Auto gas tanks: the great barrier grief.

Makers of plastic gas tanks face new rules and new fuels. They need new barriers with near zero emissions to fight a resurgence of steel tanks. Here's a look at how far barrier technologies have come and how far they have to go.

Plastics processors and materials suppliers have been understandably flapped about gas tanks. Following Europe's early lead, plastic tanks in the U.S. were expected to have grown by leaps and bounds from about 20%, or some 2.3 million, of all new car tanks at present to 50% by 1995. In Europe, 60% of all cars have plastic tanks now, and that number is expected to go to 90% by '95.

But now all bets appear to be off in the U.S. for the '95 model year and beyond. General Motors Corp. and Ford Motor Co. both recently switched some '95 vehicles from plastic to metal tanks. Even if that becomes only a temporary blip in plastic tank progress, that blip will still mean a loss of millions of plastic tanks over a period of years, since a gas-tank design is typically used for five to 10 years and in many car models.

The problem is tight new emissions standards for cars, combined with uncertainty about future fuels containing methanol. There are plastic barriers in development that can meet the new standards for conventional fuels. In combination, they may even contain methanol vapor, but double barriers would cost more than present plastic tanks and probably more than coated metal ones. "Steel with coatings is still cheaper than even a monolayer HDPE tank," says GM's Phil Yaccarino, manager of chassis for Chevrolet/Pontiac of Canada. And a recently published study on gas tanks and fuel systems by consulting firm Phillip Townsend Associates in Livonia, Mich., says, "New barrier technologies for plastic fuel tanks could increase production costs by as much as 15% (based on 100,000 tanks/yr) compared with producing today's plastic fuel tanks."

Meanwhile, Ford's supervisor of fuel-tank and emissions engineering, John Thorn, says it's still hard to determine whether plastic tanks will cost more, since "new steel processes aren't tested or costed out."

IT STARTED IN CALIFORNIA

Tighter emissions standards began as a regulation from the California Air Resource Board (CARB), and similar rules may be adopted by 11 other states, or about a third of the U.S. car market. CARB rules will eventually require all new cars sold in California to give off no more than 2 g/24 hr of hydrocarbons. Ten percent of new cars must comply by 1995, 30% by 1996, 50% by 1997, and 100% by 1998. Previously, cars had to comply with federal Environmental Protection Agency limits of 2 g/2 hr, effectively a twelve-fold higher limit.

Now the federal EPA is playing catch-up. New national EPA standards, authorized by the 1990 Clean Air Act, may be published by next month and could endorse CARB or be even tighter. One draft of the EPA regulation suggested 0.1 g/4 hr "resting loss," or evaporation of hydrocarbons from a standing car, which OEMs say would effectively require zero emissions from the gas tank, since tubing, gaskets, even upholstery also emit hydrocarbons. When CARB standards came out in 1990, there was no known technology that could meet them, and the same may be true of the EPA's rules.

Even steel tanks aren't safe from CARB emissions scrutiny. CARB measures emissions from a whole car, not just the tank. Steel tanks, which theoretically emit zero hydrocarbons, may emit plenty if grommets, welds and elastomeric fittings are considered. In fact, Ford and Solvay Automotive Inc., a leading plastic tank blow molder, tested steel tanks and found emissions of 0.38-0.5 g/24 hr. The test uses a stringent "sealed housing for evaporative emission" (SHED) method, in which a whole car is sealed into a box or other container, heated and cooled to simulate a day/night cycle, and hydrocarbon vapors are measured by a flame ionization detector. In Europe, emissions are measured less exactingly by weight loss, and European test fuels don't include methanol, though up to 8% methanol is allowed in gasoline there.

There is so much uncertainty among car makers, with internal design factions espousing either metal or plastic tanks, that no one knows from day to day what's happening. Some OEMs ordered tooling for steel tanks for some 1995 model-year programs as a precaution because the lead time is so much longer for steel than for plastic. For the next several months, molders say, the OEMs could still switch to plastic tooling without delaying their 1995-model programs, but they'd have to walk away from the money spent on the "parallel path" of steel tooling. Tooling for metal stamping costs $3-5 million, versus under $1 million for plastic.

Environmental issues aside, the big arguments in favor of plastic tanks remain weight savings for fuel efficiency and safety in case of collision, since plastic tanks don't explode the way steel can. OEMs dispute both points. In some car models, a plastic tank might be safer, but generally GM maintains that steel is as safe as plastic. A plastic tank may also require a heat shield, which adds back much of the weight the plastic saves. Overall, the big benefit of plastic, OEMs say, is its ability to be molded into complex shapes.

METHANOL, ANYONE?

Methanol is what's gumming things up for the plastic fuel tank. (Methanol is also hard on steel, so new epoxy/metal and plasma coatings for steel tanks are being developed, too.) If gasoline contains no more than the current legal limit of 5% methanol, existing plastic tank barriers can contain it, molders say. But some OEMs say none of the plastic barriers work well enough with methanol. And fuels with higher methanol content are being considered--10, 15 and even 35% methanol (M-10, M-15 and M-35). M-35 is unlikely to catch on because it would corrode car engines, but the legal limit could be pushed up to M-10 or M-15. In California, there is already a commercial "Clean Fuel" made of 85% methanol and 15% gasoline (M-85, or flex fuel) to be used only in new "clean fuel vehicles," which car makers are supplying with stainless-steel gas tanks.

Several molders say they're now getting good emissions test results even with M-15--notably Solvay and Kautex in Windsor, Ontario. But molders have a tough sell in Detroit because CARB requires emissions standards to be met for a car's 10-yr lifespan or 100,000 miles. And some of these barrier developments are so new that there's insufficient data on how they hold up over time, especially with methanol.

TIMING IS EVERYTHING

As Big Three auto companies plan programs beyond '95, GM switched 270,000 tanks that were to go on its Corsica/Beretta cars in plastic and will use an existing steel tank for which GM already has tooling. The reasons stemmed only partly from barrier issues--the rest of the equation was economic. The steel tank is cheaper for GM because it uses existing equipment, labor force and tooling. GM has several major '96 programs that are on the fence. "Pending satisfactory barrier technology," they'll be plastic, but decisions will need to be made in a couple of months, GM's Yaccarino says.

Ford, which makes both steel and plastic tanks, intends its WIN 88 van program, starting in the '94 model year, to have a plastic fuel tank, but has started a "parallel path" in metal just in case. "New data from test programs at Ford are tending to converge on the plastic fuel tank. The decision to go to steel on one or two platforms is not necessarily a long-term trend," says William Mayes, executive engineer for Vehicle Exterior Systems at Ford, adding that in six months he expects to have enough performance data to decide.

Worried about a flight from plastic, the Society of the Plastics Industry in Washington, D.C., convened last November an unusual plastic gas-tank committee representing nearly every company with an interest in the U.S. gas-tank market. The 50-member group, which includes competing U.S. gas-tank molders and car makers, was formed with the purpose of avoiding duplicate testing and achieving comparability in test results by setting uniform standards for reformulated fuels, performance data and durability testing of whole tanks.

GETTING BETTER ALL THE TIME

In North America, only three processors make plastic automotive gas tanks: Ford's Milan, Mich., plant molds tanks with and without fluorination. Kautex in Canada molds tanks for all of the Big Three, using fluorination; and Solvay (which acquired gas-tank molding operations formerly belonging to Kuhlman Plastics and Hedwin Corp.) also molds tanks for Big Three vehicles using either fluorination or sulfonation.

All production tanks in North America have been monolayer tanks, but coextrusion is coming. Nissan Motor Manufacturing Corp. in Smyrna, Tenn., next month launches the first commercial coextrusion of gas tanks here. Ford has also ordered a coextrusion lab machine and coex retrofits for six of its 12 continuous-extrusion gas-tank blow molders from Krupp Kautex Div. of Krupp Plastics & Rubber Machinery in Edison, N.J. (not related to Kautex in Canada).

Following is a status report on the gas-tank barrier technologies used in commercial production and a look at technologies being developed for barriers against methanol fuels.

FLUORINATION HAS THE LEAD

Fluorination, used on 60% of U.S. plastic tanks, takes compressed nitrogen gas containing 1-2% fluorine (a poisonous and very reactive material) and injects it into the parison as a tank is blown. Fluorine reacts with the inside surface of the HDPE tank by substituting fluorine for some hydrogen atoms. This turns a few microns of the surface depth into a dense fluorinated hydrocarbon, analogous to PTFE. Fluorination has been used in production for so long that the first U.S. patents have expired. (In-line fluorination started with equipment licensed from Air Products & Chemicals Inc., Allentown, Pa., whose U.S. patents ran out last January.)

The advantage of fluorination coating is that it is done on a monolayer tank. The big disadvantage is environmental. The toxicity of fluorine means that blowing air must be withdrawn and scrubbed before release, and the mold area must be specially hooded to protect workers. Off-line fluorination of tanks after blow molding is also done. Plastic Omnium in France licenses such a process from Fluoro-Seal Inc. in Houston. Fluoro-Seal has also treated test tanks for Ford.

Several companies are developing enhanced fluorination to improve the present barrier limits. Air Products has improved its fluorination technology and shared new developments with two of the three North American molders (probably Ford and Kautex Werke). Solvay claims process improvements of its own and has applied for patents on "optimized fluorination," which, in combination with a new Solvay barrier resin, is said to make a more uniform barrier than plain fluorination. At the annual Society of Automotive Engineers show and conference in February, Solvay announced that in SHED tests, its newest HDPE tanks have hydrocarbon permeation rates of less than 0.1 g/24 hr. Solvay says these results, based on testing hundreds of tanks, are corroborated in independent testing by Ford and Chrysler. Kautex Werke in Germany (parent of the Kautex gas-tank operation in Canada) also has enhanced fluorination that achieves 0.1 g/24 hr of gasoline for a production tank, and is supplying tanks to OEMs for testing, says Andy Puempel, manager of product engineering in Canada.

Molders enhance fluorination barrier several ways: by removing all oxygen from inside the parison with a nitrogen purge before blowing fluorinated nitrogen into the part; controlling mold temperatures; optimizing the amount of fluorine in the blowing gas, as well as hold time and pressure; repeating treatment as much as three times; or blowing with a pulsating movement of fluorinated gas.

Pigmentation in the HDPE also has an effect. Solvay Polymers Inc. in Houston has an enhanced barrier resin, which it calls "P resin" (for "permeability"). It's precolored with a special small-particle-size carbon black, instead of blending pelletized black masterbatch into natural HDPE at the blow molder. Because black master-batches--used for uv protection--contain a standard HDPE, they reportedly blend unevenly and make an uneven barrier. Coextruding a natural unpigmented layer of HDPE a few mils thick on the inside of a tank is also said to make fluorination more even.

P resin is a higher-molecular-weight variation of Solvay's XF-714 HDPE. Gas tanks here traditionally use 10 HLMI resin, while European tanks use tougher 5-6 HLMI resins with higher molecular weight, and P resin is tougher still at 4 HLMI.

SULFONATION: OLDIE BUT GOODIE 

A thousand tanks for Pontiac wagons were sulfonated as far back as 1970 by Dow Chemical Co., Midland, Mich. Sulfonation is an off-line process in which a tank is rinsed with sulfur trioxide, neutralized with ammonia, and again rinsed with water. The barrier is good enough by today's standards, and the advantage has been that, like fluorination, it allows a monolayer tank to TABULAR DATA OMITTED be used. But because the process is off-line, it's vulnerable to variations in ambient temperature or humidity and can't be statistically controlled. Also, the chemical agents are toxic and messy. Solvay uses sulfonation and a "supersulfonation" process, which basically means a thicker sulfide coating.

Coalition Technologies Ltd. in Midland, Mich., is developing the first in-line sulfonation. Coalition uses 10-20% sulfur trioxide gas, recycles the unused portion, and neutralizes with calcium, which becomes incorporated into the sulfonated surface. Coalition says this gives a better barrier than the company's previous "sulfometalization" process, which deposited metal molecules as a third step. In-line sulfonation has the reported advantage of being statistically controllable. (Both Solvay and Coalition derive their technologies originally from Dow patents, and Dow owns 30% of Coalition.)

Initial gas-tank trials by several car companies, however, have produced mixed results. And Coalition is criticized for testing with helium, which may not duplicate the permeation behavior of methanol or gasoline. Another concern is that both sulfonation and fluorination make surface layers that could potentially crack or flake. Coalition says accelerated aging tests haven't shown this effect with sulfonation. Coalition analyzed 10 five-year-old sulfonated tanks and one of the original 1000 sulfonated 1970 Pontiac tanks, and found no diminution of barrier properties. Solvay also says it has seen neither problem in "several million tanks including fluorination samples over 10 years old and over 50,000 miles."

LAMINAR-BLEND BARRIERS

Selar laminar-barrier materials from Du Pont Co., Wilmington, Del., are based on nylon and other resins, which are blended in an unhomogenized manner into monolayer tanks, where they form multiple layers of platelets that overlap and retard passage of hydrocarbons. It's sort of like coextrusion with a single layer. It requires a special mixing screw, a license and know-how from Du Pont, but is far simpler in equipment terms than the elaborate conversions needed for co-extrusion, fluorination or sulfonation. Selar RB901, a nylon material, is used commercially in gas tanks for Volvo, Lotus, Renault and Opel and will soon be in production at Fiat in Italy and on a Japanese car.

By European ECE-34 permeation tests, 7% Selar RB901 in HDPE shows emission rates of 0.3 g/24 hr on a production 90-liter tank in France; 0.4 g/24 hr on a 75-liter tank in Sweden to be introduced this summer; and 0.35 g/24 hr on a tank for an English car.

Selar 2, Selar plus fluorination, and coextruding with Selar are three additional possibilities for raising the barrier against M-15 fuel. Selar 2, also called Selar RBM (methanol), is based on either EVOH or modified polyvinyl alcohol (PVOH). Du Pont's Selar product manager in Brussels, Richard Bell, says PVOH is the better M-15 barrier. The same equipment and processing techniques can be used with Selar PVOH as with commercial Selar RB901, and up to 60% regrind can be used in production.

Tanks made with 7% Selar RBM and tested with M-15 fuel showed 0.5 g/24 hr with the European ECE-34 emissions test. In a SHED test, the same tank emits 0.25 g/24 hr, which isn't quite as good as OEMs want (they say they want 0.1 g/24 hr), but it's close. If fluorination is added to the Selar RBM tank, it gives SHED results of 0.189-0.3 g/24 hr even with M-15 fuels, Bell says. Ford, Solvay and Kautex are all looking at fluorinating monolayer tanks containing Selar 2. And Bekum Maschinenfabriken in Germany has developed a six-layer coextrusion machine to make developmental tanks with a layer of Selar barrier resin. (Contact Bekum Plastics Machinery, Inc., Williamston, Mich.)

COEXTRUSION WITH NYLON 6

Already the choice of several Japanese molders who, for environmental reasons, won't use sulfur trioxide or fluorine, nylon 6 coextrusion with HDPE makes an excellent gasoline barrier, but is generally thought not to stop methanol vapor. Because nylon is hygroscopic, there's also concern that it could absorb moisture from methanol, swell and cause delamination. Ford and Nissan say their results indicate no such problem.

Nylon is also said to make regrind difficult to use, but Fuji Heavy Industries Ltd. and Nissan reportedly use nylon-containing coex regrind by repelletizing it first to make a homogeneous blend.

Coextrusion of gas tanks starts in the U.S. next month when Nissan launches a five-layer tank with nylon 6 barrier, like tanks Nissan already makes in the U.K. and Japan. Nissan has found that a thick nylon layer is adequate even with methanol fuels. A source close to the project says, "On a 20-lb tank there might be up to 2 lb of nylon. The nylon is effective. Nissan has done its homework."

Nissan uses Japan Steel Works' accumulator-head multilayer blow molding machines (See PT, Feb. '91, p. 62) and a high-molecular-weight nylon 6, CM 1061 from a joint venture of Monsanto Chemical Co., St. Louis, and Toray Industries Inc., Tokyo. Mazda, Honda and Subaru/Fuji also make five-layer coex tanks with nylon 6 in Japan for domestic specialty cars, and Toyota is planning a coex tank. Fuji uses accumulator-head blow molders from Ishikawajima Harima Heavy Industries Co. (IHI) in Japan (which also has offices in N.Y.C.).

Outside of Japanese companies, coex tanks are only made in labs, but that's changing. Besides Ford's planned retrofitting of blow molding machines for coex, Solvay has a newly installed five-layer coex machine in Blenheim, Ontario. It has a coex head but no extra extruders, so at present it's making only monolayer tanks. Kautex Werke has coex lab capabilities in Europe and is designing a dual-clamp, shuttle-type coextrusion blow molder. And Walbro Automotive Corp. in Auburn Hills, Mich., is considering entering the field and has ordered a multilayer continuous-extrusion machine from Krupp Kautex for gas-tank development.

Rikutec Richter Kunstofftechnik in Alten Kirchen, Germany (represented in the U.S. by Gould & Eberhardt, Webster, Mass.), offers an unusual head technology for large multilayer containers, including gas tanks. This technology has been tested by some U.S. auto companies and others (for description, see New Products, p. 77). Two 1000-liter production machines in Germany are doing five-layer testing with gas tanks, Rikutec says.

Coextrusion with EVOH isn't in production anywhere yet, but is being qualified by GM and others as a barrier against gasoline with methanol. Rikutec says its tests show tank drop test and impact properties aren't affected by an EVOH layer (1% of wall thickness), and regrind can be used as the outer layer.

TWIN-SHEET THERMOFORMING?

The idea is simultaneously to form coextruded barrier sheets of HDPE into two tank halves and seal them together. Laser welding might be used, because a pressure weld could provide insufficient barrier integrity. Advantages claimed for this approach over blow molding would be a more uniform barrier and lower cost. On the other hand, thermoforming might be slower and might not achieve as intricate tank designs as blow molding. Another concern is whether the welds would be strong enough to prevent bursting in impact tests. Ford tried twin-sheet thermoforming gas tanks over 10 years ago, together with Brown Machine Div. of John Brown Inc. in Beaverton, Mich.

Today, the twin-sheet process is used commercially to make holding tanks for recreational vehicles by Penda Corp. in Portage, Wis., also on Brown formers. Penda, Brown and Solvay Polymers are independently said to be looking into the possibility, as is toolmaker Portage Casting and Mold. R&D interest in the process has also been piqued in Japan and Europe. And several more thermoforming machinery companies are said to be studying the possibility.

BETS ON PLASMA TREATMENT

Plasma polymer deposition may not be commercial for a while, but several molders say it holds the greatest potential for gas-tank barriers. A plasma is an ionized gas, electrically charged in a vacuum, which can be chemically quite reactive with plastics at room temperature. Plasmas of monomers like ethylene can polymerize and coat a plastic surface with a dense PE film. Because the plasma-deposited PE is of super-high density--up to 1.7 g/cc in experiments at the German Plastics Processing Institute, or IKV--it presents a higher barrier to hydrocarbons (PT, May '89, p. 14). And because no harmful chemicals are used, plasma coating presents no safety or environmental disposal problems.

Plasma technology companies in the U.S. say they've had visits and inquiries from several car companies and automotive suppliers interested in fuel-tank applications. Ford, Kautex Werke, Bekum and Finnish chemical company Neste OY are among a dozen industry supporters of a two-year development program in plasma polymerization barriers for gas tanks at the IKV in Aachen. IKV results, said to be promising, have so far been performed only on bottles, not large, challenging shapes like gas tanks.

Kautex Werke has a gas-tank development project involving plasma-polymerized barriers with BMW AG in Munich. Plasma Elektronics in Filderstadt, Germany, is said to be plasma treating tanks for the project. The Technics Plasma division of Krauss-Maffei AG (with U.S. offices in Florence, Ky.) is said to be doing R&D in Kirchheim, Germany, with a major European chemical firm in plasma polymerization on gas tanks. Technics Plasma uses microwaves to achieve "considerably higher" polymer deposition rates than low-frequency plasma processes can, according to the company.

Himont/Plasma Science in Foster City, Calif., is working with several clients on automotive barrier applications.