US 20020189719 A1
A nickel titanium hafnium copper thin film shape memory alloy having a composition (TiHf)5-55(NiCu)45-50 comprising about 2 atomic percent to about 10 atomic percent copper and the fabrication method of said shape memory thin film by magnstron sputtering using Kr as working gas and conducting the deposition at elevated substrate temperature.
1. A nickel titanium hafnium copper thin film shape memory alloy having a composition (TiHf)50-55(NiCu)45-50 comprising about 2 atomic percent to about 10 atomic percent copper.
2. The nickel titanium hafnium copper thin film shape memory alloy composition of
3. The shape memory alloy thin film composition of
4. The shape memory alloy thin film of
5. The shape memory alloy thin film of
6. The shape memory alloy thin film of
7. A magnetron sputter deposition process for producing a (NiCu)(TiHf) thin film alloy exhibiting shape memory characteristics comprising:
conducting the deposition using krypton as process gas; and,
heating a substrate onto which the thin film alloy is deposited to maintain a temperature ranging from about 350 ° C. to about 500° C. during said deposition.
8. The process of
9. A magnetron sputter deposition process for producing a thin film alloy exhibiting shape memory characteristics comprising:
conducting the deposition using krypton as process gas;
heating a substrate onto which the thin film alloy is deposited to maintain a temperature ranging from about 350 ° C. to about 500° C. during said deposition; and,
using a hot pressed powder sputter target comprising components selected from the group consisting of nickel, titanium, hafnium, and copper in proportions so as to produce a thin film alloy having a composition (TiHf)50-55(NiCu)45-50 comprising about 2 atomic percent to about 10 atomic percent copper.
10. The process of
11. The process of
 This invention relates to shape-memory alloy films. In particular, it relates to high temperature shape-memory nickel titanium hafnium copper films having low hysterisis, narrow temperature difference between Af and Mf, and more stable shape memory recovery.
 Various metallic materials capable of exhibiting shape-memory characteristics are well known in the art. These shape-memory capabilities occur as the result of the metallic alloy undergoing a reversible crystalline phase transformation from one crystalline state to another crystalline state with a change in temperature and/or external stress. In particular, it was discovered that alloys of nickel and titanium exhibited these remarkable properties of being able to undergo energetic crystalline phase changes at ambient temperatures, thus giving them a shape-memory. These alloys, if plastically deformed while cool, will revert, exerting considerable force, to their original, undeformed shape when warmed. These energetic phase transformation properties render articles made from these alloys highly useful in a variety of applications. An article made of an alloy having a shape memory can be deformed at a low temperature from its original configuration, but the article “remembers” its original shape, and returns to that shape when heated.
 For example, in nickel titanium alloys possessing shape-memory characteristics, the alloy undergoes a reversible transformation from an austenitic state to a martensitic state with a change in temperature. This transformation is often referred to as a thermoelastic martensitic transformation. The reversible transformation of the NiTi alloy between the austenite to the martensite phases occurs over two different temperature ranges which are characteristic of the specific alloy. As the alloy cools, it reaches a temperature (Ms) at which the martensite phase starts to form, and finishes the transformation at a still lower temperature (Mf). Upon reheating, it reaches a temperature (As) at which austenite begins to reform and then a temperature (Af) at which the change back to austenite is complete. In the martensitic state, the alloy can be easily deformed. When sufficient heat is applied to the deformed alloy, it reverts back to the austenitic state, and returns to its original configuration.
 Shape-memory materials previously have been produced in bulk form, in the shape of wires, rods, and plates, for utilities such as pipe couplings, electrical connectors, switches, and actuators, and the like. Actuators previously have been developed, incorporating shape-memory alloys or materials, which operate on the principal of deforming the shape-memory alloy while it is below its phase transformation temperature range and then heating it to above its transformation temperature range to recover all or part of the deformation, and, in the process of doing so, create moments of one or more mechanical elements. These actuators utilize one or more shape-memory elements produced in bulk form, and, therefore are limited in size and usefulness.
 The unique properties of shape-memory alloys further have been adapted to applications such as micro-actuators by means of thin film technology. Micro-actuators are desirable for such utilities as opening and closing valves, activating switches, and generally providing motion for micro-mechanical devices. It is reported that the advantageous performance of micro-actuators is attributed to the fact that the shape-memory effect of the stress and strain can produce substantial work per unit of volume. For example, the work output of nickel-titanium shape-memory alloy is of the order of 1 joule per gram per cycle. A shape-memory film micro-actuator measuring one square millimeter and ten microns thick is estimated to exert about 64 microjoules of work per cycle.
 The most well known and most readily available shape-memory alloy is an alloy of nickel and titanium. With a temperature change of as little as about 10° C., this alloy can exert a force of as much as 415 MPa when applied against a resistance to changing its shape from its deformation state.
 Although numerous potential applications for shape-memory alloys now require materials featuring phase transformation temperatures above about 100° C., the martensite start point for the common commercially available nickel-titanium alloys barely exceeds about 80° C. In order to meet higher temperature applications, ternary alloys have been investigated, using various additional metallic elements. For example, substitution of noble metals (Au, Pd, Pt) for Ni in NiTi alloys successfully accomplishes higher temperature phase transformations, but the costs introduced are somewhat prohibitive for many commercial applications. Ternary nickel-titanium shape-memory alloys including a zirconium or hafnium component appear to be potentially economical high temperature shape-memory candidates. However, such ternary nickel-titanium shape-memory alloys including zirconium or haffiium exhibit intrinsically larger hysterisis which prolongs phase transformation response times, making these alloys less favorable in some applications.
 Now an improved high temperature, low hysterisis shape-memory alloy thin film has been developed. According to the present invention, there is provided a nickel titanium hafnium copper shape-memory alloy film comprising about 2 to about 10 atomic percent (“at %”) copper.
 It has been discovered that while nickel titanium hafnium alloys can be utilized in applications requiring high temperature shape memory thin film alloys, the addition of copper to the ternary Ni—Ti—Hf system can markedly reduce transition hysteresis. Furthermore, copper addition tends to decrease the temperature differences between the martensite start and finish temperatures and that between austenite start and finish temperature. Accordingly, thin film alloy applications featuring quicker actuation times can be fabricated. Mechanical properties of ternary Ni—Ti—Hf thin film also may be impacted by the addition of Cu. It has been noted that the addition of Cu to Ni—Ti—Hf thin films results in reduction in ductility and transformation temperatures. However, the maximum recoverable strain remains at considerably high level and fully recoverable stress was increased. Consequently, the quaternary Ni—Ti—Hf—Cu thin films exhibit more stable recovery during thermal cycling than the ternary Ni“Ti“Hf thin films at similar applied stress level.
 According to the present invention, an improved nickel titanium hafnium copper shape memory thin film and a method for fabrication of such an alloy has been developed. The shape memory alloy comprises a (TiHf)5-55(NiCu)45-50 quaternary composition with a Cu content from about 2 at % to 10 at % (preferably about 6 at % to about 10 at %), and an Hf content from about 5 at % to about 25 at % (preferably from about 10 at % to about 20 at %). The Ni+Cu combined content is about 45 at. % to about 50 at. %, preferably from about 48 at % to about 50 at %); the Ti+Hf combined content is about 50 at % to about 55 at %, preferably from about 50 at % to about 52 at %. Preferably, the thin film is sputter deposited onto a substrate heated to a temperature ranging from about 350° C. to about 500° C. during deposition to form an in-situ crystalline structure; and, preferably krypton or xenon (particularly preferably Kr) is used as process gas during deposition.
 The above discussed and other features and advantages of the present invention will be appreciated and understood by those skilled in the art from the following detailed description and figures.
 Referring now to the figures:
FIG. 1 is a graph reporting comparative room temperature tensile tests of a Ni49Ti33.6Hf17.6 alloy thin film and a Ni43.1Ti32.1Hf18.3Cu6.5 all thin film.
FIG. 2 shows strain-temperature curves of a Ni43.1Ti32.1Hf18.3Cu6.5 alloy thin film during thermal cycling under different applied stresses.
FIG. 3 depicts strain-temperature curves of a Ni49Ti33.6Hf17.6 alloy thin film during thermal cycling under different applied stresses.
FIG. 4 reports thermal cycling characteristics of a Ni43.1Ti32.1Hf18.3Cu6.5 cycling was conducted by Joule heating and the sample was stressed at 200 MPa during cycling.
FIG. 5 is a graph reporting thermal cycling characteristics of a Ni49Ti33.6Hf17.6 alloy thin film, wherein thermal cycling was conducted by Joule heating and the sample was stressed at 200 MPa during cycling.
 In a sputtering process, as in the present invention, the sputtering deposition generally takes place in a chamber, such as a Perkin Elmer chamber. The particular process parameters for sputtering deposition are dependent on the specific sputtering equipment employed. An initial base vacuum pressure first is established; this pressure ranges from about 2×10−6 torr or lower. Preferably, the base vacuum pressure ranges from about 5×−7 torr or lower.
 During the ion sputtering process, ionization process gas should be maintained at a pressure ranging from about 0.5 to about 5.0 mTorr. Preferably, process gas pressure ranges from about 0.5 mTorr to about 2.0 mTorr. The ionizing process gas may be any typical sputtering process gas such as argon, krypton, or xenon. The preferred process gas is kyton or xenon; krypton is particularly preferred. Power applied during sputtering should range between about 300 watts to about 6 kilowatts for 8″-diameter target (corresponding to power density from about 9.2 kW/m2 to about 185 kW/m2); preferably the power applied ranges from about 500 watts to about 3 kilowatts (corresponding to power density about 15.4 kW/m2 to about 92.5 kW/m2).
 Thin alloy Ni—Ti—Hf—Cu films having a wide range of compositions and thicknesses can be deposited using the present process. Preferred shape memory alloy thin film compositions are nickel titanium hafnium copper alloys featuring compositions in the range: (TiHf)50-55(NiCu)45-50, with a Cu content from about 2 at % to 10 at %, and an Hf content from about 5 at % to about 25 at %.
 Film thicknesses ranging from about 0.5 micron to about 10 microns are preferred; particularly preferred are deposited film thicknesses ranging from about 3 microns to about 5 microns.
 Pursuant to the present process, metal targets having an elemental composition close to that of the desired alloy film to be deposited are utilized, based on the empirical results of composition shift between deposited film and the target. The targets preferably are hot pressed metal powder targets, which can be fabricated using standard hot-pressing procedures.
 The following examples are provided to further describe the invention. The examples are intended to be illustrative and are not to be construed as limiting the scope of the invention.
 8″-diameter targets with different Ni—Ti—Hf—Cu compositions were prepared by hot pressing method. Based on the empirical composition shift between thin film and target, compositions of the targets were selected in order to make slightly Ti+Hf rich films. Hot pressing was accomplished by standard procedure and the final density of the target was measured to determine fraction of porosity. Table A shows the composition of targets that were used for this study, their theoretical and measured density, and the fraction of porosity calculated accordingly. To protect targets from oxygen contamination, they were sealed in Ar until ready to be put into vacuum chamber.
 Ni Ti Hf Cu films with different compositions were sputter deposited on 5″-diameter oxide passivated Si substrates from the hot pressed targets. Deposition parameters included: base pressure 5×10−7 torr before deposition, working gas Kr pressure 1.5 mTorr during deposition, target-to-substrate distance 3.7 inches, deposition power 2.5 kW, substrate temperature about 450° C. during deposition and deposition time 25 minutes. The thickness of thin film was about 4.8 μm. Chemical composition of the film around wafer center was measured by x-ray Energy Dispersive Spectrum (EDS) equipped on a Scanning Electron Microscope (SEM), and the transformation temperatures were measured by Differential Scanning Calorimetry (DSC). The composition and the transition temperatures of film deposited from each target are shown in Table B. The data indicates that, compared with a sample without copper addition, Ni—Ti—Hf—Cu quaternary films have lower transformation temperatures, and increased copper additions lowered the transformation temperatures further. However, copper additions to Ni—Ti—Hf thin film have decreased the transformational hysteresis considerably. At the same time, the temperature intervals of both forward transformation (Ms-Mf) and reverse transformation (Af-As) are also decreased, especially for the sample with 6.5 at. % copper, which records only 11° C. of the temperature difference while sample without copper has 30° C. to 45° C. of the temperature difference. As a result, the temperarture difference between Af and Mf temperatures, a parameter that determines response rate from high-temperature state to low temperature state, was decreased by 40% or more.
 Mechanical and shape-memory properties of the deposit were evaluated by tensile testing at room temperature and thermal cycling under constant load. Thin film was mechanically delaminated from the substrate and cut to 3.0 mm×30.0 mm dimension. Tensile testing was conducted at room temperature using an Instron tensile machine. As shown in FIG. 1, at strain rate of ˜0.08/min, the fracture strength stress was close to above 900 MPa, and the room-temperature failure strain was about 6%. As a comparison, FIG. 1 also includes the room-temperature tensile test result of the sample without copper addition, which has over 8% of ultimate elongation and fracture stress is about 1300 MPa. It is demonstrated that while copper additions appeared to make the sample less ductile, the ultimate elongation still was favorable, considering that the majority of applications require strain ranging from 0.5% to 2%.
 Shape-memory properties were evaluated by thermomechanical tests which are accomplished by thermal cycling the thin film between 50° C. and 250° C. under constant load. FIG. 2 shows the results of Ni43.1Ti32.1Hf18.3Cu6.5 film schematically. Shape-memory strain is produced at minimal applied stress from 150 MPa to 175 MPa, and the recovery of shape-memory strain still takes place at applied stress as much as 800 MPa. Also, perfect recovery of nearly 2.5% strain at applied stress of 300 MPa.
FIG. 3 presents the results of thermomechanical tests for Ni49Ti33.4Hf17.6 film, which was also deposited at elevated substrate temperature so that the as-deposited film is crystalline structure and possesses shape memory effect. Comparison between the results for samples with copper or without copper indicated that, samples with copper need higher stress to effect a shape-memory strain (For example, at 150 MPa applied stress, Ni—Ti—Hf—Cu barely produced shape memory strain on cooling, but Ni—Ti—Hf has about 0.5%). However, the prefect recovery strains for both samples were almost the same. Furthermore, Cu-added SMA film appears to be advantageous on some other aspects. First, the temperature change required to obtain maximum shape-memory strain and that to recover the strain are considerably reduced by addition of copper. This characteristic is related to the fact that Ni—Ti—Hf—Cu thin film has narrow temperature intervals for both forward martensitic transformation and reverse transformation, and it is particularly desired when quick completely releasing or closing action is required and “ajar” condition needs to be avoided. Second, less temperature change is required to complete a strain-temperature loop, or an activation/deactivation cycle for copper-added SMA film. This property is related to the fact that the temperature difference between Af and Mf is reduced by adding copper and it is particularly useful for fast cycling. Third, the stress for complete recovery seems to be increased by addition of copper. This property is related to the improvement of critical stress for slip and it should ensure more stable cyclic characteristics, as shown below.
 Thermal cycling was conducted by Joule heating, i.e., applying pulsed d.c. power to the SMA sample to heat it (at power-on state) and to allow the sample naturally to cool down in ambient air (at power-off state), while constant load is applied to the sample during thermal cycling. FIG. 4 and 5 show the thermal cycling results for Ni43.1Ti32.1Hf8.3Cu6.5 and Ni49Ti33.4Hf17.6 samples at 200 MPa applied stress, respectively. It is indicated that copper addition improves the stability of the thermal cycling, especially the stability of recovery. For examples, after 5000 cycles, about 0. 1% residual strain was developed for copper-added sample, at the same time, about 0.55% residual strain is observed for sample without copper addition.
 In contrast to the sputtering method wherein the substrate is hot enough to produce in-situ crystalline structure, quaternary shape-memory thin films have also been sputter deposited from Ni37.5Ti34.5Hf18Cu10 and Ni39.5Ti33.5Hf18Cu8 targets at lower substrate temperature (100° C. to 350° C.) Thin films contained similar compositions as listed in Table 2. The as-deposited sample did not show shape memory effect, but that is not surprising because crystalline structure, particularly B2 parent phase or monoclinic or orthorhombic martensite, is required for shape-memory effect. Heat treatment was applied in order to crystallize the thin films. The procedure of the heat treatment includes heating the sample to a temperature above crystallization temperature (about 500° C.) and holding the sample at that temperature for a period of time (from several minutes to hours). Although the phase transformation was detected by thermal analysis (such as DSC) or electrical resitivity measurement, the sample was very brittle. At room temperature, it was almost impossible to impart an observable amount of plastic strain without breaking the sample. Therefore, shape-memory effect could not be determined. However, the brittle nature of the annealed sample indicated that the annealing method was not suitable to produce the Ni—Ti—Hf—Cu shape memory thin film with satisfactory mechanical properties.
 In contrast to produce the Ni—Ti—Hf—Cu shape memory thin film by using krypton as working gas, argon as working gas was also investigated. As described in Example 1, the substrate temperature was heated to about 450° C. during deposition. Indeed, crystalline structure was obtained without further heat treatment. However, as-deposited film appeared rather brittle, and it was almost impossible to conduct shape-memory property evaluation with those thin films.
 Various other embodiments or other modifications of the disclosed embodiments will be apparent to those skilled in the art upon reference to this description, or may be made without departing from the spirit and scope of the invention defined in the appended claim