Sometimes everything can seem to happen at once. The new game in town is active invisibility cloaks (AIC), which use electronics and antennas to generate a cloaking field to hide an object. Two types of active cloaks have just been revealed (excuse the pun). While being impressive feats of technology, such cloaks could easily be defeated in practice.

Earlier this week, a research paper from Andrea Alu's group at the University of Texas demonstrated that any passive invisibility cloak (e.g., one just depending on the properties of metamaterials) would not only fail to achieve invisibility under broadband illumination (for example, white light), but would actually cause the cloaked object to stand out more strongly than when it is uncloaked.

A natural question to ask is if there are any other possibilities. In fact, there are. An active cloak uses sensors, electronics, and antennas to actively generate an electric field near the surface of the cloak that interferes destructively with the radiation scattered from the cloak and contents. This electric field will be called a cloaking field. In the past few days, Professor George Eleftheriades' EE research group at the University of Toronto has revealed a working model of an active cloak, while Prof. Alu's group has worked out a new approach for making active cloaks.

In essence, if the cloak transmits the exact opposite of the light being scattered from the cloak, it will appear as if there is no object there. Even the shadow behind the cloak, which results from the illumination that is scattered from the front of the cloak, is removed. Such a cloak is as broadband as the sensors, electronics, and antennas can handle.

Schematic of the University of Toronto active invisibility cloak (Photo: University of Toronto)

Sounds good, but what's the catch? Turns out there are two catches. First, to successfully use an active cloak you need to characterize the light that is scattered from the target. Until this is known, the cloaking field required to cancel out the scattered illumination can't be determined. Second, you can't design your cloaking field from local measurements at the site of the antenna. Measurements from around the cloaking device are needed before the signals to be sent to the antennas can be properly calculated and generated.

Both of these requirements cause a response delay, the time interval between first sensing a source of illumination and the cloak's generation of an effective cloaking field. Even given infinitely fast electronics, the response delay is going to be similar in size to the time it takes light to cross the longest dimension of the cloak.

An example of where an active cloak would function quite well is to hide from a Doppler radar system. The Doppler effect tells us that light scattered off a moving target changes in frequency by an amount proportional to the relative speed between the source of the waves and the target. A change in speed of 15 m/s (34 mph) will change the frequency by one part in ten million.

Fourier analysis tells us that it takes time to notice so small a change in frequency. In particular, to be able to detect that 15 m/s change in speed, a 10 GHz Doppler radar would have to have a pulse containing ten million oscillations to be sure we would notice that one of them was missing. As a result, the pulse would have to be about a millisecond in length.

Given that the Doppler radar has a sensitivity set to measure a pulse a millisecond long, it won't notice that an active cloak takes a handful of nanoseconds to sense the pulse, and generate an effective cloaking field for the target. The portion of the pulse that is not effectively cloaked is simply too short to be detected.

What sort of cloaking job is difficult for an active cloak? It is when the radar (also including light) avoids sending out any pattern or similarity which can be predicted by the cloak's electronics. As the cloaking field cannot be generated until the pulse is sufficiently understood to erase its scattering from the cloak, any foreknowledge of the radar pulses makes the job of generating a cloaking field easier.

Assume that the radar illumination takes the form of pulses having constant interval, frequency, strength, and shape. In that case, the cloak's electronics could learn that each pulse is the same, so it only needs to pump out a predetermined cloaking field at the fixed pulse interval to cloak the target. This does assume that the target doesn't change shape or position – unlike the passive metamaterial cloaks, the external fields do penetrate into the region containing the target. Accordingly, we want to avoid giving the cloak such cues.

Arguably the worst case is if the illumination appears at random intervals, the short pulses have randomly varied frequency, width, and shape, and the wavelength is short compared to the volume that is being cloaked. These characteristics accomplish the befuddlement of an active cloak in separate ways. Clearly, by avoiding random intervals the cloak cannot initiate a cloaking field in advance of the arrival of an illumination pulse – even it it knows the shape of the pulse in advance, it doesn't know when the pulse will arrive, causing an unavoidable and perhaps lengthy response delay.

Similarly, if the pulses have different nominal frequencies, the cloak will take some time to detect what the frequency is, and figure out an appropriate cloaking field. If the pulses have different widths, they will include different bandwidths of frequencies around the nominal frequency. This is very confusing, as you may have to monitor the entire pulse to see if there are hidden signals within it.

This brings us to the effect of pulses of variable shape. If the pulse has a constant intensity save for a narrow region at the middle of the pulse, there will appear a short burst within the pulse where the frequency spread of the pulse suddenly increases. This cannot be prepared for, so an active cloak cannot effectively hide a target from such a doctored pulse – some of the scattered signal (or an improperly matched cloaking field) will leak out of the cloak, making it visible. Remember that the cloaking field, if not properly matched to a scattered field, is roughly as easy to detect as would be the scattered light from the object; if they were not the same size, the one couldn't erase the other.

To sum up, the Achilles heel of an active cloak is its response delay – it cannot respond to randomly changing radar pulses quickly enough to effectively cloak a target from detection. To reduce the ability of a radar to find a cloaked object is likely possible, to achieve perfect cloaking is not. What we still have to learn is how closely an active cloak can approach perfection against an attack which is random in several different directions. At this point, however, the new active cloaks are not, as many in the media are announcing, "an invisibility cloak that works." At least not in my book.

Source: Physics Review X and Physics Review Letters preprint