Evolution of Non-Volatile Memory: Eli Harari’s system level floating EEPROM revolutionized solid-state memory

By Steve Brachmann
March 18, 2017

Eli Harari

Eli Harari

Imagine living in a world where every time you wanted to use the computer to format a document or edit a video file, you had to complete your work in one sitting and transfer it to a physical recording medium, like paper or magnetic tape. You couldn’t save your work for later because the computer’s memory is wiped every time the power is turned off. That world would be a world without non-volatile memory, the aspect of computing architecture which allows for the retrieval of stored information and which is unaffected by power cycles. A recent market research report from consulting firm MarketsandMarkets predicts that the global market for all forms of non-volatile memory, including flash, embedded and many other forms of memory, will increase by a compound annual growth rate (CAGR) of 9.96 between 2016 and 2022, when the entire sector is expected to be worth $81.51 billion USD.

This Wednesday, March 22nd, gives us an opportunity to chronicle an important anniversary in the development of one seminal form of non-volatile memory. Some 23 years ago, the U.S. Patent and Trademark Office (USPTO) issued a patent for a flash electrically erasable programmable read-only memory (EEPROM) technology invented by Eli Harari, a 2017 inductee into the National Inventors Hall of Fame and a co-founder of the American flash storage developer SanDisk Corporation (NASDAQ:SNDK). Harari’s career spans a pair of the most revolutionary discoveries enabling data storage on portable battery-powered devices like smartphones and digital cameras and many of the computing devices on which this article is being read.

Harari’s EEPROM is World’s First Electrically Programmable and Erasable Memory

Eli Harari earned his Ph.D. in mechanical and aerospace engineering from Princeton University in 1973 and shortly thereafter was employed by the microelectronics division of Hughes Aircraft Company. Harari’s doctoral thesis focused on the use of thin film silicon dioxide (SiO2) to create a floating gate in transistors for the long-term retention of electrons. Harari was also focused on the concept of electron tunneling, a quantum mechanical effect in which the wave properties of an electron may give it a limited ability to sneak past a barrier.

Harari’s work between the years of 1976 and 1978 laid the foundation for the invention of the floating gate EEPROM. Erasable programmable read-only memory (EPROM) was known at the time but that type of memory is not erased electrically, it was erased when exposed to ultraviolet light, making it problematic for use in embedded electronics. Harari’s work demonstrated that thin film SiO2 was suitable for the efficient and reliable conduction of electrons for both program and erase functions in a memory. Further, Harari’s device incorporated the quantum mechanical principle of Fowler-Nordheim tunneling, or field emission, a process by which electrons can tunnel through a barrier in the presence of a high electric field. Fowler-Nordheim techniques increase the probability of being able to measure electrons trapped within a floating gate and allows for data erasure to occur by reversing the voltage of the electric field.

Much of this pioneering work into 0-volatile semiconductor memory is reflected in U.S. Patent No. 4115914, entitled Electrically Erasable Non-Volatile Semiconductor Memory. Issued in September 1978, it claimed a process for fabricating a nonvolatile field effect transistor storage device including the steps of forming active field effect transistor regions including source, channel and drain regions, forming a first layer of dielectric material on top of the field effect transistor regions, forming a layer of resistive material on top of the dielectric material to form a first gate electrode, forming a second layer of dielectric material on top of and around the resistive layer to make the first gate electrode a floating gate electrode and then forming a second gate electrode on top of the second dielectric material layer and covering the entire length of the channel region. The close proximity of the floating gate electrode to the semiconductor body improves the reliability of using the tunneling techniques as a conduction mechanism for transferring charges between the floating gate and the semiconductor.

System Level Flash EEPROM Architecture Replaces Hard Disk Drives

Harari’s invention of EEPROM non-volatile memory kicked off a decade of major discoveries in solid-state memory for computing devices. Harari would leave Hughes to continue development in EPROM and EEPROM devices at Intel Corporation. Across the Pacific Ocean, engineers at Japanese electronics company Toshiba developed what became known as flash memory, so-called because the quick removal of blocks of data from a memory reminded one Toshiba engineer of a camera’s flash. Flash memory didn’t allow for the erasure of individual bytes in the way that EEPROM devices developed after Harari’s invention could but that simplified architecture allowed flash devices to store data more compactly. Also, although EEPROM devices allowed for the erasure of individual bytes, those devices must be totally wiped before new data could be stored while flash allowed individual blocks to be erased and rewritten without having to wipe the whole memory. Toshiba built a flash EEPROM device which employed thick SiO2 and did not employ Fowler-Nordheim tunneling, and invented NAND-gate flash memory which did employ thin SiO2 and Fowler-Nordheim tunneling.

The conventional flash technologies of the late 1980s had drawbacks other than the inability to erase individual bytes. The reliability of flash devices was much lower in data storage write and erase activities compared to code storage as data storage applications had very low tolerance of error bits occurring over multiple write/erase cycles. Individual cells in flash memory had a tendency to break down over time, contributing to this problem. This only becomes exacerbated as flash memory devices were scaled up to meet the requirements of more complex computing architectures.

Harari, still steeped in the development of non-volatile memory devices, thus recognized that a need existed to create a flash memory which could work on a system-level computing architecture. Having co-founded the SanDisk Corporation in 1988, Harari set about developing a flash memory array that could provide system-level reliability and reduce bit-errors to zero with a technology that proactively detected and replaced defective cells before any failures could develop. The flash EEPROM memory developed by Harari worked cooperatively with a controller for the detection and correction of defective memory cells.

This important advance in solid-state memory is reflected in U.S. Patent No. 5297148, titled Flash EEPROM System and issued to Harari and two other inventors on March 22nd, 1994. It claimed a memory card connectable to a computer system having an array of EEPROM cells partitioned into a plurality of flash sectors, each flash sector a group of cells erasable together as a unit and having a portion reserved for redundant cells, a memory controller controlling operation of the EEPROM cells, error detection means within the memory controller that detects defective cells in the array, defect pointers generated by the memory controller linking a defective cell’s address to a corresponding redundant cell substituting for the defective cell and a defective cell substitution means responsive to the defect pointers. This results in a flash memory array which can last through a longer lifetime of write/erase cycles, thus providing enhanced performance and the capability to serve as non-volatile memory in a computer system and thus obviate magnetic disk storage devices which were conventionally used.

Like many other innovations in computer architecture have been in the past, SanDisk’s System Flash architecture was too expensive to become much of a consumer success in the tech industry. However, as the National Inventor Hall of Fame’s bio for Harari notes, SanDisk continued to develop technologies which drove down the price of flash EEPROM devices by 100,000 thanks to the progress of Moore’s Law. One of these innovative technologies noted by the Hall of Fame were multi-level cells (MLCs) which stored two or three bits of data on each flash EEPROM transistor.

Harari retired as CEO of SanDisk in 2010, one year after he received the IEEE Robert N. Noyce Medal awarded for exceptional contributions to the microelectronics industry. In 2012, the floating gate EEPROM developed by Harari at Hughes was declared to be an IEEE Milestone. That same year, Harari was awarded the National Medal of Technology and Innovation, which is awarded to innovators who have contributed to the economic, environmental and social well-being of the United States. Over the course of his career, Harari was listed as an inventor on more than 180 U.S. patents.

A 2012 interview with Harari published by science and tech publication The Register reflected the inventor’s views that flash NAND memory would largely replace hard disk drives in all computing platforms by the end of this decade. Current flash memory developers are working on building 3D NAND chips with more than 32 gigabytes of data storage for smartphone platforms.

The Author

Steve Brachmann

Steve Brachmann is a writer located in Buffalo, New York. He has worked professionally as a freelancer for more than a decade. He has become a regular contributor to IPWatchdog.com, writing about technology, innovation and is the primary author of the Companies We Follow series. His work has been published by The Buffalo News, The Hamburg Sun, USAToday.com, Chron.com, Motley Fool and OpenLettersMonthly.com. Steve also provides website copy and documents for various business clients.

Warning & Disclaimer: The pages, articles and comments on IPWatchdog.com do not constitute legal advice, nor do they create any attorney-client relationship. The articles published express the personal opinion and views of the author and should not be attributed to the author’s employer, clients or the sponsors of IPWatchdog.com. Read more.

Discuss this

There are currently No Comments comments.