New microscopy technique accidentally connects to the dark side (of plasmons)

Some papers read a bit like a roller coaster. The title sucks you in with a promise of a revealed truth and excitement. As you proceed though the introductory fluff and get to the anticipated revelation, it begins to dawn on you that what excited them is not going to excite you. It’s kind of like going to a showing of The Princess Bride expecting to see a romance film. But you’ve invested this much time already, so it can’t hurt to see it through to the end, can it? And then, like Inigo Montoya, the whole point of the paper jumps out, waves its sword in your face and shouts at you, leaving you wondering how dumb you could have been to not have seen it coming. 

So goes a recent paper in Science. Its title advertises the work as an interesting new imaging technique. The imaging, while kinda-a-sorta interesting, is nothing compared to the physics that enabled it, which preserves short-lived oscillations through a peculiar kind of coupling. I realize that “peculiar kind of coupling” sounds mysterious, but I could think of no other succinct way of describing it.

A new imaging technique, yippee

All my regular readers know that I get pretty excited about new imaging techniques, so let’s get that part of the paper out of the way first. Basically, the researchers have combined a particular type of electron microscopy with a particular type of optical spectroscopy to create a microscope that allows you to visualize state populations and state coherences with the resolution of the electron microscope.

What do I mean by that? Well, atoms and molecules are made of energetic states. That is, their electrons can be excited from low-lying energies to higher energies by light pulses with the right frequency or color. Since there are lots of atoms within the spot of a laser, you get populations of atoms in a particular state, depending on how bright the laser pulse was. The flow of population between these states can be coherent, incoherent, or a mixture of the two. These coherences take the form of oscillating electric fields, since all transitions between states involve electrons oscillating.

To look at both population and coherence, the researchers use a series of light pulses with carefully timed delays and phases between the pulses. The laser pulses (or a separate measuring pulse) are energetic enough to knock a few electrons free, allowing them to be detected by an electron microscope with 50nm resolution. The rate at which these photoelectrons are produced depends on the energetic state the atoms are in and whether the timing and phasing of the pulses matches their coherence. 

Think of it like playing with a child on a swing with the goal of ejecting him or her from the swing and sending them flying through the air. First you need to push at the right frequency. Then, once the swing is in motion, you also need to push at the right time or phase, otherwise you will be slowing the swing down again. And, finally, if the swing is in motion, you don’t need to work as hard to achieve ejection.

To bring it back to electrons. Those in excited states are easy to eject, but, if you time the pulses wrong, you don’t manage to eject them as easily.

Having developed a nice technique, the researchers needed something to take a look at. They chose something that they thought would be quite challenging: a roughened silver surface. These rough surfaces support localized surface plasmon resonances. But, the very thing that allows them to be excited—the irregular surface of the metal scatters the light into the plasmon—also damps the oscillations out very fast. In other words, these oscillations can have large populations, but very very short lived coherences: typically less than 10fs (a femtosecond is 10^-15s).

Oh, be still my beating heart

At this point, I was about ready to give up, because the technique is interesting, but surely roughened silver isn’t going to offer new insights, right? In how many languages can you say “wrong”? It turns out that the researchers discovered a spot on the metal where the plasmon coherence lasted for around 100fs, more than ten times longer than expected. That was a bit of a puzzle. After some thought and a few calculations, the researchers concluded that this particular plasmon could couple to something called a dark plasmon mode.