8.02SC Supplementary Problem Solutions: Vector Integral Calculus in Space

6. Vector Integral Calculus in Space 6A. Vector Fields in Space 6A-1 a) the vectors are all unit vectors, pointing radially outward. b) the vector at P has its head on the y-axis, and is perpendicular to it 6A-2 1 2 (−x i − yj − zk) 6A-3 ω(−z j + y k ) 6A-4 A vector field F = M i + N j + P k is parallel to the plane 3x − 4y + z = 2 if it is perpendicular to the normal vector to the plane, 3 i − 4 j + k : the condition on M, N, P therefore is 3M − 4N + P = 0, or P = 4N − 3M . The most general such field is therefore F = M i + N j + (4N − 3M ) k , where M and N are functions of x, y, z. 6B. Surface Integrals and Flux xi + yj + zk ; therefore F · n = a. We have n = a �� Flux through S = F · n dS = a(area of S) = 4π a3 . 6B-1 S 6B-2 Since k is parallel to the surface, the field is everywhere tangent to the cylinder, hence the flux is 0. 1 n 1 i+j+k √ is a normal vector to the plane, so F · n = √ . 6B-3 3 3 √ √ √ 1 1 3 2) area of region 1 2 (base)(height) 2 ( 2)( 2 √ √ √ Therefore, flux = = = = . 2 3 3 3 xi + yj + zk y2 ; F · n = . Calculating in spherical coordinates, a � π� π a � π� π 1 y2 3 4 sin3 φ sin2 θ dφdθ. a sin φ sin2 θ dφ dθ = a3 dS = flux = a a 0 0 0 0 S �π Inner integral: sin2 θ(− cos φ + 31 cos3 φ) = 43 sin2 θ; 6B-4 n= �� 0 Outer integral: 4 3 1 3a (2θ − 1 4 sin 2θ) �π 0 = 2 3 3 πa . 1 2 S 1 1 2 E. 18.02 EXERCISES i+j+k z √ ; F·n= √ . 3 3 �� � 1 � 1−y �� 1 dx dy z dx dy √ (1 − x − y) dx dy. (1 − x − y) √ = =√ flux = 3 |n · k | 3 1/ 3 0 0 S S �1−y Inner integral: = x − 12 x2 − xy = 12 (1 − y)2 . 6B-5 n= 1 0 Outer integral: = 6B-6 � 1 0 1 1 1 (1 − y)2 dy = · − · (1 − y)3 2 2 3 z = f (x, y) = x2 + y 2 (a paraboloid). �1 = 0 n 1 . 6 S 1 1 By (13) in Notes V9, x+y=1 dS = (−2x i − 2y j + k ) dx dy. (This points generally “up”, since the k component is positive.) Since F = x i + y j + z k , �� S F · dS = �� R (−2x2 − 2y 2 + z) dxdy , where R is the interior of the unit circle in the xy-plane, i.e., the projection of S onto the xy-plane). Since z = x2 + y 2 , the above integral �� � 2π � 1 1 π 2 2 =− r2 · r dr dθ = −2π · = − . (x + y ) dxdy = − 4 2 R 0 0 The answer is negative since the positive direction for flux is that of n, which here points into the inside of the paraboloidal cup, whereas the flow x i + y j + z k is generally from the inside toward the outside of the cup, i.e., in the opposite direction. xi + yj y2 , F·n= . a a In cylindrical coordinates, since y = a sin θ, this gives us F · dS = F · n dS = a2 sin2 θ dz dθ. � �π/2 � π/2 � π/2 � k θ sin 2θ π sin2 θ dθ = a2 h a2 sin2 θ dz dθ = a2 h − Flux = = a2 h . 2 4 2 −π/2 −π/2 0 −π/2 6B-8 6B-12 On the cylindrical surface, n = Since the distance from a point (x, y, 0) up to the hemispherical surface is z, �� z dS average distance = ��S . dS S �� � 2π � π/2 a cos φ · a2 sin φ dφ dθ. In spherical coordinates, z dS = S Inner: = a3 Finally, �� � π/2 sin φ cos φ dφ = a3 ( 0 0 0 sin2 φ 2 �π/2 0 = a3 . 2 Outer: = a3 2 dS = area of hemisphere = 2πa2 , so average distance = S � 2π dθ = πa3 . 0 a πa3 = . 2πa2 2 6. VECTOR INTEGRAL CALCULUS IN SPACE 3 6C. Divergence Theorem 6C-1a 6C-2 div F = Mx + Ny + Pz = 2xy + x + x = 2x(y + 1). Using the product and chain rules for the first, symmetry for the others, (ρn x)x = nρn−1 x x + ρn , ρ adding these three, we get (ρn y)y = nρn−1 div F = nρn−1 Therefore, div F = 0 ⇔ n = −3. y y + ρn , ρ (ρn z)z = nρn−1 z z + ρn ; ρ x2 + y 2 + z 2 + 3ρn = ρn (n + 3). ρ 6C-3 Evaluating the triple integral first, we have div F = 3, therefore ��� 2 div F dV = 3(vol.of D) = 3 πa3 = 2πa3 . 3 D To evaluate the double integral over the closed surface S1 + S2 , the normal vectors are: xi + yj + zk (hemisphere S1 ), n2 = − k (disc S2 ); n1 = a using these, the surface integral for the flux through S is �� �� �� �� x2 + y 2 + z 2 dS + −z dS = a dS, F · dS = a S2 S1 S S1 since x2 + y 2 + z 2 = ρ2 = a2 on S1 , and z = 0 on S2 . So the value of the surface integral is a(area of S1 )= a(2πa2 ) = 2πa3 , 1 which agrees with the triple integral above. �� ��� 6C-5 The divergence theorem says F · dS = div F dV. S D Here div F = 1, so that the right-hand integral is just the volume of the tetrahedron, which is 31 (base)(height)= 13 ( 21 )(1) = 16 . 6C-6 The divergence theorem says �� S F · dS = ��� 1 1 div F dV. D Here div F = 1, so the right-hand integral is the volume of the solid cone, which has height 1 and base radius 1; its volume is 31 (base)(height)= π/3. 6C-7a Evaluating the triple integral first, over the cylindrical solid D, we have ��� div F = 2x + x = 3x; 3x dV = 0, D since the solid is symmetric with respect to the yz-plane. (Physically, assuming the density is 1, the integral has the value x̄(mass of D), where x̄ is the x-coordinate of the center of mass; this must be in the yz plane since the solid is symmetric with respect to this plane.) To evaluate the double integral, note that F has no k -component, so there is no flux across the two disc-like ends of the solid. To find the flux across the cylindrical side, n = xi + yj, F · n = x3 + xy 2 = x3 + x(1 − x2 ) = x, 4 E. 18.02 EXERCISES since the cylinder has radius 1 and equation x2 + y 2 = 1. Thus � 2π �� � 2π � 1 cos θ dθ = 0. cos θ dz dθ = x dS = 0 S 0 0 6C-8 a) Reorient the lower hemisphere S2 by reversing its normal vector; call the reori­ ented surface S2′ . Then S = S1 + S2′ is a closed surface, with the normal vector pointing outward everywhere, so by the divergence theorem, �� �� �� ��� S1 F · dS = F · dS + F · dS = div F dV = 0, S S1 S2′ D since by hypothesis div F = 0. The above shows �� �� �� F · dS = − F · dS = S1 S2′ S2 S2 F · dS, since reversing the orientation of a surface changes the sign of the flux through it. b) The same statement holds if S1 and S2 are two oriented surfaces having the same boundary curve, but not intersecting anywhere else, and oriented so that S1 and S2′ (i.e., S2 with its orientation reversed) together make up a closed surface S with outward-pointing normal. 6C-10 If div F = 0, then for any closed surface S, we have by the divergence theorem �� ��� F · dS = div F dV = 0. S Conversely: �� S D F · dS = 0 for every closed surface S ⇒ div F = 0. For suppose there were a point P0 at which (div F)0 6= 0 — say (div F)0 > 0. Then by continuity, div F > 0 in a very small spherical ball D surrounding P0 , so that by the divergence theorem (S is the surface of the ball D), �� ��� F · dS = div F dV > 0. S D But this contradicts our hypothesis that 6C-11 flux of F = �� S F · dn = ��� �� S F · dS = 0 for every closed surface S. div F dV = D ��� 3 dV = 3(vol. of D). D 6D. Line Integrals in Space 6D-1 a) C : x = t, dx = dt; y = t2 , dy = 2t dt; z = t3 , dz = 3t2 dt; � � 1 (t2 )dt + t3 (2t dt) − t(3t2 dt) y dx + z dy − x dz = C 0 �1 � 1 1 2 3 1 t3 2t5 3t4 2 + − = + − = − . = (t + 2t4 − 3t3 )dt = 3 5 4 0 3 5 4 60 0 � � 1 1 b) C : x = t, y = t, z = t; y dx + z dy − x dz = t dt = . 2 C 0 6. VECTOR INTEGRAL CALCULUS IN SPACE 5 c) C = C1 + C2 + C3 ; C1 : y = z = 0; C2 : x = 1, z = 0; C3 : x = 1, y = 1 � � � � 1 y dx + z dy − x dz = 0+ −dz = −1. 0+ C C1 C2 0 � d) C : x = cos t, y = sin t, z = t; zx dx + zy dy + x dz C � 2π � t cos t(− sin t dt) + t sin t(cos t dt) + cos t dt = = 0 2π cos t dt = 0. 0 6D-2 The field F is always pointed radially outward; if C lies on a sphere centered at the origin, its unit tangent t is always tangent to the sphere, therefore � � perpendicular to the radius; this means F · t = 0 at every point of C. Thus C F · dr = C F · t ds = 0. 6D-4 a) F = ∇f = 2x i + 2y j + 2z k . b) (i) Directly, letting C be the helix: x = cos t, y = sin t, z = t, from t = 0 to t = 2nπ, � � 2nπ � 2nπ 2t dt = (2nπ)2 . M dx + N dy + P dz = 2 cos t(− sin t)dt + 2 sin t(cos t)dt + 2t dt = 0 C 0 b) (ii) Choose the vertical path x = 1, y = 0, z = t; then � 2nπ � 2t dt = (2nπ)2 . M dx + N dy + P dz = C 0 b) (iii) By the First Fundamental Theorem for line integrals, � F · dr = f (1, 0, 2nπ) − f (1, 0, 0) = 912 + (2nπ)2 ) − 12 = (2nπ)2 C 6D-5 By the First Fundamental Theorem for line integrals, � � � � � F · dr = sin(xyz)� − sin(xyz)� , C Q P where C is any path joining P to Q. The maximum value of this difference is 1 − (−1) = 2, since sin(xyz) ranges between −1 and 1. For example, any path � C connecting P : (1, 1, −π/2) to Q : (1, 1, π/2) will give this maximum value of 2 for C F · dr. 6E. Gradient Fields in Space 6E-1 a) Since M = x2 , N = y 2 , P = z 2 are continuously differentiable, the differential is exact because Nz = Py = 0, Mz = Px = 0, My = Nx = 0; f (x, y, z) = (x3 + y 3 + z 3 )/3. b) Exact: M, N, P are continuously differentiable for all x, y, z, and Nz = Py = 2xy, Mz = P x = y 2 , My = Nx = 2yz; f (x, y, z) = xy 2 . c) Exact: M, N, P are continuously differentiable for all x, y, z, and Nz = Py = x, Mz = Px = y, My = Nx = 6x2 + z; f (x, y, z) = 2x3 y + xyz. � � � i j k �� � 6E-2 curl F = �� ∂x ∂y ∂z �� = (xz 2 − y) i − yz 2 j − x2 k . � x2 y yz xyz 2 � 6 E. 18.02 EXERCISES 6E-3 a) It is easily checked that curl F = 0. b) (i) using method I: f (x1 , y1 , z1 ) = = Therefore � � (x1 ,y1 ,z1 ) F · dr = (0,0,0) x1 x dx + 0 f (x, y, z) = � � C1 y1 y dy + 0 1 2 2 (x F · dr + � 2 � z1 z dz = 0 C2 F · dr + � C3 (x 1, y1, z1) F · dr 1 2 1 2 1 2 x + y + z . 2 1 2 1 2 2 C3 C1 x1 C2 2 + y + z ) + c. (ii) Using method II: We seek f (x, y, z) such that fx = 2xy + z, fy = x2 , fz = x. fx = 2xy + z ⇒ fy = x2 + g y = x2 fz = x + h′ (z) = x f = x2 y + xz + g(y, z). ⇒ gy = 0 ⇒ g = h(z) ⇒ h′ = 0 ⇒ h = c Therefore f (x, y, z) = x2 y + xz + c. (iii) If fx = yz, fy = xz, fz = xy, then by inspection, f (x, y, z) = xyz + c. 6E-4 Let F = f − g. Since ∇ is a linear operator, We now show: ∇F = ∇f − ∇g = 0 ∇F = 0 ⇒ F = c. Fix a point P0 : (x0 , y0 , z0 ). Then by the Fundamental Theorem for line integrals, F (P ) − F (P0 ) = � P P0 ∇F · dr = 0. Therefore F (P ) = F (P0 ) for all P , i.e., F (x, y, z) = F (x0 , y0 , z0 ) = c. 6E-5 F is a gradient field only if these equations are satisfied: Nz = Py : 2xz + ay = bxz + 2y Mz = Px : 2yz = byz M y = Nx : z 2 = z 2 . Thus the conditions are: a = 2, b = 2. Using these values of a and b we employ Method 2 to find the potential function f : fx = yz 2 ⇒ f = xyz 2 + g(y, z); fy = xz 2 + gy = xz 2 + 2yz ⇒ gy = 2yz ⇒ g = y 2 z + h(z) fz = 2xyz + y 2 + h′ (z) = 2xyz + y 2 ⇒ h = c; therefore, f (x, y, z) = xyz 2 + y 2 z + c. 6E-6 a) M dx + N dy + P dz is an exact differential if there exists some function f (x, y, z) for which df = M dx + N dy + P dz; that, is, for which fx = M, fy = N, fz = P . b) The given differential is exact if the following equations are satisfied: Nz = Py : (a/2)x2 + 6xy 2 z + 3byz 2 = 3x2 + 3cxy 2 z + 12yz 2 ; Mz = Px : axy + 2y 3 z = 6xy + cy 3 z My = Nx : axz + 3y 2 z 2 = axz + 3y 2 z 2 . Solving these, we find that the differential is exact if a = 6, b = 4, c = 2. (x1, y1 ) 6. VECTOR INTEGRAL CALCULUS IN SPACE 7 c) We find f (x, y, z) using method 2: fx = 6xyz + y 3 z 2 ⇒ f = 3x2 yz + xy 3 z 2 + g(y, z); fy = 3x2 z + 3xy 2 z 2 + gy = 3x2 z + 3xy 2 z 2 + 4yz 3 ⇒ gy = 4yz 3 ⇒ g = 2y 2 z 3 + h(z) fz = 3x2 y + 2xy 3 z + 6y 2 z 2 + h′ (z) = 3x2 y + 2xy 3 z + 6y 2 z 2 ⇒ h′ (z) = 0 ⇒ h = c. Therefore, f (x, y, z) = 3x2 yz + xy 3 z 2 + 2y 2 z 3 + c. n 6F. Stokes’ Theorem 6F-1 � a) For the line integral, since the differential is exact. � � i � ∇×F = �� ∂x �x For the surface integral, b) Line integral: � C F · dr = ydx + zdy + xdz = C Using x = cos t, y = sin t, � xdx + ydy + zdz = 0, C j ∂y y � S C � �� k �� ∂z �� = 0, and therefore ∇×F·dS = 0. S z � ydx, since z = 0 and dz = 0 on C. C � 2π 0 2 − sin t dt = − � 2π 0 � � i � Surface integral: curl F = �� ∂x � y �� S 1 − cos 2t dt = −π. 2 � j k �� ∂y ∂z �� = − i − j − k ; n = xi + yj + zk z x� �� ∇ × F) · n dS = − (x + y + z) dS. S To evaluate, we use x = r cos θ, y = r sin θ, z = ρ cos φ. r = ρ sin φ, dS = ρ2 sin φ dφdθ; note that ρ = 1 on S. The integral then becomes � 2π � π/2 [sin φ(cos θ + sin θ) + cos φ] sin φ dφ dθ − 0 0 � 1 φ sin 2φ ) + sin2 φ Inner: − (cos θ + sin θ)( − 2 4 2 � � 2π � 1 π − − (cos θ + sin θ) Outer: dθ = −π. 2 4 0 �π/2 0 � π 1 ; = − (cos θ + sin θ) + 4 2 � 6F-2 The surface S is: z = −x − y, so that f (x, y) = −x − y. n dS = h−fx , −fy , 1i dx dy = h1, 1, 1i dx dy. (Note the signs: n points upwards, and therefore should have a positive k-component.) � � � i j k �� � curl F = �� ∂x ∂y ∂z �� = − i − j − k � y z x� �� �� Therefore curl F · n dS = − 3 dA = −3π, where S ′ is the projection of S, i.e., S S′ the interior of the unit circle in the xy-plane. As for the line integral, we have C : x = cos t, y = sin t z = − cos t − sin t, so that 8 E. 18.02 EXERCISES � 2π � � − sin2 t − (cos2 t + sin t cos t) + cos t(sin t − cos t) dt ydx + zdy + xdz = C 0 � 2π � � 2π � 1 3 2 2 2 = (− sin t − cos t − cos t) dt = −1 − (1 + cos 2t) dt = − · 2π = −3π. 2 2 0 0 � 6F-3 Line integral: yz dx + xz dy + xy dz over the path C = C1 + . . . + C4 : � C � = 0, since z = dz = 0 on C1 ; � 1 � = 1 · 1 dz = 1, since x = 1, y = 1, dx = 0, dy = 0 on C2 ; 0 � C2 � 0 ydx + xdy = xdx + xdx = −1, since y = x, z = 1, dz = 0 on C3 ; 1 � C3 = 0, since x = 0, y = 0 on C4 . C1 C 3 (1,1,1) C 4 C 2 C 1 (1,1,0) C4 Adding up, we get � � i � curl F = �� ∂x � yz j ∂y xz � C F · dr = � + C1 � + C2 � + C3 � = 0. For the surface integral, C4 � �� k �� curl F · dS = 0. ∂z �� = i (x − x) − j (y − y) + k (z − z) = 0; thus xy � Let S1 be the top of the cylinder (oriented so n = k ), and S2 the side. � � � i j k �� � a) We have curl F = �� ∂x ∂y ∂z �� = −2x j + 2 k . h � −y x x2 � �� �� For the top: curl F · n dS = 2 dS = 2(area of S1 ) = 2πa2 . 6F-5 S1 S1 S2 C S1 xi + yj For the side: we have n = , and dS = dz · a dθ, so that a �2π �� � 2π � 2π � h −2xy a dz dθ = −2h(a cos θ)(a sin θ) dθ = −ha2 sin2 θ = 0. curl F · n dS = a S2 0 0 0 0 �� �� �� Adding, curl F · dS = + = 2πa2 . S S1 S2 b) Let C be the circular boundary of S, parameterized by x = a cos θ, y = a sin θ, z = 0. Then using Stokes’ theorem, �� � � 2π curl F · dS = −y dx + x dy + x2 dz = (a2 sin2 θ + a2 cos2 θ) dθ = 2πa2 . S C 0 6G. Topological Questions 6G-1 a) yes b) no 6G-2 Recall that ρx = x/ρ, etc. Then, using the chain rule, curl F = (nρn−1 z c) yes d) no; yes; no; yes; no; yes z x z x y y − nρn−1 y ) i + (nρn−1 z − nρn−1 x ) j + (nρn−1 y − nρn−1 x ) k . ρ ρ ρ ρ ρ ρ 6. VECTOR INTEGRAL CALCULUS IN SPACE 9 Therefore curl F = 0. To find the potential function, we let P0 be any convenient starting point, and integrate along some path to P1 : (x1 , y1 , z1 ). Then, if n = 6 −2, we have � � P1 � P1 1 F · dr = ρn (x dx + y dy + z dz) = ρn d(ρ2 ) 2 P0 C P � � P10 n+2 n+2 n+2 P1 ρ ρn+2 ρ ρ − 0 = 1 + c, since P0 is fixed. = ρn+1 dρ = = 1 n + 2 P0 n+2 n+2 n+2 P0 ρn+2 Therefore, we get F = ∇ , if n = 6 −2. n+2 � P1 dρ If n = −2, the line integral becomes = ln ρ1 + c, so that F = ∇(ln ρ). ρ P0 6H. Applications and Further Exercises 6H-1 Let F = M i + N j + P k . By the definition of curl F, we have ∇ × F = (Py − Nz ) i + (Mz − Px ) j + (Nx − My ) k , ∇ · (∇ × F) = (Pyx − Nzx ) + (Mzy − Pxy ) + (Nxz − Myz ) If all the mixed partials exist and are continuous, then Pxy = Pyx , etc. and the right-hand side of the above equation is zero: div (curl F) = 0. 6H-2 a) Using the divergence theorem, and the previous problem, (D is the interior of S), �� ��� ��� curl F · dS = div curl F dV = 0 dV = 0. S D D b) Draw a closed curve C on S that divides it into two pieces S1 and S2 both having C as boundary. Orient C compatibly with S1 , then the curve C ′ obtained by reversing the orientation of C will be oriented compatibly with S2 . Using Stokes’ theorem, �� �� �� � � curl F · dS = curl F · dS + curl F · dS = F · dr + F · dr = 0, S S1 S2 C C′ since the integral on C ′ is the negative of the integral on C. Or more simply, consider the limiting case where C has been shrunk to a point; even as a point, it can still be considered to be the boundary of S. Since it has zero length, the line integral around it is zero, and therefore Stokes’ theorem gives �� � curl F · dS = F · dr = 0. S C 6H-10 Let C be an oriented closed curve, and S a compatibly-oriented surface having C as its boundary. Using Stokes’ theorem and the Maxwell equation, we get respectively �� � �� �� �� 1 d 1 ∂E ∇×B·dS = ·dS = B·dr and ∇×B·dS = E·dS. c dt S C S S S c ∂t � �� 1 d Since the two left sides are the same, we get B · dr = E · dS. c dt S C In words: for the magnetic field B produced by a moving electric field E(t), the magneto­ motive force around a closed loop C is, up to a constant factor depending on the units, the time-rate at which the electric flux through C is changing. S1 C C S2