Like exploding bottles, these episodes also find their way to court. Sometimes the cases involve cigar butts, cigarettes, a condom, a tampon, or a bandage. One case involved a cigar and an insect, in the same drink. Many cases involve slivers and flakes of metal, or bits of rust. Plaintiffs have won cases for drinks containing ants, bees, centipedes, cockroaches, flies, grubs, maggots, moths, worms, roaches, a yellow jacket, a wasp nest, a black widow spider, and a snake.
Plaintiffs have found in their drinks: a roach egg, “a partially decomposed worm or cocoon,” insect larvae, a fly covered in fungus, a putrefied mouse, a mouse skeleton, “a rat with the hair sucked off,” a piece of flesh, something identified as part of a mouse or bird, rotten old meat, and “blood vessels of unknown origin.” In three dozen cases involving dead mice, the court has found for the plaintiff. Plaintiffs have vomited blood, excreted blood, had ulcers and dysentery. Many have been bedridden for weeks or felt “deathly sick.” Many have been hospitalized. One had stomach surgery; another lost thirty pounds in five weeks. Most never drink soda again, out of the perfect container or otherwise.
It’s hard to say how many of these cases are legit, because there’s also the story of an Illinois man named Ronald Ball. Mr. Ball claimed that on November 10, 2008, he bought a can of Mountain Dew from a vending machine on the outskirts of St. Louis, opened it, took a sip, spit it out, vomited, then poured the rest of the can’s contents into a Styrofoam cup, whereupon he discovered a dead mouse. He called the number on the can, complained, and, as requested, sent the mouse and the remaining Mountain Dew to PepsiCo, its manufacturer.
From the serial number on the can, Pepsi determined it had been filled seventy-four days before Mr. Ball opened it. To assess the mouse, Pepsi sent it, in the Mountain Dew, in a glass jar, to a veterinary pathologist in Salt Lake City named Lawrence McGill. McGill, who had conducted thousands of necropsies, and claimed familiarity “with the effects of an acidic fluid” on mice and other animals, opened up the jar, and put the mouse in formaldehyde, to arrest its decomposition. The next day, he cut open the mouse, and went hunting for clues. In the mouse’s leg and head, he found bones; in the mouse’s unruptured abdominal cavity, he found a liver, intestine, and stomach; and in the mouse’s lung, he found cartilage cells—indicating that the mouse had been in the Mountain Dew no more than a week. Because he was unable to open the mouse’s eyelids, he deduced that the mouse was young, at most four weeks old. The Mountain Dew, he determined, had a pH of 3.4. The conclusion McGill reached was simple, based on unanimous, irrefutable evidence: the mouse did not exist at the time the can was filled and sealed, and had not been stewing in Mountain Dew for more than a week, let alone seventy-four days. The pathologist said as much in a signed, notarized affidavit dated April 8, 2010.
Three weeks later, by which time Mr. Ball complained that the dissected mouse (which had been returned to him) had been rendered unfit for further testing, he sued for $325,000. Pepsi’s lawyers got straight to the point. Citing McGill’s findings, they said that Mountain Dew was so corrosive that, had a mouse been subjected to the concoction for seventy-four days, as Mr. Ball alleged, it would have dissolved into a “jellylike substance,” leaving no identifiable mouse parts behind. The judge vacated that case. Mr. Ball sued again, seeking $50,000 in relief. The judge dismissed that case too. Before it was all over, the news-of-the-weird story was on every channel in America. Eric Randall, in the Atlantic’s blog, summed it up best. He wrote, “This seems like a winning-the-battle-while-surrendering-the-war kind of strategy that hinges on the argument that Pepsi’s product is essentially a can of bright green/yellow battery acid.”
Therein lies the crux of designing the perfect beverage container: either the container does not want to cooperate, or the beverage makes containment difficult. Or both.
In the case of the aluminum can, it’s both and then some. For the country’s largest can maker, the aluminum that is mined in Dwellingup, Australia, and smelted in Evansville, Illinois, and manufactured in Golden, Colorado, into twelve-ounce cans behaves throughout the can-making process begrudgingly. It tries to jam the machines that bring it into existence, and strays toward a number of disastrous scenarios: it may stretch, crumple, fracture, collapse, pleat, buckle, or blister. If it emerges as a healthy can, it still aspires to misbehave, by refusing to protect whatever beer, soda, energy drink, or other “product” ends up inside it. It wants to interact with the product, and change its taste. Worse, the can still finds more ways to throw a tantrum. Beyond exploding, it may leak, or somehow corrode: from the inside out, or the top down, or the bottom up. Rust is a can’s number one enemy. Manufacturing strong, healthy aluminum cans, in fact, is so challenging, and requires such a vast amount of study, design, and precise machining, that many consider cans the most engineered products in the world. This notion—that the ubiquitous aluminum can, which seems anything but amazing, is in fact incredible—was the first thing I learned at Can School.
The second thing I learned at Can School is that of all the operations used to stave off aluminum’s suicidal tendencies, wheedle the metal into submission, and avert what more than one can-making employee called time-bomb behavior, the corrosion-related procedures are so sensitive and shrouded in secrecy that asking lots of questions about them is a good way to get kicked out of Can School.
Can School was put on by America’s largest can maker. Over three days, in the spring of 2011, engineers, chemists, and managers from the company discussed “improved pour rates” and “recloseability” and the overall “experience from a can.” Except they didn’t call the common 12-ounce can a can. They called it a 202 (because the diameter across the top is 2 2/16"). Most wore cell-phone holsters, and many sported mustaches. One described cans as “like sunshine”; another discussed the “opening performance,” and didn’t mean the opera. Nearly sixty attendees—from Heineken (Mexico), MillerCoors, Nestlé, and Pabst—listened attentively at four long tables in a semicircular conference room just north of Denver.
On that first day—before I nearly got kicked out—to my right sat three women from Pepsi, with New York accents. To my left sat a man from Anheuser-Busch, who told me that he hedges a year out on the London Metal Exchange on $1.5 billion of aluminum annually. Left of him sat a guy from the Dairy Farmers of America. I heard them conferring about Americans’ milk-drinking habits. In front of me sat two guys from Coca-Cola. Behind me sat three more guys from Coca-Cola. One attendee wore a shirt with a patch that said “Can Solo.” Another gave me a business card with a picture of six canned beers and his title: Can Whisperer.
Displayed prominently on a screen was the motto “eat. drink. imagine.” On the left wall was a table covered in various food cans, from Crisco to Chef Boyardee. On the right, various beverage cans: Molson, Labatt, Foster’s, Pabst. At my seat: a black folder, and two posters, showing the steps of the can-manufacturing process. There was nothing rusty about it. Beside the screen stood a podium, and to its left, an American flag. To its right stood the baby blue flag of the Ball Corporation.
If you drink beer, or soda, or juice, or water, or sports drinks, or coffee, or milk, or, really, anything, or if you have ever preserved fruits or vegetables in glass jars, you may recognize the name. Grab a can of beer and start hunting. The Ball logo—cursive, underlined, and slanted upward—is minuscule, but it’s probably there. On a Pabst can, it’s about an inch below the rim, above the bar code, above the sell-by date, just above the government warning, just right of the word problems in “may cause health problems.” But it’s not on every can; it’s up to the product manufacturer to decide if it wants the Ball logo there at all. It’s on Miller Lite, Stroh’s, Heineken, Schlitz, Miller High Life, Tecate, Colt 45, Blue Moon, Honey Brown, Stella Artois, Dr Pepper, Mountain Dew, Pepsi, Coke, Schweppes, Izze, and Starbucks, but it’s not on Monster Energy, Budweiser, or Bud Light.
The Ball Corporation has been running Can School annually for twenty-five years; not quite a thousand people in the
beverage industry have graduated. To say the company is qualified to teach the course is a gross understatement. The people of the world go through 180 billion aluminum beverage cans a year. That’s four six-packs for every person on the planet. The United States and Canada gobble up more than half of them—100 billion a year—and Ball makes a third of these. (Two other companies make the majority of the rest.) Ball, which has a long history with containers, operates thirteen beverage-can plants in Europe, five in China, and five in Brazil. It also operates fourteen steel food can factories in the United States and one in Canada. In pursuit of making cans, Ball employs fourteen thousand people. The total area of just the company’s American beverage can factories approaches six million square feet. Ball is the third largest manufacturer in China, the second largest manufacturer in Europe, and the undisputed major player in the US beverage can industry. The company makes about a quarter of the world’s beverage cans.
Since Ball seriously entered the can market, in the 1980s, its stock has outperformed DuPont and American Express, and trailed just behind ExxonMobil. Since 1994, when Ball first made $1 billion worth of cans, its compound annual growth rate has exceeded 12 percent. Since 2002, when Ball bought Schmalbach-Lubeca AG for $1.18 billion, it’s held the title of largest can manufacturer in the world. In 2009 the Fortune 500 company acquired four more US plants, in Georgia, Ohio, Florida, and Wisconsin, for which it paid more than half a billion dollars, and as a result of which its market share increased to 40 percent. That same year, Ball sold $4.6 billion worth of aluminum beverage cans, earning $300 million in profit. Ball makes satellites, too—but they’re not the big moneymaker. The big moneymaker is beverage cans, at a dime apiece. Sell forty billion of them—earning two-thirds of a penny on each—and you’ll make out better than the Dow Jones Industrial Average and the S&P 500. But that doesn’t mean it’s easy.
On account of corrosion, it took engineers 125 years of tinkering with steel can designs before they figured out how to encase beer within, another quarter century for them to wheedle aluminum into service, and most of another decade for them to swap the beer with Coke. Consider a can of Coke. It’s a corrosion nightmare. Phosphoric acid gives it a pH of 2.75, salts and dyes render it still more aggressive, and the concoction exists under ninety pounds per square inch of pressure, trying to force its way out of a layer of aluminum a few thousandths of an inch thick. It sits there for weeks, months, years, often in a humid fridge, or dank pantry, or hot trunk, or stagnant warehouse. That the can doesn’t corrode is a technological marvel. That we are capable of reproducing that result hundreds of billions of times over—with a failure rate of 0.002 percent—is an unheralded corrosion miracle. And Coke is just the beginning. In the near half century since it was first canned, we’ve packed more corrosive beverages, such as San Pellegrino, V8, and Mountain Dew, into cans that have concurrently gotten thinner and more delicate.
All that protects the meager aluminum is an invisible plastic shield. Industry insiders call it an internal coating, or IC, and it’s the product of a phenomenal amount of work. This plastic must be tough but flexible. It also must be rheologically suitable in terms of viscosity, stability, and stickiness. Without this epoxy lining, only microns thick, a can of Coke would corrode in three days. Our stomachs are stronger than the aluminum. But it’s becoming increasingly evident that other parts of our bodies may not be stronger than what’s in the epoxy. That’s why the beverage container industry would rather not talk about rust, and why I nearly got kicked out of Can School.
Before coating the insides of their cans, Ball needs to know how corrosive the product—the beverage within—will be. The epoxy coating, after all, isn’t free—it costs about a half penny per can—and Ball doesn’t want to waste it. Also, some beverages are so corrosive that no amount of coating will protect their cans. Ball is not in the business of sending cans out into the world to be slaughtered by overaggressive liquids. The coating must perform. Otherwise, cans explode, and legal costs climb.
Ed Laperle, a tall, thin corrosion engineer at Ball, told me that until twenty-five years ago, Ball solved this problem by studying test packs. Five months before Can School, Laperle gave me a tour of the company’s corrosion lab, which it calls its Packaging Services lab. Laperle, whose white beard makes him look astonishingly like Bob Vila, told me how the old system worked. Until the 1990s, he would tell customers, “Okay, send me some, I’ll put it in a can, on a shelf, for six months and see what happens.” Cases just sat there. Customers had to wait. Eventually Laperle figured out whether the beverage demanded a can with a windbreaker or a down jacket, and called back and said, sure, we can put your product in our cans. It didn’t take long for customers to tire of the wait. Scott Brendecke, a corrosion engineer who works for Laperle, described the pervading logic that evolved. Customers cited historical performance, hoping to weasel out of test packs. “People would say, ‘Oh, this root beer is just like that cola,’ ” Brendecke said. Like any good engineer, he disliked the logic. “The category of ‘similar’ just broadens over time,” he said. “There’s no learning. Eventually failure.” By failure, Brendecke meant leaks and explosions.
Then Ball figured out how to do the corrosion test in about four hours. They call it a pitting scan. Using a potentiometer the size of a desktop computer, and a simple wire-frame setup that looks like it cost $100 at Radio Shack and would be suitable for a high-school chemistry lab, the engineers in Laperle’s lab apply a tiny DC current to a sliver of encased aluminum sitting in the liquid in question, which sits in a glass jar. Over four hours, the potentiometer spits out a graph of current versus time. It looks like a pyramid. The peak represents the pitting potential, or PP, of the liquid. The pitting potential reveals how much current is required to remove the outer layer of aluminum oxide from the aluminum sliver. Once that outer layer is removed, corrosion can proceed unabated.1
By plugging the pitting potential and some other measurements (salt, copper, chloride, dye, dissolved oxygen, pH) into a carefully guarded equation, Laperle’s team can determine the corrosivity of the product. The corrosivity, in turn, determines the thickness of the coating on the can it’ll be in. Beer, for example, isn’t very corrosive, so coatings on beer cans are extremely thin, and weigh in the neighborhood of 90 milligrams. Beer just happens to be amenable to coexistence with aluminum, because of its mild pH and some other convenient traits, which I’ll get to later. Coke, which is more corrosive, demands a heavier coating. Especially acidic beverages, like lemon-lime drinks, and salty, or “isotonic,” drinks, such as V8, demand greater corrosion protection, and hence thicker coatings of up to 225 milligrams. The coating thicknesses are referred to as A, B, and C. Neither Laperle nor anybody else in the can-making industry, though, would tell me exactly what they are. The most anyone revealed is that the average is 120 milligrams per can.2
Laperle has come across many products that are too corrosive to put in a can with even the thickest coating. If a beverage’s corrosivity is over the line—if, say, it’s a particularly potent “bright green/yellow battery acid”—Laperle will tell the product manufacturer that the beverage needs to be modified before it can be put into a can. When Laperle does this, over the phone, he is direct. “I tell ’em straight up, ‘Your product failed. You can reformulate. Here are some suggestions.’ ” He’ll advise raising the pH or cutting down the dyes. He has figured out that if he gives the manufacturer a goal, the conversation will go more smoothly. “If you just tell ’em ‘It failed,’ they hang up unhappy,” he said. Specific cases he refused to cite. A few times, with products that were an order of magnitude too corrosive to put in cans, no matter what Laperle tried, he called the manufacturer and said, “We’re kinda done. There’s nothing else I can do for you.” Yes, the can wags the product.
This much is clear: the higher the pitting potential of a beverage (anywhere from 100 to 500 millivolts), the fewer cans will fail. This much is also clear: sodium benzoate is bad. Copper is bad. Sugar is good
. It absorbs carbon dioxide, decreasing the pressure within a can, and it also inhibits other corrosion reactions, because sugar tends to deposit onto pores in the coating. Thus, Diet Coke underperforms regular Coke on at least two counts. Citric acid and phosphoric acid are equally bad. Red #40 is pretty bad, and high chlorides are really bad, and together, they’re really, really bad. At Can School, degrees of badness were represented on a graph of pitting potential versus chlorides, with a line asymptotically falling from the comfort zone, through the worry zone, and into the extreme worry zone. The extreme worry zone is where failures happen, where cans corrode from the inside out. When that happens, you get explosions and then lawsuits.
Brendecke later explained internal corrosion with a good metaphor. He held up two hammers, one in each hand. In his right hand, a metal hammer; in his left, a toy hammer, made of inflatable purple plastic. “You can bang on your beverage can all day,” he said, swinging his left hand, “and nothing will happen.” Or, he said, you can add corrosive elements, until the weight of the head of the plastic hammer resembles that of the metal one.
But the pitting potential is measured in a jar in a lab, rather than in an actual sealed can out in the world, so it’s not absolute. That’s why the corrosion engineers do a few more tests, to examine the package-product interaction, or PPI, of the coating and the beverage. As in the old days, they do test packs, storing eight cases for at least three months, so that they can determine if any of the products are absorbing any metal. Metal pickup must be less than two parts per million. They use a spectroscope to check. Laperle and his staff also test their coatings, using electro-impedance spectroscopy, or EIS. To do this, they modify the pitting scan setup a bit. Instead of using a sliver of uncoated aluminum, they use a two-inch square of coated aluminum. Instead of using a beverage, they use a salty and acidic corrosive liquid called liquor 85. And instead of applying DC current, they apply AC current across a range of about forty frequencies, from about 100 kilohertz to 10 millihertz. The resulting relationship between voltage and current, measured over a couple of days, reveals the values of resistors and capacitors in a model of the various electrochemical reactions taking place. Modeling this impedance, which involves deconvoluted voltages and nonfaradaic components in the inner and outer Helmholtz planes, takes most of a PhD in chemistry. Calculating it takes a good potentiostat. The result yields the strength of the coating.
Rust: The Longest War Page 10
