The achievements of Oxford Instruments Superconductivity as a world leader in the research and development of superconducting magnets recently gained them a place among the finalists of the 2003 MacRobert Award. Here, Martin Townsend outlines the latest advances in nuclear magnetic resonance, a technique that allows examination of molecules by exploiting the unique behaviour of atoms placed in a magnetic field.
Thanks to superconductivity, magnets can provide a much higher magnetic field for a much smaller footprint and lower running costs than has ever been possible with permanent magnets or even resistive electromagnets. Indeed, the properties of the superconductors forming the magnet mean that they are capable of carrying large currents with no loss of energy. However, the issues involved in building such superconducting (SC) magnets are far from simple. Oxford Instruments Superconductivity have developed the world’s largest commercial nuclear magnetic resonance (NMR) magnet and recently became the first company to produce 900 MHz magnets on a commercial basis. Oxford Instruments Superconductivity is a world leader in the supply of SC magnets and low temperature cryogenic systems to the scientific and industrial research community. The company has over 6000 SC magnets installed worldwide and supplies 50% of the world’s requirement. The 900 MHz magnets can reach maximum field strength of 21.14 Tesla, or 400 000 times the strength of the Earth’s magnetic field. This article looks at some of the applications in which SC magnets are used and the engineering needed to make this possible.
From lodestones to superconductors People have been aware of magnetism for thousands of years, although until relatively recently (around 1820), the only magnets available were naturally occurring magnetic rocks called ‘lodestones’ (magnetite). These gave rise to the word ‘magnetism’, after the district of Magnesia in Asia Minor where lodestones were originally found. Total magnetic field strength was measured in gauss, but most measurements are now given in the much larger units of Tesla (1 T = 104 gauss). To put the strength of fields produced by SC magnets into perspective, the Earth’s field is relatively weak and averages only 0.5 gauss, whilst a standard bar magnet produces a magnetic field strength of approximately 3000 gauss/0.3 T at its poles. In contrast, the most powerful commercial SC magnets currently available can achieve a stable field of up to 21.14 T.
Why superconducting magnets?
At first glance, SC magnets might seem much more complicated than electromagnets, especially with the requirement for low temperatures to keep the magnet solenoid in its superconducting state. However, they have several advantages over their permanent or electromagnetic counterparts.
First, SC technology allows users to produce extremely high magnetic fields without the many kilowatt, or even megawatt, power supplies needed for electromagnets. Once ‘brought to field’, SC magnets can be disconnected from their power source and function in persistent mode, resulting in significant savings in electricity costs.
Second, SC magnets can also generate a far higher field than permanent magnets, which are limited to 2 T. SC magnets are currently capable of up to 21.14 T. This can be extended to around 45 T with the addition of a resistive (or electro-) magnet, but obviously this removes some of the advantages, as a permanent electricity supply is now needed. Although electromagnets are capable of generating fields of up to 35 T, the power consumption required for this would be considerable. In addition, an SC magnet field can be kept extremely stable with a very low drift rate and very high homogeneity, features that are essential for applications such as NMR spectroscopy and fourier transform mass spectrometry (FT-MS).
Third, SC magnets have a very small footprint in comparison to electromagnets (typically the coils of a resistive 1 T magnet are 10 times larger than a SC 10 T magnet), and the space needed is further reduced because they do not require water cooling for the power supply or solenoid. However, a cryogen such as liquid helium is needed to cool the magnet to below its superconducting critical temperature (Tc) and liquid helium is relatively expensive in some parts of the world. For the superconductor niobium titanium, Tc is 9 K; above this the material becomes resistive.
SC magnets are used the world over in a huge variety of applications. In healthcare, magnetic resonance imaging, which is almost universally employed for clinical diagnosis, is dependent upon high quality SC magnets. Molecular biology also benefits from the SC magnets found at the heart of high-resolution NMR techniques that are essential in drug discovery and development.
Bigger and better
SC magnets have been continually evolving over the last 40 years to become larger and more powerful, culminating recently in the commercial availability of 900 MHz NMR magnets. So why have people continued to demand bigger and better magnets, and what benefits do they bring?
Several essential features of NMR spectrometer performance are controlled by the size and quality of the SC magnet found at the centre of the instrument. As the strength of the magnetic field is increased, the resolution and signal-to-noise (S/N) ratio are also increased. This leads to higher information content, allowing the detection and characterisation of smaller amounts of material and more complex molecules – proteins of up to 100 kilo-Daltons have been successfully analysed at 900 MHz.
However, in order to achieve this greater detail, the SC magnet needs to provide a magnetic field of very high homogeneity and stability. The stability of the magnetic field also needs to be kept to within 10 Hz/hr (or around ten parts per billion per hour) in order to maintain high quality results. The homogeneity has to be better than 1 part in ten billion over the sample volume (typically 5 mm diameter by 25 mm long). Many of these factors become more difficult to maintain with magnets of increasing size and strength. In fact, even the increase in magnetic field strength from 800 to 900 MHz required several critical parts of the magnet to be redesigned.
First, a conductor has to be chosen that can provide a stable field at 900 MHz, since the field drift rate has to be kept to 1 in 108 for high-resolution NMR experiments. Niobium–titanium wire was developed in the 1970s and has been a ‘workhorse’ superconductor ever since then. This is used in the lower field regions of the magnet coil, but cannot carry sufficient current in the higher field section. A unique niobium–tin wire, made up of thousands of filaments, was used to achieve this. The filament structure improves the stability of the magnet by preventing ‘flux jumping’, which dissipates energy in the superconductor. The manufacture of this wire has to include stringent quality control procedures to ensure that short samples of wire perform at field and that a constant diameter is maintained. Any fluctuation in the wire diameter will result in detrimental effects on magnet stability.
However, it is practically impossible to manufacture a single length of wire to this standard when the solenoid for a 900 MHz magnet uses in the order of 290 km of wire. Therefore, low resistance jointing techniques were developed to join lengths of both niobium–titanium and niobium–tin wire together.
The biggest commercial wide-bore 900 MHz magnet in the world, at the Pacific Northwest National Laboratory (PNNL) in Washington, USA, contains 27 MJ of stored energy when at field. The 900 MHz magnets currently being produced commercially have 17 MJ of stored energy; although this may sound like a lot less, it still equates to 4 kg of TNT exploding.
For the SC material to operate, it must be kept in a bath of liquid helium at 2.2 K. A few ìJ of energy, equivalent to a pin dropping from the height of a few centimetres, would be enough to raise the temperature sufficiently to cause the magnet to become resistive, or ‘quench’. Here, the helium boils off and 17 MJ of stored energy is released very quickly, risking damage to the magnet structure.
In addition to the amount of stored energy, the stresses experienced by the magnet are huge. Mechanical stress increases quadratically with the field strength for a given magnet. At 900 MHz, these stresses are greater than 200 tonnes, with a magnetic pressure of more than 250 MPa.
Traditional ways of reinforcing the coils (wax impregnation) were insufficient at these high strength fields. For the 900 MHz project, the coil was evacuated in a special vacuum chamber and the chamber let back up to atmospheric pressure with epoxy resin to replace the air voids. This enables the coils to withstand forces in excess of 200 tonnes.
In the event of magnet failure, how can you manage the dissipation of 17 MJ of energy without causing terminal damage to the magnet structure? The challenge is to develop ways of releasing the energy very quickly, in a manner that avoids magnet damage through thermal gradients or excessive voltages in the coil. To address this, EMS™ (Energy Management System) was developed to ensure that, during failure mode, all coil stresses and voltages are kept within design limits. Heaters are used to make the magnet coils resistive, dispersing the energy from the quench evenly and preventing sections of the coil being damaged by excessive voltages.
The cryogenic environment
At present, the majority of SC magnets are used in wet cryogenic environments (helium liquid). However, increasingly the trend is to replace this helium liquid with mechanical cryocoolers. These give more flexibility and ease of use for the end-user, removing the need for helium refills. They also allow the use of SC magnets to be expanded outside laboratories towards industrial applications where the external environment is very demanding.
High field NMR cryostats need to maintain 2.2 K while still allowing for refills at atmospheric pressure. This temperature has the effect of increasing the superconductor’s critical current characteristics, thus increasing the upper critical field limit. The most straightforward way of achieving this is to pump directly on the liquid helium reservoir. The structure also needs to be able to support the magnet mass, even under changing levels of cryogen and evaporation rates, whilst minimising any heat leak into the system that could raise the temperature above 2.2 K. In addition, it needs to be strong enough to withstand the discharge of large volumes of helium gas if the magnet quenches.
Aside from magnet performance, it is also important to take into account the positioning of the system. If a high field magnet can wipe a credit card from five metres away, it needs to be set in a controlled environment. This is where shielding can play an essential role.
Active shielding utilises a SC core wound in the opposite direction to the main solenoid, giving an opposing field that limits the stray field from the main coil. However, this is not financially viable for larger magnets, as the main solenoid needs to be made bigger to reach its target field strength, which can result in a 20–25% cost increase. Passive shielding works by placing a specially designed structure around the magnet. Systems such as Oxford Instruments’ Vectorshield™ give 800/900 MHz NMR systems protection from the negative effects exerted by external electromagnetic disturbances on NMR data quality. This helps maintain quality data as well as limiting the stray field. Shielding can also help with security and safety and enables magnets to be housed in, for example, built-up areas that would not otherwise be suitable.
What is NMR?
NMR is a unique, non-destructive tool for the examination of molecules. The method exploits the fact that many common atoms, such as the hydrogen atom, will resonate (absorb energy at a specific frequency) when placed in a strong magnetic field. Each hydrogen or carbon atom in a given molecule also resonates at a different frequency based on its local chemical environment. This ‘chemical shift’ allows the scientist to visualise not only how many hydrogen and carbon atoms are present in a molecule, but also how they are bound together. NMR therefore has a diverse range of applications, from the stabilisation of nuclear waste to new cures for cancer.
Towards 1 gigaHertz NMR
As magnet technology continues to progress, more powerful NMR systems with improved sensitivity and resolution will allow more detailed experiments to be carried out on larger molecules. Increasing the size and strength of SC magnets further, however, puts far more pressure on the conductor. Currently, there are no conductors available that can give 100 A/mm2 (the critical current density needed for superconductivity) at 1 GHz/23.1 T. Using new SC materials such as niobium–aluminium may solve this problem. Alternatively, high temperature superconductors (HTS) may prove useful. These are largely ceramic materials that retain their SC properties up to temperatures of more than 90 K. However, it is difficult to achieve the necessary conductor performance with high temperature superconductors, as they are brittle and very difficult to wind in the manner of a wire solenoid.
Nevertheless, together with cooling systems that remove the need for liquid helium, HTS materials may help take SC magnets out of the laboratory and into more demanding environments. Other future developments may include decreasing the size of magnets – although they are much smaller than electromagnets of equivalent strength, the most powerful high field SC magnets are still very large. Developments such as these should mean that NMR will remain an essential tool for researchers in structural biology, materials science and numerous other research areas for many years to come.
Oxford Instruments Superconductivity
Martin Townsend joined Oxford Instruments as an engineering apprentice in 1979. Having worked in a variety of functions he most recently project managed the development of the award winning 900 MHz NMR magnet systems. He now works in business development in the biomedical sector. Martin has never stood still in his career and he continues his studies at present with an MBA in Technology Management.