1: \documentclass[showkeys,showpacs,prb,aps,twocolumn,superbib]{revtex4}
2: %\documentclass[prb,preprint]{revtex4}
3: \usepackage{graphicx}
4: \usepackage{color}
5: %\usepackage{epstopdf}
6: \makeatletter
7: \makeatother
8:
9: \begin{document}
10: \title{
11: First-principles study of wurtzite InN$(0001)$ and $(000{\overline 1})$ surfaces
12: }
13: \author{Chee Kwan Gan$^1$ and David~J.~Srolovitz$^2$}
14: \affiliation{
15: $^1$Institute of High Performance Computing, 1 Science Park Road,
16: Singapore 117528, Singapore\\
17: $^2$Department of Physics,
18: Yeshiva University, New York, NY 10033, USA
19: }
20:
21: \date{May 19, 2006}
22:
23: \begin{abstract}
24: Density-functional calculations are used to study various plausible
25: structures of the wurtzite InN $(0001)$ and $(000{\overline 1})$
26: surfaces.
27: These structures include the unreconstructed surfaces,
28: surfaces with monolayers of In or N, several possible coverages and
29: locations of In or N adatoms and vacancies.
30: The stable structure of the $(0001)$ surface under N-rich conditions
31: is the unreconstructed, In-terminated, surface,
32: while under In-rich conditions the stable surface has a $3/4$ monolayer
33: of In atoms.
34: The stable structure of the InN $(000{\overline 1})$ surface
35: corresponds to a full monolayer of In atoms in the atop sites (directly
36: above the N atoms) over the entire range of accessible In (or N)
37: chemical potential.
38: The atomic structures of the low-energy
39: structures are also discussed.
40: \pacs{68.47.Fg}
41: \keywords{InN surfaces, surface reconstruction, group-III nitrides}
42: \end{abstract}
43: \maketitle
44:
45:
46: \section{Introduction}
47: Group-III nitrides are of considerable interest for applications in
48: laser and light-emitting
49: diodes\cite{Nakamura96v35,Nakamura97,Taniyasu06v441}. Alloying
50: different group-III elements in nitride compounds (e.g., Al$_x$Ga$_{1-x}$N or
51: In$_x$Ga$_{1-x}$N) provides a means of tuning
52: the band gap over a wide spectral range
53: (e.g., from blue to infrared in the case of In$_x$Ga$_{1-x}$N).
54: %The importance of these
55: %materials have led to many theoretical investigations of the
56: %bulk pure compounds\cite{Wright95v51,Kim96v53,Wright97v82} and
57: %their alloys\cite{Ho96v69,Teles00v62,Purton05v15,Gan06_PRB}. The
58: %effect of defects has also been studied for the pure compounds
59: %\cite{Gorczyca99v60,Stampfl00v61,VandeWalle04v95} and their
60: %alloys\cite{Ramos02v14}.
61: The growth of high quality group-III nitride compounds can be a considerable
62: challenge and a barrier to the successful transition of these materials into
63: applications. The connection between vapor phase growth
64: processes and film quality can usually be traced back to the interplay of growth
65: conditions and surface structure. While there have been several
66: theoretical studies on the surfaces of group-III nitrides, these have largely
67: focused upon AlN and
68: GaN\cite{Northrup96v53,Northrup97v55,Smith97v79,Rapcewicz97v56,%
69: Fritsch98v57,Wang01v64,Lee03v68,Timon05v72}. Relatively little
70: theoretical analysis has been devoted to the surfaces of InN, largely
71: because the level of interest in this material has been limited by
72: difficulty in producing this material.
73:
74: InN is an attractive material for electronic and opto-electronic applications
75: because it possesses a series of outstanding
76: properties\cite{Bhuiyan03v94}, such as a small effective mass\cite{Mohammad96v20} (leading to a large electron
77: mobility and a high carrier saturation velocity). The steady-state
78: drift velocity in InN is $4.2\times 10^7$ cm/s is significantly larger than
79: that of many III-V competitors, GaN ($2.9 \times 10^7$~cm/s), AlN
80: ($1.7\times 10^7$~cm/s), and GaAs
81: ($1.6\times 10^7$~cm/s)\cite{Foutz99v85,Bhuiyan03v94}.
82: Since InN exhibits a high peak overshoot velocity\cite{Foutz99v85}, it is a good
83: candidate for high-performance heterojunction field-effect
84: transistors.
85:
86: The growth of high quality InN is difficult compared to other
87: III-nitrides, such as AlN and GaN, because it has a relatively
88: low dissociation temperature and a high equilibrium
89: N$_2$ vapor pressure\cite{MacChesney70v5,Ambacher96v14}.
90: The lack of suitable
91: substrate materials also hinders the use of InN as a heteroepitaxial thin film.
92: Nonetheless, several groups\cite{Davydov02v229,Wu02v80,Nanishi03v42} have
93: successfully grown good quality InN single crystals. Interestingly, they have
94: reported bandgaps of approximately 0.9~eV, which is quite small
95: compared to the commonly accepted values of 1.8~eV to 2.1~eV.
96:
97: Since InN does not possess a center of symmetry,
98: the (0001) and $(000{\overline 1})$ surfaces are different. The first is
99: In-terminated, while the latter is N-terminated. Experimental evidence
100: suggests that polarity (termination) is an important consideration in the growth of
101: high quality group-III nitride semiconductors\cite{Saito01v228}. For example,
102: surface with opposite polarity exhibit different morphologies\cite{Daudin96v69}.
103: Radio frequency molecular
104: beam epitaxial (rf-MBE) growth of InN on sapphire\cite{Saito01v228} produces
105: different polarity surfaces depending on growth temperature (N-terminated at
106: low temperature, In-terminated at high temperature). This demonstrates that growth
107: conditions can modify the relative growth rates of surfaces with opposite polarity.
108: The observed change in the relative growth rates of surfaces with opposite polarity
109: with temperature
110: may be associated with different activation energies for growth or may be
111: indicative a change in surface structure with growth conditions.
112: In this work, we determine the equilibrium structure of the important
113: (0001) and $(000{\overline 1})$ surfaces of InN as a function of growth conditions.
114: In particular, we examine the role of In (or N)
115: chemical potential on the equilibrium surface structure. We vary chemical potential
116: rather than temperature, since the range of chemical potentials that are
117: experimentally accessible is much larger than that for temperature in
118: the most common group-III nitride growth methods (i.e., MOCVD).
119:
120: \section{Theoretical models}
121: The present calculations are performed within the local
122: density approximation (LDA) framework, as implemented within the Vienna Simulation
123: Package (VASP)\cite{Kresse96v54}, employing ultrasoft
124: pseudopotentials\cite{Vanderbilt90v41}. The
125: important\cite{Wright95v51} 4d states for In are included as valence
126: electrons in the pseudopotentials. Throughout this work, we employed
127: a relatively high cutoff energy of 435~eV. We first optimize the bulk
128: wurtzite InN structure, yielding the following lattice parameters\cite{Gan06v73}:
129: $a=3.518$~\AA, $c = 5.690$~\AA, and
130: the internal parameter $u=0.379$. These lattice parameters are in
131: very good agreement with those found from
132: experiment\cite{Davydov02v229}; $a = 3.5365$~\AA\ and $c=
133: 5.7039$~\AA.
134:
135: For the surface calculations, our supercells contained six (0001) bilayers
136: in which the lower four bilayers were fixed in the bulk
137: configuration. The upper two bilayers and any adatoms or adlayers
138: were allowed to relax. Our system is larger than those used by
139: many others in studies of other group-III nitride surfaces, where typically four
140: bilayers\cite{Smith97v79,Rapcewicz97v56, Wang01v64,Lee03v68} were
141: employed. To prevent unphysical charge transfer between the top and bottom
142: slab surfaces, pseudo-hydrogens with fractional charges are used. For the
143: In-terminated (0001) surface, the dangling bonds in the N layer are
144: saturated with pseudo-hydrogens with a fractional charge of 3/4. For
145: the N-terminated $(000{\overline 1})$ surface, we saturate the
146: dangling bonds in the In layer with pseudo-hydrogens with a fractional
147: charge of 5/4. We employ a $2\times 2 \times 1$ Monkhost-Pack set
148: sampling scheme for the surface supercell, which gives two irreducible
149: $k$-points. The structural optimizations were terminated when the
150: magnitude of the Hellmann-Feynman force on each ion is less than
151: 1~mRy/Bohr (25.7~meV/\AA). We perform calculations with vacuum
152: thicknesses of both 12.9~\AA\ and 15.7~\AA. These two sets of calculations
153: yield almost identical
154: results, showing the appropriateness of the vacuum thickness employed.
155: Hereafter we present the results obtained with a vacuum thickness of
156: 15.7~\AA.
157:
158:
159: \begin{figure}
160: \resizebox*{3.0in}{!}{\includegraphics[clip]{Gan_Fig1.eps}}
161: \caption{(Color online)
162: (a) A schematic of the atom positions in the top two atomic planes
163: of a $2\times 2$ unit cell of the $\{0001\}$ surface. The
164: larger and smaller filled circles denote indium atoms on the surface
165: and nitrogen atoms one plane down, respectively, for the In-terminated
166: (0001) surface. The same schematic describes the $(000{\overline 1})$
167: surface, but with the elemental identity of the atoms switched.
168: The locations of the H3 (hollow) and T4 (top)
169: sites are labeled by dotted circles.
170: {\bf (b)
171: A view of an In trimer residing
172: on the T4 sites from above.
173: The gray and black atoms denote the In and N atoms, respectively. Notice that
174: a nitrogen atom is exposed in the lower right corner of the image.
175: (c) The side view of the $2\times 2$
176: unit cell. Four pseudo-hydrogen atoms (shown in white) are found at the
177: bottom of the cell.
178: }
179: }
180: \label{fig:h3t4}
181: \end{figure}
182: \begin{figure}
183: \resizebox*{3.0in}{!}{\includegraphics[clip]{Gan_Fig2.eps}}
184: \caption{
185: The relative formation energies for the $2 \times 2$
186: unit cell of Fig.~\ref{fig:h3t4}(a) as a function of the In chemical potential
187: for the (a) (0001) and (b) $(000{\overline 1})$ surfaces.
188: The formation energies
189: are all referenced to that of the surface with a single vacancy on the
190: surface atomic plane.
191: }
192: \label{fig:1-bar1}
193: \end{figure}
194:
195: We consider various vacancy, adatom, trimer, and
196: monolayer surface structures that satisfy the electron
197: counting rule\cite{Northrup97v55}. It is useful to refer to Fig.~\ref{fig:h3t4}(a),
198: which shows two (0001) layers of atoms in a $2 \times 2$ surface unit cell,
199: in order to understand the structures.
200: The surfaces considered include a
201: perfect surface [as shown by the black circles
202: in Fig.~\ref{fig:h3t4}(a)], a monolayer of In (or N)
203: atoms directly above each In (or N) terminating atom [i.e., this is 4 atop
204: atoms in the $2 \times 2$ cell of Fig.~\ref{fig:h3t4}(a)], a quarter-monolayer of
205: adatoms with the adatoms directly atop one of the In (or N) atoms (T4 sites)
206: in the $2 \times 2$ cell, a quarter-monolayer of
207: adatoms with the adatoms centered between three surface atoms
208: (H3 sites), a single vacancy in the terminating plane of the
209: $2 \times 2$ cell, and a triangle of three In (or N) atoms on H3 or
210: T4 sites (i.e., a trimer). The relaxed structure of a representative case of
211: an In trimer on the T4 sites is shown in Figs.~\ref{fig:h3t4}(b) and \ref{fig:h3t4}(c).
212:
213: The relative formation energies of the different surfaces are calculated
214: within the thermodynamically allowed range of the In
215: chemical potential $\mu_{\rm In}$.
216: The chemical potentials of In and
217: N are related through\cite{Qian88v38}
218: \begin{equation}
219: \mu_{\rm In} + \mu_{\rm N} = \mu_{\rm InN(bulk)} .
220: \label{eq:1}
221: \end{equation}
222: Operationally, we can define the chemical potential
223: $\mu_{\rm InN(bulk)}$ as the total energy per
224: cation-anion pair $E({\rm InN}_{\rm bulk})$ in wurtzite InN at 0~K. The
225: chemical potential of In is bounded above by that of the pure, tetragonal
226: In metal\cite{Graham56v84}, denoted by $\mu_{\rm In}^0 = E({\rm In})$.
227: Similarly, the chemical potential of N is
228: bounded above by that of a nitrogen molecule N$_2$, denoted by $\mu_{\rm N}^0 = \frac{1}{2}E({\rm N}_2)$.
229: The enthalpy of formation of InN is simply
230: \begin{equation}
231: \Delta H = \mu_{\rm InN(bulk)} -
232: \mu_{\rm N}^0 - \mu_{\rm In}^0
233: \label{eq:2}
234: \end{equation}
235: where $E({\rm In})$, $E({\rm N_2})$, and $E({\rm InN_{bulk}})$ are determined
236: from total energy calculations within LDA.
237: The calculated enthalpy of formation of InN, $\Delta H=-0.376$~eV, agrees reasonably well with the
238: experimental value\cite{Harrison89,Kittel96} of $-0.28$~eV. The fact that
239: these calculations suggest that InN is more strongly bound than in experiments
240: is not surprising since the LDA tends to lead to overbinding.
241: The small enthalpy of formation implies
242: that InN has a low thermal stability compared with other group-III
243: nitrides such as GaN and AlN where the experimental formation
244: enthalpies\cite{Harrison89,Kittel96} are $-1.23$~eV (i.e., 4 times
245: larger than for InN) and $-3.21$~eV (i.e., 11 times larger than for InN),
246: respectively.
247:
248:
249: The formation energy of any structure is calculated from
250: \begin{eqnarray}
251: E_{\rm form} &=& E_{\rm opt} - n_{\rm In}\mu_{\rm In} - n_{\rm N}\mu_{\rm N} \nonumber \\
252: &=&E_{\rm opt} - \mu_{\rm In}(n_{\rm In}-n_{\rm N})-\mu_{\rm InN(bulk)}n_{\rm N},
253: \label{eq:3}
254: \end{eqnarray}
255: where $E_{\rm opt}$ is the total energy of the structure resulting
256: from the optimization of the structure with respect to atomic positions
257: and $n_{\rm In}$
258: and $n_{\rm N}$ are the number of In and N atoms in the structure, respectively.
259: We have used Eq.~(\ref{eq:1}) to eliminate
260: $\mu_{\rm N}$ to arrive at the second line of
261: Eq.~(\ref {eq:3}).
262: Equations (\ref{eq:1}) and (\ref{eq:2}) show that
263: the thermodynamically allowed range of the chemical potential of In is
264: $\Delta H + \mu_{\rm In}^0 \le \mu_{\rm In} \le \mu_{\rm In}^0 $.
265:
266: \section{The $(0001)$ Surface}
267: The relative formation energies for the $2 \times 2$ InN $(0001)$
268: unit cell of Fig.~\ref{fig:h3t4}(a) are shown in Fig.~\ref{fig:1-bar1}(a) for
269: the surface structures described above as a function of the In
270: chemical potential. In this figure, we reference the energy
271: to that of the $2 \times 2$ InN $(0001)$ unit cell with a single In vacancy
272: in the surface atomic plane. This simplifies
273: the comparison of energies between the different structures,
274: as well as leads to the cancellation of
275: the contribution to $E_{\rm opt}$ in
276: Eq.~(\ref{eq:3}) from the chemical potential of pseudo-hydrogen atoms.
277: We omit the curves for the structures
278: of energies greater than 0.25 eV.
279: Note that the slopes of the lines in Fig.~\ref{fig:1-bar1} simply follow
280: from the fact that the different structures have
281: different numbers of In
282: and N atoms (see Eq.~(\ref{eq:3})).
283:
284: The surface with an N adatom on the H3 site has a lower energy
285: than that with an N adatom on the T4 site by
286: 0.70~eV/$2\times 2$. This is largely due to a stronger electrostatic
287: repulsion between the T4 nitrogen adatom and the nitrogen atom below.
288: Similar behavior has been observed for the AlN\cite{Northrup97v55} $(0001)$
289: and GaN\cite{Rapcewicz97v56} $(0001)$ surfaces.
290: % energy differences of 1.51~eV/$(2\times 2)$ and 0.9028~eV/$(2\times 2)$,
291: The surface with an In adatom
292: on the T4 site has a lower energy than that with an
293: In adatom on the H3 site since there is an electrostatic attraction
294: between the In adatom on the T4 site and the N atom below. The
295: surface with a full In monolayer with atoms on the atop sites has
296: a relatively high energy compared with the other (0001) surfaces.
297:
298: Under N-rich conditions (low In chemical potential), the bare
299: $1\times 1$ (In terminated) surface [as shown in Fig.~\ref{fig:h3t4}(a)]
300: has the lowest energy.
301: Our calculations also show that surfaces in which the In trimer resides
302: on either the T4 or H3 sites yield essentially the same energy,
303: despite the fact that In atoms are directly above N atoms in
304: the T4 In trimer model. Similar behavior is observed for the
305: AlN$(000{\overline 1})$\cite{Northrup97v55} surface
306: where Al trimers on the T4 or H3 sites have the same energy.
307: This shows that arguments based solely on electrostatic interactions
308: are not sufficient to explain even the qualitative behavior of
309: many of the adatom reconstructed surfaces.
310: Figure~\ref{fig:1-bar1}(a) shows that under In-rich (large
311: In chemical potential) conditions,
312: the stable (0001) surface have In trimers, located on either
313: T4 or H3 sites.
314: The present results for the stable (0001) InN surface are
315: significantly different than the same surface in other group-III nitrides
316: across the entire range of chemical potentials.
317: For AlN$(0001)$ \cite{Northrup97v55} and
318: GaN$(0001)$\cite{Rapcewicz97v56} surfaces, the surface with an N
319: adatom on the H3 site is found to be energetically favorable under
320: N-rich conditions, while the surface structure with a metal atom
321: (i.e., Al or Ga) on the T4 site is energetically favorable under
322: metal-rich conditions.
323:
324: Examination of the unreconstructed InN $(0001)$ surface
325: [as shown in Fig.~\ref{fig:h3t4}(a)] shows that
326: the In-N bilayer spacings for the first and second bilayers are
327: 0.654~\AA\ and 0.655~\AA, respectively. These are slightly smaller
328: than the ideal, bulk spacing of 0.688~\AA. The spacing between
329: In-In atomic planes are 2.862~\AA\ and 2.825~\AA\ for the first and
330: second pairs, as compared with a bulk spacing of 2.845~\AA.
331: This shows that the spacing between the first In-In
332: layer expands, while the second contracts, relative to the bulk spacing.
333: Similar oscillatory surface relaxations are observed in other systems
334: (e.g., see Ref. \onlinecite{Sun03v68}).
335:
336: For the InN $(0001)$ surface with a single In vacancy on the surface layer,
337: we find that the top-layer In atoms undergoes
338: $sp^2$ bonding configuration by relaxing vertically. This reduces the
339: bond length between the threefold-coordinated In and the
340: threefold-coordinated N atoms from the ideal value of 2.145~\AA\ to
341: 2.050~\AA. This 4.4~\% reduction in length is comparable to that for
342: the case of AlN\cite{Northrup97v55}, where a 6\% contraction in the
343: distance between the Al and N atoms in the vacancy model of AlN$(0001)$
344: surface was observed. The vertical movement of the top layer
345: atoms reduces the top In-N bilayer spacing from the ideal value of
346: 0.688~\AA\ to 0.350~\AA.
347:
348: The surface for with an In trimer on the T4 or H3 sites
349: have the same energy and are the most stable surfaces at large
350: In chemical potential, as discussed above.
351: It is interesting to note that the interlayer spacing between the trimer
352: In atoms and the In atoms in the next plane are 2.61~\AA\ for
353: the T4 site trimer and 2.60~\AA\ for the H3 site trimer. Hence,
354: these two surface configurations are very similar both structurally
355: and energetically.
356:
357: \section{The $(000{\overline 1})$ Surface}
358:
359: The relative formation energies for the $2 \times 2$ InN $(000{\overline 1})$
360: unit cell of Fig.~\ref{fig:h3t4}(a) are shown in Fig.~\ref{fig:1-bar1}(b) for
361: the surface structures described above as a function of the In
362: chemical potential. In this figure, we reference the energy
363: to that of the $2 \times 2$ InN $(000{\overline 1})$ unit cell with a
364: single N vacancy in the surface atomic plane.
365: No results are presented for the surfaces with N adatoms or N trimers
366: since these systems are unstable.
367: Northrup {\it et al.}\cite{Northrup97v55} argued that the
368: relatively small size of the N atom makes surfaces with N adatoms or
369: trimers on the AlN$(000{\overline 1})$ highly unstable. It is
370: therefore not surprising that our calculations also show that N trimer
371: or N adatom models are not stable since the single-bond covalent
372: radius\cite{Cotton95} of In (1.48~\AA) is considerable larger
373: than that of Al (1.30~\AA).
374:
375: The relaxed unreconstructed InN $(000{\overline 1})$ surface
376: (N-terminated) has a very large formation energy, relative to most
377: of the other surface structures.
378: We find that the surfaces with an In adatom on the H3 site has an
379: energy that is 1.83~eV/$(2\times 2)$ lower than that with the In
380: adatom on the T4 site.
381: This may be attributed to the fact that there is greater electrostatic
382: repulsion between the T4 In adatom and the In atom directly below it
383: than between the H3 In adatom and the more distant In atoms two
384: atomic planes below.
385: Similar argument also holds for In trimer models, where the H3 In
386: trimer has a lower in energy than the T4 In trimer by
387: 1.23~eV/$(2\times 2)$.
388: The surface with an In monolayer with the atoms on
389: the atop sites has the lowest energy across the entire allowed range of
390: In chemical potential.
391: The next lowest energy structure under N-rich (low In chemical potential)
392: conditions is surface with a single In adatom in the $2\times2$ unit cell
393: on the H3 site.
394: These results contrast with the results for other group-III
395: nitrides $(000{\overline 1})$ surfaces. In AlN and GaN, the lowest
396: energy surface structure corresponds to an Al
397: monolayer (atop sites occupied)\cite{Northrup97v55} or Ga monolayer\cite{Gan06_unpublishedB} only under metal-rich conditions.
398: However, under N-rich conditions, the $(000{\overline 1})$ surfaces
399: with a single Al or Ga atom on the H3 site in the $2\times2$ unit cell have
400: the lowest energy.
401:
402: Examination of the InN $(000{\overline 1})$ surface with an In monolayer
403: with the In atoms on the atop sites is the lowest energy structure over the
404: entire chemical potential range, as mentioned above.
405: The In-N bilayer spacing for the top two bilayers
406: are 0.519~\AA\ and 0.510~\AA, as compared with a bulk spacing of
407: 0.668 \AA. The In-In atomic plane separations are
408: 2.482~\AA, 2.423~\AA, and 2.593~\AA\ for the first three pairs
409: (counted from the surface), as compared with the bulk value of 2.845~\AA.
410: This is a substantial contraction that oscillates as it decays slowly
411: into the bulk.
412:
413: The surface structure corresponding to an In adatom in the H3 site
414: also has a relatively low energy.
415: This adatom is situated 0.808~\AA\ above the triangle of three N
416: atoms in the plane below.
417: The corresponding In-N bond length is 1.980~\AA.
418: The three N atoms that form this triangle (i.e., nearest neighbors
419: of the In H3 adatom) are displaced upward, toward the In adatom
420: by 0.424~\AA\ relative to the other, threefold-coordinated surface
421: N atoms.
422:
423: In the InN$(000{\overline 1})$ surface structure with a single
424: N-vacancy in the $2\times2$ unit cell, the top-layer N atoms
425: undergo substantial vertical relaxation.
426: In this structure, the N-In bilayer spacing is 0.328~\AA\ as compared
427: with the bulk spacing of 0.688~\AA.
428: The bond length between the threefold-coordinated N and In atoms
429: is also reduced from the ideal value of 2.145~\AA\ to 1.898~\AA.
430: This is an 11.5\%\ reduction in bond length, which is much larger than
431: the 6\%\ contraction observed in the N vacancy surface structure
432: of AlN$(000{\overline 1})$\cite{Northrup97v55}.
433: This large contraction may be attributed to the system trying to move
434: toward a configuration that is $sp^2$ bonded.
435:
436:
437: \section{Conclusions}
438: We have performed density-functional calculations to
439: determine the stable structure of both the InN $(0001)$
440: and $(000{\overline 1})$ surfaces as a function of the
441: In (or N) chemical potential. We considered twelve plausible
442: surface structures. Several earlier studies considered
443: the structure of this surface in other group-III nitrides.
444: While structurally similar, InN is quite distinct, as indicated
445: by a much lower heat of formation than the others (BN, AlN,
446: GaN).
447:
448: For the $(0001)$ surface, we find that the relaxed,
449: unreconstructed InN $(0001)$ (In terminated) surface
450: is stable under N-rich (low In chemical potential) conditions.
451: On the other hand, under In rich (large In chemical potential)
452: conditions, the stable structure corresponds to a trimer of
453: In adatoms on either H3 or T4 sites (this is $3/4$ of an In monolayer).
454: These stable $(0001)$ surface structures differ from those
455: found for two other closely related group-III nitrides
456: over the entire range of accessible chemical potentials.
457: The stable structure of the InN $(000{\overline 1})$ surface corresponds
458: to a monolayer of In atoms directly above the surface N atoms.
459: This structure is stable over the entire range of accessible chemical
460: potentials.
461: In AlN and GaN, the same surface structure was only found to be
462: stable under metal-rich conditions.
463:
464: The present results provide guidance for the exploitation of
465: different thin film growth methods/conditions in order to exploit the
466: dependence of equilibrium reconstructions on chemical potential.
467: Unlike in GaN or AlN, the stable structure of the important $(0001)$
468: surface can be manipulated by changes in growth condition.
469:
470: \section{acknowledgments}
471:
472: The authors gratefully acknowledge useful discussions with
473: Y.~Y.~Sun on surface calculations.
474: This work was supported by Visiting Investigator Program, Agency for
475: Science, Technology and Research (A*STAR), Singapore.
476:
477:
478: %\bibliographystyle{apsrev}
479: %\bibliography{phase}
480: \begin{thebibliography}{32}
481: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi
482: \expandafter\ifx\csname bibnamefont\endcsname\relax
483: \def\bibnamefont#1{#1}\fi
484: \expandafter\ifx\csname bibfnamefont\endcsname\relax
485: \def\bibfnamefont#1{#1}\fi
486: \expandafter\ifx\csname url\endcsname\relax
487: \def\url#1{\texttt{#1}}\fi
488: \expandafter\ifx\csname urlprefix\endcsname\relax\def\urlprefix{URL }\fi
489: \providecommand*{\bibinfo}[2]{#2}
490: \providecommand*{\eprint}[1]{#1}
491: \providecommand*{\url}[1]{#1}
492: \begingroup\makeatletter
493: \@temptokena{%
494: \expandafter\ifx\csname citenamefont\endcsname\relax
495: \DeclareRobustCommand\citenamefont{\@firstofone}%
496: \global\let\citenamefont\citenamefont
497: \global\expandafter\let\csname citenamefont \expandafter\endcsname\csname
498: citenamefont \endcsname
499: \fi
500: }\if@filesw\immediate\write\@auxout{\the\@temptokena}\fi
501: \expandafter\endgroup\the\@temptokena
502:
503: \bibitem[{\citenamefont{Nakamura} \emph{et~al.}(1996)\citenamefont{Nakamura,
504: Senoh, Nagahama, Iwasa, Yamada, Matsushita, Kiyoku, and
505: Sugimoto}}]{Nakamura96v35}
506: \bibinfo{author}{\bibfnamefont{S.}~\bibnamefont{Nakamura}},
507: \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Senoh}},
508: \bibinfo{author}{\bibfnamefont{S.}~\bibnamefont{Nagahama}},
509: \bibinfo{author}{\bibfnamefont{N.}~\bibnamefont{Iwasa}},
510: \bibinfo{author}{\bibfnamefont{T.}~\bibnamefont{Yamada}},
511: \bibinfo{author}{\bibfnamefont{T.}~\bibnamefont{Matsushita}},
512: \bibinfo{author}{\bibfnamefont{H.}~\bibnamefont{Kiyoku}}, \bibnamefont{and}
513: \bibinfo{author}{\bibfnamefont{Y.}~\bibnamefont{Sugimoto}},
514: \bibinfo{journal}{Jpn. J. Appl. Phys.} \textbf{\bibinfo{volume}{35}},
515: \bibinfo{pages}{Part II, L74} (\bibinfo{year}{1996}).
516:
517: \bibitem[{\citenamefont{Nakamura and Fasol}(1997)}]{Nakamura97}
518: \bibinfo{author}{\bibfnamefont{S.}~\bibnamefont{Nakamura}} \bibnamefont{and}
519: \bibinfo{author}{\bibfnamefont{G.}~\bibnamefont{Fasol}},
520: \emph{\bibinfo{title}{The Blue Laser Diode: GaN Based Light Emitters and
521: Lasers}} (\bibinfo{publisher}{Springer}, \bibinfo{address}{Berlin,
522: Heidelberg}, \bibinfo{year}{1997}).
523:
524: \bibitem[{\citenamefont{Taniyasu} \emph{et~al.}(2006)\citenamefont{Taniyasu,
525: Kasu, and Makimoto}}]{Taniyasu06v441}
526: \bibinfo{author}{\bibfnamefont{Y.}~\bibnamefont{Taniyasu}},
527: \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Kasu}}, \bibnamefont{and}
528: \bibinfo{author}{\bibfnamefont{T.}~\bibnamefont{Makimoto}},
529: \bibinfo{journal}{Nature} \textbf{\bibinfo{volume}{441}},
530: \bibinfo{pages}{325} (\bibinfo{year}{2006}).
531:
532: \bibitem[{\citenamefont{Northrup and Neugebauer}(1996)}]{Northrup96v53}
533: \bibinfo{author}{\bibfnamefont{J.~E.} \bibnamefont{Northrup}} \bibnamefont{and}
534: \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{Neugebauer}},
535: \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{53}},
536: \bibinfo{pages}{R10477} (\bibinfo{year}{1996}).
537:
538: \bibitem[{\citenamefont{Northrup} \emph{et~al.}(1997)\citenamefont{Northrup,
539: DiFelice, and Neugebauer}}]{Northrup97v55}
540: \bibinfo{author}{\bibfnamefont{J.~E.} \bibnamefont{Northrup}},
541: \bibinfo{author}{\bibfnamefont{R.}~\bibnamefont{DiFelice}}, \bibnamefont{and}
542: \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{Neugebauer}},
543: \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{55}},
544: \bibinfo{pages}{13878} (\bibinfo{year}{1997}).
545:
546: \bibitem[{\citenamefont{Smith} \emph{et~al.}(1997)\citenamefont{Smith,
547: Feenstra, Greve, Neugebauer, and Northrup}}]{Smith97v79}
548: \bibinfo{author}{\bibfnamefont{A.~R.} \bibnamefont{Smith}},
549: \bibinfo{author}{\bibfnamefont{R.~M.} \bibnamefont{Feenstra}},
550: \bibinfo{author}{\bibfnamefont{D.~W.} \bibnamefont{Greve}},
551: \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{Neugebauer}},
552: \bibnamefont{and} \bibinfo{author}{\bibfnamefont{J.~E.}
553: \bibnamefont{Northrup}}, \bibinfo{journal}{Phys. Rev. Lett.}
554: \textbf{\bibinfo{volume}{79}}, \bibinfo{pages}{3934} (\bibinfo{year}{1997}).
555:
556: \bibitem[{\citenamefont{Rapcewicz} \emph{et~al.}(1997)\citenamefont{Rapcewicz,
557: BuongiornoNardelli, and Berholc}}]{Rapcewicz97v56}
558: \bibinfo{author}{\bibfnamefont{K.}~\bibnamefont{Rapcewicz}},
559: \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{BuongiornoNardelli}},
560: \bibnamefont{and} \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{Berholc}},
561: \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{56}},
562: \bibinfo{pages}{R12725} (\bibinfo{year}{1997}).
563:
564: \bibitem[{\citenamefont{Fritsch} \emph{et~al.}(1998)\citenamefont{Fritsch,
565: Sankey, Schmidt, and Page}}]{Fritsch98v57}
566: \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{Fritsch}},
567: \bibinfo{author}{\bibfnamefont{O.~F.} \bibnamefont{Sankey}},
568: \bibinfo{author}{\bibfnamefont{K.~E.} \bibnamefont{Schmidt}},
569: \bibnamefont{and} \bibinfo{author}{\bibfnamefont{J.~B.} \bibnamefont{Page}},
570: \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{57}},
571: \bibinfo{pages}{15360} (\bibinfo{year}{1998}).
572:
573: \bibitem[{\citenamefont{Wang} \emph{et~al.}(2001)\citenamefont{Wang,
574: Kr{\"u}ger, and Pollmann}}]{Wang01v64}
575: \bibinfo{author}{\bibfnamefont{F.~H.} \bibnamefont{Wang}},
576: \bibinfo{author}{\bibfnamefont{P.}~\bibnamefont{Kr{\"u}ger}},
577: \bibnamefont{and} \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{Pollmann}},
578: \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{64}},
579: \bibinfo{pages}{035305} (\bibinfo{year}{2001}).
580:
581: \bibitem[{\citenamefont{Lee} \emph{et~al.}(2003)\citenamefont{Lee, Dong,
582: Feenstra, Northrup, and Neugebauer}}]{Lee03v68}
583: \bibinfo{author}{\bibfnamefont{C.~D.} \bibnamefont{Lee}},
584: \bibinfo{author}{\bibfnamefont{Y.}~\bibnamefont{Dong}},
585: \bibinfo{author}{\bibfnamefont{R.~M.} \bibnamefont{Feenstra}},
586: \bibinfo{author}{\bibfnamefont{J.~E.} \bibnamefont{Northrup}},
587: \bibnamefont{and}
588: \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{Neugebauer}},
589: \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{68}},
590: \bibinfo{pages}{205317} (\bibinfo{year}{2003}).
591:
592: \bibitem[{\citenamefont{Timon} \emph{et~al.}(2005)\citenamefont{Timon, Brand,
593: Clark, Gibson, and Abram}}]{Timon05v72}
594: \bibinfo{author}{\bibfnamefont{V.}~\bibnamefont{Timon}},
595: \bibinfo{author}{\bibfnamefont{S.}~\bibnamefont{Brand}},
596: \bibinfo{author}{\bibfnamefont{S.~J.} \bibnamefont{Clark}},
597: \bibinfo{author}{\bibfnamefont{M.~C.} \bibnamefont{Gibson}},
598: \bibnamefont{and} \bibinfo{author}{\bibfnamefont{R.~A.} \bibnamefont{Abram}},
599: \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{72}},
600: \bibinfo{pages}{035327} (\bibinfo{year}{2005}).
601:
602: \bibitem[{\citenamefont{Bhuiyan} \emph{et~al.}(2003)\citenamefont{Bhuiyan,
603: Hashimoto, and Yamamoto}}]{Bhuiyan03v94}
604: \bibinfo{author}{\bibfnamefont{A.~G.} \bibnamefont{Bhuiyan}},
605: \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Hashimoto}},
606: \bibnamefont{and} \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Yamamoto}},
607: \bibinfo{journal}{J. Appl. Phys.} \textbf{\bibinfo{volume}{94}},
608: \bibinfo{pages}{2779} (\bibinfo{year}{2003}).
609:
610: \bibitem[{\citenamefont{Mohammad and Morkoc}(1996)}]{Mohammad96v20}
611: \bibinfo{author}{\bibfnamefont{S.~N.} \bibnamefont{Mohammad}} \bibnamefont{and}
612: \bibinfo{author}{\bibfnamefont{H.}~\bibnamefont{Morkoc}},
613: \bibinfo{journal}{Prog. Quant. Electr.} \textbf{\bibinfo{volume}{20}},
614: \bibinfo{pages}{361} (\bibinfo{year}{1996}).
615:
616: \bibitem[{\citenamefont{Foutz} \emph{et~al.}(1999)\citenamefont{Foutz, O'Leary,
617: Shur, and Eastman}}]{Foutz99v85}
618: \bibinfo{author}{\bibfnamefont{B.~E.} \bibnamefont{Foutz}},
619: \bibinfo{author}{\bibfnamefont{S.~K.} \bibnamefont{O'Leary}},
620: \bibinfo{author}{\bibfnamefont{M.~S.} \bibnamefont{Shur}}, \bibnamefont{and}
621: \bibinfo{author}{\bibfnamefont{L.~F.} \bibnamefont{Eastman}},
622: \bibinfo{journal}{J. Appl. Phys.} \textbf{\bibinfo{volume}{85}},
623: \bibinfo{pages}{7727} (\bibinfo{year}{1999}).
624:
625: \bibitem[{\citenamefont{MacChesney}
626: \emph{et~al.}(1970)\citenamefont{MacChesney, Bridenbaugh, and
627: O'Connor}}]{MacChesney70v5}
628: \bibinfo{author}{\bibfnamefont{J.~B.} \bibnamefont{MacChesney}},
629: \bibinfo{author}{\bibfnamefont{P.~M.} \bibnamefont{Bridenbaugh}},
630: \bibnamefont{and} \bibinfo{author}{\bibfnamefont{P.~B.}
631: \bibnamefont{O'Connor}}, \bibinfo{journal}{Mater. Res. Bull.}
632: \textbf{\bibinfo{volume}{5}}, \bibinfo{pages}{783} (\bibinfo{year}{1970}).
633:
634: \bibitem[{\citenamefont{Ambacher} \emph{et~al.}(1996)\citenamefont{Ambacher,
635: Brandt, Dimitrov, Metzger, Stutzmann, Fischer, Miehr, Bergmaier, and
636: Dollinger}}]{Ambacher96v14}
637: \bibinfo{author}{\bibfnamefont{O.}~\bibnamefont{Ambacher}},
638: \bibinfo{author}{\bibfnamefont{M.~S.} \bibnamefont{Brandt}},
639: \bibinfo{author}{\bibfnamefont{R.}~\bibnamefont{Dimitrov}},
640: \bibinfo{author}{\bibfnamefont{T.}~\bibnamefont{Metzger}},
641: \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Stutzmann}},
642: \bibinfo{author}{\bibfnamefont{R.~A.} \bibnamefont{Fischer}},
643: \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Miehr}},
644: \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Bergmaier}},
645: \bibnamefont{and}
646: \bibinfo{author}{\bibfnamefont{G.}~\bibnamefont{Dollinger}},
647: \bibinfo{journal}{J. Vac. Sci. Technol. B} \textbf{\bibinfo{volume}{14}},
648: \bibinfo{pages}{3532} (\bibinfo{year}{1996}).
649:
650: \bibitem[{\citenamefont{Davydov} \emph{et~al.}(2002)\citenamefont{Davydov,
651: Klochikhin, Seisyan, Emptsev, Ivanov, Bechstedt, Furthm{\"u}ller, Harima,
652: Mudryi, Aderhold, Semchinova, and Graul}}]{Davydov02v229}
653: \bibinfo{author}{\bibfnamefont{V.~Y.} \bibnamefont{Davydov}},
654: \bibinfo{author}{\bibfnamefont{A.~A.} \bibnamefont{Klochikhin}},
655: \bibinfo{author}{\bibfnamefont{R.~P.} \bibnamefont{Seisyan}},
656: \bibinfo{author}{\bibfnamefont{V.~V.} \bibnamefont{Emptsev}},
657: \bibinfo{author}{\bibfnamefont{S.~V.} \bibnamefont{Ivanov}},
658: \bibinfo{author}{\bibfnamefont{F.}~\bibnamefont{Bechstedt}},
659: \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{Furthm{\"u}ller}},
660: \bibinfo{author}{\bibfnamefont{H.}~\bibnamefont{Harima}},
661: \bibinfo{author}{\bibfnamefont{A.~V.} \bibnamefont{Mudryi}},
662: \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{Aderhold}},
663: \bibinfo{author}{\bibfnamefont{O.}~\bibnamefont{Semchinova}},
664: \bibnamefont{and} \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{Graul}},
665: \bibinfo{journal}{Phys. Stat. Sol. (B)} \textbf{\bibinfo{volume}{229}},
666: \bibinfo{pages}{R1} (\bibinfo{year}{2002}).
667:
668: \bibitem[{\citenamefont{Wu} \emph{et~al.}(2002)\citenamefont{Wu, Walukiewicz,
669: Yu, III, Haller, Lu, Schaff, Saito, and Nanishi}}]{Wu02v80}
670: \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{Wu}},
671: \bibinfo{author}{\bibfnamefont{W.}~\bibnamefont{Walukiewicz}},
672: \bibinfo{author}{\bibfnamefont{K.~M.} \bibnamefont{Yu}},
673: \bibinfo{author}{\bibfnamefont{J.~W.~A.} \bibnamefont{III}},
674: \bibinfo{author}{\bibfnamefont{E.~E.} \bibnamefont{Haller}},
675: \bibinfo{author}{\bibfnamefont{H.}~\bibnamefont{Lu}},
676: \bibinfo{author}{\bibfnamefont{W.~J.} \bibnamefont{Schaff}},
677: \bibinfo{author}{\bibfnamefont{Y.}~\bibnamefont{Saito}}, \bibnamefont{and}
678: \bibinfo{author}{\bibfnamefont{Y.}~\bibnamefont{Nanishi}},
679: \bibinfo{journal}{Appl. Phys. Lett.} \textbf{\bibinfo{volume}{80}},
680: \bibinfo{pages}{3967} (\bibinfo{year}{2002}).
681:
682: \bibitem[{\citenamefont{Nanishi} \emph{et~al.}(2003)\citenamefont{Nanishi,
683: Saito, and Yamaguchi}}]{Nanishi03v42}
684: \bibinfo{author}{\bibfnamefont{Y.}~\bibnamefont{Nanishi}},
685: \bibinfo{author}{\bibfnamefont{Y.}~\bibnamefont{Saito}}, \bibnamefont{and}
686: \bibinfo{author}{\bibfnamefont{T.}~\bibnamefont{Yamaguchi}},
687: \bibinfo{journal}{Jpn. J. Appl. Phys.} \textbf{\bibinfo{volume}{42}},
688: \bibinfo{pages}{Part I, 2549} (\bibinfo{year}{2003}).
689:
690: \bibitem[{\citenamefont{Saito} \emph{et~al.}(2001)\citenamefont{Saito, Tanabe,
691: Yamaguchi, Teraguchi, Suzuki, Araki, and Nanishi}}]{Saito01v228}
692: \bibinfo{author}{\bibfnamefont{Y.}~\bibnamefont{Saito}},
693: \bibinfo{author}{\bibfnamefont{Y.}~\bibnamefont{Tanabe}},
694: \bibinfo{author}{\bibfnamefont{T.}~\bibnamefont{Yamaguchi}},
695: \bibinfo{author}{\bibfnamefont{N.}~\bibnamefont{Teraguchi}},
696: \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Suzuki}},
697: \bibinfo{author}{\bibfnamefont{T.}~\bibnamefont{Araki}}, \bibnamefont{and}
698: \bibinfo{author}{\bibfnamefont{Y.}~\bibnamefont{Nanishi}},
699: \bibinfo{journal}{Phys. Stat. Sol. (B)} \textbf{\bibinfo{volume}{228}},
700: \bibinfo{pages}{13} (\bibinfo{year}{2001}).
701:
702: \bibitem[{\citenamefont{Daudin} \emph{et~al.}(1996)\citenamefont{Daudin,
703: Rouviere, and Arlery}}]{Daudin96v69}
704: \bibinfo{author}{\bibfnamefont{B.}~\bibnamefont{Daudin}},
705: \bibinfo{author}{\bibfnamefont{J.~L.} \bibnamefont{Rouviere}},
706: \bibnamefont{and} \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Arlery}},
707: \bibinfo{journal}{Appl. Phys. Lett.} \textbf{\bibinfo{volume}{69}},
708: \bibinfo{pages}{2480} (\bibinfo{year}{1996}).
709:
710: \bibitem[{\citenamefont{Kresse and Furthm{\"u}ller}(1996)}]{Kresse96v54}
711: \bibinfo{author}{\bibfnamefont{G.}~\bibnamefont{Kresse}} \bibnamefont{and}
712: \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{Furthm{\"u}ller}},
713: \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{54}},
714: \bibinfo{pages}{11169} (\bibinfo{year}{1996}).
715:
716: \bibitem[{\citenamefont{Vanderbilt}(1990)}]{Vanderbilt90v41}
717: \bibinfo{author}{\bibfnamefont{D.}~\bibnamefont{Vanderbilt}},
718: \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{41}},
719: \bibinfo{pages}{R7892} (\bibinfo{year}{1990}).
720:
721: \bibitem[{\citenamefont{Wright and Nelson}(1995)}]{Wright95v51}
722: \bibinfo{author}{\bibfnamefont{A.~F.} \bibnamefont{Wright}} \bibnamefont{and}
723: \bibinfo{author}{\bibfnamefont{J.~S.} \bibnamefont{Nelson}},
724: \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{51}},
725: \bibinfo{pages}{7866} (\bibinfo{year}{1995}).
726:
727: \bibitem[{\citenamefont{Gan} \emph{et~al.}(2006)\citenamefont{Gan, Feng, and
728: Srolovitz}}]{Gan06v73}
729: \bibinfo{author}{\bibfnamefont{C.~K.} \bibnamefont{Gan}},
730: \bibinfo{author}{\bibfnamefont{Y.~P.} \bibnamefont{Feng}}, \bibnamefont{and}
731: \bibinfo{author}{\bibfnamefont{D.~J.} \bibnamefont{Srolovitz}},
732: \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{73}},
733: \bibinfo{pages}{235214} (\bibinfo{year}{2006}).
734:
735: \bibitem[{\citenamefont{Qian} \emph{et~al.}(1988)\citenamefont{Qian, Martin,
736: and Chadi}}]{Qian88v38}
737: \bibinfo{author}{\bibfnamefont{G.~X.} \bibnamefont{Qian}},
738: \bibinfo{author}{\bibfnamefont{R.~M.} \bibnamefont{Martin}},
739: \bibnamefont{and} \bibinfo{author}{\bibfnamefont{D.~J.} \bibnamefont{Chadi}},
740: \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{38}},
741: \bibinfo{pages}{7649} (\bibinfo{year}{1988}).
742:
743: \bibitem[{\citenamefont{Graham} \emph{et~al.}(1956)\citenamefont{Graham, Moore,
744: and Raynor}}]{Graham56v84}
745: \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{Graham}},
746: \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Moore}}, \bibnamefont{and}
747: \bibinfo{author}{\bibfnamefont{G.~V.} \bibnamefont{Raynor}},
748: \bibinfo{journal}{J. Inst. Metals} \textbf{\bibinfo{volume}{84}},
749: \bibinfo{pages}{86} (\bibinfo{year}{1956}).
750:
751: \bibitem[{\citenamefont{Harrison}(1989)}]{Harrison89}
752: \bibinfo{author}{\bibfnamefont{W.~A.} \bibnamefont{Harrison}},
753: \emph{\bibinfo{title}{Electronic structure and the properties of solids}}
754: (\bibinfo{publisher}{Dover Publications}, \bibinfo{address}{New York},
755: \bibinfo{year}{1989}).
756:
757: \bibitem[{\citenamefont{Kittel}(1996)}]{Kittel96}
758: \bibinfo{author}{\bibfnamefont{C.}~\bibnamefont{Kittel}},
759: \emph{\bibinfo{title}{Introduction to Solid State Physics}}
760: (\bibinfo{publisher}{John Wiley and Sons}, \bibinfo{address}{New York},
761: \bibinfo{year}{1996}), 7th ed.
762:
763: \bibitem[{\citenamefont{Sun} \emph{et~al.}(2003)\citenamefont{Sun, Xu, Zheng,
764: Zhou, Feng, Huan, and Wee}}]{Sun03v68}
765: \bibinfo{author}{\bibfnamefont{Y.~Y.} \bibnamefont{Sun}},
766: \bibinfo{author}{\bibfnamefont{H.}~\bibnamefont{Xu}},
767: \bibinfo{author}{\bibfnamefont{J.~C.} \bibnamefont{Zheng}},
768: \bibinfo{author}{\bibfnamefont{J.~Y.} \bibnamefont{Zhou}},
769: \bibinfo{author}{\bibfnamefont{Y.~P.} \bibnamefont{Feng}},
770: \bibinfo{author}{\bibfnamefont{A.~C.~H.} \bibnamefont{Huan}},
771: \bibnamefont{and} \bibinfo{author}{\bibfnamefont{A.~T.~S.}
772: \bibnamefont{Wee}}, \bibinfo{journal}{Phys. Rev. B}
773: \textbf{\bibinfo{volume}{68}}, \bibinfo{pages}{115420}
774: (\bibinfo{year}{2003}).
775:
776: \bibitem[{\citenamefont{Cotton} \emph{et~al.}(1995)\citenamefont{Cotton,
777: Wilkinson, and Gaus}}]{Cotton95}
778: \bibinfo{author}{\bibfnamefont{F.~A.} \bibnamefont{Cotton}},
779: \bibinfo{author}{\bibfnamefont{G.}~\bibnamefont{Wilkinson}},
780: \bibnamefont{and} \bibinfo{author}{\bibfnamefont{P.~L.} \bibnamefont{Gaus}},
781: \emph{\bibinfo{title}{Basic Inorganic Chemistry}} (\bibinfo{publisher}{John
782: Wiley and Sons}, \bibinfo{address}{New York}, \bibinfo{year}{1995}), 3rd ed.
783:
784: \bibitem[{\citenamefont{Gan and Srolovitz}(2006)}]{Gan06_unpublishedB}
785: \bibinfo{author}{\bibfnamefont{C.~K.} \bibnamefont{Gan}} \bibnamefont{and}
786: \bibinfo{author}{\bibfnamefont{D.~J.} \bibnamefont{Srolovitz}}
787: (\bibinfo{year}{2006}), \bibinfo{note}{unpublished}.
788:
789: \end{thebibliography}
790:
791: \end{document}
792: