1: \documentstyle[pre,aps,epsf]{revtex}
2: %\documentstyle[11pt,epsf]{article}
3: %\textwidth=15cm
4: %\textheight=23cm
5: %\hoffset=-2.0cm
6: %\voffset=-2.0cm
7:
8: \begin{document}
9: \def\de{d\dot{\epsilon}}
10: \def\edot{\dot{\epsilon_p}}
11: \title{Relaxation oscillations and negative strain rate
12: sensitivity in the Portevin - Le Chatelier
13: effect}
14:
15: \author{S. Rajesh$^1$ and G. Ananthakrishna$^{1,2}${\thanks{E-mail:garani@mrc.iisc.ernet.in}}}
16: \address{$^1$Materials Research Center,\\
17: Indian Institute of Science,\\
18: Bangalore - 560012.\\
19: India.\\
20: $^2$Center for Condensed Matter Theory\\
21: Indian Institute of Science\\
22: Bangalore - 560012\\
23: India.}
24: \maketitle
25:
26:
27: \begin{abstract}
28: \noindent
29: A characteristic feature of the Portevin - Le Chatelier effect or the
30: jerky flow is the stick-slip nature of stress-strain curves
31: which is believed to result from the negative strain rate dependence
32: of the flow stress. The latter is assumed to result from the
33: competition of a few relevant time scales controlling the dynamics of jerky
34: flow. We address the issue of time scales and its connection to
35: the negative strain rate sensitivity of the flow stress within the
36: framework of a model for the jerky flow which is
37: known to reproduce several experimentally observed features
38: including the negative strain rate sensitivity of the flow
39: stress. We attempt to understand the above issues by
40: analyzing the geometry of the slow manifold underlying
41: the relaxational oscillations in the model. We show
42: that the nature of the relaxational oscillations is a result
43: of the atypical bent geometry of the
44: slow manifold. The analysis of the slow manifold structure
45: helps us to understand the
46: time scales operating in different regions of the slow manifold.
47: Using this information
48: we are able to establish connection with the strain rate
49: sensitivity of the flow stress. The analysis also helps us to
50: provide a proper dynamical interpretation for the negative
51: branch of the strain rate sensitivity.\\
52:
53: \noindent
54: PACS number(s): 83.50.-v, 83.50.By, 81, 05.45.-a, 62.20.-x
55: \end{abstract}
56:
57:
58: %\twocolumn
59: \section{ Introduction}
60:
61: \noindent
62: The Portevin - Le Chatelier effect \cite{portevin} or the jerky
63: flow has been an object of continued
64: interest in materials science for quite some time. The phenomenon
65: refers to an instability seen in the form of
66: repeated stress drops followed by periods of reloading observed
67: when tensile specimens are deformed in a certain range of
68: strain rates and temperatures \cite{cot53}. The effect
69: is seen in many interstitial and substitutional metallic alloys
70: (commercial aluminium, brass, alloys of aluminium and
71: magnesium \cite{bri70}).
72: Each of the load drops is related to the formation and propagation
73: of dislocation bands \cite{bri70,kub93}.
74: The traditional picture of the instability is that it stems from
75: dynamic interaction of mobile dislocations with solute atoms and is called
76: as dynamic strain ageing \cite{cot53}. It is this that is
77: expected to lead to
78: negative strain rate sensitivity (SRS) of the flow stress \cite{bri70,kub93,mcc72,van75,kub85}.\\
79:
80: \noindent
81: Plastic flow is intrinsically nonlinear and therefore methods of
82: nonlinear dynamics have a natural role to play in understanding plastic
83: instabilities \cite{kub93,ana82,NLI,NLII,NLIII}.
84: Use of these new techniques have
85: led to insights which were hitherto
86: not possible.
87: The first attempt to look at the phenomenon
88: from a nonlinear dynamical angle was taken by Ananthakrishna
89: and coworkers\cite{ana82}, which
90: offers a natural basis for the description of the time dependent
91: aspects of the PLC effect which were ignored in the earlier theories \cite{mcc72,van75,kub85}.
92: Their model allows for explicit inclusion and interplay of different
93: time scales inherent in the dynamics of dislocations.
94: These authors show that the
95: occurrence of the instability is a consequence of Hopf bifurcation as a
96: function of the applied strain rate. Many known features
97: of the PLC effect such as the existence of a window of strain rates
98: and temperatures within which it occurs, etc.,
99: were correctly reproduced.
100: Most importantly, and for the first time, the
101: {\it negative SRS was shown to emerge naturally} in the model,
102: as a result of nonlinear interaction of the
103: participating defects. It
104: also predicts the existence of
105: chaotic stress drops in a range of strain rates,
106: which has been recently verified \cite{ana95a,ana95b,nor96,nor97}.
107: Even the number of degrees of freedom
108: estimated turn out to be the same as in the model offering justification for
109: ignoring spatial degrees of freedom (See Ref. 16 also).
110: Further dynamical analysis of the model
111: for the creep
112: case has shown that the temperature dependence of the strain bursts is
113: consistent with experimental findings\cite{mul97,mul98}.\\
114:
115:
116: \noindent
117: The study of the PLC effect from a dynamical angle has
118: been useful in elucidating several features; but it has
119: also brought certain other issues into sharp focus which
120: were hitherto not investigated in depth.
121: This paper is intended to address one such issue related to the
122: time scales relevant to the dynamics of the PLC effect within the
123: framework of the above model. This is reflected in the two well
124: known attributes of the PLC effect, namely, the
125: negative strain rate behavior of the flow stress and the
126: stick-slip or relaxational nature of the dynamics reflected
127: in the stress time series.
128: In order to motivate, we will present arguments
129: showing that conflicting conclusions can be arrived at when one
130: analyses this question starting from these two angles.\\
131:
132: \noindent
133: We start with a discussion of the well
134: accepted physical picture of the PLC effect namely the dynamic strain ageing.
135: At a qualitative level, theories of strain ageing
136: already have an implicit suggestion that the occurrence of the negative SRS
137: is related to the competition of diffusive time scale and the
138: waiting time of dislocations at obstacles \cite{mcc72,van75},
139: even though, there is no dynamics involved in these theories.
140: The physical picture of strain ageing is as follows.
141: At small velocities, solute atoms have enough
142: time to diffuse to the temporarily arrested dislocations thus
143: providing additional pinning thereby
144: impeding their breakaway from localized obstacles.
145: Due to the constant
146: applied strain rate, the overall stress
147: to keep the dislocations moving increases bringing the stress to a threshold level
148: beyond which dislocations break away\cite{mcc72,van75}. At high velocities,
149: such additional pinning due to solute atoms is not possible
150: since the waiting time of dislocations
151: at obstacles is too short for the diffusion to occur.
152: A schematic diagram of the SRS is shown in Fig. 1.
153: In the language
154: of the stick-slip dynamics, the branch $A^\prime B^\prime$ corresponds to
155: the stick-state and $B^\prime C^\prime$ to the slip-state (Fig. 1).
156: The slope of stress verses velocity ('friction coefficient')
157: at low velocities is much higher than that corresponding to high velocities
158: since in the former case, solute atoms
159: have to be dragged along with the dislocations, while
160: in the latter case there is no solute atmosphere.
161: Based on physical considerations,
162: these two stable branches are {\it assumed} to be
163: separated by {\it an unstable} branch with a negative slope to
164: reflect the nonaccessible nature.\\
165:
166: \noindent
167: The occurrence of the negative flow rate
168: characteristic is not just limited to the PLC effect\cite{mau85,tos95}.
169: With particular reference to the conceptual aspects of the
170: negative branch,
171: we cite two other mechanical systems, namely, the peeling of an adhesive tape and
172: frictional sliding of a block of
173: material over another\cite{mau85,tos95} which shows the inaccessible nature of the negative slope branch.
174: However, in the case of the PLC effect,
175: there has been attempts to obtain experimental
176: points in this domain of strain rates\cite{bod67,kub86}
177: which has lead to some confusion about the measurability
178: of the negative slope branch of the
179: SRS which will be discussed later (Section
180: 5).
181: Therefore, it
182: is important to understand the meaning of the negative branch of SRS from a dynamical point of view
183: with reference to the PLC effect which
184: hopefully will lead to a better understanding of other
185: stick-slip phenomenon.\\
186:
187: \noindent
188: Returning to the PLC effect, Penning \cite{pen72} was the first
189: to recognize that the negative SRS could be
190: used to explain the strain rate jumps observed in experiments.
191: Subsequently, the negative SRS feature has been used as an input into
192: several theories \cite{jea93,leb95}.
193: In particular, it has helped to successfully
194: explain the nature of yield drops occurring in different
195: regimes of strain rate and temperature \cite{leb95}. Pertinent to the
196: our discussion of time scales inherent to the PLC effect, we
197: note that in such theories,
198: {\it two slow time scales} corresponding to the two dissipative
199: branches, $A^\prime B^\prime$ and $C^\prime D^\prime$, show up
200: along with {\it two fast time scales} corresponding to the
201: jumps ( $B^\prime C^\prime$ and $D^\prime A^\prime$) in the strain rates. A more direct
202: reflection of the time scales inherent to the dynamics of the
203: PLC effect can be deduced from stress-strain curve.\\
204:
205:
206:
207: \noindent
208: To facilitate discussion of time scales involved in an experimental
209: stress-time series, consider the so called machine equation written as
210: \begin{equation}
211: \dot \sigma_a = \kappa[ \dot \epsilon_a -\dot \epsilon_p]
212: \end{equation}
213:
214: \noindent
215: where $\sigma_a,\dot{\epsilon}_a,\dot{\epsilon}_p$ refer to the stress, applied strain rate
216: and plastic strain rate respectively and $\kappa$ is the combined elastic
217: constant of the machine and the sample. We note that only stress
218: is monitored by the load sensing devise. Apart from this,
219: it is possible to measure the plastic strain rate using strain
220: gauges or using cinematographic
221: techniques \cite{neu93}.
222: Using Eq.(1), we
223: can now identify different time scales in an experimental curve.
224: Consider a typical stress-strain curve for an applied strain rate of
225: 8.3 x $ 10^{-5} s^{-1}$ for the PLC effect in $Cu-10\%Al$ is shown in Fig.2.
226: (For our purpose, we will ignore the nonperiodic nature.)
227: From the saw-tooth shape of the stress - strain series,
228: two points emerge: (a) the positive slope of $\sigma_a - \epsilon_a$
229: curve is close to the elastic loading rate ( $\kappa \dot
230: \epsilon_a$), and
231: (b) the duration of each stress drop is very short.
232: From Eq.(1), we see that the stress drop duration is the time interval
233: during which $\dot{\epsilon_p}(t)$ larger than $\dot{\epsilon_a}$.
234: We also note that the changes in slopes, when they occur, are
235: abrupt (within the recording accuracy of 0.05$ s^{-1}$).
236: Knowing that $\dot \epsilon_p$
237: is proportional to the mobile dislocation density and using Eq. (1), we can see
238: that the mobile dislocation density should be
239: nearly constant in the rising part of $\sigma - \epsilon_a$
240: curve and therefore correspond to the stick-state.
241: Further, the short duration of the stress drop should be a result of rapid multiplication of mobile
242: dislocation and therefore corresponds to the slip-state.
243: This must be followed by the process of immobilization of dislocations.
244: However, the abrupt change in the slope ( from negative to
245: positive) also implies
246: that immobilization time scale is also fast.
247: Indeed, as is clear from Fig. 2, it is not possible to separate
248: out these two fast time scales.
249: Thus, from the stress-strain curve, we see only
250: {\it one slow time scale} and {\it two fast time scales}
251: which is in apparent conflict with what was argued from the schematic diagram of the orbit $A^\prime B^\prime C^\prime D^\prime$ in Fig. 1.
252: This discussion raises several questions relating the origin of
253: these time scales causing jumps in dislocations densities which needs
254: to understood if the above inconsistency has to be resolved. Specifically:
255: (i) What is dynamical mechanism which keeps the mobile dislocation density
256: constant and in low levels for long intervals of time?
257: (ii) What are the mechanisms for rapid multiplication and immobilization of mobile
258: dislocations ?
259: As we shall see, resolving these issues
260: will also help us to interpret the negative SRS in an
261: appropriate way. Further, associating various
262: time scales with different branches of the SRS provides a better
263: insight into the stick-slip dynamics of the PLC
264: effect. \\
265:
266:
267:
268: \noindent
269: Analysis of time scales can be best understood from a dynamical
270: point of view. It is well known that relaxation oscillations are at
271: the root of stick-slip behavior.
272: One of the standard ways of understanding relaxation oscillations is by
273: analyzing the slow manifold geometry
274: \cite{arn93,kop92,bar88,hau96,den94} of the underlying model. Following this, we shall
275: attempt to understand
276: the above issues from the point of view of relaxation
277: oscillations.
278: The paper is organized as follows. In
279: section 2, we briefly introduce the model along with the known
280: results. In section 3, we state some bifurcation features
281: relevant for further discussion.
282: In section 4, we show that the nature of relaxation oscillation
283: in the model
284: is {\it atypical} and is due to the {\it atypical nature of bent nature of the slow
285: manifold} of the model. This analysis further helps us
286: to understand the dynamical basis of different time scales relevant to the PLC effect.
287: In section 5, we discuss the concept of
288: negative SRS and its measurement in some detail to highlight the
289: meaning of the negative branch.
290: Using the geometry of the slow manifold,
291: we calculate the dependence of stress on the plastic strain rate
292: and show the connection between the various branches of the SRS
293: and the time scales operating in different regions
294: of slow manifold which in turn helps us to resolve the
295: inconsistency of time scales.
296: Section 6 is devoted to discussion and conclusions.\\
297:
298: \section{ A Dynamical Model for Jerky Flow}
299:
300: \noindent
301: The model consists of mobile dislocations and immobile dislocations and another
302: type which mimics the Cottrell's type, which are dislocations with clouds of
303: solute atoms \cite{ana82}. Let the corresponding densities be $N_m$, $N_{im}$
304: and $N_i$, respectively. The rate equations for the densities of dislocations
305: are:
306:
307: \begin{eqnarray}
308: \dot{N}_m&=&\theta V_m N_m-\beta N_m^2-\beta N_m N_{im}+\gamma N_{im}
309: -\alpha_m N_m\,,
310: \\
311: \dot{N}_{im} &=&\beta N^2_m -\beta N_{im} N_m -\gamma N_{im} +\alpha_i N_i,
312: \\
313: \dot{N}_i &=& \alpha_m N_m - \alpha_i N_i.
314: \end{eqnarray}
315:
316: \noindent
317: The overdot, here, refers to the time derivative. The first term in Eq. (2) is
318: the rate of production of dislocations due to cross glide with a rate
319: constant $\theta$. $ V_m$ is the velocity of the mobile dislocations
320: which in general depends on some power of the applied stress
321: $\sigma_a$. The second term refers to the annihilation or immobilization of two mobile dislocations. The third term
322: also represents the annihilation of a mobile dislocation with an
323: immobile one. The fourth term represents the remobilization of the
324: immobile dislocations due to stress or thermal activation ( see $\gamma
325: N_{im}$ in Eq. 3). The last term represents the immobilization of
326: mobile dislocations either due to solute atoms or due to other pinning
327: centers. $\alpha_m$ refers to the concentration of the solute atoms
328: which participate in slowing down the mobile dislocations. Once a mobile
329: dislocation starts acquiring solute atoms we regard it as a new type of
330: dislocation, namely the Cottrell's type $ N_i$,
331: i.e, the incoming term in Eq. (4). As they acquire more and
332: more solute atoms they slow down and eventually stop
333: the dislocation entirely.
334: At this point, they are considered to have
335: transformed to $N_{im}$ (loss term
336: in Eq. (4) and a gain term in Eq. (3)).
337: Indeed, the whole process can be mathematically represented by defining $N_i = \int_{-\infty}^t K(t-t') N_m(t') dt' = \alpha_m \int_{-\infty}^{t} exp(-\alpha_i (t-t')) N_m(t')dt'$ which represents the entire process of slowing down of $N_m$ in an exponential
338: fashion with a time constant $\alpha_i$. (The choice of $K$ as having a exponential form is obviously a simplification of the actual process.)\\
339:
340: \noindent
341: These equations should be dynamically coupled to the machine equation which now takes the form
342: \begin{eqnarray}
343: \dot{\sigma_a} = \kappa(\dot{\epsilon_a} - B_0 N_mV_m),
344: \end{eqnarray}
345: \noindent
346: where $V_m$ is the velocity of mobile dislocations and $B_0$ is
347: the Burgers vector.
348: A power law dependence of $V_m = V_0 \left({\sigma_a}/{\sigma_0}\right)^m$ is used.
349: These equations can be cast into a
350: dimensionless form by using scaled variables $x=N_m \left({\beta}/{\gamma}\right)$, $ y= N_{im}\left({\beta}/{\theta V_0}\right)$, $z=N_i \left({\beta \alpha_i}/{\gamma \alpha_m}\right)$, $\tau = \theta V_0 t$ and $\phi={\sigma_a}/{\sigma_0}$.
351: \begin{eqnarray}
352: \dot{x} & = & \phi^mx - ax -b_0x^2 -xy +y,
353: \\
354: \dot{y} & = & b_0\left(b_0x^2 -xy-y+az\right),
355: \\
356: \dot{z} & = & c(x-z),
357: \\
358: \dot{\phi} & = & d\left(e-\phi^mx\right),
359: \end{eqnarray}
360: \noindent
361: Here $a = {\alpha_m}/{\theta V_0},
362: b_0={\gamma}/{\theta V_0}, c={\alpha_i}/{\theta V_0}$,
363: $\kappa=(\theta \beta \sigma_0d/\gamma B_0)$
364: and $e=(\dot{\epsilon_a}\beta/B_0 V_0 \gamma)$.
365: For these set of equations there is only one steady state which is stable. There is a range
366: of the parameters $a,b,c,d,m$ and $e$ for which the linearized equations are unstable.
367: In this range $x,y,z$ and $\phi$ are oscillatory.\\
368:
369: \noindent
370: Among these physically relevant parameters, we study the behavior of the
371: model as a function of most important parameters namely the applied
372: strain rate $e$ and the velocity exponent $m$.
373: The values of other parameters are kept fixed at $a=0.7,
374: b_0=0.002, c=0.008,$ and $d=0.0001$.
375: As can be verified these equations exhibit
376: a strong volume contraction in the four dimensional phase space. We note that
377: there are widely differing time scales corresponding to $a,b_0,c$ and $d$ (in the
378: decreasing order) in the dynamics of the model. For this reason, the equations
379: are stiff and the numerical integration routines were designed specifically to solve
380: this set of equations. We have used a variable order Taylor series expansion
381: method as the basic integration technique where the coefficients are determined
382: using a recursive algorithm.
383:
384: \section{Summary of bifurcation exhibited the model}
385:
386: \noindent
387: The model exhibits a rich variety of dynamics such as period bubbling, period
388: doubling, and complex bifurcation sequences referred to as mixed
389: mode oscillations(MMOs) in literature. Here, we will briefly recall only those aspects of
390: the bifurcation diagram relevant for the discussion of relaxation oscillations.
391: The gross features of the phase diagram in
392: the $(m,e)$ plane are shown in Fig. 3.
393: In our discussion, we use $e$ as primary control parameter and $m$ as
394: the unfolding parameter. For values of $m > m_d \sim 6.8$, the equilibrium
395: fixed point of the system of equations, denoted by ($x_0,y_0,z_0,\phi_0$), is stable. Both $x_0$ and $z_0$ are $\sim {e}/{2}$ and $y_0$ and $\phi_0$ are independent of $e$. At $m=m_d$, we have a
396: degenerate Hopf bifurcation as a function of $e$. For values
397: less than $m_d$, we have a back-to-back Hopf bifurcation,
398: the first occurring at $e = e_{c_1}$ and the reverse at $e = e_{c_2}$.
399: The periodic orbit connecting these back-to-back Hopf
400: bifurcations is referred to as principal periodic orbit.
401: The dynamics of the system is essentially bounded by these two
402: Hopf bifurcations.
403: In Fig. 3, the broken line represents the Hopf bifurcation and the dotted
404: lines
405: correspond to the locus of first three successive period doubling bifurcations. The inner, continuous lines represent the locus of saddle node bifurcations corresponding to period 3, 4 and 5 which are the first three dominant periodic windows in the alter
406: nating periodic chaotic bifurcation sequence.
407: Complex
408: bifurcation sequences, characterized by
409: alternate periodic-chaotic sequences are seen in the hatched region of the
410: parameter space. A codimension two
411: bifurcation point in the form of a cusp at
412: $(e_c,m_c)$ formed by the merging of the locus of two saddle node bifurcations of the principal periodic orbit
413: (represented by bold lines) is shown as filled diamond in Fig. 2.
414: Phase plots and other diagrams have been obtained by plotting the maxima of any one of
415: the variables $x$,$y$,$z$ or $\phi$ as a function of the control parameters
416: ($e,m$).\\
417:
418: \section{Mechanism of relaxation oscillations}
419:
420: \noindent
421: One characteristic feature of the dynamics of the system is its strong
422: relaxational nature.
423: This
424: feature persists even in regions of the $(m,e)$ plane wherein
425: complex periodic-chaotic oscillations are seen (hatched
426: region in Fig. 3). The presence of relaxations oscillations and complex periodic chaotic oscillations are
427: interrelated
428: and are a result of the geometry of the slow manifold.
429: (For details see Ref. \cite{raj99}.)
430: Relaxation oscillations that manifest in the model
431: is a type of relaxation oscillation wherein the fast
432: variable takes on large values for a short time after which it {\it assumes
433: small values of the same
434: order of magnitude as that of the slow variables.} The time
435: spent by the fast
436: variable in the part of phase space where the amplitude
437: is small is a substantial portion of the period of the orbit.
438: A typical plot of
439: $x(t)$ ( continuous line) and $z(t)$ ( dotted line) are shown in the
440: inset of Fig. 4 for $e=200.0$ and $m=1.2$.\\
441:
442:
443: \noindent
444: To understand the nature of the relaxation oscillations, we first study the structure of the slow manifold ($S$) and the behavior of the trajectories visiting different regions of $S$.
445: The slow manifold of a multiple timescale dynamical system is given by the surface spanning the time invariant solutions
446: of the fast variable. In our case, it is given by
447: \begin{eqnarray}
448: \dot{x} = g(x,y,\phi) &=& -b_0 x^2 + x \delta + y = 0
449: \end{eqnarray}
450: \noindent
451: with $\delta = \phi^m -y -a$.
452: Here, the slow variables $y$ and $\phi$ (and
453: therefore $\delta$) are regarded as
454: parameters. Further, as we will see below, it is simpler to deal with the
455: structure of the slow manifold in terms of the $\delta$ instead of both $y$
456: and $\phi$.
457: Then, the physically allowed solution of the above equation is
458: \begin{equation}
459: x = \frac {\delta + \sqrt{\delta^2+4b_0 y}} {2b_0 }
460: \end{equation}
461: \noindent
462: where $\delta$ can take on both positive and negative values.
463: Noting that $b_0$ is small and therefore $\delta^2 \gg 4b_0y$, two
464: distinct cases arise corresponding to $\delta >0$ and $\delta < 0$
465: for which $x \sim {\delta}/{b_0}$ and $x \sim -{y}/{\delta}$
466: respectively. Further, since the slow variable
467: $\phi$ and $y$ take on values of the order of unity, the range of $\delta =
468: \delta(y,\phi)$ is of the same order as that of $\phi$ and $y$ (as is evident
469: from Figs. 4 and 5). Thus, we see that $x \sim -
470: {y}/{\delta}$ is small and $ x \sim {\delta}/{b_0}$ is large.
471: For values around $\delta = 0$ and positive, we get $x \sim \left( {y}/{b_0} \right)^{1/2}$.\\
472:
473: \noindent
474: The bent-slow
475: manifold structure along with the two portions of the slow manifold, namely, $S_1$ ($\delta > 0$)
476: and $S_2$ ($\delta < 0$) are shown by
477: bold lines in the $(x,\delta)$ plane in Fig. 4.
478: We have also shown a trajectory corresponding to a mono-periodic relaxation oscillation ($m =1.2$ and $e =200.0$) by a thin line.
479: As can be seen, the trajectory spends most of the time on
480: $S_1$ and $S_2$.
481: A local
482: stability analysis for points on $S_1$ and $S_2$ shows that $\partial g/\partial
483: x = \delta-2b_0x$ is negative implying that the rate of growth of $x$ is
484: damped. Hence these regions, $S_1$ and $S_2$ will be referred to as
485: attracting or "stable".
486: For points below the line $2b_0x=\delta$ ($\delta > 0$),
487: $\partial g/\partial x > 0$
488: and hence we call this region as 'unstable'
489: (shaded region of Fig. 4).
490: Even then, the trajectory starting on $S_2$ does
491: continue in the direction of increasing $\delta$ beyond $\delta=0$.
492: We note that this region is not
493: a part of the slow manifold.
494: Once the
495: trajectory is in this region, it moves up rapidly in the $x$ direction (due to the
496: `unstable' nature) until it reaches $x = {\delta}/{2b_0}$
497: line, thereafter, the trajectory quickly settles down on to the $S_1$ part of
498: the slow manifold as ${\partial g}/{\partial x}$ becomes
499: negative. As the trajectory descends on $S_1$ approaching
500: $S_2$, we see that the trajectory deviates away from $S_1$. This
501: happens when the value of $x$ is such that $2 b_0 x < \delta$,
502: i.e., ${\partial g}/{\partial x} > 0$. Thus, points on $S_1$
503: satisfying this condition are locally unstable.
504: Thus, the
505: trajectory makes a jump from $S_1$ to $S_2$ in a short time.
506: This roughly explains the origin of the relaxation oscillation in
507: terms of the reduced variables $\delta$ and $x$.\\
508:
509:
510:
511: \noindent
512: The actual dynamics is in a higher dimensional space and a proper
513: understanding will involve analysis of the movement of the
514: trajectory in the
515: appropriate space. Moreover, unlike the standard $S-$
516: shaped manifold with upper and lower attracting pleats with the
517: repulsive (unstable) branch\cite{rossler}, in our model, both branches of the bent-slow
518: manifold are connected, and there is no repulsive branch of the slow manifold.
519: {\it Thus, the mechanism of jumping of the orbit from $S_2$ to $S_1$ is not
520: clear.} In order to understand this, consider a 3-d plot of the trajectory
521: shown in Fig. 5.
522: The region $S_2$ corresponding to small
523: values of $x$ lies more or less on the $y-\phi$ plane and the region $S_1$
524: corresponding to large values of $x$ is nearly normal to the
525: $y-\phi$ plane due
526: to the large $b_0^{-1}$ factor.
527: Regions $S_1$ and $S_2$ are demarcated by the `fold curve'
528: given by $\delta = \phi^m-y-a = 0$ which dominantly lies in the $y-\phi$ plane.
529: The rapidly growing nature
530: of the trajectory
531: lying to
532: right of the `fold curve' is due to $\partial g/\partial x > 0$.\\
533:
534:
535: \noindent
536: The principal features of the relaxation oscillations that we need to explain
537: are: a) very slow time scale for evolution on $S_2$, b) fast transition from
538: $S_2$ to $S_1$ and c) evolution on $S_1$.
539: In order to understand this, it is necessary to establish how
540: the trajectory
541: (viz, $x$,$y$,$z$ and $\phi$) visits various regions of the slow manifold in a sequential way.
542: However, our emphasis is more on those aspects of relaxation oscillation pertaining to
543: the issue of time scales raised in the Introduction, i.e., the timescales involved in the stress - time curve (Fig. 1).
544: (A detailed investigation
545: on the behavior of trajectories on this slow manifold has also been carried out. See Ref. \cite{raj99}.)
546: However, to understand the dependence of stress $\phi(t)$, we would also require information of $y$ which in turn depends on $z$.
547: In order to
548: understand this, we shall
549: analyze Eqs. (7) and (9) by recasting them in terms of $\delta$ in various regions of $S$.
550: This will help to understand the
551: general features of the flow viz., on $S_2$, just outside $S_2$ and on $S_1$.
552: In the whole analysis, it would be helpful to keep in mind the
553: range of values of $x, y, z$ and $\phi$, shown in Figs. 4 and 5,
554: in particular their values as the trajectory enters
555: and leaves $S_1$. \\
556:
557: \noindent
558: First,
559: consider rewriting
560: Eq. (7) valid on the slow manifold $S$ in terms of $\delta$:
561: \begin{equation}
562: \dot{y} = b_0(x\delta - xy + az).
563: \end{equation}
564: \noindent
565: The presence of the
566: $z$ variable in Eq. (12) poses some problems.
567: Using detailed arguments based on the knowledge of the magnitude of $x$ and $z$ just inside, on and outside $S_2$, it can be shown that the trajectory enters $S_2$ at small values of $y$ and leaves $S_2$ at relatively larger values. Further one can show that there is a turning point for $y$ on $S_2$ (see Fig. 5). For details, see Ref\cite{raj99}.\\
568:
569: \noindent
570: With this information on the evolution of $y$ on $S_2$, we
571: now consider the changes in $\phi$ as the trajectory enters and leaves $S_2$.
572: From Eq.(9), it is clear that an yield drop starts when $x$ is large (i.e., when $x\sim \delta/b_0$ on $S_1$)
573: and ends when $x$ is close to minimum, when the trajectory is on $S_2$, which implies that $\phi$ is small when the trajectory enters $S_2$.
574: Using the value of
575: $x= {y}/{\vert \delta \vert}$ on $S_2$ in
576: Eq. (9), we find that $e \gg {\phi^m y}/{\left| \delta
577: \right|}$, since $y$ is near its minimum value as the trajectory enters $S_2$.
578: Thus, $\phi$ increases linearly from small values of $\phi$ at a rate close to $d e \ll 1$.
579: {\it We recall that the loading rate in the experimental
580: stress-strain curve was $\kappa \dot{\epsilon_a}
581: ( de $ in scaled variables), which now can be understood as due to the
582: structure of the slow manifold}.
583: This is a direct consequence of the fact that the magnitude of $x$ remains constant as
584: $\dot{x} \sim 0$ for the entire interval the trajectory on
585: $S_2$. This is consistent with what we argued from the stress - time
586: plot (Fig. 2), namely, the mobile dislocation density should be constant
587: during the loading period. As the trajectory moves into $S_2$, $y$ goes through a maximum whereas $\phi$ continues to increase since $x \sim y/|\delta|$ remains small.
588: However, as the trajectory is just outside $S_2$ for which $x \sim \left(
589: {y}/{b_0} \right)^{1/2}$ for $\delta >0$
590: and small,
591: $ \phi^m \left( {y}/{b_0}\right)^{1/2} \sim
592: e$, since $\phi$ and $y$ are relatively large which implies that $\phi$ is about to decrease.
593: The above discussion on $\dot{y}$ and $\dot{\phi}$ for region just outside and inside the fold curve also gives us the
594: direction of movement of the trajectory in this region, namely, it
595: enters $S_2$ in the region corresponding to small values of $y$ and $\phi$, and makes an
596: exit for relatively larger values of $\phi$ and $y$ (compared to
597: their values as the trajectory enters $S_2$).
598: Further, as $\dot{x} \sim 0$, we see that the dynamics on $S_2$ is controlled by the slow variables.\\
599:
600:
601: \noindent
602: Finally, just to the right of $\delta = 0$ line,
603: $\dot{x} \sim
604: x \delta $, with $\delta$ very small, which suggests that the time constant is
605: small. Thus, the growth of $x$ is slow in the neighbourhood of $\delta = 0$,
606: and is tangential to the $S_2$ plane even in the `unstable' region.
607: However, once the trajectory moves away from $\delta = 0$, the growth of
608: the trajectory is controlled by $\partial g/\partial x$ and hence {\it the time scale of growth of $x$ is of the order of $\delta^{-1}$ which is of the order of unity explaining the short time span of the stress drop} seen in Fig. 2.
609: This also
610: explains why the trajectory tends to leave stable portion of the slow manifold $S_2$ and move into the `unstable' region.\\
611:
612: \noindent
613: Once in the `unstable' region, the value of $x$ continues to grow in this
614: region of the phase space as can be seen from Eq. (9) until the value of
615: $x$ is such that $\phi^mx=e$ is satisfied. Beyond this value of $\phi$,
616: $\dot \phi$ is negative. Thus, the trajectory leaving $S_2$ eventually falls
617: onto the $S_1$ part of the slow manifold. We can again evaluate $\dot{y}$ and
618: $\dot{\phi}$ just as the trajectory reaches $S_1$.
619: Using $x \sim {\delta}/{b_0}$ in Eq. (7),
620: it can be shown that $y$ decreases. Now, consider the equation for $\phi$.
621: Using $x \sim {\delta}/{b_0}$ on $S_1$, we see that $\phi^m\delta/b_0 > e$. Thus $\dot{\phi} < 0$
622: when the trajectory reaches $S_1$ with a time constant
623: $\sim d/b_0$ which are
624: relatively fast.
625: (These statements are true only as the trajectory hits $S_1$.)
626: We recall here that in the experimental time series, the stress
627: drops from a peak value to its minimum in a very short time
628: span. Further, we have argued that this
629: should be the sum of contributions arising from fast
630: multiplication of dislocations (which we have already argued has
631: a time scale of $\delta^{-1}$) and subsequent immobilization.
632: The latter is reflected in another rapid time scale
633: $\sim d/b_0$. {\it This explains the difficulty
634: in separating the contributions arising from the two processes in the experimental time series.}
635: Moreover, since $x$ is a fast variable, the changes in $x$ component dominates
636: the descent of the trajectory.
637: Finally, as the trajectory approaches $S_2$,
638: ${\partial g}/{\partial x}$ becomes positive and the trajectory jumps
639: from $S_1$ to $S_2$.
640: Combining these results, we see that the
641: trajectory moves towards the region of smaller values of $y$ and $\phi$ entering
642: $S_2$ in a region of small values of $y$ and $\phi$. \\
643:
644: \noindent
645: In summary, the sequential
646: way the orbit visits various parts of the phase space is as follows. The
647: trajectory enters $S_2$ part of the slow manifold in regions of small $y$ and
648: $\phi$ making an exit along $S_2$ for relatively large $\phi$ and $y$.
649: Thereafter, the trajectory moves through the
650: `unstable' part of the phase space before falling onto the $S_1$ and quickly
651: descends on $S_1$. This completes the cyclic movement of the trajectory and
652: explains the geometrical feature of the trajectory shuttling between these two
653: parts of the manifold and the associated time scales.\\
654:
655:
656: \noindent
657: Now, the question that remains to be answered is $-$ do the trajectories always
658: visit both $S_1$ and $S_2$ or is there a possibility that the trajectory
659: remains confined to $S_1$ ? It is clear that if the former is true, relaxation
660: oscillations with large amplitude will occur and if the latter
661: is true, these are
662: likely to be nearly sinusoidal small amplitude oscillations.
663: Here, we recall that the coordinates of
664: the saddle focus fixed point are
665: $x_0 = z_0 \sim e/2 $ which is much larger than the value of
666: $x \sim y/\left|\delta \right| $ on $S_2$.
667: Thus, the fixed point located
668: on the $S_1$ will be close to the `fold' at the first Hopf
669: bifurcation which occurs at small values of $e = e_{c_1} \sim 5$.
670: Due to the unstable nature of the
671: fixed point, the trajectories spiralling out are forced onto the
672: $S_2$ part of the manifold resulting in relaxation oscillation.
673: This point has been illustrated by considering the example of
674: a period eleven orbit
675: for $m=1.2$ and $e=267.0$ shown in Fig. 6. As is clear from this diagram,
676: the small amplitude oscillations
677: are located on the $S_1$. As $e$ is further varied, the small amplitude oscillations
678: grow with $e$, but the relaxation nature does not manifest until the orbit
679: crosses over to $S_2$.
680: To the best of the authors
681: knowledge the mechanism suggested here for pulsed type relaxation oscillations
682: is new. \\
683:
684: \noindent
685: As we will see, the analysis of the slow manifold and the time
686: scales operating in different parts of the phase space will be
687: useful in providing an appropriate interpretation of the various
688: branches of the SRS.
689:
690: \newpage
691: \section{Negative strain rate sensitivity}
692: \noindent
693: At the outset, we stress that it has been recognized
694: that the negative unstable branch is not accessible to the dynamics
695: of the PLC effect. Even so, early formulations and the way experimental measurements
696: have been carried out has given rise to considerable confusion.
697: The purpose of the
698: material presented below is to briefly discuss the concept of negative
699: SRS and working methods adopted in the literature, and also
700: clear some misconceptions.\\
701:
702: \noindent
703: Theories of dynamic strain ageing (DSA) assume that the interaction
704: of dislocations with solute atoms when averaged over the specimen dimensions
705: can be represented by a constitutive relation connecting stress, strain
706: and strain rate which is conventionally written as \cite{kub89}
707: \begin{equation}
708: \sigma = h\epsilon + F(\dot\epsilon).
709: \end{equation}
710: The basic assumption inherent in Eq. (13) is that
711: stress can be split into
712: a function of $\epsilon$ and another of $\dot{\epsilon}$ alone.
713: Then, the SRS is defined as
714:
715: \begin{equation}
716: {\cal S} = \left.\frac{\partial \sigma}{\partial ln \dot \epsilon}\right|_{\epsilon} = \dot{\epsilon} \frac{d \sigma}{d \dot{\epsilon}}
717: \end{equation}
718:
719:
720: \noindent
721: Clearly, this definition uses $\epsilon$ as a state variable. This
722: unfortunately is not correct since strain is history dependent.
723: Inspite of this, conventionally, strain is fixed at a small nominal
724: value and the flow stress at that value is used
725: to obtain the SRS.
726: It is interesting to note that the existence of critical strain for the onset of the PLC effect implies that
727: when the nominal
728: strain value is lower than $\epsilon_c$, there are no
729: serrations even when the applied strain rate value is in the domain of
730: the PLC effect ( $e_{c_1} < e < e_{c_2}$). Yet, the onset of
731: serrations for higher strains is somehow reflected in the measured nonmonotonic
732: behavior of the flow stress\cite{bod67}.
733: In experiments, by fixing
734: $\epsilon$ at some nominal value less than $\epsilon_c$,
735: the flow stress ( at the
736: fixed strain) is found to increase as a function of applied strain rate
737: $e$ for $ e < e_{c_1}$, shows a decreasing trend for $ e_{c_1} < e
738: < e_{c_2}$ and again reverts to an increasing trend
739: for $e > e_{c_2}$ \cite{bod67}.
740: Thus, the flow stress has the form
741: shown in Fig. 1.
742: No explanation has been offered in the literature as
743: to why this nonmonotonic behavior should be seen. However,
744: explanation from the dynamical point of view is fairly
745: straightforward and is as follows.
746: We recall here that the model predicts the existence the critical strain
747: $\epsilon_c$, as also the existence of window of strain rates $e_{c1} < e < e_{c2}$ within which serration can occur.
748: Thus, from Eq. (9)
749: it is easy to understand the increasing order
750: in which the stress-strain curves are placed for increasing
751: values of $e$ when $e < e_{c_1}$ and $e > e_{c_2}$.
752: In this range of $e$, the fixed point is
753: stable and thus all trajectories converge to the fixed point.
754: However, for $e_{c_1} < e < e_{c_2}$, we note that serrations
755: result only for large enough strains, i.e., once the time of deformation is such that strain crosses $\epsilon_c$.
756: In our theory, serrations are equated with the existence of
757: periodic (or aperiodic) solutions when $e_{c_1} < e < e_{c_2}$. These steady
758: state solutions are usually reached only after transients die down.
759: Thus, low value of nominal strain implies {\it short evolution time} which in turn
760: implies that the {\it stress is being monitored at a transient state}.
761: Thus, the decreasing trend of the flow stress for $e_{c_1} < e <
762: e_{c_2}$ is a reflection of the impending periodic (aperiodic) steady state
763: that will be reached eventually. Indeed,
764: this was the method followed in our earlier calculation
765: since the procedure was easy to implement numerically\cite{ana82}.
766: However, in many experimental situations, it is not possible to choose
767: a nominal strain value low enough that it is
768: less than $\epsilon_c$ for the entire range of strain rate values.
769: In such a case, since the stress-strain curves
770: are serrated, there is an ambiguity in the value of stress
771: to be used. A working method
772: adopted is to use a stress value as the
773: mean value of the upper and lower stress
774: values \cite{kub86}. Then, the flow
775: stress appears to decrease for the domain of applied strain rate
776: values where the PLC effect manifests. Thus, this method
777: gives the impression of actually measuring the unstable branch.\\
778:
779: \noindent
780: The above methods are not suitable for adoption since they do
781: not permit the use of the knowledge of the slow manifold.
782: There is an alternate method which uses the relaxation
783: oscillations inherent to the dynamics of the PLC effect.
784: In this method, by analogy with
785: electrical analogues, one assumes
786: that there exists a
787: family of curves $F(\dot\epsilon_p)$ for each $\epsilon$ of the
788: form shown in Fig.1 which trigger
789: relaxation oscillations in the form of plastic strain rate bursts and
790: stress drops. By comparing the measured stress drops and strain
791: bursts, one concludes the existence of the unstable branch, but
792: {\it one never records any points in this region.}
793: This method is suitable for our study since we will use
794: relaxation oscillations arising in the model.\\
795:
796: \noindent
797: In the following we shall argue that the two slow time scales
798: in the dynamics actually translate into the two stable dissipative
799: branches of the SRS and the two fast time scales to jumps in
800: plastic strain rate across the stable branches of the negative SRS.
801: Since SRS represents $\phi$ as
802: a function of the plastic strain
803: rate $\dot \epsilon_p = \phi^m x$, in Fig. 7, we have shown a
804: projection of the phase space
805: trajectory on the $\phi - \dot{\epsilon}_p$ plane (instead of
806: $\phi - x$) corresponding to a monoperiodic relaxation oscillation. ( Here, we
807: have retained the same notation for the scaled plastic strain
808: rate as for the unscaled one.)
809: The unstable fixed point is also
810: shown. Starting from any initial value around the unstable focus, trajectories
811: spirals out converging onto the limit cycle.
812: In Fig. 7, we have identified different regions of the phase space with
813: different regions of the slow manifold, $S_1$ and $S_2$.
814: We first note that there is a considerable similarity between Fig. 7 and
815: the schematic representation of the relaxational oscillation obtained using
816: the negative SRS shown in Fig. 1.
817: Note also that in contrast to the artificial flat parts $B^\prime C^\prime$ and $C^\prime D^\prime$ of Fig. 1,
818: the equivalent parts in Fig. 7 have a finite negative slope.
819: Lastly, as in experiments, the {\it strain rate jump} from B to C
820: is over {\it two orders of magnitude}. \\
821:
822: \noindent
823: Here, we set up a correspondence between the dynamics in the phase space (Fig. 7) and the slow manifold (Figs. 4 and 5).
824: From our earlier discussion, we know that when the trajectory is on $S_2$,
825: $x$ is constant and small in magnitude. Consequently, according
826: to Eq. (9), $\phi$ should increases linearly and hence
827: this corresponds the rising branch AB in Fig. 7.
828: Further, noting that $\dot x \sim 0$ for the entire interval of time spent
829: by the trajectory on $S_2$ (see Fig. 4), the branch AB of Fig. 7 corresponds to {\it the pinned
830: state of dislocations}.
831: For this branch,
832: one can easily see that the (mean)
833: value of ${\cal S} \sim 3.5$ using Eq. (14).
834: Further, as we
835: move up on this branch towards B (Fig. 4), the value of $\delta$
836: approaches zero, and $\phi$ reaches its maximum value.
837: Once $\delta$ becomes positive, the trajectory leaves $S_2$, and
838: thus, the strain rate jump from B to C in Fig. 7,
839: corresponds to the trajectory jumping from
840: $S_2$ to $S_1$ in Figs. 4 and 5. Note that the slope $\partial \phi/\partial
841: \dot\epsilon_p$ for this portion of the orbit is
842: quite small and negative unlike the zero value
843: for the equivalent part in Fig. 1.
844: Further, we know from Fig. 4, once the trajectory reaches $S_1$, the value of $x$ decreases
845: rapidly resulting in the decrease of $\dot\epsilon_p$.
846: Thus, the region CD in Fig. 7 corresponds to the
847: movement of the trajectory on
848: $S_1$ ( in Fig. 4 for which $\delta > 0$). For this branch, one can
849: quickly check that the strain
850: rate sensitivity ${\cal S}$
851: is positive having a mean value ( $\sim 1.5$) which is
852: a factor of two less than that for the
853: branch AB, implying that the nature of dissipation is quite
854: different from that operating on AB.
855: This is consistent with known facts about the two branches as
856: mentioned in the introduction.
857: Combining this with the
858: fact that $\dot x$ is decreasing,
859: the branch CD in Fig. 7 mimics the equivalent branch $C^\prime D^\prime$ in Fig. 1, which is identified with
860: {\it the slowing down of
861: the mobile dislocations without solute
862: atmosphere}.\\
863:
864:
865: \noindent
866: We recall that the stress drop duration has contributions
867: from two fast processes, namely, dislocation multiplication and
868: it's subsequent immobilization. But, these two time scales could not
869: be separated in the stress strain curve. However, in the present phase plot
870: representation (Fig. 7), we see that the fast multiplication of dislocations
871: correspond to BC and that of immobilization to CD. This correspondence
872: has been possible due to the mapping of the relevant time scales in the
873: dynamics of the dislocations obtained from the analysis of the slow
874: manifold to the various regions in the phase plot thereby
875: allowing us to identify the individual contributions.
876: (Note also that in Fig. 7, we have plotted points of the trajectory
877: at equal intervals of time which shows that the time interval
878: corresponding to BD is small.)
879: From Figs. 4 and 5, we see that, as the trajectory descends on $S_1$ part of the slow
880: manifold and gets close to $S_2$, it leaves $S_1$, since
881: ${\partial g}/{\partial x}$ becomes positive ($x \sim 50$).
882: Further, the strain rate sensitivity parameter $\cal S$
883: changes sign at D.
884: For the corresponding DA part in Fig. 7, the slope is small and negative as
885: for the part BC.
886: Noting that B and D are the points at
887: which ${\cal S}$ turns negative, and noting that the fixed
888: point is unstable, {\it the so called "unstable branch" of the SRS,
889: not accessible to the dynamics, can be inferred by
890: drawing a (dotted) line connecting the maximum and the minimum of the
891: stress and passing it through the unstable fixed point (Fig. 7).} \\
892:
893: \noindent
894: We will now attempt to use the results of our analysis of time
895: scales in different regions of the slow manifold
896: to obtain the dependence of $\dot \epsilon_p $ as
897: a function of $\phi$. The equation for $\dot \epsilon_p $ is
898:
899:
900: \begin{equation}
901: \frac{d\dot\epsilon_p}{dt}=x m\phi^{m-1}\dot{\phi} + \phi^m\dot{x}
902: \end{equation}
903: \noindent
904: which on using Eq.(6) and (9) gives
905:
906:
907: \begin{equation}
908: \frac{d \dot\epsilon_p}{d \phi} = \frac{\dot\epsilon_p
909: (\frac{mde}{\phi} + \delta) - \dot \epsilon_p^2 (\frac{md}{\phi}
910: + \frac{b}{\phi^m}) + y\phi^m} {d(e-\dot {\dot \epsilon_p})}
911: \end{equation}
912:
913: \noindent
914: Here we note that
915: in the slow manifold description, all slow variables appear as parameters.
916: However, since SRS describes the dependence of the slow variable $\phi$
917: as a function of the (derived) fast variable $\dot \epsilon_p $,
918: we will consider the other two variables $y$ or $\delta$ or both
919: as parameters. Numerical solution of Eq.(16)
920: has been attempted using $y$ and $\delta$ as parameters.
921: Good numerical approximation
922: is obtained by noting that $y$ and hence $\delta$ is periodic.
923: Thus, any reasonable approximation for the periodicity of
924: $y$, for example, sine function with a proper amplitude and
925: phase gives a good fit with the phase plot.
926: However, our interest here is
927: to obtain approximate expressions for $\dot\epsilon_p(\phi)$ on
928: different branches. For this reason, we will use typical values of
929: $\delta$ and $y$ for the interval under question.
930: From section 4, the trajectory has different dynamics in different
931: regions of the slow manifold. These are : (1) on $S_2$ where $\dot x$
932: is nearly zero for the entire time spent by the trajectory on $S_2$, (2)
933: just outside $S_2$ where $\dot x \sim x \delta$,(3) on
934: $S_1$ where $x \sim \delta/b_0$ for $\dot{\epsilon_p} > e$ and (4) when
935: the trajectory jumps from $S_1$ to $S_2$.
936: Approximate solutions obtained for these cases are shown
937: in the phase plot by solid lines. Details are given in the
938: Appendix. It is clear that
939: these solid lines reproduce the general features of the phase plot quite well. We stress here that these lines correspond to the simplest approximation.
940:
941:
942: \noindent
943: The above analysis refers to a fixed value of $e$.
944: As a function of $e$, we find
945: that the magnitude of the stress drops
946: increases initially and then
947: decreases.
948: This feature is a direct result of the existence of back-to-back Hopf bifurcations in the model.
949: On the other hand, experimentally one sees
950: only a decreasing trend.
951: While the decreasing trend is consistent, the increasing trend seen in the model for low strain values
952: can be traced to the effect of another crucial parameter in the
953: model, namely $b_0$. We recall that this parameter corresponds to the remobilization of
954: immobile dislocations. For the value of $b_0$ used in the present
955: calculation, the bifurcation from the steady state is a
956: mildly subcritical Hopf bifurcation, i.e., across the transition the amplitude of the stress change is abrupt but the magnitude is small. However, for smaller values of
957: $b_0$, this jump can be made sufficiently large in which case
958: the amplitude of the stress drops can be made to decrease with
959: $e$ right from the onset of the PLC effect. \\
960:
961:
962: \section{Discussion and Conclusions}
963:
964:
965: \noindent
966:
967: The study of the relaxation oscillations in the model was
968: motivated by the need to explain the apparent inconsistency
969: between the time scales observed in experimental stress-time
970: series and those that could be argued on the basis of the
971: negative SRS feature commonly used in the literature. The
972: study of
973: relaxation oscillations using the geometry of the slow
974: manifold
975: has helped us
976: to identify different time scales operating in different regions
977: of the phase space, apart from showing that
978: the nature of the relaxation in the model is due to the atypical bent geometry of the slow manifold. This geometry is very different
979: from the standard $S-$ shaped
980: manifold and hence the
981: relaxation oscillations seen here
982: differs qualitatively
983: from that seen in systems with $S-$shaped slow manifold.
984: Some comparative comments between these two types of manifolds
985: may be in order here. As in the $S-$shaped
986: manifold, there are two attracting branches in our case also,
987: namely $S_1$ and
988: $S_2$. The dynamics on $S_2$ is slow as it is controlled by
989: the slow variables $y$ and $\phi$.
990: On the other hand, on $S_1$, the time dependence of the
991: trajectory is largely controlled by the fast variable $x$.
992: In this
993: sense, the dynamics on $S_2$ is slow and that on $S_1$ is fast.
994: with a large magnitude and with a much smaller magnitude from
995: $S_1$ to $S_2$. Though there are two fast jumps as in the $S-$shaped
996: manifold, in our case, there is no equivalent unstable part
997: of the slow manifold which causes these jumps.\\
998:
999:
1000:
1001:
1002:
1003:
1004: \noindent
1005: The analysis of the time scales controlling the
1006: relaxation oscillations has been directly used to reconstruct
1007: the relaxation
1008: oscillation in the $\phi - \dot\epsilon_p$ plane which bears a strong
1009: resemblance to the the relaxation oscillations resulting from
1010: the assumed form of the negative SRS. The information on
1011: different time scales operating in different regions of the slow
1012: manifold has been used
1013: to calculate the dependence of $\phi$ on $\dot \epsilon_p$
1014: for the two dissipative branches and the associated strain rate
1015: jumps between them.
1016: This has helped to identify the various
1017: regions of the slow manifold with {\it the stick state} and {\it
1018: the slip state of dislocations.} It has also helped us to clarify the inconsistency in the time scales of the dynamics.
1019: Further, several important features of the SRS derived from the
1020: model compare well with that reported in the literature. In
1021: particular, we note that
1022: the slope of the first dissipative branch AB is larger
1023: than that of the second branch CD (Fig. 7). Further,
1024: we recall that $-\delta = -\phi^m +y +a$ which
1025: is positive for AB, gradually approaches zero as B is reached
1026: followed by strain rate jump. Similarly, for the
1027: branch CD, $\delta$ approaches zero as we approach D followed by a jump
1028: in the strain rate. Thus, {\it vanishing of $\delta$ is indicative of
1029: strain rate jumps just as the strain rate sensitive parameter $S$ also vanishes}. Noting that $y$ is the immobile dislocation
1030: density, it is tempting to interpret $\delta$ as
1031: being related to some kind of effective stress. (Recall that the effective stress is
1032: $\sigma^* = \sigma_a - HN_{im}^{1/2}$, where H is the
1033: work hardening coefficient.)
1034: Thus, the points at which strain rate jumps
1035: occur correspond to points at which the effective stress vanishes which is very much like the
1036: classical explanation.
1037: Since the definition of strain rate sensitivity assumes strain as a state variable which is not true, $\delta$ may be an effective alternate parameter for defining strain rate sensitivity.
1038: Thus, it is nice to see that we can attribute
1039: a physical meaning to this parameter.\\
1040:
1041: \noindent
1042: The analysis has also helped us to provide a dynamical
1043: interpretation of the negative SRS. The analysis also shows that
1044: the large jumps in the strain rate
1045: across the stable branches are due to the relaxational nature
1046: of the dynamics which in turn is a result of bent nature of the
1047: slow manifold and the fact that the bifurcation is of the Hopf type.
1048: Using this, we have inferred the existence of
1049: the unstable branch as containing the two points, B and D, where strain rate
1050: jumps (where $\delta$ and $S$ are zero) and the unstable fixed point.
1051: In this sense, Hopf bifurcation is at the root of the `negative' SRS.
1052: Similar
1053: features of SRS were found to operate in a model
1054: designed to mimic
1055: stick-slip dynamics of tectonic faults \cite{ana94}. There are
1056: other studies on stick-slip dynamics, both experimental\cite{hes94} and theoretical \cite{car95}, which support the
1057: view that Hopf bifurcation was found to be responsible for
1058: the instability.
1059: Thus, it is likely that Hopf bifurcation is
1060: relevant to situations where stick-slip dynamics operates and
1061: wherever one measures the two stable branches and the jumps
1062: across the branches\cite{ana97}. \\
1063:
1064:
1065:
1066:
1067:
1068:
1069: \noindent
1070: The relaxation oscillations in the model are reminiscent of the
1071: {\it canard} type of oscillations in multiple type scale dynamical
1072: systems \cite{mil98,eck83}.
1073: The latter type of oscillations result from `sticking' of
1074: the trajectory to the repelling part of the $S-$shaped slow manifold before
1075: jumping to the attracting pleat of the slow manifold. In our
1076: case, although these oscillations have a similarity with {\it
1077: canard} type of solutions, the repelling part of slow manifold does not exist.
1078: Instead, the trajectories stick to the `unstable' part of the
1079: phase space where the dynamics is accelerated once the
1080: trajectory moves well into this region. This aspect coupled to
1081: the fact that there is no inherent
1082: constraint on the manifold structure leads to oscillations of
1083: all sizes. It is clear that such oscillations result from the
1084: trajectory `sticking' to the
1085: direction of the $S_2$ plane and moving into the `unstable' part of the phase
1086: space by varying amounts each time the trajectory visits $S_2$.
1087: These jumps translate into stress drops of varying sizes which
1088: are generally seen in experimental time series (Fig. 2).
1089: This also means that $\dot{\epsilon_p} - \phi$ is not a simple
1090: limit cycle, and
1091: the simplistic approach of inferring the `negative' SRS
1092: should be given up.
1093: The present analysis stresses the
1094: importance of using sound dynamical tools such as the slow
1095: manifold as the basis for studying more complex
1096: oscillations rather than phenomenological concepts such as
1097: the negative SRS.\\
1098: \appendix
1099: \section
1100: \noindent
1101: Here, we obtain approximate analytical expressions
1102: for $\dot \epsilon_p (\phi)$ for different regions of $\phi -
1103: \dot{\epsilon_p}$ phase plot (Fig. 7)
1104: using the knowledge of time scales obtained
1105: from the analysis of relaxation oscillations.
1106: For the numerical evaluation, the values of control parameters have been
1107: chosen as $e=200$, $m=1.2$,
1108: $b_0=0.002$ and $d=0.0001$.\\
1109:
1110: \noindent
1111: Region AB :
1112: When the trajectory is on $S_2$, $\dot{x} \sim 0$, for the entire interval of time.
1113: Using
1114: $x= -{y}/{\delta}$ in Eq.(15), we get
1115: \begin{equation}
1116: \frac{d \dot\epsilon_p}{d \phi} = -\frac{\phi^{m-1} m y}{\delta}.
1117: \end{equation}
1118: \noindent
1119: Noting that $\delta = \phi^m - y -a$, this equation can be integrated thereby reducing the number of parameters to one, namely $y$.
1120: Integrating, we get
1121: \begin{equation}
1122: \dot\epsilon_p = - y {\it ln}\left(\frac{\phi^m-(y+a)}{\phi^m(0)-(y+a)}\right) + \edot(0)
1123: \end{equation}
1124:
1125:
1126:
1127: \noindent
1128: where $\dot\epsilon_p(0)$ and $\phi(0)$ refer to their respective values
1129: as the trajectory enters $S_2$. In Fig. 7, we have used $\dot\epsilon_p(0) = 4.7$,
1130: $\phi(0) = 2.2$ and $y = 6.15$.\\
1131:
1132: \noindent
1133: Region BC :
1134: This region corresponds to the
1135: jump from $S_2$ to $S_1$. This happens when the trajectory is just outside
1136: $S_2$.
1137: For this region, $\phi$ is near $\phi_{max}$, and since $x \sim \left({y}/{b_0}\right)^{1/2}$, the evolution of $x$ is well
1138: described by $\dot{x} \sim x \delta $, implying that the time of evolution is very short. Thus, we can
1139: regard the evolution of $\phi$ as being mainly determined by that of $x$.
1140: (This region also corresponds to $\delta > 0$ and small $ \sim 0.2$.)
1141: Thus, we use $\phi = \phi_{max}$ on the RHS of Eq. (9). Then, \begin{equation}
1142: \frac{d \phi} {d t} = d (e - \phi^m_{max} x_{s_2} e^{\delta t})
1143: \end{equation}
1144: \noindent
1145: where $x_{s_2}$ is the value of $x$ at the time of leaving $S_2$.
1146: Integrating this equation with initial conditions at $t=0$, $\phi = \phi(0)=\phi_{max}$, we get
1147:
1148: \begin{equation}
1149: e^{\delta t}=\frac{te\delta}{x_{s_2} \phi^m_{max}} - \frac{(\phi(t) - \phi_{max})\delta}{x_{s_2}\phi^m_{max}d} + 1
1150: \end{equation}
1151:
1152:
1153:
1154: \noindent
1155: Clearly, the first term is small since the time span of
1156: evolution that we are interested is also $\sim \delta$.
1157: Now consider Eq. (15).
1158: Using $\dot{x} \sim x \delta $, we get
1159:
1160: \begin{equation}
1161: \frac{d \dot\epsilon_p}{d t} = \dot \epsilon_p \left( \frac{med}{\phi_{max}} +
1162: \delta\right) - \frac{m\dot\epsilon_p^2 d}{\phi_{max}}
1163: \end{equation}
1164:
1165: \noindent
1166: Since ${med}/{\phi_{max}} << \delta$, we drop the first term.
1167: Integrating the above equation with the initial value $\dot{\epsilon_p}(0) = e$, leads to
1168: \begin{equation}
1169: \dot\epsilon_p=\frac{\phi_{max}e(\phi_{max}-\phi(t)) \delta}{md^2\phi^m_{max}x_{s_2}(\phi_{max}\delta/md - e) - (\phi(t)-\phi_{max})emd}
1170: \end{equation}
1171: \noindent
1172: In Fig. 7, we have used the values $\delta = 0.021$,
1173: $\phi_{max}=4.98$ and $x_{s_2} = 1.7$.\\
1174:
1175:
1176:
1177: \noindent
1178: Region CD :
1179: Consider the trajectory on $S_1$ with $x \sim \delta/b_0$ and $\dot\epsilon_p \gg e$. Then, Eq. (16) reads,
1180:
1181:
1182:
1183: \begin{equation}
1184: \frac{d \dot\epsilon_p}{d \phi} = - \frac{\dot\epsilon_p (\frac{med}{\phi} + \delta)
1185: - \dot\epsilon_p^2 (\frac{md}{\phi} + \frac{b_0}{\phi^m}) + y\phi^m} {d\edot(1-\frac{e} {\dot\epsilon_p})}
1186: \end{equation}
1187: \noindent
1188: Since ${e}/{\dot\epsilon_p} < 1.0$, we expand the denominator and
1189: retain terms upto $\dot\epsilon_p^{-1}$.
1190: We note here that on $S_1$, $x$ is rapidly decreasing and therefore, using slow manifold values is not a good approximation. Even so, as a simplest approximation we use $x \sim \delta/b_0$. Then, we get
1191:
1192: \begin{equation}
1193: \frac{d \dot\epsilon_p}{d \phi} = - \frac{me^2 b_0}{\phi^{m+1} \delta} - \frac{eb_0^2 y}{\phi^m d\delta^2 } + \frac{m\phi^{m-1}\delta}{b_0} - \frac{b_0y}{d\delta}
1194: \end{equation}
1195:
1196: \noindent
1197: In this equation, both $y$ and $\delta$ appear as parameters
1198: whose values are chosen appropriate to this region.
1199: Integrating the above equation with the initial values of $\phi(0)$
1200: and $\edot(0)$,
1201: we get
1202:
1203: \begin{eqnarray}
1204: \dot\epsilon_p &=& \frac{eb_0^2y}{(m-1)d\delta^2}
1205: ( \phi^{1-m} - \phi^{1-m}(0)) + \frac{e^2b_0}{\delta} (\phi^{-m}-\phi^{-m}(0))\nonumber\\
1206: & &+ \frac{\delta}{b_0}(\phi^{m} - \phi^{m}(0)) - \frac{b_0y}{d\delta} (\phi - \phi(0))) + \dot{\epsilon_p}(0)
1207: \end{eqnarray}
1208: \noindent
1209: In Fig. 7, we have used the values of $\delta = 3.57$, $y = 1.0$ with $\phi(0)=4.3$ and $\edot(0)=7460.0$.\\
1210:
1211:
1212: \noindent
1213: Region DA :
1214: This region again corresponds to the the trajectory descending on $S_1$ but
1215: $\dot\epsilon_p < e$.
1216: Then, Eq.(16) can be written
1217: in the form
1218:
1219:
1220: \begin{equation}
1221: \frac{d \edot}{d \phi} = -\frac{\edot (\frac{med}{\phi} + \delta) - \edot^2 (\frac{md}{\phi} + \frac{b_0}{\phi^m}) + y\phi^m} {ed(1-\frac{\edot}{e})}
1222: \end{equation}
1223:
1224: \noindent
1225: In this region, the value of $\phi$ is
1226: slowly varying with its value near the minimum
1227: for which $\delta \sim - 0.15$.
1228: Since $\phi$ is near $\phi_{min}$, we use $\phi = \phi_{min}$
1229: and regard the variation
1230: as largely arising due to
1231: the changes in $\dot \epsilon_p$.
1232: Using $(1-\frac{\dot \epsilon_p}{e})^{-1} \approx (1 + \frac{\dot\epsilon_p}{e})$, we have
1233:
1234:
1235: \begin{equation}
1236: \frac{d \edot}{d \phi} = -(A_1\edot + B_1\edot^2 - C_1 \edot^3)
1237: \end{equation}
1238: \noindent
1239: where $A_1= \frac{m}{\phi_{min}} - \frac{\vert\delta\vert}{ed}$ and
1240: $B_1=-\frac{\vert \delta\vert}{e^2d} -
1241: \frac{b_0}{e\phi^m_{min}d}$ and $C_1= \frac{m}{\phi_{min} e^2} + \frac{b_0}{e^2d \phi^m_{min}}$.
1242: Since $C_1 \ll A_1$ and $ B_1$, we drop the last term.
1243: Integrating the above equation with the initial conditions, $\phi(0) = \phi_{min}$ and $\dot{\epsilon_p} = e$, we get
1244:
1245:
1246:
1247: \begin{equation}
1248: \edot=\frac{eA_1e^{A_1(\phi-\phi_{min})}}{A_1+eB_1(1-e^{A_1(\phi-\phi_{min})})}
1249: \end{equation}
1250:
1251:
1252: \noindent
1253: In Fig. 7, we have used $\phi_{min} = 1.55$.
1254:
1255:
1256:
1257: \begin{thebibliography}{99}
1258:
1259: \bibitem{portevin}
1260: %1
1261: F. Le Chatelier, Re. de Metallurgie {\bf 6}, 914 (1909);
1262: A. Portevin and F. Le Chatelier, C. R. Acad. Sci. {\bf 176}, 507
1263: (1923).
1264: \bibitem{cot53}
1265: %2
1266: A.H. Cottrell, Phil. Mag. {\bf 44}, 829 (1953).
1267:
1268:
1269: \bibitem{bri70}
1270: %3
1271: B.J. Brindley and P.J. Worthington, Metall. Reviews {\bf 145}, 101 (1970).
1272: \bibitem{kub93}
1273: %4
1274: L.P. Kubin, Y. Estrin and E.C. Aifantis, Organizers, {\it View Point Set}, Scripta
1275: Metall. {\bf 29}, (1993).\bibitem{mcc72}
1276: %5
1277: P.G. McCormick, Acta Metall. {\bf 20}, 351 (1972).
1278: \bibitem{van75}
1279: %6
1280: van den Benkel, Phys. Stat. Sol.(a) {\bf 30}, 197 (1975).\bibitem{kub85}
1281: %7
1282: L.P. Kubin, and Y. Estrin, Acta Metall. {\bf 33}, 397 (1985).
1283: \bibitem{ana82}
1284: %8
1285: G. Ananthakrishna and M.C. Valsakumar, J. Phys. {\bf D 15}, L 171 (1982).
1286: The basic model was formulated in G. Ananthakrishna and D. Sahoo, J. Phys. {\bf D14}, 2081 (1981).
1287:
1288: \bibitem{NLI}
1289: %9
1290: {\it Non Linear Phenomena in Materials Science}
1291: Solid State Phenomena, Vols. 3-4 , edited by L.P. Kubin and G. Martin
1292: (Trans Tech, Switzerland, 1988) and other references therein.
1293:
1294: \bibitem{NLII}
1295: %10
1296: {\it Non Linear Phenomena in Materials Science II}, edited by G. Martin and L.P. Kubin (Trans Tech, Zurich, 1992), and other references therein.
1297: \bibitem{NLIII}
1298: %11
1299: {\it Non Linear Phenomena in Materials Science III}, edited by G. Ananthakrishna, L.P. Kubin, and G. Martin (Scitec, Zurich, 1995), and other references therein.
1300: \bibitem{ana95a}
1301: %12
1302: G. Ananthakrishna, C. Fressengeas, M. Grosbras, J. Vergnol, C. Engelke,
1303: J. Plessing, H. Neuh\"auser, E. Bouchaud, J. Planes, and L.P. Kubin, Scripta Metall. {\bf 32}, 1731 (1995).
1304:
1305: \bibitem{ana95b}
1306: %13
1307:
1308: G. Ananthakrishna and S. J. Noronha in \cite{NLIII}
1309:
1310: \bibitem{nor96}
1311: %14
1312: S. J. Noronha, G. Ananthakrishna, L. Quaouire, C.
1313: Fressengeas and L.P. Kubin, Int. J. Bifurcation and Chaos {\bf 7}, 2577 (1997).
1314:
1315: \bibitem{nor97}
1316: %15
1317: S.J. Noronha, G. Ananthakrishna, L. Quaouire and C.
1318: Fressengeas, Pramana - J. of Physics {\bf 48}, 705 (1997).
1319:
1320: \bibitem{ana93}
1321: %16
1322: G. Ananthakrishna, Scripta Metall. {\bf 29}, 1183 (1993).
1323:
1324: \bibitem{mul97}
1325: %17
1326: M. Bekele and G. Ananthakrishna, Phys. Rev. E {\bf 56}, 6917 (1997).
1327:
1328: \bibitem{mul98}
1329: %18
1330: M. Bekele and G. Ananthakrishna, Int. J. Bifurcation and Chaos {\bf 8}, 141 (1998).
1331:
1332:
1333: \bibitem{mau85}
1334: %19
1335: D. Maugis, J. Mat. Sci. {\bf 20}, 3041 (1985).
1336:
1337: \bibitem{tos95}
1338: %20
1339: {\it Physics of Sliding Friction}, edited by B.N.J. Pearson and E. Tossatti,
1340: NATO ASI ser. E, Vol. 311, (Kluwer Academic Publishers, Dordrecht, 1996).
1341:
1342:
1343: \bibitem{bod67}%21
1344: A. Rosen and S.R. Bodner, J. Mech. Phys. Solids {\bf15}, 47 (1967).
1345:
1346: \bibitem{kub86}
1347: %22
1348: L.P. Kubin, and Y. Estrin, J. de Physique {\bf 47}, 497 (1986)
1349:
1350:
1351: \bibitem{pen72}
1352: %23
1353: P. Penning, Acta Metall. {\bf 20}, 1169 (1972)
1354:
1355:
1356:
1357:
1358: \bibitem{jea93}
1359: %24
1360: V. Jeanclaude and C. Fressengeas, C. R. Acad. Sci., Paris, {\bf 316}, 816 (1993).
1361:
1362: \bibitem{leb95}
1363: %25
1364: M.A. Lebyodkin, Y. Brechet, Y. Estrin and L.P. Kubin, Phys. Rev. Lett. {\bf 74}, 4758 (1995).
1365:
1366:
1367: \bibitem{neu93}
1368: %26
1369: H. Neuh\"auser in {\it Dislocations in solids}, edited by F.R.N.
1370: Nabarro, North Holland, {\bf 6}, 319 (1993).
1371:
1372:
1373:
1374: \bibitem{arn93}
1375: %27
1376: A. Arneodo, F. Argoul, J. Elezgaray, and P. Richetti, Physica D {\bf 62},
1377: 134 (1993).
1378:
1379:
1380: \bibitem{kop92}
1381: %28
1382: M.T.M. Koper, P. Gaspard and J.H. Sluyters, J. Chem. Phys. {\bf 97},
1383: 8250
1384: (1992).
1385:
1386: \bibitem{bar88}
1387: %29
1388: D. Barkley, J. Chem. Phys. {\bf 89}, 5547 (1988).
1389:
1390:
1391: \bibitem{hau96}
1392: %30
1393: M.J.B. Hauser, and L.F. Olsen, J. Chem. Soc. Faraday Trans. {\bf92}, 2857
1394: (1996).
1395:
1396: \bibitem{den94}
1397: %31
1398: B. Deng, Int. J. Bifurcation and Chaos {\bf 4}, 823 (1994).
1399:
1400:
1401: \bibitem{raj99}
1402: %32
1403: S. Rajesh and G. Ananthakrishna (Communicated).
1404:
1405:
1406: \bibitem{rossler}
1407: %33
1408: O.E. R\"{o}ssler, Z. Naturforsch. {\bf 31A}, 259 (1976).
1409:
1410: %\bibitem{boi76}
1411: %38
1412: %J. Boissonade, J. Chim. Phys. {\bf 73}, 540 (1976).
1413:
1414:
1415: \bibitem{kub89}
1416: %34
1417: L.P. Kubin, and Y. Estrin, Scrip Metall. Mater. {\bf 23}, 815 (1989)
1418:
1419: \bibitem{ana94}
1420: %35
1421: G. Ananthakrishna, and H. Ramachandran, in {\it Nonlinearity and
1422: Breakdown in Soft Condensed Matter}, edited by K. K. Bardhan, B.
1423: K. Chakrabarti and A. Hansen, Lecture Notes in Physics, Vol.
1424: 437, (Springer-Verlag, Heidelberg, 1994).
1425:
1426: \bibitem{hes94}
1427: %36
1428: F. Heslot, T. Baumberger,B. Perrin, B. Caroli, and C. Caroli, Phys. Rev. E {\bf 49}, 4973 (1994).
1429:
1430:
1431:
1432: \bibitem{car95}
1433: %36
1434: See C. Caroli and P. Noziers in Ref\cite{tos95}.
1435:
1436: \bibitem{ana97}
1437: %37
1438: G. Ananthakrishna in {\it Nonlinearities in Complex Systems} edited by
1439: S. Puri, and S. Dattagupta, (Narosa Publishing House, New Delhi, 1997).
1440:
1441: \bibitem{mil98}
1442: %38
1443: A. Milik, P. Szmolyan, H. L\"{o}ffelmann, and E. Gr\"{o}ller, Int. J.
1444: Bifurcation and Chaos {\bf 8}, 505 (1998).
1445:
1446:
1447: \bibitem{eck83}
1448: %39
1449: W. Eckhaus, Lect. Notes in Math. {\bf 905}, 432 (1983).
1450:
1451: \end{thebibliography}
1452:
1453: \begin{figure}
1454: %\epsfxsize=8cm
1455: \epsfbox{fig1.eps}
1456: \caption{
1457: Schematic plot of the SRS. Branch $B^\prime D^\prime$ shown by the dashed line
1458: describes the negative strain rate sensitivity of the PLC
1459: effect. Dotted lines represent the discontinuous strain rate
1460: jumps leading to serrations in the stress strain curve.}
1461: \end{figure}
1462:
1463: \begin{figure}
1464: %\epsfxsize=8cm
1465: \epsfbox{fig2.eps}
1466: \caption{ Stress-time plot for single crystal of $Cu - 10\%at.Al$
1467: deformed at
1468: constant strain rate of $\dot{\epsilon} = 3.3 \mbox{x} 10^{-6} s^{-1}$.}
1469: \end{figure}
1470:
1471: \begin{figure}
1472: %\epsfxsize=8cm
1473: \epsfbox{fig3.eps}
1474: \caption{Phase diagram of the model in $(m,e)$ plane.
1475: See text for details.
1476: The broken line corresponds to the locus of Hopf bifurcations
1477: and dotted
1478: lines to the successive period doubling bifurcations.
1479: See text for details.}
1480: \end{figure}
1481:
1482: \begin{figure}
1483: %\epsfxsize=8cm
1484: \epsfbox{fig4.eps}
1485: \caption{
1486: Evolution of a trajectory (thin lines) along with the bent-slow
1487: manifold ($S_1$ and $S_2$ shown by thick lines)
1488: structure in the $x-\delta$ plane, for $m=1.2$ and $e=200$.
1489: Inset shows the time series of the $x$ variable (continuous line) and $z$
1490: variable (dotted line).}
1491:
1492: \end{figure}\begin{figure}
1493: %\epsfxsize=8cm
1494: \epsfbox{fig5.eps}
1495: \caption{
1496: Evolution of the trajectory along with bent-slow manifold ($S_1$ and $S_2$)
1497: structure in ($x,y,\phi$) space indicated by the gray plane, for $m=1.2$ and
1498: $e=200.0$.}
1499: \end{figure}
1500:
1501:
1502: \begin{figure}
1503: %\epsfxsize=8cm
1504: \epsfbox{fig6.eps}
1505: \caption{
1506: Evolution of the trajectory along the bent-slow manifold ($S_1$ and $S_2$) structure
1507: for $m=1.2$ and $e = 267.0$.}
1508: \end{figure}
1509:
1510: \begin{figure}
1511: %\epsfxsize=8cm
1512: \epsfbox{fig7.eps}
1513: \caption{Empty circles show the phase space projection of
1514: $\sigma$ vs $\dot{\epsilon}_p$ corresponding to a relaxation
1515: oscillation. The unstable fixed point is shown by a filled
1516: diamond. The dotted line through the fixed point represent the
1517: apparent negative SRS region. The thick lines are
1518: analytical approximations of corresponding regions.}
1519: \end{figure}
1520:
1521: \end{document}
1522:
1523: