex model, you now have the ability to manipulate sulfur, aromatics, and distillation properties, for example, and be able to claim credit for changes in those properties."

Greco says this does not cause a constraint for most refineries, but stresses that it is an additional limit. Refineries that had a higher baseline are more likely to be affected by it, he said.

The complex model also evens out the playing field by comparing each refiner's RFG quality with a statutory baseline. For conventional gasoline, on the other hand, the antidumping rule, which limits many parameters to not more than 125% of 1990 levels, still apply to individual refinery baselines.

"From a formulation standpoint, there really should be very little difference. We don't anticipate any difficulties in rolling out the complex model," said Greco.

Terrence Higgins, technical director of the National Petroleum Refiners Association (NPRA), agreed with Greco's assessment.

"Some refiners were worse off than average to start with, and some were better off than average," said Higgins. He says the flexibility of the complex model will enable refiners to compensate for part of that difference.

Still, some changes were necessary. Some refiners probably solved the Phase 1 complex-model problem in conjunction with their Year 2000 plan, says Higgins.

"There are at least two firms that have used the complex model all along," says Higgins. The differences involved in using the complex model depend on several factors, he says, including the way a refinery has historically blended, the way it is configured, and its feedstock quality.

The complex and simple models were both introduced in 1994, Schneider pointed out, "so the people responsible at each site had a chance to analyze this over 3 years. They didn't start out using the complex model, with a few exceptions. So they've had a chance to look at what they're producing, look at their individual limits, and look at their import facilities. And they've had 3 years to tailor their operations."

Phase 1 NO_x

The Phase 1 complex model mandates a 1.5% reduction in calculated NO _x emissions compared with the statutory 1990 baseline. In the equation used to calculate NO _x, sulfur is the parameter that has the greatest effect.

Under the Phase 1 simple model, sulfur could not exceed the 1990 baseline. But for the Phase 1 complex model, there is a maximum sulfur specification of 500 ppm (see table [49,295 bytes], OGJ, Jan. 6, 1997, p. 21).

Higgins said that, when EPA promulgated the initial RFG rule in 1994, it included a per-gallon NO_x cap. While this cap wasn't constraining under the simple model, it would have been under the complex model.

The cap "took away a lot of flexibility," said Higgins.

After some urging from API, EPA decided to reconsider the NO_x cap. "They have proposed to eliminate that NO_x per-gallon cap and haveellipsetold us that we can expect that regulation to be out in the near future," said Higgins.

Compliance tools

Given the intricate nature of gasoline blending under the EPA emissions models, many refiners are using computer programs to help them cope with the requirements. Many major oil companies have written their own programs for this task, but at least two programs are available commercially.

One such program, offered by Houston-based Pace Consultants Inc., is called PACEctr (pronounced "pace setter"). Another, produced by EYWK, is called WKComply.

The EYWK compliance package is used at about 80 sites, says Schneider. These users include refiners, blenders, traders, independent laboratories, and corporate headquarters.

WKComply has been used with all of the federal models and with the California predictive model for California Air Resources Board (CARB) Phase II RFG. It is updated with each change in the regulations, says Schneider.

WKComply is a database program that performs three fundamental tasks: tracking, projecting, and reporting.

The tracking function keeps records for compliance purposes and produces the batch reports that must be submitted to EPA quarterly or annually, depending on the type of gasoline involved. (RFG reports are submitted quarterly, and conventional gasoline reports annually.)

The program also determines the quality of product required in order for the producer to meet its appropriate average values over a given reporting period.

In this way, says Schneider, "you know, if you're very near the end of a reporting period, what you have to produce for the remainder of the period, and how that will impact your compliance."

WKComply also will determine the absolute maximum for certain parameters.

Reporting requirements

Since RFG was introduced in 1995, refiners have been required to submit to EPA reports containing information related to the quality of the gasoline they make. Reporting requirements did not change significantly with the introduction of the mandatory complex model in 1998.

"I'm not aware of any significant changes in the actual reports or formats," said Schneider. "But the record-keeping is more complex because of the eight parameters."

Reports on RFG batches must be submitted quarterly, but refiners have a 2-month "lay-over" period, he explained. So, for gasoline produced during January, February, and March, for example, the appropriate documents must be submitted by the end of May.

For conventional gasoline, anti-dumping reports must be submitted annually, also with a 2-month delay.

Phase 2 RFG

Refiners face more onerous regulation beginning Jan. 1, 2000, when Phase 2 complex-model RFG will be introduced.

Three major changes are in store for 2000. These relate to calculated emissions of NO_x, volatile organic compounds (VOCs), and toxic compounds.

"Both the VOC reduction and the toxics reduction become more stringent," said Greco, "but the most significant change is the addition of a 5-7% NO_x reduction in summer."

To reduce NO_x, says Greco, "from a blending standpoint, in the complex model you get the most bang for your buck from reducing sulfur or olefins. Those are going to be the two primary means by which refiners comply with the NO_x reduction standard."

"It's going to be a delicate balance because, if you reduce olefins, it benefits you on NO_x but it actually increases VOCs. So there are tradeoffs associated with at least, for example, olefins."

Sulfur and NO_x

The Phase 2 complex model includes a reduction in NO _x emissions of 6.8% from the 1990 baseline. Several parameters affect NO _x in this model: Rvp, sulfur, aromatics, and olefins to a lesser degree.

"From what we can tell from that modelellipsesulfur has the biggest impact on NO_x," said Dick Arthur, who works in the process engineering department of M.W. Kellogg Co., Houston.

Cat gasoline (gasoline produced in the FCCU)-the largest contributor to gasoline sulfur-is responsible for 40-50% of the sulfur in the final product, said Arthur.

A few refiners may be able to adequately reduce their gasoline sulfur level by processing low-sulfur crudes. Most refiners, however, will not be able to meet sulfur specs by manipulating crude oil blending.

These companies have two basic options: they can hydrotreat the feed to the FCCU or hydrotreat the gasoline fraction produced by the unit.

Cat feed desulfurization has many advantages, but it is very capital-intensive. "It runs at higher pressure," compared with cat gasoline desulfurization, said Arthur. "It requires high-alloy steel to make recycle compressors and reactors." And you have to treat a larger volume of feed: 50,000 b/d, for example, vs. 5,000 b/d of heavy cat gasoline.

The advantages associated with cat feed hydrotreating include improved gasoline yield and quality and reduced SOx emissions from the FCCU.

"Treating the cat gasoline is less capital-intensive," said Bill Hillier, product director of refinery planning for M.W. Kellogg Technology Co.

Another option on the feed end of the FCCU is to install a mild hydro-cracking unit upstream of the unit. This scheme can optimize the gasoline-to-distillate ratio for swinging from summer to winter operation, says Arthur, and it increases FCC yield.

"If you're looking for a return on your investment, you would just as soon mild hydrocrack or hydrotreat the feed to the FCC," said Arthur.

If a refiner chooses the mild hydro- cracking option, there is a tremendous increase in yields of jet fuel and diesel, while gasoline output remains relatively flat, said Hillier. "So adding hydro- cracking capacity...improves the refinery's material balance, makes environmentally good diesel, and makes more jet and good cat feed," he said.

"For environmental and yield reasons, that seems like sort of a logical thing to do, Hillier said. "We have taken cat feed hydrotreaters and converted them into hydrocrackers. We just completed one such project with OMV in Europe," where diesel demand is greater than in the U.S.

"We have some proposals out on those kinds of projects, but not in the U.S.," he added.

"I think each refiner will find the most economical way. In general, I would say people are trying to minimize capital expenditures. They're not willing to put out a couple hundred million dollars for cat feed hydrotreating if they can spend $30 million on some form of sulfur reduction in the gasoline."

"If you look forward ellipsethere's going to be more pressure on sulfur," said Arthur. "You're going to want to go toward cat feed hydrotreating over time. It's a better option for the future."

Some refiners will have to use a combination of feed and product treatment, say Arthur and Hillier. "One of the majors we were working with found (that cat feed hydrotreating) wasn't high-enough pressure and enough desulfurization to meet their requirements," said Hillier, so it undertook a smaller project to help it clean up the sulfur in the tail end of the cat gasoline.

"Another example of that is with your CARB specs," said Arthur. CARB has set a flat limit on its Phase II RFG of 40 ppm sulfur for larger refiners and 80 ppm for smaller plants.

The average sulfur in gasoline in the U.S. before introduction of the Phase 1 complex model was about 340 ppm. Higgins says refiners will have to reduce sulfur levels to 100-150 ppm in order to meet NO_x limits under the Year 2000 RFG regulations.

When sulfur specifications fall below 100-200 ppm, said Arthur, a refiner will have to use a combination of cat feed hydrotreating and desulfurizing the heavy cat gasoline.

What each refinery has to do, he said, is look at its gasoline qualities, plug the values into the complex model equations, and start "playing" with the variables. In this way, they can evaluate the effect of changing certain variables against the costs to do so.

"Those that add cat feed hydrotreating will also have to add hydrogen capacity," said Arthur. Desulfurization of heavy cat gasoline, on the other hand, consumes much less hydrogen, so individual refineries may have adequate capacity to feed that kind of unit.

Cat gasoline desulfurization is more selective in how it uses hydrogen, says Arthur. "If you're pumping it into a cat feed hydrotreater, you're saturating a lot of material," he said.

Arthur summed up the trade-offs associated with the processing choices for sulfur reduction: "You can either spend the money downstream (of the FCCU) and get the toxics or NO_x credit, or you can spend a little more money upstream...and get some return on it. Longer-term, that's going to be a better alternative."

Olefins and NO_x

Hillier points out that the C ₅/C ₆ olefin content of gasoline also is an important variable in determining NO _x emissions via the Phase 2 complex model.

The alkylation unit historically has been used to process the C₄ olefin stream. But, says Hillier, "There is a potential for having to do something with the C₅ olefins, such as alkylation."

"Some refiners are already doing a little of that," said Arthur.

"It's not as easy a feed to alkylate as the C₄s," said Hillier, "but I think there could be a potential for people to start running additional C₅s into their alkylation units."

When asked if he thought U.S. refiners would increase their alkylation capacities in order to meet the Phase 2 specs, Hillier said, "I have a feeling that alkylation could play a bigger role."

Alkylate is a perfect gasoline feedstock because it has no sulfur and no aromatics. "It's just a matter of the fact that the C₅s don't act as well in alkylation as C₄s." said Hillier. "You don't get the same yield, you don't get as good an octane, there's more acid use, etc."

"There are two reasons you would want to do it," said Arthur. "One is the olefins issue. The other is that C₅ alkylation doesn't change octane really, but it changes Rvp. In other words," he explained, "C₅ alkylate is much lower Rvp than the C₅ feed to the alky unit."

This reduced Rvp would help a refiner meet both NO_x and VOC limits.

VOCs

Under the Phase 2 complex model, VOCs must be reduced by about 10% from 1990 baselines.

"The biggest knob you have to turn there in a refineryellipseis Reid vapor pressure," said Arthur.

In the northern U.S., Rvp reduction will not be as big an issue. But refiners in the southern part of the country, for the most part, have already taken the C₄ stream (butanes) out of gasoline in order to meet lower Rvp specifications.

In 2000, they are going to have to dig into C₅s in order to reduce Rvp, says Arthur. "The biggest issue that I can see that refiners are going to have to face is 'What are we going to do with this stuff?'" he said.

The stream can be used as feed to either a chemical facility or a hydrogen plant. Refiners that have an ethylene plant nearby have some flexibility, says Arthur.

Toxics

Under the Phase 2 complex model, toxic emissions will have to be 21% less than the 1990 baseline vs. 16.5% less for the Phase 1 complex model.

Factors affecting toxics emissions include Rvp, aromatics content, oxygenate concentration, sulfur, and olefins. But the equation is most sensitive to benzene, said Arthur.

"If you take a gallon of benzene out (of gasoline)," explained Hillier, "it's equivalent (in its effect on toxic emissions reduction) to taking 28 gal of aromatics out. In other words, if somebody had a choice of taking out benzene or aromatics, they'd get a bigger bang for their buck by taking out the benzene."

The maximum benzene content in today's RFG is 1.0 vol %. Some refiners may choose to try to eliminate this instead of reducing aromatics, says Hillier.

"This points out the complexity of gasoline blending at the refinery," he explained. "The blender must figure out what's the most economic way to match all these variables up and not get put in jail."

To prepare for Phase 1 federal RFG, many refiners installed pre-reformer splitters to remove benzene precursors (OGJ, Mar. 21, 1994, p. 51).

"Now," said Arthur, "in order to achieve more toxics reduction, an easy target is that light straight-run stream that they cut out last time." That stream contains 2-5% benzene.

One way to handle this stream is to isomerize it. Arthur said Kellogg has a process that allows the refiner to convert as much as 5 vol % benzene, removing it from that stream. By eliminating one of the key benzene sources, this buys the refiner a lot of toxics credits for a relatively low capital cost, he said.

Other ways to remove it include extractive distillation and benzene hydrogenation, said Hillier.

The evaporative qualities of gasoline also have a big effect on toxic emissions. In the model, these are expressed as E200 and E300-respectively, the percentages of gasoline evaporated at 200° F. and 300° F.

"The presence of E200 and E300 in the complex model equations is essentially making gasoline a lighter product," said Hillier. "That causes a yield problem."

"Generally speaking," added Ar- thur, "the severe undercutting of the cat gasoline...reduces aromatics and thus buys you toxics credits."

If a refiner moves that heaviest end into the diesel pool, everything has to be hydrotreated. Arthur conceded that could be one processing option, "ellipseif you've got (adequate) cetane and you can do that without too much of a loss. You can also undercut it to the reformer," he said.

"Both of those are very expensive options, but they buy you room for aromatics, toxics, and they buy you credits with that 300° evaporative point. Frankly, I don't think either one of them is a very good option, although California (refiners) have to use them," said Arthur.

Planning for 2000

"We're not that far away from Phase 2," said Hillier. "There's work to be done.

"Some of these changes we're talking about don't take a high-tech engineering company like Kellogg to do-the distillation changes and a few things like that," he said. "But, as you get into a little more complex technology, we haven't seen a big run on projects.

"We've done a hydrotreater on cat gasoline, but there's not a big rush of people lining up at the door. I'm a little surprised at that," he said.

NPRA Pres. Urvan Sternfels says it's a sure bet that the companies that are making RFG now are the ones that are going to continue to do so.

"Some of them have kind of phased into it already with other projects," said Higgins. "There is a tendency to try to minimize the capital layout and maybe do it with other operational changes if they can."

Higgins is involved in an organization called the Implementation Workgroup. The group, composed of EPA, affected industries, and states, was formed to ensure a smooth implementation of gasoline standards-an issue that has caused problems in the past (see story, p. 26).

"From the standpoint of someone who works on the Implementation Workgroup," said Higgins, "I think the industry is comfortable that, when the time comes, they will be ready."

"The oil industry is very innovative," said Hillier. "They find ways to do things that are much more economical than originally thought."

Potential pitfalls

Sternfels cautioned that industry preparation "is predicated on the assumption that only those companies and those areas that are now supplied with RFG will be in that business. If, as we expect might happen, there is a greater spread of RFG to other areas-either by states opting into the program or by further requirements for RFG-that obviously will have some impact on refiners that market in those areas."

EPA issued a rule late last year limiting states' ability to opt out of the RFG program after opting into it (OGJ, Nov. 3, 1997, p. 34). The rule, which forces states to choose whether to participate by yearend 1997, was designed to reduce uncertainty with regard to the size of the RFG market.

EPA also ruled to decrease the acceptable Naaqs ozone concentration but deferred implementation of the rule, says Sternfels. This change will greatly increase the number and size of nonattainment areas, thus creating a massive increase in RFG demand (OGJ, Jan. 8, 1996, p. 34).

"We don't know what the states that are affected will do," said Sternfels. "They are required to modify their SIPs (state implementation plans) and demonstrate that they're working towards attainment by taking any measures that they deem appropriate."

"My sense is that the issue of overall supply in an area will be of critical importance-and one that I would hope (EPA) would pay attention to," said Sternfels. "There are just too many undetermined factors."

Directionally, says Sternfels, a sudden increase in RFG demand in areas newly designated as nonattainment would tend to force some of the smaller companies in those regions to make difficult decisions.

"If a small niche refinery is supplying an RFG market, they're in deep trouble if they can't justify the capital investment to leapfrog into the technological status that would enable them to do so," he said. "That would further exasperate our overtaxed, high-utilization situation.

"We're really at the edge now. I think that would create more demand for imports if it backs out domestic supplies. It would mean that we are less capable of defending ourselves against economic and political blackmail from foreign interests."