Phase dependence of the thermal memory effect in polycrystalline ribbon and bulk Ni55Fe19Ga26 Heusler alloys

Phase dependence o he he mal memo y e ec in polyc ys alline ibbon and bulk Ni 55 Fe 19 Ga 26 Heusle alloys A. Vidal-C espo a , A.F. Manch´ on-Go d´ on b,** , J.M. Ma ín-Olalla a , F.J. Rome o a , J.J. Ipus a , M.C. Galla do a , J.S. Bl´ azquez a,* , C.F. Conde a a Dp o. Física de la Ma e ia Condensada, ICMSE-CSIC, Uni e sidad de Se illa, P.O. Box 1065, 41080, Se illa, Spain b Ins i u o de Ciencia de Ma e iales de Se illa, ICMSE CSIC-Uni e sidad de Se illa, C. Am´ e ico Vespucio 49, Se illa, 41092, Spain ARTICLE INFO Keywo ds: Ma ensi ic ans o ma ion The mal memo y e ec Shape memo y alloys Ni-Fe-Ga Heusle alloys Ul aslow calo ime y ABSTRACT The he mal memo y e ec , TME, has been s udied in Ni 55 Fe 19 Ga 26 shape memo y alloys, ab ica ed as ibbons ia mel -spinning and as pelle s ia a c-mel ing, o e alua e i s dependence on he ma ensi ic s uc u e and he mac os uc u e o he samples. When he e e se ma ensi ic ans o ma ion is in e up ed, a kine ic delay in he subsequen comple e ans o ma ion is only e iden in he ibbon samples, whe e he 14M modula ed s uc u e is he dominan phase. In con as , deg ada ion o he modula ed s uc u e o he p esence o he γ phase signi - ican ly educes he obse ed TME. In such cases, he magni ude o he TME app oaches he de ec ion limi s o comme cial calo ime e s, and only high- esolu ion calo ime e a e y low hea ing a e (40 mK h −1 ) can show he e ec . Following he kine ic a es and subsequen cooling, he e e se ma ensi ic ans o ma ion was comple ed a se e al hea ing a es o con i m he a he mal na u e o he phenomenon. 1. In oduc ion The p ope ies o shape memo y alloys, SMAs, ha e a ac ed conside able in e es making hem p omising o inno a i e enginee ing and mechanical applica ions, p ima ily due o hei unique shape memo y e ec , SME, and supe elas ici y, SE [1]. These dis inc cha ac- e is ics a ise om he ma ensi ic ans o ma ion, MT, ha hese ma- e ials exhibi : a solid-s a e, i s -o de phase ansi ion. MT in ol es he coo dina ed mo emen o a conside able numbe o a oms, achie ing a g ow h eloci y simila o sound wa es [2]. On cooling, a high- empe a u e aus eni e phase wi h high symme y ans o ms in o a low- empe a u e ma ensi e phase wi h educed symme y. The hea ing ans o ma ion is hus called e e se MT. In gene al, he low- empe a u e ma ensi e phase exhibi s a mul idomain mic os uc- u e, esul ing om he dis o ion o he c ys al la ice. This s uc u al dis o ion accoun s o many o he dis inc i e cha ac e is ics o hese ans o ma ions. In addi ion o showcasing SME o SE, SMAs ha e demons a ed hei capaci y o ecall no only speci ic shapes bu also he empe a u es a which he e e se MT was in en ionally hal ed [3,4]. Unlike he well-unde s ood SME, he called he mal memo y e ec , TME, is less clea , and di e en scena ios ha e been p oposed encompassing he edis ibu ion o accumula ed s ess [5,6] o he impac o geome ical cons ic ions [7]. In his phenomenon, when an in en ional a es occu ed a a s op empe a u e Ts op be ween aus eni e s a , As, and aus eni e inish empe a u es, A , i.e. Ts op ∈(As,A ), he ollowing comple e e e se MT shows a kine ic pause close o Ts op. I has been p e iously labeled as he mal a es memo y e ec [8], o as a e e sible s ep-wise ans o ma ion om ma ensi e o aus eni e [9]. In ecen yea s, a ious heo e ical app oaches ha e been p oposed o explain his phenomenon, wi h pa icula emphasis on he ela ion- ship be ween TME and ac o s such as disloca ions [10], in e ace in- e ac ions [11], elas ic de o ma ion o he ma ensi e pla es [3,8], and he con inuous dis ibu ion o s o ed elas ic ene gy, which elaxes du - ing pa ial hea ing cycles [12]. Howe e , he e is no consensus ega ding he o igin o hese e ec s, and hei physical in e p e a ion emains con o e sial. This unce ain y a ises p ima ily because local s ess e olu ion canno be easily measu ed, making i di icul o assess i s impac on ans o ma ion beha io . This is a signi ican echnological challenge, especially since mos eal SMA ac ua o s ypically expe ience only pa ial ans o ma ion cycles unde wo king condi ions [13]. Mo e ecen ly, in-si u neu on di ac ion expe imen s ha e p o ided a * Co esponding au ho . ** Co esponding au ho . E-mail add esses: [email p o ec ed] (A.F. Manch´ on-Go d´ on), [email p o ec ed] (J.S. Bl´ azquez). Con en s lis s a ailable a ScienceDi ec In e me allics jou nal homepage: www.else ie .com/loca e/in e me h ps://doi.o g/10.1016/j.in e me .2025.108695 Recei ed 15 No embe 2024; Recei ed in e ised o m 27 Janua y 2025; Accep ed 1 Feb ua y 2025 In e me allics 180 (2025) 108695 A ailable online 9 Feb ua y 2025 0966-9795/© 2025 The Au ho s. Published by Else ie L d. This is an open access a icle unde he CC BY-NC-ND license ( h p://c ea i ecommons.o g/licenses/by- nc-nd/4.0/ ). comp ehensi e desc ip ion o e olu ion o he moelas ic s ains occu - ing in SMAs du ing pa ial he mal ans o ma ion. As a esul , TME is expec ed in mul i-domain c ys als unde going s uc u al i s -o de phase ansi ions, whe e accommoda ion p ocesses lead o s ain dis- ibu ion among domains [14]. While he TME was ini ially obse ed in Ni-Ti alloys [8], i has also been de ec ed in Ni-Mn-Ga [15,16] and Ni-Fe-Ga [17,18] Heusle alloys. In he la e case, he MT occu s be ween an aus eni e phase wi h a B2 o o de ed L2 1 s uc u e o ei he a modula ed (se en-laye , 14M, o i e-laye , 10M) o a non-modula ed (L1 0 e agonal) ma ensi ic s uc u e. The speci ic s uc u e depends on ac o s such as he compo- si ion and he he mal and mechanical his o y [19,20]. The γ-phase p ecipi a es a e usually o med in con en ionally p oduced samples, esul ing in enhanced mechanical p ope ies. These p ecipi a es can be also easily induced by he mal ea men s in monophasic mel -spun ibbons [21]. This duc ile phase does no pa icipa e in he ma ensi ic phase ansi ion, and i s p esence weakens he SME. As he amoun o he duc ile phase inc eases, i hinde s he mo emen o ma ensi e a ian s, educing he ma e ial’s abili y o eco e s ain. In his sense, po osi y can a ec he ans o ma ion a o ing he mo emen o he win bounda ies due o a highe ee su ace, as epo ed o Ni-Mn-Ga [22]. This phenomenon is enhanced in oams [23]. The e o e, he di e se a ay o mic os uc u es in which MT can be obse ed calls o an examina ion o hei in luence on he TME. This s udy sys ema ically explo es he TME a ising om in e up ing e e se MT in di e en samples o Ni 55 Fe 16 Ga 26 Heusle alloys ab i- ca ed as ibbon-shaped specimens, using mel -spinning, and as pelle - shaped samples, using a c-mel ing. The s udy encompassed he execu- ion o ei he a single in e up ion o he e e se MT o a se ies o in e up ed p ocesses wi h se e al s op empe a u es. The main goal was o compa e how he ma ensi ic s uc u e and he mac os uc u e o he sample in luences he TME. The indings p esen ed in his s udy com- ple e p e ious wo ks by some o he au ho s on Ni-Fe-Ga SMAs, explo ing a ious aspec s: i) he in luence o p essu e on he phase s a- bili y and magne os uc u al ansi ion o ibbons [20], ii) dis inc ions in s uc u al and magne ic p ope ies be ween ibbon and bulk samples [21], iii) he kine ics [24] and iso he mal/a he mal na u e [18] o he MT in ibbons, and i ) he iden i ica ion o a alanches du ing he MT in bulk samples using ul aslow calo ime y [25]. 2. Expe imen al The ma e ial analyzed in his s udy is an alloy wi h a nominal composi ion o Ni 55 Fe 19 Ga 26 (a . %) ab ica ed using bo h mel -spinning and a c-mel ing echniques. Comp ehensi e de ails ega ding he p epa a ion p ocedu es and a ho ough mic os uc u al and magne ic cha ac e iza ion o bo h samples a e a ailable in Re . [21]. Mo eo e , de ailed insigh s in o he ma ensi ic s uc u e a oom empe a u e (RT), he MT p ocess, and he chemical composi ion o he Ni 55 Fe 16 Ga 26 samples u ilized in his s udy can be also ound in ou ea lie publica ion [21]. The mal p ope ies o he samples we e assessed using wo dis inc ins umen s: a comme cial di e en ial scanning calo ime e (DSC) Pe kinElme DSC7 (Pe kinElme , No walk, CT, USA) and a home-made conduc ion calo ime e able o pe o m high- esolu ion DTA analysis (1 nW) [26,27]. DSC was employed unde A low and equipped wi h a cooling sys- em. Va ious hea ing a es, β, ( anging om 5 o 40 K min −1 ) we e applied. To ake in o accoun he impac o he a ying β on he measu ed empe a u e, he mel ing empe a u e o In s anda d (429.75 K) was u ilized (e o s below 0.5 K) when he cooling sys em was ac i e, and he mel ing poin o Pb (600.65 K) was used o he wise. Fo he ibbons, an ini ial s ep hea ing up o 473 K was applied o elimina e he in luence o he e ogenei ies due o s ain ields ha migh be p esen in he sample du ing i s p ocessing, he eby ensu ing con- sis ency and epe i i eness in subsequen measu emen s. Mo eo e , o mi iga e po en ial a ia ions in he cha ac e is ic pa ame e s o he o wa d and e e se MT caused by cycling and inhomogenei ies [24], he same ibbon pieces we e consis en ly used o he same se ies o DSC measu emen s. Gi en he small mass o an indi idual piece ela i e o he c ucible dimensions o he equipmen , mul iple pieces o ibbons we e included in each expe imen o achie e a mass compa able o ha o he u ilized s anda ds. Fo sake o compa ison, a consis en p o ocol was applied o he bulk sample, bu using a single piece o all expe imen s. In TME expe imen s using DSC, samples we e ini ially hea ed a β = 20 K min −1 un il Ts op ∈(As,A )was eached and he empe a u e was hal ed o 10 min. This ime was chosen om ou p e ious esul s [18] ha show ha iso he mal ea men s in hese samples equi e 10 min o dwell ime o s abilize he ans o med phase ac ion. The e o e, a e 10 min, we expec ha empe a u e will be homogeneous and a he mal p ocesses a e comple ed. Then, he samples we e cooled below M a cons an β =-20 K min −1 . In he subsequen s ep, he samples we e ehea ed abo e A a a cons an a e, while employing a ying a es om 5 o 40 K min −1 o he second hea ing s ep o in es iga e he ki- ne ics o he p ocess. Addi ionally, di e en Ts op alues we e u ilized. To cla i y he ollowed p ocess, Fig. 1 illus a es he empe a u e- ime cu e o a comple e DSC expe imen , including an iso he mal dwell conduc ed a 381 K, in e up ed a e 10 min, as an illus a i e example. The inclusion o an iso he mal s ep is necessa y o achie e uni o m empe a u e dis ibu ion wi hin he c ucible, pa icula ly because mul- iple pieces o ibbons a e p esen . The ma ked zone co esponds o ha in which he hea ing a e has been modi ied o di e en expe imen s. The inal empe a u e o second hea ing, which exceeds A , ensu es he elimina ion o any esidual memo y e ec s, as he sample is ully ans o med in o he aus ene phase. S a ing he hea ing p ocess a e s abiliza ion o he calo ime e below M con i ms ha he sample is ini ially in he ma ensi ic s a e. The mal p ope ies we e also s udied in he conduc ion calo ime e capable o cap u ing high- esolu ion DTA aces. De ailed in o ma ion abou he equipmen and i s unc ionali y can be ound in Re s. [26,27]. Due o he subs an ial he mal ine ia o he equipmen , scanning a es a y om a ew kel ins pe hou (slow a e ange) o a ew millikel ins pe hou (ul aslow a e ange), signi ican ly slowe han hose used in comme cial calo ime e s. The sample was a pa allelepiped piece wi h dimensions 9.0 mm ×9.57 mm ×2.35 mm (leng h x wid h x heigh ) and Fig. 1. Illus a ion o a comp ehensi e he mal ea men conduc ed in com- me cial DSC expe imen s o examine he he mal memo y e ec in he in es- iga ed NiFeGa ibbon samples. In he ini ial hea ing s ep (β =20 K min −1 ), he e e se ma ensi ic ans o ma ion was hal ed a a speci ic empe a u e, 381 K, and o a du a ion o 10 min in he epo ed case (1). Subsequen ly, he sample unde wen cooling below ma ensi e inish empe a u e and was hen ehea ed o 473 K a di e en hea ing a es (2). The shaded a ea ep esen s he he mal ea men in which TME should be e lec ed. A. Vidal-C espo e al. In e me allics 180 (2025) 108695 2 1.5987(1) g in mass. Su ace obse a ions o he specimens we e pe o med using a FEI Teneo scanning elec on mic oscope (SEM) ope a ing a 20 kV. Be o e SEM imaging, he su ace o bulk and ibbon samples was mechanically polished using a se ies o sandpape s wi h a ying g i sizes, ollowed by polishing wi h ab asi e pas es. 3. Resul s In his s udy, we ha e analyzed six di e en samples, namely he as- spun ibbon (AS-R), he mally ea ed ibbon (TT-R), p essu e ea ed ibbon (PT-R), pul e ized ibbon (powde ), as-p epa ed bulk (AP-B), and he mally ea ed bulk (TT-B). Table 1 p esen s he key ansi ion empe a u es, encompassing he aus eni e s a , As, and inish, A , empe a u es, along wi h he ma ensi ic s a , Ms, and inish, M , empe a u es. Fu he mo e, he espec i e s uc u e o each sample a RT de e mined by X- ay di ac ion echnique is ou lined. The s uc u e o he AS-R sample displays a monophasic ma ensi e wi h a modula ed 14M s uc u e a RT. Howe e , when he ibbon is c ushed in o powde , an in e ma ensi ic ans o ma ion occu s, leading o a non-modula ed L1 0 s uc u e [20]. When uniaxial p essu e is applied on he ibbons, a pa ially in e ma ensi ic ans o ma ion be- ween he 14M modula ed o a non-modula ed phase occu s [20]. On he o he hand, he mal ea men s can induce he p ecipi a ion o he γ phase [21]. Con e sely, he a c-mel ing me hod ails o yield a single-phase sample, esul ing in he simul aneous p esence o a non-modula ed ma ensi e s uc u e and γ phase a RT. In he bulk sample, he modula ed phase eme ges ollowing an ex ended hea ea men (1073 K o 24 h) and subsequen apid wa e quenching, wi h a small inc ease o γ phase a e his ea men . The cha ac e is ic Table 1 Cha ac e is ic empe a u es associa ed wi h he MT measu ed by DSC scans a 20 K min −1 and he s uc u e o he analyzed samples a RT. ND: Non-de ec ed; AS: As-spun; AP: As-p epa ed; TT: The mally- ea ed; PT: P essu ed- ea ed. In he case o he PT sample, he p esen ed da a co espond o he MT o he 14M s uc u e. Sample S uc u e Ms (K) M (K) As (K) A (K) Ribbon AS 14M 370.3 346.9 379.5 399.5 TT 14M+γ360.6 338.5 371.2 388.0 PT 14M +L1 0 370.3 346.9 379.5 399.5 Powde L1 0 ND ND ND ND Bulk AP L1 0 +γ321.1 288.4 322.9 353.0 TT 14M+γ300.3 283.3 311.9 324.8 Fig. 2. SEM mic og aphs using backsca e ed elec ons a di e en magni ica ions: a) and b) as-spun ibbon, c) and d) he mally ea ed ibbon, and e) and ) powde ob ained om manually g inding he as-spun ibbon. Figu es b) and d) shows highe magni ica ions o selec ed a eas o panels a) and c), espec i ely. A. Vidal-C espo e al. In e me allics 180 (2025) 108695 3 empe a u es o he MT a e lowe in he bulk sample compa ed o he ibbon, in luenced by he al e a ion o he e/a pa ame e due o he p esence o he γ phase wi h a lowe con en in Ga [21]. The mic os uc u e o he samples was analyzed using SEM. Fig. 2 shows ep esen a i e backsca e ed elec on (BSE) images o he AS-R (panels a and b), TT-R (panels c and d), and he powde de i ed om he ibbons (panels e and ). Simila ly, Fig. 3 shows BSE images o he AP-B (panels a and b) and he TT-B samples (panels c and d). The hickness o he ibbons a ied sligh ly, anging om 30 o 40 μ m (Fig. 2a), and p edominan ly ea u ed a single-phase ma ensi ic mic os uc u e. This mic os uc u e consis ed o se e al a ian s o ma ensi e pla es wi h coa se g ains measu ing se e al mic ons in size. The ma ensi e lamellas, exhibi ing a pla e-like shape, we e clea ly isible, along wi h small black po es (indica ed wi h a ows) o med du ing he quenching p ocess om he mel . No signi ican di e ences we e obse ed be ween he ee and wheel side o he ibbons. The mic os uc u e shows ew g ain bounda ies, in which g ain g ow h was inhibi ed by he ini e hickness o he ibbons [28]. Wi hin he ma ensi e pla es, conjuga ion bounda ies, CB, we e isible, sepa a ing egions wi h di e en ly o ien ed nano wins [29]. These CBs we e commonly ound in he sel -accommoda ed s a e o ma ensi e mic o- s uc u e, pa icula ly nea g ain bounda ies whe e b anching occu s [30]. B anching is ealized by a shi in nano win o ien a ion o i s conjuga ed coun e pa wi hin a single ma ensi e pla e, o ming CBs [31]. In he TT-R, he hickness o he ma ensi e pla es inc eased, e en occupying he en i e c oss-sec ion o he ibbon. In e es ingly, despi e he ela i ely low ea men empe a u e and he sho annealing ime, his p ocess appea ed o igge he p ecipi a ion o an addi ional phase dispe sed wi hin he ma ix (indica ed by dashed a ows in Fig. 2d). The p ecipi a ed phase was disce nible by i s da k g ey con as and s ong alignmen oughly pe pendicula o he su ace o he ibbon. Fig. 2e and displays SEM-BSE images o powde s p oduced by manually milling he mel -spun ibbons. The pa icle sizes is no homogeneous and anged om 50 o 150 μ m, p edominan ly exhibi ing a polygonal shape, which esul ed om he b i leness o ibbons due o he absence o gamma p ecipi a es. Fig. 3 con i ms he signi ican a ia ion in mic os uc u e depending on he p epa a ion me hod. Speci ically, while he AS-R samples dis- played a single-phase mic os uc u e, he samples p oduced using he a c-mel ing echnique e ealed he p esence o a clea ly dis inguishable second phase. The mo phology o his second phase changed wi h he mal ea men , leading o an inc ease in he size o he p ecipi a es. The p ecipi a ion o he gamma phase, which is en iched in Fe and impo e ished in Ga, educed he Fe con en in he ma ensi e ma ix. This is quali a i ely shown in Re . [25] by EDS measu emen s. Al hough he quan i a i e changes a e oo small, quali a i e beha io is clea and may lead o an inc ease in e/a ac o , he eby lowe ing he ma ensi ic ans o ma ion empe a u es. I is also wo h no ing ha he dis ibu ion o p ecipi a es in Fig. 2d ollows he same pa e ns as hose obse ed in Fig. 3a. Unlike ibbon samples, po osi y is no obse ed in bulks. Mo eo e , he low empe a u e a which he ibbons ha e been ea ed a e no expec ed o p oduce any e olu ion in he po ous densi y. The e o e, al hough his di e ence may also in luence he a ying beha io obse ed be ween ibbon and bulk samples, i is unlikely o accoun o he di e ences obse ed speci ically among ibbon samples. 3.1. Tempe a u e memo y e ec in Ni 55 Fe 19 Ga 26 as-mel spun ibbons Fig. 4 shows (dashed lines) he di ec and e e se ans o ma ion in he AS-R samples a e con en ional hea ing ( e e se MT) and cooling uns (di ec MT). I is impo an o emphasize ha he dashed line, ep esen ing he en i e he mal cycle o a elaxed sample, is consis en ly ea u ed in he DSC plo s o compa a i e pu poses. We choose β= ± 20 K min −1 o op imize he signal- o-noise a io in DSC measu emen s [18]. No ably, he ibbon specimens unde wen wo p e- he mal cycles om RT o 473 K ( elaxed sample) o emo e po- en ial inhomogenei ies among di e en ibbon pieces, in luenced by p e ious he mo-mechanical his o ies [21]. The ac ual p ocedu e o es ing he TME in ol ed: i) Hea ing om RT o Ts op (AS <TS op <A ); ii) Kine ic s op a Ts op o 10 min; iii) Cooling down o 323 K; and i ) Hea ing om 323 K o 473 K. S eps i) and ii) could be epea ed mul iple imes wi h a ying Ts op in each incomple e e e se MT. Fig. 3. SEM mic og aphs using backsca e ed elec ons: a) and b) as-p epa ed bulk, and c) and d) he mally ea ed bulk a di e en magni ica ions. A. Vidal-C espo e al. In e me allics 180 (2025) 108695 4 Fig. 4a–b displays he cha ac e is ic cu es ob ained om a TME es wi h a single s op, whe e he Ts op was se a 381 K and 383 K, espec- i ely. No ably, kine ic s ops a e clea ly disce nible on he hea low cu es du ing he second hea ing. In addi ion o he single-s op expe i- men s, a double-s op es was execu ed. Fig. 2c displays DSC scans a e pe o ming wo consecu i e incomple e cycles du ing he hea ing phase a empe a u es o 383 K and 381 K in dec easing o de . In such case, wo kine ic s ops show up in he subsequen comple e e e se MT. These indings a e consis en wi h p e ious esul s, ha associa e he numbe o he mal a es s a anged in dec easing Ts op wi h he numbe o in- e up ions in he subsequen hea ing un. In con as , when Ts op is a - anged in an inc easing o de , hen one kine ic s op a he highes empe a u e Ts op is only obse ed [4,32,33]. Incomple e di ec MT we e also es ed on cooling expe imen s, inco po a ing di e en Ts op (be ween M and Ms) and du a ions. He e, only a ac ion o he aus eni e unde goes ans o ma ion in o he ma ensi e phase, lea ing he es o he aus eni e una ec ed. A e his in e up ed di ec MT, he ans o med ma ensi e has he po en ial o e e o aus eni e upon u he empe a u e ele a ion beyond A . Howe e , he ou comes e eal ha he subsequen cooling scans egis- e ing di ec MT does no display any indica ions o kine ic in e up ion. These esul s a e in line wi h hose p e iously epo ed [4,16]. 3.2. Tempe a u e memo y e ec in Ni 55 Fe 19 Ga 26 he mally ea ed ibbons P e ious s udies show ha hea ea men s conduc ed abo e he MT led o an i e e sible downwa d shi in MT due o s uc u al modi i- ca ions. In he case o he mal ea men s pe o med below 573 K, s ess elaxa ion accumula ed du ing ibbon ab ica ion occu ed. Annealing ea men s below 873 K cause sligh changes in he la ice pa ame e s, al e ing he b/a and des abilizing he 14M modula ed s uc u e. A empe a u es abo e 873 K, he monophasic na u e o he ibbon is los , and he ma ens ic modula ed s uc u e ans o ms in o wo dis inc phases: aus eni e and he γ-phase [21]. In o de o analyze he di e - ences in TME due o hese mic os uc u al changes, TT-R samples we e ob ained a e hea ing AS-R samples up o 623 K. Fig. 5 illus a es he DSC scans o he Ni 55 Fe 19 Ga 26 ibbon a e being hea ed up o 623 K. M and Ms ha e dec eased wi h he he mal ea men in compa ison wi h he AS-R. The sample was hea ed un il he MT e e se ans o ma ion was hal ed a 373 K o 10 min. Then, he Fig. 4. DSC scans egis e ed using he p o ocol o Fig. 1 (con inuous black lines) along wi h comple e DSC scans (dashed blue lines) ep esen ing he ansi ion om e e se and di ec MT in Ni 55 Fe 19 Ga 26 ibbons. These scans we e pe o med ollowing ei he a singula a es a Ts op1 =381 K (shown in panel a), a singula a es a Ts op2 =383 K (illus a ed in panel b), and a combina ion o a es a Ts op2 and Ts op1, each las ing 10 min a hei espec i e empe a u es (depic ed in panel c). (Fo in e p e a ion o he e e ences o colou in his igu e legend, he eade is e e ed o he Web e sion o his a icle.) Fig. 5. DSC scan egis e ed using he p o ocol o Fig. 1 using a Ts op =373 K and a dwell ime o 10 min (con inuous black line) along wi h comple e DSC scan (dashed blue line) cap u ing he ansi ion om ma ensi e o aus eni e in he Ni 55 Fe 19 Ga 26 ibbon a e hea ed up o 623 K (TT-R sample). (Fo in e - p e a ion o he e e ences o colou in his igu e legend, he eade is e e ed o he Web e sion o his a icle.) A. Vidal-C espo e al. In e me allics 180 (2025) 108695 5 sample was cooled down o 323 K and he subsequen hea ing un showed a smoo he kine ic in e up ion. This e idence sugges s ha he des abiliza ion o he modula ed s uc u e (b/a<7) was smea ing he TME. Inc easing iso he mal dwell ime up o 60 min does no enhance he TME. The phase ac ion o he 14M s uc u e in he TT-R sample can be in e ed om he en halpy change obse ed in he TT-R Sample and he AS-R sample. Assuming ha he AS-R is 100 % in 14M (ΔH =5.46 J g −1 ), hen we in e ∼90 % o TT-R is 14M phase (ΔH =5.06 J g −1 ). 3.3. Tempe a u e memo y e ec in Ni 55 Fe 19 Ga 26 mechanically ea ed ibbons The applica ion o a uniaxial p essu e on he AS-R sample induces a g adual ans o ma ion om he 14M modula ed phase o he non- modula ed L1 0 s uc u e [20]. Speci ically, a load o 2 ons was applied o a 0.785 cm 2 su ace (~250 MPa) o 5 min a RT o p oduce PT-R sample. DSC scans o PT-R sample a e depic ed in Fig. 6. The dis inc peaks a ibu ed o he MT obse ed in he AS-R exhibi a no iceable dec ease. This sugges s ha he 14M, he ans o mable phase, becomes des abi- lized unde p essu e. F om ΔH =2.64 J g −1 , we in e ha ∼50 % o he PT-R sample is 14M, he emaining hal belongs o he non-modula ed L1 0 phase, as he γ−phase is no induced by p essu e [20]. Addi ion- ally, he in e up ed e e se MT conduc ed a he same Ts op han hose employed in Fig. 1 e eals some TME. Fig. 6c depic s he DSC plo s o he Ni 55 Fe 19 Ga 26 powde s de i ed om as-p epa ed ibbons, exhibi ing a L1 0 monophasic ma ensi e s uc u e (see Re . [20]). Unde he es ed condi ions, no MT is obse ed, indica ing he elimina ion o he modula ed phase h ough mechanical ea men . These indings align wi h p io esea ch on Ni-Mn-Sn milled alloys, whe e he peak associa ed wi h MT diminishes no ably wi h p olonged milling ime, e en ually disappea ing a e 45 min [34]. 3.4. Tempe a u e memo y e ec in Ni 55 Fe 19 Ga 26 as-p epa ed bulk Fig. 7 displays he DSC scans o he Ni 55 Fe 19 Ga 26 AP-B sample, ea u ing a biphasic s uc u e (L1 0 +γ p ecipi a es [21]), ollowing an in e up ed e e se MT a a Ts op =323 K, be ween As and A du ing hea ing. In all es ed condi ions, he e is no indica ion o a kine ic in e up ion o he subsequen comple e ans o ma ion. This appa en Fig. 6. DSC scans egis e ed using he p o ocol o Fig. 1 (con inuous black line) wi h Ts op se a a) 381 K and b) 383 K, along wi h comple e DSC scan (dashed blue line) cap u ing he ma ensi ic ans o ma ion in he Ni 55 Fe 19 Ga 26 PT-R samples. Panel c) co esponds o DSC hea ing and cooling cu es o he Ni 55 Fe 19 Ga 26 powde s de i ed om as-p epa ed ibbons. (Fo in e p e a ion o he e e ences o colou in his igu e legend, he eade is e e ed o he Web e sion o his a icle.) Fig. 7. DSC scans egis e ed using he p o ocol o Fig. 1 (con inuous black line) wi h Ts op =323 K and dwell ime se a 10 min, along wi h comple e DSC scan (dashed blue line) cap u ing he ansi ion om ma ensi e o aus eni e in he Ni 55 Fe 19 Ga 26 as-p epa ed bulk. (Fo in e p e a ion o he e e ences o colou in his igu e legend, he eade is e e ed o he Web e sion o his a icle.) A. Vidal-C espo e al. In e me allics 180 (2025) 108695 6 absence o TME aligns wi h hose documen ed in Re . [17], whe e bulk NiFeGa samples did no exhibi dis inc dips linked o TME a e an incomple e e e se phase ansi ion. 3.5. Tempe a u e memo y e ec in he mally- ea ed bulk The NiFeGa TT-B sample exhibi s a biphasic s uc u e wi h 14M ma ensi e phase and γ−phase p ecipi a es. Fig. 8 shows he DSC e- sul s. Wi h a ΔH =3.30 J g −1 and om he p e ious ΔH o he AS-R sample we in e a ∼60 % 14M and ∼40 % γ−phase mix u e. The TT-B sample was also placed in a home-made conduc ion calo- ime e ha ope a es as a high-sensi i e DTA ace . The sample sus- ained wo ul aslow hea ing uns. As a e e ence, i was hea ed a a a e o 40 mK h −1 and he e e se ans o ma ion was eco ded. Resul s om his expe imen we e shown p e iously in Re . [25]. Then, he sample was cooled om he aus eni e phase and hea ed again a a a e o 40 mK h −1 . The empe a u e was hal ed o i e days a 315.7 K. The sample was hen cooled down a a a e o 40 mK h −1 un il 309 K was eached. I was hen quenched down o 280 K, hea ed back o 306 K when he s anda d hea ing p o ile was esumed, and he sample comple ed he e e se MT. The hea ing DTA aces in he empe a u e egion close o he MT a e shown in Fig. 9. A kine ic s op is clea ly disce nible on he hea low cu es du ing he hea ing a e a smoo h baseline was sub ac ed. 4. Discussion MT empe a u es can a y wi hin a ce ain empe a u e ange, depending on β [19,35]. The e o e, calo ime ic measu emen s a e he mal a es we e conduc ed wi h β anging om 5 o 40 K min −1 o iden i y he dependence o he MT empe a u e in e al a e kine ic a es . Ini ially, hea ing o Ts op ollowed by subsequen cooling o 323 K was conduc ed a 20 K min −1 . Subsequen ly, a hea ing scan up o 473 K was pe o med a di e en β anging om 5 o 40 K min −1 . The esul ing DSC scans a e illus a ed in Fig. 10 (le panels). Panels a and b depic he cha ac e is ic cu es ob ained om a single Ts op o 381 K and 383 K, espec i ely. Addi ionally, a double es was con- duc ed, as shown in panel c. A sligh shi owa ds highe empe a u es is obse ed o A as β inc eases, while β has no disce nible in luence on As. Consequen ly, he ans o ma ion empe a u e in e al ΔT=A −As inc eases wi h β. These esul s a e consis en wi h hose epo ed by Wang e al. [36], sugges ing ha β leads o a la ge a ia ion in A compa ed o As. Righ panels o Fig. 10 show he co esponding ans o med ac ions which we e app oxima ed by he no malized ans o med en halpy, X= ΔH(T)/ΔH o al, whe e ΔH(T) ep esen s he en halpy de eloped up o empe a u e T, and ΔH o al deno es he o al en halpy o he MT. All cu es collapse in o a common line below Ts op, wi hin he s anda d unce ain y associa ed o baseline de e mina ion in he DSC [37]. Howe e , de ia ions can be obse ed abo e Ts op in ag eemen wi h he β e ec desc ibed abo e. When compa ing he ob ained cu es wi h hose o a comple e ans o ma ion wi hou any in e up ion, a kine ic delay e idenced by a shi o he MT o highe empe a u es (∼5K) om Ts op is obse ed. This delay seems o be accumula i e as obse ed in he expe imen wi h wo Ts op in Fig. 10 , wi h a delay >10 K. The impac o he TME on he e olu ion o he ans o med ac ion in he o he s udied samples has also been analyzed and depic ed in Fig. 11. As an icipa ed, he shi o X in he case o he TT-R (Fig. 11a) and PT-R (Fig. 11b) samples is signi ican ly educed compa ed o he AS- R, wi h a shi o highe empe a u es o app oxima ely 1 K. Howe e , in he case o he bulk samples, his shi is wi hin he esolu ion o he equipmen , ega dless o he ma ensi e phase s uc u e (see panels c and d o he AP-B and TT-B samples, espec i ely). Fig. 12 shows he esul s o he high- esolu ion DTA expe imen co esponding o he TT-B sample. In ul aslow DTA expe imen , he ans o med ac ion was app oxima ed by he no malized ans o med en halpy om 310 K (whe e he aus eni e ac ion is ze o) and assuming ha , a 323 K, he aus eni e phase ac ion is 1. The educ ion in he wid h o e e se MT is in ag eemen wi h he hea ing a e e ec desc ibed abo e. When compa ing he ans o med ac ion wi h hose o he comple e e e se MT wi hou in e up ion (dashed-line in Fig. 12), a sligh shi o highe empe a u es can be obse ed. The e o e, i s beha io is simila o ha shown by ibbon samples a e analyzing TME (see Fig. 10), bu no in he case o he bulk. The highe esolu ion o he DTA scans a e y low hea ing a e, along wi h he e y long iso he m dwell, enables us o assess ha TME also occu s in he TT-B sample. The TME obse ed in ul aslow DTA o TT-B sample is highe han ha p esumed o DSC expe imen s. This di e ence could be due o he e y long iso he m and he e ec o educed β, which could be mo e Fig. 8. DSC scans egis e ed using he p o ocol o Fig. 1 (con inuous black line) wi h Ts op =310 K and dwell ime se a 10 min, along wi h comple e DSC scan (dashed blue line) cap u ing he ansi ion om ma ensi e o aus eni e in he Ni 55 Fe 19 Ga 26 he mally ea ed bulk. (Fo in e p e a ion o he e e ences o colou in his igu e legend, he eade is e e ed o he Web e sion o his a icle.) Fig. 9. DTA aces o TT-B sample wi hou he mal a es ( e e ence in blue) and a e he mal a es a 315.7 K (black) in he empe a u e anges a ound e e se ma ensi ic ans o ma ion. (Fo in e p e a ion o he e e ences o colou in his igu e legend, he eade is e e ed o he Web e sion o his a icle.) A. Vidal-C espo e al. In e me allics 180 (2025) 108695 7 e icien in eleasing s ain in he un ans o med ma ensi e phase. As i will be discussed below, his could be a sui able mechanism o TME. In TME expe imen s, du ing he pa ial e e se MT, he ansi ion om ma ensi e o aus eni e hal s a a speci ic empe a u e be ween AS and A . In ha si ua ion, only a po ion o he ma ensi e e e s o he aus eni e phase, while he es o he ma ensi e pe sis s. This esidual ma ensi e is commonly e e ed as M1. Upon subsequen cooling down o M , he aus eni e phase unde goes e- ans o ma ion in o ma ensi e, esul ing in he o ma ion o a new ma ensi e phase, deno ed as M2. Du ing he nex hea ing scan, M2 and M1 sequen ially ansi o he aus eni e phase, leading o a kine ic delay be ween bo h ans- o ma ions. Se e al s udies ha e ied o explain he TME by a ibu ing i o he elease o elas ic s ain ene gy du ing he i s in e up ed ans- o ma ion in he emnan M1. This s o ed elas ic s ain would ac as ins abili ies o launch nuclea ion p ocess in he e e se ans o ma ion [2]. The e o e, M1 is s abilized in a second e e se ansi ion as he elas ic s ain is lowe ed. Howe e , he p oposed mechanisms Fig. 10. DSC scans egis e ed (a,b,c) and co esponding ans o med ac ions (d,e, ) using he p o ocol o Fig. 1 (con inuous lines) along wi h DSC signal o un- ea ed sample (dashed blue lines) ep esen ing he ansi ion om e e se and di ec MT in Ni 55 Fe 19 Ga 26 ibbons a he indica ed hea ing a es. These scans we e pe o med ollowing ei he a singula a es a Ts op1 =381 K (a,d), a singula a es a Ts op2 =383 K (b,e), and a combina ion o a es a Ts op2 and Ts op1, each las ing 10 min a hei espec i e empe a u es (c, ). (Fo in e p e a ion o he e e ences o colou in his igu e legend, he eade is e e ed o he Web e sion o his a icle.) A. Vidal-C espo e al. In e me allics 180 (2025) 108695 8 conside ably a y and a e subjec o deba e. Madangopal e al. [8] sugges ed ha M2 migh accumula e mo e elas ic s ain ene gy han M1, he eby po en ially p e-posi ioning he e e se ans o ma ion o M2. Howe e , we do obse e a common As alue independen ly o he p esence o TME (o e en a delay when e y long iso he ms occu a Ts op) bu a delay ha ini ia e app oxima ely a he empe a u e whe e he hal ha e occu ed (in he absence o e y long iso he m a Ts op). Con e sely, i has been p oposed ha he TME a ises om he educ ion o elas ic s ain ene gy in M1, equi ing a highe empe a u e o ini ia e he ans o ma ion o M1 once he ans o ma ion o M2 is comple e [9]. Ou esul s a e in ag eemen wi h his second in e p e a ion. Upon cooling, he a omic a angemen adop s a winning s uc u e wi h pe iodic s acking o de , po en ially esul ing in 24 a ian s o ma ensi e [13]. The MT s o es some elas ic s ain ene gy wi hin he he moelas ic ma ensi e a ian s. The e o e, while he elease o elas ic s ain ene gy om ma ensi e a ian s is o en associa ed wi h he TME, i is impo an o no e ha he cohe en ene gy o adjacen phases also in luences he TME [38]. This e ec is clea ly obse ed in ou esul s. On he one hand, he p esence o he non-modula ed phase and he des a- biliza ion o he 14M phase es ic s he occu ence o TME (see Figs. 5 and 6). Con e sely, in bulk samples, he p esence o a ound 40 % γ p ecipi a es inhibi s TME o equi e high p ecision echniques o e i- dence i . 5. Conclusions The s udy o he he mal memo y e ec , TME, in ol ed six samples sha ing he Ni 55 Fe 19 Ga 26 composi ion bu di e ing in mic o and mac os uc u e. The esul s show ha he TME, igge ed by a pa ial e e se ans o ma ion om ma ensi e o aus eni e phase, is no a common phenomenon in he s udied shape memo y alloys. Ins ead, i appea s o be con ingen upon he speci ic mic os uc u e o he sample. Fig. 11. T ans o med ac ion as a unc ion o empe a u e o he e e se MT a e kine ic a es in he case o he a) TT-R and b) PT-R samples; and c) AP-B and d) TT-B samples. Inse s show he enla ged 0.4<X <0.6 egion. Fig. 12. T ans o med ac ion as a unc ion o empe a u e o he e e se MT a e kine ic a es in he TTB sample in he ul aslow calo ime e along as o he o iginal ans o med ac ion (dashed line) o o iginal scans. A. Vidal-C espo e al. In e me allics 180 (2025) 108695 9