Many refractory metal powders (Ta, Nb, Ti, Zr, etc) are made by hydriding an ingot of a specific material. Hydriding embrittles the metal allowing it to be easily comminuted or ground into fine powder. The powder is then loaded in trays and placed in a vacuum vessel, and in a batch process is raised to a temperature under vacuum where the hydride decomposes and the hydrogen is driven off. In principle, once the hydrogen is removed the powder regains its ductility and other desirable mechanical properties. However, in removing the hydrogen, the metal powder can become very reactive and sensitive to oxygen pickup. The finer the powder, the greater the total surface area, and hence the more reactive and sensitive the powder is to oxygen pickup. For tantalum powder of approximately 10-44 microns in size after dehydriding and conversion to a true Ta powder the oxygen pickup can be 300 ppm and even greater. This amount of oxygen again embrittles the material and greatly reduces its useful applications.
To prevent this oxygen pickup the hydride powder must be converted to a bulk, non hydride solid which greatly decreases the surface area in the shortest time possible while in an inert environment. The dehydriding step is necessary since as mentioned previously the hydride is brittle, hard and does not bond well with other powder particles to make usable macroscopic or bulk objects. The problem this invention solves is that of converting the hydride powder to a bulk metal solid with substantially no oxygen pickup.
We have discovered how to go directly from tantalum hydride powder directly to bulk pieces of tantalum a very short time frame (a few tenths of a second, or even less). This is done in a dynamic, continuous process as opposed to conventional static, batch processing. The process is conducted at positive pressure and preferably high pressure, as opposed to vacuum. The dehydriding process occurs rapidly in a completely inert environment on a powder particle by powder particle basis with consolidation occurring immediately at the end of the dehydriding process. Once consolidated the problem of oxygen pick up is eliminated by the huge reduction in surface area that occurs with the consolidation of fine powder into a bulk object.
The equilibrium solubility of hydrogen in metal is a function of temperature. For many metals the solubility decreases markedly with increased temperature and in fact if a hydrogen saturated metal has its temperature raised the hydrogen will gradually diffuse out of the metal until a new lower hydrogen concentration is reached. The basis for this is shown clearly in
Vacuum is normally applied in the dehydride process to keep a low partial pressure of hydrogen in the local environment to prevent Le Chateliers's principle from slowing and stopping the dehydriding. We have found we can suppress the local hydrogen partial pressure not just by vacuum but also by surrounding the powder particles with a flowing gas. And further, the use of a high pressure flowing gas advantageously allows the particles to be accelerated to a high velocity and cooled to a low temperature later in the process
What is not known from
Information from diffusion calculations are summarized in Table 1. The calculations were made assuming a starting concentration of 4000 ppm hydrogen and a final concentration of 10 ppm hydrogen. The calculations are approximate and not an exact solution. What is readily apparent from Table 1 is that hydrogen is extremely mobile in tantalum even at low temperatures and that for the particle sizes (<40 microns) typically used in low temperature (600-1000 C) spraying operations diffusion times are in the order of a few thousandths of a second. In fact even for very large powder, 150 microns, it is less than half a second at process temperatures of 600 C and above. In other words, in a dynamic process the powder needs to be at temperature only a very short time be dehydrided to 10 ppm. In fact the time requirement is even shorter because when the hydrogen content is less than approximately 50 ppm hydrogen no longer causes embrittlement or excessive work hardening.
The device consists of a section comprised of the well known De Laval nozzle (converging-diverging nozzle) used for accelerating gases to high velocity, a preheat -mixing section before or upstream from the inlet to the converging section and a substrate in close proximity to the exit of the diverging section to impinge the powder particles on and build a solid, dense structure of the desired metal.
An advantage of the process of this invention is that the process is carried out under positive pressure rather than under a vacuum. Utilization of positive pressure provides for increased velocity of the powder through the device and also facilitates or permits the spraying of the powder onto the substrate. Another advantage is that the powder is immediately desified and compacted into a bulk solid greatly reducing its surface area and the problem of oxygen pickup after dehydriding.
Use of the De Laval nozzle is important to the effective of operation of this invention. The nozzle is designed to maximize the efficiency with which the potential energy of the compressed gas is converted into high gas velocity at the exit of the nozzle. The gas velocity is used to accelerate the powder to high velocity as well such that upon impact the powder welds itself to the substrate. But here the De Laval nozzle also plays another key role. As the compressed gas passes through the nozzle orifice its temperature rapidly decreases due to the well known Joule Thompson effect and further expansion. As an example for nitrogen gas at 30 bar and 650 C before the orifice when isentropically expanded through a nozzle of this type will reach an exit velocity of approximately 1000 m/s and decrease in temperature to approximately 75 C. In the region of the chamber at 650 C the hydrogen in the tantalum would have a maximum solubility of 360 ppm (in one atmosphere of hydrogen) and it would take less than approximately 0.005 seconds for the hydrogen to diffuse out of tantalum hydride previously charged to 4000 ppm. But, the powder is not in one atmosphere of hydrogen, by using a nitrogen gas for conveying the powder, it is in a nitrogen atmosphere and hence the ppm level reached would be expected to be significantly lower. In the cold region at 75 C the solubility would increase to approximately 4300 ppm. But, the diffusion analysis shows that even in a high concentration of hydrogen it would take approximately 9 milliseconds for the hydrogen to diffuse back in and because the particle is traveling through this region at near average gas velocity of 600 m/s its actual residence time is only about 0.4 milliseconds. Hence even in a pure hydrogen atmosphere there is insufficient residence time for the particle to reabsorb hydrogen. The amount reabsorbed is diminished even further since a mass balance of the powder flow of 4 kg/hr in a typical gas flow of 90 kg/hr shows that even if all the hydrogen were evolved from the hydride, the surrounding atmosphere would contain only 1.8% hydrogen further reducing the hydrogen pickup due to statistical gas dynamics.
With reference to
One aspect of the invention broadly relates to a process and another aspect of the invention relates to a device for dehydriding refractory metal powders. Such device includes a preheat chamber at the inlet to a converging/diverging nozzle for retaining the metal powder fully heated in a hot zone to allow diffusion of hydrogen out of the powder. The nozzle includes a cooling chamber downstream from the orifice in the diverging portion of the device. In this cooling chamber the temperature rapidly decreases while the velocity of the gas/particles (i.e. carrier gas and powder) rapidly increases. Substantial re-absorption of the hydrogen by the powder is prevented. Finally, the powder is impacted against and builds a dense deposit on a substrate located at the exit of the nozzle to dynamically dehydride the metal powder and consolidate it into a high density metal on the substrate.
Cooling in the nozzle is due to the Joule Thompson effect. The operation of the device permits the dehydriding process to be a dynamic continuous process as opposed to one which is static or a batch processing. The process is conducted at positive and preferably high pressure, as opposed to vacuum and occurs rapidly in a completely inert or non reactive environment.
The inert environment is created by using any suitable inert gas such as, helium or argon or a nonreactive gas such as nitrogen as the carrier gas fed through the nozzle. In the preferred practice of this invention an inert gas environment is maintained throughout the length of the device from and including the powder feeder, through the preheat chamber to the exit of the nozzle. In a preferred practice of the invention the substrate chamber also has an inert atmosphere, although the invention could be practiced where the substrate chamber is exposed to the normal (i.e. not-inert) atmosphere environment. Preferably the substrate is located within about 10 millimeters of the exit. Longer or shorter distances can be used within this invention. If there is a larger gap between the substrate chamber and the exit, this would decrease the effectiveness of the powder being consolidated into the high density metal on the substrate. Even longer distances would result in a loose dehydrided powder rather than a dense deposit.
Experimental Support
The results of using this invention to process tantalum hydride powder −44+20 microns in size using a Kinetiks 4000 system (this is a standard unit sold for cold spray applications that allows heating of the gas) and the conditions used are shown in Table II. Two separate experiments were conducted using two types of gas at different preheat temperatures. The tantalum hydride powder all came from the same lot, was sieved to a size range of −44+20 microns and had a measured hydrogen content of approximately 3900 ppm prior to being processed. Processing reduced the hydrogen content approximately 2 orders of magnitude to approximately 50-90 ppm. All this was attained without optimizing the gun design. The residence time of the powder in the hot inlet section of the gun (where dehydriding occurs) is estimated to be less than 0.1 seconds, residence time in the cold section is estimated to be less than 0.5 milliseconds (where the danger of hydrogen pickup and oxidation occurs). One method of optimization would simply be to extend the length of the hot/preheat zone of the gun, add a preheater to the powder delivery tube just before the inlet to the gun or simply raise the temperature that the powder was heated to.
As noted the above experiment was performed using a standard Kinetecs 400 system, and was able to reduce hydrogen content for tantalum hydride to the 50-90 PPM level for the powder size tested. I.e. the residence time in hot sections of the standard gun was sufficient to drive most of the hydrogen out for tantalum powders less than 44 mictons in size.
The following example provides a means of designing the preheat or prechamber to produce even lower hydrogen content levels and to accommodate dehydriding larger powders that would require longer times at temperature. The results of the calculations are shown in table III below
The calculations are for tantalum and niobium powders, 1O and 400 microns in diameter, that have been assumed to be initially charged with 4000 and 9900 ppm hydrogen respectively.
The powders are preheated to 750 C. The required times at temperature to dehydride to 100, 50 and 10 ppm hydrogen are shown in the table are shown. The goal is to reduce hydrogen content to 10 ppm so the prechamber length is calculated as the product of the particle velocity and the required dehydriding time to attain 10 ppm. What is immediately apparent is the reaction is extremely fast, calculated prechamber lengths are extremely short (less than 1.5 mm in the longest case in this example) making it easy to use a conservative prechamber length of 10-20cm insuring that this dehydriding process is very robust in nature, easily completed before the powder enters the gun, and able to handle a wide range of process variation.