Octane and Knock

Knowing about the wide world of gasoline additives will make more of a difference to your ear’s well-being than you might suppose ,but springing for premium won’t


MY LAWNMOWER WOULDN’T START, SO I PUT IT IN my car and drove it over to the Tractor Barn. “You’ve got bad gas,” the mechanic said.

“Hmmm,” I said, attempting to sound as though I had suspected as much all along and, simultaneously, I knew he was pulling my leg.

“They put a lot of alcohol in gas now, even when they say they don’t,” he continued. “And when the alcohol evaporates, the gas goes bad.”

As it turned out, the mechanic wasn’t quite right about gasoline. (He also found a more expensive problem with my lawnmower.) But he got me thinking. I had always assumed that gasoline was some sort of elemental material that had certain variable properties, such as regularness and unleadedness, but that it was incapable of exhibiting either goodness or badness in the sense the lawnmower mechanic intended. As it turned out, I wasn’t quite right either.

Gasoline is different today from what it was even a couple of years ago, when I also didn’t know anything about it. Changes in automotive technology, prompted to a large extent by government regulations concerning fuel economy and engine emissions, have created some unexpected problems for petroleum refiners, car manufacturers, and drivers. Some of these problems affect automotive performance; others have implications for human health and the state of the environment. Most drivers are either misinformed about the problems or ignorant of them altogether.

What idiots!

Leaded and Unleaded

LIKE MANY PARENTS, I DREAD THE DAY WHEN MY child is old enough to ask, “Daddy, what makes a car go?” To prepare myself, I recently turned to Scott Corbett’s What Makes a Car Go?, a fascinating book with large type and well-drawn pictures. Corbett explains that at the heart of the matter are cylinders and pistons. “The cylinders are like six children, sitting at a long table, each waiting his turn to be fed,” according to Corbett. The pistons, which are “shaped like tin cans,” slide up and down inside these hungry youngsters, propelled by a burning mixture of air and gasoline—their food. “Cylinders and pistons work like popguns,” furthermore. Eventually the car pulls up to the fuel pump, toppling a neatly stacked display of oil cans. Fillerup!

Of course, it was all a good bit simpler in Corbett’s day. What Makes a Car Go? was published in 1963. If you tried to run your nice shiny 1987 Chrysler Le Baron on gasoline from the golden age of driving, you’d have a big, expensive problem on your hands almost before you could sputter over to Corbett’s house and shake your fist out the window. In the first place, gasoline usually shouldn’t be stored for more than six months or so; that twenty-four-year-old fuel would probably have turned to a gummy muck. In the second place, gasoline in the Corbett era contained additives that can cripple modern engines and emission-control devices.

The best-known of these old-fashioned additives is lead. Adding lead to gasoline is a cheap way to increase its octane rating. A few pennies’ worth (in the form of tetraethyl lead or tetramethyl lead, two compounds of lead, carbon, and hydrogen) can transform a tankful of relatively inexpensive low-octane regular gasoline into a tankful of relatively expensive high-octane premium.

Regular unleaded gasoline usually has an octane rating of 87 or 88, regular leaded measures about 89, and premium unleaded 92 or 93. Most drivers believe that these numbers are measures of a gasoline’s “power, ” comparable to the proof number on a bottle of Scotch. This is not correct. The octane number posted on a gas pump is, rather, a measure of that gasoline’s ability to prevent an undesirable engine phenomenon known as knock or ping. This is the rattling-milk-bottles sound you may hear under stressful (for your car) conditions, such as accelerating in a high gear on a steep incline with a broken lawnmower in your back seat. The higher the number on the pump, the greater the gasoline’s resistance to knock.

A properly functioning, non-knocking car engine burns fuel in a very orderly way. When a piston moves down inside its cylinder, a mixture of air and vaporized gasoline is drawn into the space above it; when the piston moves back up again, this mixture is compressed. A spark plug, located at the top of the cylinder, emits a flash of electricity that ignites the mixture. As it burns, it expands dramatically, propelling the piston back down in the cylinder. When the piston, on the rebound, moves back up again, a valve at the top of the cylinder opens to let out the exhaust, and then the process begins again. An automobile needs an electric starter to get its pistons moving in the first place, but once the process is under way, it is self-sustaining. The up-and-down motion of the pistons is transferred to the wheels by means of various rods, shafts, and gears, leaving the driver free to look for a parking space.

In a knocking engine, this orderly process goes awry. Heat generated by the compression of the fuel-and-air mixture, or glowing carbon deposits inside the cylinder, can cause parts of the mixture to ignite spontaneously, without any help from the spark plug. This trips up the neatly churning pistons. It also sets off shock waves that produce a knocking sound. If severe enough, these shock waves can damage engine parts. One rare but devastating form of knock can cause an engine to self-destruct in a matter of seconds. Called high-speed knock, this malady is little known outside Europe, where people drive just as fast as they want to. Here in America most knock is harmless.

Higher-octane gasoline resists knock by igniting less readily than lower-octane gasoline. Giving a car a higheroctane gasoline than it needs to avoid knock is almost always a waste of money. In fact, a car that never knocks is probably being overindulged. When the price of gasoline fell not long ago, many drivers began treating their cars to premium gasoline. These drivers felt like big spenders, but most of them weren’t getting anything for their extravagance except, perhaps, a crisp “Yes, sir!” from the teenager handling the pump.

There is one exception. A growing number of new cars are equipped with devices called knock sensors or knock detectors. These are little gizmos that listen for knock and, if they hear it, get rid of it by bring the spark plug a bit later. Doing this usually diminishes the performance of the engine. Drivers of such cars can sometimes improve both power and fuel efficiency by buying gasoline with a higher octane rating.

Not all octane ratings mean the same thing. My car knocks every once in a while on Texaco Super Unleaded but never on Exxon Supreme, even though both fuels have a posted octane rating of 92. There are several possible reasons for this. First, different cars respond in different ways at different times to different gasolines. Second, the octane number posted on a fuel pump is the average of two other numbers, and the precise composition of the average can make a difference to individual cars. Finally, Exxon is known in the fuel industry for “giving away” octane—that is, for selling gasoline that is slightly higher in octane than the number posted on the pump. The government tests unleaded gasoline to make sure that it measures up to its posted octane ratings, and some gasoline producers protect themselves by tossing in a little extra.

The opposite also happens. Unscrupulous dealers sometimes sell gasoline with an octane rating substantially lower than the number posted. New York City’s Department of Consumer Affairs cited 135 gas stations for octane violations in 1986. The major gasoline companies tend to police their stations fairly closely, but the problem is by no means limited to unbranded gasolines. Of the nine New York stations that had three or more violations in 1986, five sold Texaco gas, one sold Getty gas, and one sold both.

Adding a gram of tetraethyl lead to a gallon of gasoline can, at negligible cost, raise its octane rating by as much as four or five points, or about the difference between regular and premium. Unfortunately, this useful additive is also a hideous poison. Lead can cause anemia and mental retardation, damage or impair nerves, and increase blood pressure. Leaded gasolines contain chemical “scavengers” that prevent excessive buildup of lead in engines by dispelling it, in the form of lead-bromide and lead-chloride gases, in automobile exhaust. These compounds are then breathed in, consumed, and absorbed by almost everyone. Lead’s effect on health is cumulative; once it settles in the body, it usually stays there.

In 1973 the Environmental Protection Agency proposed a five-step reduction of lead in gasoline. At the time, an average gallon of regular gasoline contained about two grams of lead, most of which ended up in the atmosphere. The EPA proposed a gradual reduction to half a gram per gallon—averaged over total production of both leaded and unleaded—by the end of 1978. The petroleum industry responded in a predictably hysterical fashion. As a result, the compliance date was held up for nearly two years and small refiners were exempted. A more stringent standard, limiting lead content to half a gram per gallon of leaded gas alone, went into effect in 1985.

On January 1, 1986, the government further reduced the maximum allow able amount of lead in leaded gasoline, to a tenth of a gram per gallon. Actually, most leaded gasoline still exceeds this limit. The EPA has permitted refiners to “bank” the difference between the maximum allowance and the amount they have actually used. A refiner that used less than half a gram per gallon in 1985, for example, earned “lead usage rights” that could be used later or sold to other refiners. A lively market for lead rights arose, with prices greatly exceeding the cost of the lead. The true level of lead in leaded gasoline is often double or triple the stated legal maximum.

This won’t go on for long. The industry’s lead bank is almost depleted. In addition, the market for leaded gasoline has shrunk. Beginning with the 1975 model year, most new cars were equipped with catalytic converters. These are anti-pollution devices that transform some of the most harmful parts of automobile exhaust into relatively benign substances. One harmful substance they can’t transform is lead, which clogs them in short order.

Gasoline producers were as grumpy about catalytic converters as they were about the detoxification of leaded gasoline. Many drivers were grumpy too. Some used plastic adapters to fit leaded-gas nozzles into the smaller fuel openings on their cars. Others removed the converters from their cars and replaced them with non-functioning “test pipes.” In 1983 a nationwide survey indicated that 16 percent of cars designed to run on unleaded gasoline were actually being run on leaded.

Most fuel-switching has an economic motive. Leaded gasoline has sometimes sold for fifty cents a gallon less than unleaded. The gap has closed considerably, but there’s still a difference at most gas stations. In my area leaded gasoline usually sells for about nine cents a gallon less than unleaded.

The apparent bargain is misleading, though. A few tankfuls of leaded gasoline can disable a catalytic converter; with continued fuel-switching the disabled converter can diminish engine performance. Even in cars whose converters have been removed, the use of leaded gasoline can increase maintenance expenses. Spark plugs and motor oil need to be changed more frequently. Lead can also cut fuel economy by five to ten percent. In 1985 the EPA estimated that improper fueling cost drivers about nineteen cents a gallon, mostly in increased maintenance. More important from a public perspective, cars with missing or disabled converters spew out vastly greater amounts of pollution than cars whose emissions systems haven’t been tampered with.

Gasoline companies bear much of the blame for fuelswitching. In their various battles against regulation, they have tended to exaggerate the cost of replacing lead as an octane-booster. Lead is indeed cheap, but the alternatives to it are not all as expensive as the industry has sometimes claimed. Furthermore, the industry has maintained an artificial price difference between leaded and unleaded grades as a marketing strategy. Gasoline companies often refer to leaded regular as their “fighting grade,” the fuel whose price they post prominently in order to bring customers in off the street. For many gas stations leaded gasoline is actually a loss leader. One station in my town sells 87-octane regular unleaded for $1.15 a gallon and 89octane regular leaded for $1.06. The cost of replacing lead to raise the octane rating in the unleaded regular does not account for this difference. In fact, if you removed the lead from the leaded regular, it would still be higher in octane than the unleaded, and thus presumably more expensive to produce than the regular unleaded. When I asked an oilcompany public-relations officer about this paradox, she suggested that I avoid “drawing too much of a relationship between the cost of producing gasoline and the retail price.”

The EPA could have greatly reduced fuel-switching from the outset by requiring gas stations to price leaded gasoline less attractively. Even with all the cheating by misinformed drivers, though, the average level of lead in Americans’ blood fell 37 percent between 1976 and 1980, the last year for which national figures are available. These reductions saved the nation hundreds of millions of dollars in medical costs alone, according to the EPA. The benefits have presumably increased in recent years, as more pre1975 cars have gone out of service and as the oil companies have begun to move away from leaded gasoline altogether. (My local Exxon station now sells three grades of unleaded and no leaded gasoline.)

Do we need leaded gasoline? In pre-1975 cars lead does help prevent an engine problem called valve-seat recession: lead from burned fuel coats exhaust-valve openings and protects them from deterioration. (The valve seats in newer engines are made of tough alloys that don’t need lead to protect them.) But valve-seat recession usually occurs only when unprotected engines are run at full throttle over extended periods. Old cars are seldom operated this way. Virtually all of them would work just as well with unleaded fuel as they do with leaded. They could work better, since their maintenance costs would tend to be lower.

Motorboats and farm vehicles, which often are operated at full throttle, do gain valve protection from leaded gasoline if their engines are old enough not to be equipped with hardened valve seats. Some of the noisiest lobbying against the EPA’s lead standards has come from boat owners. In the words of an article published in Lakeland Boating earlier this year, “While the EPA phaseout prevents brain damaged progeny, our enthusiasm is somewhat tempered by the death rattles of last decade’s marine engines.” Still, the cost of boat owners’ maintenance problems comes nowhere near the social cost of lead pollution.

Alcohol and Gasoline

ONE WAY TO INCREASE THE OCTANE RATING OF GASoline without using lead is to add alcohol. Ordinary ethanol, or grain alcohol, has an octane rating of around 100. Mixing one part ethanol with nine parts 87octane unleaded regular boosts the rating by about three points and only slightly reduces the energy content of the fuel. Methanol, or wood alcohol, has a similar effect. Unfortunately, both additives can cause problems.

Several forces made alcohol very attractive as a gasoline additive in the late 1970s and early 1980s. First, the EPA’s lead-reduction schedule created a demand for new sources of octane numbers. Second, uncertainties about oil supplies created an interest in renewable fuel sources. (Ethanol is usually manufactured from corn or other agricultural products.) Third, the high price of gasoline, in combination with various state subsidies and a federal tax incentive, made ethanol (which is otherwise relatively expensive) an economically feasible fuel-extender. Gasohol, a fuel much discussed in the late 1970s, consists of 90 percent unleaded gasoline and 10 percent ethanol.

The use of alcohol in unleaded gasoline is regulated by the EPA. In order to prevent the introduction of new fuels that might harm catalytic converters, the government requires fuel manufacturers to get permission from the EPA before selling fuels that are substantially different from those that were available before 1975. (The alcohol content of leaded gasoline is not regulated, since cars that legally use leaded gasoline don’t have catalytic converters.) The EPA has thus far approved three alcohol-and-gasoline formulations: one that includes up to 10 percent ethanol and two that include combinations of smaller quantities of methanol and other alcohols known as cosolvents.

Now that gasoline prices have fallen and other octaneenhancing methods have improved, the use of ethanol by the major petroleum refiners is limited. Still, something like 785 million gallons of ethanol were blended into gasoline last year, according to the Renewable Fuels Association, an ethanol lobbying group. Three fourths of this alcohol was used by small, independent refiners who had few other means of boosting octane. Where ethanol is used in this way, it always constitutes 10 percent of the fuel mixture: the EPA doesn’t permit using more than 10 percent, and the federal government doesn’t give tax breaks for using less. Since unsubsidized ethanol is substantially more expensive than gasoline, there’s no incentive for sneaking in more.

Methanol is quite a bit less expensive than ethanol, but at the moment, at least, it’s still too costly to be very attractive to fuel producers as a blending ingredient. The federal tax credit for alcohol-blending does not extend to fuels containing alcohol made from petroleum, coal, or natural gas, and most methanol is made from natural gas. Still, some methanol does make its way into gasoline, often as an octane-increaser in leaded gasoline sold at unbranded stations.

There are other reasons, in addition to cost, why alcohol isn’t more used in gasoline these days. Alcohol is a solvent for many materials for which gasoline is not. When poured into gasoline storage tanks, it can stir up slumbering substances that one highly trained petrochemical engineer described to me as “gunk.”

Alcohol also mixes with water; gasoline does not. When a blend of alcohol and gasoline is contaminated by water (from condensation inside a fuel tank, for example), a highly corrosive water-and-alcohol mixture can separate from the gasoline and accumulate at the bottom of the tank. Methanol (when unaccompanied by cosolvents) is particularly likely to cause this phenomenon, which is called phase separation. Methanol can also play havoc with plastic engine parts. Chrysler says that its warranties may not cover damage caused by using methanol blends. The owner’s manual for my Subaru, I recently discovered, says not to use fuel containing methanol “under any circumstances.” (It also says to avoid a steady diet of gasohol.) Some octane-enhancing additives sold at auto-supply stores contain little or nothing but methanol.

(“Neat” methanol—that is, methanol all by itself—may be the automotive fuel of the future if car manufacturers are willing to redesign their fuel systems to accommodate it. It produces little pollution, and provides a use for surplus natural gas. A number of refiners, automobile companies, and others are experimenting with it.)

Boat owners are upset about alcohol, too. Alcohol increases gasoline’s natural ability to eat through rubber fuel hoses. Leaking fuel is a more serious safety problem in a boat than in a car, since the fuel ends up not on the ground but in the bilges, where explosive vapors can accumulate. According to the Coast Guard, there are about 350 fuel-related fires and explosions on boats each year. What part alcohol plays in these accidents is impossible to estimate, since the evidence generally disappears in the accident.

Determining whether a particular gasoline contains alcohol (and, if so, what kind and how much) can be difficult even under normal circumstances. Forty-three states require alcohol labeling at the fuel pump, but only twentyone require that both the type and percentage of the alcohol be specified. Most people don’t read the fine print on fuel pumps anyway. Inquisitive sorts may wish to try a funsounding at-home test that BOAT/U.S. (as the Boat Owners Association of the United States prefers to be called) recommends:

Fill a graduated beaker, or similar measuring container, half full with gasoline. Add an equal amount of water. Shake the container for a minute, then let it sit for several minutes. The water will settle to the bottom and the gasoline will rise to the top. ... If there is no alcohol in the fuel the container will still show 50 percent gas and 50 percent water. If the gasoline has alcohol in it, the division will not be 50-50.

How Crude Oil Gets to Be Gasoline

THE MAJOR REFINERS DON’T NEED ALCOHOL, TO INcrcase their octane ratings. They’ve made up for the loss of lead by refining crude oil more severely, to produce more high-octane components that can be blended into fuel.

Gasoline is not a single substance, like water. It’s a blend of various hydrogen-and-carbon compounds, called hydrocarbons. Crude oil is a veritable stew of these compounds, which range in heft and complexity from dainty methane (each molecule of which consists of one carbon atom and four hydrogen atoms) to sluggish penta-octa-contane, or C85H172 (85 carbons, 172 hydrogens). In between methane and penta-octa-contane are a large number of compounds with individual properties and interesting names: propane, butane, pentane, heptane, decane, cyclohexane, benzene, toluene, and many others. Comparatively few of the hydrocarbons in crude oil end up in gasoline. Others go on to become kerosene, diesel fuel, heating oil, lubricants, solvents, plasticizers, polishes, waxes, graphite, asphalt, and many other products, including the fuel used to run oil refineries.

Separating these compounds out of crude oil is first a matter of distillation. Each compound has its own boiling point, the temperature at which it ceases to be a liquid and turns into a gas. Methane, for example, vaporizes at about -259° Fahrenheit (unless it’s kept under pressure or dissolved in something else, as it is in crude oil); hexadecane vaporizes at about 550°. When petroleum is heated, the components with the lowest boiling points—referred to in the distillation industry as the “light ends”—vaporize first. The less volatile compounds follow. As the resulting cloud of vapor gradually cools, the different compounds reliquefy one at a time and can be drawn off separately. At oil refineries this initial cooling and separating is done in tall towers called pipe stills. In a pipe still, reliquefying hydrocarbons are drawn off in groups. The group containing hydrocarbons with boiling points between about 90° and about 220° is called straight-run gasoline.

Straight-run gasoline has an octane rating of just 63 or so. Your car wouldn’t like it. To boost that number and enhance certain properties of gasoline, refiners blend in many additional components extracted from petroleum. Some of these are retrieved or created in complicated processes whose names seem to say to the layman, “You will never understand what I am”: naphtha reforming, catalytic cracking, hydrocracking, alkylation, isomerization. In one way or another, these processes transform hydrocarbons that are of little or no use in gasoline into useful hydrocarbons. In a hydrocracker, for example, great big hydrocarbon molecules are split into smaller molecules, yielding several high-octane blending components known collectively as hydrocrackate.

One of the most important post-lead octane-enhancers is butane, the stuff in cigarette lighters. Butane vaporizes at about 31° Fahrenheit at normal atmospheric pressure, but it dissolves easily in gasoline and has an octane rating of around 92. Blending it into gasoline not only raises the gasoline’s octane but also enables refineries to improve their profit margins, since butane adds volume but costs a lot less to make than other gasoline components, and since ordinary refining processes yield more butane than cigarette lighters need.

Butane is a highly volatile substance. That is, it evaporates very readily. The wavy-looking air you see when you put gas in your car is mostly escaping butane—as good a reason as any for not smoking at gas stations.

Some volatility is necessary in gasoline. A motor fuel has to evaporate before it can be burned in an engine. The fuel that goes into cylinders is a vapor, not a liquid. Once a car is up and running, vaporizing the gasoline is easy: the heat of the engine takes care of it. But starting a cold engine, especially on a cold day, requires a fuel that evaporates fairly readily all by itself.

Too much volatility causes problems, however. On a hot day (or on a cold day in a hot engine) excessively volatile fuels can cause a condition called vapor lock. This occurs when a pocket of evaporated fuel stymies the fuel pump, which can pump only liquids, and brings everything to a halt. Excessive volatility can also cause starting and restarting problems. As a result, gasoline makers sell different blends in different geographic regions at different times of the year. Some producers alter their blends each week to compensate for changes in the weather.

Highly volatile fuels also cause greater air pollution. All cars leak fuel vapors, even when they’re not in use. The vapors seep out of carburetors, fuel injectors, and other engine parts. Such seeping vapors are known as evaporative emissions. The more volatile the fuel, the higher the level of these emissions. The increasing use of butane as an octane-enhancer and fuel-extender has worsened the problem of emissions and the pollution they cause (alcohol worsens it too), and the EPA is considering new regulations to correct it. ‘The regulations will probably put limits on the volatility of gasoline and might also require improved vapor-recovery systems on new cars.

A similar type of pollution is known as refueling emissions. These are vapors that accumulate in cars’ fuel tanks as they empty and escape into the atmosphere when new fuel displaces them. California and the District of Columbia already require that gas nozzles be equipped with plastic collars that fit over tank openings and channel escaping vapors into underground recovery tanks. Similar systems may one day be required nationwide.

Detergents Fix a Fuel-Injector Problem

CONTROLS ON AUTOMOBILE EMISSIONS HAVE GIVEN us cleaner air. They’ve also given us more interesting automobiles. Cars in the 1960s were almost unbelievably inefficient, with huge engines and enormous appetites for gasoline. If the old 450-cubic-inch monsters of fifteen and twenty years ago had been engineered as carefully as today’s dinky power plants, we’d have been able to drive them to the moon.

In building the efficient engines necessitated by environmental restrictions, automotive designers have both borrowed from and contributed to the technology of highperformance driving. My humble Subaru has a microprocessor-controlled multiport fuel-injection system that would have made almost any driver green with envy just a few years ago.

Sharper engineering has also made trouble for drivers. One problem it caused appeared three years ago. It involved precisely the sort of fuel-injection system I have in my Subaru.

Until fairly recently most cars didn’t have fuel injectors. They had carburetors. A carburetor is a small, relatively simple device that receives gasoline from the fuel pump, allows it to mix with air, and, essentially, throws it at the engine block, letting the cylinders slurp up what they want. A fuel injector dispenses fuel in much more carefully measured amounts. The latest fuel-injection systems are incredibly precise. They give each cylinder exactly the amount of fuel that it needs, as determined by a tiny computer that monitors what the rest of the car is up to. This can produce big gains in fuel efficiency, emission control, and engine performance. But the engineering tolerances are very fine. The fuel opening in one of these injectors is about two thousandths of an inch, so even a tiny deposit can gum up the works.

Multiport fuel-injection systems began appearing in American cars in significant numbers in the 1985 model year. Some drivers noticed problems almost immediately—within a few hundred miles. Their engines balked, stalled, and had trouble starting. The problem was traced to fouled injectors. General Motors sent a letter to the oil companies suggesting that their gasoline was at least partly to blame.

The big refiners and the car manufacturers work together fairly closely. Each industry keeps the other informed of anticipated product changes, and each generally follows a set of standard fuel specifications established voluntarily through a nonprofit group called the American Society for Testing and Material, or ASTM. The ASTM standard for automotive gasoline is called D-439. It spells out octane levels, seasonal and geographic ranges for volatility, and many other variables. I ordered a copy and then put it on a shelf where I keep a lot of things I never look at; it’s full of numbers and graphs.

D-439 keeps the refiners and the car makers on more or less the same wavelength, but the injector problem was a surprise to everyone. There was no doubt that a problem existed. The big gasoline makers embarked on crash programs to solve it.

One of the first to offer a solution was Mobil. Like virtually all gasolines, Mobil’s had long contained detergent additives, designed to keep carburetors and other engine parts free of deposits. You may have seen Mobil television commercials showing gasoline sloshing around, washingmachine-style, inside the O in the company’s name. Mobil found that it could clean up fouled injectors by increasing the dosage of its standard carburetor detergent. It did so in its gasolines, and promoted its Super Unleaded Plus to mechanics and service managers as a solution to the multiport problem. Sales of Super Unleaded Plus rose 47 percent from 1985 to 1986.

Exxon also tried boosting its detergents, but wasn’t satisfied with the results. An Exxon engineer told me that the company had had two goals in approaching the problem: to create a formula that would clean up injectors in a single tankful, and to make the formula available in both regular and premium. The company spent six months devising and testing new detergents, eventually settling on one called XCL-12.

XCL-12 was added to all Exxon unleaded gasoline beginning in March of last year. Other big refiners made similar changes at around the same time. These changes have greatly reduced the injector deposit problem; indeed, they have eliminated it for most drivers, many of whom weren’t aware that the problem existed in the first place.

But car makers acknowledge that fuel-related enginetrouble will become increasingly likely as automotive technology advances. A report last year on the multiport problem for the Society of Automotive Engineers referred to injector deposits as “the tip of intake-system deposit problems.” The authors of the report, researchers at Chevron, had found that some detergents intended to clean injectors were dirtying intake valves—small valves on the cylinders that open to let fuel injectors squirt in fuel.

Engineers at BMW have been researching a similar problem since 1983. A few BMW drivers had begun having performance problems with their cars. BMW engineers performed exploratory surgery and discovered dirty valves. To find out what caused them, BMW asked six major oil companies to supply samples of their gasolines. BMW promised to share its findings privately with the six companies. To maintain confidentiality, the six fuels were identified only by code numbers. Each company knew its own code but not that of any of the others. The companies weren’t even told which other companies had participated.

BMW ran the six gasolines in test engines and measured their effect on intake valves. It found that three of the gasolines produced fewer deposits than a reference gasoline with no additives, and three produced more. BMW engineers suspect that old-style carburetor detergents may be at least partly to blame. In an earlier test, gasoline containing alcohol was also found to foul the valves.

To solve the problem in its cars, BMW has offered affected customers a one-time valve cleaning (using ground walnut shells) and encouraged them to buy little bottles of a BMW detergent additive, which its dealers sell for about two dollars apiece. Drivers are supposed to add a bottle to their fuel tanks every time they fill up. A simpler and decidedly cheaper solution would be to buy Chevron gasoline. BMW’s additive is made by Chevron. It’s the sameadditive that Chevron puts in its gasoline. (It can also be bought in bottles from Chevron. The brand name is Techron.)

Other car manufacturers don’t seem to be as worried about the BMW problem as BMW is. American car makers have been looking for the same deposits in their own cars but either haven’t found them or haven’t been concerned by what they’ve seen. Still, they are looking out for similar problems and conducting studies of their own. BMW is conducting more tests, including a study of large fleets of cars in actual service. Curiously, deposits in BMWs driven in Germany are different from those in BMWs driven in the United States. The German deposits are smooth and sleek; the American deposits are round and bumpy, like cauliflower. The difference seems to be caused by differences in driving speed.

Shopping Brands

FOR MANY YEARS THE FEELING AMONG SMART SHOPpers has been that all gasoline is alike. Consumers Union has long maintained that “gasoline is gasoline, whether it bears a well-known brand name or comes from an independent service station,”as it affirmed in a brief article in the April, 1985, issue of its magazine, Consumer Reports, In support of its thesis, the article quoted Vic Rasheed, the executive director of the Service Station Dealers of America, as saying that “additives are primarily just advertising gimmicks.” He also said, “If the branded and unbranded prices are close, most people would rather go to the branded dealers because they have some assurance of quality control. But it really doesn’t matter.”

Rasheed’s opinion has changed since then. “Our current position,” he told me recently, “is that there is a need to be very careful, very selective about gasoline today.”He also said that the odds of buying inferior fuel at retail are higher “when you buy an unbranded gasoline or an independent brand, or some gasoline whose origin you really can’t trace.”

Rasheed’s new position is less a flipflop than an acknowledgment that the world has changed. Until quite recently gasolines really were more or less alike, and most of the detergents they contained had little or no effect on engine performance. Now such additives can make a significant difference to many drivers. The change has created some exhilaration at the big oil companies. An Exxon engineer told me that the push to develop XCL-12 had been very exciting.

There’s probably more excitement to come. Possible fuel-related problems are of great concern to auto makers, who have lately begun to offer five, six, and seven-year warranties on some of their cars. None of them wants to have to provide expensive fuel-related maintenance for such extended periods.

Many unbranded-gasoline dealers buy detergents and even gasoline from the big companies. But many of them don’t. At the very least, drivers of cars with multiport injectors (which will soon be standard equipment on most cars) should be using fuel containing detergents known to keep injectors clean. For most such drivers, that probably means buying gasoline from stations with familiar names. Even among well-known brands, differences exist. Drivers should shop around if they notice performance problems. Some cars simply work better with some gasolines than with others.

Lawnmowers, too, I suppose. About nine months after being told I had bought bad gas, I decided I’d better do something about it. I still had about a gallon in a can in the garage. In all likelihood it had not been bad when the mechanic had said it was; it had been bought just six weeks before. But by now it was undoubtedly past its prime. How to get rid of it?

I didn’t want to pour it down a drain or onto the ground. Gasoline is a dreadful pollutant. Because gasoline contains known and suspected carcinogens, even contamination of one part per million in drinking water could render it impotable. (Leaking gasoline is a problem of enormous proportions, and one that will worsen in coming years as more and more underground storage tanks deteriorate with age.) So I took my can of bad gas down to my local gas station and asked the owner if he would get rid of it for me. He offered to pour it into his trash hopper. This struck me as a very bad idea, so I took my gas back home and let it sit a while longer.

Then the grass started to grow again. I had to get rid of my old gas so that I could buy some new. I thought about just pouring the old gas into my lawnmower anyway. But a year’s worth of gummy residue, condensation, rust, and dirt would plug up the engine for sure.

So I did the only other thing I could think of. I looked around to be certain that no one was watching, and I put it in my wife’s car.