M A T E R I A L S C I E N C E --- P L U M E
A N A L Y S I S
Abstract -- Introduction -- Experimental Apparatus -- Results & Analysis -- Contrast Effects -- Conclusion -- Acknowledgements -- Figure Captions
Pulse-Contrast Effects on Energy Distributions of C1+
to C4+ Ions for High-Intensity 100-fs Laser-Ablation Plasmas
P. A. VanRompay, M. Nantel, and P. P.
Pronko
Center for Ultrafast Optical Science and Department of Electrical Engineering
and
Computer Science, University of Michigan, Ann Arbor, MI 48109, USA
Published in Applied Surface Science, vol. 127-129, pages
1023-1028, 1998.
ABSTRACT
top
A time-resolved spherical-sector electrostatic analyzer is used to measure the energy spectra of carbon ions from ablation plumes of graphite that are generated by high-intensity ultrafast laser pulses. The time-resolved spectra from the E/q analyzer are used to construct energy distributions of the individual charge states of various carbon ions. The measurement range of the detector is from 0.05 to 20 keV per unit charge, allowing for observation of up to 80 keV at charge state four. Data are presented for laser intensities of 5 x 1014 to 2 x 1015 W/cm2 at a wavelength of 780 nm. Charge states and energy distributions are observed to be strongly dependent on the laser pulse contrast and/or presence of a pre-pulse. A bi-modal energy distribution is obtained for high contrast pulses and interpreted in terms of a plasma hydrodynamic model for the lower ion energies (few hundred eV) and an accelerated ion contribution for the higher energies (several keV).
PACS: 52.40.Nk, 52.70.Nc, 41.75.Ak, 52.50.Jm
Keywords: laser-plasma interaction, ion energy distributions, electrostatic analyzer, plasma hydrodynamic expansion, accelerated ions, laser contrast effects
INTRODUCTION
top
A succession of theoretical papers has appeared recently that model and describe mechanisms by which high-energy ions could be generated in a laser-ablation plasma [,,,]. A few experimental papers have also appeared which examine ion energy distributions using time-of-flight methods [,,,]. With the advent of CPA high-intensity lasers [], it has become possible to achieve optical laser-field intensities that make the practical generation of high-energy ions a laboratory reality. Interest continues to build around the possibility of using such an approach as a method for designing an MeV ion beam device that would provide current densities competitive with, or superior to, modern SF6 pressure tank MeV ion accelerators. In addition to such practical ion beam applications, studies of laser plasmas, through an analysis of expanding ablation plumes, represent a subject of continuing interest in regard to basic physics processes in laser plasmas and for diagnostic control of the vapor phase in laser PVD and CVD film deposition technology.
We have undertaken a series of experiments that use a spherical-sector electrostatic analyzer to extract specific ion species and energy information. This paper presents some of the early results from that study and interprets the data in terms of the hydrodynamic expansion properties of the laser plasma and other ambipolar field effects that exist in the early stages of the plasma formation. Unique aspects of the plasma are involved as a consequence of the shortness of the laser pulse (100 fs).
EXPERIMENTAL APPARATUS AND PROCEDURES top
The present experiments were done with a CPA Ti:Sapphire-based 10-Hz laser system with s-polarized light on sample []. Chirped-pulse amplification (CPA) is a technique used to amplify ultrashort laser pulses without damaging the amplifying medium and results in subpicosecond laser pulses focusable to intensities on the order of 1014 - 1016 W/cm2 or more. For such high-intensity pulses, the contrast ratio (the ratio of peak intensity to background intensity) is an important parameter for laser-matter interactions and can be measured with a 3rd-order auto-correlator []. When focused on a solid, the energy in the background of a low contrast pulse can be high enough to perturb the material before the arrival of the main ultrashort pulse, especially when the background extends to long times (ns) []. In our experiments, we investigated pulses with both 105 contrast, where up to half of the laser energy is in the background, and 106 contrast, where only one-tenth of the energy is in the background.
The ultrashort, amplified laser pulse is transported to the sample under a vacuum of ~10-5 Torr (to eliminate self-phase modulation in air) and is focused onto the sample using a parabolic mirror. The plasma produced by the laser pulse expands in a plume directed away normal to the sample. For analysis of the plume, the ions are sent down a 75 cm drift tube to the spherical-sector electrostatic energy analyzer which is housed in a separate vacuum chamber at ~10-7 Torr. The two vacuum systems are effectively isolated using differential pumping. Low vacuum pressure in the drift region and analyzer chamber is important to eliminate spurious charge-exchange effects and to minimize noise in the microchannel plate (MCP) detectors of the analyzer. Providing a voltage on the analyzer sector plates allows the selection of ions with a known energy-to-charge (E/q) ratio. For example, a voltage on the sector plates that selects 1-keV C1+ ions will also select 2-keV C2+ ions, 3-keV C3+ ions and so forth. The charge states with a given E/q are separated by time-of-flight, since they travel down the drift tube at different velocities (kinetic energies). Therefore, the signal detected on the MCP at the end of the analyzer gives peaks at different times, one for each charge state, as shown in figure 1.
The charge-separated ion energy distribution is extracted by converting the
measured electron current from the MCP to an ion current (incorporating the MCP
gain and efficiency), and combining the set of time-of-flight spectra by charge
state. This ion energy distribution is an
average distribution per laser pulse and can be analyzed for dependencies on
laser parameters.
RESULTS
& ANALYSIS top
Figure 2 shows the charge-separated ion energy distributions for carbon ablation at a laser intensity of 1015 W/cm2. Also shown is the total (charge-averaged) ion current, as constructed from the charge-separated data, and as it would appear in a non-ion-specific time-of-flight measurement. As expected, the C1+ to C4+ signals are dominant since the 5th and 6th electrons in carbon have much higher ionization energies due to their being tightly bound in the filled 1s2 energy shell. The charge-separated spectra show two main features – distributions at low and high energy, both of whose positions depend on charge state. The charge-averaged spectrum shows the high-energy feature but hides specifics of the ion response resulting from dynamic plasma processes. The following analysis examines the dependence of these distributions on laser contrast ratio and investigates the origin of the low- and high-energy distributions.
Figure 3 shows the strong dependence of laser contrast on the C2+ ion energy distribution at a laser intensity of 1015 W/cm2 (the 3rd-order auto-correlator contrast measurements are shown in the inset). For a contrast of 105 and intensity of 1015 W/cm2, the ns-long background is at 1010 W/cm2, which is just above the threshold for plasma generation. The result is that the 100 fs main pulse interacts with a pre-formed plasma instead of the solid target. An explanation for the absence of the high-energy component under low pulse contrast will be discussed later, when its origin is discussed. This observation suggests a method of tuning the ion energy distribution in a useful and controllable way, which is of interest for ion acceleration schemes and thin-film deposition.
In order to analyze the observed ion energy distributions in relation to the hydrodynamic expansion of the laser-produced plasma, a one-dimensional numerical simulation was used []. The simulation is based on an average-atom, two-fluid description of the plasma with a one temperature, local-thermal-equilibrium model. It solves the Helmholtz equation coupled with the fluid equations to give a self-consistent description of the ultrashort laser-plasma interaction.
The simulation provides good insight into the physics of the plasma expansion process; however, one must be careful in interpreting the simulation relative to the observed ion energy distribution. The simulation describes the initial one-dimensional hydrodynamic expansion, but cannot describe the subsequent three-dimensional adiabatic expansion of the real plasma beyond a few microns from the surface. Whereas the simulation allows the plasma to thermalize, the adiabatic expansion of the real plasma leads to a "frozen ionization" or persistence of high charge state ions at long times and distances from the target []. In the simulation, this frozen ionization can be obtained by choosing a cut-off time at which the hydrodynamic expansion is superseded by adiabatic expansion of the plasma. Such a condition is observed when the average ionization vs. energy takes on a quasi-steady state in time, and when the velocity vs. distance relationship of the ions becomes monotonic and approximately linear. The ion energy spectrum at this cut-off time will effectively represent the average ionization state that persists in the expansion and is detected at the analyzer. Experimentally, this frozen ionization condition is considered to occur at a distance from the sample surface that is on the order of the laser spot radius [14].
Once the frozen state is chosen from the hydrodynamic simulation, the simulated ion energy distributions can be constructed using the experimental conditions and compared to the measured distribution, as shown in figure 4. Since the simulation is an average-atom model, only charge-averaged distributions can be compared. The measured (charge-averaged) ion energy distribution was calculated using a semi-empirical MCP efficiency curve and is for a laser intensity of 1015 W/cm2, at 106 contrast. Both curves were scaled for comparison. Since the efficiency for the MCP at low energy (below ~600 eV) is not particularly well known, there is some uncertainty in the exact position of the low-energy peak of the experimental charge-separated distributions and the associated charge-averaged distributions. The comparison shows that the hydrodynamic simulation predicts the existence and approximate shape of the low energy peak. These are the same ions that survive in the low contrast condition of figure 3.
On the other hand, the hydrodynamic simulation cannot explain the existence of the high-energy portion of the average ion distribution in figure 4. In addition, the observation in figure 2 that the position of the high-energy peak for the higher charge states occurs at increasingly higher energy qualitatively implies an electromagnetic acceleration process. A mechanism that could account for this ion acceleration is the rapid formation and expansion of a space-charge layer of electrons which creates a time-dependent ambipolar field [1,2,3,4]. This layer is formed by the escape of a discrete population of high-energy electrons from the main plasma plume. These electrons do not undergo significant energy-changing collisions and, by virtue of their escape from the plasma, set up a space-charge layer that accelerates a fraction of the plasma ions. The result is that ions of different charge states will be accelerated to different energies with higher charge states reaching higher energies.
A simple model for ion acceleration treats
the fast electrons as one side of an effective capacitor. Assuming there are Ne
fast electrons in a region with transverse area A at an average distance
d from the surface, the electric field of such a capacitor is eNed/e 0A, where e is the electronic charge and e 0
is the free-space permittivity. An ion starting at the surface of the sample
and accelerating in this electric field will receive an increase in kinetic
energy D E of Zie2Ned/e 0A where Zi is the ionization state
of the ion. Assuming the transverse area of the electron cloud is approximately
the laser spot size (p s2/4) and rearranging, the
relation becomes:
The position of the high-energy peaks in figure 2 can be closely reproduced
with Einitial = 0.60 keV and D E = 0.67Zi
keV in agreement with this simple acceleration model. For the experimental conditions
of figure 2, laser spot size s = 100 m m and D E =
0.67 keV for C1+ (Zi = 1), the product Ned
is 2.91 x 108 m m. The origin of Einitial may be
considered to be due to the transfer of energy from the electrons to the ions
before the acceleration process. For such high-density plasmas, the
electron-ion collision frequency is high enough to allow such an energy
transfer during the early part of the laser pulse. This elementary analysis is
consistent with our observed data; however, more sophisticated models will be
required for a more thorough analysis.
Contrast
Effects top
As previously noted, it is observed that
laser pulse contrast has an important effect on the characteristics of the
laser-generated ions. Figure 5 shows the measured charge-averaged ion energy
distributions for a laser intensity of 2 x 1015 W/cm2 at
four laser conditions: a) low contrast (105 peak-background ratio),
b) low contrast with ultrashort pre-pulse 9 ns before and with 1/10 of the
energy of the main pulse, c) high contrast (better than 106
peak-background ratio), d) high contrast with pre-pulse (same pre-pulse as in
condition b). For low and high contrast (cases a and c), we see the same
comparison noted in the C2+ spectra in figure 3; the low contrast
condition is missing the high-energy, accelerated component. For the high
contrast case with and without pre-pulse (cases c and d), we see that each
contains the hydrodynamic and accelerated components. However, in the pre-pulse
case, the ion energies have been shifted to lower energies (for both
components) and the signal levels have been reduced with the hydrodynamic
component suffering the most. A similar trend occurs in the low contrast case
with and without pre-pulse.
In the high-contrast case without pre-pulse,
the plasma is formed by an ultrashort pulse which creates a steep density
gradient and an out-flow of fast electrons. The fast electrons accelerate the
ions to energies in the high-energy component, and the hydrodynamic expansion
of the plasma creates the low-energy component in the distribution. With a
pre-pulse, a low-density pre-plasma begins to form 9 ns before the main pulse
which creates an absorbing medium for the main pulse resulting in a much
shallower plasma density gradient. The fast electrons still accelerate the ions
during the main pulse, but not from as steep a gradient as in the non-pre-pulse
case. The result is that the ion energies shift to lower energy as observed.
In the low-contrast case, the plasma is
formed by a low-intensity ns-long background that allows the plasma to
gradually expand before the main 100 fs pulse arrives. The pulse interacts with
a fully-formed plasma that has different reflectivity and absorption properties
than the solid case. In addition, the density gradient is shallow and any fast
electrons will form in this gradient. An acceleration process separate from the
hydrodynamic expansion of the plasma is not possible under such conditions, as
is indicated in the spectra by the absence of the high-energy component.
Figures 6 and 7 show the dependence on the
laser intensity of the charge-averaged and C3+ ion energy
distributions, respectively, for a high-contrast condition. For increasing
laser intensity, the charge-averaged distributions increase in magnitude, shift
to higher energies, broaden, and exhibit an increasing proportion of ions in
the hydrodynamic component to those in the accelerated component. In fact, for
the lowest laser intensity shown, the distribution is primarily composed of
accelerated ions. In order to further interpret these observations, the
charge-separated spectra can be examined. For example, the C3+
spectrum, as show in figure 7, has no hydrodynamic component for the lowest
laser intensity. As the intensity becomes higher, this component also
increases. For the laser intensities of 5 x 1014 and 1 x 1015
W/cm2, the shift in the C3+ accelerated component peak
(fig. 7) is not as large as the shift in the
charge-averaged accelerated peak (fig. 6). This example demonstrates one way
that the charge-averaged distribution can hide specifics of the ions’
dependence on laser parameters. The differing effects of laser intensity on
each charge state’s distribution may become important in producing specific
high-energy ion beams.
CONCLUSION top
The results from this set of experiments show that the size, shape, and charge states of the ion energy distributions in the ablation plume depend strongly on the laser intensity and even more importantly on laser pulse contrast. The presence of a pre-pulse can also modify the distributions. The observed bi-modal energy distributions are identified as coming from hydrodynamic expansion for low-energy ions (few hundred eV) and ion acceleration for the high-energy ions (several keV).
ACKNOWLEDGEMENTS
top
This work was supported in part by the National Science Foundation through the Center for Ultrafast Optical Science under grant STC PHY 8920108, as well as AFOSR equipment grant F49620-95-1-0474. We acknowledge X-Ray Specialty Instruments (Ann Arbor, MI) for loan of equipment and Spectrogon for compressor gratings. One of the authors (MN) acknowledges support from Fonds pour la Formation des Chercheurs et l’ Aide à la Recherche (FCAR, Quebec).
FIGURE
CAPTIONS top
Figure 1: A typical time-of-flight ion spectrum taken for a 100-fs, 1015-W/cm2 laser pulse with a 725-V difference on the analyzer sector plates. The spectrum shown is an average of 256 scans. The signal is read as peaks in negative voltage (since the MCP produces an electron current). The small oscillatory component after each peak is due to electronic ringing.
Figure 2: Ion energy distributions separated by charge state and charged-averaged for carbon ablation with 100-fs laser pulses at an intensity of 1015-W/cm2.
Figure 3: Ion energy distribution of C2+ for laser-contrast ratios (peak-to-background) of 105 and 106 at a laser intensity of 1015 W/cm2. The inset shows the measured contrast for the corresponding distributions.
Figure 4: Comparison between hydrodynamic simulation results for the experimental conditions and the measured charge-averaged ion energy distribution.
Figure 5: Measured charge-averaged ion energy distributions for a laser intensity of 2 x 1015 W/cm2 at four laser conditions with high/low contrast (106/105) and with or without a pre-pulse (ultrashort pulse with 1/10 the laser energy at 9 ns before main pulse).
Figure 6: Measured charge-averaged ion energy distributions for three laser intensities at high contrast (better than 106 peak-to-background ratio).
Figure 7: Measured C3+ ion energy distributions for three laser intensities at high contrast (better then 106 peak-to-background ratio).