First Use of Bubble Memory in Computers

Building on discoveries of the properties of certain ferrimagnetic materials under an applied magnetic field, scientists began to explore the use of these substances for the storage of computer information, leading to commercial products in the late 1970’s.

Summary of Event

The fanfare over the commercial prospects of magnetic bubbles ignited in the public domain on August 8, 1969, from a report appearing in both The New York Times and the Wall Street Journal. The early 1970’s would see the flames spread (at least in the computer world) with each prognostication of the benefits of this revolution in information storage technology. By the late 1970’s, however, it was apparent even to diehard advocates that magnetic bubble technology, although sound in theory, would need to benefit from further innovation in the laboratory before conquering the world of computer data storage. It became a victim of high production costs coupled with low reliability as well as the dramatic gains in performance and cost-benefit ratios of competing conventional storage technologies. A few flickers of research shone at a diminishing number of industrial research and development laboratories in the 1980’s, mostly in Japan. Computers;memory
Bubble memory, computer
Magnetic storage devices
[kw]First Use of Bubble Memory in Computers (Aug. 8, 1969)
[kw]Bubble Memory in Computers, First Use of (Aug. 8, 1969)
[kw]Memory in Computers, First Use of Bubble (Aug. 8, 1969)
[kw]Computers, First Use of Bubble Memory in (Aug. 8, 1969)
Bubble memory, computer
Magnetic storage devices
[g]North America;Aug. 8, 1969: First Use of Bubble Memory in Computers[10390]
[g]United States;Aug. 8, 1969: First Use of Bubble Memory in Computers[10390]
[c]Computers and computer science;Aug. 8, 1969: First Use of Bubble Memory in Computers[10390]
[c]Science and technology;Aug. 8, 1969: First Use of Bubble Memory in Computers[10390]
Bobeck, Andrew H.

Although it was a new disclosure to the public at large that August day in 1969, magnetic bubble technology had held the technical interest of a small group of researchers around the world for many years. The laboratory that probably can claim the greatest research advances with respect to computer applications of magnetic bubbles is Bell Telephone Laboratories Bell Telephone Laboratories (later part of American Telephone and Telegraph). Basic research into the properties of certain ferrimagnetic materials started at Bell Laboratories shortly after the end of World War II. Ferrimagnetic substances are typically magnetic iron oxides. Research into the properties of these and related compounds accelerated after the discovery of ferrimagnetic garnets in 1956 (these are a class of ferrimagnetic oxide materials that have the crystal structure of garnet). Ferrimagnetism is similar to ferromagnetism, the phenomenon that accounts for the strong attraction of one magnetized body for another. The ferrimagnetic materials most suited for bubble memories contain, in addition to iron, the element yttrium or a metal from the rare earth series.

It is an interesting story of fruitful collaboration between scientist and engineer, between pure and applied science, that produced this promising breakthrough in data storage technology. In 1966, Bell Laboratories’ scientist Andrew H. Bobeck and colleagues were the first to realize the data storage potential offered by the strange behavior of thin slices of magnetic iron oxides under an applied magnetic field. The first United States patent for a memory device using magnetic bubbles was filed by Bobeck in the fall of 1966 and issued on August 5, 1969. It is important to recognize the collaborative nature of this work and the contributions of many of Bobeck’s colleagues.

The three basic functional elements of a computer are the central processing unit (CPU), input and output (I/O), and memory. Memory is further categorized into two types: primary (typically semiconductor-chip-resident memory used by the computer for essential operations of the system software and for user applications) and secondary (additional memory used for mass storage and archival purposes). Another distinction that can be drawn between primary and secondary memory is volatility. Most implementations of semiconductor memory require a constant power source to retain the stored data. If the power is turned off, all stored data are lost. Memory with this characteristic is called volatile. Disk and tape, typically used for secondary memory, are nonvolatile. Here memory does not rely on electrical currents to sustain its existence, but on the orientation of magnetic domains.

Andrew H. Bobeck of Bell Telephone Laboratories.

(AT&T Archives)

One can visualize by analogy how this will work by taking a group of permanent bar magnets that are labeled with N for north at one end and S for south at the other. If an arrow is painted starting from the north end with the tip at the south end on each magnet, an orientation can then be assigned to a magnetic domain (here one whole bar magnet). Data are “stored” with these bar magnets by arranging them in a row, some pointing up, some pointing down. Different arrangements translate to different data. In the binary world of the computer, all information is represented by two states. A stored data item (known as a “bit,” standing for binary digit) is either on or off, up or down, true or false, depending on the physical representation. The on state is commonly labeled with the number 1 and the off state with the number 0. This is the principle behind magnetic disk and tape data storage.

Now imagine a thin slice of a certain type of magnetic material in the shape of a 3-by-5-inch index card. Under a microscope, using a special source of light, one can see through this thin slice in many regions of the surface. Darker, snakelike regions can also be seen, representing domains of an opposite orientation (polarity) to the transparent regions. If a weak external magnetic field is then applied by placing a permanent magnet of the same shape as the card on the underside of the slice, a strange thing happens to the dark serpentine pattern—the long domains shrink and eventually contract into “bubbles,” tiny magnetized spots. Viewed from the side of the slice, the bubbles are cylindrically shaped domains having a polarity opposite to that of the material on which they rest. The presence or absence of a bubble indicates a 0 or 1 bit. Data bits are stored by moving the bubbles in the thin film. As long as the field is applied by the permanent magnet substrate, the data will be retained. The bubble is thus a nonvolatile medium for data storage.

Essentially, magnetic bubble memory is the fully electronic analog to disk or tape memory. Conceivably, this device could replace both primary and secondary memory. A computer system needs quick access to data stored in primary memory, and this is provided by bubble memory because it is electronic and not electromechanical, which is the basic mechanism for disk and tape technology. The property of nonvolatility qualifies bubble memory as a secondary or archival storage medium (at least for small to moderate storage needs). As with any engineering solution, there are trade-offs. Bubble memory is not as fast as semiconductor-based memory. One reason is that bubble memory is inherently a serial device in organization. The bubbles are stored on several paths called loops, and a rotating weak magnetic field applied at right angles to the direction of magnetization of the bubbles causes the bubbles (and lack of bubbles, called “holes”) to migrate around the loops. An organizational grid of patterns is superimposed on the surface of the substrate upon which the bubbles rest to create storage locations and enable orderly tracking of the bubbles along loops (or tracks) in the medium. Access time will vary with the migration rate and the number of storage locations being addressed.


Magnetic bubble memory created quite a stir in 1969 with its splashy public introduction. Most of the chip manufacturer research and development laboratories immediately instituted bubble memory development projects. Texas Instruments, Philips, Hitachi, Motorola, Fujitsu, and International Business Machines (IBM) joined the race with Bell Laboratories to mass-produce bubble memory chips. Texas Instruments became the first major chip manufacturer to mass-produce bubble memories in the mid- to late 1970’s. By 1990, however, almost all the research into magnetic bubble technology had shifted to Japan. Hitachi and Fujitsu continue to invest heavily in this area.

Mass production proved to be the most difficult task. Although the materials are different, the process to produce magnetic bubble memory chips is similar to that for semiconductor-based chips such as RAM (random access memory). It is for this reason that major semiconductor manufacturers and computer companies initially invested in this technology. Lower fabrication yields and reliability issues plagued early production runs, however, and, although these problems have mostly been solved, gains in the performance characteristics of competing conventional memory have limited the impact that magnetic bubble technology has had on the marketplace. The materials used for magnetic bubble memories are costlier and possess more complicated structures than those used for semiconductor or disk memory.

Speed and cost of materials are not the only bases for comparison. It is possible to perform some elementary logic with the bubbles. Conventional semiconductor-based memory offers storage only. The capability to perform logic with magnetic bubbles propels bubble technology far beyond any other magnetic technology with respect to functional versatility. Although disk and tape technologies offer the same nonvolatility and rewriteability features as magnetic bubble memory, and greater capacity at a lower cost per bit of storage, they come up short with respect to access time, reliability, and convenience of packaging. Semiconductor-based memory offers tremendous speed and a refined integrated circuit fabrication process that yields low-cost, highly reliable packages. Magnetic bubble memories are packaged as chips, similar to the packaging of semiconductor memories or logic gates. This is much more convenient than disk or tape solutions that require drive units. Also, removable magnetic bubble cartridges have been introduced. Furthermore, since there are no moving mechanical parts, bubble memory is inherently more reliable. Simple interfaces to the memory have been developed that compare favorably to the control logic necessary for disk or tape technology.

A small niche market developed in the 1980’s. Magnetic bubble memory can be found in intelligent terminals, desktop computers, embedded systems, test equipment, and other, similar microcomputer-based systems. It seems best suited for applications in which a small amount of secondary memory is needed—on the order of a few megabytes (a byte is a group of eight bits; a megabyte is a group of a million bytes). In summary, the short history of magnetic bubble technology can be characterized in terms of ten-year periods, with the 1950’s and 1960’s as decades of discovery, basic science, and theoretical refinement; the 1970’s as a period of industrial investment and experimenting, leading to mass production; and the 1980’s as a time of dwindling impact, a retreat to the laboratory for further innovation. Computers;memory
Bubble memory, computer
Magnetic storage devices

Further Reading

  • Bobeck, Andrew H., and H. E. D. Scovil. “Magnetic Bubbles.” Scientific American 224 (June, 1971): 78-90. An excellent introduction to the topic of magnetic bubbles by two of the original investigators. Thorough in its explanations of the physics of magnetic domains underlying this technology and accompanied by many good diagrams and photographs. Any study of the subject should start here.
  • Chang, Hsu, ed. Magnetic Bubble Technology: Integrated-Circuit Magnetics for Digital Storage and Processing. New York: IEEE Press, 1975. Prepared under the sponsorship of the IEEE Magnetics Society; presents an overview of bubble technology (history, devices, applications, physics, and materials), followed by a compilation of reprinted articles from scientific journals or conferences published from 1971 to 1974. Also contains a complete (1950 to 1974) bibliography of published articles on magnetic bubbles and a list of issued United States patents on bubble domain devices from 1960 to 1974.
  • Eschenfelder, A. H. Magnetic Bubble Technology. 2d ed. New York: Springer-Verlag, 1981. Designed as a supplementary text for a semester course in either solid-state physics or materials science. The reader is expected to have had some background in physics and mathematics at the college junior to senior level. Chapter 1 provides a good overview of the field and is accessible to the general reader.
  • Hodges, David A. “Microelectronic Memories.” Scientific American 237 (September, 1977): 130-145. Another excellent review of the state of the art (in 1977) in information storage technology. Part of a dedicated issue on the topic of microelectronics, this article offers a complete picture of the multiple alternative primary memory technologies, along with magnetic bubbles. Highly recommended.
  • Johnson, Mark, ed. Magnetoelectronics. San Diego, Calif.: Elsevier, 2004. Comprehensive text on the general field of magnetic computer memory, explaining the relationship between bubble memory and other magnetic forms of storage. Bibliographic references and index.
  • Triebel, Walter A., and Alfred E. Chu. Handbook of Semiconductor and Bubble Memories. Englewood Cliffs, N.J.: Prentice Hall, 1982. Intended as a reference work for practicing engineers and technicians. Contains valuable in-depth information not found in other books on microprocessors or digital electronics. Chapter 1 provides an excellent overview of the subject, and chapter 10 is on magnetic bubble memories. Highly recommended.

First Electronic Stored-Program Computer Is Completed

UNIVAC I Becomes the First Commercial Electronic Computer

Floppy Disks Are Developed for Computer Data Storage