GB2420662A

GB2420662A - Electrocaloric colling device with heat switches

Info

Publication number: GB2420662A
Application number: GB0426230A
Authority: GB
Inventors: Alexandr Mishchenko; Neil David Mathur
Original assignee: Individual
Current assignee: Individual
Priority date: 2004-11-29
Filing date: 2004-11-29
Publication date: 2006-05-31
Also published as: GB0426230D0

Abstract

A heat transfer device comprises electrocaloric elements 200, 210, 220, 230 and heat switches 100, 110, 120, 130, 140. The device also includes a control input for controlling the heat switches 100, 110, 120, 130, 140 and electrocaloric elements 200, 210, 220, 230 to transfer heat from a heat source 20 to a heat sink 10. The heat switches 100, 110, 120, 130, may be formed as thermoelectric, tunnelling or microelectromechanical switches. Electrocaloric elements 200, 210, 220, 230 take the form of multilayer capacitors comprised of ferroelectric thin films or can be made from bulk materials. The device can be cascaded or paralleled to provide greater heat transfer. A method of controlling a heat transfer device is also disclosed.

Description

Solid State Electrocaloric Cooling Device with Heat Switches and Method of

Cooling

DESCRIPTION

Background

Introduction

Some observers believe the physical limits of silicon will one day slow or halt the steady advance of the computing industry. Today though, a more immediate threat to Moore's law has emerged. The same forces that enable exponential improvements in processor performance are creating enormous problems in another dimension: power. As the transistor density increases, processors consume more power and generate more heat, at an accelerating rate. If past trends continue, with no change in power management, then, over the next decade, high-perform computing devices will surpass the industry's ability to cool them economically IiJ. Despite the many novel approaches that are being tried now to develop alternative cooling methods, there are Possible Solutions that are still almost missed by the scientific community, one of them is an electrocaloric refrigeration system proposed in this invention.

Common refrigeration Systems use greenhouse gases that are hazardous to the environment. A new reliable active solid state device Without moving parts proposed in the present invention can offer some advantages over the common approach to refrigeration, both in the electronics and computer industries, and in home and industrial refrigeration systems. Being more efficient, the proposed refrigeration method can help to save energy spent on numerous refrigeration facilities all around the world.

Electrocaloric Effect and Electrocalorjc Working Element The electrocaloric effect is, broadly speaking, a change of a materlial's temperature upon an application or removal of an electric field. There are some obvious facts that make the electrocaloric effect (ECE) more attractive for applications than magnetocalo effect (MCE): * it is easier to generate and maintain electric rather than

magnetic fields,

* it is easier to organ ise the cycle of energy return after each cycle of the refrigeration, * it is easier to build a cascaded cooling device comprising several working elements separated by heat switches, etc. A standard geometry of an electrocaloric (EC) working element is shown in Fig. 1. It comprises a pair of electrodes 300 and a slab of electrocajoric material 310. So an electrocaloric element is basically a capacitor with an electrocaloric material between its electrodes The element is energised from a DC or AC voltage source 330, and the voltage is applied by Switching 320. So when the switch 320 is turned on, the electrocaloric material 310 polarises and its temperature changes due to the electrocaloric effect. When 320 IS turned off, the temperature of 310 returns to its initial value. It should be understood that the Figure is a schematic only and different circuitry can be used, as well as different structure of the electrodes (e.g. multilayer capacitor structure, similar to that described in [2], for example; layers of electrocaloric material in the multilayer capacitor can be made with thin film deposition techniques, e.g. sol-gel, as disclosed in [3], for

example).

Electrocaloric Materials The best known electrocaloric materials are perovskite ferroelectric materials. The highest electrocajoric effect at room temperature was found in PbSc0 5Ta0 503 (PST). For a more detailed description of the material see [4], for example The value of the ECE peaks at the critical temperature Tc but remains within 80% of the maximum value within about 10 C region around T. The critical temperature can be adjusted by a special heat treatment of the material during the annealing phase of its manufacturing [5]. A typical dependence of ECE for PSI after different heat treatments at electric field kVcm1 is shown in Fig. 2., which is adapted from 16].

The highest ECE was found in PbZrTiO3 (PZT) doped with Sn and Nb: Pb0 ggNb0 o2(Zro75Sn0 2Ti0 05)0 9803 [7]. The ECE peaks at about 160 C and has a maximum value of 2.6 C at 30 kVcm1. Its dependence upon temperature for samples obtained with different treatments is shown in Fig. 3.

Due to the nature of the electrocaloric effect, it is natural that its value grows with the electric field. In practice the maximum field that can be applied to a material is determined by the breakdown voltage. For people acquainted with the state of the art it is well known that thin film materials have much higher breakdown voltages than same Compounds in bulk form. A typical breakdown value for thin films is 2 MVcm1 while that for bulk materials is about 30 kV cm1. This gives grounds to the prediction that thin film materials will have a much higher electrocaloric effect. Some experimental evidence of this was provided in [8]. Thus, a multilayer capacitor system comprised of a plurality of ferroelectric thin films (deposited by sol gel, for example, as disclosed in 13]) separated with electrodes (with a geometry described in [2], for example) can manifest a much higher ECE than bulk material of the same composition due to the higher electric field applied. The compounds described above can be used in the proposed invention either in the form of a multilayer capacitor or in bulk form. However, it should be understood, that any materials with ECE can be used in the proposed invention.

The inherent efficiency of an electrocaloric element depends on several factors. Among them are hysteresis losses if the material used is ferroelectric and losses due to Joule heating due to finite electrical resistance of the elements. As numerous experiments showed, the inherent efficiency of EC elements 5 remarkably high - around 98% of the Carnot cycle f9}. The efficiency i was estimated using the formula: (1) 7l!D QIN - Q0 where QECE is the amount of heat generated in an EC element due to ECE, QIN is electrical energy supplied for the polarization if the EC element, QOUT is electrical energy withdrawn from the EC element at its depolarization and q is the efficiency of the ideal Carnot refrigeration cycle. The Carnot efficiency is given by: T 2

_SQLQ

ID

1HOT 1COLD where THOr and TCOLD are temperatures of the hot and cold end respectively.

Thus it is supposed in the following that the electrical circuits of the cooling device allow for saving of the electrical energy coming from a depolarization of the working elements. This is easy and straightforwr to implement by building RC - circuits for example, where the capacitance element is an electrocaloric element.

Heat Switches A heat switch is a device that can be Switched between two states - heat Conducting and heat insulating. The heat insulating state does not need to be fully insulating. Some examples are given below.

Thermoelectric heat switches are important parts of the proposed device in some embodiments, so they will be discussed in more detail here.

An example of a thermoelectric couple is presented in Fig. 4. As a rule, it comprises two legs - 440 (n-type semiconductor) and 450 (p-type Semiconductor) several layers of electrical conducting material 41 0, several layers of material 400 which is an electrical insulator and heat conductor.

When the element is energised with the voltage source 460, it pumps heat from heat source 420 to heat sink 430. The heat pumped at the cold surface 420 Qc (W) is equal to: Q. = 2N YJTC1d - - kzTG (3) where N is the number of couples (there is only one in Fig. 4), TcO,d is the cold surface temperature (K), a is the Seebeck coefficient (VK1), I is the electrical current (A), p is the resistivity of 440 and 450 (c2cm), k is the heat conductivity of 440 and 450 (Wcm1K1), zT- the temperature difference between 430 and 420 (K), and G is the ratio of an element's cross-sectional area to its height (cm). The first term in the brackets in (3) describes the Peltier effect, the second term describes Joule heating and the third describes parasitic heat flow from 430 to 420.

A thermoelectric element can work as a heat switch, i.e. it can either conduct heat (when I = 0) from the hot to the cold surface (an on state of a heat switch) or not conduct heat (when I I, at which the Peltier cooling just offsets the thermal Conductivity and Joule heating an off state of a heat switch). The value of l can be found from the equation Q= 0: Off = IQO *f?aq:1i _/ii (4) It should be understood that the value given by (4) is an example only, and any other close value can be used. Also, a heat insulating state (off state) does not need to be fully heat insulating.

A magnetocalj cooling system using thermoelectric and microelectromechaflical (MEMS) heat switches IS disclosed in [10]. Various arrangements of heat switches and working magnetocaJorj elements are mentioned. However due to the difficulty in generating a magnetic field, construction of such a device would be complicated Additionally its efficiency and cooling power would be questionable due to large heat leaks. One of significant differences between ihe device and method disclosed in [10] and the present invention is the nature of the working elements - they are electrocaloric in the present invention, not magnetocaiorjc as in [10]. It makes the whole system simpler and allows for direct estimations of the device efficiency with parameters of available materials.

High thermodynamic efficiency of thermoelectric heat switches was discussed in length in [10]. All the discussion along with the conclusions are relevant in the present context. For people acquainted with the field it is well known that efficiency of a thermoelectric material sharply increases with decrease of the temperature difference between the hot and cold contacts of a thermocouple. This difference is kept small throughout the whole cooling cycle of the proposed device.

An electrocaloric cooling system based on a cascade of electrocajoric working elements separated with piezoelectric heat switches is disclosed in [11]. A piezoejectric heat switch is similar to the electrocaloric element shown in Fig. 1, but material 310 has high piezoelectric effect in this case, i.e. significant contraction or elongation upon application of electric field. If placed between two electrocajoric elements so that upon elongation it touches both electrocaloric elements, and upon contraction touches none of them, such a piezoelectric element can work as a heat switch. The proposed device does not use bulk piezoelectric heat Switches. In some embodiments of the proposed device, microelectromechanical heat switches are used, but their major feature is a complex surface that increases the surface contact area, which is absent in [11].

State of the Art of the Electrocaloric Cooling Technology An electrocaloric cooling system comprised of a plurality of electrocajoric (EC) working elements (an example of one is depicted in Fig. 1) with a range of working temperatures covering a wide temperature interval was described in [12, 13]. The heat exchange between adjacent electrocaloric working elements as well as heat flow to the ambient and from the cooled body was organised with fluid and gas heat exchangers. The prototype schenajjc is shown in Fig. 5. The working elements were made of PbSco5Tao53 with various thermal treatment so that their working temperatures were different. Variations of ECE versus temperature for different working elements are shown in Fig. 2. The prototype comprised two blocks of working elements 1 and 2 (see Fig. 5), heat load 3, and heat sinks 4.

The liquid heat exchanger was pumped through the system with a pump 5.

Four blocks of PST have been combined into a single shell. Each block had a sandwich type structure comprising a number of PSI plates x lOx 0.3 mm3 each. Each block was 55 mm long with a cross section of x 5 mm2. The total mass of the PSI working material was 35 g. The results of the experiments with the prototype are shown in Fig. 6. Curve 1 shows the ECE of a single EC element, and curve 4 shows the temperature difference between the heat load and the heat sink. The maximum field for Curve 1 is kVcrn1 and could not be increased due to surface arc (the experiment was carried out in air). Due to the same reason the disclosed prototype with a gas heat exchanger, He, was studied only up to 30 kVcm1 (see Curve 2). The ECE effect of a working element placed in pentane, a heat exchanger liquid, is shown by Curve 3. The ECE value is lower due to higher heat capacity of pentane Compared to that of a gas. However, the net temperature difference developed between the ends of the whole device with liquid pentane as a heat exchange liquid is much higher and shown by Curve 4.

Electrocaloric refrigeration systems for cryogenic temperature ranges (from 20 K and below) are disclosed in [14, 15]. However, the design of the proposed systems did not comprise heat switches, like the prototype described in the previous paragraph and unlike the present invention.

An electrocaloric refrigeration system comprising a number of working elements Separated by passive unidirectional heat pipes was proposed in [16]. The proposed unidirectional heat pipes work by principle of unidirectional heat transfer accomplished by evaporation of a liquid heat carrier, e.g. N2 in a cryogenic temperature range. However, such a process is slow and has a large inertia, which makes the realization of such a device unlikely. Although the cited patent allowed for different heat pipes, it did not mention that they could require electrical energy to operate, what makes the present invention different. The heat switches proposed in the present invention differ from passive unidirectional heat pipes in that they are active elements and require electrical energy to function. In both claims of f16J concerning the heat pipes (claims 1 and 2), the pipes were not supposed to require electrical energy for operation.

Another electrocaloric cooling system based on piezoelectric heat switches was discussed in the end of the previous point.

Description of some embodiments of the proposed

device and estimation of their characteristics Description of some embodiments of the proposed device A proposed device comprises at least one electrocaloric working element 200 placed between at least two heat switches 100 and 110, see Fig. 7. Heat switches ioo, 110 have thermal contact with a heat sink 10 and heat source 20 respectively. There is also a thermal contact between 100 and 200,110 and 200.

In another embodiment a cascaded system comprising two electrocaloric working elements is proposed. An electrocaloric element 210 of an additional stage is utilised as a heat source in the previous stage of the cascaded system, see Fig. 8. In other words, the embodiment of a cascaded system shown in Fig.8 comprises two cooling devices: a first device comprising an electrocaloric element 200 and heat switches 100 and 110 and a second device comprising an electrocaloric element 210 and heat switches and 120.

Referring to Fig. 8, a heat source 20 is Connected mechanically to the heat switch 120 so that there is a good thermal contact between them. The heat switch 120 is mechanically connected to the working element 210, so that there is good thermal contact between them. However, there is no direct electrical connection between 120 and 210. In a similar fashion electrocaloric element 200 is thermally connected to heat switches ioo and 110 and electrically isolated from them. Elements 210 and 110 as well as 110 and 200 do not contact each other in Fig. 8 for the sake of clarity. Buffer layers providing thermal contact and electrical insulation between 10 and 100, 100 and 200, 200 and 110, 110 and 210, 210 and 120, and 120 and 20 are also not shown here for clarity.

In some embodiments, thermoelectric elements can be used as heat Switches, i.e. they can be switched between two states - heat insulating and heat conducting. In some embodiments an on state corresponds to heat flow fiom the hot o the cold elements by virtue of heat Conduction through the thermocouples in passive mode. In other embodiments, the heat flow in an on state is assisted by the application of a reversed current so that the time of heat exchange decreases, and this mode of operation is called active mode.

An off state is defined as zero or nearly zero heat transfer to the cold end of a thermoelectric couple/couples. In this state a direct current is passed through the couples so that the thermoelectric cooling effect precisely or almost precisely offsets the heat conduction. In some embodiments the current passed through the thermocouples in an off mode can be higher than that required to offset the heat Conduction process in order to reduce the time of heat exchange.

A thermoelectric heat switch incorporated into the proposed device can be energised either directly or by electromagnetic induction, in a similar way to that described in [10] in relation to magnetocalori systems.

In another embodiment thermoelectric switches are replaced by tunneling cooling devices, disclosed in [17], for example. In these embodiments a proposed device operates in a similar way to that described in the previous paragraph.

Preferably, thermoelectric heat switches can be manufactured by means of electrolysis disclosed in [18]. The cited patent proposes a thermoelectric device for electrical energy generation, but it can be used for the reverse purpose, i.e. heat pumping, due to Onsager relations.

Thermocouples made by electrolysis are with a flat geometry, e.g. with 1 mm x 1 mm cross section and 100 pm thickness. BiTe is the material used in the cited patent, and it can be used in the present invention. However, other thermoelectric materials can be used for heat switches as well. From the analysis below it follows that the higher the figure of merit of used thermoelectric materials the higher the efficiency of the whole cooling device.

The following estimations show a preferable range of sizes of thermoelectric heat switches. However it should be understood that this is just an example and heat Switches of other sizes can be used. For the cooling device to work efficiently, there are two conditions to be satisfied: (i) the heat capacity of each heat switch must be much less than that of an electrocaloric element, and (ii heat transfer must be accomplished for a reasonable amount of time. Although there is no strict objective limit on this time, I = 1 s seems a plausible upper value. A time scale of the device operation can be set by the polarisafion time (0. 1 s), so a heat transfer taking, say, 10 s, is far too slow.

These conditions can be expressed as CHeaw 2Nh2GCB,_Te <<C, c (5) 2N2kG where C is the heat capacity of an EC element, CH6atSW is total heat capacity of a heat switch, CBITe is the volume heat capacity of bismuth telluride, a possible material used ifl thermocouples, N is the number of thermocouples, h is the height of the thermocouples, G is the ratio of the area of one leg of a thermocouple to its height h, k is the heat diffusivity of BiTe. After a simplification one has two inequalities: __ C -. (6) 4Nki 2Nh2CBT A necessary Condition for this inequality to be true is simply that it should remain true if we take G out. Then C and 2N cancel out and we get a Condition for the height h: I--. (7) V CBZTh Substitution of the values for bismuth telluride (CB,Te 544 Jkg1K1, k= 0. 366*106 m2 1) and 1 s for I gives h << 1 mm. Thus, a height of - 100 jim is appropriate, and this gives a hint at a preferable method to use for the fabrication of the couples: electrochemical deposition [18], as described above. A typical value of the 6 factor is around 1 cm, so a cross sectional size of one leg is about = 0.32 - 1.0 mm and a total size of a typical number of thermocouples in a heat switch 2N = 2 * 540 legs (this value is used in the efficiency estimation described below, however a different number of thermocouples can be used) is = 10 - 33 mm. If a free space between the legs is of about the same size, the cross sectional size of the whoie heat switch is 20 - 66 mm. It is important to note that different values of G factor and other geometrical features can be used in the proposed device.

In some embodiments microelectromechanical systems (MEMS) can switch from a heat conducting to a heat insulating state. With this embodiment, when a first working element is subject to electric field, it generates heat and its temperature is raised above the temperature of the heat sink. At this time a heat switch between the first EC element and a heat load is in an off state, and a heat switch between the EC element and a heat sink is in an on state, thus allowing for the heat flow from the EC element to the heat sink. The heat flow goes on until the temperatures of the EC element and the heat sink are the same. In the next stage, the electric field is removed and the EC element cools down and its temperature drops below the temperature of the heat load. At this time the heat switch between the EC element and heat load is in an on state, and the heat switch between the EC element and the heat sink is in an off state. This configuration allows for the heat flow from the heat load to the EC element. The heat flow goes on until the temperatures of the EC element and the heat load are the same. By repeating these steps the device comprising only one EC element and two heat switches can pump heat from the cold to the hot end. Embodiments comprising more than one EC element can work in a similar fashion.

Operation of an embodiment comprising one electrocaloric element and two heat switches shown in Fig. 7 comprises: * cooling electrocaloric element 200 to a temperature of less than source 20; * turning heat switch 110 on to transfer heat from the source 20 to the electrocaloric element 200; * turning the heat switch 110 off; * warming electrocaloric element to a temperature of greater than sink 10; * turning heat switch 100 on to transfer heat from electrocaloric element 200 to sink 10; and * turning the heat switch 10 off.

In other embodiments of the proposed device, e.g. cascaded or paralleled systems having a plurality of devices for transferring heat from a heat source to a heat sink, the operation can be repeated in every device. Some examples are given below.

An example of an operation of a cascaded system comprising 2 electrocaloric elements with heat switches depicted in Fig. 8 is schematically shown in Fig. 9. After all transition processes finish and the device reaches equilibrium, electrocalorjc elements 200 and 210 are at their working temperatures Tw200 and Tw210 respectively, what Corresponds to an optimum electrocaloric effect of the material used. The working temperatures are chosen so that Tw210 < Tw200, Tw210 is close to the heat source (20) temperature T20 and Tw200 is close to the ambient (10) temperature T10. An optimum difference between adjacent working temperatures depends on the value of the electrocaloric effect among other factors. The difference between adjacent working temperatures of different pairs of working elements can be different. Referring to Fig. 9, an initial configuration of a cooling device is: 210 is depolarised, 200 is polarised, both are at their working temperatures, and alt heat switches (100, 110, and 120) are in an off state, i.e. they do not conduct heat. This configuration can be achieved by a proper biasing of the heat switches. For example, to cool down 200, the heat switch 100 should pump some heat from 200 to the ambient 10, etc. Stable cooling at constant cooling power is achieved by running elementary cooling cycles. Each cycle comprises four stages: * () Application of bias to 210. Its temperature rises due to the electrocajoric effect. At the same time 200 is depolarjsed Its temperature decreases due to electrocaloric effect. All heat switches 100, 110, and 120 are in an off state.

* (ii) Heat switch 110 is turned on, i.e. it conducts heat. Switches ioo and 120 are off. Heat IS transferred from 210 to 200.

* (iii) 210 is depolarised. It cools down by the electrocaloric effect and its temperature is now lower than that of the heat source 20. 200 is polarised. It heats up and its temperature becomes higher than that of the heat sink 10. All switches are in an off state.

(iv) Switches 120 and 100 are on, switch 110 is off, heat transfers from the source 20 to element 210 and from 200 to the ambient 10.

The cycle described can be generalised and used in embodiments with cascades comprising any number of electrocaloric elements. However, the described cycle is just an example and other cycles can be run on devices proposed in the present invention.

Another embodiment of the proposed device, a cascaded system comprising four electrocalorjc elements: 200, 210, 220, 230 and five heat switches: 100, 110, 120, 130, 140 is presented in Fig. 10. Electrocaloric element 210 is utilised as a heat source in the previous stage comprising the electrocaloric element 200. Electrocaloric element 220 is utilised as a heat source in the stage comprising electrocaloric elements 200 and 210.

Electrocaloric element 230 is utilised as a heat source in the stage comprising electrocaloric elements 200, 210, and 220. In other words, the system comprises four cooling devices: a first device comprising an electrocaloric element 200 and heat switches 100 and 110, a second device comprising an electrocaloric element 210 and heat switches iio and 120, a third device comprising an electrocaloric element 220 and heat switches 120 and 130, and a fourth device comprising an electrocaloric element 230 and heat switches and 140. The hot end of a fourth device is the cold end of the third device, the hot end of the third device is the cold end of the second device, and the hot end of the second device is the cold end of the first device.

Referring to Fig. 10, the end working element 200 has the highest working temperature of all elements and is separated from the heat sink 10 by the heat switch 100. The end working element 230 has the lowest working temperature of all elements and is separated from the heat source 20 by the heat switch 140. As an example, the working elements 200, 210, 220, and 230 of the present invention can have the working temperatures Tw200, Tw210, Tw220, and Tw230 shown in Table 1.

Table 1. An example of working temperatures of a cascaded device comprising four electrocaloric elements in C.

It should be understood that the number of cooling devices in a cascaded or parallel system is variable and a cascade comprising four electrocaloric elements is shown for example only. However, as the following analysis shows, devices comprising an even number of EC elements are in general more efficient and therefore preferable. it is assumed that the energy of a discharge after a deporarization of an EC element is kept in the circuit and used afterwards Due to the periodic character of the device operation, it can be designed rather easily as an RC circuit. Electrical circuitry is not depicted in Figs. 8, 10, 14 for the sake of clarity.

Some aspects concerning the efficiency and cooling power of the proposed device Device efficiency and cooling power depend on the time required to fulfil each stage of a refrigeration cycle. The time of polarisation/depolarisation of an EC element rD0, is rather difficult to estimate, it depends on the electrical capacitance of an EC element c8 and on the circuit resistivity p: = PCE,. In theory it could be made very small by reducing p at a given value of CE,. But in practice there is a minimum value of p at which the ferroelectric material does not break down on the application of voltage. So, this time cannot be adjusted easily. A reasonable experimental value for, is about 0.01 - 0.1 s [12].

The time of a heat exchange process rcan be varied by changing the heat switches' geometry. However, the faster the heat exchange, the more power is required to offset the heat conductivity between the elements when the heat switches are in an off state. A much more promising approach for reducing the heat exchange time is applying a reverse current to the Peltier elements, and this idea is explored later. For the sake of clarity and in order to consider the working principle of the device, let be about1 s until the end of the next paragraph. The time required for a heat exchange between an end element (200 or 230 in Fig. 10, for example) and a heat sink or heat load (10 or 20 in Fig. 10) is 2r if neither heat sink nor heat load change their temperature upon the heat exchange.

Referring to Fig. 9, each stage (I, ii, iii and iv) comprises either fast polarisation/depolarisation processes or slow heat conduction processes, but never both at the same time. Moreover, each heat exchange process is either between 200 and 210 or between 200 and 10 or 210 and 20. The former takes r seconds, the latter two take 2r seconds, and the polarisation takes negligible time comparing to z That is why it is preferable to use even numbers of electrocaloric elements in the proposed device - otherwise each of the four stages would involve heat transfer to the heat sink/heat load, thus slowing down the whole cycle and reducing the cooling power. If r is close to the polarisation time then it changes nothing in the principle of operation and this discussion, but the time notation in Fig. 9 (i 3z 4z. . . ) will be different. It should be understood that embodiments comprising an odd number of electrocaloric elements can also be used.

Moreover, it is possible to reduce the cycle period even more. Stage ii (see Fig. 9) can be started before stage i is finished (the same holds with stages iv and ii respectively). As soon as temperatures of 200 and 210 are equal at some moment during stage heat switch ii can be turned on to start the heat conduction. Also, applying this concept will allow a thermodynamic process to run within the working elements that is close to a Carnot cycle. A working thermodynamic cycle of an electrocaloric element is depicted in Fig. 11 using the variables Tand S, where Tis temperature and S is entropy.

If stage ii (and iv) start after stages I (and iii respectively) finish, then an element passes through the points 1 -2-3-4' of the diagram. But if stage ii (or iv) starts after point 2 (or 4) on the diagram, then the cycle will be closer to the ideal Carnot cycle 1-2-3-4. This approach will allow an increase in the inherent efficiency of the energy transfer inside the electrocaloric elements.

Examples of such energy interconversions are: electrical energy at the eiectrodes - energy of the ferroelectric, and structural ordering heat energy of the the crystal lattice.

In some embodiments, the time required for heat exchange between two adjacent EC elements r can be decreased by using an extra heat conductor parallel to the heat switch. It effectively increases the heat switch's heat conductivity and lowers its thermoelectric figure of merit. In some embodiments, these shunts are microactuators fabricated by microelectromechaflical systems technology. One of the possible designs for additional heat conductors is shown in Fig. 12, a. The thermoelectric elements of the heat switch between 200 and 210 are not shown in that Figure for the sake of clarity. Only two electrocaloric elements are shown in the Figure for the sake of clarity. However, this concept can be applied to any number of working elements. Microactuators can be made of a piezoelectric material, e.g. PZT - lead zirconium titanate. They can be either compressed (voltage applied), so that the device does not conduct heat, or stretched (no voltage applied) so that there is a physical contact between the two parts to allow heat transfer. The efficiency of such a device depends on the contact surface area SComp. It is compared to the plain surface area Spjajn in Fig. 12, b. The larger the surface area, the more efficient the microelectromechanical heat switch.

Three geometries (ladder - 61, conical - 62, and spherical - 63) for the heat transfer elements are shown in Fig. 12, b along with their effective contact surface area normalised to the plain surface area. The size of such a device can vary, but for the sizes considered in the current research, a reasonable value is 10-100 jtm in both diameter and depth. Although conical (62) and spherical (63) shapes could be hard to produce, the ladder type projection (61) can be produced by a conventional microfabrication technique depicted in Fig. 13. In some embodiments, such a device can be used as a mechanical heat switch between the working elements and end working elements and heat sink/load.

Referring to Fig. 13, a preferable way to manufacture a mechanical heat switch is presented. For those acquainted with the art, the procedure is commonplace In the first stage I, a layer of photoresjst is deposited on top of Si or other heat conducting material. A part of its surface 30 is exposed to the UV radiation. Then the exposed surface is etched away by diluted acetic acid, for example, so that a trench is formed in the photoresjst. Then at stage II a strong etchant (e.g. HF) is applied to the surface, It removes the remaining photoresist and makes trench 31 in Si (or other heat conducting material used). At stage Ill the photoresist is deposited again and a part of its surface 32 is exposed to the UV light, and similarly to step II, a strong etchant is applied at step IV to make a next trench. Finally a part of a complex surface 61 is obtained. Only one trench is shown in Fig. 13, however such a procedure can be realised to produce a number of trenches on the same surface.

In some embodiments, the work of a heat switch based on a thermoelectric element is modified in the following way. A reverse voltage is applied across the element when it is in the on state (that is no current, heat transfer goes on - for example switch 110 at the stage ii in Fig. 9). In this case the Peltier effect will effectively pump heat from the hot to the cold electrocaloric element that it connects (for example from 210 to 200 in Fig. 9).

In other words, the application of a voltage with reverse polarity increases the effective heat conductivity of the heat switch, decreases the time required for the heat transfer and increases the cooling power and efficiency. To find out the influence on the cooling efficiency, a couple of equations for the heat transfer can be considered (for more details see the Appendix). One of the main results of the calculations is that even a rather small current (0.05 A, for example) decreases r by several orders of magnitude. This method will be referred to as the active heat conductivity method, Example 1: Calculation of efficiency and cooling power of the proposed device in some embodiments The operation of some embodiments of the proposed cooling device with thermoelectric heat switches has been simulated using MS Excel . The electrocajoric effect of each electrocaloric element versus temperature was described by a Gaussian curve with a maximum of 3 C. The peaks of the Gaussiaps are centered at the appropriate working temperatures for each element Tw, and have a width of about 10 C. Embodiments of cascades comprising 2, 4, and 6 electrocaloric elements have been simulated. The working temperatures (Tw0, T10 for a cascaded device comprising 2 electrocaloric elements, Tw0, T10, Tw20, T30 for a cascaded device comprising 4 electrocalorjc elements, and Tw200, T10, Tw0, T30, TW24(J, T50 for a cascaded device comprising 6 electrocaloric elements) used in the simulation are listed in Table 2. It should be understood that the present invention is not limited to these temperatures. Any temperatures may be used.

An embodiment of the proposed device Comprising a Cascaded system of 6 electrocajoric elements and 7 heat Switches is shown in Fig. 14. The system comprises additional electrocaloric elements that are utilised as heat sinks/sources in the previous stages of the cascaded system. This embodiment comprises 6 cooling devices that are cascaded in analogy with cascaded systems described above (shown in Figs. 8 and 10). As an example, the initial temperature of the heat load 20 is 16 C for all devices, the initial temperature of the heat sink 10 is 20 C for the cascaded device comprising 2 electrocal6ric elements, 24 C for the cascaded device comprising 4 electrocajoric elements, and 28 C for the cascaded device comprising 6 electrocaloric elements.

Tw250, oc Tw240, C TW230, OC Tw220, oc Tw210, oc TW200, oc 17 19 21 23 25 27 Table 2. Working temperatures of some embodiments of cascaded cooling devices comprising various numbers of electrocaloric elements in C.

Other important characteristics of the model are listed in Table 3. The efficiency of the proposed device depends on how fast the heat is dissipated from a heat sink (component io in Figs. 7, 8, 10, and 14, for example). The rate depends mainly on the design of the heat sink. Usually it is a structure with a large surface area and a fan that creates the air flow through it. As a rule, applied heat (that is to be dissipated) increases the temperature of a heat sink. As a first approximation, the relation between the power to be dissipated W and temperature change 6T of the heat sink is linear, and the coefficient is called the heat resistivity: PHeat = ST/W A typical value for a metallic plate without a fan is PHeat = 0.3 CW* However, in the calculations presented below, the heat resistivity is assumed to be much less (O.001-O.05 CW-1) This simplification allows us to study the intrinsic thermodynamic properties of some embodiments of the proposed device. The efficiency and cooling power obtained from the simulations are compared with the parameters of a commercial Peltjer cooler with the heat resistivity of the same order of magnitude: 0.04 CW1.

2 EC 4 EC 6 EC I _____ elements elements elements aijstance of a heat sink 10, C W1 0.05 0.01 0.0001 rmber of couples per one heat switch 540 ightofacouplem 10 fitio of a couple's cross section to height, 1 cm Total couple cross section, mm2 10.3 * 10.3 of each ECE, mm3 30 * 30 * 2 Table 3. The parameters of cooling devices used in the calculation.

The simulation results for a cascaded device comprising 2 electrocalorie elements at equilibrium (after 600 cycles have passed) are shown in Fig. 15. Four stages of the cooling cycle (I, i ii and iv) are highlighted (they are the same as in Fig. 9). The temperature of the heat load (component 20 in Figs. 7, 8, 10, and 14, for example) is raised artificially at the end of stage ii of every cycle simulating the heat generation in a real device. The simulated temperature of the heat sink (component io in Figs. 7, 8, 10, and 14) is higher than that of the ambient because of its finite heat resistivity (its temperature should rise in order to dissipate the incoming heat).

It should be understood that other cycles can be run in the proposed device, and this particular cycle is simulated for illustration only. Increasing the heat conductivity of heat switches while all other parameters are held constant produces a visible effect on the overall cooling performance. Fig. 16 shows some important characteristics (for a cascaded device comprising 4 electrocajoric elements shown in Fig. 10), namely the cooling power, efficiency and the average time r for heat transfer between two adjacent electrocaloric elements (200 and 210, 210 and 220, 220 and 230 in Fig. 10) as a function of the heat conductivity of the switches. So the approach of changing the heat switches' heat conductivity when the device is working can give more options and allows for the optimisation of cooling power versus efficiency. That is, additional heat conductors provide an adjustable option for the design of a particular cooling device so that it could be used either to pump more heat less efficiently or to work at the highest efficiency but with lower cooling power. In some embodiments, additional shunts made of MEMS described above can fulfil the function to change the effective heat conductivity of the heat switches.

Example 2: Calculation of efficiency and cooling power of some embodiments of the proposed device comprising heat switches in active heat conductivity mode Embodiments comprising heat switches in active heat conductivity mode are more preferable as the calculation gives more striking results. Some of them are presented in Figs. 17 and 18. Referring to Fig. 17, the efficiency and cooling power of the proposed cooling device versus a difference between heat sink and heat load temperatures, T10 - T20, are presented for some embodiments: a cascaded device comprising 2 electrocaloric elements, a cascaded device comprising 4 electrocaloric elements, and a cascaded device comprising 6 electrocaloric elements at different values of the reversed current passed through heat switches in active heat conductivity mode. The efficiency is expressed in percentage of the ideal Carnot efficiency given by formula (2). Application of quite a small reverse current (when the heat switches are in an on state, i.e. Conducting heat) significantly improves the device performance The device parameters used in the calculation are listed in Table 3. Referring to Fig. 18, efficiency and cooling power versus the reverse current of the same devices are presented. It should be understood that the parameters used are only for the sample calculation, as well as geometric sizes of the electrocaloric elements and heat switches. Any other set of parameters and sizes can be used in the proposed cooling device.

Moreover electrocajoric elements can have different sizes, preferably the size of the working element which is closest to the heat sink should be biggest to account for the heat losses in the previous electrocaloric elements and heat switches.

A comparison of a commercial fully optimised thermoelectric device with an embodiment of the electrocaloric cooling device proposed here is made in Fig. 19. The data for thermoelectric cooling device is taken from [19] (device reference number P18-7-30 with G 0.17 cm, N 71, PHeat- 0.04 Cw-1 the temperature difference between its cold and hot ends is about 6 K). It should be understood that the comparison has been made with just one thermoelectric solution taken from [19]. Other solutions provided by this source can have different parameters. The heat resistance of the thermoelectric cooler is chosen to be very low (close to that of the cooling device suggested): 0.04 CW1. As seen from Fig. 19, the efficiency of the considered embodiment of the present invention rises with the cooling power, unlike for the thermoelectric cooler. A possible reason is as follows. In order to provide more cooling power, the current through the thermoelectric heat switch should be increased. However, this also increases Joule heating, which is proportional to the current squared. As a result, the efficiency falls. In the case of the considered embodiment, if more heat is pumped through the it, the temperature of the cold surface rises, but the energy consumption remains the same. If the heat dissipation at the hot end is fast enough, the efficiency rises with cooling power.

Appendix: Active heat conductivity mode in more detail Some facts are presented below to clarify the idea of the active heat conductivity mode. The equations for the heat transfer between two electrocaloric elements interspaced with a thermoelectric heat switch are ùflsdered here (for instance, consider two electrocaloric elements 200 and 210 interspaced with a heat switch 110 in Fig. 8, with the rest of the Figure omitted). The process can be described by two differential equations: Cd 2N1T(./d ++kG(, - t (Al) = 2N1_ T110, + - kG - and the boundary conditions are: J Coid L0 = TcOId (0), (A.2) l.T1101L0 = where N is the number of couples, C (JK1) is the heat capacity of an electrocaloric element, TcO,d (K) is the temperature of an electrocaloric (EC) element with lower initial temperature TcO/d (0)(K), THOt (K) is the temperature of another EC element with higher initial temperature TcO,d (0)(K), 1(A) is the current passed through the thermoelectric heat switch, a is the Seebeck coefficient (VK1), p is the resistivity (acm), k is the heat conductivity (Wcm1K1), G (cm) is the ratio of a heat switches' cross-sectional area to its height. These equations are general and include the "reverse" current in thermocouples that accelerate the heat exchange process. However, if we assume I 0, these equations will describe the passive heat transfer process between two electrocaloric elements.

A typical size of an electrocaloric element used in the calculations was * 30 * 2 mm3 with a total heat capacity of C = 4.86 JK1. Initially, one of the electrocaloric elements was at TcO/d(O) = 300 K, and the other was at THOt(0) 305 K. Let 540 thermoelectric couples made out of bismuth telluricie, which are connected electrically in series and thermally in parallel (see Fig. 4) have a height of 10 tim, and let the total cross section correspond to G= 1 cm: 10.3 * 10.3 mm2 (small enough to fit between the electrocaloric elements). Then two electrocaloric elements will exchange heat for about - 0.4 s, which is close to the time constant of the EC element polarisation.

The total heat capacity of the couples is 0.44 JK1, around ten times smaller than that of the electrocaloric elements. Let us now apply a reverse voltage to the couples in order to reduce z- . The result is depicted in Fig. 20, I for currents of 0, 0.05 and 0.2 A. As can be seen from the Figure, the application of even a 0.05 A current is enough to influence Larger current of 0.2 A makes a larger effect. The Joule heating is small for both currents on the timescale of one second. At a larger timescale, the Joule heating becomes important and both temperatures rise linearly versus time (see Fig. 20, II). The difference between temperatures in the linear region is determined by the competition of the Peltier heat transfer and heat conductance.

The solution for a passive heat exchange (at I = 0 A) can be deduced from (A.1) and (A.2). The temperature difference THOt - TcO/d decays exponentially with time: THOt - TcO,d exp (-t/r), where r=4NkG/C. (A.3) This formula is used to estimate presented in Fig. 16.

Figure captions FIG. 1. A schematic of an electrocajoric element.

FIG. 2. Results of the electrocaloric effect measurement of four different samples of PbSc0 5Ta0 503 used in the prototype shown in Fig. 5. The Curie point, taken as the temperature of the maximum of ECE, can be moved, which allows a much wider temperature interval to be covered than with a single element. After [6].

FIG. 3. Results of the electrocaloric effect measurement of Pb0 99Nb0 02(Zr0 5Sn0 2Tio.05)0 9803. After [20].

FIG. 4. A thermoelectric couple.

FIG. 5. A schematic of an electrocaloric fridge prototype. Some characteristics of the prototype are shown in Fig. 6. After [12].

FIG. 6. Experimental results obtained with the prototype shown in Fig. 5, see text for explanation. After [12].

FIG. 7. Schematic design of one embodiment of the proposed cooling device comprising one electrocaloric element 200 and two heat switches 100 and 110. Electronic circuits that supply the electrocaloric element and heat switches are not shown here for the sake of clarity.

FIG. 8. Schematic design of one embodiment of the proposed cooling device, a cascaded system comprising 2 electrocaloric elements 200 and 210; 3 heat switches 100, 110 and 120; and electronic circuits that supply the electrocaloric elements and heat switches. The circuits are not shown here for the sake of clarity.

FIG. 9. The working cycles of the cooling sandwich shown in Fig. 8. Another embodiments of the proposed device (e.g. cascaded devices comprising 4, 6, etc. electrocaloric elements) can use working cycles of the same principle.

Heat exchange between 200 and 210 takes rseconds and between 200 and (210 and 20) - 2'r seconds. A label ATI stands for a polarisation of an electrocaloric element (either 200 or 210) that increases its temperature by AT via the electrocaloric effect and a blue label AT - a depolarisation of an electrocaloric element (either 200 or 210) accompanied by its cooling with AT See also Fig. 15, later.

FIG. 10. Schematic design of another embodiment of the proposed cooling device, a cascaded system comprising 4 electrocaloric elements 200, 210, 220 and 230; 5 heat switches 100, 110, 120, 130, and 140; and electronic circuits that supply the electrocaloric elements and heat switches. The circuits are not shown here for the sake of clarity.

FIG. 11. Possible thermodynamic cooling cycle of an electrocaloric element (e.g. 200, 210 in Fig. 8).

FIG. 12. A possible design of a micromechanjcal heat switch which can be fabricated with a method presented in Fig. 13.

FIG. 13. Microfabrication of a ladder microniechanical heat switch shown in Fig. 12.

FIG. 14. Schematic design of still another embodiment of the proposed cooling device, a cascaded system comprising 6 electrocaloric elements 200, 210, 220, 230, 240, and 250; 7 heat switches 100, 110, 120, 130, 140, 150, and 160; and electronic circuits that supply the electrocaloric elements and heat switches. The circuits are not shown here for the sake of clarity.

FIG. 15. Results of an example calculation of a possible cooling cycle of one embodiment of the proposed device - the one comprising 2 electrocaloric elements, shown in Fig. 8. The temperatures of the heat source 20 (T20), EC elements 200, 210 (T200, T210), and the heat sink 10 (T10) are shown. The picture represents equilibrium - 600 cycles have been simulated before these curves are taken. Four stages of a possible refrigeration cycle are highlighted (see Fig. 9). The electrocaloric effect is taken to be the same in both elements: AT1 AT2.

FIG. 16. Efficiency and cooling power of one embodiment of the proposed device comprising 4 electrocaloric elements (shown in Fig. 10) versus heat conductivity of the switches. is the time for heat exchange between any two electrocaloric elements (A.3).

FIG. 17. Efficiency and cooling power of some embodiments of the proposed device (shown in Figs. 7, 8, 10 and 14): cascaded devices comprising 2, 4, and 6 electrocaloric elements in passive (the reverse current in heat switches is IRev- 0 A) and active (IRev= 0.05, 0.1 and 0.2 A) heat conducting modes.

The influence of the reverse current on the performance shown in Fig. 18 is calculated at the highlighted temperature differences: 3 K for a cascaded device comprising 2 electrocaloric elements, 6 K for a cascaded device comprising 4 electrocaloric elements, and 10 K for a cascaded device comprising 6 electrocaloric elements.

FIG. 18. The influence of the heat switches' reverse current on the performance of some embodiments of the cooling device: cascaded devices comprising 2, 4, and 6 electrocaloric elements. The particular current values can be reduced by increasing the number of couples N and decreasing the geometrical factor G - a standard approach in design of thermoelectric devices. It should be understood that the values used in the calculations are not fixed.

FIG. 19. Estimated efficiency of some embodiments of the proposed device: cascades comprising 2, 4, and 6 electrocaloric elements and a commercially available thermoelectric cooler. The devices operate in both passive (Ret' 0 A) and active (IRv = 0.05, 0.1 and 0.2 A) heat conducting modes. The average temperature difference between the hot and the cold ends is about 6-8 C for all curves. Data for the Melcor device is taken from [19] (a device No PT8-7-30 with G= 0.17 cm, N= 71, heat resistance rHeaf= 0. 04 Cw1).

FIG. 20. I. Temperature of an electrocaloric element with higher initial temperature (THOf) and an electrocaloric element with lower initial temperature (TcO/d) interspaced with a thermoelectric device with a current of 0, 0. 05 and 0.2 A versus time. The current is applied at t = 0 s, when THOt = 305 K and TcO,d = 300 K. JI. As above for a current of 0.2 A. Joule heating linear with time manifests on a 10 s timescale.

FIG. 21. Schematic design of an embodiment of a parallel configured device for transferring heat from a heat source 20 to a heat sink 10.

SUMMARY

We describe a device for transferring heat from a heat source 20 to a heat sink 10, see Fig. 20. The device comprises at least two heat switches and 110, the former to receive heat from the heat source 20, the latter to pass heat to the heat sink 10; at least one electrocaloric (EC) element 200 between the heat switches 100 and 110 and in thermal contact with the heat switches; and a control input for controlling 100, 110, and 200. One or both of the heat switches comprise a thermoelectric heat switch, tunnelling heat switch or an electromechanical heat switch.

Preferably, a device described above is a part of a cascaded system having a plurality of the devices for transferring heat from a cold end to a hot end. Some embodiments of a cascaded system are presented in Figs. 7, 9, and 13. Referring to Fig. 7, a cascaded system comprising two electrocaloric elements 200 and 210 and three heat switches 100, 110 and 120 is presented. An electrocaloric element 210 of an additional stage is utilised as a heat source in the previous stage of the cascaded system. In other words, the embodiment of a cascaded system shown in Fig.8 comprises two cooling devices: a first device comprising an electrocaloric element 200 and heat switches 100 and 110 and a second device comprising an electrocaloric element 210 and heat switches 110 and 120.

Referring to Fig. 9, a cascaded system comprising four electrocaloric elements 200, 210, 220, and 230 and five heat switches 100, 110, 120, 130, and 140 is presented. Electrocaloric element 210 is utilised as a heat source in the previous stage comprising the electrocaloric element 200. Electrocaloric element 220 is utilised as a heat source in the stage comprising electrocaloric elements 200 and 210. Electrocaloric element 230 is utilised as a heat source in the stage comprising electrocaloric elements 200, 210, and 220. In other words, the system comprises four cooling devices: a first device comprising an electrocaloric element 200 and heat switches 100 and 110, a second device comprising an electrocaloric element 210 and heat switches 110 and 120, a third device comprising an electrocalorjc element 220 and heat switches 120 and 130, and a fourth device comprising an electrocaloric element 230 and heat switches 130 and 140. The hot end of a fourth device is the cold end of the third device, the hot end of the third device is the cold end of the second device, and the hot end of the second device is the cold end of the first device.

Referring to Fig. 13, a cascaded system comprising six electrocaloric elements 200, 210, 220, 230, 240, and 250 and seven heat switches 100, 110, 120, 130, 140, 150, and 160 is presented. It should be understood that cascaded systems comprising any other number of electrocaloric elements and heat switches can be used.

Another embodiment of the present invention, a parallel configured device for transferring heat from a heat source 20 to a heat sink 10, is presented in Fig. 21. A plurality of parallel positioned stacks 500, 510, 520 and 530 transfer heat from a heat source 20 to a heat sink 10. The number of stacks can be different. Each stage in each stack 500, 510, 520 and 530 comprises an electrocaloric cooling system presented in one or several embodiments of the present invention. Other configurations and orientations of the elements shown in Fig. 21 may be used.

An off state of heat switches comprises applying a voltage in a first, forward direction to the switches, and an on state of the switches comprises applying a voltage in a second, reverse direction to the forward direction to the switches. However, the heat switches do not need to fully reject heat in an off state.

A device proposed in the present invention also comprises a control input that comprises one or more electrical connections, and means to control heat switches and electrocaloric elements (e.g. a computer program).

We also describe a method of controlling a heat transfer device for transferring heat from a heat source to a heat sink. As an example, a method is considered for an embodiment comprising one electrocaloric element 200 and two heat switches 100 and 110, see Fig. 20. The method comprises: cooling electrocaloric element 200 to a temperature of lessthan source or equal to the source 20; turning the first heat switch 110 on to transfer heat from the source 20 to the electrocalorjc element 200; turning the first heat switch 110 off; warming the electrocaloric element 200 to a temperature of greater than the sink 10 or equal to the sink 10; turning the second heat switch 100 on to transfer heat from the electrocaloric element 200 to the sink 10; and turning the second heat switch 100 off.

Turning the heat switches 100 and 110 off comprises applying a voltage in a first, forward direction to the switches, and turning the switches 100 and 110 on comprises applying a voltage in a second, reverse direction to said forward direction to a said switch.

We also describe a method of controlling a cascaded system having a plurality of devices for transferring heat from a cold end to a hot end and wherein the method is repeated for each device for transferring heat from a heat source to a heat sink in the cascaded system. Also, we disclose a method of controlling a parraleJ and stacked cascaded system having a plurality of devices for transferring heat from a cold end to a hot end wherein the method is repeated for each device for transferring heat from a heat source to a heat sink in said systems.

A device proposed in the present invention also comprises a controller configured to implement methods explained above, and a carrier to carry a computer program code (e.g. host or disk) to implement methods explained above.

References 1. www.intel.com/labs 2. M. Togashi. "Multilayer ceramic electronic device", US Patent No 2003/0026059, 2003.

3. D.A. Barrow, T.E. Petroff, and M. Sayer. "Method for producing thick ceramic films by a sol gel coating process", US Patent No 5585136, 1996.

4. E.H.Birks L.A.Shebanov, K.J.Borman, and A.R.Sternberg. Ferroelectrics. 94: p. 305, 1989.

5. E.K.H. Birks, L.A. Shebanov, K.Y.A. Bormanis, and M. Dambekalne. "Lead scanotantalate as active e]ement of microcryogenic systems", SU Patent No 1479440, 1989.

6. Y.V. Sinyavsky, N.D. Pashkov, Y.M. Gorovoi, G.E. Lugansky, and L.A.

Shebanov. Ferroelectrjcs. 90: p. 213, 1989.

7. B.A. Tuttle. "Polarization reversal and electrocalorjc measurements for field- enforced transitions in the system lead zirconate-lead titanate-lead oxide:tin oxide", PhD Thesis at University of Illinois at Urbana- Champaign, 1982.

8. L. Shebanovs, K. Borman, W.N. Lawless, and Kalvane A. Ferroelectrics. 273: p. 2515, 2002.

9. Y.V. Sinyavsky, N.D. Pashkov, Y.M. Gorovoi, G.E. Lugansky, and Shebanov L. Ferroelectrics. 90: p. 213, 1980.

10. U.S. Ghoshal. "Apparatus and methods for performing switching in magnetic refrigeration systems using thermoelectric switches", US Patent No 6595004, 2003.

11. V.M. Brodyansky, Yu.V. Sinyavsky, and N.D. Pashkov. "Thermal switch (its versions)", SU Patent No 918770, 1982.

12. Y.V. Sinyavsky and V.M. Brodyansky. Ferroelectrics. 131: p.321, 1992.

13. A.E. Romanov, Ju.V. Sinyavskij, N.D. Pashkov, and G.E. Luganskij.

"Method of cooling and device for realization of this method", SU Patent No 2075015, 1993.

14. L.W. Lawless. "Paraelectric refrigeration method and apparatus", US Patent No 3436924, 1969.

15. LW. Lawless. "Closed-cycle electrocaloric refrigerator and method", US Patent No 3638440, 1972.

16. A. Basiulis and B.L. Robert. "Solid-state electrocaloric cooling system and method", US Patent No 4757688, 1986.

17. A. Tavkhelidze, L. Koptonashvili, Z. Berishvili, and 0. Skhiladze. "Method for making a diode device", US Patent No 6417060, 2002.

18. J.-P. Fleurial, M.A. Ryan, A. Borshchevsky, W. Phillips, E.A. Kolawa, G.J.

Snyder, T. Caillat, T. Kascich, and P. Mueller. "Microfabricated thermoelectric power-generation devices", US Patent No 63881 85, 2002.

19. www.me1cor.com 20. B.A. Tuttle and D.A. Payne. Ferroelectrjcs. 37: p. 603, 1981.

Claims

1. A device for transferring heat from a heat source to a heat sink, the device comprising: at least two heat switches, a first to receive heat from the heat source, a second to pass heat to the heat sink; at least one electrocaloric element between two said heat switches and in thermal contact with said two heat switches; and a control input for controlling said at least two heat switches and said electrocaloric element to transfer heat from said source to said sink; and wherein one or both of said heat switches comprise a thermoelectric heat switch, tunnelling heat switch or an electromechanical heat switch.

2. A device as claimed in claim 1, wherein said device for transferring heat from a heat source to a heat sink is a part of a cascaded system having a plurality of devices for transferring heat from a heat source to a heat sink.

3. A device as claimed in claim 2, wherein said device for transferring heat from a heat source to a heat sink is a stacked cascaded system.

4. A device as claimed in claim 2, wherein said device for transferring heat from a heat source to a heat sink comprises a parallel system.

5. A device as claimed in claim 1, 2, 3 or 4 wherein said control input comprises one or more electrical connections.

6. A device as claimed in claim 1, 2, 3 or 4 further comprising a controller to control said heat switches and said at least one electrocaloric element, the controller comprising control means for: cooling said electrocaloric element to a temperature of less than said source or equal to said source; turning said first heat switch on to transfer heat from said source to said electrocaloric element; turning said first heat switch off; warming said electrocalorjc element to a temperature of greater than said sink or equal to said sink; turning said second heat switch on to transfer heat from said electrocaloric element to said sink; and turning said second heat switch off.

7. A method of making a device for transferring heat from a heat source to a heat sink comprising: providing at least two heat switches, a first to receive heat from the heat source, a second to pass heat to the heat sink; providing at least one electrocaloric element between two said heat switches and in thermal contact with said two heat switches; and providing a control input for controlling said at least two heat switches and said electrocaloric element to transfer heat from said source to said sink; and wherein providing one or both of said heat switches comprise providing a thermoelectric heat switch, tunnelling heat switch or an electromechanical heat switch.

8. A method as claimed in claim 7, wherein said device for transferring heat from a heat source to a heat sink is provided as part of a cascaded system having a plurality of devices for transferring heat from a heat source to a heat sink.

9. A method as claimed in claim 8, wherein said cascaded system is a stacked cascaded system.

10. A method as claimed in claim 8, wherein said system also comprises a parallel system.

11. A device as claimed in claim 6 wherein turning said heat switches off comprises applying a voltage in a first, forward direction to said switches; and wherein turning said switches on comprises applying a voltage in a second, reverse direction to said forward direction to a said switch.

12. A method of controlling a heat transfer device for transferring heat from a heat source to a heat sink, the heat transfer device comprising at least one electrocaloric element disposed between a pair of thermoelectric heat switches, a first switch being in thermal contact with said source and a second switch with said sink, the method comprising: cooling said electrocaloric element to a temperature of less than said source or equal to said source; turning said first heat switch on to transfer heat from said source to said electrocalorjc element; turning said first heat switch off; warming said electrocaloric element to a temperature of greater than said sink or equal to said sink; turning said second heat switch on to transfer heat from said electrocaloric element to said sink; and turning said second heat switch off; and wherein turning said heat switches off comprises applying a voltage in a first, forward direction to said switches; and wherein turning said switches on comprises applying a voltage in a second, reverse direction to said forward direction to a said switch.

13. A method as claimed in claim 12, wherein said device for transferring heat from a heat source to a heat sink is part of a cascaded system having a plurality of devices for transferring heat from a heat source to a heat sink and wherein the method is repeated for each device for transferring heat from a heat source to a heat sink in said cascaded systern

14. A method as claimed in claim 13, wherein said cascaded system is a stacked cascaded system.

15. A method as claimed in claim 14, wherein said system also comprises a parallel system.

16. A controfler configured to implement the method of claim 12.

17. A carrier carrying computer program code to implement the method of claim 12.