LAGRANGIAN HAMILTONIAN MECHANICS Solutions lo the ExercisesLAGRANGIAN ees |) HAMILTONIAN MECHANICS Solutions to the ExercisesPublished by ‘World Scientific Publishing Co, Pe. Ld ‘POBox 128, Fare Road, Singapore 912805 USA office: Suite 1B, 1060 Main treet, River Bdge, NI 07661 UK offce: $7 Shelton Stet, Covent Gardea, London WC2H SHE, Brish Library Cataloguing-n-Publication Data ‘A catalogue record for this book is available from the British Libary. LAGRANGIAN AND HAMILTONIAN MECHANICS: SOLUTIONS TO ‘THE EXERCISES Copyright © 1999 by World Scientific Publishing Co. Pe. Lid. All rights reserved. Ths book. oF pars thereof, may not be reproduced in any form or by any means, ‘lectronic or mechanical, including photocopying recording or any information storage and retrieval {stem now known orto be invented, withou writen permission from the Publisher. For photocopying of material inthis volume, please pay a copying fee through the Copyright ‘Clearance Cente, Inc, 222 Rosewood Drive, Danvers, MA 01923, USA. In this ease permission to ‘Photocopy isnot required from the publisher. ISBN 981-02-3782-0 ‘This book sprinted on acid-ie paper. Printed in Singapore by UtoPrintPREFACE This book contains the exercises from the intermediate/advanced classical rangian and Hamiltonian Mechanics (World Scientific Pub. Co, Pte. 1996) wogeter wit their complete slations. In few of the exercises I hhave seen fit to make minor changes in these are marked by asterisks. Also, Tice fa Be 0 weal pe ‘mini research “The present wor is intended primarily for instructors who ae using Lagrangian ‘and Hamiltonian Mechanics in thelr course. It is Roped that it Se aera nd that twill accasonly provide new {ngghts, Instructors may also wish to photocop jtions to those exercises which their sudens have had prsulardficuty "tis book may aso be ved, topeer wit Lagrangian and Hamiltonian ‘Mechanics, by hose who are stodying mechanics on their own, In this case I strongly urge the individuals to make serious efforts to work out a substantial number of the relevant exercises on completing their study of each chapter and before looking at what I tave waiten bere. Only in iis way wil ch individuals come face to face with, and hopefully overcome, the various difficulties which the exercises present. Exercises, whether ental or physical, re meant tobe done, not read about! Melvin G. Calkin ‘alifax, Nova Scotia September, 1998Preface (CHAPTER I Exercise 1.01 Exercise 1.02 Exercise 1.03 Exercise 1.04 Exercise 1.05 Exercise 1.06 Exercise 1.07 Exercise 1.08 Exercise 1.09 Exercise 1.10 Exercise 1.11 Exercise 1.12 Exercise 1.13 Exercise 1.14 Exercise 1.15 Exercise 1.16 Exercise 1.17 Exercise 1.18 CHAPTER II THE PRINCIPLE OF VIRTUAL WORK AND CONTENTS NEWTON'S LAWS D’ALEMBERT'S PRINCIPLE Exercise 2.01 Exercise 2.02 Exercise 2.03 Exercise 2.04 Exercise 2.05 Exercise 2.06 Exercise 2.07 Exercise 2.08 Exercise 2.09 Exericse 2.10 u 2 12 4 4 15 18 2 32 33 38 a a1 41 sos as 48 49viii Contents CHAPTER IIT LAGRANGE'S EQUATIONS Exercise 3.01 Exercise 3.02 Exercise 3.03 Exercise 3.04 Exercise 3.05 Exercise 3.06 Exercise 3.07 Exercise 3.08 Exercise 3.09 Exercise 3.10 Exercise 3.11 Exercise 3.12 Exercise 3.13 Exercise 3.14 Exercise 3.15 CHAPTER IV THE PRINCIPLE OF STATIONARY ACTION OR HAMILTON’S PRINCIPLE Exercise 4.01 Exercise 4.02 Exercise 4.03, Exercise 4.04 Exercise 4.05 Exercise 4.06 Exercise 4.07 Exercise 4.08 Exercise 4.09 Exercise 4.10 Exercise 4.11 Exercise 4.12 st st 92. 53 54 55 7 58 SIRTIIAS B 8 83 87 88 1 93, 94 96Contents CHAPTER V__ INVARIANCE TRANSFORMATIONS AND CONSTANTS OF THE MOTION Exercise 5.01 Exercise 5.02 Exercise 5.03 Exercise 5.04 CHAPTER VI_ HAMILTON'S EQUATIONS Exercise 6.01 Exercise 6.02 Exercise 6.03 Exercise 6.04 Exercise 6.05 Exercise 6.06 Exercise 6.07 Exercise 6.08 Exercise 6.09 Exercise 6.10 Exercise 6.11 Exercise 6.12 Exercise 6.13 Exercise 6.14 Exercise 6.15 CHAPTER VII_ CANONICAL TRANSFORMATIONS. Exercise 7.01 Exercise 7.02 Exercise 7.03 Exercise 7.04 Exercise 7.05 Exercise 7.06 Exercise 7.07 Exercise 7.08 Exercise 7.09 Exercise 7.10 Exercise 7.11 ix 101 101 103 105 106 10 10 10 112 113 113 14 116 18 120 122 125 127 128 131 134 138 138 139 140 14 143 45, 146 148. 150 151 153,x Contents CHAPTER VII HAMILTON-JACOBI THEORY Exercise 8.01 Exercise 8.02 Exercise 8.03 Exercise 8.04 Exercise 8.05 Exercise 8.06 Exercise 8.07 Exercise 8.08 Exercise 8.09 Exercise 8.10 Exericse 8.11 Exercise 8.12 Exercise 8.13 CHAPTER IX ACTION-ANGLE VARIABLES Exercise 9.01 Exercise 9.02 Exercise 9.03 Exercise 9.04 Exercise 9.05 Exercise 9.06 Exercise 9.07 Exercise 9.08 Exercise 9.09 Exercise 9.10 Exercise 9.11 Exercise 9.12 CHAPTER X _NON-INTEGRABLE SYSTEMS Exercise 10.01 Exercise 10.02 Exercise 10.03 Exercise 10.04 157 157 158 161 164 165 168 im 174 7 178 181 184 188 192 192 194 197 199 207 2u1 213, 214 215 217 221 221 222 25 27CHAPTERI NEWTON'S LAWS Exercise 1.01 ‘A particle of mass m moves in one dimension x ina potential well Ve Vo tan®(rx/28) ‘where Vo and a are constants. Find, for given total energy E, the position x asa function Of time and the period + of the motion. In particular, examine and interpret the low energy (E << Vo) and high energy (E>> Vo) limits of your expressions. Solution Energy conservation yields. ‘fini? + Vo tan?(x/28) = E. This canbe rearranged to give “Ff ex . af costrey2a)dx 2 J ,,JB-Vown*@ay2a) V2 J, JE- C+ Vo)sin*(mx2a) ‘To do the x-integration, we set ‘The motion is thus given by (Fig. 1) »(2)- Re)2 (Chapter I: Newton's Laws Ex. 1.01, Fig. 1 ‘The motion is periodic with period coda, ‘The turning points +A of the motion are given by For B << Vo we have A/2a <> Vo we have w(t) 0, now dropping the arbitrary phase, EAE FE ++) m_ x RE Bek r-2na oe -Z Bi @n ste wih 0,9 Oahe2-~.Exercise 101 3 ‘Thats, xu [Eta tan, or ~ FE tea ep ‘The integers n,, must be chosen appropriately; see Fig. 2. Ex. 1.01, Fig.2 In this Limit the particle oscillates back and forth at constant speed -/2E/m between rigid walls at x. 28, ‘This is to be expected, since in this limit the potential approximates the “infinite square well potenial® V=0 for inlea V-+@ for ixiva4 (Chapter I: Newton's Laws Exercise 1.02 For each of the following central potentials V(r) sketch the effective potential Ven) Ey +VO, and use your sketch to classify and draw qualitative pictures ofthe possible orbits. @ — Voy= fe? 3D leotropic harmonic oeclator ©) V@)=-V¥, for r Eo the radius oscillates back and forth between turning radii ry and r, and the orbit looks qualitatively ike Fig. 2. Ex, 1.02, Fig. 2 ‘The motion ofa particle in this potential is studied in detail in Exercise 1.14. It tums out that the orbit is actually an ellipse with geometric center atthe force center (Fig, 3). Ex, 1.02, Fig. 36 (Chapter I: Newton's Laws (0) The effective potential is 12/2? -V, for ra [U2 /2ms* for r>a For r-+0 the term L3/2m2* dominates and Vyq —*®. For r-+0 Vgy—=0 from above. Further, Veq has a discontinuity of V, at r=a. There are two posibiltes for ‘Veer, depending on the magnitude of V,. These are shown in Fig. 4. For V, > L2/2ma? the effective potential goes negative (Fig. (a) whereas for V; and then increases back to infinity. The orbit looks qualitatively like Fig. 6.Exercise 1.02 7 Ex. 1.02, Fig. 6 In fact, since in this case the particle never encounters a force, the orbit is a straight line ‘which passes by the poteatial region, Fig. 7. ty Bx. 1.02, Fig.7 For situation (4) in Fig. (4) the orbit radius decreases from infinity to a minimum 13 <4 and then increases back to infinity, The orbit again looks qualitatively like Fig. 6, In fact, since in this case the particle encounters a force only at r=a the orbit is ‘composed of staght line segments, Fig. 8. Ex. 1.02, Fig. 8 (©) The effective potential is8 Chapter I: Newion's Laws The shape of Veq depends on the magnitude of the dimensionless combination of perimeters, Ufamk. For Lfamk<1, Vay =-W/? (Pig. 9(0)) whereas for Lm 1, Vey & +1? ig. 90). Ex. 1.02, Fig. 9a) Bx. 1.02, Fig. 900) For E>0 and L?/2mk < I (situation (1)) there is no inner tuming radius and the particle spiral in tothe force center. ‘The orbit sa capture orbit like Fig. 10. Ex. 1.02, Fig. 10 Indeed, since L= bImE where b isthe impact parameter, we can write the condition for capture as b? O and L?/2mk>1 (Gituation (3), the orbit is a scattering orbit like Fig. 6. @) The effective potential is ek Veal) = aaa For 1-+0 the term H/o" dominates and. Vuq~* 2. For r—» the tem 12/2? dominates and Veg -* 0 from above. There is one axis crossing atExercise 1.02 9 that is, at 1? = 2mk/L?. There is one extremum (a maximum) at Ver Pak Naar, +e @ ae + The value of the effective potential at its extremum is ‘Veqr(max) = L*/16m7k = b*E7/4k where b is the impact parameter. The resulting effective potential is shown in Fig. 11. Ver a, Ex. 1.02, Fig. 11 For E> Vyy(max) and thus b* < 4k/E (situation (1)) the orbit is a capture orbit like Fig. 10, The capeure cross section i8 Cage = bux = 20(K/E. For E-< Var (max) and initial orbit radius less than «[4mmk/L? (situation (2) the obit has an outer turing radius but aguin no inner turning radios and the orbit spirals in to the force center. For B< Vyq(max) and inital radius greater than /4mk/L? (situation (3) the orbit is & secanerng obit like Fig. 6. (© The effective potential is B Veale sey -k For 1+0 the tr 12/2m? dominates and Vyq >. For r—~-2 the tem 12/2mi? again dominates (the exponential causes the second term to decrease faster than any ‘negative power of r, as r-» © ) and Vyq -* 0 from above. "To detemine the shape of Vag in between, we fst check whether or not ther are any ani crossings Gn view ofthe ining Bebuvior there rust be nro or an eve bet of them) These oor tradi ow10 (Chapter I: Newton's Laws al?/amk = are™. ‘The righthand side of this equation, a a function of ar, is shown in Fig. 12. (oexpar)y _(@) 03679 4 / ©) 10 c Bx. 1.02, Fig. 12 ‘The peak in Fig. (12) is at ar~1 andthe height ofthe peakis Ye = 0.3679. Thu, ifthe dimensionless combination of parameters a.?/2mk is greater than 0.3679, there ae n0 axis crossings (situation (a)), and if aL?/2mk is less than 0.3679, there are two axis ‘rossings (situation b)). i ich WC BEM check whether or not here are any extreme, These occur at radi for al? mk = (ar/2XL+ are". The sight-hand side of this equation, as a function of ar, is shown in Fig. 13. (oxRXL+aryexp(-ar) @ 0.4200 + / @) 16 Bx. 1.02,Fig. 13 ‘The peak in Fig. (13) is at ar = (1+-V/5)/2 = 1.618 and the beight of the peak is 0.4200. ‘Thus, if al?/2mk is greater than 0.4200, there are no extrema (situation (a)), and if ‘aL?/2mk is less than 0.4200, there are two extrema (situation (b). ‘The possibilities for Vgq are shown in Fig. 14,Exercise 1.03 " a? 042 < a Bx. 1.02, Fig. 14(0) Ex, 1.02, Fig. 14(0) Bx. 1.02, Fig, 14(0) ‘Sinaations (1) and (2) in Fig. 14 give bound orbits as in Ex. 1.02, Fig. 2, and situations (3) and (4) give scattering orbits as in Ex. 1.02, Fig. 6. Exerclae 1.03 ‘The first U.S, satellite to go into orbit, Explorer 1, which was launched on January 31, 1958, had a perigee of 360 km and an apogee of 2549 km above the earth’ surface. Find: (@) the semi-major axis, (©) the ecceatriity, (©) the period, of Explorer I's orbit. The earth's equatorial radius is 6378 km and the acceleration dve to sgxavity atthe earth's surface is g = 9.81m/s?. Solution ‘The minimum and maximum radi of Explorer's orbit are a(1-¢)=360+6378~6738km and a(l+e)=2549+.6378 =8927km. ‘The semi-major axis is thus a= $(6738-+8927) = 7833km and the eccentricity is 8927 - 6738 i ee © Forragrag “O1*: The period ofthe orbits given by xo Bak, Tom Rather than looking up the gravitational constant G and the mass M of the earth in a handbook, it s simpler to observe thatthe gravitational field atthe surface of the earth is = GM/R? = 9.81m/s?, and thus oo 25, (7.833 x10)5 = 6895s = L92hr. ‘osix (6378x1072 ‘Chapter I: Newton's Laws Exercise 1.04 ‘Mars travels on an approximately eliptical obit around the Sun, is minimam distance from the Sun i aboot 138 AU and fs maximum distance i about 167 AU (I AU ‘mean distance fom Eath o Sun), Find (the semmajor exis, Diecaenty, ¢ : of Mar obit. Solution ‘The semi-major axis of Mars’ orbit is a= 4.386167 -15340, nd the cccenticiy is 167=138 “Lorei3e” 8% ‘The periods given by xo pe Tou" Rather than looking up the gravitational constant Gand the mass M of the Sun, simpler to observe that for Barth year AU, snd thus for Mars ‘sin years = (a, in AU = (1.53) = 1.88, Exercise 105 ‘The most economical method of traveling from one planet to another, the Hohmann transfer, consists of moving along a (Sun-controlled) elliptical path which is tangent to the (approximately) circular orbits ofthe two planets. Consider a Hohmann transfer from Earth (orbit radius 1.00 AU) to Venus (orbit radius 0.72 AU). Find, in units of AU and year: (@) the semi-major axis ofthe transfer orbit, () the time required to go from Earth to Venus, (©) the velocity “kick” needed to place a spacecraft in Earth orbit into the transfer orbit.Exercise 1.05 13 In this problem ignore the effects of the gravitational fields of Earth and Venus on the spacecraft. Solution (2) The maximum radius ofthe trnster orbit equals the radius of Earth's rit (1.00 AUD, and the minimum radius equals the radius of Venus’ orbit (0.72 AU). The semi-major axis of the transfer orbit i = 40.00 +0.72)=0.86 AU. (©) The period « ofthe transfer orbit is (see Exercise 1.04) roa = (0.86% = 0.80 year. ‘The time to go from Earth to Venus is half the period, $(0.80) = 0.40 year. (©) The energy ofthe spacecraft, per unit mass, at Earth orbit radius rg is given by 126M ae te where v isthe speed ofthe. (On the right we have expressed the total energy of, Be pacer fr ut mika aso be ekrunjr bls Ce males Tas equation gives ‘Newton's second law shows thatthe speed vp of the Earth around the Sun is given by vy M. a 2 y 5 1 (Y\ og- fap t (3) -2-B--ae 30 v/vp =0.915. Hence the velocity "kick" Av= v- vg needed to place the spacecraft into the transfer orbit is given by Av/vg =~0.085. Since the speed of the Earth is ‘Yq = 2m AU/year, this yields Av = ~0.534 AU/year (= -2.53km/3).4 Exercise 1.06 Halley's comet travels around the Sun on an approximately elliptical orbit of eccentricity = 0.967 and period 76 years. Find: (a) the semi-major axis of the orbit (Ans. 17.9 AU), (b) the distance of close sh of Halley's comet to the Sun (Ans. 0.59 AU), (©) the time per orbit that Halley's comet spends within 1 AU of the Sun (Ans. 78 days). Solution (@ The semi-major axis ofthe orbit of Halley's comet is given by Kepler's third Inw (see Exercise 1.04), a= v% = (76)% = 17.9 AU. (®) The minimum distance of Halley's comet to the Sun is a(1~e) = 0.59 AU. © The inwantaneous disace of Halley’ comet tothe Sn andthe centre anomaly y Eat-ecasy. {In particular, the eccentric anomaly corresponding to a distance of 1 AU is given by 1 Fag 7! 0.6 Teosy, ‘so = 0.218 radians. The corresponding time is then given by Kepler's equation vresiny = @a/a}t which becomes 0.218 -0,967sin0.218 = 25/76), 80 t= 0.107 year =39 day. The total time per orbit that Halley's comet spends within TAU of the Sun is 2x39 = 78 days. Exercise 1.07 Define a "season asa time interval over which the true anomaly increases by 1/2. Find the duration ofthe honest season for earth. Take the eccentricity of ear’ crit tobe Solution ‘The tre anomaly 0 for the earth, measured from perihelion, and the eccentric anomaly 'y are related by1s where ¢ = 0.0167 isthe eccentricity of earth's orbit. This equation can be waitten, e+ cos8 Trecosé cosy =: For the shortest “season” the true anomaly increases from ~7/4 to 2/4, and the eccentric ‘anomaly increases from ~y to y where cosy = 20167 cos/) | — 140.0167 cos(74/4) ate ‘This yields y = 0.7737 radians. Kepler's equation (with tin years), 2nt = y~esiny = 0.7737 ~0.0167sin0.7737 = 0.7620, then shows that t increases from -0.1213 to 0.1213, so the duration of the shortest "season" is 2x 0.1213 = 0.2425 year = 88,59 days. "Compare this with "winter" (= 89.00 days) Exercise 1.08 A satellite of mass m moves ina circular orbit of radius ao around the earth. (2) A rocket on the satellite fires a burst radially, and as a result the satellite acquires, ‘essentially instantaneously, a radial velocity u in addition to its angular velocity. Find the seri:mujer nis the eecenoicity, and the aientaton ofthe eliptel obit into which the satelite is thrown. (b) Repeat (a), if instead the rocket fires a burst tangentially. (©) in both cases find the veloity ick required to throw th stlit into parabolic it Solution [Newton's second law shows that mvj _ ik mye, a 0 the initial speed of the satelite is vp = /k/miag. The initial energy is odavg-£--Lmg ==, Eo=>mvo 5652 ea!16 (Chapter I: Newton's Laws ‘and the initial angular momentum is Lo = mVvotg = akg - (a) A radial kick changes the energy of the satellite. Since the velocities vo and u are perpendicular to one another, the final energy is E-Ep+ imi? Expressing the inal and final energies in terms ofthe semi-major axes ofthe crit, we ve ‘This can be rearranged to give the relation between the intial and final semi-major axes, a Ba-3. ‘A radial kick does not change the angular momentum of the satelite, 0 BoB ink” mk" Expressing this in terms ofthe semi-major axes and eccentricities ofthe orbits, we have a(1-e?)= a9. Substituting the preceding expression for ag/a into this equation, we obtain the eccentricity of the final orbit em luv. ‘The equation for the final orbit is Beisecosd F ‘where 0 isthe angle from pericenter. At the kick r = ap and, for u positive, is increasing. ‘The kick thus occurs at @ = +n/2. ‘That is, pericenter is at —n/2 from the point of firing of the rocket.Exercise 1.08 ” For lul-- vo the eccentricity ¢-+1 and the semi-major axis a» such that a(1~e)-+ ag/2. The satellite is then thrown into a parabolic orbit. The energy put into the system is av, othe final tot energy is zero, as required. (®) A tangential kick changes the energy of the satellite. The final energy is 1 2 Eo By + mvgu+ tm (v9 +) = Em Bo +mvgn + 2 [Expressing the energies in terms of the semi-major axes, we have k E+ mvgu+ tu", ea ‘This can be rearranged to give the relation between the initial and final semi-major axes, Seei2. % ‘A tangential kick also changes the angular momentum of the satellite. The final angular momentum is L= m(¥o + u)ag = Lo(1 + u/vo). and thus Expressing this in terms of he semi-major axes and eccenticites of the obits, we have oll w/v). st de tg sin ho i gn ei en {Jere ade)8 ‘Chapter I: Newton's Laws For O 22 -1 (the finl velocity is then vo += £2 v) the eccentricity ¢-+ 1 and the semi-major axis a-+ o. The satellite is then thrown into a parabolic orbit. ‘The energy put into the system is Jincey ve)? mv -my3, so the final total energy i 270, a required. tis also of interest to consider the case u/vo = 1 (for which the velocity of the satellite immediately after rocket firing is zero). We then have a ao/2 and e=1, and the elipae degeoeites into a stnight ine, the satelite faling sought in towards the center of heath, me * Exercise 1.09 ‘Show that the following ancient picture of planetary motion (in heliocentric terms) is in sccord with Kepler’s picture, ifthe eccentricity eis small and terms of order ¢? and higher (@) the earth moves around the sun ina circular orbit of radius a; however, the sun is not atthe center of this circle, but is displaced from the center by a distance ea; () the earth does not move uniformly around the circle; however, a radius vector from a point which is on a line from the sun to the center, the same distance from and on the Opposite side of the center as the sun, tothe earth does rotate uniformly.Exercise 1.09 19 Solution @ Earth vs Zz o Center Sun Ex. 1.09, Fig. For an off-center circular orbit (Fig. 1) we have a? ar? seta? + 2earcosd, 1s follows from the trigonometric cosine law. This yields EAfeTaATG~eons0 1 -eon8 feat ‘On the other hand, for an elliptic orbit we have (Lagrangian and Hamiltonian Mechanics, pages 12 and 13) 2 1=t 21 ~ecos0~e*sin 0+ a” Trecos8 “The two expansions agree to onder © Earth a ‘Center "Sun Ex. 1.09, Fig.2 Applying the trigonometric sine law to the two smaller triangles in Fig. 2, we have sin($-or)=esinot and sin(@-4)=esind.Newton's Laws ‘These yield, to order €”, Grotsesinate- and 0-9 sexing s fe?sin2g 0 forthe “ancient picture” of planetary orbits we have = t+ 2esinut +6? sin2ate--- On the other hand, for a Keplerian orbit we have (Lagrangian and Hamiltonian ‘Mechanics, page 18) @-2esind + fe” sin26+-- Inversion gives On ots 2esinuts fePsindans-, which agrees with the "ancient picture" to order e. Exercise 1.10 (@) Show that 20-1 50088 (the standard form for a conic section, on setting the eccentricity € = 1 and the semi-latus- rectum p= 2ro) is the equation of a parabola, by translating it into cartesian coordinates ‘with the origin at the focus and the x-axis through pericenter. (©) A comet travels around the Sun on a parabolic orbit. Show thatthe distance r of the ‘Comet from the Sun i related tothe time t from peribelion by Besng fae ‘where distances are measured in AU and time is measured in (©) If one approximates the orbit of Halley's comet near the Sun by a parabola with 7 = 0.59 AU, what does this give forthe time Halley's comet spends within 1 AU of the ‘Sun? (@) What is the maximum time a comet on parabolic orbit may spend within 1 AU of the Sun?Exercise 1.10 a Solution (@) The equation for the orbit, 22-1 +0090, can be transformed into cartesian coordinates (x.y) by multiplying it through by r and setting =x? +y? and roos0 =x. We obtain 2to-xa qty. ‘Squaring and simplifying this, we then find a y x00 ‘which is indeed the equation of a parabola in cartesian coordinates. () To find out how r depends on time, we tum tothe energy equation and set E=0 and L? » 2mkro to obtain 1a. Ky qn +p Muliplying by +, rearranging, and integrating then gives, PE, f'n of" [yt feof ghee [ breeze Since E[ii = VGH = 25 AUM/year, this becomes ant 242i ‘with distances now in AU and time in years. fr=%.2 (Chapter I: Newton's Laws (© For rq =0.59AU and r= LOOAU the preceding equation gives t= 0.1047 year, so the time, in the parabolic approximation, that Halley's comet spends within 1 AU of the sun is 2t = 0.2094 year = 76.5days. Compare this with the actual time of 78 days (see Exercise 1.06). (@) We set r= 1.00 and adjust ro to give maximum time t. This occurs for at v2 1429) _ 23: mi Sere tp St sete mma rare E (tered) rE 3, so t= (I/3n)year = 38.7Sdays. ‘The maximum time a comet on a parabolic orbit may spend within 1 AU of the sunis thus 2t = 77,Sdays. Exercioe 1.11 [A panicle of mass m moves in a central force field F = ~(k/"*)8. (@) By integrating Newton's second law dp/dt =F, show that the momentum of the particle is given by p= po +(mk/L)®, where Po is a constant vector and L is the magnitude ofthe angular momentum. (©) Hence show thatthe orbit in momentum space (the so-called hodograph) is a circle ‘Where isthe center and whats the radus of the circle? (©) Show thatthe magnitude of po is (aak/L)e, where e i the eccentricity. Sketch the mbit in momentum space for the various cases, e=0, O 1, indicating for the lat two cases which part of the circle is relevant. (Gee: Amold Sommerfeld, Mechanics, (Academic Pres, New York, NY, 1952), tans. Marin O. Stem, p. 33, 40, 242; Harold Abelson, Andrea diSessa, and Lee Rudolph, “Velocity space andthe geometry of planetary orbits,” Am. J. Phys. 43, 579-589 (1975), Solution ‘According to Newton's second law the momentum p of « paricle moving ina central force field -(k/1)# changes at a rate oo oat aa Inthe socond equality we have multiplied and divided by m8, The reason for this is that we recognizeExercise 1.11 B Lem as the magnitude of the angular momentum, and we also note that, 2. 6g. at ‘The preceding equation becomes, ‘Since the angular momentum L of a particle moving in a central force field is constant, ‘we can integrate immediately to obtain where pi cosant vector. To identify this vector, we tke the €-componet ofthe equation. Since the @-component of pis Le, andthe @-component of py is pocas® where © isthe angle between py and 8, we Bind L_ mk F-TE + pocoss. ‘This can be written 2 L ni + PE cose, 008 which we recognize as the equation of a conic section with eccentricity © = poL/mk. ‘The magnitude of pp is thus = Pome We also recognize @ as the angle from pericenter, so the direction ofthe vector po is perpendicular toa line from the fore centr to pericenter. ‘Another way to obtain the magnitude of Po is to write mk, m3oy (Chapter I: Newton's Laws In the second equality we have used the fact that the 8-component ofp is L/r. The first two terms on the right are 2mE. where E is the toal energy. The equation thus becomes mk? ee ; whee on fi 2D sete emt sin tr ie nyt ep ‘momentum, This then retums us to our previous expression for po, We now note that mk pppoe. ‘This is the equation of a circle, in momentum space, with center at pp and with radius mk/L. Since po = (mk/L)e, the circle has its center at the origin for circular (e = 0) orbits (Fig. 1(8)), encloses the origin for elliptic (0 I) orbits Fig. 1(). CO@@ O I whereas waves slow down. (©) Suppose now thatthe particle is incident at impact parameter bon an attractive square ‘well potential (ig 2). 80s. Ex. 1.17, Fig.2 The incident angle Oy is given by sin = b/a and the refracted angle 8, by sind = bjna ‘The scattering angle is6 Chapter I: Newton's Laws © = 26 -0)), and thus con 0/2)= conty end + sntsin ob? . he -a +P. ‘The scattering angle ranges from zero at impact parameter zero to a maximum of 2cos"'(/n) at impact parameter a. We can instead express the impact parameter in terms of the scattering angle. Rearranging and squaring the preceding equation, we find e (9-35) (3-2) which simplifies to bt atsin? ya a?” n+ 1-2nc0s@/2 (with 0 < @ <2cos"*(I/n)). For small impact parameter this becomes o=S-4 Letf, the “focal eng. be the distance from the force centr to he point at which the scatiered parle crosses the ais of symmety. To find this distance, we drop been rom ie foe emer ont tio he incoming and egg Eajeetores. ‘The length of that onto ibe incoming injector is the in er, tn syne imple ta te lngh of exe perpendclas weequa Ge Bx. 1.17, Fig.3 We see that f= bjsin®,Exercise 117 37 2030/24) ncos@/2 | For small impact parameter the focal length becomes fo X=) tad is independent ofthe impact parameter, The particles are then brought o@ focus 8 distance f from the force center. ‘These remarks on the focal length hold only for weak attractive potentials. Otherwise, the geometry changes with the particle crossing the axis inside the potential region, (©) The differential scatering cross section is given by dob [ab a” Snelael" 2b dd _ n? sin(@/2)c0s(@/2)___n? sin?(@/2) a d@ ~ n? +1-2ncos(@/2) (n+ 1-2ncos(@/2)" (=? sin(@/2)(nc0s(@/2) ~ 1)(n ~ cos(@/2)) (a? +1-2ncos(@/2)) : we have do __nPa?_ (ncos(€/2)~ In - c0s(@/2)) GQ” 4e0s(872)—(n? +1 2ncos(@/2)* For large index of refraction (strong atractive potential) the differential searing cross section becomes ‘and is the same as that for scatering from a hard sphere (strong repulsive potential); see Exercise 1.16. To understand this, note that the outgoing trajectories in the two cases are in exactly opposite directions for @ given impact parameter (Fig. 4), so the cross-sections, ‘which are invariant under inversion, are equal.38 (Chapter I: Newton's Laws Repulsive Atactive Ex. 117, Fig. 4 Exercise 1.18 (@) Show that $2 cosa8 isthe equation ofthe orbit for a panicle moving in a repulsive potential V(x) = k/r?, determining andro in terms ofthe energy and angular moment. a) T° ame (Show thatthe impact parameter band saterng angle @ are related by (Ans. k (x-6F 2 * Ee@x-6) (© Show thatthe differential scatering cross section is given by do tk x-8 GO” Esind 8 2x- OF aExercise 118 39 Solution (@) The orbit equation is given by . ee 2 Pae= Paes where a? = 1+ 2mk/L?. To do the rintegraton, we first set u= jr, du =—dr/x? to obain c éu @),, Pamela We then set u=(V2mE/aL)cosA, du = ~(V2mE/aL)sinAdA. Integration gives @ = A/a, so the equation of the orbit is. us l/r (V2mE/aL)cosa0; thatis, ro/t cosa where rp = ol/-YimE. (©) Pericenter is at @ = 0, and for r+ the angle a0 -+ =3/2, 0 0, = n/2a. Expressing @. in terms ofthe scattering angle © by using © = x~2., and «tin terms ‘of the impact parameter b by using L = bv2mE., we find 1 x ae E 2 2+ Keo Salvingforb we aban yak a0? “E@Qx-8)40 (Chapter I: Newton's Laws (©) The differential scaring cross section is given by do __b |ab| a0” Fn6|a0) Since we have x-@ 42” Esin@ @72x- OF For small scattering angles this becomes ‘whereas for @ +m it becomesCHAPTER II ‘THE PRINCIPLE OF VIRTUAL WORK AND D'ALEMBERT'S PRINCIPLE Exercise 2.01 10) i F 1 yms Use d'Alembert principle to find the condition of static equilibrium. Solution Imagine a virual displacement in which the angle O increases by a small amount 30. ‘The mass moves horizontally a distance 8x = £03668, and the applied force F does work F5x = F£cos060. The mass moves vertically upwards a distance by = fsin080, and the applied force mg does work ~mg6y=-mgésin80. ‘There are no inertial forces, so d'Alember's principe gives £cos050-mgésin880 = 0, which simplifies to F=mgind ‘This is the condition of static equilibrium, Exercise 202 mg ‘Use d’Alemben's principle to find the condition of static equilibrium.a (Chapter 1: Principle of Virtual Work ‘Solution Imagine a viru displacement in which the end of the string is mised a distance dx and ‘the weight is thus raised a distance &x/2. The applied force F does work F4x and the applied force mg does work -mg5x/2. There are no inertial forces, so d'Alembert's ‘rnciple gives Fox-mgdx/2 =O. ‘This yields F = mg/2, the condition of static equilibrium. Exercise 203 at ut my Use d’Alemben's principle to find the acceleration of my. Solution Imagine a virtual displacement in which m, moves downwards a distance 5x and mz moves upwards a distance 5x. The applied force, gravity, does work myg8x on m and =myg6x on mz. The acceleration of my is X downwards, and thet of my is upwards. ‘The inertial force on my is m,% upwards and the inertial work done on my is m8 ‘The inertial force on ma is m% downwards and the inertial work done on ma is ~m,X8x. D’Alembers principle gives (2m, ~ mg)g8x ~ (my + m_)%8x= 0, $0 the acceleration of my is xe™M=™y, mem;Exercise 2.05 4a Exercise 2.04 4 sm, i Use d'Alembert’ principle to find the acceleration of my Solution Imagine a viral displacement in which m, moves downwards a distance 8x and mz thus moves upwards a disunce 81/2. The applied fore, gravity, does work m,g6x on m, and ~m2g6x/2 on mz. The acceleration of m, is downwards, and that of ma is 5/2 upwards. The inertial force on m, is m& upwards, and that on ma is my /2 downwards. These inertial forces do work -m,%8x on my and ~(t 5/2)(6%/2) on mp. DiAlemben’s principle gives ( ~ $ma)g8x- (en, +4m2)¥5x-=0, so the acceleration of my Exercise 2.05 2 es Use d'Alembert’ principle to find the acceleration of mj. Note that in this case the polley has an upward acceleration A. “Acceleration” means "acceleration relative to the earth“ (Chapter I: Principle of Virtual Work Solution Imagine a virtual displacement in which m, moves downwards a distance &x and mz moves upwards a distance 8x. The epplied force, gravity, does work mjg6x on m, and ‘=mgg0x on ma. The acceleration of m, is (X- A) downwards, and that of my is (+A) upwards. The inertial force on m, is m,(X- A) upwards, and that on m3 is m,(%+A) downwards. These inertial forces do work -m,(%-A)bx on m, and =m,(3 + A)&x on my. DAlember’s principle gives smgdx- mzg8x - m,(j - A)bx- m3 + A)Bx= 0, which simplifies to xe MoMesa). In the frame of the pulley this system, and thus the downward acceleration X of m, the same as an "Atwood! machine” (Exercise 203) in a gravitational field g+A. Inthe frame of the earth the downward acceleration of my is s-n- icmadE=tmah, Suppose m,>m. Then, if the upward acceleration A of the pulley is small, m, accelerates downwards and ma accelerates upwards. However, if A> MaMa, both masses accelerate upwards. Exercise 2.06 |: ‘A mass mis attached to alight cord which wraps around frictionless pulley of mass M, ragius R, and moment of inertia I= fr?dM. Gravity g acts vertically downwards. Use <'Alemberts principle to find the acceleration of m.Exercise 2.07 45 Solution Imagine a virtual displacement in which m moves downwards a distance 8s. The applied fore, gravity, does work mg8s on m. The inertial force on m is mi upwards and the ‘work it does is ~mi6s. There is no applied work done on the (uniform) pulley, but there is inertial work. To find this, consider a litle piece dM of the pulley which is at a distance + from the axle. This piece undergoes a virtual displacement, in the angular direction, of (1/R) Bs. The acceleration of dM, in the angular direction, is (s/R)8 so the inertial force on dM, in the angular direction, is ~dM(1/R)8. The inertial work done on 4M is thus -4M(1/R)?86s_(4M also has a radial centripetal acceleration, but this does ‘aconuibue to thee wert). Summing, we hve forthe tot ineal wrt done on y =f? a/R? 88s = ~(/R?)3d8, where I~ fr2dM is the moment of inertia of the pulley. ‘This expression forthe inertial ‘work done in a fixed axis rotation can also be written in the useful general form ~1a8@ where a =3/R is the angular acceleration and 50=s/R is the angular displacement. Returning othe orginal problem, e'Alembert’s principle now gives imgbs~ mits ~(1/R2)ibs «0, ‘0 the acceleration of m is A cylinder of mass M, radius R, and moment of inertia I~ fr?dM rolls without slipping down an inclined plane. Use d’Alembert's principle to find the acceleration of the cylinder.46 ‘Chapter 1: Principle of Virtual Work Solution Imagine a virtal displacement in which the cylinder rolls a small distance 6s down the plane. The only applied force, gravity, does work swered « Masinads. To find the inertial work is more difficult. In this virtual displacement the center of mass of the cylinder is translated s and, because the cylinder rolls, it rotates about its center of mass through an angle 80 s/R. A litle piece dM of the cylinder at (r,0) thus ‘undergoes a displacement &s down the plane together with a displacement £60 = (&s/R) in the “angular direction." This latter displacement has components ~r(6s/R)cos@ down the plane and r(8s/R)sin perpendicular to the plane (Fig. 1). “down the pane 180 sind] fant 188 cos8 Ex. 2.07, Fig. 1 4M thus undergoes @ net displacement &s(1~(r/R)cos®) down the plane and {s(r/R)sin8 perpendicular to the plane. To find the inertial force on dM, we need its ‘cceleration. ‘The acceleration of the center of mass of the cylinder is § down the plane, and the acceleration of 4M with respect to the center of mass has components x(3/R) in the "angular direction” and -r(8/R)* (centripetal acceleration) in the “radial direction.” Resolving these into their components parallel and perpendicular to the plane, we find the ‘et acceleration of dM to be 3 ~1@/R)cos0 + r/R)*sin@ down the plane and r@/R)sin® +r(/R)Fcos0 perpendicular tothe plane. ‘A more analytical way to obtain these results isto note that the cartesian coordinates of OM are xes-rsin@, y= R~reos0, where x is "down the plane” and y is "perpendicular to the plane.” Differentiating once, ‘we obtain for dM a virtual displacement bx=8s~rc0s080, by rsin080.Exercise 2.08 a Differentiating twice, we obtain for dM an acceleration = 5- rhcos6 + 167sind, ¥~ rOsin® + 16? cose. ‘The inertial force on dM is -dM times the acceleration, so the inertial work done on MM is AMIS ~ 1(/R)cos0 + (YR)? sinBTSs(1~ (s/R)cos0] ~dM[s(/R)sinO + r(4/R)? cos OL8s(*/R)sin 6) ‘= -¢MS[1 - 2(¢/R)cos0 + (1/R)*]8s - dM(7/R)(4/R)sin ds . (On integration over the cylinder the angle-dependent terms go out and we obtain awl. Ms (1 + 1/MR?)5s where I= {174M is the moment of inertia ofthe cylinder about its center of mass. This expression can be writen in the general form ~MS6s—1a68, where a is the angular acceleration of the cylinder and 88 is its angular displacement. “Applying d Alember’s principle tothe original problem, we find MgsinaBs ~ Mi(L+1/MIR?)8s~= 0, so the acceleration of the cylinder down the plane is +m Exercise 2.08 IRS a =Q] = Use d'Alemben’s principle to find the acceleration of m, down the (stationary) plane. Solution Imagine a virtual displacement in which m, moves down the plane a distance &s. The ‘mass m, then undergoes a downward vertical displacement Assina. and the appli48 (Chapter I: Principle of Virtual Work force, gravity, does work myg6ssina on it. As well, ma moves vertically upwards a distance &s, and gravity does work ~m2g8s on it. Ifthe acceleration of my is" a” down the plane, the inertial force on it is m,a up the plane, and the inertial work done on itis ‘mats. The acceleration of m2 is then "a" vertically upwards, the inertial force on itis ‘maa vertically downwards, end the inertial work done on it is ~m,ads. D'Alemben’s principle gives mygsina8s~m,g8s-m,abs-m,a8s = 0, ‘0 the acceleration of m, down the plane is msina-m, as i m+, ‘A block of mass m slides on a frictionless inclined plane, horizontally, the displacement of the plane at time t being some known function x(t). Use d’Alember’s principle to find the equation of motion of the block, taking as generalized coordinate the displacement s of the block down the plane. Note that the acceleration of the block is not "down the plane.” Solution Imagine a virtual displacement in which s increases by &s. The mass m undergoes a downward vertical displacement Sssina. and the applied force, gravity, does work rmg8ssina. The acceleration a of m is the vector sum of % down the plane and % horizontal (Fig. 1.Exercise 2.10 49 ‘The inertial foree on the mass is -ma. We need the component ofthis in the direction of the vinwal displacement, namely -m(+ cosa). The inertial work done is thus ~m(S~ Xeosa)és.. D'Alember's principle gives mgsinads~_m(+ kcosa)as = 0, ‘0 the equation of motion ofthe block is Sm gsina ~xcosa. For i= 0 this reduces to the usual result for an inclined plane. For %> gtana., however, the acceleration § is negative and the mass m accelerates up the plane. Exercise 210 ‘A block of mass m slides on a frictionless inclined plane of mass M which in tur is free to slide on a frictionless horizontal surface. Use d'Alembert's principle to find the equations of motion of the block and the plane, taking as generalized coordinates the displacement sof the block down the plane andthe horizontal displacement x of the plane. Solution ‘The system in this exercise has two degrees of freedom, and there are two independent virtual displacements. ‘The fist can be taken as in Exercise 2.09, an increase of s by &s. ‘The application of d'Alembert’ principle for this virtual displacement then proceeds as in Exercise 2.09 and leads to the fst equation of motion m(+ cosa) = mgsina. ‘The second independent virtual displacement can be taken to be an increase of x by Ox. In this virwal displacement both m and M move horizontally, so the applied force, ‘gravity, does no work. The incline M has & horizontal acceleration , so the inertial Work done on it is -M&dx. ‘The mass m has an acceleration a asin Exercise 2.09, with @ horizontal component (cosa +%). The inertial work done on m is thus =m@cosa + )8x. D’Alembert’s principle gives -Mibx—_mGcosa + 3)8x = 0,50 (Chapter it: Principle of Virtual Work which simplifies to (cosa + 8) = -Mi ‘This second equation of motion can be integrated to yield micosa +(m +M)% = constant. In this form the equation states the fact that the horizontal component of the total linear ‘momentum ofthe system is constant.CHAPTER II LAGRANGE'S EQUATIONS Exercise 3.01 fhe m ‘A bead of mass m slides without frietion along a wire which ha the shape ofa parabola y= Ax? with axis vertical inthe eanh’s gravitational Geld g Fin de Lagrangian king a geerned coor the hsm spleen x Solution (@) The kinetic energy ofthe bead is To bm? +32)= dma + 4a2e?ye?, and the potential energy is Ve=mgy=mgAx?, ‘The Lagrangian is thus ‘L(x,8) = fm(1 + 4A2x?)x? - mgAx?, @) We have Remassarys, Eo pm(@atsyx?-2mgas, Say 2.2)8 + mQA2K? Sele) 7m AAP + BA ‘0 Lagrange’s equation of motion is m(L+ 4A%x2)x = -m(4A%x)x? - 2mgAx,2 (Chapter It: Lagrange's Equations Exercise 3.02 ‘The point of suppor of a simple plane pendulum moves vertically according to y= h(t), ‘where h(t) is some given function of time. (@) Find the Legrangian, taking as generalized coordinate the angle @ the pendulum ‘makes with the vertical. (b) Write down Lagrange's equation of motion, showing in particular thatthe pendulum behaves like a simple pendulum in a gravitational field g +, Solution (@) The cartesian coordinates ofthe bob are (Fig. 1) xelsind, — y=h(t)£cos0. y he e teosa \e ho ~ £0086 | sino wind x Ex. 3.02, Fig. 1 ‘The cartesian components of the velocity ofthe bob ae thus ke Micns8, pha tdsind, and the kinetic energy is To (i? +9) = $m(6? + 2iebsin +h), ‘This result can instead be obtained by using the trigonometric cosine law to add the velocity fin the vertical dretion tothe velocity £8 in the angular direction Fig 2). Ex. 3.02, Fig. 2Exercise 3.03 3 ‘The potential energy is V-=mgy = mg(h- ¢e0s6). ‘The Lagrangian is L=T-V = 4m(0°6? + 2h¢0sind + i?) - mg(h - £c0s6). © We have aL. Chas Fe- mC hesine), — T= miedcosd-mgtsind, —!—} ~ m(06 + iesin® + hedcos®), 0 Lagrange's equation is mb = -m(g + fesin8. “This is the same as the equation of motion of a simple pendulum in a gravitational field ah. Exercise 3.03 ‘A mass m is attached to one end of a light rod of length ¢. ‘The other end of the rod is pivoted so that the rod can swing in a plane. The pivot rotates in the same plane at angular velocity « in a circle of radius R. Show that this "pendulum" behaves like a ‘simple pendulum in a gravitational field g = Ro? forall values of ¢ and all amplitudes of ‘oscillation. Solution ‘The cartesian coordinates ofthe mass are (Fig. 1) K=Reosot+ Loos(at+0), y= Rsinat+ fsin(ot +6). Ex. 303, Fig. 154 (Chapter Ill: Lagrange's Equations ‘The cartesian components of the velocity of the mass are thus ie -Rosinat-Ko+Osin(or+0), y= Rocosut+ Mw +6)cos(ot +8), and the kinetic energy is T= }m(i? +5?) }mfR7a? + 2RL0(e + 6)e080+ (w+ 6} ‘There is no potential energy, so the Lagrangian is simply T. We have at pl ar Fp TMRwcos0+mea+6), FF 2D). mR twBsind + mes arlgp) 7 mRCodsing. m6, mRtoo(w +6)sind, so Lagrange's equation is mé*§ = -mRw*tsind. ‘This isthe same as the equation of motion of a simple pendulum in a gravitational field Ra? Exercise 3.04 A pendulum is formed by suspending a mass m from the ceiling, using a spring of ‘unstretched length 4 and spring constant k. (@) Choose, and show on a diagram, appropriate generalized coordinates, assuming that the pendulum moves in a fixed vertical plane. (b) Set up the Lagrangian using your generalized coordinates. (©) Write down the explicit Lagrange's equations of motion for your generalized coordinates. Solution Ex, 3.04, Fig. 1Exercise 3.05 55 If we choose as generalized coordinates the cartesian coordinates (x.y) with ‘origin atthe point of suspension end with x horizontal and y vertically down (Fig. 1), the Lagrangian is Lm dee? + 5?) + may - diy? + y? - 09)?. ‘The second term is (minus) the gravitational potential energy and the third term is, (minss) the potential energy due to the stretch of the spring. Lagrange's equations are mie -kx(l-fo/r), my =mg-ky(1- 6/2), z Ray’ If, instead, we choose as generalized coordinates the polar coordinates (r,0) with + the distance of the mass from the point of suspension and @ the angle the spring makes ‘with the vertical Fig. 1), the Lagrangian is with © L= fm(@? +196) + mgrcos fh(r— 0). Lagrange's equations are then if = m6 + mgcos0- K(r~ f0)s 4(mus6)/dt = mr6 + 2mrif = -mgrsin® . Exercise 3.05 {A double plane pendulum consists of two simple pendulums, with one pendulum ‘suspended from the bob of the other. The “upper” pendulum has mass m, and length £,, the "lower" pendulum has mass my and length f, and both pendulumes move in the same vertical plane. (@ Find the Lagrangian using as generalized coordinates the angles ©, and 0, the Pendulums make withthe verical {© Waite down Lagrange’s equations of motion56 (Chapter It: Lagrange's Equations Solution (@) The canesian coordinates of m, are yn hsind,, yy = £6050), 0 the cartesian components of the velocity of m, are Hy 48,0050,, jy = -L0, sing. ‘The kinetic energy of m, is T= }mc3 +5) = 4m 462. The canesian coordinates of my are y= G sind, + sind, yg = £40050, + .0058;, so the cartesian components ofthe velocity of my are ig = 0,080, + £48,008), fy =i sind, ~ 48, sine, The kinetic energy of mz is Ta = hail + 92) = dma CFO + 26 620,64 c09(0, ~ 8,) + G69). ‘The total potential energy is v -mgt, cos®, ~m2g(¢,cos®; + £,c0s0;). ‘The Lagrangian is L= 496} +4, (C6 +26,20,6, cox, -0,)+ G6) 4+ mgt cos0, + myg(l,c0s6, + acos®,). (b) We have FE = (+ m Fd, + mhlsb, 010, - 0), \ 4 at : : ale) mz)676, +m6;¢,62.c08(82 ~ 8;)~ m24y¢,63(6, ~ 6,)sin(®, ~ 8, = mglyl0,0, sin(@, ~0,)~ (mm, + m3 )gf,in0,,Exercise 3.05 ” and aL atts 3B, 7 Maede + ml L8 c08(02 -0,), 2 4( aL) aj ea 4) = mz63b, + m3ly628,c08(0, ~ 04) mat t364 (6, ~ 6,)sin(, -8,), aL Fer matsbiasin(s- 0) matin Lagrange's equations are (om, +,)5, + my ly ls8a cos ~0,)~mgfyts64 in, ~0,) = -(m, + mf, sn0, aati + m3 ts8 cos, ~6,)+ mal, inp ~0,) = -mafgsinds. In the small angle approximation these become (om, +m), + malate = Lom, +m Bh. mall + matis8, ~-m hea, and are the equations of motion of two linearly coupled simple harmonic oscillators. | m ‘A bead of mass m slides on a ong straight wire which makes an angle a with, and rotates with constant angular velocity « about, the upward vertical. Gravity g acts vertically downwards. (2) Choose an inte generalized coordinate and find the Lagrangian. (b) Write down the explicit Lagrange's equation of motion.58 (Chapter il: Lagrange's Equations Solution, (@) Choose as generalized coordinate the distance r of the bead along the wire from the axis of rotation. ‘The kinetic energy of the bead is then To}? +0%?sinta), andthe gravitational potential energy is V=mgreosa. “The Lagrangian is L-4m(?? 40% sina) mgrcosa. (©) Lagrange’s equation is mi'= mo*rsin? a - mgcosa. ‘This isthe same asthe equation for one-dimensional motion in an effective potential Veq = ~}mo?e sin?a.+ mgrcosa. is 7 & cosa is This 1 has a maximum at r= + corresponding to an unstable is potential has tre te ponding to tat equilibrium point. Exercise 3.07 | be ‘A particle of mass m slides on the inner surface of a cone of half angle a. The axis of the cone is vertical with vertex downward. Gravity g acts vertically downwards. ) Choose and show on a diagram suitable generalized coordinates, and find the (®) Write down the explicit equations of motion for your generalized coordinates.Exercise 3.08 39 Solution (@) Choose as generalized coordinates the distance r of the mass from the apex of the ccone together with the angle @ measured in a horizontal circle around the axis of the cone. The coordinates r and § are simply two of the usual spherical polar coordinates. te 6 is, in this problem, fixed at O-a. The Le fim(? + 262sin?a) ~ mgrcosa. (©) Lagrange’s equations are mi ms@*sinta-mgcosa, d(mr%9sin?® a)/ét = 0 ‘The second of these equations yields mx%}sin® a= L,, where L, is a constant which can be identified as the angular momentum of the particle in the vertical (z-) direction, ‘Substituting this into the first equation, we obtain B ate ees Hse nares) ‘This is the same as the equation for one-dimensional motion in an effective potential 2 Vert = +mgrcosa. Ima sin? Vag has a minimum at 1? »—y— 4, comesponding to suble horizontal migsin? acosa circular orbit. Exercise 3.08 Using spherical polar coordinates (r,8,¢) defined by xarsin6cosp y=rsinOsing 2=rc0s8, write down the Lagrangian and find the explicit Lagrange’s equations of motion for a particle of mass m moving in a central potential V(r).6 Chapter Il: Lagrange's Equations Solution The components of the velocity of a particle in spherical polar coordinates are ¢ in the r- direction, 10 in the 8-direction, and rpsin@ in the @-direction. Since these are mutually ‘perpendicular, the speed v of the particle is given by vai? +76? 4 r7GPsin?0, snd its kinetic energy by To dmv? = }m¢e? +176? + 124? sin?6), ‘Alteratively this can be obtained by transforming the expression Tadm(i? +? +2) for the kinetic energy in cartesian coordinates, by setting x= rsin0cosp, i= tsinOcosg + 16cosOcos§ - rpsinOsing, =rsindsing, j= sinGsing + fDcosOsing + psino0s4, 210080, im tcos0- rind. If the particle moves ina central potential V(r), the Lagrangian is La T-V = 4d(i? +176? + 179? sin?6)- Ver). Webave ee ni. ra 2. metGsinte, a eee Meee Ls mar? +§2sin2@)- 2, Ls msG2sin cost a. Le emrdt santo), LamAPsinvons, Hao, ‘so Lagrange's equations in spherical polar coordinates for a particle moving in central potential are ae vo 4 (mi) = ma(6? + fPsinte)- VO, Sa%) - m#sind00s0, 4 ruin?) « qt sin’ 8), 0.Exercise 2.09 GI Exercise 3.09 For some problems paraboloidal coordinates (E,1,¢) defined by K=Encoss y= Ensing 2=4@?-n7) tum oot to be convenient (a) Show that the surfaces & = const. or n= const. are paraboloids of revolution about the zaxis with focus atthe origin and semi-latus-rectum &° or 1. (©) Express the kinetic energy of a particle of mass m in terms of paraboloidal ‘coordinetes and their first time derivatives. (Ans. T= 4mm(8? + 1? (8? + i?) + 4men?64) ‘Solution (@ Paraboloidal coordinates (&,7,.¢) are defined by x=Encos¢, y=Ensing, 2=4(6?-n7). ‘Note also that the cylindrical coordinate p (the distance from the z-axis) is pada ty? =tn ‘and that the spherical polar coordinate r (the distance from the origin) is reqitayted = }@t +n?) We have reek? and r-zen, Setting r= yp +2” in these and solving for 2, we find ey 7 Th ee fixed-§ ones open in the negative z-direction and the ixed-n ones open in the positive z- direction (Fig. 1). efa (Chapter Ill: Lagrange's Equations = constant Ex. 308, Fig. 1 (2) The element of distance in paraboloidal coordinates is obtained most easily from that in cylindrical coordinates, (ds)? = (dp)? + p7(49)? + (a2), by setting dp=Bén+nds and dz=Edk—ndn together with p= Eq. This gives (49? = © + (68)? + Cony?) +8n2(09)?. ‘The kinetic energy of a particle of mass m is then todn( St) Lala? entfe+it) +88") in parnboloidal coordinates. Exercise 3.10 ‘The motion of a particle of mass m is given by Lagrange’s equations with Lagrangian L=exp(at/myT - V) where a isa constant, T = }m(X? + j? + #2) isthe kinetic energy, and. V = V(x.y2) is the potential energy. Write down the equations of motion and interpret.Exercise 3.11 6 Solution We have cms, a : 2. 4 (-av/ax), s0 Lagrange’ equation for xis mi = ak - aV/ax with similar equations for y and z. These equations can be combined into the single vector equation ni = -af-VV. This is the equation of motion for a particle of mass m which is acted on by a conservative force -VV together witha frictional force a which is proportional tothe ‘velocity ofthe panicle. Exercise 3.11 ‘A system with two degrees of freedom (x.y) is described by a Lagrangian La fim(ai? + 201g + 692) pk(ax? + 2bxy +69) where a, b, and c are constants, with b? ‘motion and thereby identify the system. Con: in particular the cases a=c=0,b» 0 and ac, b=0. Solution Weave a a oo a 3 Lemire, H--rexsey, a oy . so Lagange’s equations ee (ak + bj) = ~k(ax + by), (bi + 69) = -k(bx +cy).“a (Chapter Il: Lagrange’s Equations ‘These can be written in matrix form of clit bh 1 we multiply both sides ofthis equation by the matix reciprocal to [t 4] (ote that bc the reciprocal exists ifthe determinant ac ~ b? of the matrix is nor-zero), we obtain mi = -kx, my = ~ky. ‘These are the equations of motion of a two-dimensional harmonic oscillator (a particle of ‘mass m constrained to move in the xy-plane and pulled towards the origin by a spring of, ‘zero unstreiched length and spring constant k). The usual Lagrangian Ly = fi? + §2)— dk(a? +) (kinetic energy minus potential energy) for this system is obtained by seting a= ¢= 1 and b=0. Itis clear from this example, however, that it is sometimes possible to find ‘other Lagrangians which lead tothe same equations of motion. ‘Thus, for example, if we take a= 0 and b= we getan equivalent Lagrangian Ly-mijy-bay, Whereas if we take a= = Land b= 0 we get an equivalent Lagrangian mi? ~ §2)~ pe(x? ~ 2). Exercise 3.12 ‘The Lagrangian for two particles of masses m, and m, and coordinates r, and 3. interacting via a potential V(e, ~F2), is L= bmi + $mafal - Vers -12). (2) Rewrite the Lagrangian in terms ofthe center of mass coordinates R= TUSL* Sata m +m and relative coordinates = ry — Fa. (©) Use Lagrange's equations to show that the center of mass and relative motions separate, the center of mass moving with constant velocity, and the relative motion being like that of a particle of reduced mass —™22_ in potential Vi). +mExercise 3.12 65 Solution (@) The coordinates +, and r2 of the particles, in terms of the center of mass coordinates R and the relative coordinates r, are (Fig. 1) Re Mor and ry eR-— Boe, m+n mm +m, my Mr m m1 T Ongin Ex. 312, Fig. 1 i m sna tal +AU? -vER) 2m+m, (©) Lagrange's equations for R and r ere : mm, 2) (my +m) = ig VO, (rm tmit~0 and uae ‘The first of these shows that the center of mass moves with constant velocity, while the second shows thatthe relative motion isthe same as that of a particle of “reduced mass” EABL-in pce Veo.CI (Chapter Mt: Lagrange's Equations Exercise 3.13° Consider the motion ofa free particle, with Lagrangian Lea f(a? + 9? +2), as viewed from a rotating coo te system X'mxcos@+ysin®, y'=-xsind + ycos0, ‘where the angle @ = 6(1) is some given function of time. (@) Show that in terms of these coordinates the Lagrangian takes the form Lio fafa? + 97? 42%) 4 2007" — y't)¢0%(K? +>] where «= d0/dt isthe angular velocity. (©) Write down Lagrange’ equstions of motion and show that they look like those for & particle which is acted on by a "fore." The part ofthe "force" proportional to w is called the Coriolis force, that proportional tow? is called the centrifuge! force, and that proportional to do/dt is called the Euler force. Identify the components of these forces.” Solution (2) To determine the Lagrangian to be used in a routing cartesian frame, we transform variables, setting in L x=x'cos6-y'sin®, 'c0s0~ y'sin® ~ ox'sinO - wy'cos®, = x'sind + y'cos0, in + ¥'c0s8 + ax’cos®-wy'sin®, ‘where © = d/t isthe angular velocity ofthe rotating frame, We thus find nf 'c0s0- sind ~on'sind -wy'c0s0)? + +(X'sin® + ¥'cos6 + cox’ cos - wy’ sind)? + (2'* edly? 42% e209 -y)4 whee sy) (0) We have aL nk! - muy’, 7 moy', | Ee my +mox’, omExercise 3.14 or so Lagrange’s equations in the rotating frame are mi’ = 2moy' + mo*x’ + m(da/di)y', mj! = -2mui' + mo?y’ - m(du/d)x’, mi'=0. The righthand sides ofthese equations are she components ofthe "inert fre” which in the rou say acts on cle. ‘The terms proportional to © are the "Coriolis fore” _ pes oe F (Coriolis) = @moj’,~2mot’,0)= -2me x’; the terms proportional to «* are the “centrifugal force” F'(centrifugal) = (meo?x’,mio*y',0) = ~meo x (a xr"); and the terms proportional to du/dt are the “Euler force" F'Eules)- (me ‘The Euler force is zero ifthe rotation is uniform. Exercise 3.14 () Write down the equations of motion resulting from a Lagrangian Lm f(a? + 92 + 22)- Vin) + (CB/2ONaF -Y¥), and show that they are those for particle of mass m and charge ¢ moving in a central potential V(¢) together with a uniform magnetic field B which points in the z-direction. (©) Suppose, instead of the inertial cartesian coordinate system (x,y,2), we use a rotating system (x',y'.2’) with Xexcoswt+ysinot, y'=-xsinot+yeoswt, 2’ =z. ‘Change variables, obtaining the above Lagrangian in terms of (x',y',2') and their first time derivatives.” Show that we can eliminate the term linear in'B by an appropriate choice of « (this is Larmor’s theorem: the effect of a weak magnetic field on a system is to induce a uniform rotation at frequency w, the Larmor frequency).68 (Chapter Ill: Lagrange's Equations Solution Weave z Sent oe a * SfY ng By 88) ag atlae) 7 2c alae)" ‘These can be combined into the single vector equation V+ (/o)ExB (where B= Bk) and are the equations of motion, in an inertial frame, of a particle of ‘mass m and charge e in a potential V together with a uniform magnetic field B in the 2- direction. (®) The Lagrangian in a rotating frame can be obtained by transforming the variables as in Exercise 3-13. The first term in L is the same as the Lagrangian in Exercise 3-13 and ‘so transforms the same; the second term V(r) is invariant; and the third transforms as AY -yRea xy yk! +00"? +), ‘The Lagrangian in the rotating frame is thus and the Lagrangian reduces to 1? ag? ob 1 eo? (x? + y"2 Lim 5mGi?? +97? +27) Vi — mak (x? +". It now contains only second order terms in the magnetic field. ‘These are small if the ‘magnetic field is weak.Exercise 3.15 co) Exercise 3.15 Show that the equations of motion of an electric charge ¢ interacting with a magnet of ‘moment m can be obtained from a Lagrangian La EMve +4Mavis + (/e)ve where is the vector potential atthe charge due to the magnet. (Y. Aharonoy and A. Casher, “Topological Quantum Effects for Nevtral Particles,” Phys. Rev. Let. $3, 319-331 (1984). Solution First consider the charge. We have a omy +Ea, ice Af aL) ay dre Cdk yy Wve [dm rr, B/E) oy, Me 89M oy, Me [am +, VA ) at edt at {{z irgatght 10s Ya) Ve | xe AY]. Sale ¥en) Al * Elle ~ ¥en) VeA + (We ~ e e Lagrange’s equations forthe charge are ay, 1dm fr, ota of Mdm. Mee [edt “ieee 2x ext)]Srexexh : Now By -VexA the magnetic field due to the magnet atthe location of the charge, The term in square brackets, is the electric field due to the magnet at the location of the charge. It consists of two pars: the first isthe electric field due to the time-dependence of the magnetic moment;0 (Chapter Ill: Lagrange's Equations the second is the electric field due to the motion of the magnet. The charge thus moves according to Newton's second law, M, Me wey +20, xBay a ¢ bere the gh hand sd isthe sproprinte Loren force wich the magnet exer on the charge. Now consider the magnet Although Lagrange's equations canbe obtained inthe usual way, its simpler to observe thatthe Lapraglan is unchanged on interchanging the Subscript © and m everywhere (but the quantty A. changes sign). Applying this sein 1 Lagrange guns forthe che, we obth Lagange eqn for magnet, am, (.ta-te)_¢ AEP (o tate) Ling 2) (0 XA). Note that, Mn Oe, Me dt ao at ‘0 the total mechanical momentum M,¥, +M,¥. is constant. The preceding equation ‘needs to be rewritten so we can interpret it. First note thatthe quantity Byres in the first term on the right is the electric field due to the charge at the location of the ‘magoet. Then in the second term set Aw -mxE,, ‘and hence find OVA = Va x (mx E,) = (mV Ey. Here we have used the vector identity Vx (ax) = a(V-b)-b(V-a)+ (b-Vja~(a-V)b together with the fact that V,,E, = 0. We then have Mn de m9) Me Ge SG EH vale ‘The first term in the round brackets,Exercise 3.15 n 1 rere is the magnetic field due to the moving charge at the location of the magnet. We thus obtain at) ‘The term in round brackets is the magnetic field in the rest frame of the magnet. This is the correct equation of motion for the magnet, as is discussed more fully in Lagrangian ‘and Hamiltonian Mechanics. My Ma = 2S KE, + (o-Vq)(B,~CHAPTER IV ‘THE PRINCIPLE OF STATIONARY ACTION ‘OR HAMILTON'S PRINCIPLE, Exercise 401 Consider a modified brachstochrone problem in which the particle has non-zero inital speed vo. Show that the brachistochrone is again a cyclo, but with cusp h = v3/2g higher than the inital poin. Solution In this modified brachistochrone problem the speed of the particle when i has fallen a distance yi vedo 2ey, and the expression forthe travel time becomes * Ire Gy/axy? ate [feel ay, I, We+2y ‘The integrand ofthis expression, is yaa? Foy.dy/dx)= sores 7 does not depend explicily on x, so forthe required cuve the quantity = v8 + Deyn + @y/dn) is constant. Itis convenient to set vj = 2gh and then set (b+ yXL+ (dy/dx)?) = 28 where h and a are constants with the dimensions of "length." Rearranging this and integrating, we findExercise 4.02 B earn. ‘The y-imtegration can be performed by setting hty (cos) and dy =asingdp toobaain xe aff (-cosg)dp = af-sing)~@ -singa)] where > is determined by the condition h = a(1-cos¢p). The preceding two equations describe cycloid which has a cusp at ($= 0.x =-a(@0~singg).y =A) and which passes through the initial point (@ = $o,x = 0,y = 0) (Fig. 1). The constant a must be chosen so that the cycloid passes through the final point (x,,y;). Guy) Ex. 4.01, Fig. 1 Exercise 4.02 ‘A bead of mass m slides without friction slong a wire bent inthe shape of a cycloid xn ag-sing) y= a(l~coss). Gravity g acts vertically down, parallel to the y axis. (4) Find the displacement s along the cycloid, measured from the bottom, in terms of the parameter @. (®) Write down the Lagrangian using s as generalized coordinate, and show that the motion is simple harmonic in s with period independent of amplitude. Thus the time required for the bead, starting from rest, to slide from any point on the cycloid to the bottom is independent ofthe starting point. What is this ime?" (Chapter 1V: Principle of Stationary Action Solution (@) The element of distance ds along the cycloid is given by (4s)? = (6x)? + Gy)? = 23[ ~cosey? + sin? @dg)? = 202(1- cosg)(do)? = 44? sin? (9/2\(49)" ds = 2asin(9/2)46, ‘The displacement s, measured along the cycloid from the bottom (6 = x), is thus = 2aftsin(g/2)db = ~4acos(@/2) ‘and ranges from —4a to 4a as we move along the cycloid from cusp (= 0) to cusp (@= 22). () The kinetic energy ofa bead of mass m which slides along the cyeloid is To 4m ‘The gravitational potential energy of the bead, measured from the bottom of the cycloid, ‘V = mg(2a- y) = mga(1 +.cosq) = 2mgacos?(/2) = 4m(g/4a)s?. “The Lagrangian is thus L = fms? ~}m(g/4a)s?, and Lagrange’s equation is i ~ -m(g/4a)s. ‘This is the equation of motion ofa simple harmonic oscillator with angular frequency = g/4 and period x= 2n/n = 2n/40/g. The time required for the bead, starting from rest to slide from any point onthe éycloid tothe bottom is one-quarter period, af andi independent ofthe sat point.Exercise 4.03 15 Exercise 4.03 Novelists have long been fascinated with the idea of a worldwide rapid transit system consisting of subterranean passages crisscrossing the earth.! Public interest in ‘sublerrancan travel rose sharply when Time magazine? commented on a paper by Paul W. Cooper, “Through the Earth in Forty Minutes"® This paper, while repeating some easler ‘work. Served asa catalyst for a numberof other papers onthe subjec® to which you may ‘Wish to refer in working the present exercise. ‘Take the gravitational potential within the carth to be $(g/R)r? where g isthe gravitational field atthe surface and R isthe radius ofthe earth (thereby neglecting the non-uniform density of the earth. (@) Fist show that a panicle stating from rest and sliding without friction through a straight tunnel connecting two points on the surface of the earth executes simple harmonic motion, and that the time to slide from one end to the other is to = xYR/E (~42.2min) independent ofthe location ofthe end points. (®) Now consider the curve 1(6) the tunnel must follow such thatthe time for the particle to slide from one end to the other is minimum. Set up the appropriate variational principle, and show that al %& Tea Pe isa first integral of the resulting Euler-Lagrange equation. Here r= fo at the bottom of the tunnel (9 is the minimum distance to the center of the earth), Rearrange this and {integrate to obtain the equation of the curve, ‘where 0 is measured from the bottom of the tunnel. The angular separation between the tend points on the surface ofthe earth is thus given by 49 = x(1~r9/R). (©) Introduce a parameter @ with ‘See Marin Gaoer, Scleifc American, Seplember 1965, p. 10-12, commenting on an ate by LK. ‘Edwards, Hlgh Speed Tube Tanjoaton” Selenite Anerson, Avg 195, 930-40, 1966, pp, 42-43 3Paal W. Coopes, “Through te Eat in Forty Mules." Am, Phys. 34, 68.70 1960). ‘See Phlip G. Kinmee, "An Example of te Need for Adequate References," Am, J Poy 34,701 (1966). SGiullo Vecalan, “Teresvial Bracistochrone,” Am. J. Phys. 34,701 (1966): Russell L. Mallet, Commeots on “Through tbe Earth in Forty Minus,” Am, J. Pays. 34, 702 (1966); L. Jackson Laslet, “Trajectory fo Minium Transit Tune Through the Earth" Am. J. Pops. 34, 702-705 (1966); Paul W. (Cooper, "Purber Commentary on "Trough the Earth in Fety Miouies," Ar. Phys. 34, 703-704 (1966).16 (Chapter 1V: Principle of Stationary Action so $=0 atthe bottom and $= 27 at the ends of the tunnel. Show that the equation of the curve takes the form Aasfeeed)-He Bho o- (End) Be. ‘Show that this isthe equation of a hypocycloid, which is the curve traced by a point on the circumference of a circle which rolis without slipping on another circle. ® In this case the larger circle is the great circle route, of radius R, connecting the end points on the surface of the earth, and the smaller circle has radius a= 4(R ~f9) (its circumference is thus the distance between the end points onthe surface). The parameter 4s the angle shown in the figure. (© Now consider the time dependence of the variables. Show in particular that 6 varies linearly with time, 4 = 2x(W/t), where <= xpy/1— (Fo/R) is the time to slide through the sinimum-time-tunnel from one end to the other. Compare + with tg for end points 700 Jkm apart on the surface. Solution (a) For a particle of mass m which starts from rest atthe surface and which slides without friction through a tunnel through the earth, conservation of energy gives inv? + $m(g/R)P? = 0+ dm(a/R)R?.Exercise 4.03 n Ex, 4.03, Fig. 1 First suppose thatthe tunnel is straight and let s denote the displacement along the tunnel from the bottom, where r assumes its minimum value rp (Fig. 1). We thea have ves and Posters, and the conservation of energy equation becomes $ms? + $m(g/R)s* = }m(g/RMR? 15). ‘The leftchand side of this equation, asa function of s, has the same form as the expression forthe total energy of a simple harmonic oscillator with angular frequency wn eR ‘The motion ins is thus simple harmonic. The time to for the particle to slide from one end of the tunnel tothe other is half a period, te afa= x (Rg = 42.2 min, (b) Now consider the curve the tunnel must follow to make the travel time a for the particle a minimum. Itis convenient to use plane polar coordinates (r,8), in terms ‘of which the element of distance ds is (ds)? = (4x)? + 17(40)*. ‘The speed v ofthe particle is given by energy conservation (see par (8) vv? = (g/RYR? -F?), ‘The expression forthe travel time becomesB (Chapter IV: Principle of Stationary Action “, ane (e100 + ag, ae ‘We wish to find the curve r= r(8) which makes this a minimum. The integrand, pa [R [eeddor +? 8 vr does not contain the independent variable © explicitly, so for the required curve the quantity ea: £ 1 2 9(dr/48) 48 BR =P Yanaor ee is constant, We set 1 2 to We? fay er Roe ‘At the bottom of the tunnel dr/d@ = 0 and this equation gives r= ro, 50 the constant ro is the minimum value ofr. Solving for dr/d8, we find a gt frag G0 To VRE= ‘Rearranging this and integrating, we then find Ro [2 eae “| FVF to . where we have chosen the constant of integration so that @=0 at the bottom of the ‘tunnel (ig. 2). Ex. 403, Fig. 2Exercise 4.03 9 ‘To do the integration, we set ‘The variable wis zero for r= and ittends to 2: for r+ R. The integral becomes 1 1 R RO* ro] (—=5—yp- | du = Ry) -ro tan” I cease eae) soe ‘0 the equation ofthe curve the minimum time tunnel must follow is aR [P81 [P= Oran 8 ead tow fs ‘The ends of the tunnel are at r= R, so the angular separation ofthe ends is 80 x(l~to/R). (© tis convenient to set 278 VR ‘where the parameter 6 is zero at the bottom and 7 at the ends of the tunnel. Solving. fort, we find #? = R?sin®(@/2) + 15 c0s"(6/2) = H(R? +13) ~ }(R? ~ 15 cos. FFunher, from the equation of the tunnel in part (b) we have. = tan Braet). on (Ban§) 4. See ee eects ts octane tr ee eee ee Se Ge ce pm ae nel pa in Be we circle which rolls on another circle, In this case a smaller circle of radius a = }(R~ fo) Sota ean80 (Chapter IV: Principle of Stationary Action Ex. 4.03, Fig. 3 If we apply the trigonometric cosine law to the triangle in Fig. 3, we have 7 R~0)? +4? -24(R ~a)cosg= HR? +13)-H(R? + 13)c086, which i one ofthe preceding equations. To obtain the second equation, the relation between @ and 6, we proceed as follows. Let Qp denote the angle tothe center ofthe smaller rolling circle and let @ denote the angle to the point on the circumference of the ‘smaller circle which traces the curve (Fig. 4) point center of. oncurve_ smaller circle Ex. 403, Fig. 4 ‘Then from Fig. 4 tan(@~0,)=——s8i0@_,__(R=rg)sing__ (R=rg)an(/2, (R=a)-acosp ” RU-cos9)+ ro(l + cos@) ” Rtan?(9/2)+ ro Now note that y+ (p/2)) = tAMO=80)* C812) cr a0 = + 72) = a ante ay” ONIN,Exercise 4.03 where to obtain the second equality we have substruted tan(@~ 0p) from the preceding equation. The angle 8 is related to @ by the condition tht the smaller circle rolls without slipping on the larger. This means thatthe arc Rp of the larger circle over which the point of contact moves must equal the arc a of the smaller circle, RO = a9 = (R= 1016/2), and thus 85+ (9/2) (o/RXA/2) ‘Substituting this into the preceding equation and rearranging, we obtain = tan Ryan) — 10, naw (oe$) ake “which is the required relation between @ and 4. (@) The time, measured from the bottom of the tunnel ° eganree = £ J [eer/40? +? ag, g}o¥ R= ‘We have seen in part (b) that de gt [Pag # gt [Pag 97 ro VRE=r slong the minimum time tunnel, o the expression for the time becomes, 1 fRa8 dr : aR YU) ‘Now introduce again the parameter @ with ee ee oa ae 2ede Saar ee Integration gives82 (Chapter IV: Principle of Stationary Action ‘0 ¢ isa linear function of the time. The total time forthe particle to travel from one end (of the tunnel to the other is ton, EF colo, so the parameter 6 can be written g-2myn). Note that t=0 for ro =R (short tunnel near the surface of the earth) and that + = tp for fo ~0 (straight tunnel through the center ofthe earth). For end points 700 km epart on the surface the angular separation is. 700 0-20 Th Whichgives 1-52 = ‘Hence the travel time along the minimum time tunnel connecting these two points is eto t-(f2)' = to EERE 0.26% =I Kn.Exercise 4.04 8 Exercise 4.04 ‘An instructive exercise in the calculus of variations is the "minimum surface of ‘evolution problem” (@) Find the plane curve y= y(x) joining two points (0,yo) and (x93) such thatthe area of the surface formed by rotating the curve about the x-axis is minimum (the Euler- Lagrange answer is y = acosh((x~b)/2), where &and b must be chosen so thatthe curve passes through the end points). {@) Using a computer or otherwise, draw representative members ofthe (one-parameter) family of such curves which star at (0,1). Hence convince yourself that if the final point i near the y-axis, o Euler-Lagrange curves passthrough the given end poins, whereas it the final point is nea the x aks, no Euler Lagrange curves pass tough the given end (@) In this laner case the solution is the discontinuous Goldschmidt solution composed of straight line segments (0,1)» (0,0) -»(x,0)~» (X,Y). In the region where there are two Euler-Lagrange solutions, calculate and compare the area given by the Goldschmidt ‘solution with the reas given by the Euler-Lagrange solutions. Which of the three gives minimum area? ‘Does ths depend on wher in the region the endpoint lies? (TRS lst par is difficult; for guidance see Gilbert Ames Bliss, Calculus of Variations, (published for The Mathematical Association of America by The Open Court Publishing Company, ‘Chicago, inois, 1925), Chap. IV, pp. 85-127.) Solution (2) Consider the surface of revolution generated by rotating the curve y = y(x) around the x-axis. A litle strip around the surface has an area A = 2ayds = nyt + (dy/dx) dx, ‘so the total surface area between limits xo and x, is An dnf' y+ Gy/xF ex.on (Chapter IV: Principle of Stationary Action We wish to find the curve y= y(x) wl imegrand Fe 2nyyl+ (€y/ax)? does not depend explicitly on the independent variable x, so for the required curve the ‘quantity makes this surface area a minimum. The 4p 2: “sae Traian” isconstant We set ‘The constant of integration bis the value of x at whi is aminimum. Thus, according. to Euler-Lagrange, the curve which generates the ce of revolution with minimum y cosh (x b)/a. ‘Such a curve is called a “catenary" (Fig. 1). The constants a and b must be chosen so that the eatenary passes trovgh the specified end points. As we shal seein part () sis ‘not always passible. What has gone wrong? "The difficalty i thatthe Euler-Lagrange ‘pprosch impliidly assumes that the required curve s smooth and this isnot always the ‘ase. Tetums out that forthe egion where there is no Euler-Lagrange curve, and indeed ven in part ofthe region where there are Euler-Lagrange curves, the solution isthe non- ‘Suooth Goldschmidt eure.Exercise 4.04 Ex. 4.04, Fig. 1 (®) For catenaries which start at (0,1) the constants a and b are related by T= acosh(b/a). ‘Since “cosh” is always greater than or equal to 1, the constant a lies between O and 1, and b= sacosh""(y ‘The family of catenaries obtained as ais varied from 0 to 1 is shown in Fig. 2. As we can see, this family has an envelope such that two catenaries pass through any point above the ‘envelope, and no catenaries passthrough any point below the envelope. 923 28 3 eg $33 3 8 26 See & & a4 y 2 1s 1 os: os 7 15 2 Bx. 4.04, Fig. 286 (Chapter IV: Principle of Stationary Action (©) The area of the surface of revolution generated by rotating one of these catenaries around the x-axis is Ax 2n y+ y/dx)? dx = 2naf" cosh®((x~by/a)dx -mi[Be sinn220=2 «sinha AI where -acosh”'(I/a) still. We have seen that two such catenaries pass through any fend point above the envelope. How do the areas ofthe surfaces of revolution generated by these compare? Also, how do they compare withthe area Ago = x(1)? +09? generated by the non-smooth Goldschmidt curve? ‘We shall not atempt to give a complete solution to this problem, Rather, we ake 1 few end points and see what happens, the aim being to bring out atleast some of the general features of the problem. In pariculer, we take as end points the intersections of the a= 0.2 catenary (the "lower" catenary) with the a= 0.4,0.6,0.8,and1.0 catenaries (the “upper™ catenary). The end points and the areas of the surfaces of revolution ed bythe upper and lower anaes and by the Goldschmidt cane re given in able Table I: Areas of surfaces of revolution 2-020 ye Ye Auger tower © Aga = with =0.4 0.729359 0.413241 3.930575 4.005856 3678077 =0.6 0818405 0.621255 4.370380 4.740310 4.354114 =0.8 0.888642 0.870795 §.059222 5.955004 5.523810 a=L0 1.005808 1.549932 8925927 1.194006 10.688607 We gt that, (0) the area generated bythe oper canary is es than that generated by the lower catenary in all cases, and () the area generated by the Goldschmidt curve is Jess than that generated by the upper catenary if the end point is “near” the envelope but {s greater than the area generated by the upper cetenary ifthe end point is "far from” the eavelope. These tum out o be general features ofthe minimum surface of revoltion problem.Exercise 4.05 87 Exercise 4.05 ‘The motion of a “free* particle of mass mon a surface is described by Lagrange’s equations with Lagrangian L=T'= fm(ds/d)?. Show thatthe resulting equations of tation ah gun fra pods, og which he pail moves comune where the x? (p= 1,2) are two independent coordinates on the surface and the Bp(X) are the components of the metric tensor. We have aL ax? AL _ 1 S8po dx? dx? a(dx"/dt) mBop ae ee 2 axe dt a” ' " 8 (gs) ~ me ot a 3x50)” Par? so Lagrange’ equations ae ey B00 GE where the 'a.p9(x) are the components of the Christoffel symbol. ‘These equations are the same as the equations of a geodesic found in Lagrangian and Hamiltonian ‘Mechanics, page 66. Further, since the Lagrangian does not depend explicitly on t, the ‘quantity wig te tee BEL ade at) dt 2 ae dt 2 ee {is constant time; thats, the speed ds/dt of the particle is constant in time.8 Chapter IV: Principle of Stationary Action Exercise 4.06 Find and solve the equations for geodesics on a plane, using plane polar coordinates (6.8) in terms of which the element ds of distance is given by ds? = dr? + r7492 Solution ‘The non-zero components of the metric tensor are Beth Bye ‘The non-zero components of the Christoffel symbol are then Tae Tay = Trae and the equations for a geadesic become &r -(%) 7 84 de do me ed ae “The second ofthese equations can be writen (a8) 9 al >. and gives PB, tery where fo is a constant. Using this to eliminate do/ds from the first equation, we find arin a(g gala) This is like the equation of motion of a particle of unit mass ina potential 13/212, so “energy conservation” yields 4)” 5 consume. 2\q5 ‘The value of the constant can be found from the constraint89 ee =[ . Ee ‘where we have chosen the constant of integration so that s=0 at r= ro. This can be written 243 Peres. ‘Substituting the result into the ¢-equation, we find Integrating, we obtain where we have chosen = 0 st s=0. This canbe writen serotang. ‘The two equations 12 =r-+4? and s~ roland, which express the polar coordinates (6) in terms ofa parameters are the equations of a suaight line (Fg. 1). If desire, the parameter § can be eliminated to give the explicit relation rcos® = fp between the coordinates. To Ex. 4.06, Fig. 19” (Chapter 1V: Principle of Stationary Action Exercise 4.07 (@) Find and solve the equations for geodesics on the surface of a cone of half angle a, using as coordinates the distance r from the apex of the cone and the azimuthal ange ¢. (b) Show that ifthe cone is cut along a line 4 = constant and flattened out onto a plane, the geodesics become straight lines. Solution {q) The distance ds between wo y boring points on the surface of the cone is given (ds)? = (45)? + sin? (dp)? “This has the same form as the expression for (ds)? in Exercise 4.06, except thatthe @ in Exercise 406 is here replaced by @sina. As a result, the calculations and indeed the equation fora geodesic, rcos@ = fo, in Exercise 406 can be easily modified to yield the ‘equation fora geodesic, Feos(@sina) = F9, in the present problem, We have taken $0 at minimum radius r ro: this geodesic ‘can, of course be rotated about the axis of the cone to give other geodesics. () Now cut the cone along a line $= Go and fatten it out onto a plane. Straight lines 4+ constant on the cone become radial straight lines on the plane. Circles r = constant ‘on the cone become circular arcs on the plane; “arcs” because the distance around these Circles on the cone is 2arsina1, so the angle they subtend on the plane is 2xsina. The {wo sides of the cut become a wedge of angle 2n(1~ sina) which is cut out of the plane. If we introduce polar coordinates (R,) on the plane, then Rer and Ongsina. We have taken = 0 at $= 0. As 6 ranges from ~(2x— Gq) through 0 t0 gp, the angle ranges from ~(2n ~9)sinat through 0to gosina (Fig. 1).Exercise 4.08 1 Ex 407, Fig. 1 ‘The equation fora geodesic on a cone, when the cone is cut and flattened out onto & plane, becomes Reos@ = ro ‘As we have seen in Exercise 4.06, ths is the equation of a straight line on the plane. Exercise 4.08 Consider two points on the surface of a sphere. Without loss we may take them on the ‘equator at (@ = 7/2, 0) and (@= 7/2, =a). The geodesics joining these points ere the two arcs of the equator. Nearby curves can be represented by = n/2+ Fragsingand/a) where the deviation from the equator has been represented by a Fourier series chosen to vanish at the end points. Evaluate the distance between the two points along such a ‘curve, valid to second order in the small quantities a. Show that for 0h~x2/4h there azeno real solutions. The curve y=h~x/th i the envelope of the family of trajectories which sat atthe origin with fixed energy E = mgh but arbitrary angle of launch. Exercloe 4.12 A particle moves vertically in the uniform gravitational field g near the surface of the earth. The Lagrangian is L- pi? - mez ‘Suppose that at time O the particle is at z=0 and at time t, itis at z= 2, . For any ‘motion 2(¢), actual or virtual, between these two points the action is SKel0)}~ f' Lead). Pretend you don't know what the actual motion is. You might then guess that it can be adequately represented by the first three terms in a power series int zero tvotefat, where z9 and vo are chosen so that 2() passes through the end points, and a is an adjustable parameter. Evaluate S for this form of 2(t) and note the dependence of S on a. For what value of ais $a minimum? Solution The action along the path z7+vot+ fat? from t=0 to t= ty, for a particle which moves vertically in a uniform gravitational field Bis100 (Chapter 1V: Principle of Stationary Action Se fl[bm(vo +20? ~mglao + vot $a} = }rcvdt + vat + fa2dy—matagy 0 fvot + fat. For paths which start at (t= 0,2 = 0) and end at (t= 2%) we have Oe2 and neve dath, so the initial velocity i Using these relations to eliminate 2 and vp from our expression for S, we find 2 mt meat | al ened So Gi RBA mc +2 ‘The action S for these paths between the given end points is a quadratic function of the adjustable parameter "a." Minimum action, and the “best” path of the form chosen, ‘occurs for a= —g. Indeed, this gives the actual path, which happens to be a special case of paths ofthe form chosen.CHAPTER V INVARIANCE TRANSFORMATIONS AND CONSTANTS OF THE MOTION Exercise 5.01 ‘Show that a function of (1), (1), and t satisfies Lagrange's equations identically independent of q,(1)) if, and only if, itis the total time derivative dA/dt of some function A(q(t).). Solution Theit let an a Now ae aay aA 4 i). eA a4,” aq," £2,004 ‘Comparing, we see that F satisfies Lagrange's equations, 4(28) 2 dt og.) ay identically (independent ofthe for of A and ofthe specific tie dependence of) ‘The “only if": conversely, suppose that Fis some function of q, q. and t, which satisfies Lagrange's equations identically. These equations can be written aS TE gO aE Danek aan, a” ‘The only place the 4's appear is in the first term on the let. Since we want the equations. to be satisfied independent of q, the coefficients of each of the &'s must be zero, or 84094 ‘This means that F i a linear function ofthe 4's,102 (Chapter V: Invariance Transformations Fagn= J fola.0i + 806.0. a F Lat, ag ter 2a oe, ad FS tng, a ala) 3m Seine care arnt ‘Substituting these into Lagrange's equations, we find ‘ S(t Ba 2 aay 294) at” an ‘The only place the 4's appear isin the first term on the left. Since we want the equations 1 be satisfied independent of g, the coefficients of each of the q's must be zer0, ha Ay Oe ‘These are the conditions for the f's to be expressible inthe form EXCH} f@- where A(q,t) is some function of the coordinates q and the time t. Substituting this into Lagrange's equations, we find aA ag aa” an 2. ‘Thus inorder to satsy Lagrange’ equations identical, a function F(q.t) most ave Sak, | ah _ dae, F- yg +H. ; Jahr maExercise 5.02 103 Exercise §.02* ‘The motion of a particle of mass m which moves vertically in the uniform gravitational field g near the surface of the earth can be described by an action principle with Lagrangian L=}mi?— mg. (@) Show that the action principle is invariant under the transformation 2'=2+a where a is any constant, and find the associated constant of the motion. (©) Show that the action principle is invariant under the transformation 2' = 2+ Bt where Bis any constant, and find the associated constant of the motion, ‘Solution (@) The new Lagrangian L'(2,2,t) for the new variable 2’ is obtained by transforming variables, setting zea and 0d! inthe old Lagrangian L(22,t). We have Lele) -Lei) = }mi? mee = }me? ~mg(e' -2) = }mz? —mgz'+ mga. = Lez) + dna Where A-=mgia. Thus the action principle and system are invariant under this transformation. "The corresponding infinitesimal invariance transformation is obtained by replacing @ by 8a and setting Srez-2=8a and 5A=mgiba. ‘Te morte cnt be sons Saesthemitarmatn = m(z+gt)da. Tecoma emmy MACE te in vey fe Pe104 (Chapter V: Invariance Transformations (b) The new Lagrangian L'(2',i,t) for the new variable 2' is obtained by transforming variables, setting zev-Bt and 2-'-B in the old Lagrangian L(2,i,t). We have Le ‘L(z,@) = 4ni? - mge = dm¢e'- By? ~ mgt’ - Br) = fi? - mgr’ - mi'B + mp? + m8 Lee) sane where A-=-m(2'~}g1°)8+ }mB%. Thus the action principle and system are invariant ‘The comesponding infinitesimal invariance transformation is obtained by replacing by 8B and setting brez'-z—t6B and bA=-m(e-4gt")6p. ‘The associated constant of the motion is eter th = mitdp— med V68 mG ~2+ det7)08 ‘The constant m(it=2+ $s1°) equals -mzy where 29 is the inital position of the ne expressions dagievo and -2+}gr?=-29 for the two constants of the motion can be inverted to obtain the velocity and position of the particle as functions of time, ievo-gt and cer evot-det’, thus solving the equations of motion.Exercise 5.03 105 Exercise 5.03 The motion of a simple harmonic oscillator is described by an action principle with Lagrangian La fms? -}mo*x? (@) Show thatthe action principle is invariant under the two-parameter (A,B) family of transformations X ax+ Asinat + Beoswt. (©) Find the two independent constants of the motion associated with the infinitesimal transformation, and identify them physically. (©) Use the rests of (b) to write down the general solution tothe equation of motion. Solution (@) The new Lagrangian L'(x'Xt) forthe new variable x’ is obtsined by transforming ‘variables, setting Kex'-Asinot-Beosat and i =i’ -wAcosat +uBsinot inthe old Lagrangian L(x,%). We have Ct) LO), = }mi? —}mo*? = }mG'-@Acosat + @Bsinat)? - Fmo*(x'- Asinast ~ Beoset)? = }mi? + mi'(-wAcosat + wBsinat) + }m(-wAcosist + wBsinwt)? ~}mo*x? - mo?x'(-Asinat ~ Beoswt) - }mw?(~Asinat ~ Beoswt)? LOW) + dA/at Le where A= mox'(-Acosut + Bsinwt) + ma(A? - B)sin2wt. ‘Thus te action principle and system are invariant under this transformation. () The corresponding infinitesimal invariance transformation is Bx bAsinut + 8Beosot and 6A = mux(-BAcosut + BBsinwt), ‘The associated constant of the motion is106 (Chapter V: Invariance Transformations acta mibasina Bose) + mex(-Bhcsst Bsn) = m(ksinwt - oxcoset)SA + m(Xcosat + wxsinwt)SB Since 5A and 6B are independent, the coefficient of 8A. and the coefficient of 8B are separately constants of motion. We can identify these constants by looking at t= 0. At this time the coefficient of 8A becomes -muxg where Xp isthe initial position, and the coefficient of 8B becomes mvg where Vo isthe initial velocity, m(Xsinat - wxcoswt) -maxy, (cost + oxsinaat)= mv . (©) These expressions for the two constants of the motion can be inverted to obtain the position and velocity as functions of time, x= x9cosut + (vo/w)sinot, i= -cxpsinat + vocoswt, thus solving the equations of motion. Exercise 5.04 ‘The Lagrangian fra particle of mass m and charge e moving ina uniform magnetic field which point inthe direction i (sce Bxerese 3.18) L= f(a? + ¥? +22) + (€B/20)(xy- yx). (@) Show that the system is invariant under spatial displacement (in any direction) and find the associated constants of the motion, (b) Show that the system is invariant under rotation about the z-axis and find the ‘associated constant of the motion Solution (@ First consider spatial displacement. The new Lagrangian L'(q/,4’) for the new ‘variables’ is obtained by transforming the variables, setting and ny’, in the old Lagrangian L(q.q). We haveExercise 5.04 107 L@.@)-L@4) = }m(e? + ¥? + 2?) + (CB/2e)ixy - yk) a di(i'? + 9"? +27) + (€B/2MG ~0'-8)8) = fe? + 9"? +27) + (CB/20)(x'y' ~ y'x') + (€B/2e)(-ayy' + a,x") = LQ’.9') + dA/at where A= (eB/2cX-a,y'+a,x’). Thus the system is invariant under spatial displacement. “The corresponding infinitesimal transformation is Breda, dye Bay, S2~ba,, and BA~ (CB/2c)(-yba, +xb4,), ‘The associated constant of the motion is t= (ms—£25) 6, (m5 +, tte, Byon o8) = {ok £25) oa «(ny £2s) ba, omit. ‘Since the displacements in the x-, y-, and z-directions are independent, the coefficients of ba,, bay, and Ba, m=mi-(eB/o)y, vemy+(eBjc)x, and py mi, are separately constants of the motion. (®) Now consider rotation about the z-axis, The new Lagrangian L'(q(,4") for the new variables q’ is obtained by transforming the variables, setting x=x'cos6+y'sin®, k= i'cos0+y'sind, ya-x'sinO+y'cos8, and j= ~i'sin® + j'c0s0, in the old Lagrangian L(q.d). We have L@.a)= La.) = f(a? +57 +22) + (CB/2O(0Y-¥¥) = }m[(i’cos6 + j'sin®)? + (~*'sinO + ¥'cos8)? +27) +(€B/2c)(x'cos0 + y'sinX-%'sin0 + ¥'cos0) ~(-x'sin0 + y'c0s0)(%'c0s8 + sin®)) = fm? +97? +27) + (CB/DONR'-y'X) ~La'a)108 (Chapter V: Invariance Transformations ‘Thus the system is invariant under rotation about the z-axis. "The corresponding infinitesimal transformation is bra -y88, by=x80, b2-0, and 8A-=0. ‘The associated constant of the motion is 89, +54 = (mi-$2y)(-y60)+ (my +$2x) x60) = [meay - yar + (8/2000? +94 Let om (ny ~ yi) + (€B/20)(x? + y*) denote this constant of the motion. It's of interest to use the constants ofthe motion and, previously obtained, to climinate mi = u + (€B/c)y and my = v~ (eBJc)x from the expression for A. We have A= x(v~(€B/e)x) ~ y(u + (€B/e)y) + (€B/2e)(x? + y?). ‘This equation can be rewritten 2 ae 2-2) (ys) a(2) (yea (ea) ++) ~(Ga) ( and shows thatthe orbit lies on a circular cylinder with sx center at x= cv/eB, y= ~ci/eB, and with radius R = (eB) identifies, physically, the constants u,v, and A. Since the Lagrangian does not depend explicily on the time, there is another constant of tx motion, he energy Wamitona) a ~(ot- Seavey omit os no imi? +P +2) 2 However, the energy is not independent of te constants already obiined. Indeed, weExercise 5.04 109 1 By? By? Bae lle) +(v-s) +] 2 = sh(v? ov +p) + 2 wy-ve) 2 (P) (ey?) 3 = heey orf). ‘The radius R of the circular cylinder on which the orbit ies canbe expressed simply in ters of the energy E- R = (c/eB)/2mE - p? = v, /o. In the second equality v, = Ji? +j is the velocity perpendicular to the magnetic field and « = eB/me is the cyclotron frequency.CHAPTER VI HAMILTON'S EQUATIONS Exercise 6.01 ‘Assystem with one degree of freedom has a Hamiltonian Hp) = 7+ A@p+ Bea) where A and B are certain functions of the coordinate q, and p is the momentum conjugate tog. (@ Find the velocity 4 (b) Find the Lagrangian L(q.4) (note variables). Solution (@ The velocity d is given by the first of Hamilton's equations, oH Bh hs aG e @). (b) The Lagrangian is then given by L(q.d) = p(@.494 - H(G. p(4.4)) = m(q - A)? - ‘This illustrates how one can go from Hamiltonian mechanics back to Lagrangian ‘mechanics. Exercise 6.02 ‘We have seen (Exercise 5.01) that two Lagrangians L’ and L which differ by the total time derivative dA/ét of some function A(Q,t), Lis L+dA/at, ‘are equivalent, leading to the same Lagrange’s equations of motion. (@) What is the relation between the generalized momenta p’ and p which these two yield? (0) Wha ise reledon between the Haritentane Hi’ and H which these two Sider erdtcly that Hanitton’s ‘equations of motion in the primed quantities are equivalent to those in the unprimed quantities.Exercise 6.02 rth Solution (@) and (Staring with a Lagrangian L, we introduce a new Lagrangian L’ by 2 transformation of the form ae u and the new Hamiltonian His ‘ ' : we Yoa-e-¥(n+ LZ 2hG Benen at) 28, is anfomnton i sometimes cle “tage wamformaton” Compe te Behavior under jc gauge wanformatons, ofthe generalized momentum of fhurped parte inan etomsgoete Bel (©) Hamitons equations fr the new variables are Se ag (22) dt Op, tend at 244) 5 Looking at our expressions for H’ and p', we see that the right hand side of the first of ‘these equations is equal to aH/ap,, so this equation becomes. S4y OH at” ap, “which isthe first Hamilton equation forthe old variables. The second Hamilton equation for the new variables becomes 4 # a ‘Expanding both sides of this equation, we find 24 A day, A at” Zoqyaqy dt” 30g, (#) -3@), “aR2 ‘Chapter VI: Hamilton's Equations ‘which, in view of the first Hamilton equation for the old variables, reduces to the second Hamilton equation A forth ld vale. anions eqstions fr te new variables ee thus equivalent o toe forthe ld vals Exercise 6.03, A parce of mass m moves in a cena frce fld with potential Vr). The Lagrangian jin terms of spherical polar coordinates (r,0,9) is Leda? +8? + Psin?0f?)-vo. (@) Find the momenta (p,.PesP4) conjugate 0 (1.0.4) ((b) Find the Hamiltonian H(r,,9,P,.Pe,P4)- (© Write down the explicit Hamilton's equations of motion. Solution (@) The momenta (P,,PorP4) conjugate to (F,8,6) are Ls mutsin?9 and py= Lo ma*sin?§. Pe e i. () The Hamiltonian H(£,0,6,P,,PosP4) is H= pit + pol + Py L- cs 2 Bi Phe Pt vee 2m 2mr? * ms? sin?6 oe (© Hamilton's equations of motion are ar cs a xo, « a8 a f= BB th « oy aH at at 90 “a 3 “ 2 aH at at 99Exercise 6.05 n3 Exercise 6.04 ‘The Lagrangian for a free particle in terms of paraboloidal coordinates (E,1,4) is (see Exercise 3.09) L= fm(E* +o? (8? +a?) + de??? (a) Find the momenta conjugate to (149). (b) Find the Hamiltonian. Solution (@) The momenta (Pj.Pq.P4) conjugate to (E.m.@) are = he mgt ent = Le me enh, and py 2h a meth Pe GErme oP. py Feamel enka and py a Sem mb ns. (©) The Hamitonian is H= ppt palit pg -L 1 “A(kS- #y- Bxerdise 605* ‘The Lagrangian for a fee particle of mass m, refered to cartesian coordinates (x',y,2") ‘which are rotating about an inertial 2-axis with angular velocity, is (See Exercise 3.13) Lim df i? +9 + 2%) + 200x'y'~y'2) +07 + y?)] (@) Find the momenta (p3.05.Pz) conjugate t (x'y'.2"). (©) Find the Hamiltonian H’Q’.y'.2'.P4.Pj PO) (Aas. B= (02 + 9? + p2)- of -y'et)) Solution (@) The momenta (P,P) Pi) conjugate to (x'y',2') are poem oy), pee emi sox Panay ( oy). Py a uy’ ), and4 Chapter VI: Hamilton's Equations Note that (se Exerise 3.13) Pk = m(%cos0(t) + ysinO(t)), py = m(-LsinO(t)+ yeos6(t)), and p; = ma, ‘0 the (canonical) momentum vector p' in the rotating frame is P= pil + py +P = mi({'eo80(t)~J'sind(0) + mi{"sind(t)+ Jeos0(0))+ mix” = mils mjl-+mik- p. equals the (canonical and kinematic) momentum vector p= mi inthe inerial frame. (©) The Hamiltonian W'(x,y'2'p PysP2) is Hea pie'+pyi" +p! -L! =p +n? +p2)-oler5 23) ‘The first term in His ip'®/2m and equals the Hartonian Hip? /2m inthe inert frame, The second tem oaless <7) olen, - whee L isthe angular momcanm inthe ineal fame. So the Hamitonlan Inthe Tastag fares HisH-aL, ‘As pointed out in Lagrangian and Hamiltonian Mechanics, H’ is like a “free energy.” Exercise 6.06 ‘The equations of motion for a particle of mass m and charge e moving in a uniform Imagetic ie B which pons i the -ieeon canbe obtained fom a Lagrangian (ee L = f(a? +7? +22) + (€B/20)hxy - yi). (a) Find the momenta (p,.P,.P;) conjugate to (x,y,z). (©) Find the Hamiltonian, expressing your answer fist in terms of (x,y.244,j,2) and then interms of (Xy.2,PysPy:Pe)- (©) Evaluate the Poisson brackets @) (mi.m0§] Gi) (mi,Exercise 606 us Solution (2) The momenta (sy. Ps) conjugate to (x,y.2) are Pye Ee em- + ok (©) The Hamiltonian is H= pyi+pyj+pe-L ~(m-2 Anca? 4? 422 ym? +j? +i) }éo (nie Bayon ee ee GyY.2,PasPyoPe)e We have B. B. zd ste (nc o$B) (ny -SBe) ach (©) The Poisson brackets of the components of the "kinematic momentum” with one another are together with [mj,mé]=[mami]-0. ‘The Poisson brackets of the components of the "kinematic momentum* with the Hamiltonian ae then {mist} = [mip m(e? + +#))-Imiasp-By, (mint) = [my pm(e? +9? +24)] [ami - Bi, [met] 6.16 Chapter VI: Hamilton's Equations ‘These lead directly to the equations of motion for a charged particle in a uniform magnetic field, namely a a ) = [mie] - 2, Hoi) =[mj.n]-- 23, and Hi) -[mi.tt]-0. Exercise 6.07 Consider the one-dimensional motion ofa particle inthe followi each case sketching representative trajectories in (x,p) phase space: (@)an infinite square well V(x)= 0 for Oa @)abouncing ball VGX)=mgx for x>0 Ve forxs0 (©)a simple harmonic oscillator V(x) = Fx? potential wells, in (@)adouble well V(x) =-}hx? Hey Solution (@ A particle of mass m and energy E in an "infinite square well potential” moves with constant momentum /2mE from x=0 to x=a where it reflects, its momentum changing discontinuously to -V2mE.. It then moves with constant momentum ~J2mE from x= a back to x= 0 where it reflects, its momentum changing discontinuously back to VamE, andthe cycle is repeated. A represenative phase space tnjectory is shown in ig. 1. Ex. 6.07, Fig. 1 (©) A particle of mass m and energy E in a "bouncing ball potential” moves, for x > 0,50 as to keep the total energy constant, ae E+mp-e.Exercise 607 nT ‘The trajectories in phase space are parabolas with x-intercept E/mg and p-intercepts sV2mE. At x=0 the particle reflects, its momentum changing discontinuously from ~V2mE to +V2mE. A representative phase space trajectory is shown in Fig. 2. o | a Ex. 6.01, Fig.2 (©)A panicle of mass m and energy E in a “simple harmonic oscillator potential” moves ‘0 as to keep the total energy constant, 2 Fyhotee. 2m 2 ‘The uajectories in phase space are ellipses with x-intercepts 2Y/EE and p-intercepts -2VImE. A representative phase space trajectory is shown in Fig. 3. Ex. 6.07, Fig.3 (@) A particle of mass m and energy E in a “double well potential” moves so as to keep the total energy constant, 2 ‘ weP lat Sak tag zaee ane ‘The equilibrium points satisus ‘Chapter VI: Hamilton's Equations and are hyperbolas; (0,0) is a hyperbolic point. The trajectories near (Fa,0) are given by Fstien? Bebe Fo tkxsa? B+ E and are ellipses: (Fa,0) are elliptic points, Representative phase space trajectories are shown in Fig. 4. The trajectories for which E» 0 are the separatrices. Exercise 6.08 {In the potential wells of Exercise 6.07 the motion is periodie but not necessarily siraple hharmonic. The action variable lis defined by 1 teh foun ‘where p is the momentum, and the integration is over a single period of the motion (see Chapter IX, Lagrangian and Hamiltonian Mechanics, for further details). (@) Show that the action variable isthe area enclosed by the orbit in phase space divided by 2x and is given by it VERE VOD) dx,Exercise 608 9 ‘where E is the total energy, V(x) is the potential energy, and x,(E) and x3(E) are the Tower and upper turning points of the motion. (b) Show that the period + ofthe motion is given by Solution (@) Consider a particle of mass m and energy E in a potential well V(x). If V(x) = E for x=; and x= xz, and V(x) OE ‘The second and third terms, which come from the dependence ofthe limits on B, give ‘zero contribution since the turning points satisfy «/2m(E~ V(x,)) = 0. We are left with 2p es *)e ¥@) where v(x) = /2(E= V@))/m is the speed of the particle at position x. We recognize the last integral as the time for the particle to go from x, to Xz, half a period <, so we have al a2 G.120 Chapter VI: Hamilton's Equations Exercise 6.09 (@) Evaluate the action variable I(E) (see Exercise 6:08) for: @ an infinite square well VG@)=0 for 0a a bouncing ball V(x)=mgx for x>0 Vo) forxs0 ii) simple harmonic oscillator V(x) = Lex? and use your results to find the periods of the motions. @) In the “old quantum mechanics" of Bohr and Sommerfeld the action variable T was, quantized in units of & (Planck's constant divided by 2x). What does this give for the energy levels of the systems of part (@)? ‘Solution (0 (Fora panicle of mass m in an “infinite square well potential” the action variable is 1-4 Vim ox «2. Jim, ‘and the period is a, fm rote oa PB. To check, note that this equals 2a/v, where v= /ZE/m is the constant speed of the particle, and is indeed the time required to make the round trp. Gi) For a particle of mass m in a "bouncing ball potential” the action variable is A gt a= mp dx = peer, and the period is al rere, cs To check, note that inthe half period "top to bottom” the particle falls a distance E/mg in atime 1/2. These are related by the well-known equation for free fall, ‘which, on rearranging, gives the preceding expression fort.Exercise 6.09 11 (Gi) For a particle of mass m in a "harmonic oscillator potential” the action variable is a PamE— Fix?) dx. ‘The integration can be performed by setting \2EFEsing and dx = 2E7E cosa togive 1028 fae? cost guy = {E7ia is the angular frequency. The period is dl 2x roan 28 which should be familiar. (b) @ For the infinite square well, Bohr-Sommerfeld quantization says ToS ymEy 0k where = 0.12 which gives the energy levels again n= 0,1,2, ee 2 Vim _ oe Te SUEY ook where n=0, E,(BS)12 (Chapter VI: Hamilton's Equations Compare this with the result given by the WKB approximation in quantum mechanics, E,(WKB) = (3n(n+3)/2)""(me?a?/2)"” where n=0,1,2,--. Also, compare it with the exact result given by quantum mechanics, E, (exact) = h(mg?#?/2)"” where the ‘hy are the roots ofthe Air function, Ai(-2,)= 0. See Table 1. “able 1, Energy of quantum “bouncing ballin unis of (ng??/2)! E,(BS)= E,(WKB)= E,(exact) = 0 23203 "2.3381 28108 = 40818 «4.0879 44618 «= «55172, 5.5206 58467 6784567867 7082779425 7.9441 wn RUNHoF 1261 119353119360 (Gi) For the simple harmonic oscillator, Bohr-Sommertela quantization says Tye Ekenh where n= 0,12-~ ° Which gives the energy levels E, = nho. Compare this with the result given by quantum mechanics, E=(n+4)ho where n=O, Exercise 6.10 (2) Use the definition of the Poisson bracket to establish Poisson bracket properties 1, 2, 3, and also the Jacobi identity 4, ‘(Show that these four properties also hold for the commutator bracket. Solution (@) For Poisson brackets we haveExercise 6.10 13 (fe a sale av -Hare) +u) Sued 2) “A(R Be e-2) a) aa 2p4 * pe) 29 -S(A 2). $(t et) 2902. @u O46)” A204 30x ap a, =foile[on). a (cu) dv _ Xcu) ev | ud 2 2. ap, 2, “Yea “Qa, -doy), bon S{ee-ney aren 2 le Si) on alin ~e) o[ovher toon Tis proves te first three properties of Poisson brackets. To prove the fourth property, the Jacobi identity, we stat with bios oe bol Se Re and work out14 (Chapter VI: Hamilton's Equations boo $3 aaa dua av 2 2u)) Pp Ay 344 Pe 9Ps 24y, ee ae ay aw av Sw) a5 apvaa, ps BpaPae 3a av aw ay aw av aw }} “apy {aya @s * Ba, 3452, Saye, O44 Ps dayda,)) Permuting u, v, and w cyclically, we obtain similar expressions for [v,[w,u]] and for [w.{o.v]} Adding these, we ind that all ers onthe right-hand side cancel and we thas sbtain the Jacobi identity [ofv.w]}+fofosul} (b For commutator brackets we have ewv])=0. [os + vaw]= (oy +05)¥= (0, +0) = (oxy v0) + (09¥- via) =[o.v}+[us.v]. [ev,v] = (cuyv - v(cu) (uv ~ vu) doy), [uw] uc) -Cowyu = (ov- vu)w+ v(uw— wu) = [owes ofo.w]. ‘To prove the Jacobi identity for commutators, note that ufv.w]]= ueow- wy) =(ow= we), ‘fava = v(wra— uw) = (ou uwyy, w{u.v]}= wav - vu) (av vaw. ‘Adding these, we find that all terms on the right cancel and we obtain the Jacobi identity.Exercise 6.11 125 Exercise 6.11 Let £(@(1),p(t) be some function ofthe canonical variables, and fo its value at time 0. {2 Stow, ite Hamiltonian H is ine-independent, thatthe funtion fat time ts gven y f= fo tlfosHl + ((7/20ifo. HLH) + (© /39Mllfo,H} HLH) (b) A particle of mass m moving in one dimension x is acted on by a constant force F. “The Hamiltonian is H= p?/2m~ Fx. Suppose that at time Othe panicle is at xo with momentum po. Use the result of (ao find the position x and momentum pat ime t (@ A piticle of mass m moving in one dimension x is in a simple harmonic oscillator well. The Hamiltonian is H= p?/2m+1m%x2/2. Suppose that atime Othe particle is Arg with momentum pp. Use the esl of () wo find the positon x and moment pat time Solution (8) We expand fin a Taylor series, ere totye(4) 2(8 ‘The successive time derivatives of fare given by af at fae rum, oe H]- [lau : sowe have 2 fefor font} Sfltonpie- as required (®) Consider a particle of mass m moving in one dimension and acted on by a constant force F. The Hamiltonian is Faw ne Bor, ‘Taking f =x we have126 (Chapter VI: Hamilton's Equations ‘Since this last Poisson bracket is a constant, the higher order Poisson brackets give 2210 and we have Taking f= p we have [pH] - fp aA Hp.x]-F. Since this Poisson bracket is a constant, the higher order Poisson brackets give zero and we have p=po+Ft. ‘These are the well-known expressions for x(t) and p(t) for motion at constant acceleration F/m. (©) Consider icle of mass m moving in one dimension in a simple harmonic Shiate et Fe Haitians : Etats? 2 2m" 2 . Taking f =x we have 2 bette fa Grote] =i feo] Blea) 2 Usha] [22 + das] 2o'fpa?) obfpa]=-o%, [e}u}n]--0%s.H]--07 2, nd so forth, We thus find renin dere =ocorors Bsinot,Exercise 6.12 127 ‘Taking f =p we have [p.tt} pee ~mo'{x,H]=-0"p. p.H} HH] - ~«*[p.H]~ mon, and so forth. We thus find ~moxosinat + pocoswt. ‘These are the well-Anown expressions for x(t) and p(t) for simple harmonic motion, Exercise 6.12 ‘The Hamiltonian for a simple harmonic oscillator is Introduce the complex quantities a= 2 (e+ (@) Express H in terms of a and a (&) Evaluate the Poisson brackets (2,2), (8.H1], and (8",H). (© Write down and solve the equations of motion for aand a". Solution (@) We have mo (2 Z eee 2(e 5), so the Haitonan fora simple hamonieoxilstr canbe writen ‘ ca Psd mote? un " 2m" 2128 ‘Chapter VI: Hamilton's Equations (©) The required Poisson brackets are uate Ban Bi, [an]= fa’ Jno ae wa, [ot] =f aa] a] (©) The time rates of change of a and of a” are thus aa at ‘These can be integrated to give a HI] = ~i a ]=~iwa and “fet ange and at majo where a and a3 are the initial values of a and a “This provides yet another way to obtain the general solution to the harmonic coxcillator problem, Exercise 6.13 (a) Evaluate the set of Poisson brackets for a component of the radius vector F= (x,y,2) ‘with a component ofthe angular momentum L= (Ly,Ly,L,). Also evaluate those for a ‘component of the linear momentum p= (p,,Py.Pz) with a component of the angular ‘momentum. Show that the results can be put in the form f.L-n}=nxr [p.L-n]=nxp ‘where mis an arbitrary constant vector. (b) Use the results of (8) to show thatthe Poisson bracket of « component of the angular ‘momentum with an arbitrary scalar function of r and p, of the form a(s?.e-p.p*), is (© Use the results of (a) to show that the Poisson bracket of a component of the angular ‘momentum with an arbitrary vector function ofr and p, ofthe form Avareap+arxp,Exercise 6.13 129 is given by [AL-n}=nxA. (@ Show that the Poisson bracket of the square of the angular momentum L? = 11 + LS +12 withan abivary vector function A of rand pis given by [AL] =2Lx A, Solution (@) We have GLgl=(uyPe-2Pp]= 0. (Yibg]=[¥.¥Ps ~2Py)*~2, fee] [2.yPs~2Py1= ys (eLyl= Exe, xPg]=2, [¥Ly]=Ly.2p_—sPe]= 0, l2aby]= C229, —xpp]*—*, Geb fxn, - ypu] -y. Dike = Ly.ap,—yPad=%, [sLz]=L2.39, yPa)= 0. Now let n= ngl+nyj+n,k be an arbitrary constant vector. The preceding equations can then be written (uL-n}=nyz-ny=(axr),, Dik-n)en,x-n,2~ (0x1, [.L-n}=n,y-nyx=(ax0),. ‘These, in tum, can be summarized as [n.L-n)=nxr. ‘The set of Poisson brackets for a component of p with a component of L can be found similarly. Itis, however, simpler to obtain these by interchanging r and p in the above uations. There are two compensating sign changes, one inthe definition ofthe Poisson ‘nce and on inthe defini ofthe angular momentum. We ths find (BL-n}=axp. () To evaluate the Poisson bracket of an arbitrary function of r*, p?, and r-p with [L-n, we first consider the Poisson brackets of 17, p?, and r-p with L-m. We have (2.L-n)=(er.L-n)=2e-(r.L-m]=2r-nxr=0, [p3,L-a}~[p:p.L-n}~2p-[p.L-a}~2p-nxp=0, (e-pL'a)=r-[pL-n)+p-(e.L-n]~r-nxp+ p-nxr =O. We then need the Poisson bracket (u,¥] where w is some function u(Gy.E20~-) of fanetions &,83,~ of the canonical variables. For this we have130 (Chapter VI: Hamilton's Equations LER eha) Lee the second equality following from the chain rule for partial derivatives, Tous 2 -p.p?),Len)= 28 Lens Sa me a (@e-pp? Lem) St? La] fr pala] (eh Lal = 0 ap) ©We have [A.Lon)=(ar+2,p+a3rxp.L-0) = a{t,L-n] + ag{p.L-n} + aslexp.L-n) maymcr+agnxp+ayex(p,L-n}+agfe,L-n)xp mer + apn x p+ ay(r (ax p)+(0xF) xP) = aynxr+aynxp+ajax(exp) anxA. ‘This result can be obtained much more easily by looking ahead to some results in Chapter Vil, Lagrangian and Hamilionian Mechanics: The change 8A in a vector A under an infinitesimal rotation 80 is BA =80%A. (On the other hand, this is a canonical transformation generated by L-80, so 5A =[A.L-56]. Comparing these expressions, we find [A.L- 80) = 80% A 8s required, (We have (A127) =A. +05 +0) = DLglAsLg]+ 2Ly[AsLy] + 2L {AL} = 2L (LA) + 2Ly(Jx A) + 2L, (ex A) =2LxA.Exercise 6.14 131 Exercise 6.14 Consider motion of a panicle of mass m in an isotropic harmonic oscillator potential ‘Vir? and take the orbital plane to be the x-y plane. The Hamiltonian is then Lots piye deat sy HS. = 5h +pp)+3kO? +92) Introduce the tee quansies 1 1 1 Si LOE phe duty, S.-Logp, +kay, S067, -999) with w= Fm (@ Show that [8.5] =0 80 (S,.5.S3) are constants of the motion, (b) Show that [S,S2]= 2083, (S2,S,]=20S,, [83.S)} = 2053, ‘80 (2w)"(S,,S;,S3) have the same Poisson bracket relations as the components of a “three-dimensional angular momentum.” ©) Show that Sh -Sf+Sh+Si. (The corresponding quantum relation has (hw)? added to the right-hand side.) Solution (a) We have 1 ae ee 1 1 [SoSi}= [pret +79)+ 5HO? + ¥?dg GPP) + GHG? -¥)] elie)132 (Chapter VI: Hamilton's Equations [SoSa]- [RO +9) +380? +7 ny +b -E (sie) +f] +[090)]eb.005) = (erp ler] +2304[oy2]+2504L.P2] +230 y.Py)) = E (typ. - 200, + 2p, +2304) a [So8s]- lw +P + fk? +y?).009, - yp 0] 2 pion} fm) Sel fen.) ~ 2 onolner]-2erfes)oAhol-29fsJ+20fy0y) o 1 EAP, +2p,py)+ qho(-2ey + 2x9) <0" So (S,S;,8) ae constants ofthe motion. (>) We have [s.2]= fact -H)+ 302-2 pny +] = Efetny-[oho}s [enue 00) ~£(2vpdloe]-20Py[oy-y]+250,Lspe]-2y04{y-79) =2E (om. +20y) = 208;, [s.s]- [dpe ++ kxyse(py - wo] +2 pay |-[.ny 0s) +400099,]-fo.30)) = 2 (o3focr]-rifeyo) »0fy.0,]- Esa) = £(¢h-1f) +e0f2-»") = 208;Exercise 6.14 133 [SS\]-[oe7, v0 dgR Ct -p+302-¥4) - Zl, i]+ feet} diel ¥ +f.) = 2 osrylups]+20.0,[v.0,)-Leo(20fr,.1]+ 20.) = 2 pap, + 2kaxy 208, Ss) have the same Poisson bracket relations as the components of an abstract "hree-dimensional angular momentum." (©) Rather than approaching this part of the exercise head on, itis beter to begin by observing that Sue (ghatetee) (oot ede’). s-(fork+ de?) -(het do?) and thus Ssh nalgeebe gee) gars"): Also note thet S+5-(Low, +b)" «Xbm-a0AF = darted oidetyt Earp} 3 + ¥°p2) -(latew (de). Comparing, we se tat S-St + S3+S3, Sh = St+S3 +35 as required,134 (Chapter VI: Hamilton's Equations ‘In quantum mechanics the set of three quantities (2w)"(S,,S2,S3) have the same Poison ice wom ih oe tne incase metic ey weap Neen tt rlecnarenare he cae ese fiat ie Sen taser monet in uaen mec Ae Pee Su)? (Sa)? (Ss)? (2)? (=) (38) +(33) ~(Go) ~“r where E is the energy of a two-dimensional harmonic oscillator with angular frequency © (note the “extra” term ~A?/4). Angular momentum theory shows that the eigenvalues ofthis quantity are A(4 + 1A? where X= 0.4.1.4. so the eigenvalues ofthe energy of a two-dimensional harmonic oscillator are BE, =AwQA+1)=Awln+1) where n= 0,1,2, Exercise 6.15 Consider motion of a particle of mass m in a gravitational potential V=—k/r and take the orbital plane to be the x-y plane. The Hamiltonian is then 1 gry pay_k n= hoterp-* ‘The angular momentum vector points in the z-direction and has (-) component La py-¥P x and the Laplace-Runge-Lenz vector (see Exercise 1.12) lies in the x-y plane and has components Ky=pyb-mkyr, Ky =-pyL-mky/r. (@) Show that (LH]-0, {K,H)-0, [K,,H}=0, 50 L, K,, end Ky are constants ofthe motion. (©) Show that (KyL]=-Ky, (KyL]=Ky,(K,.Ky]+-(mk)LExercise 615 135 (see Exercise 6.13 for some useful Poisson brackets). Now restrict yourself to bound sty tn ew (pg!) re bracket relations ‘three-dimensional angular momentum." Slow ta ihe squue of te engi of his “hey dimensional eagular momenta” is BE the comes relation is BE 12, + (he comespondng quantum relations ME — 442.) Solution (2 Sine is the 2 component of the angular momentum and H is a seaar, the Poisson bracket of L with H is zero (see Exercise 6.13), (LH) -0. SoLis a constant ofthe motion. “The Poisson brackets of K, and K, with Hare given by Us] [eye-muk cE (ot +3)-4] =n ford] Ef oto] 3Q)-m2(6)an 26) ~¥b0)-05)(3)- ) won (!-2 54} -0. [pai] [-vat— med (ot +08) fp! 2[rP* +] AG) maa) =e -sX3) 493) “193-5 se SF) =o. So K, and K, are constants of the motion (see also Exercise 1.12).136 (Chapter VI: Hamilton's Equations (b) Ky and K, are the x- and y-components of the three-dimensional Laplace-Runge- Lenz vector, K-=pxL-mkr/r, and L is the z-component of the three-dimensional angular momentum vector L = rx. Exervise 6.13 shows that for any vector K, [Ku-n}-axk. Setting n= k and taking the x- and y- components of the resulting equation, we obtain the Poisson brackets of K, and K, with L, [Kut]-(kxk),--K,, [k,.L]- (ek), -K,. ‘The Poisson bracket of K, and Ky with each other is given by ty) yt-mt2-rit-m2] . rytorat]-mefoyt2]+ mi Frat] ~-pytttare)-patfeyet]- Ey ftoy]- pao) -mi{[Py-2]- Exercise 6.13 then shows that 2 ay =-(ot eri). BEL mi £37) tap? o-an( 28h [x pil pis BE(oyx~pas) «mic . je “ant. For bound states H= -IEI where E is the energy and we can write the preceding results in the form, fete site [ital sade: bate eatExercise 6.15 137 ee ny ot te uns (dest) ete te oon ‘bracket relations with one another as the components of an abstract “three-dimensional angular momentum." (Now consider Ki +K}-(pyt-mke)"+(-psl-mb2)” ~(05+ 2-20, yp Jeet Ae wanf BEE Ee ga? cas = 2mHL? + mk? . rw avout tund sas, or wich = 18 wh st energy, we can Phat (Ke J 7 \s 2 me (anne) * (Jame) * ar “This says that the square of the length of the abstract “thre-dimensional angular momenta” mk? K In quantum mechanics he set of tee quantities (FRe>, ter.) defined in an appropriately symmetrized way, have the same Poisson bracket relations with one ‘another as in classical mechanics, so they are again like the components of "three- dimensional angular momentum.” “The square of the length of this “three-dimensional angular momentum” in quantum mechanics is kK, Z kK, mk? 9? (aise) +(e) + Fa 4 (Note the “extra” term ~A?/4). Angular momentum theory shows that the eigenvalues of this quantity are 2(4.+1)A? where 2 = 0,4.1.3,~, $0 the eigenvalues of the energy of & ‘two-dimensional hydrogen atom are pre eee GH Gaeta ane NE BOONCHAPTER VIL CANONICAL TRANSFORMATIONS Exercise 7.01 ‘Toe motion of a particle of mass m undergoing constant acceleration a in one dimension is described by xoxo + Pesta? = po +mat. ot Pte sa, pe po +mat ‘Show thatthe transformation from present “old" variables (x,p) to initial “new” variables (Kq,Po) isa canonical transformation (@) by Poisson bracket test (by finding (for t» 0) the type 1 generating function F,(x,x9,). Solution (#) The initial "new" variables (xo,Po) are given in terms of the present “old” variables (xp) by the inverse transformation Po = P~mat. ‘The Poisson bracket of the initial variables with respect to the present variables is then rool, = 20220. 280.2 «14 lrorohey = Sap ap ae to ‘Since it equals 1, the transformation from (x,p) to (Xo,Po) is canonical. (b) In a type 1 transformation the coordinates x and Xo are taken as the independent variables; the momenta p and po are expressed in terms of these, =m 22204 -m32%0_1, pemic70y Smet, py =m5—32—> mut, [Now consider the differential formExercise 7.02 139 Pax podkg = (m2—22 + Lat) dx ~(m2—20— Lat) dxy ~o BSA 4 mas) ‘Since itis an exact differential, this again shows that the transformation is canonical. ‘This approach also gives the type 1 generating function, Fro) BESADE 4 Lanner (9 where f(t) is an arbitrary function of time. Exercise 7.02 (2) Show that Qe-p, Page Ap? (where A is any constant i a canonical transformation, (by evaluating [Q,Pl, (Gi) by expressing pdq - PdQ as an exact differential dF(q,Q). Hence find the type 1 generating function of the transformation. To do this, you must first use the transformation equations to express p, P in terms of q, Q. () Use the relation F, = F, + PQ to find the type 2 generating function F,(q,P), and check your result by showing that F indeed generates the transformation, Solution (© @ The Poisson bracket of the new variables with respect tothe old is Since it (Suppose hat we regis a type 1 transformation, king he coordinates, and Q as the independent variables. The momenta p and P are then given by p=-Q. P=q+ AQ’. Now consider the differential form dq - PdQ= (~Q)dq~(a+ AQ?)4Q= 6{-aQ- 440").140 (Chapter Vit: Canonical Transformations Since itis an exact differential, this again shows thatthe transformation is canonical. The type 1 generating function is Fi. =-aQ-4Ag°, (©) Suppose, instead, that we regard this as a type 2 transformation, taking the old ‘coordinate q and the new momentum P as the independent variables. We can find the type 2 generating function from the type 1 generating function by setting F,(Q.P)=F, +PQ=-aQ-FAQ? +94 AQQ~FAQ"= 2a a ‘To check that the final expression is indeed the type 2 generating function, we evaluate its derivatives, tp eg ea (a pea (@), TxPoa=p and (P), yah =a ‘These areas required, Exercise 703 ‘The Hamiltonian fora particle moving vertically in a uniform gravitational field g is +mzq. 2m (@) Find the new Hamiltonian for new canonical variables Q, P as given in Exercise 7.02. Show that we can eliminate Q from the Hamiltonian (make Q cyclic) by choosing the ‘constant A appropriately. (b) With this choice of A write down and solve Hamilton's equations for the new ‘canonical variables, and then use the transformation equations to find the original variables q, p as functions of time. Solution (@) We introduce new canonical variables Q and P by seting, Exercise 7.02, a=P-AQ, — p=-Q. ‘The Hamiltonian forthe new canonical variables is QF K(QP) = H((Q.P),p(Q.P)) = Fs mp - AQ) = map +( 1 - mga)?Exercise 7.04 41 Une ale t . ia ‘we can simplify this to omg. (b) Hamilton's equations for the new canonical variables (Q,P) are then ms, £0, with stn Q= Qo +mgt, P= Po. Here Qo and Pp are constants, the initial values of the new variables. The equations of the canonical transformation then give the original variables as functions of time, qw= Po sary Qo met? - (0-5%,) ~ ‘This is the well-known solution to the free fall problem. We see that ~Qy isthe initial ‘momentum and Pp isthe maximum height reached. Exercise 7.04 (@) Show that Q=qcos0--2-sing, P= magsin6 + poos8, is canonical transformation, ‘) by evaluating (Q-Pl,9 ii by expressing pdq- PdQ as an exact differential dF,(q,Q.t). Hence find the {ye | geoeuing fncon of he eansfornton. ‘To dos, you mest Fist we transformation equations to express p, Pin tras of, Q. (©) Use the relation F =F; +PQ to find the ype 2 generating function F,(@,P), and check your result by showing that F indeed generates the transformation.142 (Chapter VIl: Canonical Transformations Solution (@@ The Poisson bracket of the new variables with respect tothe old is = c0s0 e080 -(-— sind) (masind) en pepe ape eel od ine (Gi) Suppose that we regard this as a type 1 transformation, taking the coordinates «qand Q asthe Independent varabes, ‘The momenta p and Pare then given by Pa mo(=25-deut). pda P ad mo g0t0--2:) eq -mo( 2 -Qetd)ag mo(q? +Q°)e010 -mo-$2) Sine ins am exae differential, this again shows that the transformation is canonical, The ‘ype I'generating function 9 F@.O- imo? + Q%)cot-mo (b) The type 2 generating function F,(q,P) can be obtained by setting A muog? = zug? - Q? e010. ‘We must sil express this in terms of the appropriate type 2 variables, q and P, by setting ‘This givesExercise 7.05 143 P,Q.) arf (35 Zand) Joo cos8 mma ae ae ‘al 2 (0 mar) cos ‘To check this expression, we evaluate its derivatives a) |? (a) a P ") al ‘The first term on the left is @ function only of x, the second is a function only of y, and the third is a function only of z. Since the sum of the three terms is a constant, each must ‘equal a constant. We set 1 (4X)? 1) ama) *28°-8e Where ci, dy, and a are separation constants with a, + ay +a, =E. These give Xa f(a abd, Y=fyfima,— mig dy, Z= fY2ma,—mks! de, so complete integral to the time-independent Hamilton-Jacobi equation is ston ate) ta? =a, dy,Exercise 806 169 W = fx[2ma, - mix? dx + f,[2ma, ~mky? dy + f-f2ma, ~ mkz? dz ‘The canonical transformation generated by W gives toon ts plus two similar equations for the y and z coordinates. The integration over x can be performed by setting Ray, = one prsings, d= Ph c0sh, dy. en Espo so x0 + ft Ec Similarly, we find Pray fe Pa gn [E 7S sin [Ecce 2) = fe sin fEces HOY PEsia FE Ce+,) and 2(0)= |Sesin Rees By. (©) The Hamiltonian for a three-dimensional isotropic harmonic oscillator is, in spherical polar coordinates, We find ute L(pbechabecrace) et ‘The time independent Hain Jacobi equtons 1 1_ pay anlar) C38) *zamtal an) |S -E Separation of variables proceeds a in Lagrangian and Hamiltonian Mechanics (pages 157, 158), and we find a complete integral 1 ae = wea fee Ea f= gss ‘The canonical transformation generated by W gives170 (Chapter VIIl: Hamilton-Jacobi Theory 2 1a aw fm ar See Fy 72 a am Aye | a [ cates) | ae bee bs Ae [tat Sata) ** “The integration ofthe third of these equations is done in Lagrangian and Hamiltonian Mechanies (pages 159,160) withthe est is Bu, sini sinO sin(@~ By, )= cosi cos0. ‘This says that the orbit isin a plane through the origin. ‘The 6-integrtion in the second ‘equation is also done in Lagrangian and Hamiltonian Mechanics (pages 161, 162), $0 this equation becomes at “ O-B. ~ tl; pal yaualae 2 2m? Here @ i the angle, measured inthe plane of the ort, between the radi vetor and a reference direction. "The rintogration here is done in Bxeeise 1.14 with the result 1 ne} pe 3-3 p= Beco} ‘This says that the orbit is an ellipse with geometric center at the force center. The r- integration inthe first of the canonical transformation equations is also done in Exercise aca 3 Atlan Feet) ‘This says how the particle moves along the orbit in time,Exercise 807 m Exercise 8.07 Use the Hamilton-Jacobi method to study the motion of a particle in a dipole field with (aon-centra!) potential vk, (#) Write down the time-independent Hamilton-Jacobi equation for W in spherical polar coordina tes. (©) Show that this equation can be solved by the method of separation of variables, and obtain an expression for W of the form W = W(¢.0.9:E.a3,a3). Your answer will also involve certain integrals; you need not evaluate these at this stage. (©) Interpret physically your separation constants a,c by oblaining p,,Pe.Py in terms ‘of 1,0,4,E.a.0. Hence show thatthe z-component L, of the angular momentum of the panicle is constant, and further thet 1? + 2mikeos® is constant, where 1? is the ‘Square of the total angular momentum of the particle. By considering the equation aw Feo th find how r varies with time. Solution (a) The Hamiltonian for a particle moving in a dipole field is, in spherical polar ‘coordinates, 1 oi) «Keg ne L (+5 8 + tart) + Bm\PF* EPO YT agg P*) ST so the time-independent Hamilton-Jacobi equation is A {rawy?, 1/awy?, (ay) tte. Ey 2m(\ar) *Va0) * Panel Te (®) We try a solution of the form W-R()+00)+06). ‘The Hamilton-Jacobi equation becomesm (Chapter Vill: Hamilton-Jacobi Theory _ 2a) 4 “This can be rearranged in the form aa anto ‘The left-hand side is a function only of rand 8, and the right-hand side isa function only ‘of. Both sides must equal a (negative) constant. We set arate (SB) +L, (S0)' ote}, (20 ‘The separation constant L,, tums out to be the z-component of the angular momentum. ‘The r0-equation can be rearranged in the form owe) 2) (2) “hae ‘The left-hand side isa function only ofr, and the right-hand side is a function only of @. Both sides must equal a constant. We set seas) (8) hyeamonee Note that the separation constant a is not necessarily positive. The variables are now ‘completely separated, and we can integrate to obtain Ref me Sar, e-| aH ~2mkcos0ee, onLe ‘Acomplete integra tothe Hamilton-Jacobi equation is thus w= fise-S a [fog antennasExercise 8.07 173 (©) The complete imegral W, is the type 2. generating function of « canonical transformation. Half ofthe transformation equations reads Ww a w 3 a fame-§, S78 gate ~2mkeos0, re Since py is the z-component ofthe angular momentum (See Lagrangian and Hamiltonian Mechanics, page 158), this identifies the first separation constant L,. Further, since B+ p6/sin®@ isthe square of the total angular momentum, L?, the second separation constant ais a=L? +2mkcos®. (€) The other half of the transformation equations reads Bert Som “, "| me -af™ nota (asf uf fo "aa" 2) Yome-a/e? ” 2) Ya 3/sin?0- amkcos6" a be | Tastes taacon ants) ** ‘The -integration in the first ofthese is easily performed, and the result can be rearranged to give Gee ga PSs Eee Be) ‘The nature ofthe radial motion depends onthe signs of E and a. If both are positive, the panicle comes in from infinity, eaches a minimum radius of Ya/2mE at t+Bg =0, end ‘goes out again to infinity. If Eis positive but a negative, the particle, starting at t= ~, ‘comes in from infinity and reaches the origin r = 0 at t Bg = ~Val/2E (or it starts at the origin r=0 at +B = vial/2E and goes out to infinity t= =). If both E and ot are negative, the particle is confined to the region rR and the circle does not enclose the origin (Fig. (a). If cy is negative, po > Vo) limits of your expressions (refer to Exercise 1.01). Solution ‘The momentum of the particle, at position x, is p= f2m(E- Votan?) ‘The particle oscillates back and forth between limits =A where ‘The action variable Is given by Hef a(E- vows?) x 2a Pa Se Sh era cost ~#| alata 7 [Ea else eof em wate where @= 70/24 and Ong =A/28. The first integral can be performed with the substitution E E sins fer yjsines — eo0dd= eA eosada,Exercise 9.01 193 and the second can be performed withthe substitution o _E tand= J=sinp, — —“F-« | Ecos. Vo cast Vo ‘The results Atal EFVo - Wo} Solving for the energy, we find (ype) -v,-2 Morr (Wot aim) “YO"TVam!* ama? “The angular frequency of oscillation is e Vv. x J—2 +. I. $2 [ees ee A = hs fe - 8 pave| fo -i)et Bee we ot 12am fo fet Ben at on oe TX) ‘nVp)_ 2 wormne(E) 3S fh ee xVo/2a2. At high energy (E>> Va) the action variable and angular fequency become 1-42 ame and ont EZ. an a\m In tis limit the system behaves like a particle in an infinite square well potential (V = 0 for Ixia and V—» © for Ixl> a).194 Chapter IX: Action-Angle Variables Exercise 9:02 A panicle of mass m moves in two dimensions (x,y) in a non-isotropic simple harmonic oscillator well Voxy) = fmaix? + dmasy? where in general ©, # 05. (0) Find the action variables (1,,1,), and express the energy in terms of these. (©) Find the angle variables (@,.6,). and express the cartesian coordinates in terms of the sction-angle variables. (©) Write down the angle variables and the cartesian coordinates as functions of time. (@) Sketch the trajectories ofthe particle in (x,y) space and in (,6,) space. Solution (@) The Hamiltonian is nL fptent]edmots? Lmaty?, ‘0 the time-independent Hamilton-Jacobi equation is jaw)? (aw)” alae) (3p) [tated smi" We aya slain ofthe form W=X(@)+ YQ). ‘The Hamilton-Jacobi equation becomes [st do] tay dae ‘The first term on the left is a function only of x, the second term is a function only of y, ‘and their sum is the constant E. Both terms musi equal a constant. We set 2 ed imw2e? cee ke aay mols? mo? ade te contons being consieed er te Bead mes many eu 8 th hoops gh ty, one Foaion We can then, to 8 good approximation, the Fahrhand sles of Hanton's eqns by the averases ov he postion ofthe bead ‘round the hoop. This gives S1Q)-Blonean(2} M2 elf Bae where A=dfirsinads isthe area enclosed by the hoop. Integrating with respect to time, we see thatthe conjugate momentum pis approximately constant, p=, and the change sss Ts 4 Pp 2A, tse fate B24 40 ‘where 49 = ([Odt isthe angle through which the hoop is turned. The frst tem in As is the dynamical change, and the second term, ~(2A/£)A8, is the Hannay change.Exercise 9.12 217 Exerciae 9.12 (Consider the "generalized simple harmonic oscillator” with Hamiltonian H = 4(Xq? +2Yqp+Zp*) where (X.Y,2) are parameters with XZ > Y?, (@) Show that the trajectories in phase space are ellipses and hence find the action ‘variable, showing that itis I= E/io where « = VXZ~ Y? isthe frequency. () Express the variables (q,p) in terms of the action-angle variables (6,1). (There are various ways to do this; one way is to solve the Hamilton-Jacobi equation to find the “approprite generating function.) (©) Suppose thatthe parameters R = (X.Y,Z) are changed adiabatically (but always with XZ_> ¥?) so as to take the system around a closed circuit in parameter space. Show that the resulting Hannay angle is sone [fe 0 H, Haanay, “Angle vrable Holonomy in adiabatic excursion of an integrable Saaniuotane, Phys a 18: 221-250 C983), MV Bey, Clase adiabatic ay fd quan aba pase". Phy. A 15.27 985) - Solution ttwe et E>] ese a cana retro te anneal vase, te Hania an be written HeETME where 7 =[q__p] is the transpose of §, and M is the symmetric matrix mote Zh ‘Suppose we introduce new canonical variables 8 by setting =RE, EeRTE where Re cosa sina’ sina cosa, ‘This is a rotation of phase space through an angle a. The Hamiltonian for the new ‘variables is218 (Chapter IX: Action-Angle Variables w=(e7 Ty) <2 (awn"y =e! ‘where M' = RMRT, We can choose R so thatthe new matrix M’ is diagonal, Ip, 0 walt 2] where 2, /2 and 2/2 are the (eal) eigenvalues. Since the properties “determinant” and “trace” of a matrix ae invariant under such transformations, we have X+Y. Ade XZ-¥? and Ay +h. ‘The new Hamiltonian becomes 1 ga, da? + di aeeeiea. Now if XZ.-Y? >0, the signs of A, and 2_ are the same. The trajectories H’ = E in (@’sp’) phase space are then ellipses with axes along the q’ and p’ axes. The semi-q’- axis is ZEA, , the semi-p’ -axis is /2E/R_, and the ellipse encloses an area sea» [2B BERT = 2n8/ YR Aa = 2nb/ RDP. ‘The action variable is thos and the angular frequency is on E.xz-¥, ‘Another way to approach this problem is to solve the Hamilton-Jacobi equation [ost eaves {S))-« 4 2 4 to obtain the generating function of the canonical transformation from the original variables (q,p) to action-angle variables (6,1). The Hamilton-Iacobi equation gives Wet aed, 7 Gq Zt ZOE OLYExercise 9.12 219 Integration then gives o-tget = RZ Yq? we-34 +7 fee (XZ~ ¥")q? dq. ‘The remaining integration can be performed by setting. HE 2 an Yaz yeint a ge prom. ze (0 +singcos4). ¥ ‘As q is taken around one cycle, the variable @ increases by 2x and W increases by 2b /YXZ—Y" . The action variable is then AWE tn Yxz-vF where we have inroduoed the angular frequency «= 22 = VXXZ-Y7. These results are the same a we bined previously, Dut we, ca tow connue on tp find he ‘comesponding angle variable” We replace, in W, the energy E by is exprestion in terms ofc mon vale, obaing Z 1 = Lat +tgesingcosy) where sing a) -™ WOad=-354" +1¢+singcoss) where sing q/>>-. ‘The resulting function W(q,) is the generating function of a canonical transformation from (q,p) variables to action-angle variables. The old momentum is (Bea (2m Vat), ‘The angle variable is that is, the angle variable is simply the variable which we introduced previously for purposes of integration. The transformation from the original (q,p) variables to action- angle variables (9,1) is thusmat any sine, ‘We now suppose that the parameters R= (X,Y,Z) are varied adiabatically so as to take the system around a closed circuit in the three-dimensional ‘space. We wish to find the resulting Hannay angle. For this we need the "flux density" B-(Vax¥p) where ( ) denotes angle averaging and where van sAisnge({Z), vpe—vatange( (2) «atooer{ (22) is zero and that of sin? is 1/2, so we need only ‘The angle average of singcoss consider 4E Avz-s25(xwz+Zvx-2YV¥) = shs(-a0x+2vzvy + az-274yv2) ‘0 the expression forthe “flux density” becomes WX VY VE a--2, Z 2YZ XZ- 2y1|= (xvx+vvy+2vZ)-15, 0 yz -y/2 ‘The Hannay angle is then given by “urs fisCHAPTER X NON-INTEGRABLE SYSTEMS Exercise 10.01 Investigate the surface of section (y=0, py>0, E fixed) for the two-dimensional oeilator with Hamiltonian (refer to Exercise 9.02) a Phy Bes Lmuts? + tmuty?, Potatoes In parca, examine the nature ofthe sequences of points resulting from various stats, (@if 0, /o, isaninrational number, O)if o,f, is araional number, 4/0, «1/5. Solution ‘According to Exercise 9.02 the cartesian variables (x.y:P,.Py) are related tothe action- angle variables (yt, by the canonical transformation ref iam, p= ftaaTeth. To Yo fast, Bye RROGT a. ‘The Hamiltonian, in terms of the action-angle variables, is. Hoyt, +0yly. Hamilton's equations show that the action variables 1, and 1, are constant in time and thatthe angle variables @, and ¢, increase uniformly with time, $= oO 40,8, 450-4 Or 0,0 Let us tke a surface of section y=0 with py >0. Coordinates on the surface are then (ops). Suppose we sta the system at t=O on the surface of section at (2,p,) and with nergy E. The coordinate y is zero and its conjugate momentum is py = fimE-ph-mohe.22 (Chapter X: Non-integrable Systems ‘The various starts for given energy E must lie within the ellipse pi+mtufs? <2mB. ‘The angle variable ¢, is initially zero, As time goes by, the angle variable @, increases, and at time 2njoa, it equals 2x and the system is back at the surface of section. The angle variable 4, is then 4:() = 4,(0)4 2n(0,/a,), and the coordinates ofthe point onthe surface of section are x0) [Be sslps» 2n(004/9,)), px(1) = 2m0,T, £09{04(0) + 2, /0y)). ‘This is repeated over and over, and we thus get a sequence of points. x(0)= Fess. + 2xn(e,/t,)), pan) = fm ,F, cos, (0)+ 20m /0,)). on the surface of section. These points lie on the elipse riers taut <0, Pita)+ mols) 0. If the frequency ratio «,/o, is an irrational number, the sequence never repeats and the points eventually cover the ellipse densely. If, onthe other hand, the frequency ratio is a rational number, say c/n, =1/5 where r and s are integers, then after § passages ‘through the surface of section the angle variable ¢, increases by 2ar and the system is back at its star. We have a finite cycle of s points. For the two-dimensional oscillator these features do not depend on the start. Exercise 10.02 A particle of mass m moves in a (two-dimensional) central force with potential a 3 Using a computer or otherwise, plot the sequences of points (x,p,) in the surface of section (y = 0, py > 0. E fixed) which result from representative stats (refer to Exercises 1.13 and 9.06).Exercise 10.02 223 Solution ‘According to Exercise 1.13 the equation of the orbit is 2 = 1 +ecosa(0~) where a= fi+2mi/I?. Here r and @ are polar coordinates in the plane of the orbit with 8p being the direction ofa pricenter. The parameters a and e ae the “semi-major- tis" and “eccentricity” They are related tothe energy E and angular momentum L by Bai? nae a as and en ie ERE = 1 -SE Solving for p,, we find mk ot? fmk [_,, 2a a?d-e%) nfm EE [SS ‘and substituting forr from the orbit equation, we find ar where we take the upper sign if L is positive and the lower sign if L is negative. The ‘conjugste momentum p, is the linear momentum in the r-direction. The momentum ‘conjugate to @ is the angular momentum L. The linear momentum in the 0-direction is Us. ‘Our surface of section is in terms of cartesian variables. ‘These variables ‘are obtained from the polar variables by the transformation equations xereos®, py =p, c0s0-(L/s)sind, yersind, py =p,sin8 + (L/r)eos8, ‘The surface of section is defined by yO with py>0. ‘These conditions imply that on the surface of section24 (Chapter X: Non-integrable Systems O=0,2n,4m for L>0 or O==n,~3n,-5n,-~ for L<0. Carson oe re of ton een = ts =a) wi pnw le ‘oper signif Lis postive andthe lower sign if Lis negative” Sequences of points on the sue of section be given by . et omen a-e2) mk ec SAEED. ssecoszaan), —py(n)= 2). . Tap 7s ecos2aan), pylon ay rindnan, where n= 0,1,2,-- and we have taken the stat, for either sign of L, at pericenter. That is, we have taken @5 =°0 for L positive and @y = ~n for L negative. The energy E and hence the "semi-major-axis" a on a given surface of section is constant. We can thus measure x in units of a and py in units of /mi/a , and the sequences of points become term in the potential (ce Exercise 1.13). We pick an eccentricity e and let n= 0,1,2.~: to obtain one Sequence of points, change e and obtain another sequence, etc. The result is Fig. 1 in which we have ken 9) = 0.05. Cu oes ne Os: ' €=0.0,0.2,04,0.608 Ex. 10.02, Fig. 1 Eccentricity ¢=0 gives the elliptic fixed points (21,0), and eccentricities ©=02,04,0.6, and 0.8 give the suecessive surrounding “curves.” The x>0 set is forExercise 10.03 28 L>0, and the x <0 setis for L <0. Note that the outermost "curve" is almost a finite ‘oycle of 17 points, so this is almost a periodic orbit. The reason is that (for these values, of 1 and e) 17a. = 20.003845 which is almost an integer (7 = 0.04995 and e= 0.8 give = 20/17 exactly). Exercise 10,03 (Consider a system with Hamiltonian 2 as higtetige " 2m por 3 ‘This is, of course, a simple harmonic oscillator with spring constant kp +1, and is exactly soluble. Suppose, however, that we regard the term }k,x” as a perturbation. Find, to first order inthe (@) the canonical transformation from the unperturbed action-angle variables to the perturbed sction-angle variables; (Ans: the generator is G = (kyl /4kg)sin24) (©) the Hamiltonian forthe perturbed action-angle variables. Hence find the first order Conrection tothe frequency of oscillation. Compare with the exact result. Solution (@) We write the Hamiltonian as HoH +H were ee Ho 2m 2ke isthe unperturbed Hamiltonian and 1 Hy = Fen? isthe perturbation, The unperturbed system is a simple harmonic oscillator for which we ‘can introduce sction-angle variables (6,1) by setting x= [Eisns, p= ae,26 (Chapter X: Non-integrable Systems where 0g = {f/m is the unperturbed frequency. In terms of these action-angle variables the unperturbed Hamiltonian is Ho) = wot and the perturbation is vite a, ‘We wish to introduce new action-angle variables (@,1') such thatthe new Hamiltonian is 1 function of the new action variable I’ slone, H'(I'). To find the generator of the appropriate canonical transformation, we expand the perturbation, sqn= Boe -038), ‘0 the Fourier components of Hy are bows, me, : Fxg Tmo, a, The foe components of te aproprinte generator ae then (Lagrangian and Hantonian chants, page aaa) "NE late s0 the generator is - Hl G@.D. By ‘The canonical transformation generated by G is By, ong Hse, (©) The new Hamiltonian is to first order inthe perturbation, HQ) = Ho) + (H@.1) ‘where (H;) isthe average of the perturbation over a cycle, For this we can either take hy from part (8) or proceed directly,Exercise 1004 2 (Hcg) a) cB. — H()-09' + EL, ‘Another way to find the new Hamiltonian is simply to transform the old Hamiltonian, ‘Since from part (a) we know the appropriate canonical transformation. This gives (a) ey fa H¢.1) ~0,(t'+ BE cosag) « BE sng’ (1+ 3) ‘in agreement withthe preceding approach. "The perturbed frequency of oscillation is hich canbe compared withthe exact elt fear % Ps wn JE wf Be nuglt+ thes PB elie oleae) Exercise 10,04 Find the continued fraction expansion of the following numbers, and write down the first five or so continued fraction approximates. Verify that these are closer to the number than (denominator), (a) 157/225; (b) ¥/2; (c) the golden ratio, y = (5 ~1)/2; (4) the base of natural logarithms, Solution (@ We have the continued fraction expansion 157 1 wt 1+——;— —- 34 ate 5208 (Chapter X: Non-integrable Systems ‘and continued fraction approximations 1 1 ut 2 ‘These differ from 157/225 by Bes Be bo} pt tt teak, [225 Te 25 *3~ 9° |as~ ial” 430 “T0* ~ 100 [157 _ 30 1 1 ns fast sars

También podría gustarte

• Mecánica Teórica: Fundamentos y Aplicaciones

  

  Aún no hay calificaciones

  Mecánica Teórica: Fundamentos y Aplicaciones

  380 páginas
• Introducción a la Física Newtoniana

  

  Aún no hay calificaciones

  Introducción a la Física Newtoniana

  198 páginas
• 2da Guia Resueltos Mecanica Analitica

  

  Aún no hay calificaciones

  2da Guia Resueltos Mecanica Analitica

  23 páginas
• (Calculo) Puig Adam - Ecuaciones Diferenciales

  

  67% (6)

  (Calculo) Puig Adam - Ecuaciones Diferenciales

  440 páginas
• Curso Completo de Mecánica Universitaria

  

  Aún no hay calificaciones

  Curso Completo de Mecánica Universitaria

  152 páginas
• Metodos Matematicos Avanzados para Cientificos e Ingenieros

  

  Aún no hay calificaciones

  Metodos Matematicos Avanzados para Cientificos e Ingenieros

  504 páginas
• 101 Problemas de Mecanica Teorica Nuevo Libro de La UGR

  

  Aún no hay calificaciones

  101 Problemas de Mecanica Teorica Nuevo Libro de La UGR

  3 páginas
• Guía de Termodinámica para Ingenieros

  

  Aún no hay calificaciones

  Guía de Termodinámica para Ingenieros

  188 páginas
• Problemas y Soluciones de Mecánica FI2001

  

  Aún no hay calificaciones

  Problemas y Soluciones de Mecánica FI2001

  96 páginas
• Fisica General Unad PDF

  

  100% (1)

  Fisica General Unad PDF

  394 páginas
• Fisica Chiappero

  

  100% (1)

  Fisica Chiappero

  178 páginas
• Mecanica Newtoniana PDF

  

  Aún no hay calificaciones

  Mecanica Newtoniana PDF

  65 páginas
• Mecánica para Estudiantes de Física

  

  100% (1)

  Mecánica para Estudiantes de Física

  209 páginas
• Sistemas Hamiltonianos - Variables Acción Ángulo PDF

  

  Aún no hay calificaciones

  Sistemas Hamiltonianos - Variables Acción Ángulo PDF

  33 páginas
• Rey Pastor - Analisis Matematico Vol 3

  

  Aún no hay calificaciones

  Rey Pastor - Analisis Matematico Vol 3

  367 páginas
• 100 Problemas de Fisica Estadistica Carlos Fernandez Tejero PDF

  

  Aún no hay calificaciones

  100 Problemas de Fisica Estadistica Carlos Fernandez Tejero PDF

  268 páginas
• Educación Abierta en Física y Estadística

  

  Aún no hay calificaciones

  Educación Abierta en Física y Estadística

  442 páginas
• Leyes de Newton y Dinámica Fundamental

  

  Aún no hay calificaciones

  Leyes de Newton y Dinámica Fundamental

  35 páginas
• Exámenes Predoctorales de Electrodinámica

  

  100% (1)

  Exámenes Predoctorales de Electrodinámica

  247 páginas
• Berkeley Physics Course Tomo 2 - Electricidad y Magnetismo (PDFDrive)

  

  Aún no hay calificaciones

  Berkeley Physics Course Tomo 2 - Electricidad y Magnetismo (PDFDrive)

  492 páginas
• Transformada de Laplace: Teoría y Aplicaciones

  

  Aún no hay calificaciones

  Transformada de Laplace: Teoría y Aplicaciones

  22 páginas
• Apunte de Teoría Física

  

  100% (1)

  Apunte de Teoría Física

  110 páginas
• Manual de Prácticas de Física Sólida

  

  Aún no hay calificaciones

  Manual de Prácticas de Física Sólida

  100 páginas
• Problemas de Mecánica Clásica

  

  Aún no hay calificaciones

  Problemas de Mecánica Clásica

  10 páginas
• Mecánica Clásica: Fundamentos y Ejercicios

  

  Aún no hay calificaciones

  Mecánica Clásica: Fundamentos y Ejercicios

  423 páginas
• Meccel PDF

  

  100% (1)

  Meccel PDF

  321 páginas
• Mecánica Clásica: Ecuaciones y Teoremas

  

  Aún no hay calificaciones

  Mecánica Clásica: Ecuaciones y Teoremas

  312 páginas
• Ecuaciones Diferenciales y Laplace

  

  100% (4)

  Ecuaciones Diferenciales y Laplace

  110 páginas
• Problemas de Mecánica - RAEC

  

  Aún no hay calificaciones

  Problemas de Mecánica - RAEC

  252 páginas
• Fundamentos de Mecánica Clásica

  

  Aún no hay calificaciones

  Fundamentos de Mecánica Clásica

  266 páginas
• Mecanica - Clasica. Aguero - Garcia.uaem

  

  Aún no hay calificaciones

  Mecanica - Clasica. Aguero - Garcia.uaem

  203 páginas
• Termodinamica

  

  Aún no hay calificaciones

  Termodinamica

  7 páginas
• Introducción a la Mecánica Física

  

  100% (1)

  Introducción a la Mecánica Física

  234 páginas
• Física II: Campos y Ondas

  

  100% (1)

  Física II: Campos y Ondas

  258 páginas
• El universo y su composición cuántica

  

  Aún no hay calificaciones

  El universo y su composición cuántica

  4 páginas
• V. López Rodríguez, Problemas Resueltos de Electromagnetismo

  

  100% (3)

  V. López Rodríguez, Problemas Resueltos de Electromagnetismo

  772 páginas
• Notas de Clase sobre Electromagnetismo

  

  Aún no hay calificaciones

  Notas de Clase sobre Electromagnetismo

  67 páginas
• Elementos de Física Matemática - R. Götze - Tomo 2

  

  100% (2)

  Elementos de Física Matemática - R. Götze - Tomo 2

  202 páginas
• Física II. Dinámica Fluidos, Astronomía Astrofísica. Edit Santillana Polimodal

  

  Aún no hay calificaciones

  Física II. Dinámica Fluidos, Astronomía Astrofísica. Edit Santillana Polimodal

  192 páginas
• Apuntes FI2001 2019 Mecánica - Patricio Cordero PDF

  

  Aún no hay calificaciones

  Apuntes FI2001 2019 Mecánica - Patricio Cordero PDF

  277 páginas
• Ondas Electromagnéticas y Maxwell

  

  Aún no hay calificaciones

  Ondas Electromagnéticas y Maxwell

  66 páginas
• Mecánica Clásica: Lagrange y Acción

  

  Aún no hay calificaciones

  Mecánica Clásica: Lagrange y Acción

  319 páginas
• Mecánica Clásica: Lagrange y Acción

  

  Aún no hay calificaciones

  Mecánica Clásica: Lagrange y Acción

  38 páginas
• Introducción a la Mecánica Clásica

  

  Aún no hay calificaciones

  Introducción a la Mecánica Clásica

  29 páginas
• Ilovepdf Merged 1 40

  

  Aún no hay calificaciones

  Ilovepdf Merged 1 40

  40 páginas
• Introducción a la Mecánica Lagrangiana

  

  Aún no hay calificaciones

  Introducción a la Mecánica Lagrangiana

  5 páginas
• Problemas de Mecánica Lagrangiana y Hamiltoniana

  

  Aún no hay calificaciones

  Problemas de Mecánica Lagrangiana y Hamiltoniana

  8 páginas
• Notas MTII

  

  Aún no hay calificaciones

  Notas MTII

  163 páginas
• Ecuaciones de Movimiento en Mecánica Lagrangiana

  

  Aún no hay calificaciones

  Ecuaciones de Movimiento en Mecánica Lagrangiana

  2 páginas
• Ecuaciones de Movimiento en Mecánica Lagrangiana

  

  Aún no hay calificaciones

  Ecuaciones de Movimiento en Mecánica Lagrangiana

  2 páginas
• Oscilaciones Acopladas y Lagrangiana

  

  Aún no hay calificaciones

  Oscilaciones Acopladas y Lagrangiana

  62 páginas
• Manual de Mecánica Clásica

  

  Aún no hay calificaciones

  Manual de Mecánica Clásica

  170 páginas
• Principios Variacionales en Mecánica Clásica

  

  Aún no hay calificaciones

  Principios Variacionales en Mecánica Clásica

  825 páginas
• Introducción al Formalismo Hamiltoniano

  

  Aún no hay calificaciones

  Introducción al Formalismo Hamiltoniano

  117 páginas
• Notas Analitica

  

  Aún no hay calificaciones

  Notas Analitica

  478 páginas
• Mec Nica Anal°tica PP

  

  100% (10)

  Mec Nica Anal°tica PP

  425 páginas
• Fundamentos de Física III - Curso 2024-2025

  

  Aún no hay calificaciones

  Fundamentos de Física III - Curso 2024-2025

  19 páginas
• Temario Completo Atrio 2022

  

  100% (1)

  Temario Completo Atrio 2022

  182 páginas
• Apuntes

  

  Aún no hay calificaciones

  Apuntes

  227 páginas
• 100 Problemas de Mecánica - Pérez García

  

  Aún no hay calificaciones

  100 Problemas de Mecánica - Pérez García

  272 páginas
• Termo Est Parte II PDF

  

  Aún no hay calificaciones

  Termo Est Parte II PDF

  44 páginas
• Constancia Movimientos

  

  100% (1)

  Constancia Movimientos

  3 páginas
• Silabo de Física Estadística - Maestría

  

  Aún no hay calificaciones

  Silabo de Física Estadística - Maestría

  5 páginas
• Metodología de Investigación Científica

  

  Aún no hay calificaciones

  Metodología de Investigación Científica

  5 páginas
• Práctica de Mecánica para Ingeniería

  

  Aún no hay calificaciones

  Práctica de Mecánica para Ingeniería

  1 página
• Cálculo de Fuerzas en Estructuras Físicas

  

  0% (1)

  Cálculo de Fuerzas en Estructuras Físicas

  3 páginas
• Simulación y Visualización

  

  Aún no hay calificaciones

  Simulación y Visualización

  4 páginas