The invention relates to a cryostat assembly, for example for use in nuclear magnet resonance (NMR) or dynamic nuclear polarization (DNP) apparatus.

In certain applications such as NMR and DNP, there is a need to be able to cool an object such as a magnet or sample to very low temperatures. Conventionally, this has been achieved using liquid cryogens such as liquid helium and/or mechanical coolers such as pulse-tube refrigerators. However, the lowest temperature achievable with these mechanisms is about 4.2K and there is a need to be able to achieve even lower temperatures. One approach to solving this problem has been to make use of an adsorption pump assembly. These are described, for example, in EP-A- 1387133.

An adsorption pump assembly comprises an adsorption pump formed by an adsorbing material such as charcoal which is in fluid communication with a pumping chamber, typically an elongate pipe defining a closed volume. Within the pumping chamber is provided a working fluid such as helium. As the adsorption pump is cooled, it adsorbs gaseous helium thus reducing the pressure within the pumping chamber and causing any liquid phase working fluid at a lower end of the pumping chamber to evaporate. This evaporation then cools the lower end. In a subsequent regeneration phase, the adsorption pump is heated so as to cause the previously adsorbed gas to be desorbed thus allowing the cooling cycle to repeat.

The problem with these known systems is that the adsorption pump must be heated in order to regenerate the gaseous working fluid. Conventionally, this is achieved by providing an electrical heater adjacent the adsorption pump but the disadvantage of this is that additional heat must be introduced into the cryostat assembly which is undesirable, as it reduces the thermal efficiency of the system.

In accordance with the present invention, a cryostat assembly comprises: a thermal shield; a cooling system including a cold member which is maintained in use at a temperature lower than that of the thermal shield; a first adsorption pump assembly located at least partly within the thermal shield, the first adsorption pump assembly comprising a pumping chamber containing a working fluid, and an adsorption pump in fluid communication with the pumping chamber; a first switch for selectively, thermally connecting the adsorption pump to the thermal shield so as to cause the working fluid to desorb from the adsorption pump,- a second switch for selectively, thermally connecting the adsorption pump to the cold member so as to cause the working fluid to be adsorbed by the adsorption pump; and a first thermal link coupling the pumping chamber to the cold member so as to condense at least a portion of the working fluid desorbed from the adsorption pump, wherein during adsorption of the working fluid by the adsorption pump, a portion of the working fluid evaporates so as to cause the temperature within the pumping chamber to fall below the temperature of the cold member.

We have realised that it is possible to heat and cool the adsorption pump using existing heat sources within the cryostat. In particular, the adsorption pump can be heated by connecting it to the thermal shield and can be cooled by connecting it to the cold member. This avoids the need for any additional electrical heater and overcomes the problems set out above.

The cooling system can be provided by any conventional system such as a mechanical or cryo cooler or a liquid cryogen containing vessel, for example containing liquid helium. In the case of a mechanical cooler, e.g. a PTR or Gifford-McMahon refrigerator, this would typically have at least two stages with a first stage coupled to the thermal

shield and the second stage defining, or coupled to, the cold member. A liquid cryogen containing vessel could also be cooled by a mechanical cooler and boil off from that vessel could be used to cool the thermal shield. - The pumping chamber typically has an elongate form, for example comprises a pipe, and will normally extend downwardly from the adsorption pump. The full operation of the adsorption pump assembly will be described in more detail below. However, briefly, when starting from a state in which the pumping chamber contains working fluid in both gaseous and liquid form, the second switch is closed in order to cool the adsorbtion pump which then adsorbs gaseous working fluid causing the liquid phase to evaporate. This evaporation cools the adjacent part of the pumping chamber. At the end of the evaporation stage, the second switch is opened and the first switch closed thus connecting the adsorption pump to the thermal shield. This warms the adsorption pump which then desorbs gaseous working fluid, at least part of which is then condensed due to contact with the thermal link between the pumping chamber and the cold member, the liquid working fluid falling to the bottom of the pumping chamber. The cycle is then repeated.

It will be appreciated that in the case of a single adsorption pump assembly, during the regeneration phase when gas is being desorbed, there will be a rise of temperature in the pumping chamber. In some applications, this is acceptable and an example is dynamic nuclear polarisation (DNP) where continuous cooling is not required.

In other cases, however, such as NMR, for example NMR spectroscopy, there is a need to cool a system coupled with a cryostat substantially continuously. In these cases, therefore, the assembly further comprises a second adsorption pump assembly located at least partly within the thermal shield, the second adsorption pump assembly comprising a pumping chamber containing a working fluid,

and an adsorption pump in fluid communication with the pumping chamber; a third switch for selectively, thermally connecting the adsorption pump of the second adsorption pump assembly to the thermal shield-so-as to cause the working fluid to desorb from the adsorption pump,- a fourth switch for selectively, thermally connecting the adsorption pump of the second adsorption pump assembly to the cold member so as to cause the working fluid to be adsorbed by the adsorption pump; and a second thermal link coupling the pumping chamber of the second adsorption pump assembly to the cold member so as to condense at least a portion of the working fluid desorbed from the adsorption pump, wherein during adsorption of the working fluid by the adsorption pump, a portion of the working fluid evaporates so as to cause the temperature within the pumping chamber to fall below the temperature of the cold member.

With this arrangement, while one adsorption pump assembly is regenerating, the other is adsorbing gas and thus cooling within its pumping chamber. By connecting both adsorbing pump assemblies to the same system, that system can be kept at the lowest temperature.

It will be appreciated, however, that it is important to minimize any thermal link between the adsorbing pump assemblies. This can be achieved by suitably connecting the respective pumping chambers, typically via heat exchangers, with the system to be cooled. For example, a mechanical linkage via a thermal switch could be provided, that switch being operable to close when the temperature of the pumping chamber is less than that of the system to be cooled but otherwise to open. Alternatively, a "thermal diode" arrangement can be provided, typically using a cryogen fluid such as helium which is caused to condense on the pumping chamber and fall towards the system to be cooled and then evaporate. The flow of heat from the system to be cooled back to the pumping chamber is much

less efficient than the cooling power in the opposite direction so that while an adsorbing pump assembly is regenerating, little heat passes to it.

Typically, the thermal shield will at least partly, preferably__fully, surround the cold member. For example, in certain applications no thermal shield is provided within the inner bore of a magnet cooled by the cryostat assembly, or no thermal shield is provided at the bottom of the assembly. An outer annular thermal shield extending only about the first (and second) adsorption pump assemblies is also feasible.

The switches can have a conventional form and conveniently comprise gas-gap heat switches. These are activated by supplying helium gas to a gap between a pair of copper plates so as to close the switch, the gas supply being terminated to open the switch.

The first and second thermal links are preferably permanently connected between the pumping chambers and the thermal shield although they could be selectively connected via respective switches since the cooling power is only required during the regeneration stage.

Some examples of cryostat assemblies and systems utilizing such assemblies in accordance with the invention will now be described with reference to the accompanying drawings, in which:-

Figure 1 is a schematic section through a first example of an NMR system utilizing a cryostat assembly;

Figure 2 is a view similar to Figure 1 but of a modified NMR system; and,

Figure 3 is a view similar to Figure 1 but of a DNP system.

The assembly shown in Figure 1 comprises an outer vacuum chamber 1 surrounding a thermal shield 2. Within the thermal shield 2 is located a cylindrical helium containing chamber 3 which communicates via a conduit 4 with a source of liquid He4 (not shown) . The chamber 3 also communicates via a conduit 5, which is wrapped around

the outside of the thermal shield 2, with a non-return valve 6 so that helium boiled off from the chamber 3 passes adjacent the thermal shield in order to cool it before being recovered.

A further - helium- chamber —10 is surrounded by the chamber 3 and contains a superconducting magnet 11. Liquid helium is provided in the chamber 10 as shown at 12 so as to cool the magnet 11. Above the magnet 11, the helium will exist in gaseous form. In practice, a (room temperature) bore will extend into the magnet 11 coincident with its axis, the bore being accessible from outside the vacuum chamber 1. This has been omitted for clarity.

The upper wall of the chamber 10 is formed into a pair of tubular extensions 15,16.

Extending into each tubular extension 15,16 is a respective pumping chamber 17,18 in the form of an elongate pipe which is closed at its lower end by a heat exchanger 19,20 and is also closed at its upper end. Each pumping chamber 17,18 contains He ⁴ as will be described in more detail below.

The upper section of each pumping chamber 17,18 is defined by an inner bore of a respective adsorbing pump, or sorb 17A, 18A, each pumping chamber and associated sorb defining a respective adsorption refrigerator Sl, S2.

Permanent thermal links 30,31 are provided between the chamber 3 and each of the pumping chambers 17,18 at positions below the sorbs 17A,18A and a further thermal link between the two pumping chambers as shown at 32.

A further thermal link 33 is coupled via a switch HS4 with a lower end of the sorb 17A and a thermal link 34 is connected via a thermal switch HS3 with the sorb 18A.

The upper end of each sorb 17A,18A is connected via a respective thermal switch HS2,HS1 with the thermal shield 2.

The thermal switches HS1-HS4 have a conventional form and conveniently comprise gas-gap heat switches which broadly comprise a pair of metal plates, typically

interleaved, defining a gas path therebetween. In order to close the switch, cryogen gas is supplied between the plates so as to form a thermal link between them, while to open the switch, the gas supply is terminated. An example of a suitable switch can be found at http: //irtek.arc.nasa.qov/HeatSwitch_page/GGapHtSwDesc.html

Operation of the switches is controlled by a microprocessor 40 as shown schematically in Figure 1.

As will be explained in more detail below in connection with the Figure 2 example, by suitably operating the thermal switches HS1-HS4, it is possible to reduce the temperature within the chamber 10 below the boiling point of helium i.e. 4.2K.

The example shown in Figure 2 is substantially the same as that shown in Figure 1 and so those components which are equivalent have been given the same reference numbers. Again, the room temperature bore has been omitted.

There are two principle differences. Firstly, instead of utilizing a supply of liquid helium to provide the primary cooling power, the chamber 3 is coupled to a second stage 50 of a pulse-tube refrigerator (PTR) 52. The first stage 54 of the PTR 52 is coupled to the thermal shield 2.

The second difference is that the magnet 11 is provided in a dry chamber 60 located within the chamber 10. Consequently, a layer of liquid helium 62 forms above the chamber 60 which is made of a material which provides a suitable thermal link to the magnet 11.

In order to understand the operation of the assemblies shown in Figures 1 and 2, we will now explain the function of the adsorption refrigerators and associated heat exchangers and switches.

The function of the sorb materials such as charcoal in each adsorption refrigerator Sl, S2 is to adsorb gas when it is cooled and to regenerate or desorb gas when it is heated. Thus, considering the Figure 2 example and adsorption refrigerator Sl, when the switch HS4 is closed

as shown in Figure 2, the sorb material 17A is cooled and this increases its adsorption capacity so that it will adsorb gas from the pumping chamber 17. At the base of the pumping chamber 17 adjacent the heat exchanger 19, the helium will exist in liquid- form. As the gaseous helium is adsorbed, this will reduce the pressure above the heat exchanger causing the liquid to evaporate and in turn causing a reduction in temperature on the heat exchanger 19.

Following adsorption, the switch HS4 is opened and the switch HS2 closed so as to connect the sorb material 17A to the warmer temperature of the thermal shield 2. This will cause gas to desorb from the sorb material and pass down the pumping chamber 17 past the permanent, thermal link 30 to the chamber 3. This will cause the gas to liquify and drop further down onto the heat exchanger 19 after which the cycle is repeated.

Thus, during the adsorption phase, the heat exchanger 19 is cooled below the temperature of 4.2K and this will be thermally communicated to gas within the tubular extension 15. The gas will therefore be cooled below its liquifying temperature so as to form a liquid which then drops down onto the chamber 60 as shown at 62. During the regeneration phase, the heat exchanger 19 will be warmer and therefore will not cause gas to liquify within the tubular extension 15. However, by operating the adsorption refrigerator S2 in an opposite cycle, it is possible to be liquifying helium within the chamber 10 either within the tubular extension 15 or the tubular extension 16 at all times so as to maintain a liquid layer 62.

Table 1 below sets out in more detail the steps for operating the Figure 2 example.

TABLE 1

In the case where there is no PTR and 4.2K cooling is provided by a boiling bath of liquid helium reference to the second stage of the PTR can be replaced by reference to the liquid helium 4.2K reservoir.

It will be appreciated that the gaseous helium in the upper section of the chamber 10 in each example and in the tubular sections 15,16 acts as a thermal diode so that there is minimal heat flow between the tubular extension 15 and -tubular extension 16 -and no or minimal heat flow from the heat exchanger 19,20 during the desorption phase to the magnet 11.

In the examples, a permanent link is provided at 30,31 between the chamber 3 and the pumping chambers 17,18. This could be replaced by a thermal switch so as to reduce the risk of heat flow to the chamber 3 during the adsorption phase.

The examples described above are suitable for applications such as NMR or MRI. Figure 3 illustrates an example suitable for other applications such as DNP. In this case, an outer vacuum chamber 100 surrounds a thermal shield 102 which in turn surrounds a bore 104 into which a DNP insert (not shown) can be located. The DNP insert (not shown) contains hardware for applying microwaves to a sample to cause dynamic nuclear polarization and hardware for dissolving the sample. Suitable apparatus is described in " ^v Increase in signal-to-noise ratio of > 10,000 times in liquid state NMR ¹¹ , JH Ardenkjaer-Larsen et al, PNAS, Sept 2, 2003, Vol. 100, no. 18, and in WO-A-02/37132. The bore 104 contains liquid helium 106 into which the sample is immersed.

A magnet 108 located in a liquid helium chamber 110 surrounds the bore 104 and this is cooled by a second stage 112 of a two stage PTR 114. The first stage 116 of the PTR 114 is coupled to the thermal shield 102.

A pumping chamber 120 is located within the thermal shield 102 and is coupled at one end 122 via a thermal switch 124 with the bore 104. The pumping chamber 120 is closed and contains a coolant He ⁴ (not shown) . The other end 126 of the pumping chamber 120 is defined by the inner bore of a sorb material 128. The sorb 128 is connected via

a thermal switch 130 to the thermal shield 102 and via a thermal switch 132 to the chamber 110.

A weak thermal link 140 is provided between the chamber 120 and the chamber 110.

--The principle—of--operation of this system is similar to that described in connection with Figures 1 and 2. By suitably connecting the sorb material 128 with the low temperature chamber 110 or the higher temperature thermal shield 102, gas can be caused to adsorb or desorb respectively thus varying the pressure within the pumping chamber 120 and during the absorption process causing liquid to evaporate in the region 122 thus cooling that region and hence cooling the helium 106 in the bore 104 to a temperature below 4.2K. The helium will stratify with colder liquid below a layer of 4.2K liquid. The helium vapour in the bore remains at atmospheric pressure.

It will be understood that in this case where only a single adsorption refrigerator is provided, while gas is desorbed from the sorb material 128, the heat switch 124 is open and the bore 104 will not be cooled to the very low temperature whereas while the gas is adsorbed the switch 124 is closed and the contents of the bore are cooled. In the case of applications such as DNP, it is not necessary to maintain the temperature substantially constant since the sample is removed before the regeneration stage.

A particular advantage of this embodiment is that the helium in the bore 104 is entirely separate from that used to cool the magnet 108. This is an advantage in applications in which the sample must be kept entirely free of contamination, for example if it is to be used as an MR contrast agent.