METHOD AND APPARATUS FOR SUPERCOOLING OF METAL/ALLOY MELTS AND FOR THE FORMATION OF AMORPHOUS METALS THEREFROM

FIELD OF THE INVENTION

This invention relates to the creation of glassy or amorphous metals and more particularly to the maintenance of a crystallization-free supercooled melt through the injection of energy into the melt during the cooling process in sufficient amounts to prevent crystallization to and permit the formation of a stable supercooled melt used as a precursor to the creation of glassy or amorphous metals when cooled to ambient temperature.

BACKGROUND OF THE INVENTION

The ability to produce amorphous metals, also called metallic glasses, from the liquid phase in significant sizes has long been pursued. However, practical production limitations imposed by the need for a combination of rapid cooling, immaculate process environments and alloy compositions have limited the applicability of known production processes.

Perhaps the most touted quality of metallic glasses is their combination of mechanical strength, elasticity, hardness and toughness. Crystalline metals/alloys above a very small scale have lattice defects disrupting the long range atomic ordering. These defects are generally the initiation sites of mechanical failure. Without crystals and such crystal defects, amorphous metals tend to outperform their crystalline counterparts in strength and elasticity. In addition to their mechanical strength, the lack of grain boundaries and lattice defects makes the amorphous alloys resistant to corrosion and wear, rendering them suitable as components in harsh chemical/mechanical environments. Moreover, since amorphous alloys can maintain flow at relatively low temperatures without crystallizing they can be molded into complicated shapes using techniques similar to thermoplastic molding.

A metallic glass is expected to have an electrical conductivity two orders of magnitude lower than the metal/alloy in its crystalline structure. As a result, efforts are being made to achieve not just the mechanical strength of metallic glass but also improved electrical conductivity. Moreover, it has been observed that metallic glasses of ferromagnetic materials can exhibit soft magnetization, i.e. almost no hysteresis in the B-H diagram as the magnetic field is cycled above and below zero. This property translates into very low losses when employed as magnetic cores in transformers or other magnetic components.

Supercooling, also known as undercooling, is the process of lowering the temperature of a liquid below its melting point without it becoming crystalized. Thermodynamically, the preferred state for most materials is a crystalline solid if the temperature is below the melting point of the particular material. The crystallization process is always initiated by one or more nucleation events in the liquid. The nucleation process is categorized as either heterogeneous or homogeneous, where heterogeneous nucleation is aided or catalyzed by a foreign element, for instance entrained impurities or the container wall, and homogeneous nucleation is induced by the base metal itself. For either category, nucleation is a random process and the driving force increases with undercooling. Once a nucleus of sufficient size has formed crystal growth ensues. However, if a liquid can be sufficiently supercooled the kinetics of crystallization becomes prohibitively slow and the liquid becomes frozen in an amorphous solid state without a crystalline structure. The temperature range where this occurs is called the glass transition range and it differs from one material to the next.

Generally, in order to reach glass transition for metallic liquids the liquid needs to be cooled sufficiently fast from the melting point down to glass transition in order to avoid nucleation and crystal formation. The necessary cooling rate depends on the material and most efforts in the prior art are concerned with finding good glass formers, that is alloy compositions that have an inherently slow crystallization kinetics and/or a glass transition that is close to the liquidus temperature of the system.

There are several empirical rules for creating a good glass former. Among these rules is the notion that good glass formers tend to include at least three different elements and that these should differ by at least 12% in atomic radius. The stoichiometry of such glass forming compositions also tend to lie close to deep eutectics. Such compositions tend to have a lower mobility when undercooled and therefore require a more modest critical cooling rate. Cooling a melt at a rate that is higher than this critical rate will bypass crystallization and the melt will solidify as glassy. Indeed, limited to techniques known in the prior art, many metallic glasses can only be made with a thickness on the order of millimeters. Additionally, in order to achieve significant supercooling it is generally considered necessary to operate in immaculate process environments to remove foreign substances and external nucleating agents in the melt. If such nucleating agents are present the melt tends to undergo heterogeneous nucleation.

SUMMARY OF THE INVENTION

It has been found that significant levels of supercooling of metals, pure metals as well as various alloys, can be achieved without the need for either dramatic cooling rates or immaculate process conditions. The result is that one can create glassy metals without quenching or stringent processing. Specifically, it has been found that when a melt is subjected to electric currents, either induced magnetically or directly applied, crystallization can be suppressed during cool down and significant levels of undercooling can be achieved without crystallization. As a result of the subject process, one can obtain amorphous metals without quenching when bringing the temperature of the supercooled melt down to ambient temperature.

This method can either alleviate altogether the need for quenching of a melt in order to reach an amorphous state for the making of bulk parts, or can be used in conjunction with existing processing methods with a reduced need for rapid quenching.

Moreover, the subject system does not rely on immaculate process environments nor the use of specialized alloy compositions for achieving and maintaining a supercooled state. Additionally, by increasing the strength of the currents during cooling, deeper supercooling can be achieved.

It has also been found that maintaining a reducing environment of the metal melt improves the process. Such an environment can involve for example hydrogen in the atmosphere or additives such as carbon to the melt in small quantities. The choice of reducing agent for a specific metal depends on the thermodynamic equilibrium of the metal oxide and the particular reducing agent at the operating temperatures, up to 200 degrees Celsius above the melting point.

As will be described, a method for supercooling a melt without rapid quenching involves heating the metal above its melting point and then cooling the molten metal while injecting electromagnetic energy or electric currents into the melt of sufficient magnitude to suppress crystallization during cool down. It is thought that the electromagnetic field strength and the resulting current density, or direct current injection is a key parameter in suppressing nucleation and prevention of crystal growth during cool down below the melting point of the metal at which crystallization normally occurs.

In one embodiment, a metal such as nickel, cobalt, copper or iron is placed in a crucible surrounded by a coil which is driven at a frequency and with a current designed to induce eddy currents in the metal of sufficient magnitude to melt the metal. In a preferred embodiment a nickel niobium alloy and a nickel boron alloy were found to yield an amorphous phase at room temperature.

It is preferable that the metal be exposed to a reducing agent, such as for example hydrogen or carbon in sufficient quantities to remove oxides present in the metal. Thereafter, the current in the coil is reduced resulting in a cooling of the melt. Energy from the coil at the reduced current keeps the atoms of the metal in the molten state just sufficiently so that the supercooled melt does not pass into the crystalline stage or exhibit recalescence at temperatures below the melting point.

In an alternative embodiment, the magnetic or inductive coupling with the metal and the power source is replaced with a conductive coupling. That is, a current is run directly through the metal. The power source drives a current of a magnitude designed to melt the metal. Once the metal is molten the magnitude of the current is reduced to sufficient levels such that the atoms of the metal in the supercooled melt do not pass into the crystalline stage or exhibit recalescence.

Moreover, for certain metal systems it has been found that the degree or level of supercooling can be increased if during cool down the magnitude of the supplied currents can be increased. The increased amount of Ohmic heating that follows from increasing the magnitude of the supplied currents is counteracted by also increasing the rate of heat extraction from the system. This can be accomplished by placing a conductive heatsink in contact with the primary containment of the metallic melt. The ability to remove heat from the melt utilizing thermal conductivity requires an identical amount of heat be added to the melt when maintaining a supercooled steady state.

It is theorized that crystallization is inhibited due to the mismatch in electrical conductivity between the nucleated solid state and the background liquid. With the solid state typically having a higher electrical conductivity, the nucleated phases will locally experience a higher current density compared with the bulk liquid phase as predicted by standard electrodynamics. The higher current density will generate additional heat where crystal growth would occur. This additional heat limits growth rates and melts the nucleated seeds to prevent crystallization.

In contrast to a nucleus of a pure metallic phase, metal oxides can have a lower electrical conductivity than the bulk liquid. The presence of such oxides can have the opposite effect. That is, current densities at the oxides are locally lower than in the surrounding melt which consequently reduces ohmic heating locally. This in turn may result in crystallization.

This problem can be overcome or mitigated by removing the oxides present in the melt with a reducing agent such as hydrogen or carbon which will reduce the oxides to pure metals and a gaseous component, steam or CO2/CO in these cases, where the latter will leave the system.

The above-described methods of supercooling metals require neither extreme cooling rates, nor the need for immaculate environments, nor the need for complex alloy structures. This removes many of the practical limitations in supercooling metals which are necessary in making metal glasses.

More specifically, the three problems solved by the subject invention are: 1) the ability to control the supercooling of metals without resorting to complex alloy compositions, 2) the ability to supercool metals in a “dirty” reactor, i.e. a reactor that offers numerous nucleation sites, 3) the ability to significantly supercool metals without requiring rapid quenching.

Provided sufficient strength of the applied currents during cooling this method can be used to generate amorphous metals without quenching. Even if an amorphous state cannot be reached for a given metal or alloy composition with this method alone, due to the steady-state level of supercooling available significantly below its melting temperature, the additional quench supplied by prior art methods can have a reduced quench rate because the starting point is significantly below the melting temperature. This enables greater thickness of the produced glassy material.

Another benefit of this process is that if crystallization is allowed to occur, either at the normal solidification temperature or in a supercooled state, when the melt is subjected to electric currents the crystal structure can be manipulated. As a result, crystal size and orientation can be tuned with the applied field during crystallization isotropy of a given material. Thus this technique provides a new process to tune structure.

BRIEF DESCRIPTION OF THE DRAWINGS

These and other features of the subject invention will be understood in connection with the Detailed Description in conjunction the Drawings, of which:

FIG. 1 is a diagrammatic illustration of a system to achieve supercooling without using rapid quenching;

FIG. 2 is a current versus time graph of the current applied to the coil of FIG. 1;

FIG. 3 is a temperature versus time graph showing an example of temperature conditions during supercooling in the apparatus of FIG. 1;

FIG. 4 is a diagrammatic illustration, similar to FIG. 1, but where a heat sink is placed in thermal contact with the primary containment to increase the heat extraction from the metal during the process, which requires a higher coil current to reach the same temperature as without the heatsink present, which in turn results in increased crystallization suppression;

FIG. 5 is a diagrammatic illustration of an alternative embodiment in which the metal to be melted and supercooled is a metal rod carried in a nonconductive mold;

FIG. 6 is a current versus time graph of the power applied across the metal rod of FIG. 5;

FIG. 7 is a temperature versus time graph of the temperature of the metal rod after melting indicating passage to a supercooled state without crystallization;

FIG. 8 is an alternative embodiment illustrating the utilization of a metal plate across which is applied current to melt the metal, with a nonconductive mold supplied with a cooling medium;

FIG. 9 is a diagrammatic illustration of the utilization of a single coil for the induction melting and subsequent treating of a metal, indicating the magnetic lines of flux and a height to depth ratio h/d=1;

FIG. 10 is diagrammatic illustration of the utilization of a split coil for the melting and processing of the sample FIG. 9 illustrating the magnetic lines of force and a height to depth ratio of h/d greater than 5;

FIG. 11 is a diagrammatic illustration of the utilization of coils to either side of the sample to be melted in which two different currents are applied to the coils with a phase relationship between the first and second coils to produce the indicated magnetic flux;

FIG. 12 is a front view of the experimental setup with an induction coil utilized in the process of supercooling metals;

FIG. 13 is a close-up of the front of the secondary containment enclosure of FIG. 1;

FIG. 14 is a close-up of the front of the secondary containment enclosure of FIG. 4;

FIG. 15 is a side view of the secondary containment enclosure of FIG. 13;

FIG. 16 is a diagrammatic illustration the water columns utilized in regulating back pressure on primary and secondary containment enclosures FIG. 13;

FIG. 17 is a diagrammatic illustration of experimental apparatus using a direct current to melt and supercool a metal;

FIG. 18 is a series of graphs of temperature versus time and coil current versus time for an experiment repeatedly supercooling nickel plus a small amount of carbon involving thermal contact with a water-cooled aluminum heatsink according to FIG. 4, only the coil current during cooling being displayed;

FIG. 19A is a graph plotting the amount of time the metal spent in a state of supercooling versus coil current during the cooling step for each of the cooling steps shown in FIG. 17;

FIG. 19B is a graph plotting the maximum degree of supercooling during each cooling step in FIG. 17 versus coil current during the cooling step for each of the cooling steps shown in FIG. 17;

FIG. 20 is a series of graphs of temperature versus time and coil current versus time for an experiment using the induction coil to repeatedly supercool 4.39 g of pure nickel without the use of a heat sink;

FIG. 21 is a time-temperature-transformation diagram in the top half of the cycles shown in FIG. 20 and in the lower half a time-temperature-transformation diagram of a similar experiment to the one shown in FIG. 20 but where a heat sink was employed and the sample size was 0.45 g pure nickel;

FIG. 22 is a series of graphs of temperature versus time and coil current versus time for an experiment repeatedly supercooling iron plus carbon without the use of a heat sink showing only the coil current during cooling;

FIG. 23A is a graph showing a plot of the amount of time the metal spent in a state of supercooling versus coil current during the cooling step for each of the cooling steps shown in FIG. 22;

FIG. 23B is a graph showing the maximum degree of supercooling during each cooling step in FIG. 22 versus coil current during the cooling step for each of the cooling steps shown in FIG. 22;

FIG. 24 is a series of graphs of temperature versus time and applied current versus time for a nickel sample treated in the DC reactor of FIG. 17;

FIGS. 25A and 25B are X-ray diffraction patterns of the same sample, with FIG. 24A showing the pattern made from a scan made on the surface of the metal and FIG. 24B showing a pattern from a scan made of powder filed from the same sample;

FIG. 26 is a series of graphs showing the last cooling step of a sample where crystallization is triggered by turning off the coil current completely;

FIG. 27 is a photograph of the sample extracted following the solidification in FIG. 26;

FIG. 28 is a photograph of the sample extracted utilizing a field to limit crystallization;

FIG. 29 is an XRD diffraction plot of an NiNb sample showing the formation of an amorphous material; and,

FIG. 30 is an XRD diffraction plot of an NiB sample showing the formation of an amorphous material.

DETAILED DESCRIPTION

The main scientific discovery disclosed herein reveals that when exposing a melt to either an electromagnetic field or to direct current the natural process of nucleation and crystal growth is suppressed allowing the melt to be significantly supercooled, ie without crystallization over substantial periods of time, thus to provide a stable supercooled melt. Moreover, the stronger the applied field the greater the extent or depth of supercooling, temperature below the equilibrium melting temperature that can be achieved. It is noted in the literature that the viscosity of the melt increases with supercooling. If the melt is sufficiently supercooled the viscosity reaches such a magnitude that crystallization is kinetically inhibited and the melt will solidify in an amorphous or glassy structure. The temperature at which point crystallization is made impossible is called the glass transition temperature. The glass transition temperature is different for different materials. The invention disclosed herein can therefore be used to supercool a metal/alloy melt completely through the glass transition temperature, thus producing a glassy material without the use of any methods in the prior art. Alternatively, it can be used in conjunction with prior methods to substantially reduce the amount of heat that needs to be extracted through rapid quenching in order to produce a glassy material.

The method to supercool disclosed herein has been proven robust enough to work in environments replete with nucleation sites, where otherwise heterogeneous nucleation would result in crystallization. Moreover, the method has been proven for more than one metal system, thereby also proving the versatility of the method.

In general, the change in heat content with time of the treated metal sample can be denoted U. Assuming that the applied electromagnetic (EM) field is the only source of heat in the metal, the heat source per unit time can be denoted S, where S is a function which increases with the strength of the applied field. For a given set-up, the heat losses from the metal sample comprises conductive losses, convective losses and radiative losses, collectively denoted Q. Combined, the heat balance can be stated as

ΔU=−Q+S.

When the metal is at steady-state at some temperature T, there is no change in heat content of the metal, i.e. ΔU=0, and the losses are balanced by the source, i.e. Q=S. Increasing the heat extracted per unit time, that is, increasing Q, requires a stronger field, that is, a greater S, to achieve a steady state at the same temperature T. Experimental evidence disclosed herein shows that the field strength determined by coil current in embodiments depicted in FIGS. 1,4, 9, 10, and 11, and by supplied current in embodiments depicted in FIGS. 5 and 8, in addition to supplying heat to the metal sample suppresses crystallization.

Referring now to FIG. 1, In order to supercool a metal in one embodiment a metal 10 is first contained in a primary containment vessel 1, such as a vitreous quartz crucible, and is heated inductively by passing a current through a work coil 2 that surrounds the metal 10 to be melted, although the metal may be melted by other means without departing from the scope of the invention. In one embodiment, the current to the coil is generated by a 250-300 kHz generator and is controlled by current control module 8 such that the current to the coil 2 is of a sufficient amplitude to heat the metal 10 to its molten state, at which time the current to the coil 2 is reduced to create conditions for a cool-down process.

Operating at lower frequencies would increase the skin depth, i.e. the penetration of the field into the metal and is therefore foreseen to increase the crystallization suppression. In order to purge the oxygen from the system, a bath of either helium, argon, nitrogen,

neon or other inert gas is introduced into the primary containment vessel 1. It has been documented in the literature that metal oxides present in the melt can impede the ability to supercool. Therefore, a reducing agent can be introduced into the primary containment vessel 1 in amounts sufficient to remove any pre-existing oxides from the metal as well as keeping new oxides from forming. Examples of such a reducing agent include, but are not limited to, graphite powder or a small addition of hydrogen in the purge gas.

In another embodiment, a secondary containment vessel such as an external chamber 3 may surround the vessel 1 and may be used to contain an optional temperature control medium 5 such as water, steam, alcohol or oils to control the temperature exterior to the vessel 1.

In one embodiment, a temperature sensor including an optical fiber 6 runs from an optical pyrometer 7 to the melt since it is desirable to measure the temperature of the melt as it transitions from a molten state to a glassy or solid state.

As illustrated in FIG. 2, the current to the working coil in FIG. 1 begins high enough to bring the metal 10 to a molten state and allows for the reducing agent to remove any oxides present. This current is then decreased to allow for cooling. In the case of nickel, the initial current to working coil 2 of FIG. 1, in one embodiment is 175 A and is then reduced in steps. Note that the frequency of the RF energy in the coil is between 250 and 300 kHz regardless of the current magnitude. In this step function decrease of current, the current does not go to zero, but rather steps down to a level where the metal is supercooled without crystallizing.

During experimentation, it was found that reducing energy in the coil from 175 A to 107 A resulted in a supercooling to a thermally steady state. In one experiment, a thermal steady state was achieved at 290° K below the melting point of nickel, for a duration of 400 seconds.

It is found that this supercooling is significantly below the melting or fusing temperature of the metal, with crystallization that normally occurs when the temperature of the melt drops by as little as 2° Celsius. The injection of energy into the melt as described herein is found to suppress crystallization.

Thus, it is a finding of this invention that supercooling of a metal can be made to occur without the use of rapid quenching. In one embodiment, this is accomplished by the introduction of electromagnetic energy into the melt during the cooling process which takes the molten metal down to supercooled temperatures below the melting temperature of the metal. It is a specific finding of this invention that the supercooled temperature can be maintained without recalescence for significant periods of time in an environment where heterogeneous nucleation is expected to rapidly crystallize the metal.

As illustrated in FIG. 3 and documented more thoroughly with experimental data presented in FIG. 20, the temperature corresponding to the current levels of FIG. 2 goes from the melting temperature of pure nickel, to 280° K degrees below the melting temperature without crystallization as evidenced by recalescence. Normally the melt would start to crystallize when going below the melting temperature by as little as 2° Celsius. It is been found that by maintaining the coil 2 as described above during a radiative cool down of the melt, a supercooled state for the nickel is maintained, in one embodiment for over 2000 seconds. Thus, a supercooled thermal steady state for the cooling melt was achieved at supercooling levels of ΔT>280° K for an extended period of time.

As illustrated in FIG. 4, what is described is a system for further increasing the degree of supercooling utilizing the equipment and procedures associated with FIG. 1. This is done by providing an additional cooling modality and by increasing the energy injected into the melt by an amount equal to the energy removed by the cooling modality. By being able to add more energy into the melt as the melt cools down, crystallization is even further suppressed. This makes possible an increase in the depth of the supercooling ΔT, the difference in temperature between the normal melting point of the metal and the temperature associated with a supercooled steady state. This crystallization suppression effect is documented more thoroughly with experimental data presented in FIG. 21 in which apparatus similar to that shown in FIG. 4 is provided with a heat sink in the form of an aluminum surface in thermal contact with the bottom of the primary vessel utilized to hold the melt. This heat sink or additional cooling modality is called a “cold foot”, and is placed in thermal contact with the primary containment vessel 1 of metal 10. Current totaling 400 A is now required to bring the nickel into the molten state. It is found that a current of around 250 A now allows the system to settle at a thermal steady state at a supercooling temperature of more than 295° K below normal solidification temperature for an extended period of time. Note that the heat sink 7 includes a chamber 9, into which a coolant 11 is introduced through conduit 13, exits the conduit as illustrated at 15 and is ported out of chamber 9 as illustrated at 17.

In one embodiment of the subject invention, pure nickel at approximately 1.7 g was used for metal 10. The experiments were performed in a quartz crucible in an argon-hydrogen mixture atmosphere at ambient pressures. Conventional theory implies that supercooling a sample of this size to levels ΔT of more than 300° K should not, as a practical matter, be feasible unless all heterogeneous nucleation sites are removed or rapid quenching is employed. Nonetheless, modest cooling rates of less than 20° K./s have been found to achieve the stated levels of supercooling and there are no special actions taken to remove nucleating agents other than the reducing environment. Moreover, the melt was then maintained at steady state at this level of supercooling for extended periods of time.

The method of supercooling metals requires neither: 1) extreme cooling rates, nor 2) the need for immaculate environments, nor 3) the need for complex alloy structures, nor 4) the need for ultra-small or ultra-thin samples, thereby removing many or most of the practical limitations to make metal glasses.

Significant levels or depths of supercooling of various metals such as pure nickel, pure cobalt, pure copper, as well as nickel plus carbon or iron plus carbon have been achieved. The actual method is very straightforward. Simply exposing the melt to injected energy at power levels below those corresponding to the melting temperature of the metal acts to suppress crystallization of the molten metal. Other energy sources may be utilized in place of the energized coil. It is clear that the strength of the applied field relative to heat removal rate is a key determinant of the levels of supercooling that can be attained. This can be seen in comparing FIGS. 17, 18A and 18B. These Figures correspond to the setup in FIG. 4 where a heat sink requires a stronger field and actually allows the liquid to reach lower temperatures without crystallizing.

Referring to FIG. 5 what is shown is a still further alternative embodiment. Rather than using inductive heating to melt the pure metal, a pure metal rod 20 is inserted into a nonconductive mold 22. The metal rod within the nonconductive mold 22 is melted in a reducing environment through the utilization of alternating or direct current from a source 26 such that the metal rod 20 is melted within the nonconductive mold 22. The current applied to the rod 20 is shown in FIG. 6 to be a step function such that when the current is reduced the molten metal cools in the presence of injected energy from current source 26. The temperature profile is shown in FIG. 7, in which after melting and the decrease of current to bring the metal to the fusion temperature, a supercooled state ensues due to the injection of the current. The melt is supercooled by controlling the current through the rod 20 similar to the way the current through the coil 2 was controlled relative to FIGS. 1 and 4.

Referring to FIG. 8, rather than using a rod, a metal plate 28 is contained within a rectilinear nonconductive mold 32. The metal plate is contacted at either end by current from a current generator 34 and operates in the same manner as that described in connection with FIGS. 5, 6 and 7.

Referring to FIG. 17, a diagram of the experimental apparatus that refers to the embodiment in FIG. 5, in which the metal sample 1 is placed in a primary containment vessel made of vitreous quartz. The metal is contacted by two graphite electrodes 2 supplying an electric current through the metal. The choice of electrode material depends on thermal stability at the operating temperature i.e. between 50 and 200° C. above the melting point of the primary metal, dissolution in the melt as well as thermal and electrical conductivity. The primary containment vessel is housed in a secondary containment vessel 3 sealed against one end of each of the electrodes. The electrodes are connected to leads 4 which connect to a DC power supply. The secondary containment vessel is purged with either an inert gas or an inert gas with added hydrogen through the inlet/outlets 5. The electrodes not only conduct an electrical current they also conduct heat away from the molten metal. This requires cooling at the leads which is provided by blowing compressed air through the nozzles 6.

As to the embodiment shown in FIG. 1, and referring now to FIG. 9 a so-called short coil may be utilized in which height of the coil h is approximately equal to the diameter d of the coil.

Alternatively, and referring to FIG. 10, in a long coil embodiment, the ratio of h/d is greater than 5.

In a further embodiment and referring to FIG. 11, two coils are utilized, with a sample being placed between the two coils. The applied currents are applied as illustrated with a possible phase shift between the currents in the two sets of coils from zero through π/2.

Experimental Setup

Referring to FIGS. 12-17, what is shown is a description of the experimental setup used in the subject invention.

Referring now to FIG. 12, primary containment vessel 1 comprising a tube of non-suspecting material has a working coil 2 located at the distal end thereof. Coil 2 is coupled to an induction power supply 20 as illustrated. The primary containment vessel is located in an enclosure 42 that constitutes a secondary containment vessel 3, with the tube comprising primary containment vessel 1 in communication with a concentric gas delivery conduit 22 through which an optical pyrometer fiber optic probe 6 extends. Conduit 22 includes a gas inlet fitting 24 to which is attached a gas inlet line 26 carrying an air purifying medium such as helium, argon nitrogen or neon. Conduit 22 contains concentric internal conduits such that gas inlet fitting 24 is connected to the inner most internal conduit, whereas a gas outlet fitting 28 connected to a gas outlet line 30 communicates with the outer internal conduit such that purging gas is directed downwardly into the primary containment vessel 1 and is removed from the primary containment vessel via gas outlet line 30. It will be noted that fiber optic probe 6 contains a viewport 29.

Referring to FIG. 13, it will be seen that secondary containment vessel 3includes a cabinet 42 that encloses primary containment vessel 1 and coil 2 with an optional control medium introduced into the cabinet at orifice 30 which circulates water, steam, alcohol or oil through the cabinet around primary containment vessel 1 and is exhausted out through exit orifice 32.

Referring to FIG. 14, primary containment vessel 1, with working coil 2 is provided with aluminum heatsink 34 at the base 36 of the primary containment vessel tube. An inlet cooling medium 11 is introduced into the aluminum heatsink 34 and exits through an outlet cooling medium hose 36 coupled to an outlet fitting 38 which communicates with the interior of heatsink 34. The cooling medium therefore provides a heat transfer medium to cool the heatsink.

Referring to FIG. 15 in which like elements have like reference characters, the secondary containment inlet line 40 is shown connected to orifice 30 in the secondary containment vessel 3 cabinet 42 here shown with a closing cover 44 secured to the primary containment vessel cabinet 42. Also shown is induction work head 20 coupled to the work coil through an electrical cabling conduit 46 and a pass-through connection 48 in cabinet 42. Note that a lens 50 is located atop an optical viewport 52 which is coupled by fiber optic cable 6 to the optical pyrometer.

Referring to FIG. 16, what is shown are pair of pressure regulators 56 and 58 to regulate the pressure in the primary and secondary containment vessels and are respectively coupled to primary containment inlet line 26 and secondary containment inlet line 40 to regulate the pressure of the gases introduced into the primary and secondary containment vessels.

Referring to FIG. 17, the direct-current embodiment of the subject invention includes a primary containment vessel 1 in which a metal bar 20 is located in an open-ended tube comprising an open-ended containment vessel 1 and is contacted at either end by electrodes 60 in turn electrically connected to electrical terminals 62 and DC connection cables 64 for supplying current through metal bar 20. Fitting 66 is utilized to connect a gas inlet 70 carrying a reducing gas to a containment vessel 72 shown in dotted outline which surrounds primary containment vessel 1, whereas fitting 68 couples a gas outlet 76 for the removal of the reducing gas in the containment vessel. Due to the open-ended nature of tube 1 the gas in containment vessel 72 circulates within this tube.

It will be appreciated that fittings 66 and 68 as well as electrical terminals 62 will be exceedingly hot and are cooled in one embodiment by compressed air cooling nozzles 80 and 82.

More particularly, and as to the operation of the various elements what is described is the function and the characteristics of the indicated parts of the system.

Primary containment 1. Containment vessel 1 consists of a tube of a non-suspecting material with a closed bottom end. Fused quartz with its excellent thermal shock resistance and high operating temperature has been the material of choice but zirconia and alumina have also been used with successful results. An additional tube (open-ended) made of non-suspecting material (e.g. alumina or quartz) mounted inside the quartz tube acts as a gas outlet to achieve gas circulation closer to the metal surface. The suspecting metal sample, together with any possible additive elements, is placed at the bottom of the quartz tube which is collinear with the center axis of the coil. Moreover, the sample inside the primary containment is placed at the center of the work coil. Several different dimensions on the closed-end tube have been employed to date, including but not limited to: 1) o.d. 19 mm, i.d. 13 mm, 2) o.d. 18 mm i.d. 15 mm and 3) o.d. 12 mm i.d. 9 mm.

Work coil 2. The work coil 2 depicted is made out of copper, allowing for internal water circulation for cooling purposes. The coil is painted with a dielectric material to prevent short circuiting. The coil consists of four windings and has an inner diameter measuring 22 mm and a length of about 20 mm. For most of the samples treated the coil operated at 160-180 A during the heating steps (around 60 s) and around 110-140 A during the cooling step when no heat sink is present. With a heat sink present the current levels to the coil during the heating step is around 350-450 A and 180-250 A during the cooling step

Secondary containment 3. The working coil and primary containment is housed in a secondary containment cabinet made of non-suspecting material capable of maintaining a small to moderate positive pressure in terms of inches of water.

Primary containment gas circulation 4. The gas circulation for the primary containment vessel is provided by inlet flow regulated with a mass flow controller connected to a PLC/desktop computer. This circulation allows for convective cooling of the top side of the sample. Various inert cooling media have been employed (e.g. He, N2, Ar) as well as non-inert gases such as CO and a mixture of Ar/H2 (4% H2) at different flow rates. Most experiments are performed using Ar or Ar/H2 mixture at a modest flow rate of 0.25 l/min. The back pressure on the primary gas flow is kept slightly higher at about an inch as compared to that on the secondary containment in order to detect cracks and possible leaks in the primary containment.

Secondary containment gas circulation 5. Gas circulation for the secondary containment vessel involves circulating a cooling medium in the secondary containment that offers some means of regulating the temperature outside the primary containment vessel and therefore also heat transfer out of same. Those experiments that have utilized the secondary containment are run with 1-4 l/min flow of N2. The back pressure on the secondary gas flow is kept slightly lower at about an inch as compared to that on the primary flow in order to detect cracks and possible leaks of the primary containment.

Optics for IR pyrometer 6. The infrared radiation of the suspecting sample is collected via a light pipe sapphire or quartz and transmitted to a detector via a quartz fiber optic cable. The light pipe is directed down through the open-ended interior tube in the primary containment vessel. Since quartz has a high transmissivity in the operating infrared spectrum of the detector, a lens can also be aimed from outside the primary containment vessel. Aiming from the outside alleviates some of the interference caused by turbulence in the melt.

Optical pyrometer 7. A dual-wavelength IR pyrometer made by Williamson™ with a nominal operating range of 480-1750° C. may be used to observe the temperature of the melt. Note that crystallization of a supercooled melt can be observed visually by the sample illuminating rapidly indicating recalescence, as well as a spike in the temperature measurements.

Power source 8. In one embodiment the power sources made by Ambrell™ and is capable of delivering up to 10 kW (max 600 A) to the work coil. The frequency is internally modulated for circuit balance depending on the load, generally between 250-300 kHz.

Heatsink 9. Placing a water-cooled aluminum heatsink in thermal contact with the primary containment vessel increases the conductive heat loss from the metal. Thus to reach the same sample temperature as in a setup without the heatsink a stronger field or higher coil-current is required. Aluminum is used because of its excellent thermal conductivity.

Circulating water 10. The aluminum heatsink is cooled by circulating water in thermal contact with the aluminum. The incoming water temperature is regulated at around 22 degrees C. using a PolyScience constant temperature circulator and the flow rates used were around 0.4-1 gallons per minute.

Method

The supercooling experiments performed to date mainly utilize the apparatus described in FIGS. 1 and 4 in which a sample is placed in vessel 1 and is inductively heated through the utilization of induction heating coil 2, followed by reducing the current to the inductive heating coil to allow cooling. It is noted that in this embodiment, the influence produced by the coil prevents crystallization or recalescence. All of this was done in an oxygen free environment due to an oxygen-purging inert or reducing gas stream 4 which filled the vessel 1 above the metal 10. A temperature control stream occupies the space between vessel 1 and surrounding enclosure 3 to manage the heat exterior to vessel 1 safely, although such temperature control medium is considered optional.

After the metal had been brought to its molten state due to the induction heating associated with the application of RF current from current control module 8 at 250-300 kHz and 175 A, or 400 A if the heat sink is employed, the current is reduced. The magnitude of the current supplied to the induction coil is low enough to allow the metal to cool below the melting point but at the same time sufficient to inhibit crystallization. As a result, the current can keep the melt in a supercooled state for an extended period of time. For a 4.39 g nickel sample of 4N5 purity a current of 108 A during cooling resulted in a steady state supercooling of more than ΔT=290° K.

More particularly, a 10 kW power source is used to generate a radio frequency AC current in an induction coil, 22 mm i.d, 20 mm height with four windings. The frequency is load-dependent and is fixed to balance the internal circuitry of the power source at around 250-300 kHz. The power source is controlled remotely through a PLC/desktop computer. The metals being supercooled were placed at the bottom of a closed-end vitreous quartz tube, which subsequently was mounted with the suspecting material, metal sample, at the center of the coil. This assembly allows for visual inspection of the sample during the process.

The metal temperature was measured with a dual-wavelength pyrometer connected via a quartz fiber optic cable to the detector. The temperature readings were fed to a computer via a PLC, allowing for readings of the temperature trend in real time.

Definitive confirmation of a glassy state cannot be done in real time. However, temperature readings supported by visual inspection of the metal during cooling can offer strong indications as to whether crystallization has occurred or not. Crystallization of the sample is either evidenced by recalescence, both a visually noticeable event involving bright illumination, as well as detectable as a spike in temperature readings, or through an observation of a thermal arrest of the cooling temperature curve at the equilibrium melting point.

The induction heating is controlled through the current delivered to the coil. The high sensitivity with respect to position of the metal in the alternating field requires a careful calibration of the current at the start of each run. The quartz tube is purged with an inert or reducing gas, e.g. argon or argon-hydrogen mixture prior to initial heating and a minor gas flow is maintained throughout the experiment. After the metal has been melted initially, the system is allowed to soak in the reducing environment before the power to the coil is dropped to levels corresponding to temperatures below the melting point. Once a run is terminated the samples are left to cool to approximately room temperature before being extracted from the reactor. Weight measurements of the sample pre and post experiments along with XRF analysis of the sample allows for a determination of any possible foreign elements present.

Experimental Results

Normally the heating and cooling through the phase transition shows clearly the latent heat of the phase transition on both the heating step and the cooling step in the cycle. This is evidenced by an approximate isotherm or thermal arrest in the temperature trend at the equilibrium melting/solidification temperature. When an electromagnetic field is imposed during cooling, the phase transition can be suppressed as evidenced by a lack of thermal arrest when passing through the melting point as well as recalescence below the equilibrium melting temperature, leading to supercooling as shown in FIGS. 17, 20 and 22.

In FIGS. 18 and 20 temperature/current readings establish a supercooled state for extended period of time. As shown in FIG. 20, the readings were recorded from an experiment performed with nickel, 4N5 (99.995%) purity, whereas as shown in FIG. 18 the readings were recorded from an experiment in which carbon was added at an amount of 0.5 wt. % The experimental setup used in this FIG. 18 experiment involved a water-cooled aluminum heat sink in thermal contact with the bottom of the primary containment as the illustrated in FIG. 4. The dotted vertical lines in the temperature plot indicate times when current to the coil is decreased. Similarly, dashed vertical lines indicate an increase of current to the coil. The current plot in FIG. 18 shows only the current during cooling to offer higher resolution. The current during heating was maintained at 360-365 A and regulated only to keep the temperature of the superheated melt consistent cycle to cycle. As can be seen in FIG. 18 by increasing the coil current during cooling the melt reaches deeper levels of supercooling before recalescing and eventually the current is high enough to prevent recalescence altogether. Further increases in the current beyond this minimum critical current raises the minimum temperature reached during the cooling step since more power is being transferred to the melt. FIG. 19B illustrates this trade-off.

FIG. 19A shows the time that the metal melt spends in a state of supercooling during each cooling step. It is noticeable that a higher current allows for longer time in a supercooled state. Note that the maximum time in each cooling step was capped at 90 seconds which is why the trend in FIG. 19A plateaus.

A system comprising pure nickel behaves qualitatively different from that of nickel with added carbon. In order to demonstrate the significant effect of the crystallization suppression provided by the induced electrical currents comparisons were made to nucleation rates at various temperatures for pure nickel as found in the prior art. For pure nickel and returning to FIG. 20, at temperatures of around 300° K below the melting point of nickel homogeneous nucleation is expected to occur. In the reactor environments of embodiments 1 and 4 heterogeneous nucleation is expected at temperatures well above this level and the fact that the indicated level of undercooling is achieved with no other measures than a reducing environment is considered remarkable. Nonetheless, it is conventionally believed that a single nucleation event is enough to trigger crystallization of an entire sample meaning that the probability for crystallization scales with the volume/mass of the sample.

Referring to FIG. 21, the time, as inferred from published data in the prior art, before homogeneous nucleation is expected to crystallize a sample of a given mass and at a given temperature is marked with a dashed line. As seen from the solid line, for a 0.45 g nickel sample the observed nucleation rates are significantly slower. Interestingly, the deviation from expected rates is greater for a smaller sample. This observation is attributed to the penetration depth of the electromagnetic field and the corresponding spatial distribution of the induced electrical currents. Using the same frequency of the field for different sample sizes the proportion of the sample that is exposed to electrical currents is greater for a smaller sample. Since the induced electrical currents are believed to be the main factor in suppressing crystallization this explains why the deviation from expected behavior is greater for a smaller sample.

Data from another experiment performed on a system comprised by iron and carbon can be seen in FIGS. 22, 23A and 23B. The same effect of a minimum critical current to reach a maximum stable level of supercooling as in FIGS. 18, 19A and 19B for nickel plus carbon is noticeable here as well. Thus the method is not specific to only nickel but applicable to other systems as well.

Data from an experiment on nickel using direct current can be seen in FIG. 24. Two cooling cycles are shown in the temperature versus time and current versus time graph. The behavior is markedly different from data gathered using the induction coil. For instance there is a noticeable thermal arrest on both cooling curves indicating partial normal solidification. However, and importantly, after this thermal arrest if the current through the sample is cut completely or stepped down there is a noticeable spike in the temperature readings which is proof of solidification of an undercooled melt. This offers a proof of concept in that at least part of the sample was maintained in a liquid state below liquidus by the supplied currents.

The main scientific and expected difference between the DC and the induction setup is that the current density in the former should be uniform through the metal as there are no frequency or skin depth issues with a direct current. That is an advantage of the DC as the currents reach even the center of the sample. The disadvantage is that the electrodes touching the melt introduce potent nucleation sites that a vitreous quartz tube does not, at least not to the same extent. An interpretation of the temperature graphs is that there is a part of the metal that actually solidifies normally as indicated by the thermal arrest. However, the observed recalescence in the graphs suggest that part of the liquid can nonetheless stay supercooled.

From an engineering perspective there is also a heat transfer issue to take into account. Whatever electrode material that is chosen, graphite in the illustrated case, will not only be a good electric conductor but also a good thermal conductor. Thus, with the relatively small dimensions that have been tested so far there are significant thermal gradients.

Moreover, experimental experience tells us that the ability of the sample to supercool is sensitive to both the integrity of the quartz crucible and the amounts of oxides present in the sample. If the quartz tube chips and releases grains of quartz into the melt, supercooling is difficult to achieve. The role of oxides in catalyzing nucleation is well known in the literature. A common mitigation practice for this problem is to use some form of gettering system, a component that removes trace impurities, e.g. oxygen, from a gas stream, to assure a very low oxygen partial pressure in the reactor. Also, ingots are typically etched with acids to remove surface oxides present on the ingot. Instead, a reducing agent in the form of carbon, eg. graphite, is introduced in the reactor or hydrogen is mixed into the purge gas. The experimental procedure is initiated by up to 30 minutes of soaking at a temperature above the melting point of the metal to ensure homogeneous distribution of the carbon, if added, and reduction of most oxides present.

Lastly, if a metal is allowed to crystallize when subjected to an electromagnetic field the crystal structure can be manipulated, for instance crystal size and orientation. In FIG. 25A an XRD-pattern from a scan taken on the surface of a treated metal sample reveals a directional solidification as evidenced by the missing (1,1,1), (2,0,0) and (2,2,2) diffraction peaks of nickel. As shown in FIG. 25B, those peaks always show up when scanning powder filed from the same sample. Thus, the field aligns the crystal structure during solidification. Such manipulation can result in enhanced properties like electrical and thermal conductivity.

As seen in FIG. 26, a sample was allowed to solidify at significant undercooling but with no current applied. In this instance, the current to the coil sample was cut causing instant solidification as exhibited by recalescence of the sample as seen by the spike. As described hereinabove, the application of electromagnetic current significantly eliminates or delays the onset of recalescence.

Referring to the photographs of FIGS. 27 and 28, of the tops of samples in which a field was not applied and in which a field was applied are shown. The main distinguishing feature is that by cutting the field completely one removes the pinching effect on the metal caused by Lorentz forces on the induced current. As a result, and is shown in FIG. 27, the sample starts collapsing under gravity and crystallization happens before the surface has acquiesced giving the jagged appearance on the top of the sample in FIG. 27 compared to the smooth surface in FIG. 28.

Referring to FIG. 29, and XRD plot of a sample of NiNb is shown indicating a substantial amorphous phase in the material as indicated by the amorphous hump at low 2theta angles. This amorphous phase was obtained utilizing the apparatus of FIG. 4 in which the melted sample was cooled in its supercooled state down to ambient temperature. The result is at least a portion of the sample having amorphous phase characteristics.

Referring to FIG. 30, likewise an XRD plot of a sample of NiP is shown indicating a substantial amorphous phase in the material as indicated by the amorphous hump at low 2theta angles. This amorphous phase was obtained utilizing the apparatus of FIG. 4 in which the melted sample was cooled in its supercooled state down to ambient temperature. The result is at least a portion of the sample having an amorphous phase characteristic.

In short, for these two samples the result is the same in that a substantial amorphous phase exists for a molten metal cooled down while at the same time injecting energy sufficient to prevent crystallization during supercooling, with the supercooled material being cooled to ambient temperatures to achieve and amorphous metal at room temperature.

In summary, the results show the ability to cool a melt in a controlled fashion to significant levels of supercooling without crystallization as evidenced by absence of both thermal arrest and recalescence. Moreover, XRD analysis on different samples show presence of an amorphous phase at room temperature. If a strong enough electromagnetic field can be applied during cooling of an appropriate material such that the metal melt reaches the glass transition temperature then four of the main problems in producing glassy metals are solved: 1) the ability to generate amorphous metals in a “dirty” reactor containing numerous nucleation sites, 2) the ability to make amorphous metals without requiring rapid quenching and 3) the ability to increase the size of the glassy metal produced since no quenching is required. In addition to these three problems 4) it is likely that this method allows for a greatly increased variety of compositions of the glassy metals to be produced including pure metals like those used in the above experiments.

Even if the strongest field attainable is not enough to reach the glass transition temperature for a given metal system with this method alone, combining this method with known methods will lower the threshold of the amount of heat that needs to be extracted through quenching. Thus, the three problems mentioned above will at least be mitigated by combining the subject method with prior art methods of producing glassy metals.

Repeated experiments have shown the ability to supercool a system comprising nickel, nickel plus small amounts of carbon, iron plus carbon, pure cobalt and pure copper respectively by controlling the amount of energy injected into the melt. This opens the door to making glassy metals from far more starting metals and metal alloys than currently thought possible. The data presented were obtained with the metal in contact with a quartz crucible meaning that the environment offers a multitude of possible nucleation sites. Prior efforts to make amorphous metals often require much more complex reactors which attempt to limit or eliminate heterogeneous nucleation sites. Finally, the possibility of sustaining a metal in thermal steady state at significant supercooled temperatures further indicates the possibility to control the processing of glassy metals.

The use of electromagnetic fields during cooling strongly appears to be the primary factor suppressing normal solidification as well as recalescence. The hypothesis at this point is that the anisotropy of a supercooled melt with small solid clusters could preferentially absorb the energy from the injected EM field at the solid/liquid interface, thereby melting the formed clusters.

While the current efforts have employed an induction coil to expose the melt to an electromagnetic field, a current applied directly to the material has shown to qualitatively yield the same effect of suppressing crystallization. Such a direct coupling would substantially increase the ability to mold glassy metals into a given shape, e.g. sheets, rods, beams and other geometries.

While the present invention has been described in connection with the preferred embodiments of the various Figures, it is to be understood that other similar embodiments may be used or modifications or additions may be made to the described embodiment for performing the same function of the present invention without deviating therefrom. Therefore, the present invention should not be limited to any single embodiment, but rather construed in breadth and scope in accordance with the recitation of the appended claims.

METHOD AND APPARATUS FOR SUPERCOOLING OF METAL/ALLOY MELTS AND FOR THE FORMATION OF AMORPHOUS METALS THEREFROM

Information

Publication Number

Date Filed

Date Published

Inventors

Original Assignees

CPC

International Classifications

Abstract

Description

Claims

RELATED APPLICATIONS

Provisional Applications (1)