Ni2MnGa(100) ferromagnetic shape memory alloy: A surface study

This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

http://www.elsevier.com/copyright

Author's personal copy

Ni2MnGa(100) ferromagnetic shape memory alloy: A surface study

S.W. D′Souza a, J. Nayak a, M. Maniraj a, Abhishek Rai a, R.S. Dhaka a, S.R. Barman a,⁎, D.L. Schlagel b,T.A. Lograsso b, Aparna Chakrabarti c

a UGC-DAE Consortium for Scientific Research, Khandwa Road, Indore, 452001, Madhya Pradesh, Indiab Ames Laboratory U. S. DOE, Iowa State University, Ames, Iowa 50011-3020, USAc Raja Ramanna Centre for Advanced Technology, Indore, 452013, Madhya Pradesh, India

a b s t r a c ta r t i c l e i n f o

Article history:Received 26 April 2011Accepted 29 July 2011Available online 1 September 2011

Keywords:Photoemission spectroscopyLow energy electron diffractionFerromagnetic shape memory alloyDensity functional theoryMartensite transition

Ni2MnGa(100) single crystal studied using low energy electron diffraction (LEED) and ultraviolet photoemissionspectroscopy (UPS) exhibits interestingmodification of the surface properties that aremainly influenced by sur-face composition as well as intrinsic effects. In the martensite phase, the LEED spot profiles show presence of anincommensurate modulation for the stoichiometric surface. In contrast, a commensuratemodulation is observedforMn-excess Ni–Mn–Ga surface. A pre-martensite phase is identified at the surface. Both the surfacemartensiticand pre-martensitic transition temperatures decrease as the Mn content increases. The UPS spectra in the aus-tenite phase exhibit systematic change in shape as a function of surface composition that can be related tochanges in the hybridization between Ni and Mn 3d states. The spectra in the martensite phase exhibit interest-ing modifications near the Fermi level, which has been compared to density of states calculated for a modulatedstructure by ab-initio density functional theory. Intrinsic surface properties dissimilar from the bulk are en-hanced hysteresis width of the martensite transition and increased pre-martensitic transition temperature.

© 2011 Elsevier B.V. All rights reserved.

1. Introduction

There is a surge of interest in the basic physics of Ni–Mn–Ga ferro-magnetic shape memory alloy because of its potential as a functionalmaterial with large magnetic field induced strain [1,2], as an energymaterial with giant magnetocaloric effect [3] and as a magnetic sen-sor with large negative magnetoresistance [4]. Ni–Mn–Ga exhibitsmagnetic field induced strain because of its fundamental propertieslike large magnetocrystalline anisotropy and high mobility of thetwin boundaries. The latter is related to low twinning stress, whichis associated with a modulated structure in the martensite phase[1,2]. Neutron scattering studies of stoichiometric Ni2MnGa singlecrystal showed that the modulation is related to a phonon anomaly[5]. Shapiro et al. identified well-defined phasons associated withcharge-density-wave (CDW) excitation below 200 K in themartensitephase [6]. Theoretically, it has been shown from density functionaltheory (DFT) that in the martensite phase, an instability in the TA2phonon mode related to Fermi surface nesting assists the formationof CDW [7]. Recently, an alternative explanation of the modulationhas been suggested from X-ray diffraction studies on a Ni excessNi2MnGa film [8]. This is based on the adaptive martensite modelthat comprises of the branching twin variants with small surface ener-gy [9]. Recent photoemission study on Ni2MnGa showed temperature

dependent variation of the spectra in the pre-martensite phase andevidence of pseudogap was observed at 0.3 eV below the Fermi level(EF) [10].

Ni2MnGa, which is ferromagnetic with Curie temperature of375 K, has a cubic L21 structure in the austenite phase. Around200 K, it undergoes a first order structural transformation to thelower symmetry martensite phase. Generally, Ni–Mn–Ga composi-tions that display modulated structure have low twinning stressand exhibit magnetic field induced strain [2]. However, whether themodulation exists at the surface of the bulk single crystal is an inter-esting question that has not been studied till date. In general, how abulk first order transition manifests itself at the surface constitutesan interesting study. Furthermore, surface study of Ni2MnGa is im-portant because it could modify the magnetic and structural proper-ties of the adlayers grown on it. Recently, we have reported lowenergy electron diffraction (LEED) and photoemission studies on stoi-chiometric surface of Ni2MnGa and Mn2NiGa at room temperaturewhere the surface was found to be Mn–Ga terminated [11]. Very re-cently, the (001)-oriented surface of epitaxial off-stoichiometric Ni–Mn–Ga ferromagnetic shape memory alloys was studied and theyalso observed that the surface is Mn–Ga terminated [12]. In thiswork, we report temperature dependent LEED study on the (100)surface of a bulk stoichiometric Ni2MnGa single crystal and identifythe pre-martensite and martensite phases. Periodic modulation inthe martensite phase of Ni2MnGa is observed. The nature of the mod-ulation depends on the surface stoichiometry: it is incommensuratefor the stoichiometric surface, while it is commensurate for the Mn

Surface Science 606 (2012) 130–136

⁎ Corresponding author. Tel.: +91 731 2463913x123; fax: +91 731 2465437.E-mail address: [email protected] (S.R. Barman).

0039-6028/$ – see front matter © 2011 Elsevier B.V. All rights reserved.doi:10.1016/j.susc.2011.07.023

Contents lists available at SciVerse ScienceDirect

Surface Science

j ourna l homepage: www.e lsev ie r .com/ locate /susc


excess surface. The width of the hysteresis in the martensitic transitionat the surface is found to be five times larger than the bulk value. Thesurface martensitic and pre-martensitic transition temperatures de-crease as the Mn content increases. The photoemission spectra exhibitsystematic change in shape as a function of surface composition.

2. Experimental

Ni2MnGa single crystal was grown at the Materials PreparationCenter of the Ames Laboratory by the Bridgman method [13]. Theingot was oriented in the austenite phase by Laue back reflectionand the sample was spark cut. The polishing was done mechanicallyusing quarter micron diamond paste followed by electropolishing innitric acid and methanol. The bulk composition was confirmed bywavelength dispersive X-ray spectroscopy. LEED was performed at abase pressure of about 4×10−11 mbar using ErLEED optics fromSpecs GmbH, Germany. The stoichiometric Ni2MnGa(100) surfacewas prepared by sputtering with 1.5 keV Ar+ ions and annealingat 770 K for 1 h. The cleanliness and the surface composition wereensured using X-ray photoelectron spectroscopy (XPS). The photo-emission studies were performed by using a commercial electronenergy analyzer (Phoibos 100 from Specs GmbH, Germany). For ultra-violet photoemission spectroscopy (UPS)measurements, He I (21.2 eV)radiation was used. The analyzer was set at a pass energy of 2 eV andmedium area lens mode with collection angle of ±7° was used. To ob-tain a single variant state in the martensite phase, the crystal was cutalong [10] direction (corresponds to [100] direction of body centeredtetragonal unit cell with I4/mmm space group in the bulk) and clampedalong the vertical i.e. [01] direction in a sample holder designed forstudying complex metal surfaces [14]. To check this, we recordedLEED patterns as a function of position over a grid on the crystal at 12different positions. In each case, the patterns in both martensite andaustenite phases remained unaltered. In this sample holder, the ther-mocouple is placed below the sample for reliable temperaturemeasure-ment. The martensite transition temperatures for the bulk crystal havebeen determined using differential scanning calorimetry. The latentheat is 0.2 kJ/mole in agreement with literature [15].

3. Results and discussion

3.1. LEED of stoichiometric Ni2MnGa(100) in martensite phase

Ni2MnGa is in the austenite phase at room temperature since itsMs (martensitic start temperature) is 206.5 K. Thus, the (100) surfaceexhibits a four-fold LEED pattern (Fig. 1(a–b)), as reported earlier[11,12]. A spectacular modification of the LEED pattern occurs in themartensite phase with the appearance of an array of satellite spotscorresponding to each of the spots in the austenite phase. Thesespots are observed over the range 85bEpb120 eV in the [10] direction(Fig. 1(c–f)). In the corresponding intensity profiles shown in Fig. 2,the spots appearing along (00)→(

P20) are marked as 1–5, where 1 is

closest to the (00) spot. Similarly, spots that appear along(00)→(20) are designated by numbers 1′–5′. The intensity profilesin Fig. 2(b) show that all the satellite spots move toward the (00)spot as Ep increases. Comparison of the profiles shows that 5 and 5′spots in the martensite phase nearly coincide with (2

P0) and (20)

spots in the austenite phase, indicating that these are the fundamen-tal reflections in the martensite phase (Fig. 2(a)). As mentioned in theprevious section, to compare with the bulk, note that the direction ofcompression (vertical i.e. [01]) corresponds to the shorter c axis of thebulk martensite phase with respect to the b.c.t. cell in the austenitephase. The b axis of the bulk is along [10], and this is the directionin which the modulation appears in the bulk.

To understand the origin of the satellite spots in the martensitephase, we first examine their separations. If the four spots (1–4) be-tween the fundamental reflections (00) and spot 5 are equispaced,

this would mean a five times increase of the surface unit cell in thereal space. However, detailed analysis of the line profiles bringsforth an interesting result that the satellite spots are not equispaced.This is observed in all the profiles shown in Fig. 2(b): the separationbetween spots 2 and 3 is clearly smaller than the separation betweenother spots. This indicates that the surface structure is not simple andgenerally unequal separation of the diffraction spots is indicative of amodulated structure. The theoretical formulation of Bragg reflectionsfrom modulated structures shows that the scattering vectors areg±nq, where g is a reciprocal lattice vector, q is the modulation wavevector and n is an integer [16]. Thus, each fundamental spot of thebasic structure given by g is surrounded by a series of equally spacedsatellite reflections of diminishing intensity at distances ±nq. The in-tensity of the satellite reflections is less if the deviation of atom posi-tions from the basic structure is small, and all the satellites might notbe observed [16]. Unequal separation is observed between satellite re-flections corresponding to the different fundamental reflections whenthe basis vector (gp) of the reciprocal lattice is not an integral multipleof q, indicating presence of modulation. If q/gp is rational (irrational),the nature of the modulation is commensurate (incommensurate).

To quantify the spot separations, we have considered 37 profilesfor averaging, including those shown in Fig. 2(b). The separationsare obtained as a fraction of the basis vector of the reciprocal latticein the non-modulated martensite phase [17]: bMS* (=g10) obtainedas half of the separation between spot 5 and (00) (Fig. 1(f)). Thespot separations turn out to be 0→1: 0.413±0.002; 1→2: 0.421±0.002; 2→3: 0.287±0.001; 3→4: 0.430±0.003 and 4→5: 0.431±0.002, where the error is given by the standard deviation divided by

85

c

90

dA

B

95

e

1005

(00)

5'

f

b*

MS

austenitic phase (300 K)

martensitic phase (100 K)

100

(00)(20)

(20)

b

(11)

85

a

[10]

Fig. 1. Low energy electron diffraction (LEED) patterns of Ni2MnGa(100) in (a–b) theaustenite phase at room temperature and in (c–f) the martensite phase at 100 K. Theelectron beam energies (Ep) in eV are shown in the bottom right corners. The first Bril-louin zone(white dashed square) and the direction AB along [10] outside first Brillouinzone are shown in (d). In (f), dashed lines indicate the 5 and 5′ spots and the whitearrow represents bMS* .

131S.W. D′Souza et al. / Surface Science 606 (2012) 130–136


square root of the number of data points. From the above analysis, it isclear that the fundamental reflections like spot 5 and (00) are sur-rounded by satellites corresponding to a surface modulation wave vec-tor qS with maximum n=2, as shown in Fig. 2(b). Averaging over 50spot separations excluding the 2→3 separation, we obtain qS=(0.4246±0.0012)bMS* . Note that although the proximity of 2′ and 3′spots is observable in Fig. 2(a), the prime numbered spots have notbeen considered in the averaging because 1′ is hidden by the shadowof the electron gun and 3′ and 4′ peak positions are uncertain due topresence of a dark patch on the screen.

The characteristics of the surface modulation can be determinedfrom the value of qS and related to the bulk modulated structure. Itis well known that the bulk modulation is associated with a shuffleof the (010) planes along the [100] direction of the tetragonal primi-tive cell of the austenite phase [17,18,20]. Thus, bulk truncation atthe (100) surface would result in surface modulation along [10] i.e.bMS* direction and this is indeed observed here. There is disagreementin literature about the character of the modulation in the bulk. Basedon neutron diffraction studies [18], it was found that the structure ofstoichiometric Ni2MnGa is 7M (i.e. 7 layered) type with a commensu-rate bulk modulation wave vector (qB). qB is a rational fraction of thebasis vector of the reciprocal lattice given by 3

7 bTM i.e. 0.4286bM* . How-

ever, incommensurate modulation was reported for Ni2MnGa fromneutron scattering studies by Zheludev et al. with qB=0.43bM* [5].A recent transmission electron diffraction study showed that the sat-ellite spots have unequal spacing and the modulation was reported tobe incommensurate [19]. Powder X-ray diffraction study using super-space approach by Righi et al. also reported an incommensurate qBvalue of 0.4248(3)bM* [17], which differs from 3/7 by 0.0038. Thus,although qB is close to 3/7 implying a nearly 7 fold increase in unitcell length i.e. 7M modulation as per our definition, on the basis ofthe convention that observation of four superlattice spots betweenthe main spots indicates 5M modulation, Righi et al. described their

structure to be 5M. The above discussion shows thatmost of the studieson bulk Ni2MnGa so far conclude that the modulation is incommensu-rate. In the present study, =(0.4246±0.0012)bMS* for the stoichiomet-ric Ni2MnGa(100) surface is very close to the incommensurate qB value(0.4248) reported for bulk Ni2MnGa [17]. The similarity of qS to qB dem-onstrates that the magnitude of the modulation wave vector isunchanged at the surface, and on the basis of the closeness of qS to theincommensurate bulk value, we assign the surface modulation to beincommensurate.

The intensity of the satellite spots is expected to decrease as theorder (n) increases and such variation has been observed in bulk dif-fraction studies [2,19–22]. However, this is not the case here (Fig. 2).For example, the fundamental reflection (spot 5) is lower in intensitythan spot 4 (n=1) and the relative intensities of the satellite spotschange with Ep: spots 1–3 become intense, while spots 4 and 5show a maximum at Ep=95 and 100 eV, respectively (Fig. 2(b)).This could result from surface buckling, which is a possibility inNi2MnGa because it has large Mn local moment [23,24]. Recent scan-ning tunneling microscopy studies on Ni–Mn–Ga film grown on MgO(001) also suggest outwards relaxation of one atomic species (Mn orGa) [12]. Thus, the displacement of the atoms from the equilibriumpositions that gives rise to modulation would be more complicatedat the surface compared to the bulk. Since in Ni2MnGa, the constitu-ent atoms are close in atomic number and the unit cell consideringmodulation is large, determination of the atomic positions at the sur-face would be a challenging task for future.

The modulation has recently been explained by an adaptive mar-tensite phase where nano-twinning of the non-modulated tetragonalphase occurs [9]. For example, (5

P2)2 twinning was shown in a Ni

excess Ni–Mn–Ga thin film grown on MgO [8]. Notably, diffractionpattern showing incommensurate modulation, as observed here, canbe simulated by random faulted nano-twins in the adaptive martens-ite model. The difference in the diffraction pattern between a real in-commensurate phase and random faulted nano-twins is that for thelatter the satellite spots vary significantly between different Brillouinzones [9]. The satellite spot separations outside the first Brillouin zonein the [10] direction in Fig. 1(d) remains unchanged (for examplealong AB). Thus, we observe from the LEED patterns that the satellitespot positions in the [10] direction in different Brillouin zone areunchanged within our measurement accuracy (Fig. 1). This indicatesthat the incommensurate modulation observed here does not arisefrom random faulted nano-twins. However, it is possible that theexpected variation as per the adaptive martensite model is smallerthan our measurement accuracy.

In the martensite phase, as the temperature is decreased from200 K to 100 K, we find that the satellite peaks do not shift in positionbut become more intense with respect to the fundamental spot(Fig. 3). For example, the satellite spot 4 is of similar intensity asthe fundamental spot 5 at 200 K, whereas spot 4 is more than doubledin intensity at 100 K. Similarly, intensities of satellite spots 1–3 in-crease as the temperature is lowered. This is an indication that astemperature is lowered, surface phonon vibrations are suppressed.

3.2. LEED of stoichiometric Ni2MnGa(100) as a function of temperature

In this section, we study the evolution of the LEED pattern as afunction of temperature for the stoichiometric Ni2MnGa(100). Thus,the transition temperatures have been determined and the pre-martensite phase has been identified at the surface. The martensitetransition is a first order transition, but the width of the hysteresisis small for thermoelastic materials like Ni2MnGa, where the twinboundaries are mobile. From heat flow data for the bulk crystal, theMs is 206.5 K and the austenitic start temperature As is 215.7 K.Thus, the width of the hysteresis in the bulk is 9 K (=As−Ms). Atthe surface, appearance of the martensite satellite spots during cool-ing gives the surface martensitic start temperature (MsS) to be 200K

inte

nsity

(ar

b.un

its)

25020015010050

123451' 2'

3'4' 5'

(00)a (20)(20) austenite martensite

E p= 95 eV

inte

nsity

(ar

b.un

its)

1601401201008060

pixel number

b

95

855 4 3 21

(00)

90

100

Ep

qS

2qS qS

2qS

(eV)

Fig. 2. (a) Comparison of the austenite and martensite phase intensity profiles ofNi2MnGa(100) in [10] and [

P10] directions. (b) A series of intensity profiles in the mar-

tensite phase at 100 K along [P10] for different Ep. The modulation wave vector qS is

shown by red arrows. 1 pixel corresponds to 0.0226bMS* .

132 S.W. D′Souza et al. / Surface Science 606 (2012) 130–136


(Fig. 4). From the heating cycle, the surface austenitic start tempera-ture AsS is 243 K. Thus, the width of the hysteresis is 43 K (AsS−MsS),which is about five times larger than in bulk value (9 K). This obser-vation is of general importance for first order transitions at surfaces,and indicates lesser mobility of the twin boundaries at the surfacecompared to the bulk. Broader thermal hysteresis at the surface com-pared to the bulk was observed earlier for martensite transformationin cobalt metal, which was related to higher thermal equilibriumtransition temperature of the surface or that the sliding of the planesis better pinned at the surface [25].

Another interesting observation from Fig. 3 is the appearance ofsatellite spots above MsS that does not correspond to the austenitephase, for example between 270 and 200 K. These spots are absentboth in the austenite (above 270 K) and martensite (below 200 K)phases. Moreover, these spots cannot be explained as co-existingmartensite and austenite phases, rather it depicts a different surfacephase. This phase appears in a temperature range aboveMsS. Interest-ingly, the temperature range where this phase appears shifts to lowervalues concomitant with decrease of MsS with surface composition(discussed in sub-section D). In this phase, four satellite spots appeararound the fundamental reflection of the austenite, two each in [10]and [01] directions, respectively. Since it appears as a precursor tothe martensitic phase, in the literature related to bulk studies, thisphase has been termed as the pre-martensitic phase [5,18,19,26–29]. The pre-martensite phase was shown to be a first order structuraltransition driven by magnetoelastic coupling [27]. Inelastic neutronscattering experiments on stoichiometric Ni2MnGa showed the exis-tence of a soft TA2 phonon mode at q=0.33 that incompletely con-denses at the pre-martensitic transition temperature, TP=260 K[5,26]. Earlier TEM studies showed existence of similar satellites be-tween fundamental reflections in the (

P110) direction, and two satel-

lites in the (110) direction that were assigned to the pre-martensitephase [19,29].

3.3. LEED of Mn excess Ni2MnGa(100) surface in the martensite phase

To examine the effect of non-stoichiometry on the modulation, wehave studied a Mn rich Ni2MnGa surface annealed at 830 K. For thissurface, the LEED satellite spots appear at 157 K (=MsS) while cool-ing. The surface composition from XPS is Ni1.8Mn1.1Ga1.1 and the cor-responding Mn:Ni ratio is 0.6. LEED patterns recorded at 100 K on thissurface clearly exhibit non-equispaced satellite spots indicating pres-ence of modulation, as shown by the line profiles in Fig. 5. The lineprofiles using different Ep and from different parts of the surface

have been analyzed and found to be similar. qS has been determinedby averaging about 66 qS values and we obtain the value to be0.4282±0.0016. Surprisingly, this value is significantly differentfrom the incommensurate qS value (=0.4246) for the stoichiometricsurface. On the other hand, it is almost equal to the commensuratevalue of 3

7 (=0.4286). Thus, it can be concluded that the Mn richNi2MnGa surface exhibits a commensurate 7M modulation. Notethat for Mn excess Ni–Mn–Ga (Ni2Mn1.05Ga0.95), Ranjan et al.reported from X-ray diffraction study that the bulk structure exhibitscommensurate 7M modulation [30].

inte

nsity

(ar

b.un

its)

160140120100806040

pixel number

T(K)

200

150

100

54

3

2

1

(00)

125

175

Fig. 3. LEED intensity profiles along [P10] for Ni2MnGa(100) surface with Ep=90 eV at

different temperatures in the martensite phase. 1 pixel corresponds to 0.019bMS* .

270

260

247

205

201

195

180

160

140

120

100

300

Fig. 4. LEED patterns for Ni2MnGa(100) with Ep=100 eV, as a function of temperatureduring cooling.



3.4. Transition temperatures as function of surface composition

For Ni–Mn–Ga, the dependence of Ms on bulk composition is wellknown in literature [15,31,32]. This motivated us to determine thesurface transition temperatures from LEED and examine their varia-tion with composition. In a previous work [11], we have shown thatthe surface becomes Ni-rich and Mn deficient after sputtering. But,as the annealing temperature is increased, Mn segregates to the sur-face and for temperatures higher than 800 K, the surface becomesMn rich. Mn enrichment at the surface by annealing has been ob-served in other Mn containing alloys like quasicrystalline Al–Pd–Mn[33], and half Heusler alloy NiMnSb [34], where annealing to 600−900 K restores the approximate bulk composition at the surface.Thus, by annealing at different temperatures after sputtering, theMn to Ni ratio can be varied at the surface and its influence on MsS

can be studied by LEED. As in our earlier work,[11] the Mn:Ni ratiois obtained by recording the XPS spectra for the Mn 3p, Ni 3p andGa 3d region under same analyzer settings and sample geometry,finding the areas of the respective peaks and dividing by the photo-emission cross-sections. The surface composition corresponds to anapproximate depth of about 20 Å, which is the mean free path ofthe photoelectrons. In Fig. 6, MsS, as determined by observing theLEED pattern, is plotted against Mn:Ni ratio. The accuracy of MsS iswithin ±2K, since the LEED pattern changes sharply at the transitiontemperatures. MsS clearly decreases from 200 K to about 120 K as Mncontent increases. Interestingly, in bulk Mn excess Ni–Mn–Ga, de-crease of Ms has also been reported in the Mn:Ni ratio range of 0.5to 1; and for even higher Mn, Ms increases [32,35]. But, the decreasein Ms is more pronounced in the bulk (Fig. 6). For example, Ms is aslow as 50 K for Mn:Ni=1 i.e. for the composition Ni1.5Mn1.5Ga[32,35], whereas for a similar surface composition MsS is 120 K. Thereasons for this difference could be that the electronic and magneticproperties of Mn excess Ni–Mn–Ga depends sensitively on the posi-tion of the excess Mn atoms, i.e. whether they occupy equivalent orinequivalent sites of the L21 structure. In the bulk, for inequivalentMn atom positions as nearest neighbors, it has been shown that a fer-rimagnetic alignment will be energetically favored and the decreasein magnetic moment with Mn content has been explained [36,37].This might be different for the surface. Moreover, the surface mighthave more disorder than the bulk, and the presence of disorder atthe surface is evident from the background intensity of the LEED pat-terns. So, for the Mn excess Ni–Mn–Ga surface, the surface magnetic

properties and electronic structure might be somewhat differentfrom the bulk that could result in higher MsS.

Neutron diffraction studies reported that the pre-martensitephase exists between 200 and 255 K and has a three fold modulatedorthorhombic structure [18]. We have determined the surface pre-martensitic start temperature (PsS) as the temperature where thepre-martensite LEED spots appear. We find that PsS decreases sub-stantially as Mn:Ni ratio increases (Fig. 6). Moreover, the differenceof PsS and MsS also decreases for the Mn excess surface. PsS=270 Kfor the stoichiometric surface is substantially higher than the bulkvalue (255 K) [18], Our finding of enhanced PsS at the surface is inagreement with a X-ray scattering study that reported an increaseof 15 K in the near surface region [38].

3.5. Photoemission spectra in austenite phase for different surfacecompositions

A proof that the surface composition changes with annealing tem-perature is obtained from the UPS valence band spectra at room tem-perature that show dramatic changes in shape with Mn:Ni ratio(Fig. 7). A broad rounded peak appears around 0.9 eV for a Ni-richsurface with Mn:Ni=0.27. For Mn:Ni=0.3, this peak shifts tohigher binding energy (BE) of 1.15 eV, and a shoulder is observed ata lower BE (0.6 eV). For Mn:Ni=0.35, the main peak shifts to1.27 eV, and the shoulder decreases in intensity. As the Mn:Ni ratioapproaches the stoichiometric values of 0.48–0.52, the main peakappears at 1.32 eV, and the shoulder is observed as a small peak at0.6 eV. Thus, the main peak position shifts from 0.9 eV (Mn:Ni=0.27) to 1.32 eV (Mn:Ni=0.42). If compared to the Ni excessspectrum (Mn:Ni=0.27), the width of the main peak for thestoichiometric surface is substantially reduced. As the Mn contentincreases above the stoichiometric value (Mn:Ni=0.84), a sharppeak at 0.3 eV is observed. The states close to the Fermi level havesubstantial Mn 3d contribution [36,39]. So, for the Mn excesssurface, this is manifested as an intense peak at 0.3 eV. Densityfunctional theory calculations performed by us for Mn2NiGa indicatethat this peak is related to hybridized 3d states between Ni and theexcess Mn that occupies the Ni site [36,39]. While the valence bandspectra changes with the Ni:Mn ratio, the LEED shows an austenitephase at room temperature for all the compositions.

inte

nsity

(ar

b. u

nits

)

1601401201008060

pixel number

1

23

4

5

100

95

90

(00)

Ep

(eV)

qS

2qS

qS

2qS

Fig. 5. Intensity profiles of Mn excess (Mn/Ni=0.6, Ni1.8Mn1.1Ga1.1) Ni2MnGa(100) sur-face in the martensitic phase at 100 K along [

P10] for different Ep. 1 pixel corresponds to

0.0226bMS* .

300

250

200

150

100

50

MsS

, Ms

, PsS

(K

)

1.00.90.80.70.60.50.4

Mn:Ni ratio

7.50 7.40 7.32 7.25 7.18 7.13

(e/a)

Pre-martensite surface Martensite surface Martensite bulk

Fig. 6. Variation in the surface martensitic transition temperature (MsS), surface pre-mar-tensitic transistion temperature (PsS) and bulk martensitic transition temperature (Ms)[32,35] as a function of Mn:Ni ratio. The dashed lines serve as guide to the eye. The hori-zontal (vertical) error bars indicate the uncertainty of composition (transition tempera-ture) measurement. The top axis shows the valence electron to atom (e/a) ratio.



3.6. Comparison of photoemission spectra between austenite and martensitephases

In Fig. 8(a), the UPS spectra for stoichiometric Ni2MnGa surface inthe austenite (300 K) and martensite phases (100 K) are compared.The spectra exhibit a broad main peak centered around 1.3 eV,which arises from the hybridization between the Ni 3d and Mn 3dstates [40]. A clear suppression of spectral intensity is observed nearEF in the martensite phase. The difference spectra show a redistribu-tion of spectral weight to higher BE (suppression at 0.3 eV and en-hancement at 0.8 eV) and evidence of a pseudogap at EF (boldarrow). For the Mn excess Ni–Mn–Ga surface (Fig. 8(b)), the mainpeak is shifted to slightly higher binding energy (1.4 eV) and a peakis observed at 0.3 eV in the austenite phase. While the main peak re-mains unchanged, the 0.3 eV peak is suppressed in the martensitephase, as clearly observed in the difference spectrum. In a previoustemperature dependent photoemission study, large modification ofthe photoemission spectra was observed and a decrease of intensity

at 0.3 eV and an increase at 0.8 eV BE was reported in both the pre-martensite and martensite phases [10]. This is in agreement withthe present data for the martensite phase.

The change in the spectral shape in the UPS spectra (Fig. 8) couldbe related to the modulated structure of the martensite phase and if astructure induced pseudogap is present, that could be easily identi-fied by density functional theory, as shown in other systems likeHume–Rothery alloys or in elemental metals like gallium [41,42].For this reason, we have calculated the total and partial density ofstates of both austenite and martensite phases of Ni2MnGa using allelectron ab-initio full potential linearized augmented plane wave(LAPW) method. We calculate the austenite phase using the L21structure. For the martensite phase, we use the commensurate 7Mmodulated structure obtained from neutron diffraction study [18].This calculation involves an orthorhombic Pnnm unit cell with 56atoms, and the parameters used are similar to our earlier theoreticalstudies [37,40,43].

The main peak of the total DOS constitutes of mainly Ni 3d states.In the austenite phase, it is centered around 1.5 eV, in agreement withexperiment. But, in the martensite phase, the DOS main peak shifts toabout 1.8 eV, while the experimental peak does not shift between thetwo phases. The DOS at EF is similar between the two phases and thusthe pseudogap at EF is not observed in the martensite phase. Howev-er, in a small energy range below EF (0.1 to 0.3 eV) suppression of DOSis observed (Fig. 9). The enhancement in the intensity of the spectrumthat is observed in the experiment at 0.8 eV in the martensite phasecould be related to the extra Ni 3d–Mn 3d hybridized states in theDOS around 0.8 eV. However, any enhancement corresponding to apeak in the DOS at 0.5 eV is not observed. Thus, the overall agreementbetween experiment and theory is not good, although a modulatedstructure has been used in the theory. The reason could be that theactual nature of modulation is different from that given in Ref [18].In fact, this has been pointed out in a later X-ray diffraction study[17], who showed that the modulation is incommensurate for stoi-chiometric Ni2MnGa. On the other hand, the adaptive martensitemodel shows that a 52

–modulation exists for a Ni excess Ni–Mn–Ga

film [8]. It is possible that better agreement with experiment mightbe achieved if calculations are performed with the incommensuratestructure proposed for stoichiometric Ni2MnGa [17]. However, at pre-sent, we have not performed this calculation because incommensu-rate modulation can only be approximately considered using a verylarge unit cell. Another possible reason could be that surface relaxa-tion effect has not been considered in the theory, whereas experi-mental evidence point towards surface buckling. Finally, theobserved changes in the spectral shape could also be related to theelectron–phonon coupling in the pre-martensite phase, and this has

3 2 1 0

b300

100

inte

nsity

(ar

b. u

nits

)

3 2 1 0

T (K) a300

100

binding energy (eV)

Fig. 8. UPS spectra for (a) stoichiometric Ni2MnGa(100) and (b) Mn excess (Mn/Ni=0.84) Ni2MnGa(100) surfaces in the austenite (300 K) and martensite (100 K)phases. The 300 K spectra have been shifted along the vertical axis for clarity of presen-tation. The difference spectra shown in the bottom of the graph have been obtained bysubtracting the 100 K spectrum from the 300 K spectrum after normalizing them tosame height at the main peak.

20

15

10

5

0

DO

S (

stat

es /

eV f.

u.)

3 2 1 0 -1

binding energy (EF=0; eV)

Ni2MnGa

martensite austenite

Ni 3d

Mn 3d

total

Fig. 9. Total DOS, Ni 3d and Mn 3d partial DOS for Ni2MnGa in the austenite (red dot-dashed line) and modulated martensite (blue solid line) phases. The total and Ni 3dDOS are shifted along the vertical axis for clarity of presentation, the zero is indicatedby horizontal lines on the right axis.

Inte

nsity

(ar

b.un

its)

3 2 1 0

binding energy (eV, EF=0)

0.27

0.30

0.35

0.42

0.48

0.52

0.84

Mn:Ni

Fig. 7. Ultraviolet photoemission spectroscopy (UPS) valence band spectra for Ni2-MnGa recorded with He I (21.2 eV) radiation at room temperature for different surfacecompositions given by the Mn:Ni ratio. The spectra are staggered along the vertical axisfor the clarity of presentation.



been observed in a recent temperature dependent photoemissionstudy [10].

4. Conclusion

In this paper, we report surface study of Ni2MnGa(100) single crys-tal surface using low energy electron diffraction (LEED) and ultravioletphotoemission spectroscopy (UPS). The properties of the surface areinfluenced by both intrinsic effect related to the free surface as well assurface composition. The changes in surface properties due to composi-tion are as follows: In the martensitic phase, an incommensurate mod-ulation is observed for the stoichiometric surface that is moreprominent at lower temperature. In contrast, a commensurate modula-tion is observed for the Mn excess non-stoichiometric surface. A pre-martensite phase is identified from temperature dependent LEED. Thepre-martensitic transition temperature (PsS=270 K) for the stoichio-metric surface is substantially higher than the bulk (255 K). The surfacepre-martensitic as well as the martensitic transition temperatures de-crease as the Mn content increases. The hysteresis width of the mar-tensite transition is found to be five times larger than the bulk. Thephotoemission spectra in the austenite phase exhibit a clear variationin line shape with surface composition that can be related to thechanges in the hybridization between Ni and Mn 3d states. In the mar-tensite phase, the photoemission spectra show distinct modification inline shape compared to the austenite phase. The theoretical DOS be-tween the austenite and the 7M modulated martensite phase do notfully explain the photoemission data, and the possible reasons are dis-cussed.We envisage that the presentworkwill stimulate further exper-imental research on Ni2MnGa surface.

Acknowledgments

K. Horn, M. Schönberg, S. Singh, A. M. Awasthi, and C. Kamal arethanked for support and useful discussions. Fundings from MaxPlanck Institute, Germany and Department of 20 Science and Tech-nology, Government of India are thankfully acknowledged. S.W.D.and J. N. thank C.S.I.R. for research fellowship.

References

[1] K. Ullakko, J.K. Huang, C. Kantner, R.C. O'Handley, V.V. Kokorin, Appl. Phys. Lett.69 (1996) 1966;S.J. Murray, M. Marioni, S.M. Allen, R.C. O'Handley, T.A. Lograsso, Appl. Phys. Lett.77 (2000) 886.

[2] A. Sozinov, A.A. Likhachev, N. Lanska, K. Ullakko, Appl. Phys. Lett. 80 (2002) 1746.[3] J. Marcos, L. Mañosa, A. Planes, F. Casanova, X. Batlle, A. Labarta, Phys. Rev. B 68

(2003) 094401.[4] C. Biswas, R. Rawat, S.R. Barman, Appl. Phys. Lett. 86 (2005) 202508.[5] A. Zheludev, S.M. Shapiro, P. Wochner, L.E. Tanner, Phys. Rev. B 54 (15) (1996)

045.[6] S.M. Shapiro, P. Vorderwisch, K. Habicht, K. Hradil, H. Schneider, Europhys. Lett.

77 (2007) 56004.[7] C. Bungaro, K.M. Rabe, A. Dal Corso, Phys. Rev. B 68 (2003) 134104.[8] S. Kaufmann, U.K. Röβler, O. Heczko, M. Wuttig, J. Buschbeck, L. Schultz, S. Fähler,

Phys. Rev. Lett. 104 (2010) 145702.[9] A.G. Khachaturyan, S.M. Shapiro, S. Semenovskaya, Phys. Rev. B 43 (1991) 10832.

[10] C.P. Opeil, B. Mihaila, R.K. Schulze, L. Mañosa, A. Planes, W.L. Hults, R.A. Fisher, P.S.Riseborough, P.B. Littlewood, J.L. Smith, J.C. Lashley, Phys. Rev. Lett. 100 (2008)165703.

[11] R.S. Dhaka, S.W. D'Souza, M. Maniraj, A. Chakrabarti, D.L. Schlagel, T.A. Lograsso, S.R. Barman, Surf. Sci. 603 (2009) 1999.

[12] P. Leicht, A. Laptev, M. Fonin, Y. Luo, K. Samwer, N. J. Phys. 13 (2011) 033021.[13] D.L. Schlagel, Y.L. Wu, W. Zhang, T.A. Lograsso, J. Alloys Compd. 312 (2000) 77.[14] R.S. Dhaka, A.K. Shukla, M. Maniraj, S.W. D'Souza, J. Nayak, S.R. Barman, Rev. Sci.

Instrum. 81 (2010) 043907.[15] S. Banik, R. Ranjan, A. Chakrabarti, S. Bhardwaj, N.P. Lalla, A.M. Awasthi, V. Sathe,

D.M. Phase, P.K. Mukhopadhyay, D. Pandey, S.R. Barman, Phys. Rev. B 75 (2007)104107.

[16] S. Van Smaalen, Incommensurate Crystallography, Oxford University Press, Oxford,2007;M.A. Van Hove, W.H. Wienberg, C.-M. Chan, Low-Energy Electron Diffraction,Springer Verlag, Berlin, 1986.

[17] L. Righi, F. Albertini, G. Calestania, L. Pareti, A. Paoluzi, C. Ritter, P.A. Algarabel, L.Morellon, M.R. Ibarra, J. Solid State Chem. 179 (2006) 3525;L. Righi, F. Albertini, L. Pareti, A. Paoluzi, G. Calestani, Acta Mater. 55 (2007) 5237.

[18] P.J. Brown, J. Crangle, T. Kanomata, M. Matsumoto, K.-U. Neumann, B. Ouladdiaf,K.R.A. Ziebeck, J. Phys. Condens. Matter 14 (2002) 10159.

[19] T. Fukuda, H. Kushida, M. Todai, T. Kakeshita, H. Mori, Scr. Mater. 61 (2009) 473.[20] V.V. Martynov, V.V. Kokorin, J. Physique III 2 (1992) 739.[21] J. Pons, R. Santamarta, V.A. Chernenko, E. Cesari, J. Appl. Phys. 97 (2005) 083516.[22] N. Lanska, O. Söderberg, A. Sozinov, Y. Ge, K. Ullakko, V.K. Lindroos, J. Appl. Phys.

95 (2004) 8074.[23] B.L. Ahuja, B.K. Sharma, S. Mathur, N.L. Heda, M. Itou, A. Andrejczuk, Y. Sakurai, A.

Chakrabarti, S. Banik, A.M. Awasthi, S.R. Barman, Phys. Rev. B 75 (2007) 134403.[24] M. Wuttig, Y. Gauthier, S. Blügel, Phys. Rev. Lett. 70 (1993) 3619.[25] M. Erbudak, E. Wetli, M. Hochstrasser, D. Pescia, D.D. Vvedensky, Phys. Rev. Lett.

79 (1997) 1893.[26] U. Stuhr, P. Vorderwisch, V.V. Kokorin, P.-A. Lindgard, Phys. Rev. B 56 (1997)

14360.[27] A. Planes, E. Obrado, A.G. Comas, L. Manosa, Phys. Rev. Lett. 79 (1997) 3926.[28] V.V. Khovailo, T. Takagi, A.D. Bozhko, M. Matsumoto, J. Tani, V.G. Shavrov, J. Phys.

Condens. Matter 13 (2001) 9655.[29] V.V. Kokorin, V.A. Chernenko, E. Cesari, J. Pons, C. Segui, J. Phys. Condens. Matter

8 (1996) 6457.[30] R. Ranjan, S. Banik, S.R. Barman, U. Kumar, P.K. Mukhopadhyay, D. Pandey, Phys.

Rev. B 74 (2006) 224443.[31] A.N. Vasilev, A.D. Bozhko, V.V. Khovailo, I.E. Dikshtein, V.G. Shavrov, V.D. Buchelnikov,

M. Matsumoto, S. Suzuki, T. Takagi, J. Tani, Phys. Rev. B 59 (1999) 1113.[32] G.D. Liu, J.L. Chen, Z.H. Liu, X.F. Dai, G.H. Wua, B. Zhang, X.X. Zhang, Appl. Phys.

Lett. 87 (2005) 262504.[33] Z. Shen, M.J. Kramer, C.J. Jenks, A.I. Goldman, T.A. Lagrasso, D. Delaney, M. Heinzig,

W. Raberg, P.A. Thiel, Phys. Rev. B 58 (1998) 9961;A.K. Shukla, R.S. Dhaka, S.W. D'Souza, M. Maniraj, S.R. Barman, K. Horn, Ph. Ebert,K. Urban, D. Wu, T.A. Lograsso, J. Phys. Condens. Matter 21 (2009) 405005.

[34] J.S. Correa, Ch. Eibl, G. Rangelov, J. Braun, M. Donath, Phys. Rev. B 73 (2006)125316.

[35] S. Banik, S. Singh, R. Rawat, P.K. Mukhopadhyay, B.L. Ahuja, A.M. Awasthi, S.R. Bar-man, E.V. Sampathkumaran, J. Appl. Phys. 106 (2009) 103919;S. Banik, R. Rawat, P.K. Mukhopadhyay, B.L. Ahuja, A. Chakrabarti, P.L. Paulose, S.Singh, A.K. Singh, D. Pandey, S.R. Barman, Phys. Rev. B 77 (2008) 224417.

[36] S.R. Barman, S. Banik, A.K. Shukla, C. Kamal, A. Chakrabarti, Europhys. Lett. 80(2007) 57002.

[37] A. Chakrabarti, S.R. Barman, Appl. Phys. Lett. 94 (2009) 161908.[38] G. Landmesser, U. Klemradt, R.L. Johnson, T.R. Finlayson, Appl. Phys. Lett. 90

(2007) 021907.[39] S.R. Barman, A. Chakrabarti, Phys. Rev. B 77 (2008) 176401.[40] S.R. Barman, S. Banik, A. Chakrabarti, Phys. Rev. B 72 (2005) 184410.[41] G. Trambly de Laissardière, D.N. Manh, L. Magaud, J.P. Julien, F. Cyrot-Lackmann,

D. Mayou, Phys. Rev. B 52 (1995) 7920.[42] S.R. Barman, D.D. Sarma, Phys. Rev. B 51 (1995) 4007.[43] S.R. Barman, A. Chakrabarti, S. Singh, S. Banik, S. Bhardwaj, P.L. Paulose, B.A.

Chalke, A.K. Panda, A. Mitra, A.M. Awasthi, Phys. Rev. B 78 (2008) 134406.


Ni2MnGa(100) ferromagnetic shape memory alloy: A surface study

Documents