Transformation behavior and inverse caloric effects in magnetic shape memory Ni44-xCuxCo6Mn39Sn11 ribbons

Accepted Manuscript Transformation behavior and inverse caloric effects in magnetic shape memory Ni44xCuxCo6Mn39Sn11 ribbons A. Wójcik, W. Maziarz, M.J. Szczerba, M. Sikora, A. Żywczak, C.O. Aguilar-Ortiz, P. Álvarez-Alonso, E. Villa, H. Flores-Zúñiga, E. Cesari, J. Dutkiewicz, V.A. Chernenko PII:

S0925-8388(17)31974-6

DOI:

10.1016/j.jallcom.2017.05.337

Reference:

JALCOM 42068

To appear in:

Journal of Alloys and Compounds

Received Date: 28 February 2017 Revised Date:

25 May 2017

Accepted Date: 31 May 2017

Please cite this article as: A. Wójcik, W. Maziarz, M.J. Szczerba, M. Sikora, A. Żywczak, C.O. Aguilar-Ortiz, P. Álvarez-Alonso, E. Villa, H. Flores-Zúñiga, E. Cesari, J. Dutkiewicz, V.A. Chernenko, Transformation behavior and inverse caloric effects in magnetic shape memory Ni44-xCuxCo6Mn39Sn11 ribbons, Journal of Alloys and Compounds (2017), doi: 10.1016/ j.jallcom.2017.05.337. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT

Transformation behavior and inverse caloric effects in magnetic shape memory Ni44-xCuxCo6Mn39Sn11 ribbons A. Wójcika*, W. Maziarza, M.J. Szczerbaa, M. Sikorab, A. Żywczakb, C. O. Aguilar-Ortizc, P. Álvarez-Alonsod, E. Villae, H. Flores-Zúñigac, E. Cesarif, J. Dutkiewicza, V.A. Chernenkog,h Institute of Metallurgy and Materials Science, Polish Academy of Sciences, 25 Reymonta Street, 30-059 Kraków, Poland b

RI PT

a

Academic Centre on Materials and Nanotechnology, AGH University of Science and

Instituto Potosino de Investigación Científica y Tecnológica A.C., Camino a la Presan San

M AN U

c

SC

Technology, Al. Mickiewicza 30, 30-059 Kraków, Poland

José 2055, Lomas 4ª sección, San Luis Potosí 78216, Mexico d

Departamento de Física, Universidad de Oviedo, 33007 Oviedo, Spain e

CNR ICMATE, Unità di Lecco, 23900 Lecco, Italy

f

Departament de Física, Universitat de les Illes Balears, Ctra. De Valldemossa, km 7.5,

g

h

TE D

E-07122 Palma de Mallorca, Spain

Ikerbasque, Basque Foundation for Science, 48013 Bilbao, Spain

BCMaterials & University of Basque Country (UPV/EHU), 48080 Bilbao, Spain

EP

*

Corresponding author: [email protected]

AC C

Abstract

An influence of the Cu doping on structural, magnetic and thermoelastic properties of the Heusler Ni44-xCuxCo6Mn39Sn11 (x=1–4 at.%) ribbons has been investigated. It is found that the addition of Cu stabilizes austenite phase. Martensite transformation (MT) temperatures generally decrease when Cu concentration increases, which is attributed to the atom size effect. The inverse magnetocaloric and elastocaloric effects in the vicinity of MT under moderate magnetic fields and stresses have been evaluated. Small Cu addition enhances both

1

ACCEPTED MANUSCRIPT effects as compared to quaternary Ni–Co–Mn–Sn alloy. The magnetic entropy change under low magnetic field of 15 kOe increases from 2.9 J kg-1K-1 for Cu0 to 6.3 J kg-1K-1 for Cu2. Keywords: Ni-Co-Mn-Sn Heusler alloys; melt-spun ribbons; martensitic transformation; magnetocaloric effect; elastocaloric effect

RI PT

1. Introduction

Many studies have been performed on Ni–Co–Mn–Sn metamagnetic shape memory alloys (MMSMAs) (see, e.g., Refs.[1-7]). These materials undergo a first order martensitic

SC

transformation (MT) from ferromagnetic austenite to weak magnetic martensite. Due to a large magnetization difference (∆M) between the high temperature parent austenite phase and

M AN U

low-temperature martensite phase a field induced reverse martensitic transformation (FIRMT) occurs. This phenomenon results in a metamagnetic shape memory effect (MMSME) and an inverse magnetocaloric effect (IMCE). For IMCE, a positive magnetic entropy change (∆SM≥0) and a negative adiabatic temperature change (∆Tad≤0) are observed, which means

conventional MCE) [8].

TE D

that the sample cools down when magnetic field is adiabatically applied (in contrast to the

From practical point of view, a suitable magnetocaloric material should be characterized by

EP

the MT occurring near room temperature and should exhibit appropriate magnetocaloric properties, i.e. high values of the magnetic entropy change, ∆SM, as well as large refrigerant

AC C

capacity (RC) at low magnetic fields [9,10]. Cong et al. have studied the influence of Co addition on magnetic properties and phase transformation temperatures in Ni50-xCoxMn39Sn11 alloys. It was found that Ni44Co6Mn39Sn11 bulk alloy exhibits the largest ∆M across MT. However, the MT temperature is quite high for such alloy equal to about 342 K [11]. On the other hand, it is known that the MT temperature and magneto-structural properties of Ni–Mn based Heusler alloys are very sensitive to chemical composition and multiple doping. We assume that, particularly, addition of Cu into the aforementioned Ni44Co6Mn39Sn11 alloy could

2

ACCEPTED MANUSCRIPT be an effective tool to fine control the temperature range of MT and improve its magnetostructural properties, as it is known that Cu substitution for Ni and Mn in the ternary Ni–Mn– Sn and Ni–Mn–Ga alloys could effectively tailor the magneto-structural transformation by affecting the magnetic coupling between Mn atoms [12]. As an example, the addition of 2

RI PT

at.% of Cu into Ni46Mn43Sn11 decreases both the MT and Curie temperature of austenite (

)

while increases the value of ∆M resulting in an increase of ∆SM [12]. Huu et al. have also reported a decrease of the MT temperature with Cu addition in Ni50-xCuxMn37Sn13 (x=0, 1, 2,

case of Ni50-xCuxMn36Sn14 (x=2, 4, 6 at.%) alloys, the

SC

4, 8 at.%) alloys exhibiting relatively large values of ∆SM (5.4 J kg-1K-1 at 15 kOe) [13]. In the temperature slightly increases,

M AN U

while the MT temperature decreases upon Cu addition [14]. Note, that the decrease of MT temperature as a function of Cu doping is opposite to what it is expected from the increase of valence electron concentration (e/a) [15]. This decrease can be explained by the enlargement of the unit cell volume caused by replacing smaller atoms (Ni) by the larger ones (Cu) [16].

TE D

Furthermore, the addition of Cu into MMSMAs can enhance their ductility like in the case of Ni–Mn–Ga [17]. Castillo-Villa et al. have investigated the inverse magnetocaloric and conventional elastocaloric (eCE) effects in Ni43Mn40Sn10Cu7 [18]. The moderate values of

EP

both two effects have been obtained in the last work due to a reduced magnitude of the

AC C

magnetization/strain change at MT induced by the applied field/stress limited by 5T/5.24MPa. The influence of a fifth element on the magnetocaloric effect and magneto-structural properties of Ni–Co–Mn–Sn has been reported in Refs. [19-21]. The results show that manufacturing of quinary alloys allows a fine tuning of MT temperature and improves their physical properties such as thermal hysteresis [19] and ductility [21]. Another factor affecting the transition temperatures and magneto-structural properties of MMSMAs is the fabrication method. Many reports concern just bulk alloys prepared by the arc- or induction melting methods. However, bulk alloys require a time consuming heat

3

ACCEPTED MANUSCRIPT treatment in order to obtain a good homogeneity and desirable properties. An alternative is the melt-spinning process, which allows obtaining polycrystalline ribbons of good homogeneity leading to shortening of heat treatment time while enhancing MCE [22-24]. Furthermore, the

between caloric material and the cold and hot sinks.

RI PT

ribbon shape is favorable for refrigerant applications, as it improves the heat exchange

This work presents a systematic study of the influence of Cu substitution for Ni on the structural, transition and magnetic characteristics exhibited by the Heusler Ni44(x=1–4 at.%) ribbons. Furthermore, both inverse magnetocaloric as well as

inverse elastocaloric effects have been investigated.

M AN U

2. Experimental

SC

xCuxCo6Mn39Sn11

A series of Ni44-xCuxCo6Mn39Sn11 alloys with x=0, 1, 2, 3, 4 at.%, hereafter referred to as Cu0, Cu1, Cu2, Cu3 and Cu4, respectively, were prepared by induction melting from the high purity (>99.99) elements in argon atmosphere. An extra 2 wt.% of manganese was added to

TE D

compensate losses connected with its evaporation during fabrication process. The as-cast ingots were subsequently annealed at 1220 K for 5h under argon atmosphere in order to ensure a good homogeneity. Then, the ribbons pieces with 30-50µm of thickness, 3-5 mm of

EP

width and different lengths were fabricated by melting and ejecting with 0.25 MPa

AC C

overpressure onto a copper wheel rotating at the linear speed of 30 m s-1 which was optimal in our case to get the adequate quality ribbons. The microstructure and chemical composition of ribbons were examined by using Philips XL30 Scanning Electron Microscope (SEM) and Tecnai G2 Transmission Electron Microscope (TEM) equipped with energy dispersive X-ray microanalyser (EDX) and High Angle Annular Dark Field (HAADF) detector. Thin foil samples for TEM observations were prepared with TenuPol-5 double jet electropolisher using an electrolyte of nitric acid (1/3) and methanol (2/3) at temperature of about 243 K. Room temperature X-ray diffraction (XRD)

4

ACCEPTED MANUSCRIPT patterns were collected by a Bruker D8 diffractometer using a Co radiation. Lattice parameter of paramagnetic austenite was studied by a Bruker D8 at temperature of 430 K. Transformation behavior and thermal effects were investigated by a differential scanning calorimetry (DSC) using Mettler DSC 823 instrument with cooling/heating rate of 10 K min-1.

RI PT

Magnetic properties as well as the characteristic temperatures of martensitic and magnetic transitions were examined by a LakeShore 7407 vibrating sample magnetometer (VSM) equipped with a Janis made LN2 cryostat. Temperatures of both martensitic and magnetic

SC

transitions were estimated from the temperature dependences of physical properties by the two-tangents method. The thermomechanical measurements were carried out with a stress-

M AN U

control option of operation in Q800 TA Instruments Mechanical Analyzer (DMA) using the dynamic mode (temperature rate of 2 K min-1, strain amplitude of 10-4 and frequency of 1Hz) and the static one (temperature dependence of strain ε(T) was measured under load applied along the ribbon axis with temperature rate of 5 K min-1). Isothermal stress-strain (σ-ε) curves

TE D

were registered with a rate of 2.5 MPa min-1. The samples with a gauge length of 7-10 mm were clamped between grips faces through a thin Al foil to better distribute the compressive stresses and prevent samples fracture at the ends. Other possible clamping effects on the

EP

results of measurements were minimized by a relatively small ribbon area subjected to compression. It is worth noting that in DMA machines the thermocouple is not attached to the

AC C

sample whereby producing misfit in absolute values between characteristic temperatures determined by DMA method and other methods, e.g., DSC (see also Sec. 3.3). 3. Results and discussion

3.1 Microstructure and structure characterization The microstructure and chemical composition of Ni44-xCuxCo6Mn39Sn11 (x=0, 1, 2, 3, 4 at.%) melt-spun ribbons have been investigated by SEM and TEM microscopy. Figure 1 demonstrates SEM secondary electron (SEM SE) images of Cu1 ribbon taken at the free side

5

ACCEPTED MANUSCRIPT (the side which was not in contact with the wheel). All ribbons possess a similar microstructure consisting of two types of grains, independent of their chemical composition. The first type is given by small equiaxed grains with the size of about 1–2 µm while the second type is represented by larger cells with various shapes being arranged in clusters. A

RI PT

similar result was already observed in the Ni48Mn39.5Sn12.5 ribbons by Czaja et al. [25]. It was also found that the free surface exhibits coarser morphology in comparison to the wheel side. Furthermore, one can see the difference in the microstructure between the initial alloy with

SC

the composition Ni44Co6Mn39Sn11 (referred to as Cu0), investigated in our previous work (Ref. [26]) and Cu-doped ribbons. It was shown that the microstructure of the quaternary

M AN U

alloy is composed of fine equiaxed grains, and cells being more complicated in shapes and large grains of about 10–15 µm with martensite plates inside. The divergence between ribbons seems to be connected with the addition of the fifth element which affects the characteristic temperatures of MT as well as the crystal structure and microstructure.

TE D

The chemical compositions of Ni44-xCuxCo6Mn39Sn11 (x=0, 1, 2, 3, 4 at.%) melt-spun ribbons were analyzed by the EDX method. The results taken from the free side are listed in Table 1. Moreover, detailed investigation shows that there is no difference in the distribution of

EP

elements between the free and the wheel side surfaces of the ribbons as well as along their cross-section. This was also shown by Santos et al. in the case of Ni50Mn37Sn13 ribbons [27].

AC C

Based on the EDX results the valence electron concentration per atom, e/a, was estimated as shown in the last column of Table 1. This was calculated from the sum of d and s electrons for Mn (7), Ni (10), Co (9), Cu (11), and s and p electrons for Sn (4). Table 1 shows that the substitution of Cu for Ni leads e/a to increase with respect to the initial alloy. Figure 2 presents a set of scanning transmission electron microscopy (STEM-HAADF) images of Cu1 ribbon. Fig. 2a with low magnification shows a cellular microstructure composed of the dark cells of about 2 µm in cross-section and bright intercellular areas with 6

ACCEPTED MANUSCRIPT size from 100 to 500 nm. The martensite plates can be easily recognized in the intercellular areas. It can be deduced that the crystallization of ribbons follows the mechanism of a cellular growth type. Fig. 2b shows high magnification image with the EDX chemical analysis taken from the region of cells and intercellular areas. The difference in the chemical content

RI PT

between these two areas is seen. In the intercellular areas, higher amount of Mn and Co but lower concentration of Sn in comparison with cell interiors are found. This results in different values of e/a, which are equal to 8.15 and 8.20 for the cells and intercellular areas,

SC

respectively. One can see, that a favourable nucleation of martensite at the grain boundaries

transformation temperature on e/a [15].

M AN U

can be explained in accordance with the well-known dependence of martensitic

TEM bright field (BF) images captured from the cells (a) and intercellular areas (c) of Cu1 ribbon at room temperature is presented in Figure 3. The corresponding selected area electron diffraction patterns (SAEDPs) are indexed in terms of the cubic L21 structure with [100] zone

TE D

axis (b) and six-layered 12M martensite with [210] zone axis (d). Five additional spots between two fundamental maxima are easily obseved. This type of martensite is not often observed in MMSMAs, contrary to four-, five- and seven-layered structures. However, it has

EP

been found in the alloys, such as Ni44Co6Mn39Sn11 [26], Ni50Mn40Sn10 [28], and

AC C

Ni44.5Co5.5Mn39.5Sn10.5 after homogenization for long time [29]. The coexistance of austenite and martensite as a consequence of local compositional fluctuation has been reported in the case of Ni46Mn41.5-xFexSn12.5 [30] and Ni50-xMn37.5+xSn12.5 [31]. Cu2 also reveals mixed microstructure consisting of austenite and martensite. Figure 4 shows BF images and corresponding SAEDPs for Cu3. Besides L21 austenite small amount of martensite phase, being indexed as the 4O orthorhombic four-layered structure is also visible. Cu4 ribbon is found to be in a single austenite form. Thus, one can conclude that at room temperature the crystal structure of the investigated ribbons evolves in the following sequence: 12M+L21 7

ACCEPTED MANUSCRIPT (Cu0)

L21+12M (Cu1, Cu2)

L21+4O (Cu3)

L21 (Cu4). It is obvious that the

substitution of Ni by Cu results in the stabilization of austenite. XRD measurements confirm TEM observations. Figure 5 depicts X-ray diffraction patterns of Ni44-xCuxCo6Mn39Sn11 (x=1–4 at.%) ribbons at room temperature taken from the free side. It

RI PT

is found that Cu1 and Cu2 ribbons reveal the mixture of L21 structure with traces of martensite being identified as a 12M six-layered martensite. Ribbons containing 3 and 4 at.% of Cu have pure L21 austenitic structure. No extra peaks indicating the existence of additional

SC

phases are observed. In the case of Cu3, martensite phase is not detected by diffraction patterns. This might be connected with a small amout of martensite which is below detection

M AN U

limit of XRD.

Furthermore, the lattice parameter ac and unit cell volume Vc of L21 paramagnetic austenite for all ribbons were estimated and plotted in Figure 6. Calculations were made based on X-ray measurements carried out at 430 K in order to obtain fully austenite phase. It is seen that the

TE D

lattice parameters as well as the unit cell volume rise up with the increase of Cu concentration in Ni44-xCuxCo6Mn39Sn11 ribbons. It is attributed to the substitution of smaller Ni atoms with atomic radius of 0.125 nm by larger Cu atoms with atomic radius of 0.128 nm [32].

EP

3.2 Transformation behavior and magnetic characterisation Figure 7 depicts DSC cooling and heating curves for Ni44-xCuxCo6Mn39Sn11 (x=0, 1, 2, 3

AC C

at.%) melt spun ribbons. One can see exothermic and endothermic peaks corresponding to the forward and reverse martensitic transformation with start and finish characteristic temperatures (Tms, Tmf, Tas, Taf), respectively. The transformation temperatures TM and TA are determined as (Tms+Tmf)/2 and (Tas+Taf)/2, respectively. The width of thermal hysteresis, being a consequence of the first-order nature of MT, was calculated as a difference between austenite finish and martensite start temperatures (∆T=Taf-Tms). DSC curves enable also an estimation of the averaged MT enthalpy change (|∆H|), calculated as the area under the

8

ACCEPTED MANUSCRIPT exothermic and endothermic peaks. The entropy change of MT (∆S) was estimated as |∆H|/T0, where T0 is (Tms+Taf)/2 [33]. All results are listed in Table 2. Note that the DSC curves for Cu4 are not shown here, because MT takes place below the studied temperature range. The thermomagnetization dependences, M(T), for all ribbons are shown in Figure 8. They

RI PT

were measured under external magnetic field of 200 Oe during cooling (FC) and heating (FH). Upon cooling, samples undergo the magnetic transition from paramagnetic austenite to ferromagnetic austenite. During further cooling, the martensitic transformation from austenite

SC

to martensite with characteristic hysteresis is observed. The values of the characteristic MT temperatures (see Table 2) are in good agreement with data of the calorimetric measurements.

M AN U

Small variations between results can be ascribed to the influence of factors such as different thermal contacts of the samples with thermocouples in the DSC and VSM instruments, twotangent procedure of determination of the characteristic temperatures etc. It is well-known that the MT temperatures in the Heusler type SMAs are mainly controlled

TE D

by: (i) electron valence concentration [15, 34], (ii) unit-cell volume and (iii) microstructure. In the case of studied Ni44-xCuxCo6Mn39Sn11 (x=0–4 at.%) ribbons, an influence of the microstructure on the MT temperatures has not been considered since experimental results

EP

revealed that the Cu doping does not affect the microstructure in the condition that the same technique with constant production parameters was used to obtain ribbons.

AC C

Table 2 shows that the MT temperatures are drastically shifted to lower temperatures with the addition of 1 at.% of Cu. Then, slightly increase for Cu2 ribbon. Further addition of Cu leads again to their decrease. Such a complex behavior was observed in Ni50Mn40-xSn10Fex as a result of the precipitation of additional phases [35]. However, in the studied alloys no obvious signs of precipitates were found. The aforementioned irregular behavior may be attributed to the e/a decrease from 8.15 to 8.13, with regard to Cu0 [26], and further increase up to 8.16. Nevertheless, the general tendency for the MT temperatures in the studied ribbons is their

9

ACCEPTED MANUSCRIPT decrease with increasing of Cu concentration being opposite to the e/a rule. Figure 6 shows that the difference in atomic radius can be responsible for the reduction of the MT temperatures. As already mentioned, the atomic radius of Cu and Ni are 0.128 and 0.125 nm, respectively [32]. Thus, replacing smaller Ni atoms by larger Cu atoms results in the decrease

RI PT

of the MT temperature. Many reports concerning Cu-doped Ni-Mn based Heusler alloys show similar results pointing out that the unit cell volume factor prevails over e/a influence [12-14, 36]. It is worth noting that the concept of hybridization between Ni 3d and Mn 3d states was

SC

also put forward as an additional factor [37]. Accordingly, any change in Mn and Ni content as well as Cu addition may lead to the weakening the hybridization resulting in a reduction of

xCoxSn14

M AN U

the MT temperatures. This behavior has been observed in Mn50Ni40-xCuxSn10 [38], Ni50Mn36[39], Ni50+xMn37-xSb13 and Ni50Mn37+xCrxSb13 [40].

Magnetic and DSC measurements, as well as data from Table 2, show that

also shifts to

lower temperatures with the replacement of Ni by Cu, which correlates with the increase of

TE D

the volume cell of austenite. This is in line with fact that the hydrostatic pressure increases the Curie temperature of austenite [41]. Aguilar-Ortiz et al. confirmed that the substitution of smaller Ni atoms by larger Fe atoms leads to the enlargement of the austenite volume cell to decrease [42].

EP

which causes

AC C

Figure 9 shows the magnetization curves, M(H), at different temperatures around the martensitic transformation. The measurements were performed up to 15 kOe after every time zero-field cooling to temperature below Tmf in order to achieve fully martensite structure. The samples show metamagnetic behavior manifested through the jump of magnetization (∆M) and appearance of the hysteresis loop (one example is indicated by arrows in Fig. 9(c)) in the studied temperature ranges of 248 K – 299 K, 242 K – 302 K, 216 K – 288 K and 165 K – 210 K for Cu1, Cu2, Cu3 and Cu4, respectively. These magnetization jumps reflect the magnetic field induced reverse MT with the formation of some volume fraction of austenite.

10

ACCEPTED MANUSCRIPT The magnetization of austenite (MA) at 15 kOe increases from 52 Am2 kg-1 for Cu1 to 66 Am2 kg-1 for Cu2, 72 Am2 kg-1 for Cu3 and 67 Am2 kg-1 for Cu4 ribbons. At the same time, the magnetization of martensite (MM) equals to: 25 Am2kg-1, 17 Am2kg-1, 21 Am2kg-1 and 51 Am2kg-1 for Cu1, Cu2, Cu3 and Cu4, respectively. Thus, the investigated ribbons tend to

RI PT

increase ∆M with Cu concentration terminated by its sudden drop for Cu4. The enlargement of magnetization difference between austenite and martensite results in higher Zeeman energy

change, as discussed in the next section. 3.3 Inverse magnetocaloric and elastocaloric effects

SC

being a driving force for the FIRMT and also facilitates the higher value of magnetic entropy

using Maxwell’s relation ∆

( , )=

M AN U

Magnetic field induced entropy change (∆SM) has been estimated from magnetic isotherms , where M and H are magnetization and

magnetic field, respectively. Temperature dependences of ∆SM under different magnetic fields are shown in Figure 10. Peak values of ∆SM at 15 kOe for Cu0 [26], Cu1, Cu2, Cu3 alloys are

TE D

2.9 J kg-1K-1 (345 K), 2.1 J kg-1K-1 (275 K), 6.3 J kg-1K-1 (293 K) and 4.3 J kg-1K-1 (258 K), respectively. The result for Cu4 ribbon is not shown here, due to its negligible value. The values of ∆SM are obviously correlated with the magnetization difference between austenite

EP

and martensite but no correlation is observed with the values of transformation entropy, ∆SMT,

AC C

listed in Table 2. One can see that the highest values of ∆SM are obtained for the largest ∆M in Cu2 and Cu3 which at the same time exhibit a much smaller values of ∆SMT than Cu0. These experimental results are in line with both the concept that for the first order magnetostructural transitions the upper bound of ∆SM is limited by the value of ∆SMT, and with the results of ab initio calculations [43] and thermodynamic approach [44] which show that the M(T) anomaly at MT is the parameter quantitatively governing ∆SM at different applied fields. From the practical viewpoint, refrigerant capacity (RC) is an important parameter, which is proportional the total amount of thermal energy that can be transferred between hot and cold 11

ACCEPTED MANUSCRIPT sinks in one thermodynamic cycle [45]. RC has been calculated by using the following equation: RC= ∆

×

is the maximum at ∆SM(T) curve and

, where

is the full-width at half maximum of the same curve obtained under magnetic field change of 15 kOe. Furthermore, due to the first order nature of MT, the average hysteresis

RI PT

loss (HL) was calculated by the estimation of the area of M(H) curves at different temperatures. The values of HL are 3.8 J kg-1, 4.8 J kg-1, 8.4 J kg-1 and 7.9 J kg-1 for Cu0, Cu1, Cu2 and Cu3, respectively. The effective RCeff has been obtained by subtracting HL from the

SC

total RC [45-47]. The quantified values of RCeff in magnetic field of 15 kOe for Cu0, Cu1, Cu2, Cu3 ribbons are 19 J kg-1, 31.4 J kg-1, 41.9 J kg-1, 50.8 J kg-1, respectively. One can see

M AN U

that, the addition of 2 at.% of Cu increases the value of ∆SM for more than two times in the magnetic field of 15 kOe. Moreover Cu addition leads to the enhancement of RCeff (Cu2, Cu3), while tuning the MT temperature close to ambient temperature. Figure 11 (a,b) depicts the temperature dependences of strain, ε(T), recorded under various

TE D

applied tensile stresses for Ni44Co6Mn39Sn11 and Ni42Cu2Co6Mn39Sn11 ribbons. The application of stresses higher than 20 MPa – 30 MPa causes ribbons to fail due to their high brittleness. The measurements were performed only for two ribbons due to limitation related

EP

to the required geometry of ribbons such as length and width. The results obviously show

AC C

anomalous behavior: instead of elongation during forward MT the ribbons exhibit contraction and vice versa during reverse MT. When the two-tangent is applied to the graphs in Fig. 11(a,b), the obtained characteristic temperatures appear different as compared with ones determined by DSC or M(T).This is because the thermocouple in DMA machine is not attached to the sample (see Sec.2). Therefore, the shifts (not absolute values) of characteristic temperatures produced by applied stress are shown in Fig. 11(c). The negative slope of the curves in Fig.11(c) reflects an anomalous behavior already observed and discussed previously for the Ni50Mn40Sn10 ribbon [48]. This phenomenon is associated with the internal

12

ACCEPTED MANUSCRIPT compressive stress introduced in the ribbons during fabrication process. This compressive stress opposes the external tensile stress, whereby instead of the elongation during forward MT the ribbons exhibit contraction (see also Ref.[48]). The enhanced brittleness of the ribbons did not allow to get the conventional stress-strain-temperature relations expected at

RI PT

higher stresses (see Ref.[48]). It is worth noting that ∆T slopes are much larger for Cu0 than for Cu2 ribbons, probably, due to much larger value of transformation entropy exhibited by the former ribbon.

"

! "

, where σ and ε are uniaxial stress and corresponding

M AN U

relationship: ∆ ( , ) =

SC

The stress-induced isothermal entropy change (∆S) was calculated by means of Maxwell’s

strain, respectively. Results for Cu0 and Cu2 ribbons are shown in Figure 12. The maxima in ∆S versus T correspond to about 0.2 J kg-1K-1 under the stress of about 20 MPa for Cu0 and 0.5 J kg-1K-1 under the stress of about 30 MPa for Cu2 ribbon (0.4 J kg-1K-1 when the stress is about 20 MPa). Both values of ∆S are above zero indicating the inverse elastocaloric effect

TE D

(see Ref.[48]). This effect is associated with the anomalous stress-induced reverse MT that should occur in both ribbons and can be considered as an analog of the inverse

EP

magnetocaloric effect in Ni–Mn–Ga Heusler alloys [48]. The inverse elastocaloric effect has been also found in Fe-Rh [49] and in the decomposed Ti-Ni [50] alloys. In the last two cases,

AC C

it was attributed to the stress-induced structural transformation due to the volume contraction across the structural transformation and internal stress caused by precipitates, respectively. The inverse eCE in the studied ribbons, although being quite small, represents interesting topic to be explored in the future work. In this regard, one should bear in mind the level of internal stress in the film/substrate structures can be achieved up to 1 GPa. 4. Conclusions In summary, we have systematically studied the influence of the substitution of Ni by Cu in Ni44-xCuxCo6Mn39Sn11 melt-spun ribbons, with the aim to enhance their caloric properties and 13

ACCEPTED MANUSCRIPT shift MT close to room temperature. It is found that the MT temperature decreases with the Cu addition, which is an opposite behavior to the well-known e/a rule. However, it was experimentally confirmed that the unit cell volume effect is responsible for such a behavior. The substitution of Ni by Cu causes the unit cell volume to expand resulting in the decrease of

RI PT

the characteristic temperatures of MT. The cellular mechanism of grains growth was disclosed being also a reason of the chemical gradient across grain. Additionally, the inverse magnetocaloric and elastocaloric effects in the vicinity of MT under moderate magnetic fields

SC

and stresses have been evaluated. The estimated stress-induced entropy change turned out to be a result of the inverse elastocaloric effect which is a consequence of the internal

M AN U

compressive stress (for more detail, see Ref.[48]). In order to realize and assess a conventional elastocaloric effect in the ribbons, more work is needed to enhance their mechanical properties. Small addition of a fifth element, such as copper, strongly influences the structural behavior and magnetic properties of the studied ribbons, whereby improving

TE D

their magnetocaloric properties important for applications. Acknowledgements

The results have been obtained within the frame of project PBS/A5/36/2013. Partial financial

EP

support of the Spanish Ministry of Economy and Competitiveness (project MAT2014-56116C4-1-3-4-R, AEI/FEDER) and Basque Government Department of Education (project IT711-

AC C

13) and the CONACYT Mexico (CB-2010-01-157541) is acknowledged. References [1]

A.

Ghosh,

K.

Mandal;

J.

Alloys

Compd.

579

(2013)

295.

DOI:

60

(2012)

5335.

DOI:

10.1016/j.jallcom.2013.06.062. [2]

D.Y.

Cong,

S.

Roth,

L.

Schulz;

Acta

Mater.

10.1016/j.actamat.2012.06.034. [3] A. Ghosh, K. Mandal; Eur. Phys. J. B 86 (2013) 378. DOI: 10.1140/epjb/e2013-40489-0.

14

ACCEPTED MANUSCRIPT [4] R. Kainuma, Y. Imano, W. Ito, H. Morito, Y. Sutou, K. Oikawa, A. Fujita, K. Ishida, S. Okamoto, O. Kitakami, T. Kanomata; Appl. Phys. Lett. 88 (2006) 192513. DOI: 10.1063/1.2203211.

033903. DOI: 10.1063/1.2761853.

RI PT

[5] T. Krenke, E. Duman, M. Acet, X. Moya, L. Mañosa, A. Planes; J. Appl. Phys. 102 (2007)

[6] N.M. Bruno, C. Yegin, I. Karaman, J.-H. Chen, J.H. Ross Jr., J. Liu, J. Li; Acta Mater. 74 (2014) 66. DOI: 10.1016/j.actamat.2014.03.020. P.J.

Shamberger,

F.S.

Ohuchi;

Phys.

Rev.

B

79

(2009)

144407.

DOI:

SC

[7]

10.1103/PhysRevB.79.144407.

M AN U

[8] T. Krenke, E. Duman, M. Acet, E.F. Wasserman, X. Moya, L. Ma#$osa, A. Planes; Nat. Mater. 4 (2005) 450. DOI: 10.1038/nmat1395.

[9] A.K. Pathak, I. Dubenko, C. Pueblo, S. Stadler, N. Ali; J. Appl. Phys. 107 (2010) 09A907. DOI: 10.1063/1.3335893.

TE D

[10] M.E. Wood, W.H. Potter; Cryogenics 25 (1985) 667. DOI: 10.1016/00112275(85)90187-0.

[11] D.Y. Cong, S. Roth, M. Pötschke, C. Hürrich, L. Schultz; Appl. Phys. Lett. 97 (2010)

EP

021908. DOI: 10.1063/1.3454239.

[12] R. Das, S. Sarma, A. Perumal, A. Srinivastan; J. Appl. Phys. 109 (2011) 07A901. DOI:

AC C

10.1063/1.3540327.

[13] D.T. Huu, N.H. Yen, P.T. Thanh, N.T. Mai, T.D. Thanh, T.-L. Phan, S.Ch. Yu, N.H. Dan; J. Alloys Compd. 622 (2015) 535. DOI: 10.1016/j.jallcom.2014.10.126. [14] I. Dincer, E. Y%& z%& ak, Y. Elerman; J. Alloys Compd. 506 (2010) 508. DOI: 10.1016/j.jallcom.2010.07.066. [15] T.Krenke, M. Acet, E.F. Wassermann, X. Moya, L. Ma#$osa, A. Planes; Phys. Rev. B 72 (2005) 14412. DOI: 10.1103/PhysRevB.72.014412.

15

ACCEPTED MANUSCRIPT [16] L. Pan-Pan, W. Jing-Min, J. Cheng-Bao; Chin. Phys. B 20 (2011) 028104. DOI: 10.1088/1674-1056/20/2/028104. [17] J. Wang, C. Jiang; Scripta Mater. 62 (2010) 298. DOI: 10.1016/j.scriptamat.2009.11.023. [18] P.O. Castillo-Villa, L. Ma#$osa, D.E. Soto-Parra, J.L. S'́ nchez-Llamazares, H. Flores-

RI PT

Zúñiga, C. Frontera; J. Appl. Phys. 113 (2013) 053506. DOI: 10.1063/1.4790140.

[19] Z. Guo, L. Pan, M. Yasir Rafique, X. Zheng, H. Qiu, Z. Liu; J. Alloys Compd. 577 (2013) 174. DOI: 10.1016/j.jallcom.2013.04.102.

SC

[20] B. Emre, N.M. Bruno, S.Y. Emre, I. Karaman; Appl. Phys. Lett. 105 (2014) 231910. DOI: 10.1063/1.4903494.

10.1007/s12598-013-0100-7.

M AN U

[21] F. Chen, Y.-X. Tong, B. Tian, L. Li, Y.-F. Zheng; Rare Met. 33 (2013) 516. DOI:

[22] Y. Zhang, Q. Zheng, W. Xia, J. Zhang, J. Du, A. Yan; Scripta Mater. 104 (2015) 41. DOI: 10.1016/j.scriptamat.2015.04.004.

TE D

[23] J.L. Sánchez Llamazares, T. Sanchez, J.D. Santos, M.J. Pérez, M.L. Sanchez, B. Hernando Ll. Escoda, J.J. Suñol, R. Varga; Appl. Phys. Lett. 92 (2008) 012513. DOI: 10.1063/1.2827179.

EP

[24] B. Hernando, J.L. Sánchez Llamazares, J.D. Santos, Ll. Escoda, J.J. Suñol, R. Varga, D. Baldomir, D. Serantes, Appl. Phys. Lett. 92 (2008) 042504. DOI: 10.1063/1.2838356.

AC C

[25] P. Czaja, W. Maziarz, J. Przewoźnik, A. Żywczak, P. Ozga, M. Bramowicz, S. Kulesza, J. Dutkiewicz; Intermetallics 55 (2014) 1. DOI: 10.1016/j.intermet.2014.07.001. [26] A. Wójcik, W. Maziarz, M.J. Szczerba, M. Sikora, J. Dutkiewicz, E. Cesari; Mater. Sci. Eng. B 209 (2016) 23. DOI: 10.1016/j.mseb.2016.03.002. [27] J.D. Santos, T. Sanchez, P. Alvarez, M.L. Sanchez, J.L. S'́ nchez Llamazares, B. Hernando, Ll. Escoda, J.J. Suñol, R. Varga; J. Appl. Phys. 103 (2008) 07B326. DOI: 10.1063/1.2832330.

16

ACCEPTED MANUSCRIPT [28] H.F. Tian, J.B. Lu, L. Ma, H.L. Shi, H.X. Yang, G.H. Wu, J.Q. Li; J. Appl. Phys. 112 (2012) 033904. DOI: 10.1063/1.4740458. [29] A.M. Pérez-Sierra, J. Pons, R. Santamarta, P. Vermaut, P. Ochin; Acta Mater. 93 (2015) 164. DOI: 10.1016/j.actamat.2015.04.027.

782 (2014) 23. DOI: 10.4028/www.scientific.net/MSF.782.23.

RI PT

[30] W. Maziarz, P. Czaja, J. Dutkiewicz, R. Wróblewski, M. Leonowicz; Mater. Sci. Forum

[31] W. Maziarz, P. Czaja, M.J. Szczerba, L. Lityńska-Dobrzyńska, T. Czeppe, J. Dutkiewicz,

SC

J. Alloys Compd. 615 (2014) S173. DOI: 10.1016/j.jallcom.2013.12.164.

DOI: 10.1016/j.actamat.2013.04.035.

M AN U

[32] H. Zheng, W. Wang, S. Xue, Q. Zhai, J. Frenzel, Z. Luo; Acta Mater. 61 (2013) 4648.

[33] C.M. Wayman, H.C. Tong; Scr. Metall. 11 (1977) 341. DOI: 10.1016/00369748(77)90263-0.

[34] V.A. Chernenko; Scripta Mater. 40 (1999) 523. DOI: 10.1016/S1359-6462(98)00494-1.

TE D

[35] Z. Wu, Z. Liu, H. Yang, Y. Liu, G. Wu, R.C. Woodward; Intermetallics 19 (2011) 445. DOI: 10.1016/j.intermet.2010.10.010.

[36] W.J. Feng, L. Zuo, Y.B. Li, Y.D. Wang, M. Gao, G.L. Fang; Mater. Sci. Eng. B 176

EP

(2011) 621. DOI: 10.1016/j.mseb.2011.02.003.

[37] K.R. Priolkar, D.N. Lobo, P.A. Bhobe, S. Emura, K. Nigam; EPL Europhys. Lett. 509

AC C

(2011) 38006.

[38] H.C. Xuan, P.D. Han, D.H. Wang, Y.W. Du; Intermetallics 54 (2014) 120. DOI: 10.1016/j.intermet.2014.05.022. [39] L.H. Yang, H. Zhang, F.X. Hu, J.R. Sun, L. Q. Pan, B.G. Shen; J. Alloys Compd. 588 (2014) 46. DOI: 10.1016/j.jallcom.2013.10.196. [40] M. Khan, J. Jung, S.S. Stoyko, A. Mar, A. Quetz, T. Samanta, I. Dubenko, N. Ali, S. Stadler, K.H. Chow; Appl. Phys. Lett. 100 (2012) 172403. DOI: 10.1063/1.4705422.

17

ACCEPTED MANUSCRIPT [41] T. Kanomata, K. Shirakawa, T. Kaneko; J. Magn. Magn. Mater. 65 (1987) 76. [42] C.O. Aguilar-Ortiz, D. Soto-Parra, P. )*lvarez-Alonso, P. L'́ zpita, D. Salazar, P.O. Castillo-Villa, H. Flores-Zú#$iga, V.A. Chernenko; Acta Mater. 107 (2016) 9. DOI: 10.1016/j.actamat.2016.01.041.

RI PT

[43] D. Comtesse, M.E. Gruner, M. Ogura, V.V. Sokolovskiy, V.D. Buchelnikov, A. Gr%& nebohm, R. Arr+́ yave, N. Singh, T. Gottschall, O. Gutfleisch, V.A. Chernenko, F. Albertini,

S.

F'& hler,

P.

Entel;

Phys.

B

89

(2014)

184403.

DOI:

SC

10.1103/PhysRevB.89.184403.

Rev.

013902. DOI: 10.1063/1.4939556.

M AN U

[44] V.A. L'vov, A.Kosogor, J.M. Barandiaran, V.A. Chernenko; J. Appl.Phys. 119 (2016)

[45] K.A. Gschneidner Jr., V.K. Pecharsky, A.O. Pecharsky, C.B. Zimm; Mater. Sci. Forum 315-317 (1999) 69. DOI: 10.4028/www.scientific.net/MSF.315-317.69. [46] V. Provenzano, A.J. Shapiro, R.D. Shull; Nature 429 (2004) 853. DOI:

TE D

10.1038/nature02657.

[47] F. Chen, Y.X. Tong, L. Li, J.L. S'́ nchez Llamazares, C.F. S'́ nchez-Vald,́ s, P. Müllner; J. Alloys Compd. 691 (2017) 269. DOI: 10.1016/j.jallcom.2016.08.240.

EP

[48] P. Álvarez-Alonso, C.O. Aguilar-Ortiz, E. Villa, A. Nespoli, H. Flores-Zúñiga, V.A.

AC C

Chernenko; Scripta Mater. 128 (2017) 36. DOI: 10.1016/j.scriptamat.2016.09.033. [49] S.A. Nikitin, G. Myalikgulyev, M.P. Annaorazov, A.L. Tyurin, R.W. Myndyev, S.A. Akopyan, Phys. Lett. A 171 (1992) 234. DOI: 10.1016/0375-9601(92)90432-L. [50] F. Xiao, T. Fukuda, T. Kakeshita; Scripta Mater. 124 (2016) 133. DOI: 10.1016/j.scriptamat.2016.07.016.

18

ACCEPTED MANUSCRIPT

Figure captions Figure 1. SEM SE micrograph taken from the free side of the Cu1 ribbon.

RI PT

Figure 2. STEM-HAADF images of Cu1 ribbon taken at low (a) and higher (b) magnification with EDX results of chemical composition taken from the area of two cells and the intercellular area. Arrow points to the martensitic plates.

SC

Figure 3. TEM BF image (a) and corresponding SAEDP (b) taken from the cells; TEM BF image (c) and SAEDP (d) taken from intercellular area for Cu1 ribbon.

M AN U

Figure 4. TEM BF images (a), (c) and corresponding SAEDPs (b), (d) for Cu3 ribbon. Figure 5. X-ray diffraction patterns of Ni44-xCuxCo6Mn39Sn11 (x=1, 2, 3, 4 at.%) measured at the free side of the ribbons at room temperature.

Figure 6. Dependence of lattice parameter ac and unit cell volume Vc of paramagnetic

TE D

austenite at temperature of 430 K on the concentration of Cu for all ribbons. Figure 7. DSC curves for Ni44-xCuxCo6Mn39Sn11 (x=0, 1, 2, 3 at.%) ribbons. Figure 8. Thermomagnetization curves measured under magnetic field of 200 Oe for (a) Cu1,

EP

(b) Cu2, (c) Cu3, (d) Cu4 ribbons. The two-tangent method of determination of characteristic temperatures is illustrated in the panel (a).

AC C

Figure 9. Magnetization curves measured at different temperatures in the vicinity of the martensitic transformation in the ribbons: (a) Cu1, (b) Cu2, (c) Cu3 and (d) Cu4. Figure 10. Temperature dependences of magnetic field induced entropy change for Cu1, Cu2 and Cu3 ribbons at different magnetic fields. Figure 11. Temperature dependences of strain under various tensile stresses applied to Cu0 (a) and Cu2 (b) ribbons, showing that instead of elongation during forward MT the ribbons exhibit contraction and vice versa during reverse MT; (c) the shifts of characteristic

19

ACCEPTED MANUSCRIPT temperatures of MT produced by the applied stress. The filled and empty symbols correspond to Cu0 and Cu2, respectively. The linear fits are used to guide the eye. Figure 12. Temperature dependence of isothermal entropy change, due to inverse elastocaloric effect, for Cu0 upon cooling (a) and for Cu2 upon heating (b). The calculations have been

AC C

EP

TE D

M AN U

SC

RI PT

made using results in Fig. 11(a,b).

20

ACCEPTED MANUSCRIPT Ni

Cu

Co

Mn

Sn

e/a

Cu0

44.5±0.9

-

6.1±0.2

38.7±0.7

10.7±0.8

8.15

Cu1

42.4±0.3

1.3±0.1

6.1±0.1

39.6±0.1

10.6±0.1

8.13

Cu2

42.0±0.5

2.4±0.1

6.0±0.3

39.0±0.5

10.7±0.3

8.16

Cu3

41.0±0.4

3.3±0.1

6.2±0.1

38.7±0.3

10.7±0.3

8.16

Cu4

40.0±0.2

4.3±0.1

5.9±0.1

39.0±0.5

RI PT

Ribbon

11.0±0.2

8.17

Table 1. Actual chemical composition (in at.%) and electron concentration per atom (e/a) for

AC C

EP

TE D

M AN U

SC

Ni44-xCuxCo6Mn39Sn11 (x=1–4 at. %) ribbons.

ACCEPTED MANUSCRIPT

Cu0

Cu1

Cu2

Cu3

Cu4

Method

Tms (K)

Tmf (K)

TM (K)

Tas (K)

Taf (K)

TA (K)

∆T (K)

(K)

|∆H| (J g-1)

|∆S| (J kg-1K-1)

DSC

329

301

315

332

345

339

16

-

9.3

27.7

VSM

332

308

320

331

351

341

19

399

-

-

DSC

251

209

230

263

284

274

33

-

2.2

8.2

VSM

251

205

230

259

286

273

35

393

-

-

DSC

260

209

235

281

292

287

32

-

3.8

13.8

VSM

267

218

243

287

293

290

26

393

-

-

DSC

230

192

211

236

267

252

37

-

3.3

13.3

VSM

230

187

209

232

265

249

35

384

-

-

DSC

-

-

-

187

208

198

-

-

-

-

VSM

166

145

156

187

211

199

45

376

-

-

RI PT

Ribbon

SC

Table 2. Forward (TM, Tms, Tmf) and reverse (TA, Tas, Taf) martensitic transformation

M AN U

temperatures, determined by DSC and magnetic measurements, thermal hysteresis (∆T), Curie temperature of austenite (

), the averaged value of enthalpy for the forward and reverse MT

(|∆H|) and the averaged entropy change at the transition (|∆S|) of Ni44-xCuxCo6Mn39Sn11 (x=0,

AC C

EP

TE D

1, 2, 3, 4 at.%) ribbons.

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT •

Cu substitution for Ni stabilizes austenite

•

TM and

•

Addition of 2 at.% of Cu enhances both magneto- and elastocaloric effect

•

Inverse elastocaloric effect is related to stress-induced structural transformation

AC C

EP

TE D

M AN U

SC

RI PT

decrease with the replacement of Ni by Cu