A super(conducting) career
Joe Brown explains why he is still enthusiastic about designing and manufacturing superconducting magnets after nearly 40 years in the industry.
Shortly after the discovery of superconductivity in 1911, many scientists believed that it would soon be possible to construct electromagnets that could generate high fields without the high power requirements of conventional resistive windings. Those hopes were, however, quickly dashed when it was discovered that the presence of magnetic fields of ~30 mT destroyed a material's ability to carry current without resistance. It would be another 25 years before researchers found materials such as PbTl2 that retained some ability to carry current without resistance in the presence of a magnetic field, and it was not until the late 1950s and early 1960s that materials such as Nb3Sn and NbTi were developed into forms that would allow superconducting magnets to be manufactured commercially.
One of the first firms to take advantage of these developments was Oxford Instruments, which was formed in 1959 as the first spin-out company from the University of Oxford. Today, superconducting magnets have applications that range from the "big physics" of the Large Hadron Collider through to magnetic resonance imaging (MRI) machines used in medical diagnosis, and producing or maintaining them is still very much a part of Oxford Instruments' activities. As a consultant magnet engineer in the firm's nanoscience division, I am involved at every step of the process, from understanding customers' requirements, through design and manufacture, to delivery, installation and technical support.
From welder to magnet engineer
My involvement with superconductivity began almost by accident when, in 1972, I applied for a job as a welder at a company in Oxfordshire. I had done my apprenticeship as a fitter/welder at the UK Atomic Energy Research facility in nearby Harwell, and then spent five years developing advanced welding technologies related to nuclear-fuel and medical-isotope containment. During those years, I had also studied applied physics part-time at what was then Oxford Polytechnic (now Oxford Brookes University).
Perhaps because of this experience, instead of offering me a position as a welder, the company, Thor Cryogenics, asked if I was interested in a role as a superconducting magnet technician – someone responsible for winding and assembling superconducting magnets. I had always been interested in science, and the opportunity to work in a company using superconductivity was very attractive to me, so I said yes.
My role as a magnet technician developed, and by the time I joined Oxford Instruments in 1986 I had moved into project engineering, where I mixed technical activities such as designing magnets and cryogenic systems with project-management work like ensuring equipment was built on time and within commercial constraints. In 1999 I progressed to my current role, which is biased towards the technical aspects of magnet design. However, most magnet engineers do have some project-management responsibilities, and I am also involved in mentoring and training junior colleagues, visiting customer laboratories and speaking at conferences.
My career path has not been a typical one: most of my colleagues who have joined Oxford Instruments in recent years have taken the more academic route of full-time university education to first degree or even PhD level. However, there is little formal training on the specifics of magnet design available, so much of the required knowledge has to be gained "on the job". This means that a good general education to degree level in physics or engineering is adequate as a foundation because it gives you the information you need to understand the concepts and processes involved in magnet design and construction.
The design process for a superconducting magnet is a marriage of mathematical modelling and engineering. Today, much of the modelling is done via computer programs, most of which have been developed to provide the specific information needed by the magnet designer. There is, however, a large part of magnet design that is based on empirical data that have built up over the years, related to the processes used to build a working magnet. This is where the engineering comes into play, with the need to understand how to work with the materials and structures used in magnet construction.
A good example of the type of magnet I am currently working on is one designed for use in neutron-scattering experiments. This magnet has two windings separated by a gap through which researchers can fire a neut_ron beam at the sample being studied and observe the resultant scattered neutrons. This type of magnet is known as a "split pair" and several factors make it particularly challenging to design. One is the huge attractive forces between the two halves when the magnet is energized, which can be as high as a few hundred tonnes. Such forces present challenges for the mechanical structure and the interfaces between coils and supporting structure; if the magnet is not designed and manufactured correctly, its performance can degrade over time. The usual solution is to separate the two halves of the magnet with a series of aluminium alloy rings, which, while sufficiently transparent to neutrons, are strong enough to support the attractive force.
Another design challenge with this type of magnet is that there are trade-offs between the particular geometry of the coils that would minimize the superconductor volume (and hence cost), and geometries that produce the required uniformity of magnetic field over the sample volume. Resolving this problem normally comes down to a compromise depending on the individual circumstances: financial versus technical.
A third consideration is the need for sufficient operating margins in terms of flux density, current density and temperature for the superconducting wires used within the magnet coils. There are also design considerations related to dimensional constraints such as the size of the samples and the neutron-scattering angles. The magnet's overall size, both mechanically and in terms of "magnetic footprint" (stray flux density), can be a problem in many applications because of restricted access or proximity to other equipment.
When we design magnet structures we do so with the aid of finite-element modelling, where the magnet assembly is computer modelled under its loaded condition to determine stress and strain magnitudes and distributions. This is an iterative process in which the structural components are optimized to provide the required structural integrity. After a magnet has been designed, the next steps are to manufacture and test it. Here, engineers like me are involved at every stage, from defining manufacturing processes and testing strategies to analysing test results and presenting them in the form of operating instructions and manuals.
With my roots firmly in the practical side, I find my role very satisfying because it means that I get to be involved with the complete process – from the customer's first ideas of what they require to a piece of hardware that allows the experiment to be performed. I also get a great sense of pride when I read an article or paper detailing the experimental results obtained using a magnet I have designed and helped to manufacture. Most of the magnets we manufacture at Oxford Instruments are for laboratory-based research and tend to be "one-offs" specially designed to suit a particular set of experimental requirements.
When it comes to job satisfaction, I think the fact that I have been in the business for almost 40 years says it all. It has not always been easy because, even after 50 years of superconducting-magnet manufacture, there are still times when a new behaviour of a magnet will catch you out, leading to sleepless nights when you are trying to work out what is going on. However, it is very rewarding, and for anyone looking for a challenging role in a hi-tech industry that is still developing, I cannot think of a better place to be.
About the author
Joe Brown is a consultant magnet engineer at Oxford Instruments NanoScience, UK, e-mail firstname.lastname@example.org.
This article appears in the April 2011 issue of Physics World.
last edited: November 05, 2014