Storing data in molecules

A new way to meet the challenge of burgeoning data storage demands...
09 July 2019

Interview with 

Brenda Rubenstein, Brown University


Coloured chemical burettes


My first computer hard drive held 10 megabytes of data and it was the size of a shoebox; these days we routinely fit systems with discs that hold a million times that amount but are a fraction of the size. Even so, at the rate we're going, we'll soon be churning out so much data each day that the physical space we need to store it - and the electricity needed to power it - will become a huge problem. At the moment, we convert the data we want to store into a series of 0s and 1s, a bit like computer Morse code, and this sequence is written onto a hard disc. To store more noughts and ones, you need bigger - or more - discs. But Brenda Rubenstein, at Brown University in the US, has come up with a much better way of doing it: using molecules. In this system, the presence of a specific type of molecule in a droplet of liquid equals a 1; the absence of that molecule signals a zero. And because she can mix many different types of molecules together in the same droplet, she can store much more information and in a fraction of the space. To read the information back, she feeds the droplets into an analyser called a mass spectrometer that detects whether a molecule is there or not. She’s using it at the moment to store small images. Ankita Anirban heard how it works...

Brenda - Typically, when we are thinking about data which can be images, it could be text, we usually store that data in a series of zeros and ones. And when we want to store those in different media, basically what we need are two states to represent the zero and the one. We are trying to think about how we can actually store those states, those zeros and ones in the smallest possible volume that we can think of. What we do is we store information in the presence and absence of molecules. If we have a solution, we can either put a very specific identifiable molecule in the solution or we can leave it out. And so by choosing which molecules we put in or out we can encode a string of bits.

Ankita - So how big are your molecules?

Brenda - Our molecules are roughly several angstroms by several nanometres across, so we're talking about areas of nanometres squared or so. However, we can use all different types of molecules so those specifications are for some of the smallest molecules that we're thinking about.

Ankita - So in terms of size, how does that compare to your typical hard drive?

Brenda - The comparison that I love to make is a little 'back of the envelope' calculation that I did a while ago. If we're able to store bytes in absolutely every molecule in a glass of water, for example, that glass of water would be able to store about 10 to the 28 different bits. Now if we scaled that up to what that would mean in terms of hard drives, that same amount of information would require about 200 Empire State Building's worth of hard drives in order to store it.

Ankita - And so how are you actually picking up these molecules which are absolutely tiny and mixing them up?

Brenda - So this is the part that makes me super excited. We've gotten to the point where engineers can essentially create very automated, very fast liquid handling robotics, so our robot can very quickly mix all the molecules that we need to make our information.

Ankita - Once you've mixed up the solution, you've got your image in a liquid form, how do you then get that image back from that?

Brenda - Yeah. So here is where we use some excellent modern technology that's really come along over the past decades, which is mass spectrometry. And the whole idea of mass spectrometry is that this very nice machine can read out the individual masses of all of the molecules in a solution. We know which molecules that we're using to store the information. We run our solution through the mass spectrometer, and then we see if the masses that match up with those molecules actually appear in the spectra that we take, and that allows us to read out exactly what was in our solution to begin with.

Ankita - How long does it take?

Brenda - We've stored images of the scale of tens of bits, all the way to hundreds of thousands of bits, and in order to make these images, it takes on the order for the liquid handling tens of minutes to hours. Some of the limitations there are the speeds it takes to actually mix that many molecules, and the other limitation is really how much can we resolve using mass spectrometry? As we advance those techniques we’ll be able to not only mix more molecules, we also will be able to read out more molecules accurately.

Ankita - This sounds like quite a sophisticated setup you've got in your lab, do you think realistically it's something that everyday people could use in the way that we use say, USB sticks?

Brenda - Yes. Eventually, I foresee that. Now obviously, the technology is young and we are doing this for the purpose of doing basic science to figure out what it is that we need, what are the engineering barriers and so forth. And so once we work these things out like we have, it's actually not that difficult to do and I certainly see these things being miniaturised in the future. Exactly what direction that will take, we'll see as the science proceeds...


Add a comment