Atomic Physics at Sussex
Mike Pendlebury, 10 January 2011
NOTE ADDED: The following is a draft account by Mike of how the Atomic Physics group at Sussex was set up. It is dated 10 January 2011 and was written specifically for this wiki. Characteristically, before posting the final version Mike wanted to check various details. Knowing of his illness, I tried (unsuccessfully) several times to obtain a version that could be mounted. It was also labelled Strictly Confidential at this stage, although neither I nor others who have seen it since Mike's death can see anything in it that is even slightly controversial, nor any reason why it should not now be mounted. Somewhat guiltily, nevertheless, this is what I am now doing.
David Bailin, 25 September 2015
FURTHER NOTE: I have made some formatting changes in the article, and corrected some obvious typos, without changing any of Mike's essential content.
Robert Smith, 29 September 2015
The preparatory experiences of Ken Smith
The subject of Experimental Physics at the University of Sussex was first established in 1961 and grew with truly remarkable rapidity. Within five years it was able to boast a total research income that at that time exceeded that of any other physics department in the UK and 21 permanent faculty members. For this progress it had mainly to thank the tremendous energies invested by Professor Ken Smith with his ability to draw on very relevant previous experience. The favourable national climate at that time for expansion of university education in the physical sciences was also helpful.
In 1944, Ken Smith and Brian Flowers (later to become Lord Flowers) had just completed their undergraduate degrees in Cambridge. The course had been shortened to two years by the exigencies of the war. Immediately, they were recruited by Sir John Cockroft, then Jacksonian Professor of Physics in the Cavendish Laboratory, to join a team of scientists that he was assembling to man the new Laboratory of Nuclear Studies at Chalk River in Canada. This new laboratory was to be directed by Cockroft and to be brought into use for the Anglo-Canadian contribution to the atomic bomb project. The first nuclear reactor outside of the United States was under construction at Chalk River and it was commissioned in September 1945. Once employed there, one of Ken Smith’s principal tasks was to develop methods of making large numbers of quartz fibre radiation dosimeters that could be used routinely by those subject to radiation in the course of their work for the project. Thus, Ken was able to witness the very rapid bringing into operation of a large new physics based laboratory in a new and initially empty building and also to gather experience of working with radioactive substances. The latter experience would later help him to make one of his most outstanding contributions to physics. By 1946 the war was over and the British members of the bomb project were brought back to the UK and were set to work to establish the UK Atomic Energy Authority also directed by Cockroft on the site of an old airfield at Harwell. There were two principal groups there. Experimental Physics was headed by Otto Frisch who had with Rudolf Peierls in Birmingham first drawn general attention in what became known as ‘The 1940 Memorandum’ to a feasible method of making a super-bomb based on nuclear fission. The other group was Theoretical Physics headed by Klaus Fuchs who some years later was imprisoned for spying. By 1946 most of the important information gathered from the wartime bomb project by the assembled British scientists had been archived and, being suitably debriefed, they were free to resume their normal lives. Frisch acquired the post of Jacksonian professor in the Cavendish, since Cockroft was to remain head of AERE Harwell, and Ken Smith went to Cambridge to become a research student supervised by Frisch. Cambridge physics had suffered considerably from the effects of the war and Frisch was anxious to build up the department again. To this end, he set Ken, and another new PhD student, Ted Bellamy, to work on a completely new line of work for the Cavendish called Atomic Beam Radiofrequency Spectroscopy. It was not Frisch’s own interest, which remained mainstream nuclear physics, and he left Ken and Ted with much freedom to get on with the allotted task. Between them they built a large and complex atomic beam machine with several interconnected vacuum chambers and a refined system of atomic beam optics terminating in a mass spectrometer. To achieve this Ken had to spend many hours driving lathes and other machines in the mechanics workshops and also in acquiring skills in the fast developing field of electronics which was of ever increasing importance in physics research. This experience was invaluable when he became responsible for establishing the mechanical and electronics workshops for Physics at Sussex.
Within three years Ken and Ted had a fully working beam machine and had each made some publishable atomic physics measurements. Indeed just within the three years Ted had submitted his thesis and left the Cavendish. In a further three months Ken had completed his own thesis and immediately his measurements produced a major reaction of admiration, even incredulity, in the field of atomic beam spectroscopy – he had measured directly the spins and magnetic moments of four radioactive nuclei (24Na,134Cs, 131Cs and 86Rb) and each one was an example of a type of measurement which had been regarded by more established persons in the field as being impossible. From his experience of radioactivity gained in Canada, Ken had seen that these measurements were, on the contrary, certainly possible! From then on, he became de-facto head of the atomic beams group at the Cavendish. The group continued very productively for more than a decade until 1962 when he brought the group with him to Sussex.
Ken’s other relevant experience for tasks at Sussex came when in 1959 he was given the responsibility for interacting with the architects at the Cavendish Labs, who had been engaged to insert floors and rooms to fill the interior of a cavernous hall. This hall had previously housed the accelerator that had made possible the Nobel Prize winning pioneering experiments on man-made nuclear reactions carried out by Cockroft and Walton. Twenty-five years later their accelerator had been superseded by more advanced designs elsewhere and the space was to be filled with laboratories and offices suitable for those engaged in other scientific research. Soon after that, Ken found himself engaged on a similar task of dealing with architects in relation to the buildings Pevensey 1 and Pevensey 2 at Sussex. In Pevensey 1 the design and construction had really gone too far for Ken to rectify all its problems for providing physics research labs, but the plans for Pevensey 2 benefited greatly from his experience and it is most appropriate that Pevensey 2 now houses the entire Physics and Astronomy Department at Sussex. There was nevertheless, as soon as Ken was appointed, a great deal of temporary modification work to be overseen in Pevensey 1. For the two years 1962-64 Pevensey 1 was the only science building at Sussex and in the first instance it was planned to be a physics building. However, in the interests of starting the sciences as soon as possible a last minute decision was taken that Physics and Chemistry would share Pevensey 1 for the period 1962-64 until the first Chemistry building was completed two years later. This required among other things the temporary installation of work-benches suitable for practical chemistry and fume cupboards and a huge water drainage tank so that concentrated fluids such as acids would be well diluted before entering the main drains. During these initial two years, working space in that building was extremely tight with physics research apparatus being set up in all kinds of unforeseen places such as lecture theatre preparation rooms and in parts of what was ultimately to be the mechanical workshops.
Atomic beam magnetic resonance spectroscopy
Returning now to the activities of the Atomic Beams Group, the decade from 1951 to 1963 at the Cavendish saw the development of a programme with three main strands, which transferred to Sussex in 1962 and continued in that form at Sussex till 1966. The staff who came to Sussex with Ken Smith were Mike Pendlebury, DSIR Research Fellow, Peter Unsworth, Lecturer, and Geoffrey Rochester, Post-doctoral Fellow. Each was mainly responsible for one of the strands just mentioned. The first strand was the continuation of the measurements of the nuclear spins and magnetic moments of radioactive nuclei. The results continued to be relevant to the development of the shell model of the nucleus. Following Ken Smith’s initial work, Peter Nutter (thesis 1955) built a second beam machine and measured the nuclear spin and magnetic moment of the isomeric nucleus 116In*, then Ruben Title (thesis 1956) measured the spin of the radioactive 210Bi (Radium E) and also all the details of the ground state hyperfine structure of the stable isotope 209Bi. The latter measurement started the second strand of work, which was to measure the magnetic hyperfine splitting constants of the stable isotopes of the Group V elements, P, As, Sb, and Bi. These were of particular interest in relation to the development of our knowledge of how to calculate the electron structure of atoms.
In 1955, the work on radioactive nuclei was taken over by then graduate student Geoff Rochester who obtained results for 121Sb,122Sb, 123Sb, 124Sb and 72Ga. At the final stage of the radioactive nuclei programme the technique was stretched to work with substances with the shortest possible nuclear lifetimes. To this end Rochester built a third atomic beam machine, which was operated from 1963 to 1966 at the Harwell site a few metres distance form the nuclear reactor that was used to created the nuclei to be examined in the machine. Pellets of a stable isotope of the element of interest were pushed by compressed gas down a stainless steel tube into the core of the reactor where they would be irradiated with neutrons for about 30 mins. The pellets now containing newly created radioactive isotopes were then blown out again to a point where they dropped automatically into the already hot beam source of the atomic beam machine. Within half a minute, an atomic beam was established and magnetic resonance could be applied to the relevant atoms with measurements continuing for the next half hour. This last part of the programme was directed from Sussex and results were obtained for 206Tl, 66Cu, 108Ag, and 110Ag which have the half-lives 4.2 min, 5.3 min, 2.3 min, and 24 s respectively [C.J. Cussens, G.K. Rochester and K.F. Smith, J.Phys. A 2 (1969) pp 658-665.] The machine was described in J. Sci. Instrum. 41 (1964) 629. This strand of the studies ended in 1966 when Rochester obtained a lectureship in the Cosmic Ray Group at Imperial College.
The third strand of atomic beams work was an investigation of atoms of the rare earth elements. These have atomic configurations with unfilled shells of f-electrons with l = 3. The work was started by Ian Spalding in 1956 and continued at Sussex by Peter Unsworth [Proc Phys Soc 79 (1962) 787 & Proc Phys Soc 86 (1965) 1249]. There was an interest to know better the properties of the atomic states of these most complex of ground state atomic configurations. There was also a nuclear physics interest provided by the opportunity to measure nuclear quadrupole moments in a region of giant nuclear quadrupole moments that occurs among the rare earth elements and also to measure the nuclear magnetic moments. At first this work was carried out on the machine built by Nutter, which had a hot wire ionising beam detector. This was a relatively simple type of detector that worked very efficiently, but it only worked with atoms with low ionisation potentials such as alkalis and rare earths. At Sussex, the latter machine was modified by Unsworth incorporating more powerful Stern-Gerlach deflecting magnets and, for the magnetic resonances, a larger uniform field magnet (Newport Instruments).
The second stream of work involving the Group V elements required the fitting of ionisers using electron beams transverse to the atomic beam, in the original machine built by Ken Smith. Firstly in 1955 a simple ioniser built by Title was used. Then, in 1961, this was replaced by a more elaborate ioniser built by Pendlebury [J Sci Inst 43 (1966) 6]. These devices provided detection for beams of all elements and also of molecules. The more elaborate ioniser used mainly at Sussex was thought to have the highest signal to background of any in operation anywhere at the time.
The first working atomic beam machine at Sussex set up in a room destined for welding and sheet metal work!
The elements of Group V of the periodic table have half full p electron shells. The three p electrons have aligned spins in the ground state making a total spin S = 3/2, but being p electrons they have no electron density at the nucleus and therefore no direct magnetic interaction with the nucleus and thus no contribution to the magnetic hyperfine splitting of the ground state. Nevertheless, the group showed that these atoms do have finite magnetic hyperfine interactions, but only about 2% of the magnitude one might observe for an alkali atom. The idea for the origin of this splitting at the time was that the valence p-electrons have exchange interactions with the s electrons of the atomic core, which are not the same for spins parallel to the valence electrons and those with antiparallel spins. This ‘core polarisation’ destroys the equality in density between the spin-up and spin-down pairs of electrons in the core causing them to have net hyperfine interactions with the nucleus that are linked to the valence electron spin. Only an atomic wave function that is a sum of many configurations can explain the size of the resulting hyperfine level splitting. The 1960s was a time when the techniques for calculating such wave functions for atoms were being developed. There was thus considerable interest among the theorists concerned to see whether the wave functions, that they were producing, could explain the measured hyperfine interactions for the Group V elements. A French theorist, Carl Moser, made a special visit to the group at Sussex at the time so that he could try to judge for himself whether the results which the Sussex group had just obtained [Proc Phys Soc 84 (1964) 849] for the hyperfine structures of ground states of 31P and 75As were reliable. He was particularly interested in how the group measured the sign of the energy splitting since their results for 31P agreed quite well with the magnitude obtained in the experiments but not with the sign. As it happened, unbeknown to the Sussex group, an American group had, at almost exactly the same time, measured the hyperfine structure of the ground state of 31P using the method of optical pumping. It was a relief to find out later that their sign and the sign obtained at Sussex were both positive and the magnitudes agreed to well within the experimental errors of 0.01%. The disagreement with theory was not removed until 1968 when a group at the University of California [Phys. Rev. Lett. 21 (1968) 1139] showed that the dominant mechanism for explaining the hyperfine splitting in 31P was that of the Goldstone-Bruekner many body theory. The latter gave them a contribution to the splitting in 31P that was opposite in sign and approximately twice as big as that due to core polarisation.
The machine with the electron cross beam ioniser was also used to make hyperfine structure measurements on several transition elements [Proc Phys Soc 86 (1965) 1145 and 1249 and J. Phys. B 5 (1972) 386, thesis K.H. Channappa (1967) and thesis D.B. Ring (1968)]. The J Phys paper reporting on the hyperfine structure of 95Mo and97Mo demonstrated the success of a molten ball atomic beam source for refractory metals developed at Sussex. Later it was pointed out by another group that due to an unusual combination of parameters for this atom the neglect of the mixing of fine structure J states by the magnetic hyperfine interactions perturbed the extraction process for the very small nuclear quadrupole interaction constants by enough to make a significant error in the conclusions about the quadrupole moments.
Measurements were also made on the rotational states of S2 molecules for all rotational states between J = 10 and J = 70 leading to the extraction of several coupling parameters. This molecule is paramagnetic with S = 1 just like O2 which is in the same group of the periodic table [Colloques Internationaux…...]
Other fundamental atomic and neutron physics investigations
On learning of the startling announcement of the detection of CP violation Kaon decays in 1964, Roger Blyn-Stoyle, head of theoretical physics at Sussex, came immediately to the Atomic Beams Group members to ask whether they could think of measurements that they could make that would contribute to this new and important topic. The reply to him was to suggest making more precise measurements of the electric dipole moment of the neutron (nEDM). At that time just one such measurement had been made in 1951 and there seemed to be considerable scope for improvement in the precision.
In fact, the nEDM became a programme of work at Sussex that has lasted up to the time of writing this account in 2010. It was also the first stage of a shift in the interests of the Atomic Beams Group towards a new range of fundamental physics measurements. Within a few years, for example, Ken Smith had carried out a project on the time variation of photon emission rates in spontaneous decay [thesis M.J.R. Armistead 1971]. Peter Unsworth became engaged in an experiment to understand light shifts [thesis P.T.Woods 1972] and then a challenging experiment to measure the Lamb shift in the n = 2 state of hydrogen, which by the time of the final publication, was the most precise of all such measurements ever made [G Newton, D.A. Andrews and P.J. Unsworth, Phil. Trans. Roy. Soc., Series A, 290 (1979) 373-404]. Also in this period, Pendlebury, in addition to the neutron work, studied some aspects of atomic charge-exchange reactions [theses, A.P. Cluley 1972 and L. F. De Souza-Coelho 1981]. Starting in about 1969 and throughout the period just mentioned, the activities of what was now better called the Atomic Physics Group were reinforced by the contributions of two long-term research fellows, Gavin Newton [Sussex thesis 1970] and Robert Golub who came from Brandeis University, USA.
The first Sussex neutron EDM measurement was carried out in 1969 using a neutron beam at the reactor near Aldermaston and was written up in the Sussex thesis of K.N. Baird . The Sussex group also collaborated with American and French groups in the neutron EDM measurement made at the Institut Laue Langevin (ILL), Grenoble [W.B. Dress, P.D. Miller, J. M. Pendlebury et al. Phys. Rev. D 15 (1977) 9-21].
In the period 1975-7 the Sussex group was able to participate in the same collaboration, to make the most precise measurement of the neutron magnetic moment [Phys. Rev. D 20 (1979) 2139-53]. In 2010 this still stands as the most precise result. The use in the latter experiment of the method of Ramsey magnetic resonance with protons in flowing water, was suggested by Ken Smith and Mike Pendlebury [Rev. Sci. Instrum. 50 (1979) 535-40]. There is an example of it in the Sussex Year 4 Undergraduate Physics Laboratory. This flowing water method played an important part in the success of the neutron MDM measurement.
The attempts to measure the neutron EDM that took place in the period 1965-69 highlighted the advantages of using neutrons with the lowest speeds possible. Thus, when results were announced by groups at TU Munich and at JNIR Dubna in 1968 and 1969 that they had observed the bottling of very slow neutrons, with velocity ~5 m/s, called ultra-cold neutrons (UCN) for periods of the order 100 s, the news was immediately of great interest to the Sussex group [review article, R. Golub and J.M. Pendlebury, Contemp. Phys. 13 (1972) 519-558]. Eventually the Sussex group became the first to measure the neutron EDM using completely trapped neutrons [Phys Lett. B136 (1984) 327]. In the meantime, there had been much work to establish a source of such slow neutrons in the ILL, Grenoble and to learn how to make suitable detectors and polarizers etc. See, for example, the Sussex theses of [N. Mufti 1977, A.R. Taylor, 1977 and S.M. Burnett, 1982] and the review [Rep Prog. Phys. 42 (1979) 439-501]. There was for a period of several years at this time an accelerator-based neutron generator that was operated from time to time in the Pevensey 1 building. It was used to produce small numbers of UCN that heped us to develop efficient low background detectors and thin film polarizers for UCN. Other special studies were required in order to develop a suitable multilayer mu-metal magnetic shield and magnetometers for the nEDM project [thesis T.J. Sumner 1980 and J. Phys. D 20 (1987) 1095-1101].
In 1975, Golub and Pendlebury [Phys. Lett. 53A (1975) 113-5 and Phys. Lett. 62A (1977) 337-9] identified what they called a ‘superthermal source’ principle for the production of ultracold neutrons. Now, 35 years later, this principle is being employed in all the new sources of UCN coming on stream. The name derives from the fact that a higher density of UCN is obtained by a short period of cooling of the incoming cold neutrons than would be the case if they had time to come into thermal equilibrium with the ultra-cold moderator.
In 1979 with the hydrogen Lamb shift experiment very successfully completed and fully written up, Peter Unsworth, stimulated by the concerns about the world’s energy supplies as highlighted by the 1970s oil crisis, decided he would like to change the direction of his research to one involving the development of intelligent electronics for energy saving purposes. He started out in this direction by concerning himself with the problem of more efficient control of the frequent stopping and starting of the millions of electric motors in use in industry.
As a result of Peter’s departure, the Atomic Physics Group was reduced to just two faculty KFS and JMP who were by then concentrating on the neutron physics program. However, two members of the Nuclear Physics Group at Sussex, Jim Byrne and Peter Dawber, were at that time also planning to carry out neutron physics experiments, in their case concentrating on the measurement of neutron decay properties. There was also at that time an offer to Sussex of collaboration by an Italian group and by a group at the Rutherford Appleton Laboratory who were proposing to carry out an experiment to investigate whether there are neutron to antineutron oscillations in a free neutron beam. In addition, an American group wished to collaborate in experiments to measure parity violating neutron spin-rotation, which occurs when polarized cold neutrons pass through unpolarized solid targets. This all added up to a neutron physics group of viable size working in a well defined, and not overpopulated, field of physics. The spin-rotation collaboration soon began to produce results, e.g., [Phys. Lett. 119B (1982) 298 and Phys. Rev. C29 (1984) 2389, etc.] and a little later the neutron to antineutron oscillation measurement [Phys. Lett. 156B (1985) 122]. These were followed by an experiment using UCN to measure the neutron decay lifetime [Phys. Rev. Lett. 63 (1989) 594].
Returning to the account of the neutron EDM project, the 1984 result cited above was followed in 1986 by the coming on stream at the ILL of a new 50 times more intense source of UCN. This made possible a new phase of data taking that took place in the period 1986 to 1989 with essentially the same apparatus as used previously. The result was published as [Phys Lett 234B (1990) 191]. In parallel, there were further studies of UCN polarizers [thesis N. Crampin 1989] and on the construction and uses of a UCN monochromator [thesis D.J. Richardson 1989]. Studies of 3He gas as a magnetometer for the nEDM measurements were also made [thesis P W Franks 1986]. The group also wrote reviews on ‘ultracold neutrons’ [Golub and Pendlebury Rep. Prog. Phys. 42 (1979) 439-501] and on ‘molecular beams’ [Pendlebury and Smith, Contemp. Phys. 28 (1987) 3-32].
Also during the new data-taking period just mentioned, the group started to work on the development of a new magnetometer for the neutron EDM experiment that was to be based on free precession of nuclear spin-polarised 199Hg atoms, which would simultaneously be stored with and occupy the same volume as the stored UCN [Theses Y. Chibane 1990 and M. Chouder, 1993]. This magnetometer took five years to develop, but was ultimately very successful [Nucl. Instr. and Meth. A 404 (1997) 381-393]. Fortunately the delay incurred largely overlapped with the delay caused when the ILL reactor went out of service in 1991. The fault was such that the reactor’s central tank had to be completely replaced and neutron beams were not available again until 1996.