nexoncn.com

文档资料共享网 文档搜索专家

文档资料共享网 文档搜索专家

当前位置：首页 >> >> Structure, elastic moduli and thermodynamics of sodium and potassium at ultra-high pressure

Structure, elastic moduli and thermodynamics of sodium and potassium at ultra-high pressures

M. I. Katsnelson1 , G. V. Sinko2 , N. A. Smirnov2 , A. V. Tre?lov3 , and K. Yu. Khromov3

Institute of Metal Physics, Ekaterinburg 620219, Russia Federal Nuclear Center “Institute of Technical Physics”, Snezhinsk 456770, Russia 3 Russian Science Center ”Kurchatov Institute”, Moscow 123182, Russia (Today)

2 1

arXiv:cond-mat/9912245v2 [cond-mat.mtrl-sci] 30 Mar 2000

The equations of state at room temperature as well as the energies of crystal structures up to pressures exceeding 100 GPa are calculated for Na and K . It is shown that the allowance for generalized gradient corrections (GGA) in the density functional method provides a precision description of the equation of state for Na, which can be used for the calibration of pressure scale. It is established that the close-packed structures and BCC structure are not energetically advantageous at high enough compressions. Sharply non-monotonous pressure dependences of elastic moduli for Na and K are predicted and melting temperatures at high pressures are estimated from various melting criteria. The phase diagram of K is calculated and found to be in good agreement with experiment. 64.30.+t, 64.70.Kb, 71.25.Pi

The theoretical and experimental studies of the matter properties at ultra-high pressures arouse a great interest in the connection with the possibility to obtain phases with uncommon properties as well as geophysical and astrophysical applications. As an example, the problem of metallic hydrogen can be mentioned1 . In the high pressure studies the alkali metals can be conveniently used as model objects. This is due, ?rst, to their high compressibility and, second, to the variety of physical phenomena occurring in their compression and numerous structural and electron phase transitions (see, e.g.2–9 ). For heavy alkali metals it is the famous s?d isostructural FCC-FCC transition (see, e.g.,10 and references therein) as well as the transitions to uncommon distorted phases at higher pressures11 . Recently it was supposed, basing on the electron structure calculations, that lithium can transform at high enough pressures into “exotic” phases similar to that of hydrogen12. Thus, further theoretical investigations of structural properties of alkali metals at ultra-high pressures seem to be interesting and important. Despite a lot of considerations, this is still an open problem. The most of early attempts used computational approaches which were not accurate enough from the contemporary point of view. It is well known (see, e.g.,13 ) that the highly accurate quantitative description of the electronic and, especially, lattice properties of metals needs the consideration of the real form of potential in the crystal and going beyond the frame of local approximation in the density functional, in particular, the 1

allowance for generalized gradient corrections (GGA)14 . In the present work a consistent theoretical study of the relative stability of crystal structures of Na and K under pressure as well as a variety of related lattice properties, is performed basing on these ?rst-principle calculations. The most interesting result obtained is that, contrary to the traditional concepts (see, e.g.,9 ) neither structure, which is characteristic of metals under normal conditions (BCC, FCC and HCP), is stable at high enough pressures even in Na where there are no electron transitions. The ab initio calculations of electronic structure, thermodynamical potential, equilibrium lattice parameters and elastic moduli at temperature T = 0 were carried out using the FP-LMTO method15 with allowance for the GGA in the form proposed in14 . A careful optimization of the parameters of this method15 made it possible to carry out the calculations of the total energy with an accuracy within the limits of 0.1 mRy/atom. Parameter c/a for the HCP lattice was determined by the minimization of the total energy for a ?xed speci?c volume, and the elastic moduli — by numerical di?erentiation of the total energy with respect to tetragonal and trigonal deformations (see, e.g.,9 ). Up to now, there are only few works devoted to the ?rst-principle calculations of elastic moduli of metals under high pressures (see, e.g.,16 for Mo and W). The expressions for the free energy F (connected with Gibbs thermodynamical potential G by the relationship G = F + P V , P = ?dF/dV , where P is the pressure, V is the volume) and elastic moduli Cαβγδ at ?nite temperatures can be presented in the following form: F (V, T ) = Ee? (V ) + Fph (V, T )

0 dCαβγδ Pph (V0 , T ) dV0 B0

(1)

0 Cαβγδ (V, T ) = Cαβγδ (V0 ) + V0 ? +Cαβγδ (T )

(2) is the free energy

where Fph = T

ξ,q

?n 2sh

hωξ (q) ? 2T

of the phonon subsystem in the harmonic approximation, Ee? (V ) is the total energy of the electron subsystem obtained from the FP-LMTO calculations15 , Pph = ??Fph /?V is the phonon pressure, B0 is the bulk modulus of the electron subsystem at T = 0, ξ, q, ω are the

number of phonon branch, wave vector and phonon frequency, respectively. In the expression for the elastic moduli Cαβγδ (V, T ) the ?rst term corresponds to the electron contribution at T = 0, second one - to the quasiharmonic contribution due to the e?ects of thermal expansion, and third one - to the phonon contribution obtained from the di?erentiation of Fph with respect to the corresponding deformation parameters. For the calculation of phonon contributions to thermodynamical functions the pseudopotential model described in17 , which describes with a high accuracy a wide range of lattice properties of alkali metals, was used. Figs. 1,2 show the results of calculations of Gibbs potentials at T = O for the BCC, FCC and HCP phases of sodium and potassium, respectively. It should be noted that in the case of Na the phonon contribution to △G (contribution of zero-point vibrations, △Gzp ) are Gf cc ? zp Gbcc = 1.31 × 10?5 Ry/atom, Ghcp ? Gbcc = 1.35 × 10?5 zp zp zp Ry/atom. This is well comparable with the electron contributions to △G under the normal conditions. However, already at P > 1 GPa for Na and practically at all the pressures for K the contribution of zero-point oscillations to △G can be neglected. Generally speaking, energy di?erences of order of 10?5 Ry/atom is too small to be accurately derived in our ?rst-principle calculations; nevertheless, we have obtained correct phase diagram even for sodium at low pressures. In accordance with the results of calculations, at P = 0 the BCC phase and HCP phase have the lowest energy for K and Na, respectively (actually, under these conditions Na has not HCP but 9R structure whose energy, however, is very close to HCP9 ). It is important to emphasize that this di?erence of Na from K is purely quantitative: according to the results, shown in the insert to Fig.2, K would have to transit to the hexagonal closepacked phase at the negative pressure of the order of several kilobars. As the calculations show potassium, unlike sodium, transits from the BCC to FCC structure at P ≈ 11.6 GPa, which is in excellent agreement with the experimental data5,6 . In this case the relative change in the volume △ = (Vbcc ? Vf cc )/Vbcc ≈ 0.0067 takes place at V0 /V = 2.14. Here and below V0 is the experimental value of the speci?c volume at the atmospheric pressure and temperature 10K, equal to 484.12 a.u.18 . Our calculations make it possible to suppose that the di?erence between Na and K is associated with the electronic topological transition occurring in the BCC potassium at V0 /V ≈ 2 and destabilizing the BCC structure. The similar situation takes place in Li9 while in the BCC sodium, within the whole range of pressures, the Van-Hove singularity goes away from the Fermi level under the compression. Generally, sodium seems to be a unique metal in the Periodical Table: in the whole region of the existence of BCC structure it has no singularities of electronic structure near the Fermi level, and the Fermi surface remains approximately spherical. The calculation results of equations of state for sodium

and potassium are shown in Fig.3 along with the experimental data available. It should be pointed out that the experimental data agree with the theoretical results within their accuracy limits (≈ 10%). This creates the prerequisites for the development of pressure scale based on sodium as a reference substance. Note also that at room temperature the role of the phonon contribution to pressure falls under the compression, and this contribution in itself is small (of the order of 0.3 GPa at full pressure 20-30 GPa). Figs. 4,5 display the calculation results of the dependence of elastic moduli Cij (V ) on the compression for sodium and potassium, respectively. A drastically nonmonotonous behavior of shear moduli associated with the tetragonal (C ′ = (C11 ? C12 )/2) and trigonal (C44 ) deformations in both BCC and FCC structures is noticeable. It should be pointed out that at least at compressions V0 /V < 2 the calculation results of Cik (V ) in the pseudopotential model and in the ?rst-principle approach are close. In this region the equations of state coincide in these two approaches with an accuracy up to several percents. This con?rms a su?ciently high reliability of our use of pseudopotential model for the calculation of phonon contributions to thermodynamical values. Nevertheless, the phonon contributions to the shear moduli do not exceed 10% within the whole pressure range studied. It should be noted that softening of modulus C ′ is a typical pre-transition phenomenon connected with the structural transitions between the BCC and close-packed structures19 . However, the softening of modulus C44 in the FCC structure of K at high pressures (Fig.5) is rather surprising. It appears to be similar to the softening of this modulus, taking place in the FCC structure of Cs near the electron s → d transition20 and is due to the crawling of the Fermi level over the peak of d-state density. Fig.6 shows an experimental phase diagram of potassium and the phase diagram built on the basis of our calculations. The dependence of the melting temperature on pressure, Tm (P ) in the BCC and FCC phases was obtained using the phonon spectra and di?erent melting criteria. First of all, we use the Lindeman criterion x2 (Tm )/d2 = const, where x2 (T ) =

ξq h|qeξq | ? 2Mωξq

2

(3) is the mean square

coth

hωξq ? 2T

of atom displacement, eξq is the polarization vector, M is the atom mass, d is the distance between the nearest neighbors. Although the Lindeman criterion is empirical, it may be expected that its use for ?nding the melting temperature at high pressures would be as successful as at low temperatures21 . Nevertheless one can see from Fig.6 that it is not too accurate in a broad pressure region. Varshni melting criterion25 which is based on the temperature softening of the shear moduli, namely C44 (Tm ) /C44 (0) = 0.65, (4) appears to be much more accurate. Here we use the method of the calculation of the temperature dependence 2

of elastic moduli from the phonon spectra described in3 . Note also that we describe with high accuracy the BCCFCC phase boundary. We also present the results obtained in generalized Debye model26 when all the thermodynamical quantities are calculated in the Debye model but with the Debye temperature found from ab initio elastic moduli. One can see that this description is also rather accurate for potassium. Fig.1 shows that the BCC phase of sodium becomes energetically unfavorable as compared with the HCP at a pressure about 80 GPa. At P > 100 GPa, however, this phase demonstrates anomalies in the equilibrium value of parameter c/a (Fig.7). A sharp decrease in the ratio c/a to 1.2-1.3 at V0 /V > 4.35, which is necessary to maintain the HCP lattice in equilibrium, is doubtful actually, and seems to be indicative of transition to some non closed packed phase with a large number of atoms per cell. These phases are observed in K, Rb and Cs at high pressures5,6,11 . It is usual to associate their appearance in heavy alkali metals with the s → d transition. Thus, according to our results, all the three ”typically metallic” structures - BCC, FCC, HCP do not correspond to the lowest energy in Na where no electronic transitions are observed within the pressure range considered. In order to understand qualitatively the cause of appearance of ”nonstandard” metallic phases, let us use the above mentioned pseudopotential model for the estimations. In this model the radius of ”hard” ion core is described by the pseudopotential parameter r0 17 . As the estimates show the compression V0 /V ≈ 4 corresponding to the instability of the close- packed phase coincides with the condition of overlapping ion cores 2r0 ≈ d for Na. Hence, the concept of well determined ion cores, being at the base of standard metallic bond description, becomes inapplicable at ultra-high pressures. As a result, as we have seen, the substance transforms into exotic non close-packed phases. These results are in qualitative agreement with the results12 for lithium. In conclusion, note that it would be interesting to study the structure of sodium at ultra-high pressures, which, as follows from the results obtained, may prove to be surprising. Another result of this work, permitting a direct experimental check is non-monotonous behavior of Na and K shear moduli at pressure. At last, precision theoretical description of the equation of state of sodium would make possible to use it for the development of an accurate pressure scale up to 100 GPa. Although the contemporary ?rst-principle calculations can provide high enough accuracy also for another substances (see, e.g., recent calculations27 for Si) a very high compressibility of sodium and the absence of phase transitions in a broad range of pressures make it probably the most suitable for these purposes. The authors are grateful to D. Yu. Savrasov and S. Yu. Savrasov for the permission to use the author’s version of the code realizing the method15 in their work as well as to D. Yu. Savrasov and E. G. Maksimov for useful discussions of the details of this method. 3

S. T. Weir, A. C.Mitchell, and W. J. Nellis, Phys.Rev.Lett. 76, 1860 (1996); R .J. Hemley and N. W. Ashcroft, Nature 380, 671 (1996); P. P. Edwards and F. Hensel, Nature 388, 621 (1998); W. C. Nargyara et al, Nature 393, 46 (1998) 2 V. G. Vaks, S. P. Kravchuk, and A. V. Tre?lov, Fizika Tverd. Tela 19, 1271 (1977) 3 V.G.Vaks et al, J.Phys. F 8, 725 (1978) 4 B.Olinger and J. W. Shaner, Science 219, 1071 (1983) 5 K. Takemura and K. Syassen, Phys.Rev.B 28, 1193 (1983); H.Olijnyk and W. B. Holzapfel, Phys.Lett. 99A, 381 (1983) 6 M.Winzenick, V.Vijayakumar, and W. B. Holzapfel, Phys.Rev.B 50, 12381 (1994) 7 C.-S. Zha and R. Boehler, Phys.Rev. B 31, 3199 (1985) 8 H. L. Skriver, Phys.Rev. B 31, 1909 (1985) 9 V. G. Vaks et al, J.Phys.: Cond. Matter 1, 5319 (1989) 10 A. K. McMahan, Phys.Rev. B 29, 5982 (1984) 11 U. Schwarz et al, Phys. Rev. Lett. 81, 2711 (1998); Phys. Rev. Lett. 83, 4085 (1999); Solid State Commun. 112, 319 (1999) 12 J. B. Neaton and N. W. Ashcroft, Nature 400, 141 (1999) 13 Y.-M. Jvan and M. Kozling, Phys. Rev. B 48, 14944 (1993); V. Ozolins and M. Kozling, Phys. Rev. B 48, 18304 (1993); D. L. Novikov et al, Phys. Rev. B 56, 7206 (1997); Phys. Rev. B 59, 4557 (1999); J. E. Ja?e, Z. Lin, and A. C. Hess, Phys. Rev. B 57, 11834 (1997) 14 J. P. Perdew et al, Phys. Rev. B 46, 6671 (1992) 15 S. Yu. Savrasov and D. Yu. Savrasov, Phys. Rev. B 46, 12181 (1992) 16 N. E. Christensen, A. L. Ruo?, and C. O. Rodriguez, Phys. Rev. B 52, 9121 (1995); K. Einarsdotter et al, Phys. Rev. Lett. 79, 2073 (1997); A. L. Ruo?, C. O. Rodriguez, and N. E. Christensen, Phys. Rev. B 58, 2998 (1998) 17 S. V. Vonsovsky, M. I. Katsnelson, and A. V. Tre?lov, Phys. Metal. Metallography 76, 247 (1993) 18 R. Berliner et al, Phys. Rev. B 40, 12086 (1989) 19 M. I. Katsnelson, I. I. Naumov, and A. V. Tre?lov, Phase Transitions B 49, 143 (1994) 20 V. G. Vaks et al, J. Phys.: Cond. Matter 3, 1409 (1991) 21 A. M. Bratkovsky, V. G. Vaks, and A. V. Tre?lov, Zh. Eksp. Teor. Fiz. 86, 2141 (1984) 22 S. N. Vaidya, I. J. Getting, and G. C. Kennedy, J. Phys. Chem. Sol. 32, 2545 (1971) 23 A. A. Bakanova, I. P. Dudoladov, and R .F. Trunin, Fizika Tverd. Tela 7, 1615 (1965) 24 D. Stroud and N. W. Ashcroft, Phys. Rev. B 5, 371, (1972) 25 Y. P. Varshni, Phys. Rev. B 2, 3952 (1970) 26 G. V. Sinko and A. L. Kutepov, Phys. Metal. Metallography 82(2), 129 (1996) 27 N. E. Christensen, D. L. Novikov, and M. Methfessel, Solid State Commun. 110, 615 (1999)

1

Figure captions Fig.1. Pressure dependence of the di?erences of Gibbs potentials between BCC and FCC as well as HCP and FCC structures for Na. Fig.2. Pressure dependence of the di?erences of Gibbs potentials between BCC and FCC as well as HCP and FCC structures for K. Fig.3. Equations of states for sodium and potassium at T = 295K. Solid line corresponds to FPLMTO calculations15 , dashed line - to the calculations by the pseudopotential method17 , Empty (solid) triangles, circles and asterisks (squares) are the experimental data22–24,6 for Na and K, respectively. Fig.4. The dependence of elastic moduli C ′ and C44 on the compression U for BCC Na; empty circles and squares show, respectively, the data from2 ; the solid ones - the values obtained in the present work. Fig.5. The dependence of elastic moduli C ′ and C44 on the compression U for BCC and FCC phases of K. Solid (empty) circles and squares show C ′ and C44 values for BCC (FCC) phases, respectively. The dashed line — the C ′ values for the bcc phase from2 . Fig.6. Phase diagram of potassium. The solid line experimental data7 , dashed line- the calculations using Varshni criterion (4) dashed-dot line - the calculations using Lindeman criterion (3), dotted line - generalized Debye model (see the text). Solid circles - BCC-FCC phase boundary from our calculations. Fig.7. Dependence of the total energy of HCP structure for Na on the ratio c/a for various compressions: the solid line — U = 0.75; dashed line — U = 0.76; dasheddot line — U = 0.765. The insert shows the equilibirum values of the parameter c/a for HCP structure depending on U is shown. Solid (empty) circles denote the values taken in the global (local) minimum of the total energy, correspondingly.

4

Pressure (GPa)

4 20

0 1 3 7 15 31 63

0,0 0 25 30 3 15 -0,1

Pressure (GPa)

bcc-fcc 25 12,5

50 100 200 400

hcp-fcc

Pressure(mRy) D G (GPa) C ', CC', C44 (GPa) (GPa) 44

20 -0,2 2 10 20 15

Na

bcc-fcc

0 2 4 P (GPa)

hcp-fcc

1 5 10 10 0 05

K

700 600 500 400 300 200

0

0.2

0 -1 0 -5 0

0.3

0,0 0,0

0.4

10 0,2 0,2

u= V/V0

20 0,4

0,4

0 0

0.5

30

0,6 0,8

-5

U=1-V/V liquidPressure (GPa) 0,6 U=1-V/V 0,0 0,2 0,4

0,6

40

0,8

U=1-V/V0

T, K

bcc

100 0 0 2 4 6 8 10

fcc

12 14

P, GPa

更多相关文档:

Introduction This task is about the *structure* *of* ...*of* mostly iron *and* nickel *at* the center *of* the...effective *elastic* *moduli,* *and* S *and* P wave velocit...

gels *and* glasses stems from their *structure*. While...The data points are taken *at* late stages *of* ...*elastic* *moduli* measured in the same sample *and* ...

The independent *elastic* constants (Cij), bulk *moduli,* compressibility*,* *and* ...*of* the crystal did not exceed 0.01 mRy as calculated *at* consecutive steps,...

dynamics *and* the electron *structure* *of* BiO6 ...*of* *elastic* *moduli,* as well as hysteretic ...Doping *of* BaBiO3 with *potassium* is equivalent to...

In this connection it is *of* interest to know how the *elastic* *modulus* *of* the coating layer is affected by the mineral pigments *and* the *structure* *of* the...

Piezoelectric bending response *and* switching behavior *of* ferroelectric paraelectric bilayers_工程科技_专业资料。Available online *at* www.sciencedirect.com Acta ...

更多相关标签:

相关文档

- Elastic and Seismic Properties of Dabie.Sulu Ultrahigh Pressure Metamorphic Rocks
- Differential associations of dietary sodium and potassium intake with blood pressure a focus on puls
- Pressure Dependence of the Elastic Moduli in Aluminum Rich Al-Li Compounds
- 电感耦合等离子体发射光谱法同时测定奶粉中钙铜铁钾镁锰钠锌和磷
- Theoretical Study of Sodium and Potassium Resonance Lines Pressure Broadened by Helium Atom
- Association of Urinary Sodium and Potassium Excretion with Blood Pressure
- 《镍钴锰酸锂化学分析方法第2部分锂、镍、钴、锰、钠、镁、铝、钾、铜、钙、铁、锌和.
- 钠泵活动造成的细胞内高钾是许多代谢反应进行的必要条件
- 从几起事故谈甲硫醇钠(钾)生产中的防毒措施
- β-七叶皂苷钠与10%氯化钾、甘露醇联用致血管副反应的观察与护理

文档资料共享网 nexoncn.com
copyright ©right 2010-2020。

文档资料共享网内容来自网络，如有侵犯请联系客服。email:zhit325@126.com