2 Consequences of the Potential It may be no surprise that human minds can deduce the laws of falling objects because the brain has evolved to devise strategies for dodging them. (Paul Davies) Only mathematics and mathematical logic can say as little as the physicist means to say. (Bertrand Russell) In Chapter 1, we learned that a conservative vector field F can be ex- pressed as the gradient of a scalar @, called the potential of F, and conversely F is conservative if F = Vo. It was asserted that such po- tentials satisfy Laplace's equation at places free of all sources of F and are said to be harmonic. This led to several important characteristics of the potential. In the same spirit, this chapter investigates a number of additional consequences that follow from Laplace's equation. 2.1 Green’s Identities ‘Three identities can be derived from vector calculus and Laplace's equa- tion, and these lead to several important theorems and additional in- sight into the nature of potential fields. They are referred to as Green's identities.+ } ‘The name Green, appearing repeatedly in this and subsequent chapters, refers to George Green (1793-1841), a British mathematician of Caius College. Cambridge, England. He is perhaps best known for his paper, Essay on the Application of Mathematical Analysis lo the Theory of Electricity and Magnetism, and was ap- parently the first to use the term “potential.” 1920 Consequences of the Potential 2.1.1 Green’s First Identity Green's first identity is derived from the divergence theorem (Appendix A), Let U and V be continuous functions with continuous partial deriva- tives of first order throughout. a closed, regular region R, and let U have continuous partial derivatives of second order in R. The boundary of R is surface S, and fi is the outward normal to S. If A = VVU, then f{y Adv [V-(vWU) de R R = [wv VU + VVPU du & Using the divergence theorem yields fv vu +vvu\s= [ A-asy a 3 - [vow ads s Ou - fv nts [vvives fou-vvae= [ Rk Rk Equation 2.1 is Green's first identity and is true for all functions U and V that satisfy the differentiability requirements stated earlier Several very interesting theorems result from Green's first identity if U and V are restricted a bit further. For example, if U is harmonic and continuously differentiable in R, and if V = 1, then VU = 0, VV =0, and equation 2.1 becomes that is, (24) Thus the normal derivative of a harmonic function must average to zero on any closed boundary surrounding a region throughout which the func- tion is harmonic and continuously differentiable (Figure 2.1). It also can be shown (Kellogg (146, p. 227]) that. the converse of equation 2.2 is2.1 Green's Identities 21 Fig. 2.1. Region R subject to force field F. Surface S bounds region R. Unit vector fi is outward normal at any point on S. true; that is, if U and its derivatives of first order are continuous in R, and 34 integrates to zero over its closed boundary, then U must be har- monic throughout R. Hence, equation 2.2 is a necessary and sufficient condition for U to be harmonic throughout the region. Equation 2.2 provides an important boundary condition for many geo- physical problems. Suppose that vector field F has a potential J which is harmonic throughout some region. Because 24 = F - f on the surface of the region, equation 2.1 can be written as ca [¥-nas and applying the divergence theorem (Appendix A) yields [v-Pao Rr In words, the normal component of a conservative field must average to zero on the closed boundary of a region in which its potential is harmonic. Hence, the flux of F into the region exactly equals the flux leaving the region, implying that no sources of F exist. in the region. Moreover, the condition that V - F =0 throughout the region is sufficient to conclude that no sources lie within the region. (2.3) 0.2 Consequences of the Potential Steady-state heat flow, for example, is harmonic (as discussed in Chap- ter 1) in regions without heat sources or sinks and must satisfy eqta- tion 2.3. If region R is in thermal equilibrium and contains no heat sources or sinks, the heat entering R must equal the heat leaving R. Equation 2.3 is often called Gauss's law and will prove useful in subse- quent chapters. Now let U be harmonic in region R and let V = U. Then, from Green's first identity, four do= [ue ds. (24) R 5 Consider equation 2.4 when U = 0 on S. The right-hand side vanishes and, because (VU)? is continuous throughout R by hypothesis, (VU)? = 0. Therefore, 7 must be a constant. Moreover, because U = 0 on S and because U is continuous, the constant must be zero. Hence, if U 18 harmonic and continuously differentiable in R and if U vanishes at all points of S, U also must vanish at all points of R. This result. is intuitive from steady-state heat flow. If temperature is zero at all points of a region’s boundary and no sources or sinks are situated within the region, then clearly the temperature must vanish throughout the region once equilibrium is achieved. Green's first identity leads to a statement about uniqueness, some- times referred to as Stokes’s theorem. Let U, and U2 be harmonic in R and have identical boundary conditions, that is, U,(S) = U2(S)- The function Uj —Uz also must be harmonic in R. But Uy—U? vanishes on S and the previous theorem states that Uy —Up also must vanish at every point of R. Therefore, U; and Uz are identical. Consequently, a function that is harmonic and continuously differentiable in Ris uniquely deter- mined by its values on S, and the solution to the Dirichlet boundary- value problem is unique. Stokes's theorem makes intuitive sense when applied to steady-state heat flow. A region will eventually reach thermal equilibrium if heat is allowed to flow in and out of the region. It seems reasonable that, for any prescribed set of boundary temperatures, the region will always attain the same equilibrium temperature distribution throughout the region regardless of the initial temperature distribution. In other words, the steady-state temperature of the region is uniquely determined by the boundary temperatures.2.4 Green's Identities 23 The surface integral in equation 2.4 also vanishes if 84 = 0 on S. A similar proof could be developed to show that if U is single-valued, har- ‘monic, and continuously differentiable in R and if 22 = 0 on S, then U is a constant throughout R. Again, steady-state heat flow provides some insight. If the boundary of 2 is thermally insulated, equilibrium tem- peratures inside R must be uniform. Moreover, a single-valued harmonic function is determined throughout R, except for an additi the values of its normal derivatives on the boundary. constant, by Exercise 2.1 Prove the previous two theorems. These last theorems relate to the Neumann boundary-value problem and show that such solutions are unique to within an additive constant. The uniqueness of harmonic functions also extends to mixed boundary- value problems. Jf U is harmonic and continuously differentiable in R and if au +h = dn t= 9 on S$, where h and g are continuous functions of S, and h is never negative, then U is unique in R. Exercise 2.2 Prove the previous theorem. We have shown that under many conditions Laplace's equation has only one solution in a region, thus describing the tmiqueness of harmonic functions. But can we say that even that one solution always exists? The answer lo this interesting question requires a set of “existence theorems” for harmonic functions that are beyond the scope of this chapter. Inter ested readers are referred to Chapter XI of Kellogg [146, p. 277] for a comprehensive discussion. 2.1.2 Green's Second Identity If we interchange 7 and V in equation 2.1 and subtract the result from equation 2.1, we obtain Green’s second identity: : : av av oy vv" =f Wy! as, . fwvv-veujde= flu -ve) as, (2s) ht 3 where it is understood that U and V’ are continuously differentiable and have continuous partial derivatives of first and second order in R.wu Consequences of the Potential A corollary results if / and V are both harmonic: BV WU) yg _ fee ve] ds =0. (2.6) S ‘This relationship will prove useful later in this chapter in discussing certain kinds of boundary-value problems. Also notice that if V = 1 in equation 2.5 and if U is the potential of F, then [vw dv [F-ads. 8 In regions of space where U is harmonic, we have the same result as in Section 2.1.1, [Faas =o, $ namely, that the normal component of a conservative field averages to zero over any closed surface. 2.1.3 Green's Third Identity The third identity is a bit more difficult to d We begin by letting V = } in Green’s second identity (equation 2.5), where r is the distance between points P and @ inside region R (Figure 2.2): f[lovt-iee| ao= f [og t ie dS, P#Q. (27) R 5 Onr or én so long as P # Q. To insure that P # Q, we surround P with a sphere o and exclude it from R. Equation 2.7 becomes vu a1 _10U al - Taw f [ost eles fy =-- R > (2.8) Exercise 2.3 Show that } is harmonic for any region where r # 0. What happens at r = 0?2.1 Green’s Identities 25 Ss WQ dS Fig. 2.2. Derivation of Green’s third identity. Point P is inside surface S but is excluded from region R. Angle d® is the solid angle subtended by dS at point P. First consider the integral over o (Figure 2.2). We use the relationships, a_a on" Br al 1 Ber cos dS = 17 dQ, where dQ is the solid angle subtended at P by dS and @ = 0. The last. integral of equation 2.8 becomes {[ze+ 2 Baal Geir ees cht As the sphere becomes arbitrarily small, the right-hand side of the pre- vious expression approaches 4rU(P), and — 2.8 becomes 1 vu 1 te einer aa R Equation 2.9 is Green’s third identity.26 Consequences of the Potential ‘The significance of Green's third identity will become clear in later chapters, In Chapter 3 (equation 3.5), for example, we will show that a mass distribution described by a density p(Q) has a gravitational po- tential at point P given by uP) =71 fa R where @ is the point of integration, r is the distance from P to Q, and + i constant. This integral has the same form as the first integral of Green’s third identity if p = — 2, V7U. Similarly, the second integral of Green’ identity has the same form as the potential of a surface distribution of mass ¢, where ¢ = ;1, %. We will show in Chapter 5 (equation 5.2) that the magnetic potential of a distribution of magnetization M is given by V(P) = Cn [™ Vor dv, ® where Cp, is a constant, and this has the same form as the third integral of Green’s third identity if M is spread over S and directed normal to 8. But remember that no physical meanings were attached to U in deriving Green’s third identity; that is, U was only required to have a sufficient degree of continuity. Green’s third identity shows, therefore, that any function with sufficient differentiability can be expressed as the sum of three potentials: the potential of a volume distribution with density pro- portional to —V2U, the potential of a surface distribution with density proportional to 3%, and the potential of a surface of magnetization pro- portional to —U. Hence, we have the surprising result that any function with sufficient differentiability is a potential. An important consequence follows from Green's third identity when U is harmonic. Then equation 2.9 becomes up) — z/(% -vz4] as. (2.10) s ron onr This important result shows that a harmonic function can be calculated at any point of a region in which it is harmonic simply from the values of the function and its normal derivatives over the region's boundary. This equation is called the representation formula (Strauss [274]), and we will return to it later in this chapter and again in Chapter 12. Green’s third identity demonstrates an important limitation that faces any interpretation of a measured potential field in terms of its causative2.1 Green’s Identities 27 sources. It was shown earlier that a harmonic function satisfying a given set of Dirichlet boundary conditions is unique, but the converse is not true. If U is harmonic in a region R, it also must be harmonic in each subregion of R. Likewise equation 2.10 must apply to the boundary of each subregion. It follows that the potential within any subregion of R can be related to an infinite variety of surface distributions. Hence, no unique boundary conditions exist for a given harmonic function. This property of nonuniqueness will be a common theme in following chapters. 2.1.4 Gauss’s Theorem of the Arithmetic Mean Another consequence of Green’s third identity occurs when U is har- monic, the boundary $ is the surface of a sphere, and point P is at the center of the sphere. If @ is the radius of the sphere, then equation 2.10 becomes 1 ou 1 up)= 2 [Zas-z fu(-z) as. 3 5 The first integral vanis es according to Green's first identity, so that uP) = ga [ vas. (2.11) 5 Hence, the value of a harmonic function at any point is simply the av- erage of the harmonic function over any sphere concentric about the point, so long as the function is harmonic throughout the sphere. This relationship is called Gauss’s theorem of the arithmetic mean. We discussed the macimur principle by example in Section 1.3: If U is harmonic in region R, a closed and bounded region of space, then U attains its maximum and minimum values on the boundary of R, except in the trivial case where U is constant. Now we are in a position to prove it. The proof is by contradiction. Let 5 represent a set of points of R at which U attains a maximum M (Figure 2.3). Hence, u(s)=M = cannot equal the total of R because we have stated that U is not constant, and E must be closed because of the continuity of U. Suppose that © contains at least one interior point of R. It can be shown that if © has one point interior to R it also has a frontier point interior to R, which we call Py. Because P is interior to R, a sphere can be constructed centered about P that lies entirely within R. By the definition of a28 Consequences of the Potential R Fig. 2.3. Region R includes a set of points at which U attains a maximum, Point Py is a frontier point of ©. Any sphere centered on P) must contain at least one point of SZ and one point of R not in ©. frontier point (Section 1.1.2), such a sphere must pass outside of 3! where U ) 1 f FQ) 3 WP) = 5 / dv, (2.13) where Q is the point of integration, 7 is the distance between P and Q, and the integral is taken over all space. Each of the three cartesian components of W has a form like ~1f we=z | ‘At this point, we borrow a result from Chapter 3: Equation 2.14 is a solution to a very important differential equation, Poisson’s equation: dv. (2.14) VW, = —Fe (2.15) The relationship between equations 2.14 and 2,15 follows from Green’s third identity because the integration in equation 2.14 is over all space, and we have stipulated that F and, therefore, the three components of F vanish at infinity. Exercise 2.4 Show that equations 2.14 and 2.15 are consistent with Green's third identity. With W defined as in equation 2.13, the relationship between equa- tions 2.14 and 2.15 suggests that, vw =-F. (2.16) where each component of F leads to an example of Poisson's equation. A vector identity (Appendix A) shows that V?W can be represented by a gradient, plus a curl, that is, — VW =-V(V-W) 40 «(Fx W), (2. 7) and hence F is represented as the gradient of a sealar (V - W) plus the30 Consequences of the Potential curl of a vector (V x W). We define ¢ = —V-W and A = V x W and substitute these definitions along with eqnation 2.16 into equation 2.17 to get the Helmholtz theorem, F=Vo+VxA. Exercise 2.5 Prove the vector identity V(V -W)- V x V x W = V?7W. Hence, the Helmholtz theorem is proven: If F is continuous and vanishes at infinity, it can be represented as the gradient of a scalar potential plus the curl of a vector potential. The Helmholtz theorem is useful, however, only if the scalar and vec- tor potentials can be derived directly from F. This should be possible because of the way @ and A were defined, and the relationships can be seen by taking the divergence and curl of both sides of equation 2.12. The divergence yields Vo=V-F, which, comparing with equations 2.14 and 2.15, has the solution 1s VF o=-E/ > (2.18) The curl of equation 2.12 provides VA=V(V-A)-VXE. For convenience, we define A so that it hus no divergence, and conse- quently VWA=-VxF. Comparing this result with equations 2.13 and 2.16 leads to 1 fUxF A sf “— dv. (2.19) Exercise 2.6 Show that equation 2.19 implies than V-A Consequently, the scalar potential @ and vector potential A can be de- rived from integral equations taken over all space and involving the di- vergence and curl, respectively, of F itself. Exercise 2.7 Prove the last statement of the Helmholtz theorem; that is, show that V¢ and V x A, both vanishing at infinity, are orthogonal under integration over three-dimensional space,2.2 Helmholtz Theorem 31 2.2.2 Consequences of the Helmholtz Theorem ‘The Helmholtz theorem shows that a vector field vanishing at infinity is completely specified by its divergence and its curl if they are known throughout space. If both the divergence and curl vanish at all points, then the field itself must vanish or be coustant everywhere. In addition to this statement, the following important observations follow directly from the Helmholtz theorem and from the integral repre- sentation for scalar and vector potentials. Irrotational Fields A vector field is irrotational in a region if its curl vanishes at each point of the region; that is, F is irrotational in a region if V x F = 0 throughout the regiou. Such fields have no vorticity or “eddies.” For example, if the flow of a fluid can be represented as an irrotational field, then a small paddlewheel placed within the fluid will not rotate. Examples of irrotational fields are common and include gravitational attraction, of considerable importance to future chapters. Consider any surface S entirely within a region where V x F = 0. Integration of the curl over the surface provides [oven -nas= s and applying Stokes's theorem (Appendix A) provides fF-as=0, where the closed line integral is taken around the perimeter of S. Because the integral holds for any closed surface within the region, no net work is done in moving around any closed loop that lies within an irrotational field, that is, work is independent of path, and F is conservative, a sufficient condition for the existence of a scalar potential such that V@. Hence, the condition that V x F = 0 at each point of a region is sufficient to say that F = V@. Furthermore, a field that has a scalar potential has no curl because YV x F = V x Vo vanishes identically (Appendix A). Hence, the property that V x F = 0 at every point of a region is a necessary and sufficient condition for the existence of a scalar potential such that F = Vo.32 Consequences of the Potential Solenoidal Fields ‘A vector field F is said to be solenoidal in a region if its divergence vanishes at each point of the region. A physical meaning for solenoidal fields can be had by integrating the divergence of F over any volume V within the region, { V-Fdv ¥ and applying the divergence theorem (Appendix A) to get [Fads s where S is the closed boundary of V. Hence, if the divergence of F yanishes in a region, the normal component of the field vanishes when integrated over any closed surface within the region. Or put another way, the “number” of field lines entering a region equals the number that exit the region, and sources or sinks of F do not exist in the region. For example, gravitational attraction is solenoidal in regions not occupied by mass. Tt was stated in Section 2.1.1 that if a function ¢ can be found such that F = V¢, then the condition expressed by equation 2.20 is necessary and sufficient to say that ¢ is harmonic throughout the region. From the Helmholtz theorem, V-F = V?76+V-V x A. The last term of this equation vanishes identically (Appendix A), and V-F = V7¢. Hence, if the divergence of a conservative field vanishes in a region, the potential of the field is harmonic in the region. Note that if F = V x A, then V - F = 0; that is, the divergence of a vector field vanishes if the vector can be expressed purely as the curl of another vector. Furthermore, the converse can be shown to be true by taking the curl of both sides of equation 2.19. Hence, the property that V-F =0 is a necessary and sufficient condition. for F = Vx A (2.20) 2.2.3 Example Equation 2.18 is important to the geophysical interpretation of gravity and magnetic anomalies caused by crustal masses and magnetic sources, respectively, To see this, we use the magnetic field as an example and anticipate the results of future chapters.2.2 Helmholtz Theorem 33, A set of differential equations, called Maxwell’s equations, describes the spatial and temporal relationships of electromagnetic fields and their sources. One of Maxwell's equations relates magnetic induction B and magnetization M in the absence of macroscopic currents: Vx B= pV xM, where jg is the permeability of free space. Magnetic field intensity His related to magnetic induction and magnetization by the equation B = o(H+M). (2.21) Hence, in the absence of macroscopic currents, VxH=0, and it follows from the Helmholtz theorem that magnetic field intensity is irrotational and can be expressed in terms of a scalar potential, that is, H = —VV, where the minus sign is a matter of convention as discussed in Chapter 1. Moreover, equation 2.18 provides an expression for that scalar potential: 1 sVeH Ez { adv, (2.22) where it is understood that the integration is over all space. Another of Maxwell's equations states that magnetic induction has no divergence, that is, V-B = 0. This fact plus equation 2.21 yields v-H=-V-M, (2.23) and substituting into equation 2.22 provides 1 svem An fH dv. (2.24) Again integration is over all space. Equation 2.24 provides a way to cal culate magnetic potential and magnetic field from a known (or assumed) spatial distribution of magnetic sources. This is called the forward prob- lem when applied to the geophysical interpretation of measured magnetic fields. Equation 2.24 also is a suitable starting point for discussions of the inverse problem: the direct calculation of the distribution of magne- tization from observations of the magnetic field. We will return to this equation in Chapter 5 and subsequent chapters.34 Consequences of the Potential 2.3 Green’s Functions We now turn to Green’s functions, important tools for solving certain classes of problems in potential theory. A heuristic approach will be used, first considering a mechanical system and then extending this result to Laplace’s equation. 2.3.1 Analogy with Linear Systems We begin with the differential equation describing motion of a particle subject to both a resistance R and an external force f(t), m fue =-Ru(t)+ £O, (2.25) where v(t) is the velocity and m is the mass of the particle, respectively. One conceptual way to solve equation 2.25 is to abruptly strike the particle and observe its response; that is, we let the force be zero except over a short time interval Ar, £, ifrT+Ar. (2.27) The coefficient A can be found if the velocity of the particle is known at the moment that the force returns to zero; that is, v(7 + At) = A. To find this velocity, we integrate both sides of equation 2.25 over the duration of the force thar rtar (7 + Ar) —a(7)] =-R { veya + [ #. we ar The first integral can be ignored if Ar is small and the particle has some mass. Also v(r) = 0. Hence, mo(r+Ar)=T,2.3 Green’s Functions 35 Tt tat Time ——> Fig. 2.4. Velocity of a particle of mass m resulting from an impulsive force of magnitude I. and A = I/m for small Ar. Combining this result with equation 2.27 provides Fe Rtn, itt> rj v(t) = (2.28) 0, ift = emt), t> ty. (2.29) Past If the blows become sufficiently rapid, the particl continuous force. Then Ix + f(r)d7 and is subjected to a ‘ v(t) = a [toe ar, t>, which can be rewritten as vo) = f vtrserar, (2.30)36 Consequences of the Potential 0, ift<7; W(t7) = where LenB), ift>r. Equation 2.30 is the solution to the differential equation 2.25. It pre- sumes that the response of the particle at each instant of impact is independent of all other times. Given this property, the respouse of the particle to f(t) is simply the sum of all the instantaneous forces, and the particle is said to be a linear system. Many mechanical and electrical systems (and, as it turns out, many potential-field problems) have this property. ‘The function #(¢,r) is the response of the particle at time ¢ due to an impulse at time 7; it is called the impulse response or Green's function of the linear system. The Green’s function, therefore, satisfies the initial conditions and is the solution to the differential equation 2.25 subject to the initial conditions when the forcing function is an impulse. Equation 2.26 is a heuristic description of an impulse. In the limit as Ar approaches zero, the impulse of equation 2.26 becomes arbitrarily large in amplitude and short in duration while its integral over time remains the same. The limiting case is called a Dirac delta function (2), which has the properties f sar - Har= Fer) 31) These definitions and properties are meaningless if 5(t) is viewed as an ordinary function. It should be considered rather as a “generalized function” characterized by the foregoing properties. Green's functions are very useful tools; equation 2.30 shows that if the Green’s function q is known for a particular linear system, then the state of the linear system due to any forcing function can be derived for any time.2.3 Green's Functions 37 2.3.2 Green’s Functions and Laplace’s Equation The previous mechanical example provides an analogy for potential the- ory. In Chapter 3, we will derive Poisson’s equation WU = —4aryp ¢ ) This second-order differential equation describes the Newtonian poten- tial U throughout space due to a mass distribution with density p. Clearly VU = 0 and U is harmonic in regions where p = 0. We seek a solution for U' that satisfies the differential cquation and the boundary condition that U is zero at infinity. The density distribution in Poisson’s equation is obviously the source of U and in this sense is analogous to the forcing function f(t) of the previous section. We know from the previous scetion that the response to an impulsive forcing function f(t) = 6(t) is the Green’s function, so we could try representing the density distribution in & as an “impulse” and see what happens to U. An impulsive source in three dimensions can be written as 6(P,Q), where far. Q)dv=1, (P,Q)=0 if PQ, and where @ is the point of integration as in Section 2.2.1. Hence, we let the density be 6(P,Q) and the potential be 41 in equa- tion 2.32, Vuh = —4r78(P,Q). Then from the Helmholtz theorem and equations 2.14 and 2.15, &(P, w(PQ=9 f PQ) a, =i (2.33) where r is the distance between P and Q. This is a very interesting result, We see that 7/r is the solution to Poisson's equation when p is, an “impulsive” density distribution located at Q. Indeed, we will show in Chapter 3 that 7/7 is the Newtonian potential at P due to a point mass at Q. Hence, 7/r is the “impulse response” for Poisson’s relation; with it, the potential due to any density distribution can be determined with an integral equation analogous to equation 2.30: u(P)= [ vilP.Q\AQ)do 34)38 Consequences of the Potential 7 f 2), (2.35) where it is understood that integration is over all space. This fundamen- tal equation relating gravitational potential to causative density distri butions will be derived in a different way in Chapter 3. The important point to be made here is that the function ¢ = 7/r is analogous to the Green’s function of the mechanical example in the previous section: It satisfies the required boundary condition, that ¢1 is zero at infinity, and is the solution to Poisson's differential equation when the density is an “impulse.” Half-Space Regions ‘The representation formula, which followed from Green’s third identity, shows that the value of a function harmonic in R can be found at any point within F strictly from the behavior of U and its normal derivative on the boundary of R, that is, U(P) = “a le oa | ds, (2.36) where Q is the point of integration and r is the distance from P to Q. In practical situations, we are unlikely to have both the potential and its normal derivative at our disposal, and elimination of ge make this equation much more useful. We should expect that such a simplification is possible because earlier results have shown that the potential is uniquely determined by its boundary conditions. To eliminate gw from the third identity, we begin with Green’s second identity. Let both U and V be harmonic in equation 2.5 so that 1 spay _ au af ex a a 7 Adding this equation to equation 2.36 provides an + would 02.3 Green’s Functions 39 Bs Vi Wf Poanyrde) P(x,yAz) Fig. 2.5. Function U is harmonic throughout the half-space z < 0 and assumes known values on the surface z = 0. Parameter r is the distance between points P and Q; r’ is the distance between P’ and Q. Point P' is the image of point P such that r =r’ when Q is on the surface z = 0. If we select a harmonic V such that V+ 2 =0 at each point of 5, then 1 a 1 U(P)=-— | U=—|V+-]ds. 2.37, ine-bfok (vet) an Hence, if for a particular geometry we can find a function V such that (1) V is harmonic throughout R and (2) V + 1 = 0 at each point of S, then U can be found throughout the region, and only values of U on the boundary will be required. ‘he function V + t is called the Green’s function for Laplace’s equation in restricted regions. It satisfies Laplace’s equation throughout the region (except when P = Q) and is zero on the boundary. In principle, equation 2.37 provides a simple way to solve Laplace's equation from specified boundary conditions. Unfortunately, the func- tion V is very difficult to derive analytically except for the simplest sorts of geometrical situations, such as half-spaces and spheres. As an exam- ple, consider the half-space problem, where U is harmonic for all z <0 and is known on the planar surface z = 0 (Figure 2.5). Boundary $ then consists of the z = 0 plane plus the 2 < 0 hemisphere, as shown in Fig- ure 2.5. We construct a point P’ below the z = 0 plane that is the image of point P. The necessary properties are satisfied if we let V = —4, where r’ is the distance from P’ to Q: namely, V is always harmonic40 Consequences of the Potential since Q is always above or on the z = 0 plane, V +3 = 0 when Q is on the z = 0 plane, and V + } = 0 when Q is on the infinite hemisphere. Hence, V defined in this way satisfies the necessary requirements to be used in equation 2.375 that is, uP) (2.38) U(x,y,-Az) A: 7 U(w,8,0) On Tih ewitne eae te where Az > 0. Equation 2.39 provides a way to calculate the potential at any point above a planar surface on which the potential is known, Such calculations are called upward continuation, a subject that. will be revisited at some length in Chapter 12. Terminology ‘The Green’s function for Poisson’s equation throughout space is usually derived from a general form of Poisson's equation, V2 = —f, and thus is given by G = 74. (eg., Duff and Naylor [81], Strauss [274]). Here, we started with the gravitational case of Poisson's equation, V?U7 = —47p, and derived a slightly different form for the Green’s function, ‘The present approach led to equation 2.34, that is, UP) = [owe p(Q) dv, where integration is over all space. This simple integral expression for the potential in terms of density and the Green’s function will prove useful in following chapters ‘We also showed that if a function V can be found satisfying just two properties (V is harmonic throughout a region, and V + } is zero on the boundary of the region), then the representa very simple form, vey--2 fuZ(v+}) as ! The function V + 2 is the Green’s function for Laplace's equation in restricted regions. In future chapters, however, we will use a somewhat ion formula reduces to a2.4 Problem Set 4 looser terminology. If we let v2(P,Q) = - 2 2(V +2) in this equation, then U(P) = fu wn(P,Q) dS, s which has a form similar to equation 2.34. For this reason, we also will refer to U2(P,S) = — 2 Z(V + 4) as a kind of Green’s function, one which provides U at points away from boundaries on which U is known, 2.4 Problem Set 1. In the following, T is temperature in region R bounded by surface S, and f is the unit, vector normal to S. If ¥?T = —f(P) in region Rand 22 = g(8) on surface $. show that [tas [ods~ s and interpret the meaning of this equation. Show, starting with Green’s first identity, that if U is harmonic throughout all space. it must be zero everywhere 3. Show that. if two harmonic functions Uj and V2 satisfy Uy < Uy at each point of the boundary of a region, then U, < Ug throughout the region. 4. A radial field is described by F = ar"é. wy (a) In regions where r 4 0, find the values of n for which F is solenoidal. (b) For what values of n (r # 0) is the field irrotational? 5. Maxwell’s equations state that magnetic induction B in the absence of moving charge is both solenoidal and irrotational, that is, V-B=0, VxB=0 Show that the three cartesian components of B are each harmonic in such situations. 6. Function U satisfies the two-dimensional Laplace’s equation at every point of a circle. Find a Green’s function that will provide the value of42 _ Consequences of the Potential U at any point inside the circle from the values of U' on the boundary of the circle. Function U is harmonic everywhere inside a sphere of radius a. Find a Green’s function that will provide the value of U at any point inside the sphere from values of U on the surface of the sphere.