physics0006068/ics5.tex
1: \ProvidesFile{esapub.tex}[1999/12/02 1.01 (PWD)]
2: \documentclass[a4paper,twocolumn]{esapub} % European paper
3: %\documentclass[letterpaper,twocolumn]{esapub} % American paper
4: 
5: %\usepackage{times}
6: \usepackage{natbib}
7: \usepackage{graphicx}
8: 
9: \title{Enhanced Pitch Angle Diffusion due to Electron-Whistler 
10: Interactions during Disturbed Times}
11: \author{W. Wykes}
12: \author{S. C. Chapman}
13: \author{G. Rowlands}
14: \affil{Space and Astrophysics Group, University of Warwick, Coventry, UK}
15: 
16: 
17: \newcommand{\btx}{\textsc{Bib}\TeX}
18: \newcommand{\vperp}{v_{\perp}}
19: \newcommand{\vpar}{v_{\parallel}}
20: \newcommand{\dpsi}{\Delta \psi}
21: \newcommand{\etal}{{\em et al.}}
22: \newcommand{\back}{\!\!\!\!\!}
23: \newcommand{\filename}{esapub}
24: 
25: \begin{document}
26: \keywords{Electron, Whistler, Substorms, Diffusion}
27: 
28: \maketitle
29: 
30: \begin{abstract}
31: During periods of increased magnetospheric activity, whistler wave emissions
32: have been observed with increased wave amplitudes. We consider a pitch angle
33: diffusion mechanism that is shown to scale with whistler wave amplitude and 
34: hence is 'switched on' during these periods of intense activity.
35: 
36: We consider the interaction between relativistic electrons and two oppositely 
37: directed, parallel propagating whistler mode waves. We show that for intense 
38: whistlers this process is stochastic and results in strong electron pitch 
39: angle diffusion.
40: 
41: We show that the interaction is rapid, occur on timescales of the order of 
42: tens of electron gyroperiods and that the interaction is sensitive to wave 
43: amplitude, wave frequency and electron energy.
44: 
45: \end{abstract}
46: 
47: \section{Introduction}
48: 
49: Electron whistler interactions have long been considered as a mechanism
50: for pitch angle scattering. Gyroresonance processes with near-parallel 
51: propagating waves (e.g. \citet{kennel}, \citet{lyons:84}) have been shown to 
52: produce pitch angle diffusion for electrons that are at resonance with a 
53: series of whistler waves (see \citet{gendrin}). We generalize resonant 
54: diffusion to include all phase space dynamics, i.e. as well as considering 
55: the resonant diffusion of trapped electrons we consider diffusion of 
56: untrapped electrons (we refer to this process as `off-resonance' diffusion). 
57: Therefore we maximize the area in phase space contributing to pitch angle 
58: diffusion.
59: 
60: The underlying dynamics of the interaction between electrons and a single 
61: whistler are inherently simple, as demonstrated by \citet{laird} who derived 
62: the Hamiltonian for relativistic electrons interacting with a whistler with a 
63: single wave number. However, for a single monochromatic whistler the process 
64: is not stochastic. We introduce stochasticity by including an identical, 
65: oppositely directed whistler mode wave.
66: 
67: We initially consider a simplified system consisting of monochromatic 
68: whistlers in order to understand the underlying behaviour. This treatment is 
69: then extended to consider whistler wave packets, i.e. a group of waves with a 
70: range of frequencies.
71: 
72: We derive approximate equations in the limit of low whistler wave amplitudes 
73: and consider the degree of pitch angle diffusion for waves of different 
74: frequencies and bandwidths and for electrons of different energies.
75: 
76: \section{Equations of Motion}
77: 
78: We derive full relativistic equations of motion and approximate them in the 
79: limit of low wave amplitudes, for the case of monochromatic whistlers and 
80: wave packets. We normalize time to the electron gyrofrequency 
81: $\Omega_e=eB_0/m_e$, (where $B_0$ is the background magnetic field), the 
82: wave amplitude is normalized to the background magnetic field, wave frequency
83: is normalized to the gyrofrequency and we normalize the electron velocity to 
84: the phase velocity of the waves, given by the electron dispersion relation 
85: (ignoring ion effects):
86: \
87: \begin{eqnarray}
88: \frac{k^2c^2}{\omega^2}=1-\frac{\omega_{pe}^2}{\omega(\omega-\Omega_e)} 
89: \label{eq:disp}
90: \end{eqnarray}
91: \
92: where $\omega_{pe}$ is the plasma oscillation frequency. Electrons can become 
93: trapped by either of the two whistlers if they are traveling close to the 
94: resonance velocity, given by:
95: \
96: \begin{eqnarray}
97: \omega-{\bf k} \cdot {\bf v_r}=n \Omega_e/\gamma \label{eq:res}
98: \end{eqnarray}
99: \
100: where $n$ is an integer and $\gamma = (1-v^2/c^2)^{-1/2}$ is the relativistic 
101: factor.
102: 
103: \begin{figure*}[phtb]
104: \centering
105: %\includegraphics[width=1.0\linewidth]{dcircle.ps}
106: \includegraphics[width=1.0\linewidth]{ics5_mono.ps}
107: \caption{\label{fig:mono} Stroboscopic surface of section plots for the 
108: monochromatic whistler interaction. For quiet time wave amplitudes, panel a), 
109: all trajectories are regular, with resonances given by the resonance condition
110: Equation~(\ref{eq:res}). For high wave amplitudes observed during disturbed 
111: times, panels b), phase space is dominated by stochastic trajectories with 
112: regular trajectories confined to close to the resonances. The stochastic 
113: region is bounded above and below by the first regular, untrapped, 
114: trajectories away from resonance, therefore there is a limit on the diffusion 
115: of electrons in phase space.}
116: \end{figure*}
117: 
118: \subsection{Monochromatic Whistlers}
119: 
120: In the limit of low wave amplitudes the full equations of motion can be 
121: reduced to:
122: \
123: \begin{eqnarray}
124: \frac{d^2 x}{d t^2} &=& \frac{2b {\vperp}_0}{\gamma_0}
125: \sin\left[(1/\gamma_0-\omega)t\right] 
126: \cos\left[kx\right]\label{eq:red}
127: \end{eqnarray}
128: \
129: where $x$ is the distance along the background magnetic field, $b$ is the
130: normalized wave amplitude, ${\vperp}_0$ is the initial perpendicular velocity 
131: and $\gamma_0=1/\sqrt{(1-{\vperp}_0^2/c^2)}$.
132: 
133: \subsection{Wave Packet Approximation}
134: 
135: Instead of a pair of waves it is more realistic to consider the interaction 
136: of a wave packet, ie a small group of waves with a range of wave frequencies
137: and wave numbers. We assume the wave amplitude is non-zero over wave 
138: frequency range $\Delta\omega$:
139: \
140: \begin{eqnarray}
141: b(\omega)=\left\{ \begin{array}{r@{\quad:\quad}l}
142:      0 & \omega<\omega_0-\Delta\omega/2 \\
143:      b & \omega_0-\Delta\omega/2<\omega<\omega_0+\Delta\omega/2 \\
144:      0 & \omega>\omega_0+\Delta\omega
145:      \end{array} \right. \label{eq:that}
146: \end{eqnarray}
147: \
148: where $\omega_0$ is the central wave frequency of the wave packet. Integrating
149: the monochromatic ehistler equation~(\ref{eq:red}) over the frequency 
150: range $\Delta\omega$ gives the following wave packet equation:
151: \
152: \vspace{1cm}
153: 
154: \begin{eqnarray}
155: \!\!\!\!\!\!\!\!\frac{d^2 x}{d t^2} \!\!\!\!\! &=& \!\!\!\!\!
156: \Omega_e \!\!\frac{2bv_{\perp 0}}{\gamma_0}\!\!
157: \sin\!\!\left[\!(\!\frac{1}{\gamma_0}\!-\!\omega)t \!+\! k_0 x \!\right]\!\!
158: \frac{\sin\!\left[\!(t\!-\!\beta x)\!\Delta\omega\!/\!2\right]}
159: {(t-\beta x)} \nonumber \\
160: &+& \!\!\!\!\!
161: \Omega_e \!\!\frac{2bv_{\perp 0}}{\gamma_0}\!\!
162: \sin\!\!\left[\!(\!\frac{1}{\gamma_0}\!-\!\omega)t \!-\! k_0 x \!\right]\!\!
163: \frac{\sin\!\left[\!(t\!+\!\beta x)\!\Delta\omega\!/\!2\right]}
164: {(t+\beta x)} \label{eq:packet}
165: \end{eqnarray}
166: \
167: where $1/\beta=d\omega/dk$ is the group velocity of the waves. The wave packet
168: equation~(\ref{eq:packet}) yields the monochromatic whistler 
169: equation~(\ref{eq:red}) in the limit $\Delta\omega \rightarrow 0$, with 
170: amplitude $b^{\prime}=b\Omega_e\Delta\omega$. 
171: 
172: \section{Numerical Results}
173: 
174: %The monochromatic and wave packet equations were solved numerically using a 
175: %variable order, variable stepsize differential equation integrator. We 
176: %consider physical parameters for the terrestrial magnetosphere at $L=6$: 
177: %gyrofrequency, $\Omega_e=3.2~kHz$, plasma frequency, $\omega_{pe}=56~kHz$, 
178: %background magnetic field, $B_0=18~nT$ and wave amplitude 
179: %$B_{\omega}=20~pT$, giving a normalized wave amplitude, $b=0.001$ 
180: %(see \citet{nagano}, \citet{parrot} and \citet{summers}).
181: 
182: The monochromatic and wave packet equations were solved numerically using a 
183: variable order, variable stepsize differential equation integrator. We 
184: consider physical parameters for the terrestrial magnetosphere at $L=6$: 
185: gyrofrequency, $\Omega_e=25.3~kHz$, plasma frequency, $\omega_{pe}=184~kHz$, 
186: background magnetic field, $B_0=144~nT$ and wave amplitude 
187: $B_{\omega}=0.5~pT$, giving a normalized wave amplitude, consistent with 
188: quiet times in the terrestrial magnetosphere, $b=10^{-5}$ (see for example 
189: \citet{nagano}, \citet{parrot} and \citet{summers}).
190: 
191: The phase plots in Figure~\ref{fig:mono} are comprised of stroboscopic 
192: surfaces of section \citep{bene} to sample the full electron phase space. The 
193: initial parallel velocity was varied over the range $[-v_r,v_r]$, where $v_r$ 
194: is the resonance velocity, given by the resonance condition, (\ref{eq:res}), 
195: for $n=1$. All electrons were given a constant initial perpendicular velocity,
196: with $\vperp \approx 20 v_r$ as it was found that a high velocity anisotropy 
197: was required for stochasticity. 
198: 
199: In Figure~\ref{fig:mono} we plot parallel velocity $\vpar=dx/dt$ against phase
200: angle $\psi$, where $\psi$ is the angle between whistler propagating in a 
201: positive direction along the background field and the perpendicular velocity,
202: $\vperp$. In panel a) we consider a whistler wave amplitude consistent with
203: quiet times $(b=10^{-5})$ all trajectoriesin phase space are regular. There 
204: is little change in $\vpar$ and hence only weak pitch angle diffusion. As the 
205: wave amplitude is increased, stochastic trajectories are introduced, as the 
206: regular trajectories between the two resonances are progressively broken down.
207: In panel b) we consider the case of intense whistler wave activity during
208: substorms ($b=10^{-3}$, see for example \citet{parrot} and \citet{nagano}).
209: The stochastic region grows to encompass the resonances as the wave amplitude 
210: is increased. Regular trajectories are confined to KAM surfaces 
211: (near-integrable trajectories with an approximate constant of the motion 
212: \citep{tabor}). The stochastic region is bounded by the first untrapped 
213: (regular) trajectories away from the resonances, thus there is a limit on 
214: diffusion in phase space. 
215: 
216: As well as resonant diffusion of trapped electrons, there is diffusion 
217: of untrapped electrons throughout the stochastic region of phase space. Since,
218: for sufficient wave amplitudes, the stochastic region can encompass the 
219: resonances, the diffusion of untrapped electrons, which we refer to as 
220: `off-resonance' diffusion, may be enhanced over resonant diffusion. In 
221: addition we achieve pitch angle diffusion from a larger area of phase 
222: space.
223: 
224: \begin{figure}[phtb]
225: \centering
226: \includegraphics[width=1.0\linewidth]{ics5_nw.ps}
227: \caption{\label{fig:nw} Change in pitch angle from an initial pitch angle of 
228: $90^{\circ}$, for quiet time wave amplitudes ($b=10^{-5}$) and narrow and 
229: wide whistler wave packets. For narrow wave packets there is little change in 
230: pitch angle. For wide wave packets there is a large change in pitch angle 
231: ($\Delta\alpha\sim25^{\circ}$) occuring within a few tens of electron 
232: gyroperiods ($\sim10~ms$), hence the interaction is rapid. Changes in pitch 
233: angle attenuate with time and the pitch angle reaches a constant value.}
234: \end{figure}
235: 
236: Due to the time dependent nature of the wave packet equation~(\ref{eq:packet})
237: it is not possible to construct phase diagrams as in Figure~\ref{fig:mono} for
238: the monochromatic whistler case. Instead we can consider the dynamics of 
239: single electrons. In Figure~\ref{fig:nw} we show a single trajectory solution 
240: of the wave packet equation~(\ref{eq:packet}), for quiet time wave amplitudes 
241: and wide ($\Delta\omega=\Omega_e/50=500~Hz$) and narrow 
242: ($\Delta\omega=\Omega_e/500=50~Hz$) whistler wave packets (see for example 
243: \citet{carpenter}).
244: We consider the change in pitch angle from an initial pitch angle of 
245: $90^{\circ}$. For narrow wave packets there is little change in pitch angle 
246: and the trajectory is regular. For wide wave packets the trajectory is 
247: stochastic with a large change in pitch angle ($\Delta\alpha\sim25^{\circ}$) 
248: occuring within a few tens of electron gyroperiods ($\sim10~ms$). We can now
249: achieve strong pitch angle diffusion for wave amplitudes consistent with the 
250: quiet time magnetosphere.
251: 
252: \begin{figure}[phtb]
253: \centering
254: %\includegraphics[width=1.0\linewidth]{circle.ps}
255: \includegraphics[width=1.0\linewidth]{ics5_10.ps}
256: \caption{\label{fig:10} Log change in pitch angle, 
257: $(\log_{10}|1+\Delta\alpha|)$, as a function of wave 
258: frequency, $\omega$, and initial pitch angle, $\alpha_0$,  for 10~keV 
259: electrons ($v= 0.2c$) and quiet time (low amplitude),wide band whistler wave 
260: packets ($\Delta\omega=\Omega_e/50$). For high to moderate initial pitch 
261: angles ($\alpha_0=50^{\circ}-90^{\circ}$) there is a change in pitch angle of 
262: up to $40^{\circ}$. For low pitch angles ($\alpha_0=5^{\circ}-10^{\circ}$) 
263: the change in pitch angle is of the order of a few degrees.}
264: \end{figure}
265: 
266: \section{Pitch Angle Scattering}
267: 
268: Using the wave packet equation~(\ref{eq:packet}) with quiet time whistlers
269: ($b=10^{-5}$) and a relatively wide band whistler 
270: ($\Delta\omega=\Omega_e/50=500~Hz$, \citep{carpenter}), we can estimate the 
271: degree of pitch angle 
272: scattering. In Figures \ref{fig:10} and \ref{fig:100} we estimate the  log 
273: change in pitch angle, $(\log_{10}|1+\Delta\alpha|)$, as a function of wave 
274: frequency, $\omega$, and initial pitch angle, $\alpha_0$. We consider the 
275: interaction between $10 keV$ electrons in Figure~\ref{fig:10} ($100 keV$ 
276: electrons in Figure~\ref{fig:100}) and wide band whistlers 
277: ($\Delta\omega=\omega_e/50$). For high to moderate initial pitch angles 
278: ($\alpha_0=50^{\circ}-90^{\circ}$) there is a change in pitch angle of up to 
279: $40^{\circ}$. For low pitch angles ($\alpha_0=5^{\circ}-10^{\circ}$) the 
280: change in pitch angle is of the order of a few degrees. In 
281: Figure~\ref{fig:100} we see a similar degree of diffusion except that lower 
282: frequency whistler wave packets are required.
283: 
284: 
285: \begin{figure}[phtb]
286: \centering
287: %\includegraphics[width=1.0\linewidth]{circle.ps}
288: \includegraphics[width=1.0\linewidth]{ics5_100.ps}
289: \caption{\label{fig:100} As for Figure \ref{fig:10} except for 100~keV 
290: electrons ($v=0.5c$). Pitch angle scattering from a few degrees up to 
291: $40^{\circ}$ occurs although lower frequency whistler wave packets are 
292: required.}
293: \end{figure}
294: 
295: \section{Discussion}
296: 
297: We have considered electron-whistler wave particle interactions to investigate
298: diffusion over all phase space, to include both resonant and `off-resonance'
299: diffusion. We have considered a simplified interaction with monochromatic 
300: whistler wave to understand the underlying behaviour and have shown that the 
301: presence of the second whistler wave introduces stochastic effects into the 
302: system. For wave amplitudes consistent with disturbed times we have shown 
303: that `off-resonance' diffusion occurs and that resonant diffusion is unchanged.
304: 
305: We have considered a more realistic case of whistler wave packets and have 
306: shown that for relatively wide band whistler wave packets strong pitch angle 
307: diffusion occurs for wave amplitudes consistent with quiet, undisturbed, 
308: times. For high initial pitch angles we estimate a change in pitch angle of up
309: to $40^{\circ}$, while for low pitch angles a change of a few degrees is 
310: estimated.
311: 
312: The effectiveness
313: in scattering electrons of different energies is dependent on the wave 
314: frequency. Electrons with low energies ($10~keV$) are readily scattered by 
315: waves of around half the electron gyrofrequency, while electrons at higher 
316: energies ($100~keV$) are scattered by lower frequency wave 
317: ($\omega\sim\Omega_e/10$). $MeV$ electrons would require extremely low 
318: frequency waves for efficient scattering, hence our mechanism is most 
319: efficient for electrons in the $10-100~keV$ range.
320: 
321: \section*{Acknowledgements}
322: The authors would like to acknowledge PPARC for the funding of this work.
323: 
324: \begin{thebibliography}{}
325: 
326: \bibitem[Benettin \etal(1976)]{bene} 
327: Benettin G., L. Galgani, J. M. Strelcyn.
328: Kolmogorov entropy and numerical experiments. {\em Physica A, 14, 2338, 1976.}
329: 
330: \bibitem[Carpenter and Sulic(1988)]{carpenter}
331: Carpenter D. L., and D. M. Sulic.
332: Ducted whistler propagation outside the plasmapause.
333: {\em J. Geophys. Res., 93, 9731, 1988.}
334: 
335: \bibitem[Gendrin(1981)]{gendrin} 
336: Gendrin R. General Relations between wave amplification and particle 
337: diffusion in a magnetoplasma. {\em Rev. Geophys., 19, 171, 1981.}
338: 
339: \bibitem[Kennel and Petschek(1966)]{kennel} 
340: Kennel C. F. and H. E. Petschek.
341: Limit on stably trapped particle fluxes. {\em J. Geophys. Res., 71, 1, 1966.}
342: 
343: \bibitem[Laird(1972)]{laird} 
344: Laird M. J. Cyclotron resonance in an inhomogeneous plamsa.
345: {\em J. Plasma Physics, 8, 255, 1972.}
346: 
347: \bibitem[Lyons and Williams(1984)]{lyons:84} 
348: Lyon L. R. and Williams, D. J. Quantitative Aspects of Magnetospheric Physics. 
349: {\em D. Reidel, Hingham, Mass., 1984.}
350: 
351: \bibitem[Nagano \etal(1996)]{nagano}
352: Nagano I., S. Yagitani, H. Kojima, H. Matsumoto. Analysis of Wave Normal and 
353: Poynting Vectors of the Chorus Emissions Observed by GEOTAIL. 
354: {\em J. Geomag. Geoelectr., 48, 299, 1996}
355: 
356: \bibitem[Parrot(1994)]{parrot} 
357: Parrot M. and C. A. Gaye. A statistical survey of ELF waves in geostationary 
358: orbit. {\em GRL 21, 2463, 1994.}
359: 
360: \bibitem[Summers and Ma(2000)]{summers}
361: Summers D. and C. Ma. A model for generating relativistic electrons in the 
362: Earth's inner magnetosphere based on gyroresonant wave-particle interactions. 
363: {\em J. Geophys. Res., 105, 2625, 2000.}
364: 
365: \bibitem[Tabor(1989)]{tabor} 
366: Tabor M. Chaos and integrability in nonlinear dynamics - an introduction. 
367: {\em New York, Chichester: Wiley, 1989.}
368: 
369: \end{thebibliography}
370: \end{document}
371: 
372: