Bursting one tiny bubble of emerging military tech hype

“Two University of Michigan scientists are on the way to developing night vision contact lenses,” read the tweet, posted by a senior fellow for defense policy at a top think-tank, and retweeted by dozens, including a level-headed defense technology journalist famous for her own investigations into federally-funded fringe science and folly. I figured there must be something to this, so I clicked on the link to a story at defensetech.org, whose headline made the even more surprising claim “Scientists Develop Night Vision Contact Lens.”

Below the headline, a photo showed some guy pinching a contact with what looked like printed wiring on it. The photo was my first clue that something was screwy, because I’d seen that photo before; it was the Google lens (google it), a contact lens outfitted with a glucose sensor for diabetics.

The story claimed that the scientists had “developed a prototype contact lens that enhances night vision by placing a thin strip of graphene between layers of glass. The graphene — a form of carbon — reacts to photons, which makes dark images look brighter.” What were they claiming—that the graphene, like a laser, simply amplified the light? That didn’t make any sense, and neither did any other interpretation, because a contact lens is roughly located at the aperture of the eye, not the focal plane. The wearer would see only a blur [Note 1].

I thought this had to be an April Fool’s joke. But a quick googling of “Michigan graphene contact lens” found essentially the same story reported at an indeterminable number of sites around the web. Most of these stories seemed to rely on two sources: The University of Michigan’s press release from March 16, and IEEE Spectrum’s report from the next day, which drew on the press release and also referenced the paper in Nature Nanotechnology by professors T. B. Norris and Zhaohui Zhong, published online on the 16th. Sadly, Spectrum seemed to be the source of the claim that what Norris and Zhong had done was to actually “create their infrared contact lens.” But unsurprisingly, Zhong himself, if accurately quoted by the press release, had been the originator of the “contact lens” hype.

What the researchers actually created was far more mundane: a single-pixel test cell using graphene. The authors state that graphene “is a promising candidate material for ultra-broadband photodetectors, as its absorption spectrum covers the entire ultraviolet to far-infrared range.” It has been studied for this role since at least 2008, according to the references. However, its sensitivity in past experiments has been poor. Norris and Zhong have improved on this with some clever quantum electronics engineering [Note 2]. The result is “room-temperature mid-infrared responsivity comparable with state-of-the-art infrared photodetectors operating at low temperature.” But not quite “superhero vision,” as Discovery News put it.

Norris and Zhong’s research is important because it shows that graphene can potentially be used to create very compact thermal imaging sensors. “We can make the entire design super-thin,” says Zhong in the press release. “It can be stacked on a contact lens or integrated with a cell phone.” But first, they have to make an actual imaging sensor. They think they might be able to do something in a few years. It will require integrating patterned, doped graphene with conventional silicon lithography, a major challenge in the lab let alone for industrial production.

And while a small camera, even one small enough to fit into a phone, is not an unreasonable goal for perhaps a decade down the road, the idea of putting this into a contact lens is hard science fiction at best. To do this, you would basically have to make the entire camera, or perhaps an array of them, small enough to fit onto the contact lens. Plus, you would need a micro-projector, or again perhaps an array, small enough to fit under the camera(s), which would project the image as visible light onto the retina. Plus electronics and a power source, all small enough to fit on a contact lens. If Norris and Zhong knew how to do all this – let alone how to manufacture such a thing – I would not hesitate to declare them the greatest engineers in history.

But you gotta have vision, and you gotta have funding. These days, that means you gotta have hype. And hype echoes through the media machine, from press release to sloppy science reporting to outlets like Popular Science, CBS News, Independent, Huffington Post, and scads of fleas with smaller fleas on them.

To be fair, not all of these reports are technically inaccurate, but they all serve up the hype, because the hype is the story. The actual news is just an incremental advance in one of many ways of converting infrared radiation to an electrical signal. If some day we have the technology to make a night vision contact lens, it might make use of Norris and Zhong’s graphene trick. But that is not particularly likely, since there are so many other possible schemes. To say that their work opens the way, or is the first step which may some day lead to such a capability, is simply not true. I’m not sure where the line lies where hype is so inflated that it becomes lying. But I feel that in this case it has been crossed.

Why is this worth my time to write, and yours, dear reader, to read? Because it’s a paradigm example of how the emerging tech, especially emerging military tech, hype machine works… and while this little bubble is rather insignificant in itself, it sits atop a churning foam of bubbles big and small, some of them big enough to have real consequences. Things like laser weapons, “Iron Man” body suits, missile defense…

The inability of professional journalists, analysts and bureaucrats to separate hype from realism and sense from nonsense in military technology, and the relative absence of critical technical review, leads not only to massive waste but also to destabilizing suspicions and needless arms races which lead to … what? Nothing good, that’s for sure.

Note 1: In a camera (eye), light arrives at the aperture (iris) from every direction, and the camera geometry sorts out light coming from different directions. A detector located at the aperture, e.g. in a contact lens, would be exposed to light from all directions, and could not form an image. Also, a tiny display screen mounted in a contact would flood the eye with its light and produce a uniform blur. So neither end of this works. What you’d need is an even tinier camera, and a tiny projector (onto the retina). Since fitting them into a contact means they have to be very small, probably you’d want arrays of each to in order to collect and project enough light. Physics does not obviously rule this out, provided you’re OK with low resolution. But making such an object is clearly well beyond present capabilities. Just having a suitable candidate for a detector is such a tiny part of the problem that it’s practically irrelevant to whether or when such a technology might be realized.

The other thing you might imagine is that the graphene acts as a laser. Light from any direction might then be amplified and keep going in the same direction. So, you put this magic graphene anywhere – in the eye, on top of it, in a pair of glasses or a car windshield – and it just multiplies the light. However, the physics of this fantasy is wrong in fundamental ways, and in fact it has nothing to do with Norris and Zhong’s device.

Note 2: Graphene works as a photodetector much like other semiconductors: photons excite otherwise immobile electrons and create mobile electron-hole pairs. This results in a increase in the electrical conductivity of the material, which can be measured by passing a current through it. Alternatively, a P-N junction can be formed from two layers with different doping, hence a different affinity for the positive holes and negative electrons. The + and – charges are thereby separated, creating a voltage and current (hence, power) source, as in a solar cell. Norris and Zhong’s invention is a two-layer device, but instead of tapping the photogenerated power directly, the upper layer is left isolated, and the conductivity of the lower layer is measured by passing a current through it. Holes then accumulate in the upper layer, and their electrostatic effect on the lower layer, like the gate in a field effect transistor, creates a large change in its conductivity.

UPDATE: Since I first posted this, WIRED has become the latest (and if not the saddest, then it’s a tie) drinker of the UMich Kool-Aid. Mostly rehashing the same lines from the original press release as the other stories, WIRED also reports that Norris and Zhong say something about “car windshields to enhance nighttime driving,” which makes about zero sense. Additionally, WIRED links to this 3-year old blurb from Military.com, which suggests that “cat vision” contacts already exist, and were used in the bin Laden raid. The story has so many things wrong with it, I won’t bother to get started. The point, again, is that the tech, and especially military tech reporting world is ripe with this stuff, and sadly, so is the world of policy analysis and of actually funded R&D.

UN Presentation

I am posting a .pdf copy of a presentation I gave to the UN Secretary General’s Advisory Board on Disarmament Matters (6 March) in New York. The designated topic was Emerging Technologies, but the presentation was mainly about the most important emerging technology topic of the day, autonomous weapons. It was very well received by the ABDM, which had originally given me a one-hour slot, but continued the session for more than another hour beyond that, with extensive discussion.

US killer robot policy: Full speed ahead

My analysis of US policy for autonomous weapons, and how we got here, has been posted at the Bulletin of the Atomic Scientists’ website. In the first days after DoD Directive 3000.09 was released in Nov. 2012, I had posted a somewhat less discerning view at icrac.net. Like many people, I did not at first fully understand how the Directive was constructed, and what its implications were and are. So, in case anyone accuses me of inconsistency, I was wrong.