Functional Materials Letters

Vol. 10, No. 1 (2017) 1740001 (8 pages)

© World Scientific Publishing Company
DOI: 10.1142/S179360471740001X

Optimizing Ni–Ti-based shape memory alloys for ferroic cooling

§
A. Wieczorek*, J. Frenzel*, , M. Schmidt†,‡, B. Maaß*, S. Seelecke†, A. Schütze‡ and G. Eggeler*
* Institute for Materials (IFM), Ruhr-Universit€
at Bochum
Universit€atstraße 150, 44801 Bochum, Germany
†Intelligent
Material Systems Laboratory
Department of Materials Science & Engineering and
Department of Systems Engineering, Saarland University
P. O. Box 151150, 66041 Saarbr€ ucken, Germany
‡Lab
for Measurement Technology
Department of Systems Engineering
Saarland University,
P. O. Box 151150, 66041 Saarbr€ ucken, Germany
§
Funct. Mater. Lett. Downloaded from www.worldscientific.com
by FUDAN UNIVERSITY on 02/28/17. For personal use only.

[email protected]

Received 9 December 2016; Accepted 15 December 2016; Published 2 February 2017

Due to their large latent heats, pseudoelastic Ni–Ti-based shape memory alloys (SMAs) are attractive candidate materials for ferroic
cooling, where elementary solid-state processes like martensitic transformations yield the required heat effects. The present work
aims for a chemical and microstructural optimization of Ni–Ti for ferroic cooling. A large number of Ni–Ti-based alloy com-
positions were evaluated in terms of phase transformation temperatures, latent heats, mechanical hysteresis widths and functional
stability. The aim was to identify material states with superior properties for ferroic cooling. Different material states were prepared
by arc melting, various heat treatments and thermo-mechanical processing. The cooling performance of selected materials was
assessed by differential scanning calorimetry, uniaxial tensile loading/unloading, and by using a specially designed ferroic cooling
demonstrator setup. A Ni45 Ti47:25 Cu5V2:75 SMA was identified as a potential candidate material for ferroic cooling. This material
combines extremely stable pseudoelasticity at room temperature and a very low hysteresis width. The ferroic cooling efficiency of
this material is four times higher than in the case of binary Ni–Ti.

Keywords: Shape memory; ferroic cooling; martensitic transformation; hysteresis width; functional fatigue; refrigeration.

1. Introduction Both rely on a reversible martensitic transformation.15,16 The

low temperature phase martensite forms during cooling be-
In the last decades, increasing efforts were made to reduce
tween the martensite start and finish temperatures, Ms and
energy consumption in various technological fields. Refriger- Mf , or during mechanical loading when reaching a critical
ation and cooling cause
16% of the world's energy need.1
stress
AM . The reverse transformation can be observed
Therefore, the development of alternative refrigeration concepts
during heating between the austenite start and finish tem-
receives significant attention.2–4 These new concepts involve peratures, As and Af , and during unloading when reaching
solid-state materials which exhibit magneto-,5,6 electro-,7,8 or

MA . It is well known that both types of transformation
elastocaloric,9,10 effects. In the present work, we focus on
cycles, cooling/heating and loading/unloading, are associated
thermal effects in Ni–Ti-based shape memory alloys (SMAs), with a hysteresis.17–19 Recently, we have documented, that a
which are often referred to as elastocaloric in literature.9–12 linear correlation exists between thermal and mechanical
Ni–Ti SMAs are known for their fascinating ability to re- hysteresis widths for various different Ni–Ti and Ni–Ti–Cu-
establish their initial geometry after a deformation much based SMAs.19
larger than typical elastic strains.13,14 Two types of shape Both martensitic and reverse transformations are associ-
memory effects can be exploited, a thermal memory (one/two ated with relatively large latent heats, ΔH.20 This makes
way effect) and a mechanical memory (pseudoelasticity). Ni–Ti attractive as a candidate material for ferroic cooling.3
Figure 1 exemplarily shows an elastocaloric cooling cycle
§ where a binary Ni–Ti SMA ribbon is (i) mechanically
Corresponding author.

1740001-1
 A. Wieczorek et al.

(a) (b)

Fig. 1. Ferroic cooling cycle of a Ni–Ti ribbon: (a) stress/strain behavior determined at a strain rate of 0.1 s 1 . (b) Infrared (IR)-images showing temperature
changes in the SMA ribbon during loading (1–3), holding (3–4) and unloading (4–6).

loaded, then (ii) held at constant strain for 1 min, and then pseudoelastic Ni–Ti focus on the thermo-mechanical behav-
(iii) unloaded. Figure 1(a) shows the underlying mechanical ior in bulk10 and thin film22,23 materials. Schmidt et al.24,25
Funct. Mater. Lett. Downloaded from www.worldscientific.com
by FUDAN UNIVERSITY on 02/28/17. For personal use only.

behavior. A stress/strain plateau starting at a stress level close and Tusek et al.26 designed demonstrator devices where
to 400 MPa is observed during loading. The plateau is related binary Ni–Ti SMAs were used for cooling. Frenzel et al.20
to the formation of stress-induced martensite. During showed that transformation temperatures and latent heats
unloading, the reverse transformation occurs at a lower stress for Ni–Ti-based SMAs strongly depend on alloy chemistry.
plateau. The area between both loading and unloading curves Binary Ni–Ti, as well as most ternary and quaternary Ni–Ti-
corresponds to the amount of work (W) characterizes one based alloy systems, show a strong coupling between trans-
cooling cycle.9,12 Both forward and reverse transformations formation temperatures and latent heats,20 i.e. alloys with
are associated with a considerable heat exchange as shown in large/low Ms temperatures exhibit large/low latent heats. In
the thermographic/infrared (IR) images in Fig. 1(b). The their publication, Frenzel et al.20 rationalize the above-
color coded temperature images document the temperature mentioned correlation and explain why Ms and ΔH so strongly
changes when passing through the different loading/unload- depend on the Ni-concentration in binary Ni–Ti.
ing stages which are labeled \1" to \6" in Figs. 2(a) and 2(b). The present work aims at a compositional and micro-
The IR images show that the exothermic formation of mar- structural optimization of Ni–Ti-based SMAs specially for
tensite results in a temperature increase from 35 ○ C to 65 ○ C, ferroic cooling. In a first step, we analyze the thermal
whereas the reverse (endothermic) transformation results in a transformation behavior of binary, ternary and quaternary
decrease from 35 ○ C to 5 ○ C. compositions to identify potential candidate alloys. In a
A recent review article on heat effects in SMAs was second step, selected alloys were processed by thermo-
published by Qian et al.9 Previous studies on heat effects in mechanical treatments to obtain thin ribbons. The (thermo-)

(a) (b)
24,25,31
Fig. 2. SMA cooling demonstrator. (a) CAD image of the setup. The SMA ribbon is subjected to cyclic uni-axial loading/unloading. The ribbon can be
horizontally transferred between two Cu-blocks in order to establish heat exchange. (b) Color coded IR-image of the test setup during operation.

1740001-2
 Optimizing Ni–Ti-based shape memory alloys for ferroic cooling

Table 1. Composition ranges considered and the experiments were performed at a displacement rate
in the present work.
of 0.5 mm/min. Mechanical tests were performed at 10 K
Alloy system x (at.%) above Af . All further experimental details are given
Ni50þx Ti50x 0.1 • x • 1.2
elsewhere.19,29
Ni50 Ti50x Crx 0.1 • x • 1.25 An effort was made to assess the cooling performance of
Ni50x Ti50 Cox 0.5 • x • 7.5 two selected alloys by using an instrumented ferroic cooling
Ni50x Ti50 Fex 0.25 • x • 3
Ni50 Ti50x Vx 0.1 • x • 6
demonstrator shown in Fig. 2.24,25,31 Figure 2(a) shows a
CAD model of the cooling setup. A SMA ribbon is cyclically
deformed by a 1.2 kN linear actuator. After each loading/
mechanical behavior of these materials was characterized unloading step, the ribbon is quickly transferred to one of the
by differential scanning calorimetry (DSC) and tensile tests. two Cu-blocks which act as heat sources (blue) and heat
We define four criteria to assess the potential of each alloy sinks (red). Temperature changes in the ribbon and in the
for ferroic cooling: Cu-blocks are monitored by a thermography camera and by
Pt100 sensors. Figure 2(b) shows an IR-image of the de-
1. Af < room temperature, monstrator setup during operation. All details on the ferroic
2. large latent heat ( ΔH), cooling setup are given in Refs. 24, 25 and 31. In the present
3. low required work input (W) and work, the setup operates at strain rates of 0:1  10 1 s 1 .
Funct. Mater. Lett. Downloaded from www.worldscientific.com

4. functional stability for a high number of cycles.

by FUDAN UNIVERSITY on 02/28/17. For personal use only.

Contact times of 3 s were applied for optimum heat transfer

All four criteria are important for ferroic cooling. Af below between the SMA ribbon and the Cu blocks.32
room temperature is required to use the pseudoelastic effect at
low temperatures. The second and third criteria yield a high 3. Results
cooling efficiency. The cooling efficiency is often expressed
by the coefficient of performance (COP) which can be deter- 3.1. Transformation Behavior
mined as: COP ¼ Δ H =W (see Refs. 9 and 27). Finally, a good Figure 3 exemplarily shows DSC curves of binary (for ref-
cyclic stability is required for long-term operation. It must be erence) and various ternary SMAs. The exothermic peaks
kept in mind that SMA-based ferroic cooling is associated (positive heat flow) correspond to the formation of martensite
with multiple loading and unloading cycles. on cooling, whereas endothermic peaks (negative heat flow)
are related to the reverse transformation on heating. The areas
below the peaks correspond to the latent heats. Figure 3(a)
2. Experiments shows that adding Cr, Ni or V to stoichiometric NiTi results
In order to obtain low Af temperatures, Ni, Co, Cr, Fe and V in a decrease in both transformation temperatures and latent
were added to stoichiometric NiTi. Details on alloy compo- heats. The same trend was observed for Ni–Ti–Co and
sitions are given in Table 1. In addition, three quaternary Ni–Ti–Fe as exemplarily shown in Fig. 3(b). We note that
alloys, Ni39:83 Ti50 Cu5Co5:17 , Ni42:87 Ti50 Cu5Fe2:13 and Ni–Ti–Co and Ni-Ti-Fe show two-step transformations on
Ni45 Ti47:25 Cu5V2:75 were investigated, which are known for a cooling and heating, Fig. 3(b). The first transformation peaks
low thermal hysteresis width.28 All alloys were prepared on cooling and heating are related to the formation of R
as 50 g ingots by arc melting using high purity elemental phase.33,34 Table 2 shows Af , Ms =Rs (onset of first transformation
raw materials as a feedstock (for alloy compositions and event observed on cooling) and Δ Htot data for stoichiometric
purities, see Ref. 29). We use an optimized melting procedure Ni–Ti–Fe and Ni–Ti–Co alloys. ΔHtot refers to the total
which is documented in Refs. 20 and 30. All samples were latent heat resulting from both transformation events during
homogenized at 1000 ○ C for 24 h followed by water heating. Characteristic transformation temperatures and latent
quenching. We use DSC to determine characteristic phase heats for Ni-rich Ni–Ti, Ni–Ti–Cr and Ni–Ti–V are given
transformation temperatures and latent heats. All details elsewhere.20
on thermal analysis are given in Refs. 20 and 30. Thin rib- In Fig. 4, the latent heats ΔH and ΔHtot are plotted as a
bons (thickness: 0.7 mm, width: 2.5 mm) were prepared function of the austenite finish temperature Af . The dashed ver-
by rolling in combination with heat treatments using pro- tical line in Fig. 4 indicates room temperature, which represents
cessing routes documented in Ref. 29. Final material states the upper Af limit for the ferroic cooling cycle. Figure 4 shows
were obtained by recrystallization heat treatments after that all different alloy systems exhibit similar trends. A decrease
applying degrees of cold work close to 0.35. The mechanical in Af is associated with a decrease in ΔH = Δ Htot . This type of
behavior of the ribbons was characterized by uniaxial tensile correlation can be rationalized by the thermodynamic scenario
testing in a climate chamber. The gauge length was 37.5 mm which was recently presented in Ref. 20. From the results shown

1740001-3
 A. Wieczorek et al.

(a) (b)

Fig. 3. DSC charts of different Ni–Ti-based alloys. (a) Binary reference alloy, ternary alloy with V and Cr (replacing Ti). (b) Binary reference alloy, ternary
alloy with Co and Fe (replacing Ni).
Funct. Mater. Lett. Downloaded from www.worldscientific.com
by FUDAN UNIVERSITY on 02/28/17. For personal use only.

Fig. 4. Correlation between heats of transformation and Af temperatures. Fig. 5. DSC charts for different Ni50 Ti44 V6 material states.
The arrows indicate three selected candidate alloys.

in Fig. 4, we selected three potential candidates (highlighted by the production of thin ribbons by rolling and annealing can
arrows in Fig. 4): Ni51:1 Ti48:9 , Ni44 Ti50 Co6 and Ni50 Ti44 V6. affect phase transformation behavior. Figure 5 exemplarily
Ni44 Ti50 Co6 appears as the most promising candidate since it shows DSC charts of the Ni50 Ti44 V6 alloy after different
shows the highest latent heats. treatments. Rolling in combination with annealing at 500 ○ C
The data in Fig. 4 and Table 2 were obtained from cast and 600 ○ C yields flat DSC curves. We assume that these
material states after homogenization annealing. We note that material states contain high dislocations densities, which
suppress phase transformations. In contrast, annealing at
Table 2. Selected transformation parameters of Ni–Ti–Co and Ni–Ti–Fe. 800 ○ C establishes one-step transformations on heating and
cooling, whereas forward and reverse transformations occur
Composition Af = ○ C Ms =Rs = ○ C ΔHtot /(J/g)
at slightly lower temperatures than in the homogenized cast
Ni50 Ti49:5 Co0:5 103.5 62.3 28.6 material state.
Ni50 Ti49 Co1 91.5 45.5 24.9
Ni50 Ti46 Co4 50.5 22.0 25.8
Ni50 Ti44:5 Co5:5 23.3 4.8 22.5 3.2. Mechanical and functional behavior
Ni50 Ti44 Co6 18.1 1.6 20.2
Ni50 Ti43 Co7 5.0 12.8 15.4 Figure 6 presents the cyclic mechanical behavior of the three
Ni50 Ti42:5 Co7:5 0.8 19.6 14.1
alloy candidates. Each material state was subjected to 10
Ni50 Ti49:75 Fe0:25 96.0 58.6 24.7
Ni50 Ti49 Fe1 70.1 42.7 26.7 loading/unloading cycles (N) in order to assess functional
Ni50 Ti48: Fe2 39.8 20.6 24.1 stability. Figure 6(a) shows stress/strain curves for Ni51:1 Ti48:9:
Ni50 Ti47:5 Fe2:5 26.0 12.0 20.7 Two different material states (final heat treatments: (500 ○ C/
Ni50 Ti47 Fe3 15.1 2.1 16.5
10 min and 550 ○ C/3 min) yield a pseudoelastic behavior.

1740001-4
 Optimizing Ni–Ti-based shape memory alloys for ferroic cooling

(a) (b)
Funct. Mater. Lett. Downloaded from www.worldscientific.com
by FUDAN UNIVERSITY on 02/28/17. For personal use only.

(c) (d)

Fig. 6. Effects of composition and final heat treatment on stress/strain behavior. All samples were subjected to N ¼ 10 loading/unloading cycles, (a) binary
Ni-rich Ni–Ti, (b) ternary Ni–Ti–Co, (c) ternary Ni–Ti–V and (d) Ni–Ti–Cu–V SMA.

However, a 500 ○ C/10 min anneal results in a relative exhibit a pseudoelastic behavior, Fig. 6(c). Annealing at
large mechanical hysteresis width (
350 MPa in the first 500 ○ C for 10 min yields a narrow hysteresis at relatively high
cycle) and low accumulation of residual strains during stresses and low maximum strains (red curves). We assume
unloading. After changing the heat treatment to 550 ○ C/3 min, that this behavior is related to the presence of high dislocation
a significantly lower hysteresis width (
220 MPa) is obtained. densities, as can be concluded from the DSC chart shown in
However, this materials state accumulates > 0:5% residual Fig. 5, where no phase transformations were detected on
strain during cycling. The two material states require 3.1 J/g cooling and heating. The other two heat treatments yield lower
and 2.5 J/g, respectively, as input work (W) for the first me- transformation stresses and relatively low hysteresis widths.
chanical cycle. However, the cyclic stability of these two material states is
Adding Co to binary Ni–Ti significantly alters the me- only moderate. Both material states-accumulate > 0:5% re-
chanical behavior. Figure 6(b) shows cyclic stress/strain sidual strain within 10 loading/unloading cycles.
curves for Ni44 Ti50 Co6, which was previously identified as Figure 6(d) shows the mechanical behavior of a
the most promising candidate alloy for ferroic cooling (due to Ni45 Ti47:25 Cu5V2:75 alloy. Annealing at 500 ○ C for 10 min
large latent heats, Fig. 4). Unfortunately, both heat treatments establishes an excellent pseudoelastic behavior. The hyste-
(800 ○ C/10 min and 500 ○ C/10 min) fail in establishing good resis width is very low (close to 100 MPa), and only very
pseudoelasticity. The stress/strain curves in Fig. 6(b) reveal small residual strains accumulate during cycling. The dissi-
very large residual strains after unloading. We assume that pative work W was determined as 0.5 J/g. This is signifi-
both material states mainly deform by conventional plasticity, cantly lower than in the case of binary Ni–Ti, Fig. 6(a).
and not by the formation of stress-induced martensite. To demonstrate the cooling behavior, binary Ni51:1 Ti48:9
All material states of the third candidate alloy Ni50 Ti44 V6, and Ni45 Ti47:25 Cu5V2:75 ribbons were employed as a cooling
which were obtained by 500 ○ C, 600 ○ C and 800 ○ C annealing, material in the demonstrator setup presented in Fig. 2. The

1740001-5
 A. Wieczorek et al.

(green ¼ good, yellow ¼ moderate, red ¼ poor) is used in

Table 3 for those cases which allowed to assign a COP value.
Important key findings, which can be derived from Table 3,
are: (1) The Af temperature for binary Ni-rich Ni–Ti strongly
depends on the final annealing treatment. Similar observa-
tions are documented in Ref. 35. (2) Adding Co to Ni–Ti and
Ni–Ti–Cu promotes conventional plastic deformation, which
is indicated by a poor shape recovery during mechanical
cycling. The same effect was observed when Fe is added to
Ni–Ti–Cu. (3) A Ni45 Ti47:25 Cu5V2:75 SMA can be identified
as the most promising candidate SMA for ferroic cooling.
Although, the latent heats of this material are not so large,
this alloy combines several key parameters which outperform
Fig. 7. Ferroic cooling performance of Ni51:1 Ti48:9 (annealed at 550 ○ C for
3 min) and Ni45 Ti47:25 Cu5V2:75 in the cooling demonstrator shown in Fig. 2. those of other systems: a sufficiently low Af temperature, a
very small mechanical input work (W) and thus a very good
maximum temperature drops observed in the SMA ribbons, COP value, and an excellent functional stability. Table 3 shows
which were detected directly after unloading prior to estab- that the COP value of Ni45 Ti47:25 Cu5V2:75 is four times higher
Funct. Mater. Lett. Downloaded from www.worldscientific.com
by FUDAN UNIVERSITY on 02/28/17. For personal use only.

lishing contact with the Cu blocks, were 18 K for binary Ni–Ti than that of binary Ni-rich Ni–Ti.
and 21 K for the Ni–Ti–Cu–V SMA. During cyclic operation, Figure 8 represents an overview on the ferroic cooling
the evolution of temperatures in the heat sink and heat source performance of various SMA systems, which was recently
was monitored, Fig. 7. After 450 s, which corresponds to 60 published in an overview article by Qian et al.9 The different
loading/unloading cycles, a temperature change close to alloy systems were evaluated on the basis of the COP value
3=  3 K was accumulated in both Cu blocks. This indicates that and the maximum adiabatic temperature drop. Alloy systems
both materials can be successfully used as a solid-state cooling which are located in the upper right exhibit the best ferroic
material. We note that slightly higher temperature changes were cooling performance. Qian et al.9 show that the COP value
obtained in both Cu blocks for the Ni–Ti–Cu–V alloy. can depend on the loading condition of the material. For
example, binary Ni–Ti shows a significantly better COP
4. Discussion value in compression than in tension. These differences are
related to the well known stress/strain asymmetry in com-
The present study shows that Ni–Ti-based SMAs can be pression and tension.36 Typically, compressive loading yields
successfully used as functional materials for ferroic cooling. lower mechanical hysteresis widths, which translate into
The cooling performance depends on transformation tem- higher COP values. In Fig. 8, we add our new data point
peratures, latent heats, hysteresis widths and functional which represents our Ni–Ti–Cu–V SMA (tensile load).
stability. The data presented in Figs. 3, 4 and 6 clearly show While Ni-rich binary Ni–Ti exhibits a COP value of 5 in
that these key parameters are governed by alloy chemistry, tension and 18 in compression, 20 was obtained for the
microstructures and processing conditions. Table 3 sum- Ni45 Ti47:25 Cu5V2:75 SMA. Figure 8 shows that only two other
marizes key parameters for selected SMA material states material systems reach similar high COP values: ternary Ni–
which were investigated in the present study. A color code Ti–Cu and Cu–Zn–Al. However, these materials have other

Table 3. Compilation of ferroic cooling key parameters for Ni–Ti-based SMAs (material state: ribbon after final heat treatment).

Composition Heat treatment ( ○ C/min) Af /( ○ C) ΔHtot /(J/g) W/(J/g) COP/() Strain recovery (N ¼ 1)/% Strain recovery (N ¼ 10)/%

Ni51:1 Ti48:9 500/10 31.8 15.1 3.1 4.8 99.1 95.4

Ni51:1 Ti48:9 550/3 4.0 14.6 2.5 5.8 97.9 87.4
Ni50 Ti44 V6 800/10 1.5 10.3 0.7 14.7 96.6 85.1
Ni45 Ti47:25 Cu5 V2:75 500/10 5.2 9.2 0.5 20.0 99.3 98.7
*Ni39:83 Ti50 Cu5 Co5:17 * 500/10 0.2 18.3 2.6 7.0 53.4 35.6
*Ni42:87 Ti50 Cu5 Fe2:13 * 500/10 0.5 19.9 1.9 10.5 86.1 65.3
Ni50 Ti44 V6 500/10 — 0 0.4 — 99.9 97.8
Ni50 Ti44 V6 600/10 — 0 2.1 — 92.2 83.2
Ni44 Ti50 Co6 500/10 16.4 14.8 7.1 — 20.0 15.3
Ni44 Ti50 Co6 800/10 18.3 22.1 4.1 — 10.6 7.1

1740001-6
 Optimizing Ni–Ti-based shape memory alloys for ferroic cooling

(3) A correlation between latent heats and transformation

temperatures was observed for all alloy systems con-
sidered in the present work.
(4) Highest latent heats were observed for Ni–Ti–Co SMAs.
However, this material does not show pseudoelasticity
during mechanical loading.

Acknowledgments
The authors acknowledge the support of the DFG priority
program 1599 \Caloric effects in ferroic materials: New
Fig. 8. Comparison of cooling efficiencies (COP) and maximum adiabatic
temperature drops ( ΔTad ) of different SMAs according to Ref. 9. The red concepts for cooling" (Project Nos. EG101/29-2, SE704/2-2,
arrow indicates the new Ni–Ti–Cu–V SMA investigated in the present work. SCH2217/3-2).

disadvantages as compared to our Ni–Ti–Cu–V alloys: tern-

ary Ni–Ti–Cu alloys exhibit relatively high transformation References
temperatures, when subjected to thermo-mechanical proces-
Funct. Mater. Lett. Downloaded from www.worldscientific.com
by FUDAN UNIVERSITY on 02/28/17. For personal use only.

1. J. Bollen et al., Co-benefits of climate policy, Netherlands En-

sing routes which establish a good cyclic stability.28 Cu–Zn–Al
vironmental Assessment Agency, Report no. 500116005 (2009).
SMAs only show a good performance as single crystals.37 2. I. Sarbu and C. Sebarchievici, Energy. Convers. Manage. 105,
The cooling performances of binary Ni51:1 Ti48:9 and 403l (2015).
Ni45 Ti47:25 Cu5V2:75 during operation in the ferroic cooling 3. S. Fähler et al., Adv. Eng. Mater. 14, 101 (2012).
demonstrator merit further analysis. It was observed that the 4. X. Moya et al., Nat. Mater. 13, 4391 (2014).
quaternary Ni–Ti–Cu–V alloy provided slightly larger tem- 5. M. H. Phan and S. C. Yu, J. Magn. Magn. Mater. 308, 3251
(2007).
perature changes than binary Ni–Ti, see Fig. 7 and text
6. A. M. Tishin et al., Int. J. Refrig. 68, 1771 (2016).
above. This result was not expected because the latent heat of 7. D. Sette et al., Apl. Mater. 4, 091101 (2016).
Ni45 Ti47:25 Cu5V2:75 (9.2 J/g) is significantly lower than that 8. R. Chukka et al., Europhys. Lett. 103, 47011 (2013).
of Ni51:1 Ti48:9 (14.6 J/g), Table 3. However, there are two 9. S. Qian et al., Int. J. Refrig. 64, 11 (2016).
findings which may rationalize this observation: first, pre- 10. J. Tusek et al., J. Appl. Phys. 117, 124901 (2015).
liminary results reported in Ref. 29 indicate that the Ni–Ti– 11. E. Bonnot et al., Phys. Rev. Lett. 100, (2008).
12. J. Cui et al., Appl. Phys. Lett. 101, 073904l (2012).
Cu–V alloy has a lower heat capacity than binary NiTi.
13. K. Otsuka and M. Wayman, Shape Memory Materials, 1st edn.
Second, Ni–Ti–Cu–V shows a relatively homogeneous phase (Cambridge University Press, Cambridge, UK, 1998).
transformation behavior during mechanical cycling.38 In 14. C. Wayman and T. Duerig, Engineering Aspects of Shape
contrast, binary Ni-Ti shows macroscopic transformation Memory Alloys, 1st edn. (Butterworth-Heinemann, London,
bands with heat maxima/minima.38–40 Further work is re- UK, 1990).
15. J. Van Humbeeck, Mater. Sci. Eng. A 273–275, 134l (1999).
quired to clarify how the formation of transformation bands
16. Y. Liu et al., J. Alloy. Compd. 247, 1151 (1997).
affects ferroic cooling behavior. 17. J. M. Ball and R. D. James, Phil. Trans. R. Soc. Lond. A 338,
3891 (1992).
18. R. Zarnetta et al., Smart Mater. Struct. 19, (2010).
5. Summary and Conclusions 19. S. Jaeger et al., Int. J. Mater. Res. 106, 10291 (2015).
20. J. Frenzel et al., Acta Mater. 90, 213l (2015).
In the present work, the effects of alloy composition and
21. J. Tusek et al., Adv. Energy Mater. 5, 1500361 (2015).
processing conditions on key parameters which are relevant 22. C. Chluba et al., Science 348, 10041 (2015).
for the ferroic cooling performance of SMAs were investi- 23. H. Ossmer et al., Apl. Mater. 4, 064102l (2016).
gated. The following key results were obtained: 24. M. Schmidt et al., Int. J. Refrig. 54, 88l (2015).
25. M. Schmidt, Elastokalorisches Kühlen mit Ni-Ti-basierten
(1) Alloy composition and final heat treatment after thermo- Formgedächtnislegierungen: Thermodynamische Analyse,
mechanical processing strongly affect the ferroic cooling experimentelle Untersuchungen, Prozessoptimierung, Uni-
performance of a pseudoelastic SMA. versität des Saarlandes, Saarbrücken, Germany (2017).
(2) A Ni45 Ti47:25 Cu5V2:75 alloy with a low hysteresis width 26. J. Tusek et al., Sci. Technol. Built. Environm. 22, 4891 (2016).
27. H. Ossmer et al., Acta Mater. 81, 91 (2014).
was identified, which exhibits excellent functional sta- 28. B. Maaß, Strukturbildungsprozesse bei der Herstellung und
bility and a very high ferroic cooling efficiency. The funktionelle Eigenschaften pseudoelastischer Ni-Ti-Cu-(X)-
COP value of this SMA was determined as 20, which is Formgedächtnislegierungen, Ruhr-Universität Bochum, Bochum
four times higher than in the case of binary Ni-Ti. (2012).

1740001-7
 A. Wieczorek et al.

29. A. Wieczorek, Werkstoffwissenschaftliche Untersuchungen 33. J. Pfetzing-Micklich et al., Adv. Eng. Mater. 12, 131 (2010).
zum Kühlen mit Formgedächtnislegierungen Ruhr-Universität 34. J. Pons et al., J. Phys. Iv 5, 2931 (1995).
Bochum, Bochum (2016). 35. M. Drexel et al., The Effects of Cold Work and Heat Treatment
30. J. Frenzel et al., Acta Mater. 58, 34441 (2010). on the Properties of Nitinol Wire, ASM International 2008,
31. M. Schmidt et al., Elastocaloric cooling with Ni-Ti based Medical Device Materials IV (Conference), pp. 114.
alloys: material characterization and process variation, ASME 36. Y. Liu et al., Acta Mater. 46, 43251 (1998).
2015 Conf. on Smart Materials, Adaptive Structures and In- 37. L. Manosa et al., Funct. Mater. Lett. 2, 731 (2009).
telligent Systems (ASME, Colorado, USA, 2015), pp. 38. M. Schmidt et al., Shape Mem. Superelasticity 1, 1321 (2015).
V001T01A010. 39. C. Elibol and M. F. X. Wagner, Mater. Sci. Eng. A 621, 761 (2015).
32. M. Schmidt et al., Techn. Mess. 83, 2081 (2016). 40. M. L. Young et al., Acta Mater. 58, 23441 (2010).
Funct. Mater. Lett. Downloaded from www.worldscientific.com
by FUDAN UNIVERSITY on 02/28/17. For personal use only.

1740001-8