In the world before flash, most memory chips only retained data as long as the power was turned on. That meant important data and programs had to be stored on magnetic tape or a hard disk. But neither of those devices was ideal: They were slow to play back and susceptible to bumps and jolts, which could damage their delicate moving parts.
Masuoka's chip technology -- originally called EEPROM, or electrically erasable, programmable read-only memory -- offered a better solution. Not only could it save and erase data, it was faster than tape and disk drives and used less energy. And since it relied on electrons to hold information and had no mechanical parts, it was also less fragile.
COMING ATTRACTIONS. Today, the flash-memory chip market is growing at an astonishing rate, and Toshiba ranks second among the world's largest manufacturers, behind Korea's Samsung Electronics. The market could hit $25 billion this year, and might double by 2008. Samsung has produced the world's first flash-memory PC drives, and more such devices will surely follow, leading to light, shock-proof PCs that consume less energy. And Masuoka believes memory chips are the verge of another breakthrough that will help lower prices while boosting storage capacity.
Masuoka, 62, is now a professor at Tohoku University, in the northern Japanese city of Sendai. He spoke with BusinessWeek Tokyo correspondent Kenji Hall about the genesis of his idea, his working life, and his theory for building even better chips. Edited excerpts of their conversation follow:
How did you get started developing flash memory?
Simply put, I wanted to make a chip that could one day replace all other memory technologies on the market. In the 1980s, the market for data storage on PCs was dominated by magnetic tape and disk drives. DRAM [dynamic random-access memory, the chip that stores data and instructions for programs used on a PC] later became a key technology for PCs because of its fast data accessing and low power consumption, but tape and disk drives remained bigger. Their sales were around five times higher than DRAM.
I wanted to find a way to replace DRAM with another type of chip that could be made at a lower cost. In the mid-'80s, the market for hard disks and floppy disks exceeded 1 trillion yen ($90 billion). Now, it's even bigger. Going after that market was the obvious thing to do for me, but I didn't know anyone at Toshiba who shared my belief.
Did your idea win immediate backing from Toshiba execs?
Initially, my bosses didn't know anything about it. In the early 1980s, I was the sub-leader for a project making a 1-megabyte DRAM chip at one of Toshiba's factory. DRAM worked fine but I was determined to come up with a chip that could take over the memory market.
I started developing the idea in my free time at the factory in 1980 and filed a patent in 1981. I had informally asked four other workers on my team to help me out. We all had our jobs -- the flash memory project was like a part-time job for us.
It was a real struggle. At that time, we were using microscopes to look for blemishes and defects. The hardest part was testing whether the experimental memory actually worked, and whether the data could actually be erased the way we had designed it to. In late 1984, we announced our results at the IEEE International Electron Devices Meeting, and in February 1985, we showed a prototype 256K flash memory device to the ISCC [International Solid State Circuits Conference].
What happened next?
After we publicized our research, Intel (INTC
) asked for a sample. So we came up with a 256K flash memory chip. It was 5.7 mm by 5.8 mm [a quarter inch by a quarter inch].
That was 1985. It took another year to come up with a prototype. Even then, we were only working on flash memory on the side. After that, Ford Motor (F
) wanted a sample for its engine-controlling electronics. And that's when things really got moving -- Toshiba finally decided to do something about it. But Intel was the first to announce it would start mass-producing its own flash-memory chip, in 1987-88. [Toshiba came out later, with a different type of flash.]
Did you ever imagine that your invention might spark a revolution in digital music players and other gizmos?
Sure. Twenty-five years ago, Sony (SNE
) had just come out with its Walkman cassette-tape players. I predicted that chips would replace magnetic tape completely. In the early 1990s, you could get a 4-megabyte flash memory. Back then, flash was still mainly used for digital cameras. But I figured it would be a cinch to put flash chips into earphones, to make a headphone stereo -- much like the iPod.
Why did you quit Toshiba in 1994?
I had a boss who ordered me to take a post with no budget or staff. I suppose I might not have left Toshiba if I'd had a different personality.
You filed a lawsuit against Toshiba in 2004, alleging that you deserved compensation for your discovery. You're not the only Japanese inventor who has felt that way. Why did you sue Toshiba, and where does your suit stand now?
I can't say much about the case. My lawyer is handling it. I didn't ask for royalties. I asked that I be paid 1 billion yen [around $9 million], which I felt was my share of what the invention was worth. The Tokyo District Court hasn't reached a verdict.
I never received more than a pittance for my discovery. Yet, thanks to flash-memory chips, Toshiba is one of Japan's only thriving semiconductor makers. I'm glad Toshiba is profiting off [flash]. But I sued because I wanted to make a point: Unless Japanese companies treat their engineers better, this country will have no future. If they copied U.S. companies' incentives for engineers, Japanese tech companies might be better off than they are, now that Taiwanese and Korean companies have caught up. Many top execs in Japan just can't relate to the engineers.
Flash memory has become so widely used in portable digital music players, cameras, and cell phones, thanks to the advances in chip miniaturization. Yet analysts predict the chips will never be that cheap to make, or be able to store as much data as hard disk drives. Do you think flash has its limits?
We might see flash-memory chips approach the limits of miniaturization in five, maybe 10 years. But nobody knows for sure when or whether that will happen. I don't know either. And even if flash-memory chips reach their limits, I believe someone will come up with a breakthrough technology to get around it. In the next few years, I think we'll see the flash-memory market expand to $100 billion.
What are you working on now?
Current transistors are basically built in two-dimensional structures. My idea, which I named Surrounding Gate Transistors in 1988, is to have them three-dimensional. It would allow you to further miniaturize a chip by packing in more transistors. Memory would increase by at least a factor of 10. Power consumption would be lower, as well.
My aim is to commercialize a 50-gigabyte chip by 2010. There are plenty of people who don't believe me. But it's O.K. In technological research, the newest ideas never win immediate recognition. If everyone agreed with me, my ideas wouldn't be new.
What are your hobbies or interests outside your research?
I don't really have any. I can't separate my personal life from my research. Though I'm not really conscious of it, work is always in the back of my mind -- sometimes ideas come to me when I'm on walks or in my sleep. I devise new models in my head.
So initially, I don't use a calculator, physics formulas, or theories. Only after I've arrived at something concrete do I write it down, modify it with feedback from other researchers, and then test it in the lab.
At Toshiba, I was constantly working on my own projects -- outside the assigned work --on my days off or at the end of a day at work. I was most productive during long holidays. I suppose that makes me a workaholic.