GB2551755B - A method of manufacturing granules having a layer of metallic particles surrounding a salt core - Google Patents

Description

A METHOD OF MANUFACTURING GRANULES HAVING A LAYER OF METAL PARTICLES SURROUNDING A SALT CORE

The present invention relates to a method of manufacturing granules having a layer of metal particles surrounding a salt core.

BACKGROUND TO THE INVENTION

Metallic materials containing internal voids are commonly known as metal foams. These materials are typically very lightweight due to their porosity. Other names for such materials include cellular metal, porous metal, metallic foam and metallic sponge (or metal sponge). These types are not mutually exclusive, but the names can indicate whether the structure of the material contains open- or closed-cell voids, and whether the voids are interlinked as an open network, for example.

Examples of potential applications of metal foams are diverse and include thermal management, catalysis, electrochemistry, chemical filtering, energy storage, acoustic damping, electromagnetic shielding and for use as impact absorption materials.

One method of manufacturing an open cell foam is by powder metallurgy, where metal powder is compressed with a ‘space holder’ compound and then heated to sinter the metal. The space holder compound is then extracted to leave the metal as a foam.

An example of an existing method is disclosed in EP 1755809. Metallic particles are mixed with a carbonate additive and a liquid binder (selected to avoid reaction with the carbonate additive) to form a mixture. The preferred compounds are copper powder, potassium carbonate (K2CO3) and ethanol. The mixture is compressed and then heated to evaporate the binder and sinter the metal. Ethanol has a low boiling point (c. 78°C) which makes for simple removal by evaporation. The carbonate is then removed to leave a metal foam, which is re-heated for improved sintering of the finished product.

However, whilst the method in EP 1755809 is adequate for a research environment, it gives rise to a number of problems when scaled up for industrial use. The mixture of metal powder, carbonate and binder must be prepared in situ and used immediately, since the ethanol binder will evaporate relatively quickly at room temperature. The mixture is unsuitable for storage in excess of one day, for similar reasons. The flow characteristics of the mixture are also unsuitable for use in an automated die-filling process for compaction prior to heating, since it is wet with ethanol and the individual particles stick together.

Furthermore, the preferred carbonate (K2CO3) tends to decompose above around 890°C, which lies in the sintering temperature range for metals used in the process. Around this temperature, K2CO3 releases CO2, with the remainder quickly decomposing further into oxygen and elemental potassium during vacuum sintering. Elemental potassium is pyrophoric in air at the temperatures involved, posing a safety risk and damaging the furnace used in production. It also reacts with water to produce highly caustic potassium hydroxide, posing another safety risk and causing further damage to the furnace. Safe storage and disposal of the potassium is costly and hazardous. The resulting pore structure is irregular due to the carbon dioxide evolved during sintering. This process is only suitable for low volume, immediate, manually-controlled production of metal foams.

Research for alternatives to K2CO3 included trials using NaCI, CaCCb and KI as spaceholder salts, but only KI gave worthwhile results. However, the resulting metallic foam using KI had an irregular pore structure due to the irregularity of the KI crystals. Additionally, the precursor material including KI had poor mechanical flow properties. Tests using SrCCb were also performed but its insolubility in water made it impractical to remove the salt.

The above methods provide poor control over the properties of the resulting metal foam, and give rise to problems during industrial-scale production and automated throughput. It is an object of the present invention to reduce or substantially obviate the aforementioned problems.

STATEMENT OF INVENTION

According to a first aspect of the present invention, there is provided a method of manufacturing granules having a layer of metal particles surrounding a salt core, the method comprising the steps of a) forming a mixture from metal particles, a particulate silicate or metasilicate salt, a binding agent for binding the metal particles to the salt, and a lubricating agent; b) adding a solvent for homogenising the mixture, to form a slurry; c) mixing the slurry for homogeneity; and d) continuing to mix the slurry to allow the solvent to evaporate until the mixture is sufficiently dry that the granules form during mixing.

Optional features of the method are set out in dependent claims 2 to 10.

The mixture dries during mixing because the solvent evaporates. The granules are dry and relatively air stable, which are both advantageous properties for granules used as precursors in the industrial preparation of porous metallic materials. Slight spillage of the granules is unlikely to damage any industrial equipment or create a hazard that requires immediate clean-up. Also, the granules do not necessarily need to be prepared in situ for manufacturing the porous metallic material. They can be prepared in advance to manage production. The granules are free-flowing, i.e. have good bulk flow properties, so they can be poured or transferred easily between containers for storage or subsequent compression into a body.

In step (d), continuing to mix the mixture as it slowly dries leads to the formation of agglomerated granules, which eventually disperse into discrete individual granules upon further mixing. Mixing speed can be minimised at this stage to turn the granules over gently for further drying, without damaging their structures.

The method may further include the step of storing the granules in a sealed container for preventing excessive drying prior to use.

Since the granules include a stable binding agent (so they retain their structure when the solvent has been removed), they can be prepared in advance of sintering but stockpiled for later use. This is as opposed to being prepared as a slurry and immediately used to create sintered bodies. This allows production to be varied according to demand.

Preferably, air is removed from the container prior to sealing the container. This increases the longevity of the granules, by removing oxygen and water vapour which may otherwise lead to quicker eventual degradation of the granules. Preferably, the container volume is variable to substantially match the volume of the granules in the container. Tor example, a bag may be partially filled with granules and then the unfilled portion of the bag evacuated of air. This avoids compounds boiling at lower temperatures than under standard conditions, which could change the structure of the granules.

The mixture may include a preservative. The preservative may be added to the mixture before step (d). More preferably, there is an additional step between steps (c) and (d) of partially drying the mixture, and then mixing the preservative into the mixture.

The metal layer or surface of each granule may substantially comprise a monolayer. The metal particles may be alloy particles. The metal or alloy particles may include one of the following elements: aluminium, chromium, cobalt, copper, gold, iron, magnesium, molybdenum, nickel, silver, tin, titanium.

The salt may have a melting point substantially similar to or higher than the metal. The salt is a silicate or metasilicate salt. Preferably, the salt is sodium metasilicate.

The binding agent and/or the lubricating agent may be powdered. Either or both of the binding and lubricating agents may include a wax. The particle sizes of the metal and the salt may be selected to adjust the properties of the resulting granules.

The granules produced by the method may be used in a method of manufacturing a porous metallic material, the method comprising the steps of: a) providing metal granules by the method of claim 1; b) applying pressure to the metal granules, the pressure exceeding the yield strength of the metal particles in the metal granules, to compress the metal particles and form a body; c) heating the body below the melting point of the salt and the sintering temperature range of the metal, for substantially removing constituents other than the metal particles and salt from the body, including heating in: i) a first temperature range for increasing surface mobility of the lubricating and binding agents on or in the body and drawing the lubricating and binding agents to outer surfaces of the body for evaporation; ii) a second temperature range substantially greater than the first temperature range, for evaporating the lubricating and binding agents from the body; and iii) a third temperature range substantially greater than the second temperature range, for substantially decomposing residual lubricating and binding agents on or in the body; d) heating the body in a fourth temperature range substantially greater than the third temperature range, sintering the metal particles to form a sintered body; and e) removing the salt from the sintered body.

From this point, references to particular steps (a), (b), (c) etc. are to the steps of the method of manufacturing a porous metallic material.

This method allows for industrial production of a porous (or microporous) metallic foam or sponge material. The material has a highly uniform pore structure. There is a continuous network of regular well-defined pores throughout the volume of the material. This results from the substantially identical granules from which the material is formed. The degree of porosity is also customisable and predictable, according to the selection of metal and salt particle sizes used.

The metallic foam has a strong metal structure (or skeleton) with high conductivity and mechanical strength. Removing the salt from the sintered body frees (i.e. unblocks) the pores, and allows gases to flow through the foam, effectively dissipating thermal energy. The metallic foam product is suitable for use in free convective heat sink applications as a low profile heat sink, for example, as well as any of the other potential applications listed in the background to the invention.

One of the primary advantages of this method is that manufacture of the metallic foam can be substantially automated at an industrial scale. Rapid production of large numbers of highly similar unsintered bodies can be formed via mixing the components in step (a) and compressing the granules in step (b). The use of a free flowing powder of granules mitigates issues with known mixtures, which are unsuitable for automatic presses. Additionally, unsintered bodies formed from the granules are stronger and more stable relative to existing methods, such as when being handled or moved for heating/sintering.

Note that each of the temperature ranges of steps (c) and (d) may each independently be a temperature. In other words, heating may be done at a particular temperature rather than in a given temperature range. Note also that references to the salt may be to a plurality of salt particles unless otherwise specified. References to reagents or constituents may refer to one of the binding agent, the lubricating agent and the solvent, or to two of these in combination, or to all of these in combination.

The metal granules provided in step (a) each include a surface layer of metal particles. The surface layer may be substantially a monolayer. The method may further include the step of selecting a ratio of the metal particles to the salt for forming the monolayer. A thin even coating of metal particles encapsulating a salt particle is advantageous for sintering of the granules. If the layer of particles on each granule is too thick, then sintering does not occur as effectively. Hence a monolayer is often most effective because this ensures that, when metal particles migrate and join with those on neighbouring granules, the salt particles in the granules come into contact or face one another across gaps in the metal structure. This allows the salt in the body to be removed through the network of gaps/pores, to optimise the thermal performance of the final porous metallic material. The relative thickness of the monolayer depends on the size of the salt particle.

Note that metal particles sizes from substantially 10-200 microns may be used in different embodiments. Suitable sizes of salt particles may range from substantially 100-2000 microns. The salt size is selected to be suitably larger than the metal particle size. The ratio of metal to salt particles selected will depend on the particle sizes of each. A slight excess of metal particles (relative to that required for exact coverage of all salt particle surfaces) may be used, to ensure complete coverage. Alternatively, a slight insufficiency of metal particles may be used, giving an incomplete metal surface for a few salt particles, but avoiding multi-layering of the metal particles.

In production of the granules, the mixture dries during mixing in because the solvent evaporates. This forms composite granules with a general structure of a salt core surrounded by a layer or shell of the metal particles, bound together by the binding agent. The metal particles are coated with the binding agent, which binds to the salt. The lubricating agent can also cover parts of the salt and metal particles. The granules are essentially dry once the solvent has evaporated. The solvent is used to homogenise the mixture, rather than as the binding agent, and the granules do not separate into metal and salt particles once dry. The granules may be further mixed after they have formed to continue the drying process, allowing residual solvent to be removed by evaporation more easily. A preservative may be used during production of the granules. The preservative may be added to the mixture before or during mixing. Production of the granules may include partially drying the mixture, and then mixing the preservative into the mixture.

The preservative helps to retain the granules in a similar state to that of or following their initial formation. It can help to retain residual solvent, for example, slowing the rate at which it evaporates. It can also reduce the rate at which the binding and/or lubricating agents evaporate. Paraffin, for example, can also form a protective layer to prevent oxidation of parts of the granules.

The salt may have a melting point substantially similar to or higher than the metal. The melting point of the salt may be higher than the sintering temperature of the metal.

This ensures that the salt remains solid during the sintering process, which is below the melting temperature of the metal. Usually the sintering temperature is around 70-90% of the value of the melting point of the metal. This also avoids potential problems with deformed pore structures, if the salt particles begin to melt and change shape. Salts which are more thermally stable are preferred for this particular aspect, since they can allows access to higher sintering temperatures, which gives a better sinter result, and also permits sintering of a wider range of metals (having higher sintering temperatures).

The metal particles may be alloy particles. The metal or alloy particles may include one of the following elements: aluminium, chromium, cobalt, copper, gold, iron, magnesium, molybdenum, nickel, silver, tin, titanium. Examples of possible alloys include bronze and stainless steel. Preferably, the metal or alloy particles are around 50 to 100 microns in diameter. This can give rise to a microporous metallic foam as the final material.

Most preferably, the salt is sodium metasilicate.

Whilst silicates are generally much less soluble than carbonates, silicates are generally very thermally stable and relatively unreactive, and using a silicate as a space-holder salt can result in a highly uniform pore structure. Sodium metasilicate (Na2SiC>3) is preferred because dangerous by-products are not released during this method (such as elemental potassium of the known method of EP 1755809). Additionally, sodium metasilicate has a relatively high melting point of around 1080°C, ensuring that it remains stable for sintering temperatures up to this temperature. This is higher than the decomposition temperature of K2CO3, for example, enabling sintering at temperatures not possible with K2CO3. Sintering at temperatures above the salt melting point may be possible too, but may potentially give a less uniform pore structure in the resulting porous material. A metal silicate may be selected according to the metal particles used, and the metals in each may be different.

The salt, the metal particles, the binding agent and/or the lubricating agent may be powdered. The particle sizes of the metal and the salt may be selected to adjust the properties of the resulting granule(s) and/or porous metallic material.

Using powdered forms of the salt, metal and reagents increases the surface area of each, allowing for finer granules to be formed. Smaller granules can flow more easily in bulk. The particular grade of powder for each of the salt and metal needs to be selected carefully, tailoring the surface area and volumes of the metal and salt particles to each other in order to create granules giving a particular porosity in the final material. This is because using smaller salt particles leads to an effectively ‘thicker’ monolayer of metal particles, relative to the volume enclosed by the shell of metal particles around the salt particle, for a given granule. A single reagent may provide both binding and lubricating properties, i.e. a single reagent may be both the binding and lubricating agents. The binding and lubricating agents may comprise a single reagent. Either or both of the binding and lubricating agents may include a wax. The or each wax may be an organic wax. The binding agent, lubricating agent and solvent may each be independently selected from one or more of the following: trihydroxystearin; N,N’ Ethylene Bisstearamide; paraffin; isopropanol; kerosene. Preferably, the binding agent includes trihydroxystearin and/or microcrystalline cellulose. Preferably, the lubricating agent includes N,N’-ethylene bisstearamide. Preferably, the solvent includes isopropanol and/or kerosene.

The binding agent, lubricating agent and solvent may each be independently selected from one of more of the following: Thixcin®; Acrawax®; paraffin; Fastklean. Preferably, the binding agent includes Thixcin® and/or Pharmacel®. The binding agent may have thixotropic properties. More preferably, where Pharmacel® is used, Pharmacel® 101 is included. Preferably, the lubricating agent includes Acrawax®. Preferably, the solvent includes Fastklean. Preferably, the preservative includes paraffin. More preferably, the preservative is provided in an amount of the solvent.

Note that trihydroxystearin is a key component of Thixcin®, N,N’ Ethylene Bisstearamide is a key component of Acrawax®, microcrystalline cellulose is a key component of Pharmacel®, and isopropanol and kerosene are key components of Fastklean. The following advantages are discussed with relation to Acrawax®, Thixin®, Fastklean and Pharmacel®, but apply similarly to the relevant components.

The ratio of lubricating agent to binding agent may be substantially between 1:1 and 3:1. Preferably, the ratio is substantially around 2:1. The combination of lubricating agent and binding agent may include Acrawax® and Pharmacel®, or their equivalent key components. Preferably, the combination of lubricating agent and binding agent includes Acrawax® and Thixcin®, or their equivalent key components.

Acrawax® contains a stearate wax, N,N’ Ethylene Bisstearamide, which has moderate binding properties and good lubricating properties. This enables the granules to be compressed (or compacted) into a body at lower pressures than would otherwise be needed. The lubricant aids removal of the body from the die/mould after compression. Acrawax® tends to swell and become tacky when solvent such as Fastklean is added, further enhancing its lubricating properties.

Pharmacel® is a binding agent and contains microcrystalline cellulose, which binds metal particles to the salt. It is readily mixable with the metal particles and salt, and helps form a well-compressed body of granules.

Thixcin® is a multi-purpose wetting agent and binding agent. It contains a thixotropic castor oil derivative, and its properties change as it is worked during mixing. Applying stress makes Thixcin® less viscous, allowing it to be distributed equally when mixed into a mixture, and thickening when dried. Thixcin® can also act as a preservative for the granules prior to compression, and confers strength for the granules during compression.

Note that the use of organic waxes is preferred for granule preparation, because these waxes do not evaporate quickly at room temperature. This contrasts with conventional binding agents such as ethanol, as used in EP 1755809.

In the mixture of lubricating and binding agents, Thixcin® is preferred to Pharmacel® because the mixture has a lower sensitivity to (i.e. greater tolerance for) the proportion of solvent used to achieve successful granule formation. Additionally, Thixcin® leaves less residue in the body after heating and prior to sintering, and is preferred for automated processing.

Adding paraffin further enhances the longevity of the granules, i.e. the time between granule formation and the period of time for which they remain suitable for compression into a body. Paraffin also has moderate binding properties, to bind the metal and salt together, and reduces the rate of evaporation of reagents and solvent. This is useful because the binding agents can lose some of their binding properties when the solvent has evaporated, thus keeping some solvent in the granules gives a stronger compressed body, minimising the ease with which the granules can separate.

Using Fastklean is advantageous because it allows the mixture to dry quickly, and interacts well with the other reagents, metal particles and salt.

The ratio of lubricating agent to binding agent selected above maximises longevity of binding the metal to the salt, without over- or under-lubricating the granules (which can create problems following granule compression). There is also some benefit to the bulk flow properties of the granules. The lubricating agent enables easier ejection of the body from the die/mould after the granules have been compressed. This supports automated large-scale processing of the granules.

The solvent may be substantially removed from the granules prior to applying pressure to the granules.

This supports easy transfer of the granules from a mixing vessel to a compression vessel. It also dries them for ease of storage, if needed.

Before compression the granules may be transferred to a vessel. Step (b) may further include compressing the metal granules within the vessel to form the body. Preferably, the granules are levelled before compression. Step (b) may further include removing the body from the vessel. Preferably, compression of the metal granules is performed at a pressure substantially up to around 200 MPa. The vessel may be a die or a mould, for example.

Pressing the metal granules together forms a ‘green’ (i.e. unsintered) body. Each granule has a higher contact area with neighbouring granules following compression, maximising cohesion of the granules in the body. This is because the granules are plastically deformed. The metal deforms around the relatively harder salt particles under the high pressure. By virtue of this, small gaps or holes are created in the metal framework (or skeleton) of compressed granules. The salt particles may meet via these gaps. Interconnection of these holes or gaps forms the basis for an open pore network in the material post-sintering.

Note the compression should be performed at a pressure which exceeds the yield strength of the metal particles. EP 1755809 discusses yield strength with regard to producing porous metallic materials, the disclosure of which is incorporated herein by reference. Particular reference is made to paragraphs [0021] and [0035] of the granted specification of EP 1755809. Example yield strengths of metals and alloys which lie around or below the preferred pressure are provided below:

Aluminium: 35 MPa Copper: 69 MPa

Iron: 130 MPa Nickel: 138 MPa

Steel (1020):180 MPa Brass (70Cu-3OZn): 75 MPa

Stainless steel 316 (hot finished and annealed): 205 MPa

Aluminium alloy (2024, annealed): 75 MPa

Titanium (commercially pure, annealed): 170MPa

To elaborate, plotting stress against strain for a metal gives a linear plot from the origin for elastic deformation of the metal. Further increasing the stress can cause plastic deformation (or plastic strain). Where this additional stress causes a slight deviation by a strain of 0.2% from the linear stress-strain plot, this is considered to be the yield strength of the metal.

Levelling the granules ensures that there is even pressure distribution across the granules, and compression is uniform throughout the resulting body. A pressure of around 200 MPa is preferred because higher pressures lead to a better eventual sintered body. Salts such as K2CO3 can fracture or shatter around this pressure leading to a deformed pore structure, but sodium metasilicate, for example, can withstand 200 MPa sufficiently.

The transfer of the granules to a vessel is simpler than existing processes because the granules are dry and have good flow properties, further aided by the lubricating agent. The lubricating agent also aids removal of the body from the vessel.

The method may further include a step between steps (b) and (c) of removing surface layers of the body. This exposes the salt in the body and gaps between the compressed metal granules at the surface of the body. This may be considered as modification of the body to remove material. To remove surface layers, mechanical abrasion or chemical etching of outer portions of the body may be performed. Where mechanical abrasion is used, this may include wire brushing the outer portions.

The pressure applied in step (b) can cause a metal ‘skin’ to form on the pressed surfaces of the body. This occurs more readily for softer metals, such as copper. The skin needs to be removed (‘de-skinning’) from the outer faces of the body to open the pores, allowing later removal of the salt post-sintering, and enabling efficient thermal transfer in the final material. Mechanical abrasion can achieve this without introducing new contaminants, but chemical etching can avoid generating harmful particulates or applying excess force and breaking an unsintered body.

Step (c) may include heating the body for a period of time in each temperature range (or at particular temperatures). These temperatures are substantially below the sintering temperature of the metal. Step (d) may include heating the body for a period of time around the sintering temperature (or in the sintering temperature range) of the metal. The periods of time may be independently selected in each case.

The temperatures of steps (c)(i-iii) may each form part of a range of temperatures. These ranges may correspond to successively higher ranges of temperatures. The ranges correspond to the following purposes: drawing reagents to the surface of the body, as in (c)(i); evaporating reagents from the body, as in (c)(ii); and decomposing any remaining reagents or impurities of the body, excluding the metal and salt, as in (c)(iii).

Preferably, the body is incrementally heated over each sub-range. The body may be maintained at a peak temperature of each sub-range for a period of time.

Existing methods may use solvent to dissolve reagents from the body, but the use of a soluble salt means that this is not suitable here. Instead, a heating cycle aids removal of the majority of reagent compounds prior to sintering. Steps (c)(i-iii) remove the reagents from the body by heating below temperatures that would melt the salt.

Although these reagents are needed to form the body, they can impede sintering in step (d) by forming a barrier between the metal particles. Removal of the reagents before sintering can also minimise a waxy build-up of residue on the inside of the furnace or other heating device following sintering.

Subsequently heating the body in the sintering temperature range of the metal allows the metal particles in the granules to ‘diffuse’ or flow in the solid phase, meeting the metal particles of neighbouring granules. This is also referred to as necking. Necks between the granules grow and strengthen as the body is heated in the sintering temperature range. The resulting sintered body is a strong single piece of metal, rather than a compressed aggregation of separate metal granules as in the unsintered body.

The different heating phases or ranges of temperatures can be selected to wick or draw reagents to the surface of the body over a period of time. At higher temperatures, waxes become more fluid, and tend to migrate across the metal surface to the exterior of the body, where they can gradually evaporate. After most of the reagents have evaporated, thermal decomposition of any remaining reagents is performed to convert them to a carbon ash. It is important to do this prior to sintering the metal because the reagents can hinder sintering and/or become impurities in the metal structure, to the detriment of the properties of the metallic foam.

Incremental heating (ramping up) from room temperature to different higher temperatures allows gradual extraction of the reagents to the outermost surfaces of the body. Initially, the most volatile reagents are removed, followed by heavier less volatile reagents. Otherwise, heating to overly high temperatures initially would either vapourise or pyrolyse the reagents within the body, and the resulting residues would be extremely difficult to remove without damaging the body.

Step (c)(i) may include incrementally heating the body up to substantially around 240°C. Step (c)(ii) may include incrementally heating the body up to substantially around 310°C or 510°C. Step (c)(iii) may include incrementally heating the body up to substantially around 750°C. Step (d) may include incrementally heating the body up to substantially around 950°C. The temperature may be ramped up at a particular rate for preventing early decomposition.

The initial heating phases remove most of the reagents as described previously. Specifically, 240°C is a preferred temperature for melting and wicking lubricating agent to outer surfaces of the body. The binding agent may also be melted and wicked to the outer surfaces at 240°C. Residual solvent may also evaporate at that temperature. 310°C is a preferred temperature for evaporating the lubricating agent on those surfaces. 510°C is a preferred temperature for evaporating the binding agent from the outer surfaces. These temperatures are particularly preferred where Acrawax® is the lubricating agent and Thixcin® is the binding agent.

Subsequently heating to around 750°C is usually sufficient to thermally decompose any remaining reagents, residues or impurities to carbon. Heating to around 950°C allows copper to sinter because it lies below the melting point of copper, but hot enough for the atoms to migrate across grain boundaries of the metal particles during the sintering process. Different temperatures may be preferred for different metals, but the principle remains the same.

The method may include an additional step (c)(ii’) between steps (c)(ii) and (iii). Step (c)(ii’) may include heating the body in another temperature range substantially between the second and third ranges of temperatures, for substantially removing or evaporating the binding agent from the body. In this case, step (c)(ii) may be modified so that it has a temperature range suitable for removing only the lubricating agent from the body, e.g. 310°C. Step (c)(ii’) may include incrementally heating the body up to substantially around 510°C.

Steps (c)(i), (ii), (ii’) and/or (iii) may be performed under vacuum or an inert atmosphere. Steps (c)(i), (ii), (ii’) and/or (iii) may be performed at a pressure substantially below atmospheric pressure. Step (d) may be performed under hydrogen.

Using an inert atmosphere prevents reactions with atmospheric oxygen occurring at the surface of the metal at the high temperatures used. For example, copper forms a stable oxide layer at high temperatures, which can prevent or inhibit the sintering process, so using an inert atmosphere such as nitrogen prevents oxides forming. Similarly, the reagents could react with atmospheric gases and damage the interior of the body, and/or leave residues which have a detrimental impact on the properties of the final material.

Under vacuum, or reduced pressure, in steps (c)(i), (ii) and/or (iii), the boiling points of the reagents are lowered, making it easier to evaporate them by heating at a comparatively lower temperature relative to atmospheric pressure. Heating under hydrogen for step (d) helps to further reduce (or decompose) any persistent impurities, and clean the metal surfaces of the body. In particular, the hydrogen decomposes metal oxides on the surface, which would otherwise impair thermal performance, amongst other properties.

Step (e) may include dissolving the salt in a solvent. The solvent may be provided in a container and agitated by ultrasound. The solvent may be heated. Preferably, the solvent is water. Preferably, the solvent has a temperature of up to substantially around 70°C.

Dissolving the salt avoids using high temperatures to melt the salt, which can warp or melt the sintered body, and also avoids decomposing certain salts by evolving gases (e g. carbonates evolving CO2) which can leave insoluble residues or damage the sintered body if the salt is trapped and creates a rupture when gas is evolved. Ultrasonic agitation quickens dissolution, particularly for salts with low solubility in a given solvent. Water is a polar solvent and dissolves most salts to some extent. Other polar solvents may be suitable. Water will also not significantly change or damage the sintered body for most metals. In the case of iron, de-oxygenated water may be preferred to minimise rusting. The sintered body may be blown dry with air or another gas afterwards, to remove residual water.

Where the salt is sodium metasilicate, water at a temperature of 70°C is preferred. The solubility of sodium metasilicate in water above around 72°C begins to reduce, so higher temperatures are disadvantageous and do not lead to a higher rate of dissolution for that particular salt. Sodium metasilicate is relatively insoluble in water, so it may take several hours to fully remove it from a given sintered body. Since the salt no longer occupies internal parts of the sintered body, the sintered body is now a porous metallic foam having very high thermal conductivity, mainly due to its high interior surface area.

When the electrical conductivity of the solvent around the body substantially exceeds 20 mS, the solvent may be replaced by or replenished with new solvent.

Dissolving the salt increases the concentration of salt ions in the solvent. This increases the electrical conductivity of the solvent. The electrical conductivity can be measured, using a handheld meter, for example. To maximise the rate of dissolution, it is preferred to keep the concentration of salt ions in solution low. Furthermore, some salt ions, silicate ions in particular, can polymerise in solution. Polymers are difficult to remove from the body without causing damage. Frequently changing or diluting the solvent is worthwhile in these cases.

Preferably, the solvent is replaced (or diluted) if the electrical conductivity of that batch of solvent exceeds 20 mS. This is sufficient to avoid substantial silicate polymerisation due to excess concentration.

Preferably, the solvent is replaced (or diluted) after a period of time if the electrical conductivity of that batch of solvent is between 0.7 mS and 20 mS. The period may be 30 minutes, for example. This maximises the rate of dissolution, by reducing the number of salt ions in solution. It is also sufficient to avoid substantial silicate polymerisation at moderate concentrations over extended periods of time.

When the electrical conductivity of the solvent is below 0.7 mS after a period of time, the salt is considered to have been fully removed from the sintered body. The period of time may be 30 minutes, for example.

The method may include the further step of f) darkening surfaces of the sintered body for improving its emissivity.

Darker or blacker surfaces in the sintered body significantly improves its thermal emissivity, i.e. enabling faster radiative heat loss. This is particularly useful for sintered bodies for use as free convective pieces. The darkened layer will generally have significantly lower thermal conductivity than the rest of the sintered body, so needs to be thin for the improved emissivity to outweigh the reduced conductivity.

Step (f) may include heating the sintered body in air or oxygen for a period of time for controllably oxidising surfaces of the metal. Preferably, the temperature of the sintered body is quickly ramped up to an oxidation temperature for the particular metal, to avoid warping the sintered body.

This is appropriate where the metal oxide is darker than the native metal. The oxide layer needs to be uniform and thin to optimise the properties of the sintered body. It works particularly well where the metal is copper. Some metals, such as gold or perhaps iron, are unsuitable for this step. Other metals, such as aluminium, naturally form an oxide layer at their surface, and so post-processing to introduce an oxide layer may not be needed.

BRIEF DESCRIPTION OF THE DRAWINGS

For a better understanding of the present invention, and to show more clearly how it may be carried into effect, reference will now be made by way of example only to the accompanying drawings, in which:

Figure 1 shows a cross-sectional illustration of an ideal embodiment of a metal granule, having a salt core and a layer of metal particles surrounding the salt core;

Figure 2 shows a cross-sectional illustration of the metal granule of Figure 1 compressed together with another metal granule;

Figure 3A shows a cross-sectional illustration of a compressed body of metal granules with an outer metal skin;

Figure 3B shows a cross-sectional illustration of the compressed body of Figure 3 A, the outer metal skin having been removed;

Figure 4 shows a cross-sectional illustration of the progressive sintering of neighbouring metal granules in a compressed body; and

Figure 5 shows a cross-sectional illustration of two sintered metal granules where the salt cores have been removed.

DESCRIPTION OF PREFERRED EMBODIMENTS

Referring firstly to Figure 1, an idealised embodiment of a metal granule is indicated generally at 10. The granule 10 includes a salt particle 12 and a plurality of metal particles 14 around the periphery of the salt particle 12. The granule 10 is substantially spherical. The layer of metal particles 14 is a monolayer. The granule 10 may also be considered to be a composite granule.

The layer of metal particles 14 substantially covers the outer surface of the salt particle 12. In this embodiment, the salt particle 12 is composed of sodium metasilicate (Na2SiO3). In this embodiment, the metal particles 14 are copper (Cu) particles. The metal particles 14 are bound to the salt particle 12 by a binding agent (not illustrated). A plurality of granules 10 can be formed into a porous metallic foam by: compressing the granules into a ‘green’ body; de-skinning the body to remove the resulting metal skin (exposing the salt and gaps between the compressed granules); heating the body to draw out non-metal and non-salt constituents, i.e. compounds used in preparing the granules; sintering the metal at a higher temperature to merge the granules; and dissolving the salt out of the sintered body to free the pores in the foam.

These and other steps are considered in further detail below. Note that whilst automated processing is preferred, manual (hand) processing is also possible. Note also that Figures 1-5 are for illustrative purposes only. The granules in Figures 3 A and 3B, for example, would be much more numerous, and packed more tightly and regularly, than shown. Similarly, Figures 1, 2, 4 and 5 only show illustrative portions or sections of different parts or stages in the manufacture of a porous metallic material.

Porosity of the metallic foam

The porosity of the metallic foam formed in this process can be customised. The porosity depends on the particle sizes of the salt and metal particles used. In some cases, it may be necessary to use a particle sieve for removing particles which are too large or too small, relative to the advertised size ranges.

For Cu and Na2SiC>3, different grades of each are widely commercially available. Examples of particle size ranges and the resulting foam porosity are given below:

The required masses of each of the metal and salt particles can be calculated from the relevant particle masses, volumes and densities, and subsequent calculation of the approximate ratio of the number of metal particles needed to encapsulate a salt particle. Since there are a range of particle sizes in a given sample of metal or salt particles, there will be variation in the porosity in a given section of the metallic foam, hence the porosities given are approximate and apply to the material as a whole.

The calculations may be performed based on upper and lower limits of particle sizes, and/or with a certain distribution of particle sizes in mind. Certain correction factors may need to be applied when calculating the masses of metal and salt particles needed. A scaling factor may be used to adjust the ‘effective’ density of the salt, for example. This is because the density of the salt particles as a whole is different to that of each crystal, due to sphere packing and gaps between the particles. Sieving the salt can also change the overall density of the particles, by removing outlying sizes, and depends on the particle sizes used.

The main aim of calculating the ratio of metal to salt particles is to provide a monolayer of metal particles around each salt particle. Note that smaller particle sizes result in a relatively thicker monolayer, in terms of the thickness of the monolayer to the size of the salt particle. This in turn tends to lead to a lower porosity, and vice versa for relatively thinner monolayers.

Granule preparation

The preparation of a plurality of metal granules 10 is performed in step (a) of the method. A powdered binding agent, Thixcin®, and a powdered lubricating agent, Acrawax®, are provided in a mixing vessel. Powdered Na2SiC>3 particles and Cu particles are added to the binding and lubricating agents. In some embodiments, the metal particles are added after mixing the salt particles into the reagents. The binding agent has a high binding affinity between metal and salt particles. A solvent is then added to form a mobile slurry. In this embodiment the solvent is Fastklean, which includes isopropanol and kerosene. Although it is normally used as a degreaser, the fast-drying nature of Fastklean and its interactions with the binding and lubricating agents make it suitable for the granulation process.

In this example, the granules 10 are formed from a mixture of around 800g of Cu particles (53-100 microns), around 200g of sodium metasilicate particles (425-850 microns), around lOg of Acrawax® and around 5g of Thixcin®. Around 140 ml of Fastklean is used to form the slurry of the above components. This should be suitable for a metallic foam of around 63% porosity. The relative amounts of binding and lubricating agents vary for other metals or salts, and/or different ranges of particle size. Note that the amount of solvent is selected with the amount of lubricating agent in mind, to avoid the lubricating agent being squeezed out excessively during pressing.

The slurry is stirred for homogeneity. The mixture can then be left to dry, forming a ‘dough’ as the solvent slowly evaporates. In some embodiments, the mixture may be actively dried, by heating, for example. A pre-mixed combination of paraffin and Fastklean is then added to the dough, re-wetting it to form a slurry upon initial mixing.

Typically, the paraffin has a concentration of 1% in the solvent prior to addition to the dough. Mixing evenly distributes the paraffin with the granules being formed.

Further mixing allows the slurry to dry again, as the solvent evaporates. The metal particles 14 are bound to the salt particles 12 by the binding agent, and incorporate some of the lubricating agent and paraffin too. Initially, before fully dry, the slurry becomes an agglomerate of granules stuck together by the remaining solvent. However, further gentle mixing causes the agglomerate to break up into individual granules, like that of Figure 1. The overall processing time for these steps is around 30 minutes.

These granules 10 are essentially ‘dry’, but some residual solvent is still present. The binding agent works most effectively when wet, so predominantly drying the mixture can result in granules 10 that can be easily damaged by further mechanical work, without any solvent. However, paraffin is much less volatile than the solvent, and in place of the solvent enables the binding agent to continue effectively binding the metal and salt particles together over a longer timescale. Furthermore, since paraffin evaporates slowly in air, the granules are also relatively air stable, over a period of a day or two, before becoming too fragile for further processing.

In light of the above, the granules are suitable for medium and/or long term storage. For example, the granules can be sealed in containers or bags. Ideally the air is also removed from the containers, or in the case of bags, squeezed to minimise gas volume. This in turn minimises evaporation within the containers, further extending the useable lifetime of the granules. This is beneficial for industrial production because the granules do not need to be formed in situ prior to performing the subsequent manufacturing steps.

Granule compression into a body

The granules 10 can be compressed together to form a ‘green’, or unsintered, body. This is a rapid process for producing practically identical bodies for sintering, unlike the variability of the resulting metal foams from other methods of production. Compression aids sintering by forcing the metal in the particles to partially deform around the relatively harder salt particles. This in turn exposes gaps or ‘pinholes’ between neighbouring salt particles, or brings neighbouring salt particles into contact, which is the foundation for the network of interconnected pores in the final material.

Figure 2 shows the result of compression for two metal granules in isolation. The salt particles are in contact at a point 16. This is exemplary only, since there would normally be thousands of granules packed together in a 3D arrangement, with contact between a given granule and each of its immediate neighbours. The pressure deforms the metal around the salt particles, creating a series of pinholes in the metal structure between large cavities occupied by the salt particles.

First, the granules are poured from the mixing vessel to a die or mould, which acts as a compression vessel. The granules flow well due to their small size, and partly due to the lubricating agent. In this example, the granules are transferred from a hopper through a filling shoe and into the die. For a larger body, the shoe action should be correspondingly slower, to ensure an even fill of the die.

Once the compression vessel has been filled, the granules are levelled to ensure that the resulting body has even thickness throughout its length, so that pressure is applied evenly across the granules, avoiding sections being detrimentally over- or undercompressed.

Then, a ram is applied to the granules to apply pressure and compress them together, forming a body 18. The ram may be a hydraulic ram, for example. A higher pressure leads to a harder body and a better sintering result. However, high pressures can fracture certain salts, so the pressure applied depends on the salt used, to avoid the salt shattering and the pores being deformed. The ram applies pressure at around 200 MPa in this embodiment, which is suitable for the Cu particles and Na2SiC>3 particles in the body 18. This pressure is around the fracture point of potassium carbonate, so a lower pressure would be required for that salt, if used.

After pressing, the ram is withdrawn and the body 18 ejected from the die. During pressing, some of the lubricating agent is squeezed out of the granules in the body 18, which aids removal of the body 18 from the die. The die and/or body 18 may need cleaning if excess lubricant is present. The ram may be applied to the other end of the die to work the body 18 out, or the body 18 may be automatically punched out. Care is required in either case due to the fragile nature of the body 18, and because it may be stuck in the die after pressing.

Note that the granules in the body are plastically deformed. There is cohesion between the granules due to the higher contact surface area between the granules.

In this example, a Dorst 40t auto-press was used. In some embodiments, a 100 tonne press may be used, suitable for a body up to around 50mm x 50mm x 20mm in size. A 400 tonne press would be suitable for a body up to around 180mm x 200mm x 10mm in size. The body should be at least 2mm in thickness, otherwise there can be issues with ensuring even thickness, and the body is overly fragile. The body is typically cuboidal, due to limitations with the next step of‘de-skinning’.

Removing surfaces of the body

Unlike traditional powder metallurgy, the above pressing process creates a metal ‘skin’ 18a on the pressed faces of the body 18. This is illustrated in Figure 3 A, and happens particularly for soft metals like Cu. The body 18 then requires modifying by ‘deskinning’. This is to remove the metal skin/layers of metal at the surfaces and expose the salt in the body. It also exposes gaps/interstices between the metal granules, present due to imperfect ‘sphere packing’ of the granules. This is illustrated in Figure 3B. This in turn leaves the faces open for later removal of the salt 12, to leave an interconnected network of pores.

The metal skin 18a can be removed or stripped by mechanical abrasion or chemical etching. Care is needed when performing mechanical abrasion due to the relatively fragile nature of the unsintered body 18. Mechanical abrasion using sandpaper is possible, but wire brushing is preferred. This removes the metal skin 18a without significant impact on the salt. Wire brushing can be performed before or after the sintering step has been completed.

Note that wire brushing after sintering can lead to the brush closing off the finer pores in the material, adversely affecting the thermal properties of the final porous metallic material, for example. Consequently, wire brushing prior to heating the body 18 is preferred. Fine particulate matter from the salt 12 can become airborne when wire brushing the body, so suitable precautions should be taken.

For chemical etching of Cu, for example, treatment with ferric chloride or copper(II) chloride removes unwanted copper portions. This can be done over a period of 20-30 minutes, for example. Where possible, the etchant should be regenerated on site to optimise production. For example, when used to remove Cu, copper(II) chloride becomes copper(I) chloride, which can be regenerated using one of chlorine gas; hydrogen peroxide and hydrochloric acid; or sodium chlorate and hydrochloric acid.

Other suitable etchants can be selected based on the metal particles used. However, some metals may be too inert for chemical etching to succeed without using harsh chemicals, which could damage interior portions of the body. Therefore, mechanical abrasion is preferable in such cases.

Extracting reagents from the body

Once the unsintered body 18 has been modified to expose the interstices between granules, the reagents used to create the body 18 need to be removed as completely as possible, leaving just the salt 12 and metal 14. This involves thermal treatment of the body 18 at different temperatures. Since the reagents are organic waxes (i.e. hydrocarbons of different types), this allows them to be wicked to the surface at high temperatures.

The table below illustrates a preferred set of temperatures for achieving this where Cu and sodium metasilicate are used:

Heating of the body to the temperatures listed can be performed using a furnace. In one embodiment, a large Consarc® vacuum furnace can be used. Other types of furnace may be used instead of a vacuum furnace. The body or bodies are typically placed in the furnace and then heated from room temperature, to avoid thermal shock (where placed cold into a heated furnace).

Phases A-D are preferably done under a nitrogen atmosphere. Another inert atmosphere may be used in other embodiments. Typically, the atmosphere used is at atmospheric pressure. However, other pressures may be used in some embodiments. Note that phase A is split into two contiguous phases. Initially, the temperature is ramped up at a slow rate above 80°C. The second phase is conducted without stalling at 80°C before further heating. All of the heating phases take place sequentially in order from A to D.

Different temperatures and/or ramp up rates may be used depending on the metal used, to avoid sintering the metal before the reagents have been removed. However, the ramp up rates must not be too high, or else the body can undergo cracking or delamination. There may be some variation in the final temperatures selected and the times spent at those temperatures, depending on the components of the granules.

During phase A, the lightest compounds (in terms of molecular weight) are drawn/wicked to the outermost surfaces of the body. The lubricating agent melts, or become substantially fluid, at the initial elevated temperatures depending on the particular wax used. This facilitates wicking to the outside surfaces of the body, hence phase A is a ‘melting/wieking’ phase. It can gradually evaporate from the surface. Residual solvent is also removed.

In phase B, slightly heavier compounds are removed. This phase is typically for removal of the lubricating agent. Where the lubricating agent is a wax, this phase is a

‘de-waxing’ phase. Again, the compounds migrate across surfaces of the body until they evaporate.

In phase C, most of the remaining heavy hydrocarbons migrate from the internal parts of the body to the outer surfaces, along with residues that may be present. Typically, this corresponds to removal of the binding agent, i.e. a ‘de-binding’ phase. The temperature plateaus for a time because the binding agent is mainly trapped between the metal and salt particles, so it requires time to be fully extricated without decomposing in situ (within the body). The same applies to all of the reagents, which is why the temperature is increased slowly - to avoid decomposing the compounds whilst still within the body.

Phase D is substantially equivalent to a thermal burnout phase. This decomposes all remaining compounds or residues to carbon by pyrolysis. The final temperature in phase D is insufficient to sinter the metal, which is Cu in this embodiment. Note that the amount of carbon produced in this phase is generally low enough that there is no significant effect on the thermal properties of the final porous metallic material. However, carbon can inhibit sintering, so phases A-C are important for minimising the total amount of carbon produced.

Sintering metal in the body

Following on from phase D, a fifth heating phase E is performed. Phase E ideally occurs immediately following completion of phase D. Phase E preferably occurs under a hydrogen atmosphere. The final temperature of phase E is 950°C. Sintering typically occurs as 70-90% of the melting point temperature of a metal. Since the melting point of copper is around 1085°C, a temperature of 950°C (-87% of its melting point in Celsius; -90% of its melting point in Kelvin) in phase E is sufficient for sintering to occur.

In this embodiment, sodium metasilicate is used as the salt. The melting point of Na2SiC>3 is around 1080°C or just above, so it remains solid at the sintering temperature of Cu. Furthermore, the metasilicate does not decompose at high temperatures. Both of these characteristics mitigate against an irregular pore structure forming during sintering. Since only the Cu particles become mobile at 950°C, and the metasilicate particles retain the same size and shape, the resulting pore structure is highly uniform.

The process of sintering involves metal particles in neighbouring granules merging together to form a single metal structure. The metal particles at the surface of each granule undergo necking. The progressive formation of the necks over time is illustrated in Figure 4 for four neighbouring granules in isolation. As the necks form, metal spreads away from the pinholes, effectively widening them. This is implied in Figure 4, where the granules lose their distinctly spherical shape as the metal particles from each are merged. As a result, metal effectively recedes from the contact points 16 between the salt particles during necking. This increases the size of the gaps connecting the salt, which eases later removal of the salt. The body effectively becomes a ‘sintered body’ once sintering is complete.

Note that the sintering process can take much longer than in traditional powder metallurgy. This is because the metal particles are bound to the salt particles, and the salt particles do not expand significantly at higher temperatures. This increases the time taken for the metal particles to link via sintering.

In this embodiment, the body 18 is sintered for a period of 120 minutes. It will be appreciated that longer or shorter periods of time may be needed depending on the metal particles used, the temperature used, the pressure at which the body was compressed, and the size of the granules, amongst other factors.

It is important to choose a sintering temperature that sinters the entirety of the body 18. At too low a temperature, the body 18 sinters for outer portions but remains granular in the centre. This results in an easily breakable sintered body having poor thermal properties. The metal structure in the sintered body needs to be structurally sound to maximise thermal conductivity.

Extracting salt from the sintered body

It is necessary to remove the salt from the sintered body to yield the porous metallic foam. Due to compression of the granules before heating, the salt particles are linked via the series of contact points 16, i.e. now gaps 16a in the sintered metal structure. The gaps connect together with each other, providing a means of removing the salt. Some pockets 20 may be isolated in the sintered body as a result of imperfect sphere packing.

In this embodiment, the sintered body is allowed to cool following sintering. The salt can then be dissolved from the sintered body. The solvent used is water, to avoid corroding the metal in the sintered body. The water used here is tap water, although distilled water may also be used. However, sodium metasilicate is poorly soluble in water, so dissolution is relatively slow. To minimise the time required for dissolution, the solvent is heated. For the same reason, the solvent is agitated. Preferably, this is achieved by ultrasound (i.e. sonication), encourage solvation of the salt. Agitation also avoids pockets of water being trapped between neighbouring sintered bodies (where many are being processed simultaneously). The sintered bodies are reoriented regularly, avoiding surface discolouration.

The dissolution process is begun by immersing the sintered body in water. The water is at a temperature of 70°C. Alternatively, the water may be room temperature and heated once the sintered body is immersed. The water is contained in a sonic bath. Sonication at this temperature can take around 4 hours to fully dissolve the salt, for a typical sintered body of around 50mm x 50mm x 20mm. Sonicating the body in the water helps to dissolve the salt into sodium and metasilicate ions. The electrical conductivity of the solution is measured to keep track of the dissolution process. An electrical conductivity meter is used for this.

Note that 70°C is optimal because the dissolution profile of Na2SiC>3 tails oft above around 72°C, so there is no improved rate of dissolution for higher temperatures. Furthermore, above 70°C or so, the rate of dissolution of sodium metasilicate tends to exceed the rate of transport of its solvated ions out of the sintered body. This can form a gel or ‘mucilage’ of sorts, and may lead to polymerisation of the metasilicate at high enough concentrations, which must be avoided. This is because silicate polymers are insoluble in water and acid, and so cannot be removed without damaging or destroying the sintered body.

To avoid polymerisation of the metasilicate, the solvent is changed regularly. There are two main conditions for changing the solvent. First, if the electrical conductivity of the solution substantially exceeds 20 mS, the concentration of metasilicate in solution is too high and the solvent is replaced with a fresh batch. This applies regardless of the length of time which has passed since last changing the solvent.

Second, if the electrical conductivity of the solution remains substantially above 0.7 mS for a period of time, the concentration of metasilicate in solution is sufficient to require a fresh batch of solvent. The period of time is 30 minutes. This is because metasilicate can still polymerise at low concentrations, given enough time. In alternative embodiments, a flow of water may be used in place of a fixed volume of water, to avoid increasing concentrations of ions around and in the sintered body.

The solvent is changed as many times as necessary. The metasilicate is considered fully dissolved from the sintered body when the electrical conductivity of the solvent remains substantially below 0.7 mS for a period of time following replacement of the solvent. The period of time is 30 minutes. The total dissolved solids amount tends to be around 500ppm at this stage, similar to tap water.

The sintered body is then rinsed through with a final batch of fresh solvent (e.g. water). The sintered body is then dried, using a fan oven, for example. This means that the sintered body is now a porous metallic foam. Figure 5 illustrates two pores of the sintered body where the salt has been removed, leaving just the sintered metal.

Note that other embodiments may involve salts which do not form problematic polymers when solvated, such as carbonates. However, it is still good practice to use fresh solvent to maximise the rate of dissolution, where sintering has not decomposed the carbonate, or where the product of salt decomposition needs to be removed.

Note also that the dissolution procedure can be performed for multiple sintered bodies at the same time. The process may be automated. A Kerry Automated Washing Fine can be used, for example, which has an inbuilt electrical conductivity meter. Immersion in hot solvent and applying ultrasound dissolves the salt, and the solvent can be changed automatically based on the similar EC parameters. Once dissolution is complete, the bodies are rinsed to remove any silicate remnants on the surfaces of the bodies, and again with deionised water for the same reason. This also prevents tarnishing during drying of the sintered bodies. Drying can again be performed using a fan oven, for example.

Further processing of the porous metallic foam

Following drying, the foam can be treated to optimise its thermal properties. In the case of a Cu foam, thermal treatment can be used to create a thin oxide layer on the surfaces of the foam. Since CuO is black, this improves emissivity of the foam without being overly detrimental to thermal conductivity at its surface. Other metals may not be suitable for post-processing in this manner, depending on whether they can be oxidised, and the emissive properties of the relevant oxide.

In the case of a Cu foam, the foam is heated in air. A suitable temperature is above 350°C. Ideally, treatment at 600°C for 10 minutes gives a durable oxide layer of minimal thickness, completely covering the surfaces on the interior and exterior of the foam. The ramp up rate to reach the desired temperature should be relatively fast to avoid warping the foam, as may happen if heated for extended periods of time. In this embodiment, a continuous tunnel oven was used, and gave reliable oxidation of porous copper foams.

The porous metallic material may be cut to a particular size. This can be achieved by using a saw or grinder. This can damage the pore structure on the cut surface, but this is preferable in some cases, such as for the base of a low profile heatsink. Smearing the metal on the cut surface with a saw maximises the surface area for heat conduction at the base. This surface may be polished to flatten it for a better contact surface area. Alternatively, waterjet cutting can be preferable, minimising smearing relative to a saw. Other alternatives for maintaining good pore structure post-cutting include electric discharge machining (EDM) and wire erosion.

The porous metallic material may be connected to another object by soldering or brazing. For example, it may be attached to an impermeable plate for heat spreading, or in a sealed chamber such as in IGBT (Insulated-Gate Bipolar Transistors) coolers.

This is preferable to sintering the body when already attached to such items because of the differing rates of thermal expansion between the different materials. To avoid solder or braze paste wicking into the body of the porous metallic material, it can be advantageous to cut the material and create a closed surface by smearing the metal, as above. This creates a suitable surface with high surface area for connection to a plate, without ‘soaking up’ the solder or braze filler. Alternatively, a viscous braze paste can be used. Another alternative is to use hot solder with a relatively cold item of porous metallic material, which minimises solder ingress. Other solder-stop technology, such as the AMT process, may be used.

Note that where temperature ranges are referred to, it is possible for there to be some overlap between temperatures used to remove different reagents. For clarity, where one range is stated as being higher or greater than another, this refers to a mid-point of one range being greater than a mid-point of another range.

The embodiments described above are provided by way of example only, and various changes and modifications will be apparent to persons skilled in the art without departing from the scope of the present invention as defined by the appended claims.

Claims (10)

1. A method of manufacturing granules having a layer of metal particles surrounding a salt core, the method comprising the steps of: a) forming a mixture from metal particles, a particulate silicate or metasilicate salt, a binding agent for binding the metal particles to the salt, and a lubricating agent; b) adding a solvent for homogenising the mixture, to form a slurry; c) mixing the slurry for homogeneity; and d) continuing to mix the slurry to allow the solvent to evaporate until the mixture is sufficiently dry that the granules form during mixing.

2. A method as claimed in claim 1, including the further step of: e) storing the granules in a sealed container for preventing excessive drying prior to use.

3. A method as claimed in claim 2, in which air is removed from the container prior to sealing the container.

4. A method as claimed in any of claims 1 to 3, in which a preservative is added to the mixture prior to step (d).

5. A method as claimed in claim 4, further including a step between steps (c) and (d) of partially drying the mixture, and then mixing the preservative into the mixture.

6. A method as claimed in any of claims 1 to 5, in which the salt has a melting point substantially similar to or higher than the metal.

7. A method as claimed in any of claims 1 to 6, in which the metal particles consist of a metal or alloy including one of the following elements: aluminium, chromium, cobalt, copper, gold, iron, magnesium, molybdenum, nickel, silver, tin, titanium.

8. A method as claimed in any of claims 1 to 7, in which the salt is sodium metasilicate.

9. A method as claimed in any of claims 1 to 8, in which one or more of the binding agent and the lubricating agent, are powdered for use in step (a).

10. A method as claimed in any of claims 1 to 9, in which either or both of the binding and lubricating agents include a wax.