1: \documentstyle[floats,aps,epsfig,prb]{revtex}
2:
3: \begin{document}
4: \draft
5: % \preprint{\today}
6:
7: \twocolumn[\hsize\textwidth\columnwidth\hsize\csname
8: @twocolumnfalse\endcsname
9:
10:
11: \title{\bf Soft tetragonal distortions in ferromagnetic Ni$_2$MnGa and
12: related materials from
13: first principles}
14:
15: \author{V.V. Godlevsky and K. M. Rabe}
16:
17: \address{Department of Physics and Astronomy,\\
18: Rutgers University,
19: 136 Frelinghuysen Rd, Piscataway, NJ 08854}
20: \date{\today}
21: \maketitle
22:
23: \begin{abstract}
24: A detailed examination of the energy landscape, density of states and
25: magnetic moment of tetragonally distorted ferromagnetic Ni$_2$MnGa was performed using
26: first-principles local-spin-density (LSD) pseudopotential calculations, varying V as
27: well as c/a. The energy of tetragonal Ni$_2$MnGa is found to be {\it nearly constant}
28: for values of c/a over a wide range, with shallow minima near c/a = 1 and 1.08 in
29: addition to that near 1.2. This flat energy surface is consistent with the wide range of
30: observed values of c/a. It also explains the observation of pseudomorphic growth of
31: Ni$_2$MnGa on GaAs, despite a nominal 3\% lattice mismatch. The related materials
32: Ni$_2$MnAl, Ni$_2$MnIn and ferromagnetic NiMn were examined for similar behavior, but
33: all are seen to have a single well-defined minimum at c/a near 1, consistent with
34: available experimental information.
35: For NiMn, the ground state antiferromagnetic ordering and structural parameters are correctly predicted within the LSD approach.
36: \end{abstract}
37:
38: \pacs{71.15.Nc,81.30.Kf,75.50.-y}
39:
40: \vskip1pc]
41:
42:
43: In the design of mechanical actuator devices, both on macroscopic scales and in
44: thin-film based MEMS, there is continuing interest in identifying and optimizing new
45: high-performance materials. Materials which exhibit a martensitic transformation and
46: associated shape memory effect have been shown to be quite useful, though their
47: high-frequency applications are limited by the slow rate of the martensitic
48: transformation and poor energy conversion. It has been proposed\cite{james00} that this
49: deficiency could be addressed by using shape memory materials which are ferromagnetic,
50: and using applied magnetic fields to control the mechanical response
51: \cite{james93,james94,ullakko96}.
52:
53: Therefore, it is of particular interest to identify candidate ferromagnetic martensites
54: and to understand the transition mechanism, especially the coupling between magnetic and
55: structural degrees of freedom. The rather small number of known systems includes Fe-Pt,
56: Co-Ni, Fe-Mn-C, Fe-Ni-C, Fe-Ni-Co-Ti and Fe-Ni, with the Heusler structure compound
57: Ni$_2$MnGa being the most thoroughly investigated to date
58: \cite{ullakko96,ayuela99,webster83,kokorin93,zheludev94,zheludev95,velikohatnyi99}.
59:
60:
61: In this paper, we investigate the structural energetics of Ni$_2$MnGa and related
62: materials from first principles, with a particular focus on the tetragonal distortion
63: associated with the martensitic transition. For comparison, we examine Ni$_2$MnAl and
64: Ni$_2$MnIn (not observed to be martensitic), and NiMn, related by
65: replacing Ga with an additional Mn (exhibits a martensitic transition to an
66: antiferromagnetic tetragonal phase\cite{kasper59}).
67: While our calculations are for bulk
68: crystals, they also allow us to predict the effects of epitaxial stress, which often is
69: the dominant factor in determining the structure and properties of thin films.
70:
71: We use the self-consistent pseudopotential plane wave approach\cite{chelikowsky92}
72: within the local spin density approximation (LSDA). We use
73: Troullier-Martins\cite{troullier91} pseudopotentials and exchange and correlation
74: potential in the Perdew-Wang\cite{perdew92} form. The nonlinear core correction
75: scheme\cite{louie82} was used in the construction of the pseudopotentials. The
76: pseudopotential cut-off radii are summarized in Table I. The local components of the
77: pseudopotentials\cite{kleinman82} are chosen to be l=0 for Ni, Mn, Al and In and l=1 for
78: Ga. The energy plane-wave cut-off is 70 Ry. The unit cells consist of one formula
79: unit for Ni$_2$MnX (X = Al, Ga, and In) and two formula units for NiMn.
80: The Brillouin zone is sampled by
81: a 10x10x10 k-point Monkhorst-Pack (MP)\cite{monkhorst76} mesh with zero shift. To be able
82: to
83: calculate the magnetic moment accurately, we do not apply temperature smearing for
84: the k-point integration.
85: In NiMn, we studied two antiferromagnetic phases:
86: AF-I, with the Mn atoms with parallel moments located on alternating (1 1 0)
87: planes; and AF-II, with parallel moments on alternating (1 1 $\bar{1}$)
88: planes (in the cubic reference system) \cite{sakuma98}.
89: The
90: self-consistent calculations tend to converge to a local ferromagnetic minimum.
91: To make the system converge to the antiferromagnetic state, approximately 0.2 eV/atom
92: lower in energy, we constrain the total
93: magnetic moment to zero.
94:
95:
96: \begin{table}
97: \caption{Reference configurations and cut-off radii (a.u.)
98: used to construct the pseudopotentials.}
99: \vspace{0.2in}
100: \begin{tabular}{lddddddd}
101: \ & $r_s$ & $r_p$ & $r_d$ \\
102: \hline
103: \hline
104: \ Ni $3d^84s^24p^0$ & 2.2 & 2.2 & 2.2 \\
105: \hline
106: \ Mn $3d^54s^24p^0$ & 1.9 & 2.6 & 2.0 \\
107: \hline
108: \ Al $4s^24p^14d^0$ & 2.3 & 2.3 & 2.3 \\
109: \hline
110: \ Ga $4s^24p^14d^0$ & 3.0 & 2.6 & 3.0 \\
111: \hline
112: \ In $4s^24p^14d^0$ & 3.0 & 3.0 & 3.0 \\
113: \end{tabular}
114: \vspace{0.15in}
115: \end{table}
116:
117:
118:
119: For the high temperature cubic structures, our
120: calculations yield the lattice constants, bulk moduli and magnetic moments given in
121: Table II. For NiMn, we compare calculated results for a hypothetical ferromagnetic cubic
122: phase with
123: measurements in the paramagnetic cubic phase.
124: The results compare well with experiment and with previous first-principles studies.
125:
126:
127: \begin{table}
128: \caption{Lattice constants, bulk moduli and magnetic moments of
129: Ni$_2$MnX compounds (L2$_1$ structure) and FM-NiMn (B2 structure)
130: compared to experiment (in parentheses) and previous theoretical
131: calculations (square brackets), where available.}
132: \vspace{0.2in}
133: \begin{tabular}{lddddddd}
134: \ & a (a.u.) \\
135: \hline
136: \hline
137: \ FM-NiMn & 5.56 (5.63$^a$) \\
138: \hline
139: \ ${\rm Ni_2MnAl}$ & 10.93 (11.01$^b$) [10.98$^d$]\\
140: \hline
141: \ ${\rm Ni_2MnGa}$ & 10.91 (11.01$^b$) [10.95$^d$]& \\
142: \hline
143: \ ${\rm Ni_2MnIn}$ & 11.32 (11.47$^b$) [11.43$^e$]& \\
144: \hline
145: \hline
146: %\end{tabular}
147:
148: \ & B (GPa) \\
149: \hline
150: \hline
151: \ FM-NiMn & 155 \\
152: \hline
153: \ ${\rm Ni_2MnAl}$ & 167 [163$^d$] \\
154: \hline
155: \ ${\rm Ni_2MnGa}$ & 170 (146$^c$) [156$^d$] \\
156: \hline
157: \ ${\rm Ni_2MnIn}$ & 138 \\
158: \hline
159: \hline
160:
161: \ & $\mu (\mu_B/Mn)$ \\
162: \hline
163: \hline
164: \ FM-NiMn & 4.42 [3.8$^g$] \\
165: \hline
166: \ ${\rm Ni_2MnAl}$ & 4.22 (4.19$^b$) [4.03$^d$]\\
167: \hline
168: \ ${\rm Ni_2MnGa}$ & 4.22 (4.17$^b$) [4.09$^d$]\\
169: \hline
170: \ ${\rm Ni_2MnIn}$ & 4.31 (4.40$^b$) [3.91$^e$] [4.16$^f$]\\
171: \end{tabular}
172:
173: %\vspace{0.15in}
174: $^a$Reference \onlinecite{egorushkin83} \newline
175: $^b$Reference \onlinecite{webster83} \newline
176: $^c$Reference \onlinecite{worgull96} \newline
177: $^d$Reference \onlinecite{ayuela99} \newline
178: $^e$Reference \onlinecite{daSilva88} \newline
179: $^f$Reference \onlinecite{kilian00} \newline
180: $^g$Reference \onlinecite{sakuma98}
181: \end{table}
182:
183: Next, we present the results of total energy calculations for uniform tetragonal
184: distortions of these cubic structures, and discuss them in the context of observed
185: low-symmetry phases. At least three distinct ferromagnetic low-symmetry phases have been
186: experimentally observed in bulk Ni$_2$MnGa \cite{kokorin93},
187: two tetragonal ($\beta^\prime$ and $\beta^{\prime\prime\prime}$) and one
188: orthorhombic ($\beta^{\prime\prime}$).
189: The $\beta^\prime$ phase, obtained by cooling of the high-temperature
190: Heusler (L2$_1$) phase below T$_m \approx$ 200 K, has a tetragonal structure with c/a =
191: 0.94
192: (a=5.920 \AA~ and c=5.566 \AA).\cite{webster83} The relative volume change across the
193: martensitic transition is only 1\%. In addition, there is a shuffle of the (1 1 0)
194: planes along $[1 \bar{1} 0]$. This has incommensurate periodicity represented
195: by a wavevector of $\pi /a$[0.43,0.43,0], corresponding to approximately five
196: interplanar distances \cite{zheludev94,zheludev95}. The transition is preceded by a
197: cubic premartensitic ($L2_{PM}$) phase below T$_{PM} \approx$ 260 K, characterized by
198: softening and
199: condensation of transverse acoustic phonons $\pi /a$[0.33,0.33,0] with displacements
200: similar in
201: type to the shuffle distortion below T$_m$. Under uniaxial compression along [1 1 0]
202: the $\beta^\prime$ phase transforms into the orthorhombic
203: $\beta^{\prime\prime}$ phase (a = 5.54 \AA, b = 5.78 \AA~ and c = 6.12
204: \AA)\cite{kokorin93}. X-ray analysis shows that $\beta^{\prime\prime}$
205: martensite is also modulated with $[1 \bar{1} 0]$ displacements of the (1 1 0) planes
206: with a periodicity of seven (1 1 0) interplanar distances. The third martensite phase
207: $\beta^{\prime\prime\prime}$, obtained by further compression along [1 1 0],
208: is metastable with respect to the removal of the applied stress \cite{kokorin93}.
209: The
210: $\beta^{\prime\prime\prime}$ structure is a uniform tetragonal distortion of the cubic
211: Heusler structure (a = 6.44 \AA~and c = 5.52 \AA) with c/a = 1.18. The
212: relationships between these phases can be summarized as follows:
213:
214: \begin{equation}
215: L2_1 \;\; \stackrel{T=260 K}{\longrightarrow} \;\; L2_{PM}
216: \;\; \stackrel{T=200 K}{\longrightarrow} \;\; \beta^\prime
217: \;\; \stackrel{[1 1 0]}{\stackrel{stress}{\longrightarrow}}
218: \;\; \beta^{\prime\prime}
219: \;\; \stackrel{[1 1 0]}{\stackrel{stress}{\longrightarrow}}
220: \;\; \beta^{\prime\prime\prime}.
221: \end{equation}
222:
223: \begin{figure}
224: \centerline{\epsfig{file=01.ps,width=3.3in}}
225: \vspace{0.4cm}
226: \caption{Total energy as a function of c/a in Ni$_2$MnGa. Each curve
227: corresponds to a constant volume V per unit cell,
228: measured relative to V$_{Heusler}$ = 1300 a.u.$^3$
229: The curves are separated by intervals of $\Delta V/V=1.5$\%.
230: Each curve represents an interpolation through 26 points, shown explicitly by filled circles for V = V$_{Heusler}$.}
231: \label{Fig.ni2mnga}
232: \end{figure}
233:
234: The calculated energy of Ni$_2$MnGa as a function of c/a for V = V$_{Heusler}$ is shown in
235: Figure \ref{Fig.ni2mnga}. The region of c/a from 0.9 to 1.3 is examined in considerable
236: detail, with 26 separate calculations. Based on the experimental observations of bulk
237: crystal structures, one would expect to find a local minimum in the energy as a function of
238: c/a near 1.18 and perhaps another near 0.94. There is a shallow low-energy local minimum
239: near c/a = 1.19 with total energy 2 meV/atom higher than the L2$_1$ structure. In the
240: previous calculations of Ayuela $et.al.$, this minimum is at c/a=1.16 and is slightly lower
241: in energy than the L2$_1$ structure. The minimum can readily be associated with the pure
242: tetragonal $\beta^{\prime\prime\prime}$ phase observed by Kokorin $et.al.$ with
243: c/a=1.18\cite{kokorin93}. In contrast, there is no local minimum or any discernible anomaly
244: at c/a = 0.94, with a smooth decrease to the minimum at the L2$_1$ structure. However, our
245: fine-scale exploration of the energy surface does reveal that the entire energy surface for
246: 0.95 $<$ c/a $<$ 1.25 is remarkably flat, with the total energy varying less than 2.5
247: meV/atom. This has significant physical consequences for thin films which will be
248: discussed further below.
249:
250: First, we consider the reasons for the absence of a local minimum corresponding to the
251: ground state $\beta^{\prime}$ phase. It has been suggested that the tetragonal distortion
252: could be stabilized by a change in volume\cite{ayuela99}. To investigate this hypothesis,
253: we computed the energy as a function of c/a for varying volumes, with results shown in
254: Figure \ref{Fig.ni2mnga}. We sampled the volumes in 1.5\% increments of $\Delta V/V$. Even
255: with the possibility of volume change, the L2$_1$ structure remains the total energy
256: minimum, with no local minimum corresponding to the $\beta^\prime$ structure. We remark
257: that the minimum near c/a = 1.19 is lowered relative to the cubic structure as V increases,
258: although it never actually becomes more favorable. We have also calculated the total energy
259: of a pure orthorhombic structure at the experimental lattice constants, which give c/a=1.11
260: and b/a=1.04.\cite{kokorin93} The calculated energy of this orthorhombic phase is 4 meV per
261: atom higher than the L2$_1$ phase, consistent with previous calculations.\cite{ayuela99}
262: We conclude that
263: the observed shuffle distortions are crucial to stabilizing the observed $\beta^{\prime}$
264: and $\beta^{\prime\prime}$ phases.
265:
266:
267: A number of related materials were examined for comparison. Martensitic transitions have
268: been observed in the Ni-Mn-Al system, but not in stoichiometric Ni$_2$MnAl
269: \cite{gejima99,sutou98}. Similarly, there are no reports of low-symmetry phases of
270: Ni$_2$MnIn. NiMn exhibits a martensitic transition from a cubic paramagnetic phase at T$_m$
271: = 943 K to a tetragonal phase with c/a = 1.33 which, however, is {\it anti}ferromagnetic
272: \cite{kasper59,egorushkin83}.
273:
274: \begin{figure}
275: \centerline{\epsfig{file=02.ps,width=3.3in}}
276: \vspace{0.4cm}
277: \caption{Total energy as a function of c/a ratio of several related compounds, compared
278: with Ni$_2$MnGa. The curve for each compound is computed at a constant volume equal to the theoretical V$_{Heusler}$ given in Table II. Each curve represents an interpolation through at least 20 points in the range shown. The energies are given relative to the energy at c/a = 1, with the curves slightly offset for clarity.
279: }
280: \label{Fig.related}
281: \end{figure}
282:
283: For Ni$_2$MnAl, Ni$_2$MnIn, and ferromagnetic NiMn, the energy as a function of c/a for V =
284: V$_{Heusler}$ is shown in Figure \ref{Fig.related}. All three compounds are seen to have a
285: single well-defined minimum at c/a near 1, consistent with available experimental
286: information. In the previous calculations of Ayuela $et.al.$, there was another metastable
287: minimum for Ni$_2$MnAl at c/a=1.22. We find that for Ni$_2$MnAl the shape of this curve is
288: very sensitive to the k-point sampling density, with a local minimum near c/a=1.2 evolving
289: into a slope ``discontinuity" for grids denser than 8x8x8.
290:
291: \begin{figure}
292: \centerline{\epsfig{file=03.ps,width=3.3in}}
293: \vspace{0.4cm}
294: \caption{Total energy of NiMn (FM, AF-I, AF-II phases) as a function of c/a at constant volume.
295: For FM, AF-I and AF-II phases, we chose the volume of the observed experimental
296: structure
297: \cite{kasper59,egorushkin83}, a=3.74 \AA~ and c=3.52 \AA.
298: }
299: \label{Fig.nimn}
300: \end{figure}
301:
302: When constrained to be ferromagnetic, NiMn has a tetragonal distortion energy curve similar
303: in shape to that of Ni$_2$MnIn and Ni$_2$MnAl, though with a larger stiffness. The
304: substitution of a magnetic atom in the X site has, however, significant consequences for
305: the magnetic order, stabilizing a type-I antiferromagnetically ordered structure (AF-I) by
306: 190 meV over the lowest-energy ferromagnetic structure (FM). This is in good agreement with
307: the value of 200 meV obtained in a previous calculation\cite{sakuma98}, performed with
308: lattice constants fixed at experimental values. As shown in Figure \ref{Fig.nimn}, total
309: energy minimization for the AF-I NiMn structure leads to a predicted value of c/a = 1.33,
310: consistent with experiment.\cite{kasper59,egorushkin83} At this value a pseudogap appears
311: at the Fermi level, as shown in Figure \ref{Fig.NiMnDos}. As the tetragonal distortion
312: deviates from c/a = 1.33 the pseudogap becomes less pronounced and finally Ni d-states fill
313: the gap. The formation of a pseudogap and the associated enhanced stability does not
314: appear for a distinct antiferromagnetic ordering (AF-II), which is higher in energy by 70
315: meV and has a much shallower minimum at c/a = 1.24.
316:
317: \begin{figure}
318: \centerline{\epsfig{file=04.ps,width=3.3in}}
319: \vspace{0.4cm}
320: \caption{Density of states for (a) the AF-I and (b) the AF-II phases of NiMn for varying c/a at the same volume as in Figure \ref{Fig.nimn}.
321: The solid line corresponds
322: to the minimum energy states for both AF-I and AF-II phases.
323: }
324: \label{Fig.NiMnDos}
325: \end{figure}
326:
327: \begin{figure}
328: \centerline{\epsfig{file=05.ps,width=3.3in}}
329: \vspace{0.4cm}
330: \caption{Density of states in Ni$_2$MnX compounds (X = Al, Ga, In).
331: The atomic character (obtained by projecting the wave functions onto
332: individual atoms)
333: is indicated by appropriate labels.}
334: \label{Fig.dos}
335: \end{figure}
336:
337:
338: The densities of states for cubic Ni$_2$MnX (X = Ga, Al and In) are shown in Figure
339: \ref{Fig.dos}. For all three compounds, these are quite similar, especially near the Fermi
340: level. The total magnetic moment of Ni$_2$MnGa L2$_1$ structure is 4.228 $\mu_B$, which
341: decomposes into 0.301/3.700/-0.075 $\mu_B$ on Ni/Mn/Ga atoms. This agrees well with the
342: previous calculations of References \onlinecite{ayuela99} and \onlinecite{velikohatnyi99}.
343:
344: \begin{figure}
345: \centerline{\epsfig{file=06.ps,width=3.3in}}
346: \vspace{0.4cm}
347: \caption{Total magnetic moments of Ni$_2$MnX (X = Al, Ga, In) and FM-NiMn as a function of c/a, corresponding to the total energy calculations in Figure \ref{Fig.related}.
348: }
349: \label{Fig.magmom}
350: \end{figure}
351: The magnetic moment as a function of c/a, shown in Figure \ref{Fig.magmom}, is again
352: similar for all three compounds. It shows a sharp minimum at c/a near 1 and and a local
353: maximum at c/a $\approx$ 0.94. From the atomic decomposition of this curve for Ni$_2$MnGa
354: in Figure \ref{Fig.decomp}, we see that the features can be directly attributed to the Ni
355: contribution, most likely associated with a small Ni-derived peak in the minority density
356: of states located just below the Fermi level. It is interesting to note that the position
357: of this maximum coincides with the value of c/a for the $\beta^\prime$ martensitic
358: structure.
359:
360: \begin{figure}
361: \centerline{\epsfig{file=07.ps,width=3.3in}}
362: \vspace{0.4cm}
363: \caption{Total magnetic moment and contributions associated with individual elements
364: atoms in Ni$_2$MnGa as a function of c/a for V = V$_{Heusler}$. Note the
365: difference in the vertical scale between the first plot and subsequent plots.
366: }
367: \label{Fig.decomp}
368: \end{figure}
369:
370: The softness of the tetragonal distortion in Ni$_2$MnGa has significant implications for
371: the growth and properties of thin films. Efforts to grow single-crystal thin films on a
372: GaAs substrate (with 3.1\% lattice mismatch) have been carried out by MBE
373: \cite{dong99,dong00}. Using a Sc$_{0.3}$Er$_{0.7}$As interlayer, a tetragonal-structure
374: film with lattice parameters a = 0.565 nm and c = 0.612 nm was obtained. This is exactly
375: the structure we expect based on the total energy curves in Fig. \ref{Fig.ni2mnga}. As the
376: in-plane a is reduced to 0.565 nm to match the substrate, the c axis increases to maintain
377: the volume of the cubic Heusler phase, which minimizes the total energy over the whole
378: range of c/a considered. The energy of the resulting structure, at c/a = 1.08, is less than
379: 2 meV/atom higher than the cubic structure, explaining the observation that the thickness
380: of this pseudomorphic film, 300 \AA, far exceeds the critical thickness expected for the
381: given lattice mismatch. While the large epitaxial stress should affect the structure and
382: properties significantly, if the film can be released it should exhibit a martensitic
383: transition and shape-memory properties.
384:
385: In another recent experiment\cite{dong00} a 450 \AA~thick film of Ni$_2$MnGa was grown by
386: MBE on a GaAs substrate using a NiGa interlayer. The structure of this film was reported to
387: be tetragonal, with c/a = 1.05 (a=5.79 \AA~and c=6.07 \AA). In this case, the lattice
388: constant of Ni$_2$MnGa is a close (0.6\%) match to the interlayer, leading to the
389: expectation that the structure of the film should be cubic. Instead, the c parameter is
390: expanded, resulting in a volume 3\% larger than the volume of the L2$_1$ structure. One
391: explanation might be a slight change in the stoichiometry of the film.
392: Modelling the effect of this shift as an negative applied hydrostatic pressure, we now
393: focus on the $\Delta V/V$ = +3\% curve in \ref{Fig.ni2mnga}. The flatness of this curve is
394: similar to that for V = V$_{Heusler}$. Thus, although the expanded system is formally
395: mismatched to the NiGa interlayer, the constraint a = 5.79 \AA~is associated with an energy
396: cost of less than 2 meV/atom, which would allow pseudomorphic growth of this tetragonal
397: film far beyond the critical thickness.
398:
399: This study highlights the fact that the energy surface of Ni$_2$MnGa is far from simple.
400: Studies of additional instabilities of the cubic Heusler structure via density-functional
401: perturbation theory are in progress. In particular, the energies of shuffle distortions as a
402: function of wavevector can be obtained with this method for a better understanding of the
403: observed $\beta^\prime$ ground state structure. This information will also allow us to model the
404: properties of the high-temperature cubic Heusler phase within an effective Hamiltonian approach.
405: We expect that these properties reflect the presence of large local distortions around the
406: average cubic structure and there should be interesting differences from the properties of other
407: Heusler phases whose ground state structure is cubic.
408:
409: In conclusion, we performed first principles calculations of the total energy as a function of
410: tetragonal distortion (c/a) and volume V for Ni$_2$MnGa and related compounds Ni$_2$MnAl,
411: Ni$_2$MnIn, and NiMn. The total energy of Ni$_2$MnGa at constant volume is remarkably flat in the
412: range 0.95$<$c/a$<$1.25, varying less than 2.5 meV/atom. This provides an explanation of the
413: surprisingly high compliance of single-crystal films grown by MBE. In contrast, the related
414: materials Ni$_2$MnAl, Ni$_2$MnIn, and ferromagnetic NiMn have a single well-defined minimum near
415: c/a = 1. The densities of states and c/a dependence of the magnetic moments are quite similar in
416: all compounds studied, leading us to conclude that the unique behavior of Ni$_2$MnGa is the
417: result of fine tuning. It may be possible to achieve this in other systems through appropriate
418: alloying.
419:
420: We thank R. D. James, K. Bhattacharya, C.J. Palmstr$\o$m, J. Dong, C. Bungaro, X. Huang
421: and A. Ayuela for valuable discussions.
422: This work was supported by AFOSR/MURI F49620-98-1-0433.
423:
424:
425: \begin{references}
426:
427: \bibitem {james00} R. James and K. Hane, Acta Materialia {\bf 48}, 197
428: (2000).
429:
430: \bibitem {james93} R. James and D. Kinderlehrer, Phil. Mag. B {\bf 68},
431: 237 (1993).
432:
433: \bibitem {james94} R. James and D. Kinderlehrer, J. Appl. Phys. {\bf 76},
434: 7012 (1994).
435:
436: \bibitem {ullakko96} K. Ullakko, J. Huang, C. Kanter, R. O'Handley,
437: and V. Kokorin, Appl. Phys. Lett. {\bf 69}, 1966 (1996).
438:
439: \bibitem {ayuela99} A. Ayuela, J. Enkovaara, K. Ullako and
440: R. Niemnen, J. Phys: Cond. Matt. {\bf 11}, 2017 (1999).
441:
442: \bibitem {webster83} P. Webster, K. Ziebeck, S. Tows, and M. Peak,
443: Phil. Mag. {\bf 49}, 295 (1984).
444:
445: \bibitem {kokorin93} V. V. Kokorin, V. V. Martynov, and V. A. Chernenko,
446: Scr. Metll. Mater. {\bf 26}, 175 (1992).
447:
448: \bibitem {zheludev94} A. Zheludev, S. Shapiro, P. Wochner, A. Schwartz,
449: M. Wall, and L. Taner, Phys. Rev. B {\bf 51}, 11310 (1994).
450:
451: \bibitem {zheludev95} A. Zheludev, S. Shapiro, A. Vasil'ev,
452: S. Konoplyuk, and E. Khapalyuk, Phys. Solid State {\bf 37}, 2049 (1995).
453:
454: \bibitem{velikohatnyi99} O. Velikohatnyi and I. Naumov, Phys. Solid
455: State {\bf 41}, 684 (1999).
456:
457: \bibitem{kasper59} J.S. Kasper and J.S. Kouvel, J. Phys. Chem. Solids {\bf 11},
458: 231 (1959).
459:
460: \bibitem {chelikowsky92} J. Chelikowsky and M. L. Cohen in {\it ``Handbook
461: on Semiconductors", vol. 1}, ed. P. T. Landsberg (1992).
462:
463: \bibitem {troullier91} N. Troullier and J.L. Martins, Phys. Rev. B {\bf
464: 43}, 1993 (1991).
465:
466: \bibitem {perdew92} J. Perdew and Y. Wang, Phys. Rev. B {\bf 45}, 13244
467: (1992).
468:
469: \bibitem {louie82} S. Louie, S. Froyen, and M. L. Cohen, Phys. Rev. B {\bf
470: 26}, 1738 (1982).
471:
472: \bibitem {kleinman82} L. Kleinman and D. Bylander, Phys. Rev. Lett. {\bf
473: 48}, 1425 (1982).
474:
475: \bibitem {monkhorst76} H. Monkhorst and J. Pack, Phys. Rev. B {\bf 13},
476: 5188 (1976).
477:
478: \bibitem {sakuma98} A. Sakuma, J of Magn. Magn. Mat. {\bf 187},
479: 105 (1997).
480:
481: \bibitem {egorushkin83} V. Egorushkin, S. Kulkov, and S. Kulkova,
482: Physica B und C {\bf 123B}, 61 (1983).
483:
484: \bibitem {worgull96} J. Worgull, E. Petti and J. Trivisonno, Phys. Rev. B
485: {\bf 54}, 15695 (1996).
486:
487: \bibitem {vintaykin92} E. Vintaykin, P. Potapov, and N. Poliakova,
488: Proc. Conf. MARTENSIT-91, Kiev, Ukraina, 386 (1992).
489:
490: \bibitem {yang92} W. Yang and D. Mikkola, MRS Symp. Proc. {\bf 246},
491: 135 (1992).
492:
493: \bibitem {potapov96} P. Potapov and V. Udovenko, J. de Phys. {\bf 5},
494: C8-1059 (1996).
495:
496: \bibitem {james98} R. James and M. Wuttig, Phil. Mag. A {\bf 77},
497: 1273 (1998).
498:
499: \bibitem {kren68} E. Kren and L. Pal, J. Phys. Chem. Solids {\bf 26},
500: 101 (1968).
501:
502: \bibitem {dong99} J. Dong, L. Chen, C. Palmstr$\o$m, R. James,
503: and S. McKernan, Appl. Phys. Lett. {\bf 75}, 1443 (1999).
504:
505: \bibitem {dong00} J. Dong, L. Chen, S. McKernan, J. Xie, M. Figus,
506: R. James, and C. Palmstr$\o$m, to be published in the Proceedings
507: of 1999 MRS Fall Conference, Smart
508: Materials session.
509:
510: \bibitem {palmstrom00} C. J. Palmstrom, private communication.
511:
512: \bibitem {sutou98} Y. Sutou, I. Ohnuma, R. Kainuma, and K. Ishida,
513: Metall. Mater. Trans. A {\bf 29}, 2225 (1998).
514:
515: \bibitem {gejima99} F. Gejima, Y. Sutou, R. Kainuma, and K. Ishida,
516: Metall. Mater. Trans. A {\bf 30}, 2721 (1999).
517:
518: \bibitem {kilian00} K. A. Kilian and R. H. Victora, J.
519: Appl. Phys. {\bf 87}, 7064 (2000).
520:
521: \bibitem{daSilva88} E. Z. da Silva, O. Jepsen and O. K. Andersen, Solid State Comm. {\bf
522: 67}, 13
523: (1988).
524:
525: \end{references}
526:
527:
528: \end{document}
529: