The Evolution of Computer Memory – From Semiconductors to Proteins

Semiconductor Memory

Conventional computer memory is known as “semiconductor memory” and was invented in 1968. It’s based on technology known as the “semiconductor” which was invented in 1947. Many semiconductors grouped together is called an “integrated circuit”, more commonly known as a “computer chip”. Examples of semiconductor memory include ROM, RAM and flash memory. A big advantage of computer RAM (main memory) is price; ram is inexpensive. The main disadvantage of RAM is volatility; when you turn your computer off, the contents of RAM are lost.

Molecular Memory

Molecular memory is the name of a technology that uses organic molecules to store binary data. The Holy Grail of this technology would be to use one molecule to store one bit. For the near future, it would be more realistic to expect to have systems that use large groups of molecules to represent a single bit. Different types of molecules have been researched, including protein molecules. A more precise name of a molecular memory system that uses protein molecules is Protein Memory. Other types of molecular memory would have more precise names derived from the types of molecules on which the technologies are based.

Protein Memory

In the mid-1990s, the development of a protein-based memory system was the project of Robert Birge – chemistry professor and director of the W.M. Keck Center for Molecular Electronics. He was assisted by Jeff Stuart, a biochemist and one of Birge’s graduate students. The protein molecule in question is called bacteriorhodospin. Purple in color, it exists in the microorganism halobacterium halobium which thrives in salt marshes where temperatures can reach 140F.

The protein undergoes a molecular change when subjected to light making it ideal for representing data. Each molecular change is part of a series of many different states known as the photocycle. There are three main states: the bR state, the O state and the Q state. The O state represents binary 0 and the Q state represents binary 1 while the bR or rest state is neutral. To survive the harsh conditions of a salt marsh, the protein must be incredibly stable, a critical factor if it is to be used for representing data.

While in the bR state, the protein is placed in a transparent vessel called a cuvette, measuring 1 x 1 x 2 inches. The cuvette is then filled with a gel. The protein is fixed in place by the solidification of the gel. 2 arrays of lasers – one red and one green – are used to read and write data while a blue laser is used for erasing.

Reading, Writing and Storage Capacity

We will start in the bR state of the photocycle. A group of molecules is targeted and hit by the green laser array, also known as the Paging lasers. These molecules are now in the O state which represents binary 0. The O state allows for 2 possible actions:

• Reading – done with the red laser array set at low intensity

• Writing a binary 1 – done with the red laser array set at high intensity which moves the molecules to the Q state

The Q state allows for 2 possible actions:

• Reading – done with the red laser array set at low intensity

• Erasing – done with the blue laser which moves the molecules back to the bR state

A bacteriorhodospin storage system is slow. Although molecules change states in microseconds (millionths of a second), it’s slow when compared to semiconductor memory which has an access time measured in nanoseconds. Unfortunately, the time required to actually perform a read or write is even greater, on the order of ten milliseconds (thousandths of a second). The data transfer rate on this type of storage device is also very slow – 10 MBps (MB per second). In theory, the 1 x 1 x 2 inch cuvette could hold 1 TB of data or roughly one trillion bytes. In reality, Birge managed to store 800 MB and was hoping to achieve a capacity of 1.3 GB (billion bytes). The technology proved itself to the point that NASA was exploring methods of improving the technology during space shuttle missions, which in fact resulted in higher storage densities.

Conclusion

Birge’s quest to build a protein-based memory system for a desktop computer was unsuccessful. Although Birge’s vision failed, the development of some form of molecular memory (possibly protein memory) for desktop computers, seems possible. Scientists have also continued to work on developing other ideas involving protein memory. One idea from 2006 was to apply a layer of bR proteins to the surface of DVDs to increase storage capacity, theoretically up to 50 TB (over 50 trillion bytes). A dual layer blu-ray disc has a capacity of 50 GB (over 50 billion bytes).