How Fast Is The Photoelectric Effect?

Every student who studied modern physics in an undergraduate General Physics course would have encountered the photoelectric effect. It is a phenomenon that has a special place in the history of physics, and the theoretical description of this phenomenon gave Einstein his Nobel Prize.

So one would think that this is a done deal already, and we should know all there is to know about it. In some sense, we do. We know enough about it that we have expanded this phenomenon to be included in a more general phenomenon called photoemission. We use this phenomenon to study many things, including band structure of materials. So it is very well-known.

Yet, as with so many things in physics, the more we study it, the more we want to know the minute details of it. In this case, the current study is on how fast an electron is emitted from a material once light impinges upon it. In other words, from the moment a photon is absorbed, how quickly does the electron is liberated from the material?

This is not that easy to answer because, well, one can already guess at how would one determine (i) the exact time when one photon is absorbed into a material, and (ii) the exact time when an electron  is liberated due to that absorbed photon. On top of that, this may be a very fast process, so how does one measure a time scale that is almost instantaneous?

The authors of this latest paper[1] came up with a very ingenious method to determine this, and in the process, they have elucidated even more the various stages of what is involved in the photoelectric effect. But before we continue, let's get one thing very clear here.

The "photoelectric effect" that we know and love, and the one that Millikan studied, is the phenomenon whereby UV light is shown onto a metallic surface (cathode). We know now that this is an emission process of electrons coming from the metal's conduction band. This is important because, as this new study shows, this process is different than the emission from core levels (i.e. not from the continuous conduction band). Those of us who have done photoemission work using both UV and x-rays can attest to such differences.

The experiment in this report was done on a tungsten surface, or more specifically, W(110) surface. The hard UV light that was used allowed them to get photoemission from the conduction band and a core-level state.

What they found was that from the time that a photon is absorbed to the moment that an electron is emitted, the time for the process for a conduction electron is ~ 45 as, while for a core-level electron is ~100 as.

{as = attosecond = 1 x 10^(-18) second}

So the emission from core-level takes more than twice as long to occur. In their analysis, the authors stressed this conclusion:

These findings highlight that proper accounting for the initial creation, origin, transport and scattering of electrons is imperative for the proper description of the photoelectric effect.

Bill Spicer's 3-step model of photoemission process certainly highlighted the fact that it isn't a simple process. This paper not only reinforce that, but also included the effect of surface states in the influence to emission time and thus, possibly influencing other properties of the emitted photoelectron. 

There are many things in physics which we know a lot of. But these are also areas in which we continue to dig deeper to find out even more. There will never be a point where we know everything there is to know, even with established ideas and phenomena.

Zz.

[1] M. Ossiander et al., Nature 561, 374 (2018). https://www.nature.com/articles/s41586-018-0503-6
Summary of this work can be found here.