Structural Characterization of Cu- 13.58%Al- 3.94%Ni (wt. %) shape memory alloy elaborated by fusion

This work concerns the structural characterization of Cu -13.58% Al3.94% Ni (wt. %) shape memory alloy elaborated by fusion using various techniques such as microscopy (OM and SEM), dilatometry and diffractometry (XRD). A particular attention is lent to the thermo elastic transformation may occur depending on the content of alloying and heat treatments' adopted. In this context, measurements of transformation temperatures were carried out. The microscopy is used for the characterization of alloy after different heat treatments adopted taking into account also the chemical composition. X-rays diffraction is used in order to access information on the different phases obtained and identify some types of martensite formed after quenching of alloy. The transformation temperatures are measured by the dilatometric analysis technique. This allows studying the structural evolution that occurs during heating and cooling of the alloy studied.


Introduction
Shape memory alloys, most often rated SMA, exhibit thermomechanical remarkable properties as the superelastic effect, the shape memory effect and two-way shape memory effect; these last two are probably the best known makes this class of materials very suitable for innovative applications in various fields. [1,2] These remarkable properties are controlled by a reversible displacive structural transformation, martensitic transformation, between solid phases: the highly symmetric austenite (A) and the less ordered martensite (M). The displacive character is dominated by shearing system of austenite and resulting transformation is controlled by the strain energy. Thus, for a given chemical composition of alloy, the characteristics of this transformation depend on the microstructure and consequently the conditions of elaboration and the thermomechanical history of material. [1,2] Our study focuses on the characterization of Cu-Al-Ni ternary SMA elaborated by fusion using various techniques such as microscopy, dilatometry and diffractometry.

Elaboration of alloy
In preparing our alloy we must consult the equilibrium diagram of binary Cu-Al ( Figure 2). The ȕ phase stable at high temperatures transforms at 570 ° C to Į ductile phase and J 2 very fragile one. Rapid cooling from the stable zone of ȕ phase can keep it in the metastable state, it is transformed into martensite at relatively high temperatures, the alloy having ȕ phase the most stable (eutectoid composition of 24 atomic % or 11.8% by weight of aluminum) is transformed to martensite at 400 ° C after it is ordered DO3 to 550 ° C ( Figure.2). For a temperature Ms near ambient temperature, requires a concentration about 14% by weight of aluminum for this value, it is virtually impossible to obtain ȕ phase free of J 2 phase (very fragile) and the alloy will be unusable. [1,3] The addition of nickel moves the stability field of ȕ to higher concentrations of aluminum and changes very little Ms Temperature curve. (Figure 3) For example, the eutectoid composition goes to 13.2% by weight of aluminum to a concentration of 4% of nickel; the same M s Temperature is about 150 °C. When the concentration exceeds 5% nickel, new precipitated Ni-Al appear: for common alloys are therefore limited to 3-4% nickel. The alloys used for applications are for 4% of nickel between 13 and 14% aluminum and correspond respectively to Ms Temperatures of 200 ° C and 50 ° C approx. [3]

Choice of composition
In theoretical approach we have seen that the martensitic transformation is generally characterized by a transformation temperature Ms, related to the composition of alloy. Experimentally, metallurgists have attempted to establish optical empirical relationships to predict the levels of the respective components of the alloy through the desired start point M s and processing reports of levels of alloying elements to the content of basic element, in our case copper is the basic element. Our 29001-p.2 choice was made to obtain a copper ternary shape memory alloy (Cu-Al-Ni) with a starting temperature of transformation M s not to exceed 200 °C. For Cu-Al-Ni, as reported levels of alloying elements to the content of the base element (Cu) and the starting point of transformation Ms, the calculation gives the composition (in wt%) next: Cu -14 Al -Ni 4 (Ms § 190 ° C) [4,5] The metal used for making alloys are: The pure copper taken from plumbing pipe, pure aluminum cut from a sheet into small pieces and nickel powder very fine. The purity of these elements is acceptable to develop an alloy that can have a good memory effect.

Elaboration by fusion
The technique of preparation adopted is illustrated in Figure 4. The specimens were made following dimensions: length 130 mm, 14 mm in diameter. The samples studied are pellets (14 mm diameter, 3 mm thick).

Analysis of composition
The elaborated samples were subjected to an analysis of composition (Cu, Al, Ni) operated by a microanalyzer RONTEC mounted on the SEM. Example of analysis in Figure 5.
The analytical results seem different from one sample to another because the conditions of preparation (oxidation of Al and much more that of Cu). However we must also point out that these levels vary from one point to another, this always comes back to time keeping in the oven during the merger which determines the homogeneity of the diffusion phenomenon. For this, we have chosen samples which have compositions closer to those introduced: Al Cu -13.58% -3.94% Ni

Study of the initial-state
Before any observation, the samples were coated, polished mechanically and chemically attacked by a reagent composed of ferric chloride (FeCl3), hydrochloric acid (HCl) and ethanol. The observations made in the raw state (as successor to the merger) for various magnifications are shown in Figure 6. The sample structure consists of J 2 and Į phases. J 2 phase consists of irregular geometry and proportion increasingly important.

Study of the quenched-state
To detect the temperature of betatisation (having stabilizing the ȕ phase at ambient temperature) and check the keeping time gives the largest amount of martensite, we realized the following treatments: -We took five samples, numbered from 1 to 5, and then made a series of quenching in water at 25 °C of 850, 880, 910, 940, 970 ° C with retention of 30 mn.
-For the five samples, we redid the hardening at 940 ° C with maintain that vary from 10 mn to 50 mn with a step of 10 mn.
The test results give the temperature of 940 °C which gives the greatest amount of metastable ȕ phase at ambient temperature and 30mn which gives the greatest amount of martensite. The structure of the quenched-state sample shows the existence of martensite in the needles form.

Diffractometry (XRD)
All spectra were performed on a diffractometer XRD 3003 TT mount. The monochromatic Xradiation is produced by a copper anti-cathode rays which KĮ1 (1.5406 A °) and KĮ2 were isolated by discontinuous absorption through a nickel filter. (V = 40kV and I = 30 mA). The XRD was conducted on pellets coated. The diagrams were recorded for 2ș values between 30 ° -60 ° to observe the maximum of the main peaks. The scan speed used was set at 0.05 ° / s. [6] -Structure in initial state: The diffraction analysis shows the existence of two phases (Į and Ȗ 2 ) and for our alloy, peaks related to the Ȗ 2 phase are more intense than the peaks for Į phase. [2] -Structure in quenching state: The diffraction analysis shows that the martensite formed after quenching samples is ȕ' 1 type. [2] Fig. 8. XRD spectra. Initial-state Fig. 9. XRD spectra. Quenched-state

Dilatometry
To study the structural evolution and determine the transformation points during heating and cooling of alloy, we used the dilatometric analysis technique. Indeed, a dilatometer Dil 801 which operates in a temperature range of 20 to 1500 °C and controlled speed between 1 and 30 ° C/mn is used with cylindrical specimens (50mm long and 5mm in diameter). The linear deformation is detected by a sensor with a silica rod. The temperature is measured using a thermocouple placed near the sample. The measurements are recorded by computer.

29001-p.4
It was established variation curves of elongation versus temperature T. This quantity is very important to follow because any change of phase so of structure results in a change of length įL. A change of structure is associated with a change in the apparent coefficient of linear dilatation is reflected on the dilatometric curves by a change of slope. Thus the transformation points are determined on the locations or the curve changes slope and hence the interest to have the curve derived in order to clarify these points. [2,7]

Evolution of quenched structure
The sample was treated for 30 mn at 940 °C and then quenched in water at 25 ° C.
To study the structural evolution that occurs during heating and cooling of our sample, we used a heating speed Vc = 5 °C/mn until a temperature of ȕ phase field (800 ° C) monitoring of cooling at the same speed Vr = 5 ° C/mn until the ambient temperature t. The dilatometric curves obtained are given by figures.10; To determine the transformation points, we used the tangent method to determine these points by intersection of tangents at inflection points of the peak with the baseline extension. (Figure 13)

Influence of heat treatments
To study the effect of heat treatments on these evolutions, we conducted the following treatments: The dilatometric curves (with a speed Vc = Vr = 5 ° C / min) have the same manners in all cases described in the table (Table.1) with the presence of the same changes described in the preceding paragraph. The only difference observed concerns the differential length of the inverse transformation represented by variations in peaks height. (Figure 11)

Influence of heating temperature and time keeping
The transformation points of each evolution are determined by the tangents method and the results obtained are given in the following table: -State 1 and 5: Note that the transformation points, the speed of dilatation and coefficient alpha are different. The dilatation in state 1 is 0.0007 obtained with a speed 14.75 ȝm/mn, by cons in state 5, the dilatation is the same but it is obtained with a speed of 7ȝm/mn. Therefore, the transformation of austenite to martensite is also quick in state 1 than state 5 -State 4 and 6: The dilatation in state 4 is 0.0004 obtained with a speed of 10.80 ȝm / mn, by cons in the state 6, it is 0.0003 obtained with a speed of 4.9 ȝm / mn. Therefore, the transformation of austenite to martensite is also quick in state 4 than state 6. The same remarks in state 1, 4 and state 5, 6, Show that it is an influence of heating temperature and keeping time on this transformation.

Influence of cooling medium
The transformation points of each evolution are determined by tangents method and the results obtained are given in the following table 3: From these results, we note that the transformation points in state 3 are moving to higher values compared to state 1 and 2. The dilatation obtained with different speeds varies from one state to another. So, the cooling medium influences the formation of martensite, which transforms to austenite with different transformation points.

Determination of transformation points (A s , A f , M s and M f )
To determine the points of beginning and end of the martensitic transformation, we performed thermal cycles by dilatometry at a speed of 5 ° C / min to highlight the different transitions expressed between ambient and 700 ° C. This sample was treated to the threshold of each evolution 1, 2, 3 and 4; either, 200, 450, 550 and 700 ° C, then cooled at the same speed. (Figure 12, Table 4).   Figure 13; the transformation at heating is ȕ' 1 AEȕ 1 and manifested by a peak which shows a maximum dilatation speed. When applying heating to 550 ° C, we observe that As and Af of the inverse transformation rise, and the presence of another evolution that begins at 470 °C and finishes around 500 ° C . This transformation led to the stripping of precipitates formed previously and decomposition of the ȕ 1 phase in equilibrium constituents (Į, Ȗ 2 ). -Cooling from 550 ° C and heating to 700 ° C: One comment on the curve (c ') of Figure 13 nay that any transformation during cooling from 550 ° C, those leading to the disappearance of the ȕ 1 phase.
When applying heating to 700 ° C, we observe that there is an evolution starts at 612 ° C and finishes around 640 °C. This transformation corresponds to the recombination of equilibrium constituents Į and Ȗ 2 to form the s ȕ phase stable.
-Cooling from 700 ° C: we observe a transformation that begins at 565 °C and finishes around 482 °C, corresponding to the eutectoid transformation of ȕ phase to Į and Ȗ 2 phases.

Conclusion
For elaboration and characterization of alloy, it was based on two axes: An axis for the elaboration of our alloy by melting, overseeing everything that can hinder this elaboration (heating conditions, cooling conditions and risk watching this elaboration as oxidation). In addition, there was broad interest in heat treatments to highlight the main characteristic of these alloys (martensitic transformation). However, this mode of elaboration was not without consequences on the morphology of grains. Indeed, their growing uncontrolled led us to consider some form of refining by heat treatment method (annealing regeneration).
A second axis for the structural characterization of alloy using several techniques, Microscopy allowed to observe the structure of this alloy and to access information on the different phases observed (Į and Ȗ 2 ) we used the X-ray diffraction and based on work in references [2,6]. This technique can also determine a variant of martensite (ȕ'1.) formed after quenching. Finally, the dilatometric analysis technique permitted to determine the points of martensitic transformation and to study the various phase transformations that occur during heating and cooling of this alloy under conditions chosen.