The off-state conductance of a standard heat switch is approximately 37µW/K at a temperature of 4K. Heat switches are extremely delicate and cannot be used to provide mechanical support, indeed they MUST not be used in that way. For this reason we do not measure or specify the compressive and tensile strength. So far as lifetime is concerned, that is difficult to say as some heat switches give trouble-free operation for many years, while others do develop faults in a shorter period. We have approximately 2% failure rate overall for heat switches once they have been shipped. However we will provide free replacement under warranty for switches that fail in use due to manufacturing defects.
While it is possible to use different materials for the switch case, it can cause difficulties with the joining materials, thermal expansion and permeability to helium. These factors increase the expense of manufacture.
Activation (turning the switch ON) typically takes around 1.5 to 2.5mW of heater power, depending on the design requirements. This heat is usually dumped to the fixing point at 4K, as the transition temperature (ON/OFF) is around 15 to 20K.
Thermal conductivity in either the ON and OFF state of a heat switch is affected by temperature – the thermal conductivity of stainless steel and helium gas varies with temperature. Importantly, steel will determine the OFF-state conductivity, and helium gas the ON-state conductivity. Conductivity will also be affected by the switch’s geometry.
A typical standard He4-filled switch will have an OFF-state conductance of around 3mW between 4 and 30K, and around 10mW between 4 and 45K. It’s switching ratio will be around 500.
If you need the cool to below 2K, the heat switch will be filled with helium-3. We can build specialist switches which are optimised for low OFF-states for applications (down to 50mK or below).
For applications which need to cool down to a few mK, we have achieved a low µW conductance with high performance designs. These switches are usually at least 10mm long but can operate horizontally or vertically.
We make bespoke gas-gap switches as well as some standard types. The standard units are 2.5” (63.5mm) long and are usually mounted upright, but can operate in any orientation. Bespoke types vary in length depending on the application requirements. To design a bespoke switch, we need to know the customer’s thermal requirements for the OFF-state and an idea of ones for the ON-state (e.g. you need to cool a mass from a high temperature to a lower one on some timescale, or you need to dump so many joules of heat of magnetisation from a salt pill). With this information we can estimate the dimensions, feasibility and price of the switch.
Our most recent iteration of the mini-dilutor, tested on a powerful single-shot He7 precooler, achieved a base temperature of around 65mK. We are currently testing whether this dilution module will run satisfactorily with a continuous CC7 precooler. We hope that it will run at 100mK under at least 3µW of applied power. There should be around 100 to 200µW of additional cooling power at the still, at around 800mK or so. The complete mini-dilutor, with continuous CC7 pre-cooler, fits within a cylinder of diameter no more than 20cm, and between 25 and 30cm tall.
All of our products are full sealed and self-contained systems requiring only electrical inputs. There are no gas handling requirements or external reservoirs. A PTC or GM type 4K platform of 0.5W cooling capacity at 4K should be more than adequate to run the CC7 with mini-dilutor. Cool-down from room temperature will be faster with a larger GM/PT capacity, of course. The orientation-dependence of operation is a matter for future R&D effort. We have managed +/- 20 degrees or so, and hope to do better in future.
Please watch our videos online to understand how our coolers work.
Our continuous sub-Kelvin sorption coolers are designed to run from a PTC or GM type mechanical cryocooler. The continuous cooler has two sets of modules alongside a condenser module. A CC4 has two sets of GL4 modules, and a CC7 has two sets of GL7 modules. The continuous coolers work by alternating cycles i.e. each set of modules takes turns to cool the condenser module. While one set of modules is keeping the condenser module cold, the other side can be recycling and preparing to take over. This can fully automated so the coolers keep going, and going, and going….
A He4 system (CC4) will run under several hundred µW of load at around 1K or just below. Whereas, a He3 continuous cooler (CC7) will reach a base temperature of below 300mK, and can run under several hundred µW of load, although the run temperature does rise significantly as the load is increased.
Click here to view “Work below 300mK”, a video about how the GL7 works.
The GL7 has a mainplate that interfaces to the cold head of a mechanical cryocooler. The GL4 has just one module, whereas the GL7 has two. The larger of the modules contains Helium-4, and the smaller module contains Helium-3. The GL7 exploits the different properties of these two isotopes; most importantly, Helium-3 liquifies at a lower temperature than Helium-4.
In the GL7, we use the Helium-4 module to cool the Helium-3 module. Both modules are connected together through thermal links. To get the Helium-3 module cold, we start by cooling the Helium-4 module. (For an explanation on cooling the 4-module, please view our work below 1K .) Once the 4-module is cold, the temperature of the head will be just below 1 Kelvin.
With the Helium-4 module is cold, we can start a similar process to cool the Helium-3 module. We heat the pump of the 3-module to release the Helium-3 gas adsorbed onto the charcoal. The gas is cooled by the thermal links to below the Helium-3’s liquefaction temperature. As the Helium-3 condenses into liquid, it collects in the head of the 3-module.
When we turn off the power to the 3-pump and turn the heat switch on, the pump cools down to the same temperature as the mainplate. Meanwhile the liquid Helium-3 in the head starts to evaporate, and this evaporation cools the 3-head down to around 300 millikelvin. The Helium-3 module will remain at this temperature until all the liquid Helium-3 has evaporated and returned to the charcoal in the pump. To re-cycle the GL7, one turns off both heat switches and warms up both pumps. The process will no be ready to start again.
Our sorption coolers are quite small and we are actively working to minimise the thermal load imposed on the 4K cryocooler stage. We aim to keep the cryocooler thermal demand to a minimum. An RDK101 with 100/160mW of cooling power is adequate for a GL4 or small GL7. For a CC4/7 a PT 405 or 407 with 250/400mW should be more than adequate.
There are a number of ways to interface a CRC cryocooler into your cryostat, depending on your experiment and cryostat design.
The picture on the right below shows a cryostat with a GM cooler mounted to one side and the CRC cryocooler mounted through the 4K plate. The pumps are in the 40K space, and the heads are in the 4K space with a radiation shield around them. In this configuration the pumps do not need their own radiation shield.
The picture on the left shows a cooler with a radiation shield around the pumps. This is required when the whole cooler is mounted within the 4K space, because the pumps reach temperatures of up to 50K when the cooler is running. This design would still need to be thermally and mechanically supported to the 4K plate. It is also possible to design a radiation shield of this kind that provides the mechanical support to the 4K plate, see the interfacing page for examples. Should you need a radiation shield you may opt to build this yourself, or ask us to supply one as part of your cryocooler. (Remember we need to quote for this, factoring in a day of design time).
Have a look at our page about interfacing, there are several options on how to support our sorption coolers. Depending on which option you select, you may need a radiation shield around the hot pumps. This will also determine where the electrical connector will be best situated – either on the ‘hot’ or the ‘cold’ side of our sorption cooler mainplate.
Regarding the matter of sample mounting and support, we would not generally recommend mounting directly onto the sorption cooler. We normally suggest that the sample be mounted onto a separate ‘cold table’ with thermal contact to the sorption cooler by means of a heat strap. This decouples the mechanical support from the thermal linkage, bringing great advantages in mechanical independence at a minimal cost of thermal performance, assuming that the arrangement follows sound principles of thermal isolation etc.
We supply our coolers with the temperature sensors and heaters wired to a micro-D connector fixed on the cooler. (We can send the datasheet for the Glenair connectors we use). A GL7 has a 25-way MDM, and a GL4 has a 21-way MDM. The end user will need to either make or buy the cables to connect this plug to the world outside the cryostat. It is not possible for us to provide that cable, as every user has their own system of wiring into the cryostat. The user will also want to connect wires to their own experiment that they are cooling, and to their GM or PTC as well.
In our system, we have a cryostat cable that terminates at one end in a hermetic (vacuum) 55-way connector that fixes to the outside of the cryostat. There is a loomed ribbon cable that joins the outside connector to an internal 51-way MDM connector that we permanently fix into the inside of our cryostat, on the ‘warm’ side of the 4K plate, inside the 40K radiation shield. The loomed ribbon cable is made of manganin / constantan wires, which have high electrical conductivity, but very low thermal conductance, to minimise the flow of heat from the outside world into the cryostat. This cable was made for us by a specialist cryogenic wiring company – we have one of these fixed on each of our cryostats, it is part of the cryostat, not part of the sorption cooler. It is very important to have the right specification for the cryostat cable as otherwise it will impose a heavy heat load on the experiment.
Then, for each set-up we make an adaptor cable for the internal wiring. This adaptor connects the CRC sorption cooler (plus other sensors) to the fixed cryostat wiring cable. If the CRC sorption cooler has its electrical plug on the ‘warm’ side of the 4K plate, i.e. in the same space as the cryostat cable termination, the adaptor can use copper wires since it is entirely within a single zone of the cryostat. If it needs to pass through into the 4K space then close attention will be needed to the material, length and thermal sinking of the wires. So, if our customer already has a cryostat with a fixed wiring cable installed, they will need to make their own adaptor to connect their cryostat wiring to the CRC cooler. We are happy to advise on wiring if needed.
Our sorption coolers aren’t designed to have large masses hung from them. If you need to hang small to moderate masses, it is strongly recommend using a separate cold table to mount the object being cooled. Torsion, bending or non-axial loads will break any of our coolers.
Our sorption coolers should be used with the cold head vertically downwards i.e. at the bottom. However, the coolers can be operated up to 45 degrees to the vertical, once they are cold.
At present our sorption coolers aren’t UHV compatible, but we are keen to develop ones which are! The main hurdles involve the attachment of thermometers, heaters and wiring to the cooler. Such issues should be solvable with suitable material, assembly and production method choices. Additionally, we are considering ways of dealing with the additional pressure during a bake-out. Contact us to discuss your requirements and whether we can accommodate them.
We do not currently supply ‘turn-key’ cryogenic systems, but we are heading that way. At present we supply only the sub-Kelvin sorption cooler hardware. We can advise on where to buy a cryostat and GM/PT cryocooler to operate our sorption cooler, on thermometry and other instrumentation and on software to run our continuous systems.
The total cooling capacity of a GL7 cooler is approximately 1 Joule per NTP litre of He3 used, at 300mK (although this assumes 100% condensation efficiency). The cooling capacity of a GL4 cooler is around 3.6 Joules per NTP litre, but only around 40% of the total charge is available for cooling during the run – the rest is consumed by the ‘base load’ of the cooler.
Your cooling power requirements will affect the size of the sorption cooler’s He3 and He4 modules. A greater load requires larger modules. Single-shot ‘GL7’ sorption coolers will typically run for 24 hours under 20 to 50 µW loading.