The Cage of Trapped Electrons That Holds Every Photo on Your Phone

Hold the object in your mind first, because everything else radiates from it. Inside the phone in your pocket, a photograph has no physical existence you could point to. It is not ink, not a groove, not a magnetized stripe. It is a pattern of electric charge — a few thousand electrons here, none there, repeated across billions of tiny chambers etched into a wafer of silicon the size of a fingernail. And the electrons stay put. You can power the phone down, leave it in a drawer for a year, and the picture is still there, because each handful of electrons is sitting inside a cage it physically cannot climb out of.

The cage is called a floating gate. It is one object, tens of nanometers across, and the rest of this story is the same object seen four ways: what it is, who first drew it, who made it usable, and what that was finally worth.

The cage itself

A floating gate is a small modification to an ordinary transistor. A normal transistor is a switch: a control gate, when given voltage, lets current flow through a channel underneath it. The floating gate is a second gate slipped into the sandwich — one wired to nothing at all, completely surrounded by silicon dioxide, which is essentially glass. An island of conductor marooned inside an insulator.

A gate connected to nothing sounds useless. The point is exactly that isolation. Force some electrons onto that floating island and they have no path back out. The glass around them is a wall a few nanometers thick that, at the energies of everyday life, electrons simply cannot cross. The charge just sits there. In modern devices it sits, undisturbed, for ten years or more.

And the trapped charge does something measurable. Electrons on the floating gate push back against the control gate above them, so the transistor now needs more voltage before it will switch on. Put electrons in, and the switching threshold moves; leave it empty, and it stays low. To read the cell, you don’t have to open it — you just test it at a voltage in between and watch whether current flows. Current means empty. No current means charged. One bit, held by a pinch of electrons that have nowhere to go.

Getting electrons onto an island wrapped in glass means cheating the wall rather than breaking it. At high enough voltage, quantum mechanics allows electrons to pass through the thin oxide barrier instead of over it — a phenomenon called Fowler-Nordheim tunneling — or to be slammed across it as “hot” carriers. Either way, once they land, the high voltage is removed and the wall reasserts itself. The electrons are locked in.

That is the whole device: a sealed trap you can fill, can’t accidentally empty, and can read without disturbing. The history is the story of who arrived at it, and when.

Who first drew it

In 1967, two researchers at Bell Labs in New Jersey published a short paper in the Bell System Technical Journal with a flat, unpromising title: “A Floating Gate and Its Application to Memory Devices.” The authors were Dawon Kahng and Simon Min Sze. Kahng had already helped build the basic silicon transistor that the entire computer industry would stand on; Sze would go on to write the textbook that a generation of semiconductor engineers learned from.

What they sketched was exactly the cage above — the island of conductor sealed in glass, charged by tunneling, read by its shifted threshold. Kahng and Sze had described, in 1967, the storage cell that nearly all the world’s portable data now lives in.

But their version could not hold what it caught. A wall thin enough to let electrons in also let them seep back out — Sze later recalled that the longest any of the early devices kept its charge was about an hour — and Bell Labs, unimpressed, put the work on the shelf. Later engineers thickened the wall until the charge stayed put, but for years the floating gate lived only in read-mostly memory chips that you programmed once and rarely changed. Nobody had made it something you could fill and empty at will.

Who made it usable

Fujio Masuoka was not supposed to be working on memory you could erase.

Born in 1943 in Takasaki, north of Tokyo, Masuoka took his doctorate at Tohoku University and joined Toshiba in 1971. By the late 1970s he had built memory chips for the company and watched the existing erasable technology, EEPROM, struggle with a basic inefficiency: every storage cell needed two transistors, which meant chips stayed small and expensive. The market was modest. Toshiba’s management saw no reason to chase it harder.

Masuoka saw a way to halve the cell. If you gave up the ability to erase one byte at a time, and instead wiped a whole block of cells in a single stroke, you could strip each cell down to a single transistor. Fewer transistors meant you could pack far more of them onto the same square of silicon — which meant more storage, at lower cost, every year as the transistors shrank.

His superiors were unconvinced, and told him, in effect, to drop it. He did not. Through 1980, against instructions, he kept working the idea in his own time. That year he and a colleague, Hisakazu Iizuka, filed the founding patent.

The name came from someone else in the lab. When Masuoka described erasing an entire block of memory at once — thousands of cells dumping their electrons in one instant — a coworker named Shōji Ariizumi said it reminded him of the flash of a camera. The name stuck. The thing that would eventually store billions of camera photographs was named, before any of them existed, after a camera.

In 1984, Masuoka presented his work at the International Electron Devices Meeting in San Francisco. The design he showed — later called NOR flash — was fast to read and could run program code directly, and the audience understood immediately what they were looking at.

So did Intel. The American company seized on the architecture and, in 1988, shipped a 256-kilobit NOR flash chip of its own, turning the idea into a major commercial product.

Masuoka, meanwhile, pushed further. NOR flash was quick but bulky. He devised a second arrangement, NAND flash, that wired the cells in series like beads on a string — slower to address any single cell, harder to manufacture, but far denser and cheaper per bit. He presented it at the 1987 IEDM, and Toshiba brought the first NAND chips to market two years later. NAND was the one that would matter most. It is the architecture inside every memory card, every USB drive, every solid-state disk, and every phone — the reason a thumbnail-sized chip can now hold a terabyte.

What it was worth

For inventing the technology that now generates tens of billions of dollars a year, Toshiba gave Masuoka a bonus of a few hundred dollars.

The relationship curdled. Masuoka felt the company had failed to back his invention when it counted and failed to reward it afterward; Toshiba, by some accounts, later moved to demote him. In 1994 he left to become a professor at Tohoku University, the school where he had trained. In 2004 he sued his former employer for a share of the profits his patents had produced. The case settled in 2006 for ¥87 million — roughly three-quarters of a million dollars, against a market his work had built into one of the largest in the semiconductor industry.

The cage runs straight through it all. In 1967, two men at Bell Labs described a way to trap electrons inside a wall of glass and read their presence without disturbing them; they published it and moved on, and the idea sat mostly unused. Thirteen years later, an engineer in Japan figured out how to wipe those traps a whole block at a time, working after hours against instructions, and turned the curiosity into the substance that holds nearly everything anyone now photographs, records, or carries in a pocket.