Molecular data storage

As we generate increasing amounts of data, scientists have developed a new way to store it: in tiny molecules.

The amount of data we produce each day is truly staggering: each swipe or click is generating data and we create an estimated 2.5 million terabytes daily, and growing. And although hard drives are becoming increasingly smaller and more space-efficient, it’s hard to keep up with this pace of growth to actually store this data. 

Now, a team at Brown University are turning to molecules as a solution. Given how small molecules are, this means a large amount of data can be stored in a small physical space.

Traditionally, data are stored in “bits” of information - a bit represents a 1 or a 0, and we need a string of 8 bits - or 1 byte - to store one character, like a letter of the alphabet. So when you send a 240 character Tweet, that is 2240 bits of information you’re creating that Twitter then needs to store. Multiply that by the millions of Tweets being issued each day and you can quickly see the problem.

So how can molecules help? The answer is in their size. “If we were able to store bits in every molecule in a glass of water, that’s 10²⁸ bits of information. That same amount of information would require 200 Empire State buildings worth of hard drives to store,” estimates Brenda Rubenstein, the senior author of the study describing the new data storage technique.

Data storage is conceptually simple - as each bit is a 1 or a 0, any two-state system can be used to store data. Magnetic hard drives work on the principle of either magnetising something, or not, in order to represent a 1 or a 0, but you could equally well use whether a molecule is present or not.

“We store data in the presence and absence of molecules,” says Rubenstein. By creating a solution which contains a type of molecule, representing a 1, or lacks it, representing a 0, lots of “bits” of information can be stored.

The scientists use lots of different molecules, each with a unique mass. They have a key which defines a certain order for the molecules. This means that even when the molecules are mixed up, they can be identified, and then put in the correct order using the key to recreate the data.

A “liquid handling robot” can pick up tiny amounts of the different types of molecules. Depending on the 1s and 0s required, it then mixes the appropriate molecules into a droplet. Each droplet contains 10s-100s of different molecules, or bits, meaning a lot of data can be packed into just one drop.

How do we get this back? The scientists use a mass spectrometer, which can analyse the masses of the molecules in a solution. By knowing which molecules and absent and present, they can recreate the string of 1s and 0s - and thus read back the data.

Yet, this is just the beginning of using molecules for our data needs. Not only can they store data, but they can also perform simple calculations and the team hope to take this further.

“We’ve been able to make molecules machine-learn, now we can have molecules recognise patterns and classify signals all by themselves. But where we’d really like to take it is, even beyond that, is to have the molecules, the 10²⁸ molecules, potentially all be able to all individually calculate things for us” says Rubenstein.

So will our USB sticks soon be replaced with mini flasks of information? Possibly not yet, but this does mark a step towards a totally new type of data storage.