- •Preface
- •Contents
- •1 Nonideal plasma. Basic concepts
- •1.1 Interparticle interactions. Criteria of nonideality
- •1.1.1 Interparticle interactions
- •1.1.2 Coulomb interaction. Nonideality parameter
- •1.1.4 Compound particles in plasma
- •1.2.2 Metal plasma
- •1.2.3 Plasma of hydrogen and inert gases
- •1.2.4 Plasma with multiply charged ions
- •1.2.5 Dusty plasmas
- •1.2.6 Nonneutral plasmas
- •References
- •2.1 Plasma heating in furnaces
- •2.1.1 Measurement of electrical conductivity and thermoelectromotive force
- •2.1.2 Optical absorption measurements.
- •2.1.3 Density measurements.
- •2.1.4 Sound velocity measurements
- •2.2 Isobaric Joule heating
- •2.2.1 Isobaric heating in a capillary
- •2.2.2 Exploding wire method
- •2.3 High–pressure electric discharges
- •References
- •3.1 The principles of dynamic generation and diagnostics of plasma
- •3.2 Dynamic compression of the cesium plasma
- •3.3 Compression of inert gases by powerful shock waves
- •3.4 Isentropic expansion of shock–compressed metals
- •3.5 Generation of superdense plasma in shock waves
- •References
- •4 Ionization equilibrium and thermodynamic properties of weakly ionized plasmas
- •4.1 Partly ionized plasma
- •4.2 Anomalous properties of a metal plasma
- •4.2.1 Physical properties of metal plasma
- •4.2.2 Lowering of the ionization potential
- •4.2.3 Charged clusters
- •4.2.4 Thermodynamics of multiparticle clusters
- •4.3 Lowering of ionization potential and cluster ions in weakly nonideal plasmas
- •4.3.1 Interaction between charged particles and neutrals
- •4.3.2 Molecular and cluster ions
- •4.3.3 Ionization equilibrium in alkali metal plasma
- •4.4 Droplet model of nonideal plasma of metal vapors. Anomalously high electrical conductivity
- •4.4.1 Droplet model of nonideal plasma
- •4.4.2 Ionization equilibrium
- •4.4.3 Calculation of the plasma composition
- •4.5 Metallization of plasma
- •4.5.3 Phase transition in metals
- •References
- •5.1.1 Monte Carlo method
- •5.1.2 Results of calculation
- •5.1.4 Wigner crystallization
- •5.1.5 Integral equations
- •5.1.6 Polarization of compensating background
- •5.1.7 Charge density waves
- •5.1.8 Sum rules
- •5.1.9 Asymptotic expressions
- •5.1.10 OCP ion mixture
- •5.2 Multicomponent plasma. Results of the perturbation theory
- •5.3 Pseudopotential models. Monte Carlo calculations
- •5.3.1 Choice of pseudopotential
- •5.5 Quasiclassical approximation
- •5.6 Density functional method
- •5.7 Quantum Monte Carlo method
- •5.8 Comparison with experiments
- •5.9 On phase transitions in nonideal plasmas
- •References
- •6.1 Electrical conductivity of ideal partially ionized plasma
- •6.1.1 Electrical conductivity of weakly ionized plasma
- •6.2 Electrical conductivity of weakly nonideal plasma
- •6.3 Electrical conductivity of nonideal weakly ionized plasma
- •6.3.1 The density of electron states
- •6.3.2 Electron mobility and electrical conductivity
- •References
- •7 Electrical conductivity of fully ionized plasma
- •7.1 Kinetic equations and the results of asymptotic theories
- •7.2 Electrical conductivity measurement results
- •References
- •8 The optical properties of dense plasma
- •8.1 Optical properties
- •8.2 Basic radiation processes in rarefied atomic plasma
- •8.5 The principle of spectroscopic stability
- •8.6 Continuous spectra of strongly nonideal plasma
- •References
- •9 Metallization of nonideal plasmas
- •9.1 Multiple shock wave compression of condensed dielectrics
- •9.1.1 Planar geometry
- •9.1.2 Cylindrical geometry
- •9.3 Metallization of dielectrics
- •9.3.1 Hydrogen
- •9.3.2 Inert gases
- •9.3.3 Oxygen
- •9.3.4 Sulfur
- •9.3.5 Fullerene
- •9.3.6 Water
- •9.3.7 Dielectrization of metals
- •9.4 Ionization by pressure
- •References
- •10 Nonneutral plasmas
- •10.1.1 Electrons on a surface of liquid He
- •10.1.2 Penning trap
- •10.1.3 Linear Paul trap
- •10.1.4 Storage ring
- •10.2 Strong coupling and Wigner crystallization
- •10.3 Melting of mesoscopic crystals
- •10.4 Coulomb clusters
- •References
- •11 Dusty plasmas
- •11.1 Introduction
- •11.2 Elementary processes in dusty plasmas
- •11.2.1 Charging of dust particles in plasmas (theory)
- •11.2.2 Electrostatic potential around a dust particle
- •11.2.3 Main forces acting on dust particles in plasmas
- •11.2.4 Interaction between dust particles in plasmas
- •11.2.5 Experimental determination of the interaction potential
- •11.2.6 Formation and growth of dust particles
- •11.3 Strongly coupled dusty plasmas and phase transitions
- •11.3.1 Theoretical approaches
- •11.3.2 Experimental investigation of phase transitions in dusty plasmas
- •11.3.3 Dust clusters in plasmas
- •11.4 Oscillations, waves, and instabilities in dusty plasmas
- •11.4.1 Oscillations of individual particles in a sheath region of gas discharges
- •11.4.2 Linear waves and instabilities in weakly coupled dusty plasmas
- •11.4.3 Waves in strongly coupled dusty plasmas
- •11.4.4 Experimental investigation of wave phenomena in dusty plasmas
- •11.5 New directions in experimental research
- •11.5.1 Investigations of dusty plasmas under microgravity conditions
- •11.5.2 External perturbations
- •11.5.3 Dusty plasma of strongly asymmetric particles
- •11.5.4 Dusty plasma at cryogenic temperatures
- •11.5.5 Possible applications of dusty plasmas
- •11.6 Conclusions
- •References
- •Index
392 |
NONNEUTRAL PLASMAS |
approached. The increase of the stability parameter q (or the confining potential) causes a noticeable change of the ion beam properties. For q = 0.33 one can see an abrupt decrease in the fluorescence intensity at a certain value of ∆ω (indicated by the arrow in the figure), followed by a narrow asymmetric peak. Such behavior can be interpreted as a result of the balance between the laser cooling and the heating of ions in a r.f. field. The e ciency of the heating rapidly falls o as the ions become more ordered. The slight decrease of the stability parameter (q = 0.31) and, hence a decrease of the r.f. heating, causes the phase transition to occur at larger values of ∆ω. Because of the noise, this makes detection of the transition very di cult. Another indication of the phase transition is a sudden decrease of the beam diameter. For the experimental conditions corresponding to Fig. 10.16, the formation of the rotating ion string (viz., one–dimensional ion crystal) with the interparticle distance a 20 µm was observed. The longitudinal temperature measured in the reference frame moving together with the beam was about 3 mK, which corresponds to the coupling parameter γ > 500. By changing the parameters of the plasma and the trap (first of all, the number of ions), two– dimensional zigzag and three–dimensional helix structures have been obtained and investigated by Schramm et al. (2002). It was shown that the stability range of the crystalline structures formed in the storage ring is substantially smaller than that obtained in stationary ionic crystals and in the MD simulations.
10.3Melting of mesoscopic crystals
The nonneutral plasmas form crystalline states at su ciently low temperatures. As the temperature grows the crystal melts, yet keeping the correlation and the short–range order, and when the temperature is high enough the plasma is in a disordered gaseous state. Both the numerical simulations and experiments suggest that the stable crystalline state of the macroscopic plasma is the bcc lattice, and melting occurs at γ 173. In mesoscopic plasmas, di erent types of ordered structures can be formed. As discussed in the previous section, ions confined in a parabolic potential well (typical for ion traps) form clouds with a shell structure and a well–defined surface. In the surface layer and in each shell ions arrange themselves into a triangle lattice. As the number of ions in the cloud increases, the transition to a bcc lattice is observed in the bulk region, which is typical of infinite plasmas.
The melting of such finite–size crystalline structures has been studied by Schi er (2002) with MD simulations. In finite systems the transitions between ordered and disordered states do not occur abruptly – they develop continuously in a certain range of parameters (e.g., temperature or density). Figure 10.17 shows two outer spherical shells of the cloud containing 104 ions at two di erent temperatures. Also, the radial density distribution and the pair correlation function g(r) are shown, as obtained from simulations at three di erent temperatures corresponding to the crystalline, liquid, and gaseous states.
The potential energy and the heat capacity per particle versus the temperature are shown in Fig. 10.18. One can see the well–defined melting transition,
MELTING OF MESOSCOPIC CRYSTALS |
393 |
The upper figure shows the image of two outer shells containing 104 ions. Ions of the outmost shell are in black, the next shell is in gray. The middle figure shows the radial distribution of the ion density, and the lower one, the pair correlation function g(r) (the length is in units of the Wigner–Seitz radius) (Schi er 2002).
394 |
NONNEUTRAL PLASMAS |
kT
2.5
2.0
1.5
C
10
4
1 |
4 |
10 |
40 |
Temperature, 1000·γ-1
Fig. 10.18. Total energy U and thermal capacity C per particle for an ensemble of charges
(Schi er 2002). The calculations are for an infinite system (◦) and for systems containing 104 ions (•). Temperature is in units of γ−1.
both in a finite system of 104 ions and in an infinite Coulomb system, although in the former case the melting temperature is somewhat smaller and the transition itself is smeared out. Based on the analysis of the di usion coe cients performed by Schi er (2002), a significant increase in the ion di usion within each shell, as compared with the di usion between the neighboring shells, has been revealed in the vicinity of the melting transition.
Schi er (2002) also addressed the question: Which factor plays the more important role in the decrease of the melting temperature observed in mesoscopic crystals – the finite size or di erent types of ordered structures? Figures 10.19 and 10.20 show the heat capacity and the melting temperature calculated for a system of 100, 1000 and 10 000 ions, as well as for an infinite system. The height of the heat capacity peak decreases and its maximum shifts towards smaller temperatures as the number of ions decreases. The obtained results exhibit a smooth dependence on the size, regardless of the particular crystalline structure. Figure 10.20 shows the melting temperature, Tm, versus the number of ions in the outmost shell, Ns, normalized to the total ion number. One can see that ∆Tm = Tm(∞) − T (N ) Ns. Hence, one can claim that the decrease of the melting temperature is almost solely due to the finiteness of the system.
COULOMB CLUSTERS |
395 |
3.5Infinite Matter
3.0ions
Specific heat
1 |
3 |
5 |
7 |
Temperature 1000 γ−1
Thermal capacity of nonideal plasmas near the melting point (Schi er 2002). The calculations are for 100 ( ), 1000 ( ), and 10 000 (•) ions as well as for an infinite
system (dashed line).
10.4Coulomb clusters
In this section we discuss the properties of small (Ni 10) systems of charged particles. At su ciently low temperatures they form ordered structures called “Coulomb clusters”. Although the properties of such clusters and mesoscopic crystalline systems are similar, the Coulomb clusters exhibit a number of quite interesting features. Below we discuss the clusters trapped in a parabolic potential well, since this confinement is typical for most experiments and numerical (MD and MC) simulations. The equilibrium configurations of particles are determined by the minimum of the potential energy. Taking into account Eq. (10.3), the energy is
U = |
|
e2 |
|
|
+ |
M ω2 |
(z2 |
+ β r2). |
(10.8) |
||||
|
|
|
|
z |
|
|
|||||||
|
|
|
|
2 i |
|||||||||
i>j |
ri |
− |
rj |
| |
|
i |
|
|
i |
|
|||
|
| |
|
|
|
|
|
|
|
|
||||
By introducing dimensionless variables, |
|
|
|
|
|
|
|
|
|||||
|
|
|
|
|
|
U → |
e4M ω2 |
|
1/3 |
|
|||
r → (e2/M ωz2)1/2r, |
|
(10.9) |
|||||||||||
|
|
z |
U, |
||||||||||
|
2 |
|
|||||||||||
one can reduce the potential energy to the form which depends only on the number of particles and the asymmetry parameter β. In a spherically symmetric
trap, β = 1, we have |
|
1 |
|
|
|
|
|
| |
|
|
|
|
|
− | |
|
|||||
U = |
|
|
|
|
+ r2. |
(10.10) |
|
ri |
|
rj |
i |
|
|
i>j |
|
i |
|
|||
396
Temperature, 1000 γ -1
NONNEUTRAL PLASMAS
.
6
Infinite Matter
510 000 ions, γ = 209
4
1000 ions, γ 278
3
2 |
|
100 ions, γ |
500 |
|
|
|
|
|
|
||
1 |
|
|
|
|
|
0 |
0.2 |
0.4 |
0.6 |
0.8 |
1 |
|
Fraction of ions in surface layer
Fig. 10.20. Temperature corresponding to the maximum of the thermal conductivity,
versus the fraction of ions in the outmost shell, Ns/Ni (Schi er 2002).
Figure 10.21 shows equilibrium configurations of clusters consisting of di erent numbers of particles and confined in a spherically symmetric potential well. At a glance, one can think that all ions in clusters with Ni ≤ 12 lie on a sphere. In fact, however, this is true only for “symmetric” configurations with Ni = 2, 3, 4, 6, 8, 12, whereas for Ni = 5, 7, 9, 10, 11 the radial positions of ions are not the same (Dubin and O’Neil 1999). Starting from N = 13, the second shell is formed. As the number of ions increases further, the concentric shell structure is forming, as described in the previous sections.
Tsuruta and Ichimaru (1993) calculated the potential energy of Coulomb clusters with Ni ≤ 60. The interaction energy per particle, U/(Nie2/rs), decreases as the number of particles grows (see Fig. 10.22). This is because the role of surface e ects becomes weaker. One can see several minima in the dependence U versus Ni; the deepest ones correspond to Ni = 6, 12, and 38. Following nuclear physics terminology, Tsuruta and Ichimaru (1993) refereed to these Ni as “magic numbers” for spherical Coulomb clusters. Note that the “magic” clusters have the most–symmetric configurations corresponding to an octahedron for Ni = 6, an icosahedron for Ni = 12, and a two–block structure consisting of an octahedron inside a phase–centered icosahedron for Ni = 38 (Tsuruta and Ichimaru 1993).
|
|
COULOMB CLUSTERS |
|
397 |
||
|
a |
a |
|
a |
a |
b |
a |
|
|
|
|||
|
b |
a |
b |
b |
a |
|
|
|
|
|
|
c |
|
N = 3 |
N = 4 |
N = 5 |
N = 6 |
N = 7 |
N = 8 |
N = 9 |
a |
|
a |
a |
|
c |
|
b |
|
|
b |
|
||
c |
|
|
|
|
|
|
|
|
|
|
a |
|
|
|
|
|
|
|
|
|
N = 10 |
N = 11 |
N = 12 |
N = 13 |
N = 14 |
N = 15 |
N = 18 |
Fig. 10.21. Configurations of Coulomb clusters in a spherically symmetric parabolic po-
tential well (Rafac et al. 1991).
, |
, |
, |
, |
|
|
|
|
|
|
|
|
|
|
|
Fig. 10.22. Dimensionless energy of the electrostatic interaction for N –particle clusters
(Tsuruta and Ichimaru 1993).
In anisotropic confinement the equilibrium configuration is determined by the value of β. For β 1, the ion cloud is stretched along the z–axis, forming one– dimensional ordered structures. For β 1 ions form two–dimensional structures in the xy–plane. The ground state of two–dimensional Coulomb clusters is well studied (Lozovik 1987; Lozovik and Mandelshtam 1990, 1992; Rafac et al. 1991; Bedanov and Peeters 1994). For Ni ≤ 5 the ions lie on a circle, for 6 ≤ Ni ≤ 8 one of the ions is located in the center, and for Ni = 9 two ions are inside the ring. For Ni = 15 the second (inner) ring is completed (5, 10) and for Ni = 16 one ion again appears in the center (1, 5, 10). The two–dimensional ring structures are completely analogous to the three–dimensional shell structures in spheroidal nonneutral plasmas.
