Download Atomic Processes and High-power Radiation

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Introduction to gauge theory wikipedia , lookup

Electrostatics wikipedia , lookup

Time in physics wikipedia , lookup

Nuclear physics wikipedia , lookup

Electromagnetism wikipedia , lookup

Field (physics) wikipedia , lookup

Nuclear fusion wikipedia , lookup

Transcript
Atomic Processes and High-power Radiation
Richard M. More
National Institute for Fusion Science
Oroshi, Toki, Gifu, Japan 509-5292
and
Hitoki Yoneda
University of Electrocommunications
Chofugaoka, Chofu, Tokyo, Japan 182-8585
ABSTRACT
What scientific research can be done with strong electric fields from new pulsed laser
and x-ray sources? In this paper we examine three possibilities: first, nuclear fusion
directly driven by high-intensity laser radiation in a DT gas target; second, nonlinear
(multiphoton) x-ray processes in high-charge ions irradiated by high intensity x-ray
sources; third, the interaction of ultra-short pulse lasers with solid targets. In the third
case, some experiments have been performed and while theory agrees with experiment
for simple metal targets, unexpected behavior occurs for targets having more complex
chemical structure. These laser absorption experiments illustrate the new science that
is already emerging from experiments with high-power lasers.
1.) Strong Electric Fields
Nature makes strong electric fields. Near a uranium nucleus,
E
peak
~ 2 1019 Volt/cm
(1)
When an electron passes near the uranium nucleus, the nuclear field causes an
acceleration of ~ 10^1 times the Earth's gravity g, but doesn't create black holes or
cause other general-relativity effects. The nuclear electric field exceeds 1016 Volt/cm
over a distance of about one electron Compton length fi I me and yet it is not strong
enough to pull electrons and positrons from the vacuum. The nuclear field cannot
even pull protons out of the nucleus. Since at present the strongest laser fields are
millions of times weaker, lasers probably cannot do these exciting things either.
The nuclear field can be used for accurate scientific experiments.
For
CP634, Science of Superstrong Field Inter actions, edited by K. Nakajima and M. Deguchi
© 2002 American Institute of Physics 0-7354-0089-X/02/$ 19.00
139
example, fine-structure splitting of electron energy levels is caused by the spin-orbit
interaction,
0
f-f
**SO
— _
—
0 9
_*
7T
2m c
J7(r\ V n
^v / ^jr
n\
^ '
Eq. (2) includes Thomas precession as well as the electron spin interaction with the vxE
magnetic field experienced by an electron moving in the nuclear electric field. These
effects are well-known corrections to the non-relativistic Schroedinger equation, and Eq.
(2) shows that fine-structure splitting directly measures the electric field. The Dirac
equation and quantum electrodynamics also include smaller effects such as the Lamb
shift. Spin-orbit splitting has been carefully measured in hydrogen-like uranium ions.
The average electric field near a 2p electron is less than Eq. (1), but still large compared
to laser electric fields. Recent experiments find no observable difference from QED
theory to a precision of about a part per thousand. [1]
Comparing to this strong natural electric field we are skeptical about the
possibility of easy discoveries with lasers producing fields of 10*0 _ lo1^ Volts/cm.
The targets for laser research must be chosen carefully.
The effect of an electric field depends on the space and time scales over which
the field exists, and the scientific application depends on having access to perform
experiments. From this viewpoint lasers are interesting because laser fields extend
over a larger distance (about a micron) and there is good access to the high-field region.
Atoms are easily affected by an applied electric field. The Stark splitting of
spectral lines is well-known. Already in the 1930's the Stark effect was measured for
hydrogen emission lines at fields up to 10" Volt/cm. [2] For hydrogen-like (oneelectron) ions the energy levels are independent of angular momentum and the classical
electron orbits are ellipses fixed in space, so an orbit pointing along the field has a
different energy from an orbit perpendicular to the field. The result is a large linear
Stark splitting. For many-electron ions, the orbits are precessing ellipses, and the
precession causes the first-order effects of the applied field to cancel, leading to a
(smaller) quadratic Stark effect.
When the applied electric field is large it can pull electrons from the atom, a
process called Stark ionization. For time-dependent fields, the mechanism of
ionization depends on the frequency of the field, and one talks of tunneling ionization,
"over-the-barrier" (OTB) ionization and multiphoton ionization.[3]
Other atomic processes are affected by electric fields. For example, in
dielectronic recombination, free electrons attach to an ion in large orbits of low angular
momentum. An externally applied electric field can transfer these electrons to large
angular momentum where they do not autoionize, causing an enhancement of the
recombination. Recent experiments on electron-ion recombination in storage rings
find that radiative recombination also depends on environmental electric and magnetic
fields. [4]
140
Recent research on atomic processes in laser fields has explored above-threshold
ionization (ATI) and the generation of high harmonics. [3]
2.) Fusion in strong laser fields
In search of something more exotic, we have considered fusion reactions
induced by strong electric fields in DT gas irradiated by high-power lasers. [5] We
imgine laser intensities Ipeak ~ 10^ W/cm^ with pulse lengths of T ~ 1 psec, to be
created by future lasers in the multi-PWatt power range.
If such a laser is focused in DT gas, the electrons immediately reach relativistic
quiver energies and essentially decouple from the ions. The ions also have a quiver
motion. All deuterium ions will oscillate in phase, but because of the mass difference
between D and T ions, the unlike ions have a relative velocity which can be estimated
by
1
1
——— ———
MD
—— COS(flJf)
MT) co
(3)
If the laser intensity (I = c E2/8jc) is large, this velocity difference can be large enough
to initiate fusion reactions. At present this is a theoretical idea which we call ion
quiver-velocity (IQV) fusion. [5]
Figure 1.) IQV fusion: the laser directly shakes D, T nuclei to fusion velocities.
IQV fusion would occur against a strong background of fusion reactions from
ion acceleration by electrostatic fields generated by the target interaction. Such
electrostatic (beam-target) fusion has already been observed, so it is likely these
reactions will be numerous at the high intensities considered here. Special features of
IQV fusion may make it visible despite this strong background. IQV fusion reactions
can only occur for nuclei with unequal mass. Since IQV fusion can occur only in the
high-intensity focal-spot, the IQV reactions are strongly localized in space and time,
and this localization might help distinguish IQV fusion from background. If IQV
fusion can be measured, they may provide a way to directly verify the existence of high
intensities in the laser focal-spot.
141
To calculate IQV fusion, we must find the focal-spot diffraction pattern,
calculate fusion for oscillating Maxwell distributions, and form the space-time integral.
It is assumed that D, T ions have random spatial locations (i.e., are uncorrelated)
consistent with the high energy-density in the focal spot. We neglect reactions which
occur outside the focal-spot. At present we also neglect reactions due to fusion alpha
knock-ons, although such reactions might increase the fusion yield because the alpha
energy-loss to relativistic electrons is so small. We also neglect possible enhancement
of the fusion cross-section by the laser electric field. The following examples illustrate
the possibilities. The gas targets are taken to have a density of 80% of the critical
density.
Case 1: 100 TW laser
D-T gas target
F/2 lens
0.24 reactions
Case 2: 1 PW laser
D-T gas target
F/2 lens
8.5 10^ reactions
Case 3: 10 PW laser
D-T gas target
F/6 lens
9 10^ reactions
Case 4: 10 PW laser
DT droplet target (0.1 liquid density)
F/6 lens
5 108 reactions
From these calculations it appears that IQV fusion should be observable. It has
scientific interest as new kind of fusion at exotic conditions. The predictions deserve an
experimental test, but with the yields calculated here, this would be scientific research
rather than energy engineering.
3.) X-Ray Nonlinear Optics
We next consider atomic effects of high-intensity X-rays. New x-ray sources
(x-ray lasers, high harmonic sources and undulators proposed for electron linear
accelerators) may offer the possibility of experiments with short-wavelength radiation at
very high intensities in the coming decade.
We consider a 50 fsec pulse of kilovolt X-rays focused to 10^ W/cm^. To
reach this intensity it would be necessary to focus 50 microjoules of X-rays into a 1|4~
target area. This may not be easy, but is in the range of conditions being discussed by
scientists developing new x-ray sources. Existing x-ray laser sources have photon
energies ~ 100 eV.
Another possible approach for this research would be visible laser irradiation of
ions in a high-energy ion storage ring. We consider ions trapped in a storage ring,
having very high velocity so their motion is relativistic with a gamma-parameter of
142
about 50. Such a storage ring does not exist today but may be possible in the future.
If a visible laser shines on such ions from the correct direction, the light is Dopplershifted to 100 eV or so (and the intensity is raised) in the ion rest-frame. The moving
ion would be irradiated by high-intensity x-rays whose parameters could be controlled
very precisely.
If high-intensity x-rays interact with an ion, can they produce nonlinear optical
phenomena like those caused by visible lasers? We should not expect an exact
repetition of visible-light nonlinear optics, because the parameter 17^ is so much smaller.
For this reason we cannot expect to observe high harmonics or above-threshold
ionization resulting from x-ray irradiation.
Let us ask about the simplest nonlinear effects. Will there be two-photon
absorption? Harmonic production? How do the cross-sections for these processes
depend on photon frequency hv and target atomic number Z?
To study non-linear processes caused by high-intensity x-rays, we have solved
the time-dependent Schroedinger equation for a hydrogen-like ion irradiated by a
Gaussian pulse of 1 keV x-rays. The target ion is taken to be hydrogen-like Silicon in
its Is groundstate. The computer code includes up to 60 basis states. The X-rays are
described by time-dependent classical fields E(t), B(t). The x-ray pulse length is taken
to be 50 fsec (FWHM). For peak intensities in the range 10^ - 10^ W/cm^, we find
strong two-photon absorption Is —> 2s, a nearly resonant interaction for H-like Silicon
ions.
In these exploratory calculations, the population of the 2s state oscillates rapidly
during the x-ray pulse, a nonlinear Rabi oscillation, and the probability to remain in the
2s state at the end of the laser pulse oscillates rapidly with the peak x-ray intensity.
The average excitation probability is a few percent.
A second type of exploratory calculation uses the Thomas-Fermi theory of a
many-electron ion in strong electric field. If the x-ray frequency is low compared to
the atomic excitation frequencies, it is at least qualitatively reasonable to treat the x-ray
electric field as a slowly-varying static field. We performed 2-dimensional ThomasFermi self-consistent field calculations of ions in strong static electric fields.
The induced electric dipole moment depends on the electric field:
p = ocE + yE 3 + ...
(4)
where the functions a = a(Z, Q), y = y(Z, Q) are linear and nonlinear polarizabilities.
Using Thomas-Fermi theory, we calculated a and y for many ions and obtained general
interpolation formulas.
For an oscillating electric field, the cubic polarization oscillates at frequency 3co
and generates third-harmonic radiation. We divide the third-harmonic emission by the
143
incident intensity to obtain a cross-section for third-harmonic conversion; this crosssection is proportional to the x-ray intensity squared. The formulas are
,-
Eq. (6) predicts large cross-sections for heavy ions in low charge states.
However these ions are easily ionized by three-photon processes, and the approximation
of using the nonlinear static polarizability for x-ray response is least appropriate. For
very high charge ions, the cross-section is smaller, comparable to the Compton crosssection. Further analysis is underway.
4.) Short-Pulse Laser Interaction with Solid Targets
Short-pulse lasers can be used to heat any solid material. If we measure the
absorption (the fraction of laser light absorbed) we learn the heat energy deposited in
the target and obtain an indirect determination of the high-temperature electrical
properties of the target material. The temperature can be controlled by adjusting the
laser heating energy. The target remains near its initial solid density if the pulse is
sufficiently short, especially at the lower intensities.
Such experiments have been performed in several laboratories. We can model
such experiments by solving the Maxwell equations for laser light interacting with a
heated target material. It is necessary to make assumptions about the optical properties
(dielectric function) of the hot material and also about the density profile resulting from
target hydrodynamic expansion. Typically this expansion cannot be neglected even
for 100 fsec laser pulses unless the intensity is low.
In our calculations, we assume a local, linear relation
j = aE
(7)
At the laser frequency the (AC) conductivity is a complex number. We solve Maxwell
equations with a complex dielectric function e(x) that depends on the depth x from the
target surface,
£(x) =
i +
CO
The computer code solves the Maxwell equations for s- and p-polarized light at any
angle of incidence. If there is sufficient experimental data, including dependence of
the absorption upon the angle of incidence, polarization and laser intensity, we are able
144
to deduce consistent values for the hydrodynamic scale length and optical coefficients
of the high-temperature target material.
Even if we determine the dielectric function e = £r + i £j by analysis of
experimental data, we must ask: Is the result unique? Can we determine the
microscopic mechanisms which control the dielectric function? These are difficult
questions. However, if we can determine the dielectric function over a range of
frequencies, 8(co), these questions will have better answers. Thus experiments with a
broad-band probe pulse would be very valuable.
Accurate and detailed modeling should include the spatial averaging over the
range of intensities in the focal spot, time-dependence of the laser and probe pulses, and
electron thermal conduction in the target material. Simpler models are appropriate for
conditions of low intensity or low target temperature.
Experiments underway in the University of Electro-communications use a 300
fsec pump pulse of .248 micron radiation (at normal incidence) and a 120 fsec probe
pulse at 0.75 micron wavelength (0.65 degrees) to measure reflectivity of heated
targets. [8] The laser system has a 0.01-0.1 kHz repetition rate so it can take data
rapidly. An example of time-dependent reflectivity (taken with varying delays of the
probe pulse) for an aluminum layer (on glass) is shown in Figure 2.
SN. 1090503 dRs-6.5%, dt=326fs
I
2.8
3
3.2
3.4
3.6
3.8
4
time [ps]
Figure 2. Time-dependent reflectivity for aluminum targets measured with the
laser system described in the text.
In these experiments, the electric fields and currents induced in the targets are
not small. For example, at a typical moderate laser intensity 1 = 3 10^ W/cm^, for a
laser wavelength of k = 0.25 |i the calculations give an electric field Eo ~ 108 Volt/cm
and a current density j ~ 7 10^ Amp/cm^ in the target material. There is an
interesting question whether Eq. (7) remains valid at such high currents.
145
These experiments expose solid targets to pulsed high electric fields in a wellcontrolled manner. At high intensities, most targets have the same behavior and agree
with theory. [6] These conditions correspond to target temperatures of perhaps 50 eV
or higher. At lower intensities, we observe interesting new properties of warm
condensed matter. These properties include anomalous high absorption by ionic
solids including glass, quartz and sodium chloride, and anomalous absorption features
in the response of transition metals. [See REF. 6]
The high absorption by ionic materials is especially surprising because it differs
strongly from the behavior of these materials in the usual solid state.
In summary, we see that real-world experiments can be performed today with
enough precision to yield new scientific information about the optical properties of
matter at temperatures higher than previous accurate experiments. The challenge is to
assemble and interpret enough of this data to teach us something about the nature of the
transition from the usual solid state to hot dense plasma.
REFERENCES
[1.] J. P. Briand, P. Chevallier, P. Indelicate, K. P. Ziock and D. D. Dietrich, Phys. Rev.
Letters 65, 2761 (1990).
[2.] H. Bethe and E. Salpeter, Quantum Mechanics of One- and Two-electron Atoms,
Academic Press, Inc., New York, 1957.
[3.] See article by C. Joachain in Atoms, Solids and Plasmas in Super-Intense Laser
Fields, Ed. by D. Batani, C. Joachain, S. Martellucci and A. Chester, Kluwer
Academic/Plenum Publishers, 2001; and in Laser Interaction with Atoms, Solids and
Plasmas, Ed. by R. More, Plenum Publishers, 1992
[4.] See article by G. Dunn in Recombination of Atomic Ions, Ed. by W. Graham, W.
Fritsch, Y. Hahn and J. Tanis, Plenum Publishers, 1992.
[5.] T. Kenmotsu, T. Nishikawa and R. More, in Inertial Fusion Science and
Applications, Kyoto, September, 2001 (to be published).
[6.] D. F. Price, R. M. More, R. S. Walling, G. Guethlein, R. L. Shepherd, R. E. Stewart
and W. E. White, Physical Review Letters 75, 252 (1995).
[7.] A. Forsman, A. Ng, G. Chiu and R. More, Physical Review E 58, R1248 (1998).
[8] H. Yoneda, S. Ohta, K. Ueda and R. More, Proc. of LAMP 2002, Special Session 3,
Osaka, May 2002 (to be published in SPIE).
146