1 Introduction

In recent decades, shape memory alloys (SMAs) have made a significant contribution to the high technological application, because they have shown a remarkable noise reduction, high strain recovery, and comparably high vibration damping (Graesser and Cozzarelli 1991; Mabe 2008; Qader et al. 2019a, b). These alloys are classified into some families, such as NiTi-based and Cu-based SMAs, where they get more attention from researchers and engineers. Since their crystal structure can be flipped between two solid phases (austenite and martensite), the energy of deformation can be stored in the martensite and it can be released through austenite phase transformation.

Although NiTi-based alloys are more prominent in the industrial fields, they have a comparably high price and mostly can be used for low-temperature applications. On the other hand, Cu-based SMAs are considered as an alternative for some other applications. Cu–Al–Ni-based SMAs have demonstrated phase transformation between − 200 and 200 °C (Čolić et al. 2010), while the production technique and alloying with new elements can enhance the range of phase transformation and other some other physical characteristics of the alloy (Dagdelen et al. 2019a, b; Qader et al. 2019a, b; Vajpai et al. 2013).

Saud et al. (2016) investigated the effect of different aging conditions on microstructural and mechanical properties of quaternary CuAlNiTi SMAs. They found that the increase in both temperature and aging time could shift phase transformation temperatures to higher temperatures. The microstructural phase was affected through the aging process; thus, the tensile strength and other mechanical characteristics were influenced by keeping the alloys in different temperatures and various times. Suresh and Ramamurty (2008) found that the fraction of precipitate increased with increasing aging temperature, and thus, the damping of the alloy was decreased. There is not any investigation about the effect of aging on the microstructural and thermal characteristics of Cu–Al–Ni–Nb in the literature.

In this study, a quaternary Cu–Al–Ni–Nb shape memory alloy has been obtained using the arc-melting technique. The alloy has been investigated for a controlled isothermally aging process. The crystal structural analysis has been carried out using XRD measurement, and surface morphology has been scanned with a metallurgical microscope (optical microscope). In addition, the thermal analysis of the CuAlNiNb SMA has been studied using differential scanning calorimetry (DSC). Some other related calculations have been done to broaden the scope of the study.

2 Materials and Experimental Procedures

To produce the quaternary Cu–Al–Ni–Nb shape memory alloy, a particular amount of metallic powders with high purity were weighed with highly sensitive balance (SCALTEC Analytical balance SBC). The alloy consisted of copper, aluminum, nickel, and niobium with 79, 12, 4, and 5 weight percent (wt%), respectively. The powders were stirred well in a closed compartment to obtain a desired alloy with comparably high homogeneity. Pelletization of powders was carried out under 5 MPa compression, and the pellet was melted in an arc melter in an argon atmosphere. After the obtained ingot was cooled to room temperature in the natural atmosphere, it was cut into eight different samples to apply the aging process. In Table 1, all specimens with the experimental conditions are summarized. The alloy was aged at 1073 K for 1, 2, 3, 6, 12, and 24 h followed by quenching in the ice-brined medium.

Table 1 The nominal codes and a summary of the applying procedure on the Cu–Al–Ni–Nb shape memory alloy
Full size table

The phase transformation characteristics for as-casted and aged alloys were accomplished with differential scanning calorimetry (Perkin Elmer Sapphire). The DSC was run with 10 K/min and nitrogen gas flowed throughout the running process. Additionally, the effect of aging on the crystal structure of the Cu–Al–Ni–Nb alloy was studied through an X-ray diffraction device (Rigaku D/Max-B Geigerflex). In addition, the microstructure and the precipitation of the different constituents were investigated using optical microscopic (PRIOR Model N334 Incident Light Metallurgical Trinocular Microscope) and mapping (EVO 40XVP) technique.

3 Results and Discussion

The distribution of the alloy’s constituents is shown in Fig. 1. The dominant Cu elements have distributed all over the alloys except in some particular positions, which seems to be filled by Nb elements. Ni elements homogenously have distributed, which means nickel could be presented as a solid solution, while Al, on the other hand, could be dissolved in the matrix and presented in the same microstructures that Nb elements have created.

Fig. 1
figure 1

The obtained mapping for concentration of Cu, Al, Ni, and Nb constituents in the as-cast alloy

Full size image

The DSC measurements (Fig. 2) determined the phase transformation temperatures (TTs), which are listed in Table 2. It can be seen that the investigated alloys have TTs around 500 K which can be considered as a high-temperature shape memory alloy (HTSMA). The aging process shifted the TTs to the higher temperatures (Fig. 3), and after 24 h, the alloy evolves from a martensitic transformation to a non-martensitic transformation alloy, and the alloy lost its shape memory properties. These results show that the thermal aging method significantly affects the transformation temperatures of CuAl-based SMAs.

Fig. 2
figure 2

The DSC curves of the aged Cu–Al–Ni–Nb SMA

Full size image
Table 2 Phase transformation temperatures for all the aged Cu–Al–Ni–Nb shape memory alloys
Full size table
Fig. 3
figure 3

The obtained phase transformation temperatures for different aging times of the Cu–Al–Ni–Nb SMA

Full size image

In another study, Balo and Sel (2012) also examined the effect of aging time (0–7 h) on the CuAlNi alloy and found that the TTs fluctuated. However, Benke and his friends (2007) found that the TTs of CuAlNi SMA were increased by increasing aging time. On the other hand, Zhou et al. (2017) reported that the coherence precipitates were the main reason for the emergence of the phase transformation in FeNiCoAlTaB.

The amount of latent heat (enthalpy change) needed for the phase transformation process is obtained from the area under the endothermic troughs and exothermic peaks of the DSC curves. The value of enthalpy change (\( \Delta H \)) was obtained through the DSC software program. Mathematically, the enthalpy change (\( \Delta H \)) of the phase transformation is represented by (Elrasasi et al. 2013; Kök et al. 2020; Recarte et al. 2004):

$$ \Delta S^{{{\text{M}} \to {\text{A}}}} = \mathop \int \limits_{{A_{s} }}^{{A_{f} }} \frac{{{\text{d}}q}}{{{\text{d}}T}}\left( {\frac{{{\text{d}}T}}{{{\text{d}}t}}} \right)^{ - 1} {\text{d}}T $$

or

$$ \Delta S^{{{\text{M}} \to {\text{A}}}} = \mathop \int \limits_{{A_{s} }}^{{A_{f} }} \frac{{{\text{d}}Q^{{{\text{M}} \to {\text{A}}}} }}{{T_{o} }} = \frac{{\Delta H^{{{\text{M}} \to {\text{A}}}} }}{{T_{o} }} $$
(1)

where \( \Delta S \) is the entropy change of martensitic phase transformation and \( T_{o} \) is the temperature where Gibbs free energy for both austenite and martensite phase transformation is equal to zero (\( T_{o} = \left( {M_{s} + A_{f} } \right)/2 \)) (Dagdelen et al. 2019a, b; Kok et al. 2019). The enthalpy and entropy change values for all cases are listed in Table 3. Figure 4 presents the \( \Delta H \) and \( \Delta S \) for forward and reverse martensitic phase transformations, in which generally they decreased with increasing heat treatment. The \( \Delta H^{{{\text{A}} \to {\text{M}}}} \) and \( \Delta S^{{{\text{A}} \to {\text{M}}}} \) recorded the minimum value for 3-h aging at 1073 K. In different circumstances, Ii et al. (2009) investigated aging time and temperature for CuAl alloyed by different compositions of Mn. They have found that the time of aging increased TTs and enthalpy change of the alloys.

Table 3 Enthalpy and entropy change obtained for aged Cu–Al–Ni–Nb alloys
Full size table
Fig. 4
figure 4

The enthalpy change and entropy changes of the aged Cu–Al–Ni–Nb SMA

Full size image

The energy associated with the martensitic phase transformation is known as Gibbs free energy of transformation (\( \Delta G \)). The pushing force at equilibrium temperature can be formulated as (Dagdelen et al. 2019a, b):

$$ \Delta G^{{{\text{M}} \to {\text{A}}}} \left( {T_{o} } \right) = G^{\text{A}} \left( {T_{o} } \right) - G^{\text{M}} \left( {T_{o} } \right) = \left( {H^{\text{A}} - T_{o} S^{\text{A}} } \right) - \left( {H^{\text{M}} - T_{o} S^{\text{M}} } \right) = \Delta H^{{{\text{M}} \to {\text{A}}}} - \left( {T_{o} \Delta S^{{{\text{M}} \to {\text{A}}}} } \right) = 0 $$
(2)

where superscript A and M represent austenite and martensite, respectively. For the martensite phase transformation, Eq. (2) can be rewritten as (Dagdelen et al. 2020; Tatar et al. 2020):

$$ \Delta G^{{{\text{A}} \to {\text{M}}}} \left( {M_{s} } \right) = \Delta G^{{{\text{M}} \to {\text{A}}}} \left( {T_{o} } \right) - \Delta G^{{{\text{M}} \to {\text{A}}}} \left( {M_{s} } \right) $$
(3)

or

$$ \Delta G^{{{\text{A}} \to {\text{M}}}} \left( {M_{s} } \right) = - \left( {T_{o} - M_{s} } \right)\Delta S^{{{\text{M}} \to {\text{A}}}}. $$
(4)

Additionally, the elastic energy (\( G_{e} \)) can be calculated from subtracting Gibbs free energy for forward and reverse phase transformation (Acar et al. 2020; Kok et al. 2019):

$$ G_{e} = \Delta G^{{{\text{A}} \to {\text{M}}}} \left( {M_{s} } \right) - \Delta G^{{{\text{A}} \to {\text{M}}}} \left( {M_{f} } \right) = \left( {M_{s} - M_{f} } \right)\Delta S^{{{\text{M}} \to {\text{A}}}}. $$
(5)

The calculated results for Gibbs free energy, elastic energy, equilibrium temperature, and temperature hysteresis (\( M_{f} - M_{s} \)) for as-casted and aged alloys are listed in Table 4. There were no calculations for the CuAlNiNb alloy aged for 24 h. Figure 5a shows that Gibbs free energy and temperature hysteresis of the alloy were closely related to each other. These parameters reached maximum values for 2–3-h aging at 1073 K and then by increasing the time of aging their values diminished. Elastic energy of the alloy attained the highest value for 1-h aging at 1073 K. The value of Ge decreased and stayed nearly constant for 6 h of aging, and then, it decreased again for increasing aging time at that particular temperature.

Table 4 Temperature hysteresis (Ht), equilibrium temperatures (To), Gibbs free energy (ΔG) for martensite phase transformation (∆GA→M), and elastic energy (GE) for the Cu–Al–Ni–Nb shape memory alloy
Full size table
Fig. 5
figure 5

The calculated a Gibbs free energy and temperature hysteresis, and b elastic energy of the as-casted and aged Cu–Al–Ni–Nb SMA

Full size image

Figure 6 shows the optical microscope (OM) images. There are flower-like precipitated phases spread all over the alloys. It was found that the flower-like phases are niobium-rich phases with no copper content and the Al and Ni were dissolved in the alloys. The matrixes consist of martensite thin discrete middle phase (\( \beta^{\prime}_{1} \)). \( \beta^{\prime}_{1} \) has an 18R-type ordered structure (Araujo et al. 2017).

Fig. 6
figure 6

The optical microscopy images of the aged Cu–Al–Ni–Nb SMA

Full size image

The crystal structural and various phases included in the as-casted and heat-treated Cu–Al–Ni–Nb SMAs are given in Fig. 7. The peaks were indexed with Refs. Ercan et al. (2020), Moghaddam et al. (2017), and Qader et al. (2019a, b). They showed that the matrix of the alloys is martensite (\( \beta^{\prime}_{1} \)) phase with high-temperature cubic austenite phases, including \( \alpha \) and \( \beta_{1} \). \( \beta^{\prime}_{1} \), \( \gamma^{\prime} \), and \( \beta_{1} \) phases that have order types of DO3 (BiF3), 18R (Cu3Al), and 2H (Cu3Ti) phases, respectively (Braga et al. 2017; Otsuka et al. 1979). Additionally, there is also \( \gamma^{\prime} \) phase, which appeared for all treated times and after 24 h it disappeared. The martensitic phase transformation can be defined as \( \alpha \) and/or \( \beta_{1} \leftrightarrow \beta^{\prime}_{1} \) and \( \gamma^{\prime} \). It can be seen that the as-casted alloy shows a single peak (\( \beta^{\prime}_{1} \)), while the aged samples illustrate multipeaks with different phases. Besides, the intensity of the peaks was increased by increasing the time of aging up to 3 h, and it again decreased for more aging times. Although the 24-h aged alloy did not show shape memory characteristic, it still has a crystalline phase.

Fig. 7
figure 7

The XRD results of as-casted and the heat-treated Cu–Al–Ni–Nb SMAs

Full size image

The crystallite size of the alloys was calculated through the Scherrer equation (Buytoz et al. 2019; Shiva et al. 2015):

$$ D = \frac{0.9 \lambda }{B\cos \theta } $$
(6)

where \( \lambda \) is the X-ray wavelength that has been used for the XRD measurement (for this study it was 1.5406 Å), \( B \) is full width at half maximum (FWHM), and \( \theta \) is the Braggs angle. The results of the calculations showed that the crystallite sizes dramatically decreased by applying heat treatment on the alloy (Fig. 8).

Fig. 8
figure 8

The crystallite size of the SMAs

Full size image

4 Conclusions

A quaternary Cu–Al–Ni–Nb shape memory alloy was produced using the arc-melting technique. The alloy was aged for seven different aging times. The as-casted and other aged alloys were characterized using DSC, OM, and XRD. Additionally, some related calculations were performed. The important outcome of this study is as follows:

  • The matrix phases showed martensite-\( \beta^{\prime}_{1} \) with some flower-like precipitations distributed all over the alloys.

  • The map compositions of the alloy showed Nb elements were not dissolved in the matrix, but they precipitated some secondary phases.

  • The phase transformation temperatures increased with increasing the time of aging.

  • The enthalpy and entropy changes of the alloy recorded minimum value for 3-h aging, which could have the most stable phase transformation compared to the other cases.

  • The alloy with 24-h aging at 1073 K showed no phase transformation.

  • The crystallite size generally decreased with performing the aging process for 1 h.