In the year leading up to the release of the iPhone 6, Apple invested more than $1 billion in an effort to make sapphire one of the device’s big selling points. Making screens out of the nearly unscratchable material would have helped set the new phone apart from its competitors.  

When Apple announced the iPhone 6 this September, however, it didn’t have a sapphire screen, only a regular glass one. And a month later, the small New Hampshire-based company chosen to supply Apple with enormous quantities of cheap sapphire, GT Advanced Technologies, declared bankruptcy.

Recent documents from GT’s bankruptcy proceedings, and conversations with people familiar with operations at Apple and GT, provide several clues as to what went wrong.

Sapphire must have seemed like a perfect material for a smartphone screen. It has long been used as a cover for luxury watches, and Apple has used it to cover the cameras and fingerprint sensors in some iPhones since October 2013. But making large pieces of sapphire—enough for a smartphone screen—would normally cost 10 times as much as using glass.

In 2013, GT, a company that had manufactured and sold furnaces for growing sapphire for many years, claimed it could cut the cost by two-thirds by increasing the size of its equipment and adapting the crystal growth procedures to make cylindrical crystals—called boules—that are more than twice as large as ordinary sapphire crystals.

Apple originally offered to buy sapphire-growing furnaces from GT. But according to sources familiar with negotiations, after five months Apple demanded a major change in terms, requiring GT to supply the sapphire itself. In fact, Apple wanted GT to build the world’s largest factory to produce the stuff—more than doubling the world’s entire sapphire production capacity.

Apple said in recent court documents that GT failed to produce “any meaningful quantity of useable sapphire.” GT’s bankruptcy filing provides clues as to why the company was unable to deliver.

Producing sapphire requires a very clean environment, but ongoing construction at the factory meant that sapphire was grown “in a highly contaminated environment that adversely affected the quality of sapphire material,” according to GT. It also requires uninterrupted supplies of water and electricity to regulate the temperature of the molten aluminum oxide used to form the boule. GT said that to save costs, Apple decided not to install backup power supplies, and multiple outages ruined whole batches of sapphire.

Technical problems would have been compounded by the fact that it takes nearly a month to grow a sapphire boule. There’s no way to directly monitor if a sapphire crystal is forming correctly inside a furnace, says Eric Virey, a sapphire industry expert from Yole Développement, a French manufacturing company. Even if a problem occurs at the beginning of the process, “You only see the results at the end of the month when you open it up,” Virey says. And so, even once you’ve discovered and attempted to correct a problem, it could be another month or more before you know if a fix has worked.

The problems evidently didn’t end with producing a perfect boule. GT said in the documents that there were problems with much of the sawing and polishing equipment used to slice the boule—equipment that it says Apple selected. For example, a diamond-wire saw that was supposed to cut sapphire in 3.6 hours took 20 hours to do it and had to be replaced. According to GT, problems like these increased the costs of processing the sapphire boule by 30 percent.

The terms Apple negotiated committed GT to supplying a huge amount of sapphire, but put Apple under no obligation to buy it. In its bankruptcy documents, GT would later accuse Apple of using “bait-and-switch” tactics, and said the terms of the deal were “onerous and massively one-sided.”

It’s unclear whether the failure marks the end of Apple’s attempts to make use of sapphire in smartphone screens. Apple is using sapphire for the screen of one version of its new Apple Watch—though it will not rely on GT to produce the material. But the techniques used to grow sapphire will need further development before we’re likely to see the material widely used in smartphone screens.

GT is working on one possibility that could eliminate the need for large crystals—slicing smaller ones into ultrathin laminates that could be added to conventional glass. That would slash the amount of costly sapphire needed, while still making screens nearly impossible to scratch (see “Cheap, Scratch-Resistant Displays”).