GB2612838A

GB2612838A - Apparatus for pumping hyperpolarised gas and method of handling hyperpolarised gas

Info

Publication number: GB2612838A
Application number: GB2116444.7A
Authority: GB
Inventors: Parker Evan; Williams David
Original assignee: Hyperspin Ltd
Current assignee: Hyperspin Ltd
Priority date: 2021-11-15
Filing date: 2021-11-15
Publication date: 2023-05-17
Also published as: GB202116444D0; WO2023084248A1

Abstract

Apparatus for pumping hyperpolarised gas, and method of handling hyperpolarized gas, having a dry pump 500 comprising a pump body 520 which together with a compression member 530 forms a pump chamber 520A, the compression member being moved axially by a non-magnetic linear drive 580 coupled 585 to the compression member; inlet and outlet valves 542A, 552A, having respectively an inlet actuator 546 and outlet actuator 556, selectively allow the gas into and out of the pump chamber; the non-magnetic linear drive is arranged to provide minimal lateral forces on the pump body and the pump is made only of non-magnetic materials so as to avoid depolarisation of the hyperpolarised gas. The drive may be piezo-electric, pneumatic, or hydraulic. Inlet and outlet actuators may be electronically controlled to allow independent actuation. Suitable pump materials include non-magnetic stainless steel; alloys of titanium, aluminium; ceramics and plastics. The compression member and pump body can have an ultra-smooth coating made of titanium, aluminium, or gold to reduce the depolarising effect. The pump can be used for UHV and finds application in improving relaxation time in medical MRI and NMR for pumping noble gases such as Helium, Neon, Krypton, and isotopes of Xenon.

Description

Apparatus for pumping hyperpolarised gas and method of handling hyperpolarised gas

Field of the invention

The present invention is directed towards pumping and compressing hyperpolarised gases, such as hyperpolarised noble gases, whilst avoiding excessive depolarisation of the gas. In particular, the invention is directed towards pumping and compressing hyperpolarised gas so as to be convenient for medical and biomedical applications.

Backa round In recent years, work has been carried out on the use of hyperpolarised gases, such as hyperpolarised noble gases, to enhance the resolution and signal-to-noise ratio in nuclear magnetic resonance (NMR) applications, and in particular to provide information/contrast on state of biomedical surfaces in magnetic resonance imaging (MRI). Particular interest has been generated for the use of hyperpolarised gases in the imaging of patients with respiratory diseases such as asthma, COPD, pulmonary fibrosis and recently, particularly Covid-1 9 related conditions. Advantages of hyperpolarised gases are that the gases can be inhaled by the patient and an increase of resolution and signal-to-noise ratio of many orders of magnitude can be achieved.

In NMR, a strong magnetic field is applied to the specimen in order to align the nuclear magnetic moments of nuclei in the specimen. Not all atoms have nuclei with a nuclear magnetic moment, and only atoms with a nucleus having nuclear magnetic moment can be analysed. The most common nuclei investigated in standard NMR is 1H and 13C, as these have non-zero nuclear magnetic moments and are commonly found in organic specimens.

Both 1H and 13C have a nuclear spin of one-half, and the nucleus has two linearly independent spin states with m = +1/2 or m = -1/2. The strong magnetic field used in NMR aligns the nuclear magnetic moments of the nuclei either parallel to the magnetic field lines (m = +1/2) or anti parallel to the magnetic field lines (m = -1/2). This causes a splitting in the energy degeneracy of these two states, with the aligned state having a lower energy than the anti-aligned state. Therefore, if the population of the m = +1/2 and m = -1/2 are not equal, there will be a residual magnetic field caused by the nuclear magnetic moments of the nuclei in the two states not exactly cancelling out. This residual magnetic field is what is detected in NMR, and the higher the residual magnetic field, the larger the NMR signal that is produced.

MRI is an application of NMR and uses the NMR signal derived from different volumes of a specimen under examination to provide an image. MRI has many medical uses that will be familiar to most people. Commonly, when a patient is investigated using MRI, they must lie down within a small cavity inside a large apparatus that includes a very strong cryogenically-cooled superconducting magnet. An MRI apparatus is very expensive to purchase and to maintain, and so they are usually only found in specialist departments of hospitals. Further, they can be very stressful for the patient, as they must lie down in a small space for a relatively long time, with loud noises caused by the magnetic coils being activated and deactivated.

Whilst MRI has been very successful at imaging certain parts of the body, in particular the brain and spinal column, it has not been as successful for imaging other parts of the body, like the lungs. This is because the lungs are mostly filled with gases that do not provide a very strong NMR signal, and so provide poor contrast in an MRI image.

The difference in populations between the upper and lower states spin states of a nucleus during NMR (or MRI) is governed by thermal statistics and the populations of each state is described according to the Boltzmann distribution: Pi (Ej-E0 -= e kT Pi Where pi and pi are the probabilities of being in state i or state], respectively, and Ei-El is the energy difference between the two states.

Since the energy splitting of the nuclear spin states is quite small, only a small difference in population levels occurs, even with very strong magnetic fields (field strengths of around 1.5 Tesla are typically used, but could be up to 7 Tesla or higher). Therefore, the total nuclear magnetic moment of the atoms is usually quite small. As noted above, to image the lungs of a patient is particularly problematic, as the volume of the lungs is filled only with gas, and so only low resolution images are obtained. This is compounded by the fact that only short imaging times are possible, as the patient must effectively hold their breath when imaging. Therefore, realistic imaging times of only a few tens of seconds is possible.

However, it has been discovered that the introduction of hyperpolarised gases into the lungs can significantly increase the resolution of the MRI images, even with short imaging times. There is particular interest in the use of hyperpolarised noble gases such as He3, Kr83 and Xe129, but other gases may be used.

Hyperpolarised gases are gases that have a total nuclear spin polarization in a magnetic field that is significantly greater than what would normally be achieved under thermal equilibrium. As noted above, the populations of the spin energy levels of a nuclei that possess a nuclear magnetic moment in a strong magnetic field is determined by the Boltzmann distribution, and so under normal conditions, the difference in populations is small and there is only a small residual magnetisation. However, it has been found that it is possible to significantly alter the populations of the energy levels away from the thermal equilibrium levels, in a manner that has similarities to the optical pumping in gas lasers such as He-Ne lasers, even to the point that almost all of the nuclei are in the same spin state.

That is, each of the nuclear spins of the nuclei are aligned in the same direction (which is usually the direction of the applied magnetic field). What is more, this alignment can be maintained in suitable containers for many hours without significant relaxation back to the thermal equilibrium distribution.

There are several methods in which hyperpolarised gases can be prepared, including Spin Exchange Optical Pumping (SEOP), Metastability Exchange Optical Pumping (MEOP) and Dynamic Nuclear Polarisation (DNP). We will briefly discuss the most common of these, SEOP.

Fig. 1 shows a simplified mechanism of the hyperpolarising a noble gas using SEOP. SEOP is particularly effective in producing hyperpolarised noble gases such as bipolar 3He, 123Xe, and quadrupolar 131xe, 83Kr, and 21Ne. In SEOP, the noble gas is mixed with gaseous atoms of an alkali metal such as rubidium or caesium in an optical pumping cell. In the case of rubidium, a magnetic field is applied to the optical pumping cell such that the degeneracy of the rubidium Di line between the 5251/2 and 52P1/2 electronic energy levels is lifted. In fig. 1, the magnetic field is up the page, with the electron and nuclear spins being aligned either with in an up direction or a down direction. A buffer gas is commonly added to the mixture of noble gas and alkali metal. The buffer gas can help stabilise the excited state of the alkali metal. Common buffer gases include nitrogen, helium and hydrogen, and are optimised depending on the noble gas and alkali metal used.

Then, left-circularly polarised laser light is passed through the optical pumping cell at the wavelength of this transition. In fig. 1, the laser light comes in from the left. In some methods of SEOP, laser light is incident on the atoms from both sides of the optical pumping cell. The left-circular polarisation of the laser ensures that excitation only occurs from the m = -1/2 spin state of the 55 level to the m = +1/2 of the 5P level, such that an excess population of rubidium atoms excited in the 5P level with m = +1/2 is produced. That is, in fig. 1, the electronic transition is from spins pointing down to spins pointing up the page. Subsequent collisions of the excited rubidium atoms with m = +1/2 with atoms of the noble gas with nuclear spins with m = -1/2 lead to exchange of their respective spins. By this mechanism, the noble gas can be polarised so that the majority of the nuclear spins are in the m = +1/2 state (up the page). Optimum polarisation generally occurs when the SEOP process is carried out when the optical pumping cell is at an elevated temperature, typically between 50°C and 250°C, which may be optimised depending on the gas being hyperpolarised.

The buffer gas and the alkali metal must be removed from the hyperpolarised gas before it can be stored or transferred for measurement. Alkali metals are often removed by condensation by cooling the gases after the hyperpolarisation step. When xenon is used as the noble gas, this can be preferentially frozen using liquid nitrogen, as it is the only gas in the admixture that is a solid at this temperature. Krypton is more difficult to separate, but recent studies have suggested using hydrogen as the buffer layer and then using combustion or oxidation to separate the buffer from the mixture, for example as discussed in W02017042544 (Al) Using hyperpolarised gases in NMR applications, such as MRI of the lungs, the residual magnetic field of the gases being imaged is significantly increased, and the NMR signal is therefore much easier to detect.

In one application of MRI using hyperpolarised gases, the patient would be asked to inhale a prepared volume of hyperpolarised gas. Noble gases are particularly useful in the medical applications, as they are inert and unlikely to harm the patient. Whilst the patient holds the hyperpolarised gas in their lungs, an MRI image is taken. The patient may be imaged in a conventional MRI apparatus. However, the very high polarisation of the gas allows for a relaxation of the conditions usually required to take a higher resolution image compared to standard MRI techniques. In particular, faster imaging times are possible, as well as less constraints on the strength of the main magnetic field and the sensitivity of the NMR probes.

The increased resolution allows for less strong magnetic fields to be used, and an apparatus may be built that would allow the patient to stand up or even sit down during the MRI procedure, rather than having to lie down within the confines of a cryogenically-cooled super-conducting electromagnet. This would be less stressful for the patient. This could have enormous consequences in the field of pulmonary medicine, allowing for more accurate diagnosis whilst reducing the cost of the required apparatus. If the cost of the MRI apparatus can be reduced due to a relaxation of the requirements of the strength of the magnetic fields and the sensitivity of the NMR probes, MRI machines might be more commonly found in places like doctors' surgeries rather than only in specialised hospital departments.

The main candidates for hyperpolarisation are noble gases, and in particular there has been strong interest in 3He, 83Kr and 129Xe, but other noble gases such as 21 Ne and 131 Xe may also be used. It has been found that these gases each have their own benefits and drawbacks in MRI. For example, 129Xe can diffuse into the tissue of the lungs, providing information about the diffusivity of this tissue. Further, 83Kr has a nuclear spin I = 9/2. The associated nuclear quadrupolar moment causes rapid depolarisation when the electron shells of the noble gas atoms are distorted by external interactions, such as short term adsorption of 83Kr atoms on surfaces. This enables the presence and properties of surfaces to be probed by the 83Kr relaxation observed in the surrounding gas phase, and is of particular interest for MRI of lungs and their large surface to volume ratios (S/V) and surface quadrupolar relaxation (SQUARE) maps which could be used in diagnosing emphysema. However, the rapid depolarisation of 83Kr is particular problematic for the transportation and storage of the hyperpolarised gas.

Other gases apart from noble gases may also be used, but not all of these would be safe for medical use. However, these may still be of interest in non-medical

fields.

Whilst hyperpolarisation can be achieved at atmospheric pressure, it has been found the hyperpolarisation process is more efficient at lower pressures. Typically, the pressure of the admixture of gases in the optical pumping cell is typically around 0.1 mbar to 100 mbar (0.01 kPa to 10 kPa) and preferably around 1 mbar (0.10 kPa). However, many useful applications of hyperpolarised gases in the field of NMR require the gas to be at approximately atmospheric pressure or above. For example, in the use of pulmonary diagnostics, the hyperpolarised gases will need to be inhaled by the patients. Therefore, it is required to be able to compress the hyperpolarised gas that is at around 1 mbar (0.10 kPa) in the optical pumping cell to a pressure of at least 1 atm (101 kPa).

However, one problem that has slowed the development of hyperpolarised gases 5 for NMR/MRI applications is the need to pump and compress these gases without causing significant depolarisation.

A hyperpolarised gas will eventually lose its polarisation, as the alignment of the nuclear spins is lost. The timescale over which this occurs is known as the -11 relaxation time. Even under perfect storage conditions, the collision of gas atoms will cause depolarisations by magnetic dipolar spin relaxation, which sets the upper limit of the Ti relaxation time. However, depolarisation usually occurs with significantly shorter Ti relaxation times in practice due to various other depolarisation factors. For example, surface interactions can reduce the Ti relaxation time, especially for permeable surfaces that can trap the polarised atoms. Surface coatings have been used to try to reduce this effect, and with suitable containers, hyperpolarised gases can be kept in suitable containers for many hours without significant depolarisation.

However, it has been found that when hyperpolarised gases are pumped using a conventional pump, the gases can become fully or partially depolarised very quickly, shortening the Ti relaxation time and reducing their effectiveness for use in MRI. This is thought to be due to increases in pressure causing greater number of collisions of the atoms of the hyperpolarised gas and increased number of collisions within the surfaces of the pump. Possible further effects could be caused by turbulence of the gas during the pumping process. The depolarisation is particular problematic for 83Kr, making this gas difficult to pump for MRI applications.

Therefore, it would be desirable to provide a pump for pumping hyperpolarised gases that reduced the amount of depolarisation that occurs during the pumping process, including for 83Kr.

Previous methods of pumping hyperpolarised gases have been employed. W02004/065974 Al discloses a pumping system for hyperpolarised gases employing a reversible fluid flow against a deflectable gas transport bladder. Inflation and deflation of the gas transport bladder is operably associated with valves for directing the flow of the hyperpolarized gas. In W02017042544 (Al), a pre-evacuated storage volume serves as a pneumatically operated, single piston recompression unit. However, this cannot compress the hyperpolarised gas, only transport it at the same or lower pressure. Pneumatically driven pistons and diaphragm pumps have also been considered.

The prior art pumps have several problems that the present invention is aimed at overcoming. The use of bladders and diaphragms requires flexible organic membranes. These can lead to outgassing of organic compounds that could contaminate the hyperpolarised gas. Further, the membranes can be ruptured after sustained use. Further, the membranes are not very heat resistant, and therefore not suitable for "baking" if this is required to purge the system of excess moisture. Further, the pneumatically driven pistons require a motor that must be located at a distance from the hyperpolarising apparatus in order that any magnetic fields do not interfere with the hyperpolarised gases. Since the pump seems to be a major cause of depolarisation, some prior art pumps have tried to reduce the amount of time the gas remains in the pump chamber, the "dwell time" of the pump. If the rate of depolarisation is high in the pumping system, then it is better to reduce the amount of time the gas remains in this depolarising environment.

Summary

In a first aspect of the invention, there is provided a dry pump arranged to pump hyperpolarised gas. The pump comprises a pump body and a compression member provided within the pump body and arranged to be movable in an axial direction. The pump body and compression member define a pump chamber. The pump further comprising a non-magnetic linear drive coupled to the compression member and arranged to move the compression member within the pump body to reduce the volume of the pump chamber on a compression stroke and increase the volume of the pump chamber on an intake stroke. The non-magnetic linear drive arranged to provide minimal lateral forces on the pump body. An inlet valve is coupled to an inlet actuator and the inlet actuator is arranged to selectively allow the hyperpolarised gas into the pump chamber through the inlet valve when the compression member is performing an intake stroke. An outlet valve is coupled to an outlet actuator and the outlet actuator is arranged to selectively allow the hyperpolarised gas out of the pump chamber through the outlet valve when the compression member is performing a compression stroke. The pump is made only of non-magnetic materials so as to avoid depolarisation of the hyperpolarised gas.

Preferably, the inlet and outlet actuators and the non-magnetic linear drive are controlled by a controller, for example to reduce depolarisation of the hyperpolarised bas during operation which may be based on various parameters. One such parameter may be the determined pressure difference across the outlet valve, and the controller may ensure that this pressure difference is kept within predefined limits by control of the actuators and linear drive.

For example, during the compression stroke, the outlet actuator may only open the outlet valve when the pressure on the inner side of the outlet valve reaches substantially the same pressure, or slightly higher pressure, than the outer side outlet valve. Further, the outlet actuator may be controlled to close the outlet valve only when the compression member reaches or nearly reaches its closest approach to the outlet valve.

This has the advantageous effect that the pressure in the pump chamber of the pump can be kept as low as possible whilst still compressing the gas. This avoids excessive pressure in the pump chamber that can lead to increased depolarisation.

The actuators and linear drives should be made from non-magnetic materials. Advantageously, these may be piezo-electric actuators and drives, as these have the benefit that can be very carefully controlled and also do not work on a magnetic principle like a motor.

The pump body and compression member is made of a non-magnetic materials, and preferably this could be at least one of aluminium, titanium, ceramic or plastic, and still more preferably, the inner surfaces of these parts are treated or coated to reduce depolarisation due to surface interactions.

Many of these features combine symbiotically to provide still further advantageous effects. The reduction of the pressure in the pump chamber reduces the pressure-dependent depolarisation. This would normally cause the outlet valve of a piston pump to be opened only slightly. However, the ability to independent control of the inlet and outlet valves by the inlet and outlet actuators allows the valves to be fully opened despite the low pressure, and consequently avoid any throttling effects that would force gas through very narrow apertures bringing the gas molecules near to depolarising surfaces.

The lower pumping rate, and reduced turbulence of the pump of the present invention may also provide for a much quieter pump. This could be particularly useful in a medical setting, as loud noises could be distressing for the patient, and make communication difficult between medical personal and the patient.

In a second aspect of the invention, there is provided an apparatus for preparing and pumping hyperpolarised gas. The apparatus comprises at least a source of polarisable gas, an optical pump cell and the above-described pump. The pump is arranged to pump the hyperpolarised gas out of the optical pump cell or any ancillary chambers such as a hydrogen combustion chamber In third aspect of the invention, there is provided a method of pumping hyperpolarised gases. The method comprises providing a pump having a pump body and a compression member within the pump body. The pump body and compression member define a pump chamber. The method further comprises using a non-magnetic linear drive coupled to the compression member to move the compression member in an axial direction within the pump body to reduce the volume of the pump chamber on a compression stroke and increase the volume of the pump chamber on an intake stroke. An inlet valve is opened with an inlet actuator to selectively allow the hyperpolarised gas into the pump chamber through the inlet valve when the compression member is performing an intake stroke An outlet valve is opened with an outlet actuator to selectively allow the hyperpolarised gas out of the pump chamber through the outlet valve when the compression member is performing a compression stroke. The pump is made only of non-magnetic materials so as to avoid depolarisation of the hyperpolarised gas.

The present invention provides an improved pump for pumping hyperpolarised gases, such as hyperpolarised noble gases. These gases have to be treated carefully, because they may be easily depolarised. The depolarisation can occur due to the increased pressure in the compression chamber of a piston pump and the turbulence of the gas as they are throttled through narrowly opened valves. The materials that the pump are made from can influence the rate of depolarisation of the gases as well as the surface of the internals of the pump.

The present invention according to the above-described aspects have the advantageous effect that the rate of depolarisation can be reduced. This reduction comes about by the combination of various factors that individually help to reduce depolarisation, but combine to provide an advantage greater than the zo sum of the individual effects.

Brief description of the drawings

Embodiments of the invention may be understood with reference to the following drawings: Fig. 1 shows the mechanism of spin exchange optical pumping to hyperpolarise gases.

Fig. 2 shows a cross-section of a prior art pump during a compression stroke.

Fig. 3 shows is a schematic diagram of the four stages of the pumping process during operation of compression, exhaust, backstroke and intake.

Fig. 4 shows a graphical representation of how the pressure in the pump chamber and the outflow of gases from the pump chamber vary at during the different stages of the prior art pump of fig. 3, as the outlet and inlet valves are opened and closed.

Fig. 5 shows a cross-sectional view of a dry pump according to embodiments of the present invention.

Fig. 6 shows a graphical representation of how the pressure in the pump chamber and the outflow of gases from the pump chamber vary at during the different stages in an embodiment of the present invention, as the outlet and inlet valves are opened and closed.

Fig. 7 shows an example of a system used to hyperpolarise gases including a pump according to the present invention.

Detailed description

The present invention provides a pump that is specifically designed for pumping hyperpolarised gases, such as hyperpolarised noble gases like 3He, 83Kr and 128Xe, although other polarisable gases could be used. The present invention is based on a piston pump design, with significant adaptions to be suitable for pumping and compressing hyperpolarised gases. The pump of the present invention has several design features that provide significant technical advantages over the prior art. To explain the technical advantages, an example of a prior art pump and its limitations will now first be discussed.

Figs. 2 and 3A-D shows the key features of a known gas pump 10 forming part of the prior art (which may also be referred to as a compressor). The pump 10 has a an inlet 40 and an outlet 50 connected by a cylinder 20 at which the inlet 40 is provided and that also has a cylinder head 60 (sometimes referred to as a front end, top end or head end) at which the outlet 50 is provided.

A piston 30 is arranged is provided within the cylinder 20 and is arranged to slide within the cylinder 20. The piston 30 is coupled to a coupling rod 85 which is in turn coupled to a crankshaft 80. Therefore, as the crankshaft 80 rotates, the coupling rod 85 causes the piston 30 to move in a reciprocating manner back and forth within the cylinder 20.

The outlet 50 provided in the cylinder head 60 includes an outlet valve which comprises a valve disc 52. The valve disc 52 may be pivotally connected to the cylinder head 60 with an outlet valve pivot 54 and a valve seat, the valve seat being provided by the cylinder head 60 of the cylinder 20. A biasing means 56 urges the valve disc 52 against the valve seat.

The valve disc 52 is displaceable between a seated condition in which the valve disc 52 abuts the valve seat forming a gas tight seal therewith, as shown in figs 2, 3A, 3C and 3D, and an open condition in which gas is allowed to pass between the valve seat and the disc 52 and out from the cylinder 20, as shown in fig. 3B.

Thus the outlet valve is biased against the valve seat so that it assumes a closed condition unless sufficient pressure is applied, when it will assume an open condition. The biasing means may, for example, be an outlet valve spring 56.

The cylinder 20 may be referred to more generally as a 'pump body'. An internal 20 volume 20A of the cylinder 20 into which gas is drawn may be referred as a 'pump chamber'.

An inlet 40 is arranged to allow gas to be drawn into the pump chamber 20A and, in the described example, an inlet valve is formed by an inlet aperture 40A formed in the body of the cylinder 20, such that the inlet valve 40 is blocked and hence closed by the side of the piston 30 except as the piston 30 is withdrawn away from the outlet valve, at which point the inlet aperture 40A is revealed, and the inlet valve 40is open.

Fig. 3 shows the four stages of the operation of the pump 10. As the piston 30 executes reciprocal movement within the cylinder 20, the piston 30 alternately moves from a top dead centre (TDC) position in which a front surface of the piston 30 is at its most forward position (at or near the cylinder head 60 of the cylinder 20, see fig. 3B) and a bottom dead centre (BDC) position in which the front surface is at its most rearward position, away from the cylinder head 60 and towards a rear end of the cylinder 20 (see fig. 3D).

Starting at the stage of the cycle with the pump chamber 20A already containing gas ready to be pumped as in fig. 3A, the piston 30 moves towards the TDC position. This movement being referred to as compression stroke of the piston 30. During this compression stroke, gas in the pump chamber 20A is compressed. If the gas reaches a sufficient pressure before the piston 30 reaches the TDC position, the valve disc 52 is displaced from its seated condition as shown in fig. 3B, allowing compressed gas in the cylinder 20 to be exhausted from the cylinder 20 around a periphery of the valve disc 52 in an exhaust stage.

Once the compressed gas has been exhausted, the piston 30 returns back towards the BDC position in the backstroke stage shown in fig. 3C. In this stage, the inlet valve and the outlet valve are both closed, and the pressure in the pump chamber 20A is reduced as the volume of the chamber 20A expands.

With the piston 30 at or near to the BDC position as shown in fig. 3D, the inlet aperture 40A formed in a wall of the cylinder 20 is exposed to internal volume of the pump chamber 20A between the piston 30 and the valve disc 52, allowing gas that is to be pumped to flow into the pump chamber 20A through the inlet aperture 40A.

As this point, the crankshaft 80 continues to rotate, the piston 30 starts to move back towards the TDC position and the cycle repeats over again.

Fig. 4 shows a graph of how the internal pressure inside the pump chamber of the pump 10 of figs. 3 and 4 varies during the pumping cycle, as well as the outflow of gas from the pump chamber 20A, at the various stages identified in figs 4A-D. The upper section indicates whether the states of the inlet and outlet valves are open or closed during each stage. As can be seen, in the compression stage indicated in fig. 4A, both the inlet and outlet valves are closed, and the piston 30 move towards the outlet valve. The pressure inside the pump chamber 20A increases above inlet pressures as the piston 30 moves towards the outlet valve. Because of the way in which the prior art pump works, the outlet valve will not open until the pressure inside the pump chamber 20A is greater than the pressure outside the outlet valve. This is because the valve disc 52 is biased towards the closed position, and so the pressure in the pump chamber 20A must overcome both the outlet pressure and the biasing 56. Therefore, as seen in fig. 4, there is a noticeable "overshoot" of the pressure before the outlet valve opens. This is particularly troubling for pumping hyperpolarised gas as the rate of depolarisation increases with increasing pressure due to interatomic interactions between the molecules of the gas and also with increased surface collisions with the inside surfaces of the piston 30.

At the point that the outlet valve opens, the pressure in the pump chamber 20A is significantly higher than the outlet pressure, and there is a large initial flow of gas from the pump chamber 20A through the outlet valve as the pressures equalise.

Because the pressure of the gas itself is what opens the outlet valve, the outlet valve might not open to a very large degree, and the gas may be forced through a very small gap, causing a throttling effect and producing great turbulence. The gas will continue to flow as the piston 30 moves towards the outlet valve until the piston 30 reaches TDC in the exhaust stage, shown in fig. 3B.

The minimum volume to the pump chamber 20A is known as the dead space, and a large dead space reduces the efficiency of the pump 10. At TOO the gas in the dead space is highly compressed and the ratio of the surface that contains the gas and the volume is a maximum. This has an effect on hyperpolarised gases, since the depolarisation increases with increasing surface to volume ratio.

Once the piston 30 reaches TOO, the piston 30 is withdrawn away from the outlet valve, which will be closed by the biasing means. At this point both the inlet and outlet valves are closed, and the pressure in the pump chamber 20A will start to decrease as no gas will enter or leave chamber 20A in the backstroke stage shown in fig. 30.

As the piston 30 moves towards BDC, the inlet aperture 40A is revealed and the inlet valve becomes open. At this point, the pressure in the pump chamber 20A will be very low, and the gas can flow through the inlet 40 and into the pump chamber 20A until the pressure stabilises at the inlet pressure. This is the intake stage shown in fig. 3D. The flow of gas in this stage shown in the graph in fig. 4 as a negative outflow, as the gas is flowing into the chamber 20A. Once the pump chamber 20A is full of gas at the inlet pressure, no more gas will enter the pump chamber 20A, and the piston 30 continues to move up again towards the outlet valve. The inlet valve will be closed, and the cycle will restart with the compression stage again.

Prior art pumps 10, such as that described above in respect of figs. 2, 3 and 4, have several drawbacks that are particularly relevant to pumping hyperpolarised gases that will now be described.

As shown above, with conventional piston pumps 10, there is a significant overshoot of the pressure inside the pump chamber 20A required to open the outlet valve. This overshoot is apparent from the graph in fig. 4 as the sharp peaks in pressure just before the outlet valve opens at the start of the exhaust stage. Once the valve is open, then the gas in the pump chamber 20A is significantly higher than the outlet pressure, and so the gas will rush through the valve at high speed, producing lots of turbulence. With hyperpolarised gas excessive high pressures and turbulence are to be avoided, because they may lead to depolarisation of the gas. The turbulence and high pressures both could lead to an increased number of collisions between the hyperpolarised gas molecules, which reduces the Ti time of the gas. That is, the frequent collisions between the atoms/molecules of the gas and the turbulent nature of the flow of the gas both can reduce the level of polarisation.

To address the issue of spikes in pressure, large changes in outflow rates and subsequent turbulence in the prior art pumps 10, the present invention is designed to eliminate or at least drastically reduce these sudden changes of pressure in the pump chamber 20A.

Fig. 5 shows a cross-sectional diagram of an embodiment of the present invention. The pump 500 will be described in detail later, but the basic features and principles of the pump 500 will be discussed first. In this embodiment there is provided a cylinder 520, acting as a pump body 520, and a compression member 530, such as a piston, similar to the above-described prior art pump 10. However, an embodiment of the present invention comprises a cylinder head 560 provided at an end of the pump body 520 opposite the compression member 530 to define the pump chamber 520A, which is the volume contained within the pump 500 for collecting, compressing and exhausting a gas. The cylinder head 560 is provided with an integrated inlet valve 540 and an outlet valve 550. The inlet valve 540 is in communication with an electronically controlled inlet actuator 546 and the outlet valve 550 is in communication with an electronically controlled outlet actuator 556. The inlet actuator 546 and outlet actuator 556 are controlled electronically, for example by a control unit, and can be actuated independently of each other.

Preferably, the inlet valve 540 and outlet valve 550 open away from the pump chamber 520A, so that no part of the valves 540, 550 enter the volume swept by the compression member 530. This ensures that the top of the compression member 530 can be moved as close as possible to the cylinder head 560 and reduce the amount of dead space.

In operation, the compression member 530 moves back and forth in a direction towards and away from the cylinder head 560, in a manner similar to the prior art pump 10 describe previously. However, the movement of the compression member 530 is not driven by a crankshaft, as with the prior art, but a linear drive 580 connecting to the compression member 530 via a connecting rod 585, which will be described in more detail later.

The inlet valve 546 and outlet actuator 556 are activated to open and close the inlet valve 540 and outlet valve 550, respectively, to optimise the gas flow from the inlet to the outlet. Because the inlet valve 546 and outlet valve 556 do not depend on them being forced open by the pressure of the gas in the pump chamber 520A, they can be opened fully and independently of the difference of pressure between the inside and outside of the pump chamber 520A, and turbulence can also be avoided.

Fig. 6 shows a similar graph of the pressure in the pump chamber 520A and the outflow out of the pump chamber 520A for an embodiment of the present invention. For comparison with fig. 4, the dotted line reproduces the pressure profile of the prior art pump 10 drawn to the same scale.

In an embodiment, the outlet actuator 556 remains closed during the compression stage of each pumping cycle until the point where the pressure within the pump chamber 520A matches the pressure of the gas on the side of the outlet valve 550 outside the pump chamber 520A. At this point, or very shortly afterwards, the outlet actuator 556 opens the outlet valve 550. The pressure of gas within the pump chamber 520A matches the pressure of the outlet, or exceeds it by a small amount, and so there is no rush of gas from the pump chamber 520A through the outlet valve 550, as described above in respect of the prior art. Importantly, the maximum pressure inside the pump chamber 520A is lower than in a conventional piston pump 10, and so the pressure-driven depolarisation is reduced.

The compression member 530 continues to compress the gas within the pump chamber 520A towards the outlet valve 550, and with the outlet valve 550 now being fully open, the gas is forced through the outlet valve 550 into the outlet. The flow rate of the gas from the pump chamber 520A through the outlet valve 550 has a peak value much lower than the prior art, and remains at a steady level until the compression member 530 approaches the cylinder head 560 and the outlet actuator 556 closes the outlet valve 550.

It would be advantageous for the compression member 530 to reach TDC position close to the cylinder head 560, so that the amount of dead space is reduced to a minimum. The dead space is a measure of the minimum volume of the pump chamber 520A, which occurs when the compression member 530 reaches TDC. To optimise the flow of gas through the pump 500, it would be advantageous if the outlet valve 550 is kept open until the compression member 530 reaches TDC, and then closed at, or very close to, this point to minimise the gas that is compressed in any dead space. In this manner, it is important that the outlet valve 550 does not intrude into the pump chamber 520A.

Advantageously, the shape of the cylinder head 560 and the compression member 530 are arranged to minimise the dead space. The shape of the cylinder head 560 and the compression member 530 could be arranged to be complementary to each other. That is, they could both have planar surfaces having the same shape, or they could have complementary convex and concave shapes, such that there is minimal space volume created when the compression member 530 and the cylinder head 560 are just touching.

On the backstroke stage of the pumping cycle, the pressure in the pump chamber 520A will be reduced in the same manner as the prior art. However, unlike the prior art, rather than the inlet valve being opened at the point where the compression member 30 pulls back far enough to reveal the inlet aperture 40A, the inlet valve 540 is opened by the inlet actuator 546 at an optimised point in the backstroke. Therefore, the sudden inrush of hyperpolarised gas when the inlet aperture of the prior art is revealed can be avoided, which might otherwise lead to depolarisation of the gas Optionally, the inlet actuator 546 could delay opening of the inlet valve 540 until the point where the pressure inside the pump chamber 520A matches, or nearly matches, the pressure of the inlet gas, such that there is no rush of gas through the inlet valve 540 when the inlet valve 540 is first opened. This would be most advantageous when there is a non-negligible dead space. This is because the dead space is the volume of gas that remains when the outlet valve 550 is closed at TDC of the pump cycle. Therefore, this gas would be at the higher outlet pressure, and so if the inlet valve 540 was opened, then the higher pressure of the gas in the pump chamber 520A could cause gas to be forced back through the inlet valve 540, against the lower pressure inlet gas. Therefore, waiting until these pressures equalised would avoid any quick changes of pressure that could occur if the inlet pressure and the pressure of the pump chamber 520A were not approximately equal.

Alternatively, the inlet valve 540 could be opened just as, or shortly after, the point at which the outlet valve 550 is closed, and optionally closed just as the compression member 530 reaches BDC. This would provide the longest intake stroke possible, and so the lowest flow rate for any given volume of gas being pumped on each cycle. Therefore, this could also provide the smoothest (least turbulent) intake of gas if the dead space is negligible, such that there is very little gas in the pump chamber 520A at TDC.

As shown in fig. 6, the pressure profile of pump chamber 520A for embodiments of the present invention reduces spikes of pressure associated with the prior art pump (dotted line), with associated sudden increases and decreases that could cause turbulence in the hyperpolarised gas, and lead to depolarisation. Fig. 6 clearly shows how the timings of the opening and closing of the inlet valve 540 and the outlet valve 550 correspond with the pressure within the pump chamber is 520A.

The operation of the pump 500 is generally to draw gas from the inlet at the lower inlet gas pressure, compress the gas in the pump chamber 20A until it reaches the higher outlet pressure, then push the gas out of the outlet valve 550 at substantially the outlet pressure or slightly above the outlet pressure. An advantage of these embodiments of the present invention is that the pressure difference between the pump chamber 520A and the outlet pressure is kept low and approximately constant during the entirety of the exhaust stage. This is possible because of the use of an electronically controlled actuator to selectively open and close the outlet valve 550 at specific points in the pumping cycle, in contrast to the prior art in which an overpressure in the pump chamber 520A is required to open the outlet valve 550.

In an embodiment of the invention, a control unit (not shown) is provided for controlling the inlet actuator 546 and outlet actuator 556. The control unit determines the optimum times for activating the actuators for opening and closing the inlet valve 540 and outlet valve 550. It is noted, that if the pump 500 is used to pump the hyperpolarised gas into a rigid storage container, or the like, the pressure on the outside of the outlet valve 550 will increase during the pumping process, as the pressure in the container increases as more gas is stored. Therefore, the pressure inside the pump chamber 520A at which the outlet valve 550 should be opened would start off relatively low, and increase as the outlet pressure, that is the pressure inside the storage container, increases.

In this case, the storage container could initially be evacuated. That is the pressure in the storage container could be considerably lower than the inlet pressure of the hyperpolarised gas. Therefore, there is no requirement to compress the hyperpolarised gas in the initial stages, whilst the pressure in the storage container is lower than the inlet pressure. Therefore, the pump does not need to compress the gas entering the pump chamber 520A in the initial stages.

In one embodiment, the inlet valve 540 and outlet valve 550 could both be opened at the same time, so that the hyperpolarised gas can flow through the pump 500 and into the storage container until the pressure in the storage container matches the inlet pressure. At this point, the above described compression cycles can be commenced.

Alternatively, to avoid the sudden rush of hyperpolarised gas from the inlet straight through the pump 500 and into the storage container, the inlet valve 540 could be opened whilst the outlet valve 550 is closed, so that the hyperpolarised gas flows into the pump chamber 520A. The inlet valve 540 can then be closed and the outlet valve 550 opened to allow the gas transferred into the pump chamber 520A to flow into the lower pressure storage container in a more controlled manner.

Still another alternative is to use the pump 500 in reverse to ensure that the flow of gas between the inlet and the storage container does not cause any sudden rush of gas between different pressure zones. In this manner, the inlet valve 540 could be opened when the compression member is near to the TDC, and so the gas can flow into the pump chamber 520A at its smallest volume. The compression member 530 can then be retracted to withdraw some of the gas from the inlet valve 540. The inlet valve 540 can then be closed at some point where the compression member is between the TDC and BDC position. The continued withdrawal of the compression member expands the pump chamber 520A, and with both valves 540, 550 closed, the pressure of the gas will be reduced. The outlet valve 550 can then be opened when the compression member 530 is at, or near to, the TDC position so that the gas that is now has a pressure lower than the inlet gas pressure can be transferred to the evacuated storage container. The reduction in pressure of the gas in the pump chamber 520A prior to being allowed into the storage container reduces the pressure difference through which the gas flows during the initial stages of pumping.

Alternatively, the pump 500 may pump into a flexible storage container, such as Tedlar bag, or other suitable bag, or could be used to continuously pump, for example for inhalation by a medical patient. In this case, the outlet pressure would likely be substantially constant, and the pressure inside the pump chamber 20A at which the outlet valve 550 should be opened would be substantially the same during operation.

The pump 500 is capable of pumping a hyperpolarised gas through a range of different inlet and outlet pressures. However, of particular interest is the pumping of hyperpolarised gas generated from an optical pumping cell into either a storage container, or direct to point of use via any necessary intermediary apparatuses, such as separation chambers or combustion chambers. Under these conditions, typical inlet pressures are 0.1 to 100 mbar (0.01 kPa to 10.0 kPa), or even as low as 0.1 to 1.0 mbar (0.01 kPa to 0.10 kPa) and typical outlet pressures are about 1 standard atmosphere (101 kPa), or at least 0.5 to 2 atm (50.5 kPa to 202 kPa). Therefore, it is important that the compression ratio of the pump 500 is high enough to provide pumping through this range of gas pressures. Ideally, the compression ratio of the pump 500 should be over 100, and preferably at least about 1000.

In embodiments of the invention, sensors may be used to measure various pressures in the pump 500. Sensors (not shown) may measure one or more of the pressures in the pump chamber 520A, the inlet and the outlet. The inlet and outlet pressures could be measured on the side of the valves outside of the pump chamber 520A.

Various forms of sensors could be employed to measure these pressures. Such as resistive, capacitive, piezoelectric, optical or MEMS pressure sensors Alternatively, these pressures could be inferred from other measurements. For example, the inlet and outlet pressures may already be known prior to the pumping process. Therefore, the position of the compression member 530 within the pump chamber 520A at which the gas would be at the suitable pressure could be predetermined. Therefore, the outlet and/or valves could be opened by their respective actuators 556, 546 when the compression member reaches predetermined points in its cycle. This could itself be measured by position sensors, or could be set as part of the controlling process of the linear actuator coupled to and driving the compression member 530.

A further consideration is that hyperpolarised gases are very sensitive to stray magnetic fields that could deform the polarising field, and in particular any varying magnetic fields that could depolarise the gas. Therefore, any magnetic materials should be avoided in the production of the pump 500. Therefore, all materials that make up the pump 500 should have non-magnetic properties, or at least very weak magnetic properties. Ferromagnetic material should generally be avoided, such as steel. Although, non-magnetic stainless steel may be suitable. Of particular interest for one or more of the pump body 520, compression member 530, cylinder head 560, inlet valve 540 and outlet valve 550 are titanium and aluminium, and various alloys thereof. Other potential candidates could include ceramics and plastics.

Because of the need to avoid magnetic fields, magnetic based actuators are not suitable for the inlet and outlet actuators 546, 556. Therefore, non-magnetic actuators should be used.

Of particular interest is the use of piezo-electrical actuators. These actuators can be used to provide quick switching to open and close the inlet and outlet valves 540, 550 based on just an electrical signal, for example from the control unit. No magnetic fields are required to operate the actuators 546, 556 and they require minimal current that could produce any electromagnetic effect. Also, the materials of the actuators 546, 556 can be made from non-magnetic materials. The piezoelectrical elements tend to be made from non-organic materials, and so have the further advantage that they do not tend to outgas significantly, and so they should not contaminate the hyperpolarised gas.

Alternative non-magnetic actuators could be used instead of piezo-electric actuators. Non-limiting examples of non-magnetic actuators are pneumatic actuators and hydraulic actuators. The motors used to power these actuators could be located sufficiently far from the hyperpolarised gas that any danger of depolarisation would be avoided.

A further concern with pumping hyperpolarised gas is that lubricants, such as oils or greases, should be avoided. These could cause contaminants which could lead to depolarisation of the gas. In a medical environment, such as the use of hyperpolarised gas for MRI, contaminants must clearly be avoided.

Now, with the prior art piston pumps such as that described above, the piston 30 is usually driven by a rotating crankshaft 80 via a coupling rod 85. The coupling zo rod 85 acts to convert the rotational motion of the crankshaft 80 to a linear motion of the piston 30. The forces are sometimes known as "cosine forces" as the magnitude of the lateral force is the cosine of the angle between the coupling rod and the plane of the piston head, marked as a in fig. 2. Therefore, even with a very long coupling rod 85 this angle will never be 90 degrees, and there will always be a component of the driving force that is perpendicular to the direction of motion of the piston 30, which will cause the piston 30 to impart a sinusoidal sideways force on the wall of the pump body 20.

Since, as discussed above, lubricants are to be avoided in the pumping system for pumping hyperpolarised gases, then these cosine forces can cause excessive wear to the pump. The wear will cause contaminants or particulates which could also reduce the polarisation of the polarisation, could pose a health risk in a medical application and also reduces the lifetime of the pumping system.

To avoid the cosine forces, embodiments of the present invention use a linear actuator 580 to move the compression member 530 in a linear motion along an axial direction of the pump body 520, such that there is no, or minimum, forces imparted on the sidewalls of the pump body 520.

As with the valve actuators, there is a need to avoid magnetic fields, magnetic drives are not suitable to drive the compression member 530. Therefore, nonmagnetic drives should be used.

Of particular interest is the use of a piezo-electrical linear drives. One of these linear drives could be used to provide the back and forth action required to drive the compression member 530 to expand and contract the pump chamber 520A during the pumping process. No magnetic fields are required to operate the linear drive 580, and the materials of the linear drive can be made from non-magnetic materials. The piezo-electrical linear drives are not usually as powerful as other forms of driving mechanisms for reciprocating the compression member 530, such as a motor-driven camshaft. However, due the fact that the valves are also electronically actuated, this should be less of a problem, as the pump 500 only needs to develop enough pressure to match the atmospheric pressure of the outlet.

Advantageously, a piezo-electric linear drive 580 can be controlled to a very fine level such that the motion and position of the compression member 530 can be optimised for the effective pumping of the hyperpolarised gas. The acceleration and deceleration of the compression member 530 can be optimised for pumping the hyperpolarised gases without leading to significant depolarisation and the position of the compression member 530 at TDC can be made close to the cylinder head 560 so as to minimise or eliminate any dead space.

In contrast, the piston motion of a crank-driven pump 10 cannot be changed, as it is preset by the coupling 85 from the piston 30 to the crankshaft 80. The only control over the compression member's 30 position is the speed at which it can be driven. However, a piezo-electric drive has complete freedom over the position of the compression member 530 and speed at which it is driven. These can be arranged to optimise the pumping process, and may even be changed based on measured parameter such as the pressures across the valve heads 542, 552. That is, the speed and position of the compression member 530 could be varied during operation to optimise the efficiency of the pump 500. For example, as the pressure inside a storage container increased, the pump 500 could adjust the speed of the compression member 530 to account for the increased outlet pressure.

In prior art pumps 10, the depolarising nature of the pump 10 encourages the use of fast pump 500 speeds to reduce the dwell time that the gas molecules spend in the pump. However, with the pump 500 of the present invention, every effort is made to reduce the depolarisation of the gas in the pump 500. Therefore, longer dwell times are possible, and the pump 500 can compress the gas at a slower rate without causing significant depolarisation. The piezo-electric drive 580 can drive the compression member 530 at moderately slow speeds, for example a typical piezo-drive can drive the compression member 530 at about 1-5 cms-1.

The pump 500 could be made in various sizes to suit different situations.

However, typically the pump 500 may have a pump chamber 520A that has a volume that is between 10 and 1000 cm3, and more particularly between about 100 and 500 cm3, although different sized pumps are possible. The pump 500 is particular suitable as a compact pump, with the valve actuators 546, 556 and valves 540, 560 being integral with the pump body 520. The non-magnetic linear drive 580 could also be integrated with the pump body 520 to provide a complete pump requiring only gas inlet and outlet pipes, and control signals to the actuators 540, 550 and drive 580.

As an alternative other non-magnetic linear drives could be used instead of piezo-electric linear drive. Examples of non-magnetic linear drives are pneumatic linear drives and hydraulic linear drives. The motors used to power these types of linear drive could be located sufficiently far from the hyperpolarised gas that any danger of depolarisation would be avoided.

Now that the main features of the invention have been discussed and their importance explained, a specific example of a pump 500 according the present invention is described in further detail. The description is not meant to be limiting on the scope of the invention, and is provided as an example of how the invention could be worked. Various features of the foregoing description can be replaced or removed without departing from the invention, which is defined by the appended claims.

As shown in fig. 5, the pump 500 has a pump body (cylinder) 520 defining a pump chamber 520A within which a compression member (piston) 530 is reciprocally slidable. A connecting rod 585 is coupled to the compression member 530 and arranged to be translated back and for the motion along a longitudinal axis of the cylinder 520 by a piezo-actuated linear drive 580.

At a front end of the cylinder 520 is provided a cylinder head 560 coupled to a periphery of the cylinder 520 to seal the cylinder 520 from an external environment.

The cylinder head 560 is provided with an inlet valve 540 and an outlet valve 550.

The valves 540, 550 each have a valve stem 544, 554 coupled to a valve actuator 546, 556 arranged to open and close the valves 540, 550. Piezoactuated linear actuators are arranged to translate the respective stems 544, 554 of the valves 540, 550 along a direction parallel to the longitudinal axis of the cylinder 520. This allows each of the valves 540, 550 to be moved between an open and closed state. In fig. 5, the valves 540, 550 are arranged so that they do not protrude into the pump chamber 520A. This reduces the dead space of the pump, and avoids the chance of the piston 520 hitting and damaging the valves 540, 550.

The cylinder head 560 has an inlet aperture in gas communication with the inlet valve 540 and arranged to receive a gas to be pumped by the pump 500. Gas flowing through the head block inlet aperture enters an inlet cavity 540A within the cylinder head 560. When the inlet valve 540 is in the open condition gas may be drawn into the pump chamber 520A through an inlet valve 540 aperture.

The cylinder head 560 also has an outlet aperture in gas communication with the outlet valve 550. When the outlet valve 550 is in the open position gas is able to flow out of the pump chamber 520A through an outlet valve 550 aperture into an outlet cavity 550A within the cylinder head 560. The gas is then able to exit the cylinder head 560 through the head block outlet aperture.

The inlet valve linear actuator 546 is arranged to urge the inlet valve 540 against an inlet valve seat 542A formed in a base portion of the cylinder head 560 facing the cylinder 520 to put the inlet valve 540 into the closed state. In the embodiment shown, the inlet valve head 542 of the inlet valve 540 tapers towards its end and the inlet valve seat 542A has a corresponding complementary shape to form a snug fit between the inlet valve head 542 and inlet valve seat 542A when in the closed condition.

The outlet valve 550 has a corresponding outlet valve head 552 coupled to an outlet valve stem 554. The outlet valve head 552 is operable to be urged against an outlet valve seat 552A formed in the base of the cylinder head 560 adjacent the inlet valve seat 542A allowing the outlet valve 550 to assume a closed zo condition.

It is to be understood that because the inlet and outlet valves 540, 550 are each operable to assume open or closed conditions by means of their respective piezo-actuated linear drives 546, 556, the valves 540, 550 may be opened and closed independently of a difference in pressure of gas across the respective valves 540, 550.

This feature has the advantage that the valves 540, 550 may be opened and closed at different times and for longer or shorter periods of time than might otherwise be possible in order to improve performance of the pump 500.

Thus it is to be understood that when the piston 520 is at or near its position of closest approach to the cylinder head 560 (a position that is referred to as 'top dead centre" or TOO by analogy to crank-driven pumps), the outlet valve 550 may be closed. The precise position of the piston 520 at which the outlet valve 550 is closed is controlled in order to optimise the amount of gas expelled before the inlet valve 540 as described earlier in the description.

Furthermore, because the inlet and outlet valves 540, 550 are actuated by respective piezo-actuated linear drives 546, 556, the force with which each valve head 542, 552 is urged against the corresponding valve seat 542A, 552A may also be controlled independently of the pressure difference across the respective valves 540, 550.

Known pumps 10, like the one shown in fig. 3 and discussed above, have inlet and outlet valves that are biased in the closed condition by means of a spring element 56 and/or by a difference in pressure across the outlet valve 550.

Embodiments of the present invention are advantageous over arrangements in which spring elements 56 are employed because spring elements 56 can deteriorate in their effectiveness over time and suffer wear due to lack of lubrication. Furthermore, in order to open the outlet valve a pressure difference across the outlet valve must be sufficient to overcome the closure force of the spring 56. In contrast, embodiments of the present invention allow the inlet and outlet valves 540, 550 to be opened independently of the pressure difference and without a requirement for spring elements 56.

Spring elements 56 suffer the further disadvantage that heat is generated due to flexure of the spring elements 56 and sliding contact between the spring elements 56 and other surfaces with which they are in contact.

It is to be understood that in this embodiment, because the inlet valve 540 is actuated by an actuator 546 it is capable of opening by movement of the valve head 542 against the direction of flow of gas into the cylinder 520 when it is required to allow gas to pass into the cylinder 520. This is in contrast to conventional sprung valves in which displacement of the valve head occurs in the direction of gas flow due to a difference in pressure between gases on opposite sides of the head. Thus the inlet valve head 542 of the present invention does not intrude into the cylinder 520 thereby increasing a dead space between the piston 520 and head 560 when the piston 520 is at TDC.

A further advantage of employing piezo-actuators to open and close the valves 5 540, 550 is that the valves 540, 550 may be arranged to have high opening and closing forces even when no power is applied to the respective 546, 556.

Thus, the valve head 542, 552 may be moved to abut the valve seat 542A, 552A and to be urged against the valve seat 542A, 552A with a relatively high force. If power to the actuator 546, 556 is then terminated the valve head 542, 552 will remain urged against the seat 542A, 552A. This allows a high integrity seal to be formed between the head 542, 552 and seat 542A, 552A to prevent leakage.

Conversely, the valve heads 542, 552 may be translated away from the seats 542A, 552A with a relatively high force if required.

Furthermore, it is to be understood that inlet and outlet valves 540, 550 actuated by means of piezo-actuators are fabricated from non-magnetic materials and materials that do not outgas (or which outgas by a relatively insignificant amount).

Embodiments of the invention employing piezo-actuated linear drives have the advantage that relatively high opening and closing rates may be obtained allowing high opening and closing repetition rates.

In some arrangements, a control unit is arranged to control opening and closing of the inlet and outlet valves 540, 550 based on a signal responsive to a pressure in at least one of the pump chamber 520A, the inlet chamber 540A and the outlet chamber 550A. This signal may be derived from a pressure sensor or gauge coupled to the control unit.

Alternatively, a position of the piston 520 may be determined from a drive signal of the piezo-electric linear drive 580 driving the piston 520 or from a measurement of a position of the piston 520, for example by means of a sensor such as an optical sensor.

In some arrangements an electromagnetic, acoustic or other signal is employed to determine a position of the piston 520, for example by means of a Doppler shift between a transmitted signal and a received signal. The measurement may be made at a location external to the environment of the pump chamber 520A or within the same atmosphere as the pump chamber 520A.

It is to be understood that in the embodiment of FIG. 5 that the cylinder head 560 and actuators 546, 556 for the inlet and outlet valves 540, 550 respectively may be provided in a vacuum-tight package which will be referred to as a valve drive module. An electrical feed-through element (not shown) is provided to allow electrical power to be provided to the actuators 546, 556 within the sealed valve drive module.

By vacuum-tight package what is meant that the package is capable of preventing air from leaking from an outside of the package to an interior of the package where the actuators 546, 556are located. In some arrangements the vacuum-tight package is a helium gas leak-proof package, i.e. helium gas present on an external surface of the package is not able to leak into the package. The back ends of the compression member 530 and the inlet and outlet actuators 546, 556 could preferably be differentially pumped to remove any gas leaked out of the pump chamber 520A.

As already noted, it is very important for the materials to be used in the entirety of the pump 500 is designed to minimise the depolarisation of the hyperpolarised gas. The effect of excess pressures, and rapid changes of pressure have already been discussed above, as well as the need for using non-magnetic materials for the pump body 520, piston 530, inlet valve 540 and outlet valve 550. However, further features of the pump 500 may be required for optimised performance that have not yet been described.

There may be various seals required in the pump 500. For example, a seal 590 may be required around the periphery of the piston 530 to ensure a gas tight seal. These seals 590 will also need to be made of a non-magnetic material.

Other considerations are that the materials do not outgas, as any outgassing could lead to contaminants that could lead to increased depolarisation and also could pose a health risk if the gas is used in a medical setting. Examples of the materials suitable for making seals include ceramics, polyolefins and fluoroelastomers (such as Viton).

Preferably, a double seal 590 is used around the compression member 530, to seal the gap between the side walls of the compression member 530 and the pump body 520.

The internal parts of the pump 500 will come into contact with the hyperpolarised gas, and therefore the properties of these surfaces could play an important part in depolarisation of the hyperpolarised gas. Various mechanisms have been described as being responsible for reduction of the Ti relaxation time by interactions with the surface. For example, gas permeable surfaces could trap polarised gas atoms, and lead to depolarisation. Therefore, careful consideration should be given to the surfaces of the pump 500 that come into contact with the hyperpolarised gas.

Various coating could be used to reduce the permeability to avoid reduction of the Ti relaxation time. For example, aluminium alloy, titanium, PTFE, polymer or diamond-like carbon (DLC) coatings. Polymer coatings could be further modified by at least one of halogenation, sulfonation and carboxylation to reduce its permeability to the hyperpolarized gas or to reduce depolarisation of the hyperpolarised gas.

It has been noted that a mechanism for depolarisation of hyperpolarised noble gases is the dipolar interactions between the noble gas atoms and protons in the surface coating. Therefore, the depolarising can be reduced by replacing hydrogen atoms in the coating with deuteron. The deuterated coating have a smaller dipole interaction with the noble gas atoms and reduce the depolarisation. Therefore, in an embodiment of the invention, the inner surfaces of the pumps are treated with a deuterated coating.

Rather than a specific surface coating, the surfaces could be treated to adapt the texture to make them have less of a depolarising effect on the hyperpolarised gas. The surface of the compression member 530 and pump body 520 could be made to be ultra-smooth, for example by machining and polishing or by depositing a metallic layer, such as titanium, aluminium or gold, as a smoothing layer.

Gold has a further advantage that it acts as a dry lubricant. Therefore, wear of the moving parts of the pump 500 from friction can be reduced, for example between 10 the compression member 530 and the pump body 520.

During the preparation of hyperpolarised gases, it is common to bake the apparatus at a high temperature to drive away any moisture or contaminants that may be absorbed on the various surfaces, whilst the internal parts of the apparatus are evacuated. Therefore, a further consideration of the materials used to manufacture the pump 500 is that they can withstand high temperatures, typically up to about 250 degrees Celsius, and preferably up to 400 degrees Celsius.

zo The whole of the pump may be designed to UHV standards.

As the pump 500 may be required pumping of heated gases, since the optical pumping cell is often heated, the pump may be fitted with a cooling system, for example a water cooling system.

In some embodiments, the pump is required to compress hyperpolarised gases from a sub-atmospheric pressure to atmospheric pressure. For example from about 1 mbar (0.10 kPa) to about 1 atm (101 kPa). This would require a compression ratio of over 1000. If a single stage pump was not able to reach such a compression ratio, then multiples stages of the pump could be cascaded to increase the total pressure achievable by the pumping system. For example, two or more pumps, as described above, could be connected in series to provide increased overall compression ratio.

As already explained, the pump 500 of the present invention is directed towards the pumping of hyperpolarised gases. An important application of the use of hyperpolarised gases is in the use of NMR technology and in particular its application in medical MRI. As already discussed in the background section, the use of hyperpolarised gases has shown interesting results in improving the resolution or signal-to-noise ratio during in-vivo MRI examination of the lungs. When examining the lungs with MRI, it is advantageous if the patient inhales hyperpolarised noble gas. This gas must be delivered at approximately atmospheric pressure. Since the Ti time of some of the hyperpolarised gases can be quite low, it would be advantageous to be able to produce and store the hyperpolarised gas at the point of use, that is, in a doctor's surgery or in a hospital department. Alternatively, the hyperpolarised gas might not be stored at all and instead be pumped directly to the point of use. Therefore, a compact and cost effective pumping system that has low depolarisation effect on hyperpolarised gases, including noble gases, would be very desirable. The pump 500 of the present invention is particularly suited for such an application, and was designed with this purpose in mind.

Therefore, according to the present invention, there is disclosed an apparatus for preparing hyperpolarised gas, either for immediate use or for storing for later use.

An example apparatus 700 for hyperpolarising a gas is now described, with reference to fig. 7. The apparatus 700 comprises a source of polarisable gas 730, for example a noble gas such as 3He,Ne 12 ^1-, 83Kr and 128Xe or 131Xe. There may also be provided other gases to facilitate hyperpolarising the gas, for example alkali metals such as rubidium or caesium using the SEOP method described in the background section. Buffer gases may also be provided, such as nitrogen or hydrogen to quench the radiative de-excitation of the alkali metals, and dilute the alkali metals to avoid interactions. Other gases may also be supplied.

The apparatus also comprises a hyperpolarisation chamber 710, in which the gas is to be hyperpolarised. The hyperpolarisation chamber 710 is arranged to be in a magnetic field. This may be achieved, as an example, by arranging the chamber 710 in the centre of an electromagnetic coil configuration, such as Helmholtz Coils (not shown).

In a preferred embodiment, there is provided excitation lasers 720, arranged to pass through the hyperpolarisation chamber 710, in order to optically pump the alkali metal and initiate the SEOP process. Preferably, the lasers are high power diode array lasers so that the apparatus 700 can remain compact.

The hyperpolarised gas may be either pumped out of the hyperpolarisation chamber 710 in batches, or it may be pumped continuously. Either way, a pump 500 is needed to pump the gas through the associated tubing and to compress it to a suitable pressure for storing or administering to a patient. The pressure of the gas in the hyperpolarisation chamber 710 may be as high as atmospheric pressure. However, particularly efficient optical polarisations has been shown to occur at lower pressures, such as in the range 0.1 mbar to 100 mbar (0.01kPa to 10.1 kPa), since dipolar spin relaxation can cause a reduction in the Ti relaxation time due to magnetic dipole-dipole interactions, which increases with pressure. Therefore, the pump 500 will be required to pump the hyperpolarised gases from this range of pressures up to about atmospheric pressure without causing significant depolarisation. The aforementioned pump 500 is particularly suited for this purpose, and the apparatus 700 includes such a pump 500.

The hyperpolarised gas needs to be separated from other gases in the hyperpolarisation chamber 710 such as the alkali metal vapour and buffer layer.

Therefore, some form of purifier 740 will be coupled to the hyperpolarisation chamber 710. In the embodiment shown in fig. 7, hydrogen gas is used as the buffer layer, which can be separated from the hyperpolarised gas using combustion or oxidation in a combustion chamber 740. The pump 500 intakes the gas from the hyperpolarisation chamber 710 at low pressured, compresses the gas and pumps it into a storage container 750 at or near to atmospheric pressure. Alternatively, the pump 500 may pump the gas directly to its point of use, for example a mouthpiece 760 for inhalation by a patient Various pumps (not shown) may be used to differentially pump or evacuate the various parts of the system 700, including the back of the compression member 530 and the volume around the piezo-electric actuators 546, 556. However, only the above described pump 500 comes into contact with the hyperpolarised gas.

A filter (not shown) may be used after the outlet of the pump 500 to remove any particulates or other contaminants from the gas before being stored or used. The filter could be an integral part of the outlet valve 550 of the pump 500, or it could be a separate filter in the downstream line. Since it is important for medical applications that no unintended substances are in the final gas output, various detectors could also be incorporated into the apparatus.

The gases used for hyperpolarising can be very rare and expensive. This particularly true of some of the noble gases such as 3He. Therefore, it is desirable that the apparatus is arranged to recover the hyperpolarised gas after use, in order that it can be reused. The gas may also be purified by the apparatus, in order that it could be reused within the apparatus itself, minimising the amount of gas that needs to be replaced during usage.

All parts of the apparatus 700 would ideally be made from non-magnetic materials to avoid any non-uniformities in the magnetic field. The surfaces of all parts that the hyperpolarised gas could come into contact with is preferably treated or coated to reduce the depolarisation due to contact with the surfaces. These coatings have been discussed above in relation to the inner surfaces of the pump.

The apparatus 700 is suitable for providing on-site production of hyperpolarised gases, with the pump 500 able to compress the gas from the low pressures of the hyperpolarisation chamber 710 to approximately atmospheric pressure, either for storage 750 or immediate use 760, without significant depolarisation. Since the pump 500 is compact, with all driving means for the inlet valve 540, outlet valve 550 and compression member being integrated together within the pump 500, the hyperpolarising apparatus 700 as a whole can be reduced in size, compared to the use of other pumps, which may require motors or other forms of driving means to be located several metres away from the hyperpolarised gas.

Claims

Claims: 1. A dry pump arranged to pump hyperpolarised gas, the pump comprising: a pump body, a compression member provided within the pump body and arranged to be movable in an axial direction, wherein the pump body and compression member define a pump chamber; a non-magnetic linear drive coupled to the compression member and arranged to move the compression member within the pump body to reduce the volume of the pump chamber on a compression stroke and increase the volume of the pump chamber on an intake stroke, the non-magnetic linear drive arranged to provide minimal lateral forces on the pump body; an inlet valve coupled to an inlet actuator, the inlet actuator arranged to selectively allow the hyperpolarised gas into the pump chamber through the inlet during the intake stroke, an outlet valve coupled to an outlet actuator, the outlet actuator arranged to selectively allow the hyperpolarised gas out of the pump chamber through the outlet valve during the compression stroke, wherein the pump is made only of non-magnetic materials so as to avoid depolarisation of the hyperpolarised gas.
2. The pump of claim 1, further comprising a controller arranged to control the inlet actuator, the outlet actuator and non-magnetic linear drive.
3. The pump of claim 2, wherein the controller is arranged to independently control the inlet actuator, the outlet actuator and non-magnetic linear drive.
4. The pump of any preceding claim, wherein the controller is arranged to open the outlet actuator only when there is a specific pressure difference between the sides of the outlet valve, which could optionally be substantially no pressure difference.
5. The pump of claim 4, wherein the controller is arranged to receive a measured pressure difference between the sides of the outlet valve, or to receive a measured pressure from each side of the outlet valve and to determine the pressure difference therefrom.
6. The pump of any one of claims 2 to 5, wherein controller is arranged to time activation of the inlet and outlet actuators to avoid the pressure in the pump chamber exceeding the pressure outside outlet valve or falling below the pressure on the outside of the inlet valve by a predefined amount, or to maximise the transport gases through the pump.
7. The pump of any one of claims 2 to 6 further comprising a sensor arranged to sense the position of compression member within the pump body, wherein the controller is arranged to activate the inlet and outlet actuators based on the sensed position of the compression member.
8. The pump of any one of claims 2 to 6, wherein the controller is arranged to control the inlet and outlet actuators based on the controlled position of the nonmagnetic linear drive.
9. The pump of any preceding claim, wherein the hyperpolarised gas is a hyperpolarised noble gas, for example one or more of the group consisting of: helium-3, Neon-21 Krypton-83, Xenon-129 and Xenon-131.
10. The pump of any preceding claim, wherein the non-magnetic linear drive is a piezo-electric drive.
11. The pump of any preceding claim, wherein the inlet and outlet actuators are piezo-electric actuators.
12. The pump of any preceding claim, wherein the pump body and compression member are made from at least one of titanium, aluminium and alloys thereof, ceramics and plastics.
13. The pump of any preceding claim, wherein at least a portion of the surfaces of the pump body and compression member forming the pump chamber, and the inlet and outlet valves are treated or provided with a coating that reduces the depolarisation of the hyperpolarised gas and, optionally, the coating is at least one of aluminium, titanium, gold, PTFE, polymers, a diamond-like carbon coating.
14. The pump of claim 13, wherein the coating is a polymer modified by at least one of deuteration halogenation, sulfonation and carboxylation to reduce its permeability to the hyperpolarized gas or to reduce depolarisation of the hyperpolarised gas.
15. The pump of any preceding claim, wherein the shape of the pump chamber, piston, inlet valve and outlet valve is arranged to reduce turbulence when pumping hyperpolarised gas.
16. The pump of any preceding claim, further comprising at least one seal provided at the circumference of the compression member, so as to substantially prevent gas from escaping between the compression member and the pump body.
17. The pump of claim 16, wherein the seal is formed from at least one of ceramics, polyolefins and fluoroelastomers.
18. The pump of any preceding claim, wherein the linear actuator and the compression member are in a sealed environment, such that no air can enter the pump chamber if the seal between the compression member and the pump body fails, and optionally the sealed environment is differentially pumped.
19. The pump of any preceding claims, wherein the pump is arranged to pump hyperpolarised gas from a sub-atmospheric pressure to atmospheric pressure or above.
20. The pump of claim 19, wherein the sub atmospheric pressure is between about 0.1-100.0 mbar (0.01 kPa to 10.0 kPa).
21. The pump of any preceding claim, wherein the pump is configured to withstand baking at a temperature of at least 250 degrees Celsius.
22. An apparatus for preparing and pumping hyperpolarised gas, the apparatus comprising a source of polarisable gas, an optical pump cell and the pump according to any preceding claim, wherein the pump is arranged to pump the hyperpolarised gas out of the optical pump cell or any ancillary chambers.
23. The apparatus according to claim 22, wherein two or more pumps are connected to form a multistage pump.
24. The apparatus according to claim 22 or 23, wherein the apparatus is further arranged for carrying out in vivo imaging, spectroscopy or other biomedical applications.
25. A method of pumping hyperpolarised gases using the pump of any of claims 1 to 21, the method comprising: using the non-magnetic linear drive of the pump to move the compression member in the axial direction within the pump body to reduce the volume of the pump chamber on the compression stroke and increase the volume of the pump chamber on the intake stroke; opening the inlet valve with the inlet actuator to selectively allow the hyperpolarised gas in to the pump chamber through the inlet valve when the compression member is performing an intake stroke; and opening the outlet valve with the outlet actuator to selectively allow the hyperpolarised gas out of the pump chamber through the outlet valve when the compression member is performing a compression stroke.
26. The method of claim 25, wherein the outlet valve is opened on the compression stroke when the pressure in the pump chamber is substantially equal to the outlet pressure and closed when the compression member reaches the cylinder head.
27. The method of claim 25 or 26, wherein the inlet valve is opened on the intake stroke when the pressure in the pump chamber is substantially equal to the inlet pressure, and closed when the compression member reaches its lowest position.