In Search of New Knowledge
As told by Sir John B. Pendry

2014 KAVLI PRIZE IN NANOSCIENCE

MORE ABOUT SIR JOHN B. PENDRY

I was born in 1943 in a small town in northern England. We lived with my grandmother who owned a shop selling sweets, ice cream and chocolate – heaven for a small boy. The cellar beneath the shop was crammed with various pieces of abandoned mechanical and electrical equipment which must have stimulated my interest in science and I cannot remember a time when science was not my preoccupation. In those days experiments were of more interest that theory: loud explosions were heard in the vicinity, and neighbours’ television sets did not always function as expected. Later, to the relief of all, my interests turned to theoretical matters, largely because theorists find it easier to have a broader outlook. Music was another passion and my desire to improve my piano technique continues to this day, but without the success that science has afforded me.

I was fortunate that an excellent education in science could be had at our local grammar school where the more formal parts of my early science were put in place by my teachers. In particular I was very much helped by my physics teacher and my headmaster both of whom encouraged me to apply to their old College in Cambridge: Downing College where I matriculated in 1962 to study natural sciences which included physics, chemistry, mathematics and crystallography. The latter I hated: the three dimensional thinking required was like an hour in the gym on the parallel bars, but I have to reflect that like physical exercise, the mental training has stood me in good stead. A lesson to bear in mind: not all excellent lecture courses are popular.

The first taste of real research
After my first degree I stayed on to do my PhD in the Cavendish laboratory where I worked out the theory of low energy electron diffraction (LEED) with Volker Heine. Electrons with energies of around 100eV penetrate into surfaces only a few atomic layers and the diffraction patterns (first observed by Davisson and Germer in 1927 and for which they received the Nobel prize) are sensitive to the precise arrangement of atoms at the surface. Surface science was abuzz at that time and at the centre of several rapidly advancing fields: electronics in the physics community and catalysis amongst the chemists to name but two areas. However no one knew how the atoms were arranged at the surface, and this was regarded as the key to understanding the new fields. Frustratingly the surface information contained in the electron diffraction patterns was highly convoluted by the very strong multiple scattering of electrons. They rattled around the surface colliding with several atoms before leaving. So unlike Xray scattering which is a simple process easy to interpret, complex theory was central to extracting surface information. I succeeded in developing a workable theory. Stig Andersson working in Gothenburg had collected diffraction data from some simple adsorbate structures on a nickel surface and together we interpreted these data to produce the first surface structure determination. This was my first taste of real research which at its best combines challenging theoretical developments with interpretation of novel experiments.

Cambridge at that time was an immensely exciting place to be a scientist: Crick and Watson had a few years previously, working in what looked like a bicycle shed just below my office, solved the structure of DNA. Josephson in the neighbouring Mond laboratory had astonished the world with his predictions about superconducting junctions and I remember the susurrations of excitement that preceded the announcement of the first quasi stellar object or quasar by Tony Hewish and Jocelyn Bell. Steven Hawking was one of our classmates in Dirac’s lectures on quantum mechanics but had yet to make his famous prediction of Hawking radiation from black holes.

Moving around
An important interlude in my career came in 1972-73 with my year spent at Bell Laboratories in Murray Hill New Jersey. There I wrote my book on electron diffraction and several papers on electron states at surfaces, but the most important part of my stay was learning how this extraordinary laboratory functioned. Its list of successes in many fields is long and its demise regretted by all who worked there. This year was also my introduction to the American research scene in general which has since been an important part of my research collaborations.

Returning to Cambridge in 1973 I met my future wife, Pat, who was a post doc in the Cavendish. She had graduated from Oxford in group representation theory and now had a research fellowship at New Hall. We eventually got married in 1977 after my move north to Daresbury.

At Bell, Peter Eisenberger was working enthusiastically on the new structural technique of extended Xray absorption spectroscopy or EXAFS. In these experiments Xrays kicked out an electron from a deep core level of an atom and the Xray absorption cross section was determined by how easily that electron could escape, which in turn depended on how the surrounding atoms were arranged. This was a beautiful tool for investigating the immediate environment of the atom in question. No long range order is required, as in Xray crystallography, and the theoretical tools developed in LEED could easily be adapted to interpret the data. There was a host of potential applications in disordered materials and particularly in protein crystallography where the centre of interest contains a heavy atom, for example the iron atom at the centre of the haemoglobin complex. Patrick Lee and I wrote an influential paper on interpretation of EXAFS spectra. This was my first brush with light/matter interaction and served me in good stead when I decided to move on from Cambridge and in 1975 was offered a position as head of the new theory group at the Daresbury synchrotron radiation laboratory. There followed six very happy years at Daresbury where as head of a large group of theorists I had management responsibilities that broadened my horizons and developed skills which have been valuable in my subsequent career. The research environment was ideal with excellent resources, particularly of computing power. My theory of EXAFS spectra blossomed into a more general theory of photoemission from the valence and conduction bands which is much more complex and interesting. However after six years it was clear that further career opportunities at Daresbury were limited and it was time to move back to university. In 1981 I took my present position at Imperial College London where I have remained until the present day.

The localization problem
Establishing my research in a new environment was greatly assisted by a generous grant from Don Braben’s BP Venture Research. BP had set up the company to fund blue skies research and it was an ideal funding source for the new ideas I wanted to try out in disordered systems. Having solved the problem of electron wavefunctions interacting with complex but ordered solids I was keen to tackle the problem of waves in a disordered medium. Philip Anderson had introduced me to the localisation problem, a peculiar and little understood transition in behaviour between conducting and insulating states as the disorder is increased. I embarked on the high risk project of applying my techniques to this challenging problem. After a decade of work a complete theory of localisation in one dimension resulted. The full three dimensional theory was, and still is, missing but some interesting results came out of the work. One of them is as follows: a disordered material such as milk or snow scatters incident light waves so that very little light is transmitted through a thick layer of snow, but I was able to prove that there is always a way of shining light onto the surface so that the light is almost perfectly transmitted to the other side. This is a completely general result for any sort of waves in disordered systems and in its way astonishing. There the result sat until some twenty years after my paper, Alard Mosk and his team in Twente were able to show that the result held true. Sometimes one has to wait a long time for experimental confirmation.

The idea of metamaterials
Again, it was time to move on, not physically but research wise. Yablonovitch had startled the optics community by proving that by micro-structuring a dielectric on the scale of the wavelength of light (about ½ a micron) all light in a range frequencies could be excluded from the interior. The analogy was with electrons in a semiconductor which have similar excluded ranges of energy. My experience with electrons led me, in collaboration with Angus MacKinnon, to develop computer codes for calculating how light interacts with an ordered dielectric. In fact the way these codes were written made them immediately applicable not just to dielectrics where the permittivity is independent of frequency, but also to metals where it disperses strongly with frequency. This led me to study plasmonic systems. Electrons in metals move collectively, and wobble like a jelly if hit by light so that the electron density at the surface oscillates up and down. Exactly how they behave depends strongly on the shape of the surface and this evolved into an entirely new avenue of research on the role of structure in the optical response of a solid.

In the mid 1990’s Will Stewart invited me to the Marconi research laboratories to help them with some of their problems with electromagnetic materials. Their radar absorbing material (RAM) was very effective at protecting ships from enemy radar. Made from 3cm long carbon fibres scattered over the surface of a supporting material the RAM soaked up radar waves, absorbing them almost completely, whereas individually the fibres would strongly scatter radar. Clearly the fibres were somehow talking to one another in a way that strongly influenced their properties. This gave me the idea that is now known as metamaterials: materials whose properties depend on their internal microstructure as much as on their chemical composition. There are many experiments that you might want to do with light but are prevented from doing by the absence of materials with the right properties. Nature has been parsimonious in providing a range of naturally occurring properties and many properties allowed by the laws of physics are just not found in nature.

One of the missing properties is a negative refractive index. In the 1960’s Victor Veselago had shown theoretically that a negative refractive index would give rise to some amazing effects, but failed to realise his dream because nothing he tried had a negative index. Metamaterials came to the rescue. In 1999 there was a conference at Laguna Beach in California where I spoke for the first time about the range of properties that metamaterials could create, which included negative values for the magnetic response, or permeability. David Smith and Shelly Schultz were in the audience and realised the significance of this result. Shortly afterwards in their laboratories they utilised my ‘split ring’ structures to create a material which at radar frequencies had both a negative magnetic and electrical response. Veselago had told that this was the recipe for negative refraction. David and his colleagues had made the first material to show a negative refractive index. Their conclusions were highly controversial as they appeared to violate some fundamental rules, but the results have stood the test of time and the apparent violations reconciled.

The perfect lens
Negative refraction can be used to make a sort of lens from a flat slab of material: bending a diverging beam of light back to the axis so that it makes a focus inside the material, and also a second focus as the rays emerge from the other side. Veselago knew this, but I noticed that the focus obtained seemed unusually perfect. Further investigation through rigorous solution of Maxwell’s equations showed that this was indeed a most unusual kind of lens. All conventional lenses suffer from the limitation that they cannot see anything smaller than the wavelength the light. To my astonishment the negative refraction lens showed no such limitation. I was afraid that this result might be an error on my part and put it aside for some weeks. Colleagues were consulted who all doubted its correctness but could not find the flaw in my calculations. So with much trepidation the result was published in Physics Review Letters in 2000. As I expected there was a firestorm of interest. Most were intrigued by the result but a few were extremely angry and hostile. It was a difficult period, but eventually all the objections were answered and the result stands. However the challenge of reproducing the exacting conditions demanded by the theory has proved very great. Nevertheless proof of concept experiments have been made first at radar frequencies by George Eleftheriades in Toronto and later by Nick Fang and Xiang Zhang at UCB, and Richard Blaikie in New Zealand. It remains my most referenced paper with more than 5,000 citations, not all of them claiming that it is wrong!

The ‘perfect lens’ is so perfect that all it does is make an identical copy of the image, but mostly lenses are used to manipulate an image: for example to magnify it. However to redesign the new lens requires new design tools. Snell’s law of refraction is no help in controlling the new components of the fields in the lens, and Maxwell’s equations although exact, do not give the physical insight comparable to Snell’s law. So I went back to an old result I had used some years previously which exploited the invariance of Maxwell’s equations under coordinate transformations. Imagine that we move the light to where we want it to be by deforming space itself, thus taking the light with it. Einstein showed that as far as light is concerned a deformation of space is equivalent to a change in the refractive index and he gave us a simple formula. Thus we can imagine squeezing space to deformed the electric and magnetic field components of the light, and at the end of the process calculate a refractive index that would do the same job in a real system. I used this technique to design a perfect magnifying lens and other devices. But the most well known application was to come a year or two later.

In 2005 I was invited to a DARPA meeting in San Antonio, Texas, with the brief to ginger up proceedings. So after a quite conventional talk on metamaterials and transformation optics I mentioned at the end that I could design a cloak of invisibility and gave the formula for the material parameters required. I expected laughter but on the contrary my little joke was taken seriously and has led to a minor industry in cloaks of invisibility. In particular David Smith’s team were present at the meeting, David himself being sick with a cold, and on reporting back to David, once more came the offer to build a cloak which, incredibly, they were able to do in the space of about six months. I regard the cloak as something of a grand challenge: if you can design a cloak with transformation optics and metamaterials, you can design anything. Thanks largely to JK Rowling of Harry Potter fame the cloak is an excellent way to get science across to a wide audience, especially to school children. And we in the cloaking world have had a great time enjoying ourselves in the limelight, a spot normally reserved for astronomers.

The work on light interaction continues
Metamaterials are emerging as a new discipline and to consolidate our work in London my colleague Stefan Maier and myself applied for a Leverhulme Foundation grant to enable us to recruit more faculty and young researchers to the metamaterial cause. And so five years ago we started a new phase in our work. Today my research is largely into light interaction with the surface plasmons supported by metallic systems. The transformation optics tool brings new understanding to the many strange phenomena found in these systems, such as surface enhanced Raman scattering (SERS). Structuring metal surfaces on the nanoscale enables us to capture light and squash the optical energy into less that a square nanometre. Normally photons with a wavelength of 500nm can hardly detect an atom, whose diameter is typically less than 1nm, but our harvesting devices can concentrate the light sufficiently to allow detection of a single atom.

Throughout my career one of the great pleasures has been to meet and collaborate with other scientists, particularly students and post docs, and also colleagues from other institutions. Some of these I have mentioned above but all have played their part in helping along my efforts.