One autumn morning in 1998, a soft-spoken, ruddy-faced petroleum engineer named Dennis Phelps walked out of his office at the Atlantic Richfield Company in downtown Midland and drove a company car, a four-door Chevrolet, to M. T. Boultinghouse 11-7, an oil well that had just been drilled amid some cactus and a few scraggly mesquite trees twenty miles south of the city’s airport. He parked several yards from the wellhead and stepped inside a small RV that was outfitted with some tables and chairs, a computer, and a coffeepot. Taped to one wall was a long sheet of paper that looked like nothing but a series of squiggly lines: a seismic well log.
The RV was owned by BJ Services, a Midland firm that the Atlantic Richfield Company had hired to fracture, or “frac,” the well, a technique that involves pumping a sand-filled, gel-like fluid down the well’s pipe at high pressure. In a normal frac operation, the fluid shoots out through holes that have been bored into the pipe at certain depths, causing the adjoining rock to fracture, much as a car windshield splits into dozens of tiny cracks when struck by a hammer. The sand braces the fractures, most of them no wider than pencil lead, sort of the way timber props open a coal mine’s shafts. Then, hopefully, the oil that has been trapped in the rock flows through the fractures back to the well, seeps through the holes in the pipe, and is pumped to the earth’s surface.
In the Permian Basin, every oil well had to be fracked. It had been at least sixty years since a West Texas oilman had punched a hole into the 74,000-square-mile ancient seabed, heard a rumbling in the earth, and stood back to watch a geyser of black crude shoot into the air. In fact, by 1998, the Permian’s oil fields had been so heavily drilled that it was considered to be on its last legs. Sure, there was oil to be recovered, but it was a mere trickle compared with the glory days of the forties and fifties, when the region had the richest oil fields in the world, credited with fueling much of the Allied effort during World War II. Back then, a good Permian Basin well started out at six hundred or more barrels a day before leveling off within a couple years to one to two hundred. In the nineties, a typical well started off at forty to seventy barrels a day, then settled down to just five to fifteen.
And almost all of these Permian wells were being drilled in what is known as the Spraberry Trend, a 1.7-million-acre layer of silt and sandstone about a mile and a half underground. It was considered to be about the only place left where oilmen could count on hitting oil. Spraberry oil wasn’t difficult to find: Geologists knew the rock formations by heart, and petroleum engineers knew exactly how to frac those wells and keep them going for fifty years or more. But as one oilman put it, drilling a Spraberry well was “like watching paint dry. You know where to drill, you drill, you eventually get your ten or so barrels of oil a day, and then you drill another one.”
In 1995 the Atlantic Richfield Company, or ARCO, which was headquartered in Los Angeles, had decided to drill about three hundred Spraberry wells. According to the gossip at Midland’s Petroleum Club, ARCO was preparing to put itself up for sale, and the more oil reserves it could show on its books, the higher its value as a company would be. That’s where Dennis Phelps came in. His bosses had let him know in no uncertain terms that his job was to make sure those new wells hit oil. Phelps got the message. If he started drilling dry holes, chances were that he’d be in the next wave of ARCO layoffs.
An ARCO employee since 1970, the 52-year-old Phelps was then just one more petroleum engineer working the Permian Basin. He dressed in inexpensive button-down shirts, khakis, and loafers, and he spent most of his time in a small, spare office studying such subjects as “relative perm value,” “in situ proppant,” and “residual fracture aperture.” In his notebook, he scribbled down ideas about polymer concentrations in frac fluids. The blackboard in his office had phrases on it like “surface tilt fracture mapping.” As opposed to the swashbuckling Jett Rink-like wildcatters who had made Texas famous, Phelps had the charisma of an accountant.
All petroleum engineers, however, have a bit of a romantic streak in them. They love to spend their idle moments trying to devise some new technique that will pull more oil out of the ground. Phelps was no different. For years he had been wondering if there was a better way to frac what was known as the Wolfcamp, a layer of rock—mostly packed limestone—that ran directly below the Spraberry. Every now and then, an ambitious oilman would drill a little ways into the Wolfcamp, maybe a few hundred feet, and bore holes into his pipe because his well logs suggested there was oil lurking nearby. If he was lucky, he’d get a hundred or so barrels. But because the Wolfcamp’s limestone had such little permeability, the fracs never worked: The frac fluid simply bounced off the rock. “As far as we were concerned, the Wolfcamp was nothing but damn sorry rock,” said legendary Midland oilman Clayton Williams, who started drilling in the Permian Basin in 1957. “If you were lucky, you’d get a teaser well, but without a good frac, the oil quickly would dry up.”
It just so happened that ARCO was using a rig for its new Spraberry project that drilled wells 10,000 feet deep, straight into the heart of the Wolfcamp. Phelps received permission from his bosses to try a few experiments. In one test, attempting to put cracks in the Wolfcamp, he sent the frac fluid at a much higher speed down