You are on page 1of 16

# Differential forms

September 6, 2012

1

From differentials to differential forms

In a formal sense, we may define differentials as the vector space of linear mappings from curves to the reals,
that is, given a differential df we may use it to map any curve, C ∈ C to a real number simply by integrating:
df : C

x

=

R
ˆ
df
C

This suggests a generalization, since we know how to integrate over surfaces and volumes as well as curves. In
higher dimensions we also have higher order multiple integrals. We now consider the integrands of arbitrary
multiple integrals
ˆ
ˆ ˆ
ˆ ˆ ˆ
f (x)dl,
f (x)dS,
f (x)dV
(1)
Much of their importance lies in the coordinate invariance of the resulting integrals.
One of the important properties of integrands is that they can all be regarded as oriented. If we integrate
a line integral along a curve from A to B we get a number, while if we integrate from B to A we get minus
the same number,
ˆ B
ˆ A
f (x)dl = −
f (x)dl
(2)
A

B

We can also demand oriented surface integrals, so the surface integral
ˆ ˆ
A · n dS

(3)

changes sign if we reverse the direction of the normal to the surface. This normal can be thought of as the
cross product of two basis vectors within the surface. If these basis vectors’ cross product is taken in one
order, n has one sign. If the opposite order is taken then −n results. Similarly, volume integrals change sign
if we change from a right- or left-handed coordinate system.
The generalization from differentials to differential forms, and the associated vector calculus makes use
of three operations: the wedge product, the exterior derivative, and the Hodge dual. We discuss this in turn.

1.1

The wedge product

We can build this alternating sign into our convention for writing differential forms by introducing a formal
antisymmetric product, called the wedge product, symbolized by ∧, which is defined to give these differential
elements the proper signs. Thus, surface integrals will be written as integrals over the products
dx ∧ dy, dy ∧ dz, dz ∧ dx

1

line elements 1-forms. The wedge product of a p-form with a q-form is a (p + q)-form. We assume this wedge product is associative. Functions are called 0 -forms. a general surface integrand is Az dx ∧ dy + Ay dz ∧ dx + Ax dy ∧ dz while the volume integrand is f (x) dx ∧ dy ∧ dz These objects are called differential forms. d. Similarly. Using this. We can go further than this by formalizing the full integrand. and obeys the usual distributive laws. the general form of the integrand is a linear combination of the basis differentials. but in more than three dimensions we can have p-forms with p ranging from zero to the dimension. Since we can take arbitrary linear combinations of p-forms. Notice that the antisymmetry is all we need to rearrange any combination of forms. Λp . For a line integral. These are all the types that exist in 3 -dimensions. Similarly. we may interchange the order of a line element and a surface area. This automatically gives the right orientation of the surface. of the space. In general. differential forms come in several types. and volume forms are called 3-forms. for it l1 l2 = Adx = Bdy + Cdz 2 . Ax dx + Ay dy + Az dz Notice that we simply add the different parts. wedge products of even order forms with any other forms commute while wedge products of pairs of odd-order forms anticommute. In particular. We can always wedge together any two forms. the volume element becomes V = dx ∧ dy ∧ dz which changes sign if any pair of the basis elements are switched. functions (0-forms) commute with all p-forms. surface elements 2-forms. they form a vector space. Clearly.with the convention that ∧ is antisymmetric: dx ∧ dy = −dy ∧ dx under the interchange of any two basis forms. for if l S = Adx = Bdy ∧ dz then l∧S = (A dx) ∧ (B dy ∧ dz) = A dx ∧ B dy ∧ dz = AB dx ∧ dy ∧ dz = −AB dy ∧ dx ∧ dz = AB dy ∧ dz ∧ dx = S∧l but the wedge product of two line elements changes sign.

v) = ∂x ∂x du + dv ∂u ∂v and similarly for dy.). 3 . identify the property of form multiplication used in each step (associativity. commutation of 0-forms. For example l∧V = (A dx) ∧ (B dx ∧ dy ∧ dz) = AB dx ∧ dx ∧ dy ∧ dz = 0 since dx ∧ dx = 0. Now suppose we want to change coordinates. In general. etc. applying the usual rules for partial differentiation. y) in the plane are given as some functions of new coordinates (u. if we have more dimensions then there are more independent directions and we can find nonzero 4-forms. Of course the volume change works. showing the anticommutation of two 1-forms. in d-dimensions we can find d-forms. This is exactly the way that an area element changes when we change coordinates! Notice the Jacobian coming out automatically. too. We couldn’t ask for more – the wedge product not only gives us the right signs for oriented areas and volumes.(4). but no (d + 1)-forms. we immediately have ω ∧ ω = −ω ∧ ω = 0 In 3-dimensions there are no 4-forms because anything we try to construct must contain a repeated basis form. Notice what happens when we use the wedge product to calculate the new area element:  dx ∧ dy = = = =    ∂x ∂x ∂y ∂y du + dv ∧ du + dv ∂u ∂v ∂u ∂v ∂x ∂y ∂x ∂y dv ∧ du + du ∧ dv ∂v ∂u ∂u   ∂v ∂x ∂y ∂x ∂y − du ∧ dv ∂u ∂v ∂v ∂u J du ∧ dv where ∂x ∂u ∂y ∂u  J = det ∂x ∂v ∂y ∂v  is the Jacobian of the coordinate transformation. anticommutation of basis forms.then l1 ∧ l2 = (A dx) ∧ (Bdy + Cdz) = A dx ∧ Bdy + A dx ∧ Cdz = AB dx ∧ dy + AC dx ∧ dz = −AB dy ∧ dx − AC dz ∧ dx = −Bdy ∧ Adx − Cdz ∧ Adx = −l2 ∧ l1 (4) For any odd-order form. Then we already know that differentials change according to dx = dx (u. Of course. but gives us the right transformation to new coordinates. v). ω. The same occurs for anything we try. How does an integrand change? Suppose Cartesian coordinates (x. In eq.

we have defined a product of basis forms that we can generalize to more complicated objects. uvw) du ∧ dv ∧ dw So the wedge product successfully keeps track of p-dim volumes and their orientations in a coordinate invariant way. w) z → z (u. of x (u. v.Show that under a coordinate transformation x y → x (u. we defined the wedge product. dx ∧ dy ∧ dz = J (xyz. we can regard the integrand as being a differential form. Writing them with a boldface d indicates this vector character. w) → y (u. v. So symbols like dx or du contain directional information. Thus. Then. It vanishes along surfaces where x is constant. v. Now any time we have an integral. ∂u ∂v point in a direction normal to those surfaces. 4 . We have defined forms. v) given above is just a gradient. we write A = Ai dxi Let f (x. which enables us to write pdimensional integrands as p-forms in such a way that the orientation and coordinate transformation properties of the integrals emerges automatically. Though it is 1-forms. Many of these objects are already familiar. But all of this can go much further. have written down their formal properties. y) = axy Show that the vector with components  ∂f ∂f . Consider the product of two 1-forms. the differential. Thus. dx. w) the new volume element is just get the full Jacobian times the new volume form. and the components of the vector   ∂x ∂x . Ai dxi that correspond to vectors. We will see this in more detail below when we discuss the curl. A ∧ B = Ai dxi ∧ Bj dxj = Ai Bj dxi ∧ dxj  1 = Ai Bj dxi ∧ dxj − dxj ∧ dxi 2  1 = Ai Bj dxi ∧ dxj − Ai Bj dxj ∧ dxi 2  1 = Ai Bj dxi ∧ dxj − Aj Bi dxi ∧ dxj 2 1 = (Ai Bj − Aj Bi ) dxi ∧ dxj 2 The coefficients Ai B j − Aj B i are essentially the components of the cross product. Recall our proof that 1-forms form a vector space. Let’s sum up. ∂x ∂y  is perpendicular to the surfaces of constant f. and have use those properties to write them in components.

the coefficients of this 1-form are just the Cartesian components of the gradient. The same applies if we apply d twice to any function: d2 f = d (df )   ∂f i dx = d ∂xi   ∂f = d ∧ dxi ∂xi  2  ∂ f j = dx ∧ dxi ∂xj ∂xi ∂2f = dxj ∧ dxi ∂xj ∂xi By the same argument we used to get the components of the curl.2 The exterior derivative We may regard the differential of any function. say f (x. The operator d is called the exterior derivative. regarded as a function of x. since an arbitrary p-form in n-dimensions may be written as ω = Ai1 i2 ···ip dxi1 ∧ dxi2 · · · ∧ dxip we define the exterior derivative of ω to be the (p + 1)-form dω = dAi1 i2 ···ip ∧ dxi1 ∧ dxi2 · · · ∧ dxip Let’s see what happens if we apply d twice to the Cartesian coordinate. y. x. First consider a 1-form A = Ai dxi We define dA = dAi ∧ dxi Similarly. z). we may write this as 5 . y and z: d2 x = d (dx) = d (1dx) = d (1) ∧ dx = 0 since all derivatives of the constant function f = 1 are zero. and we may apply it to any p-form to get a (p + 1)-form.1. In Cartesian coordinates. as the 1-form: df = = ∂f ∂f ∂f dx + dy + dz ∂x ∂y ∂z ∂f dxi ∂xi Since a function is a 0-form then we can imagine an operator d that differentiates any 0-form to give a 1-form. The extension is defined as follows.

Also. suppose we have a 2-form expressed as S = Az dx ∧ dy + Ay dz ∧ dx + Ax dy ∧ dz Then applying the exterior derivative gives dS = = = dAz ∧ dx ∧ dy + dAy ∧ dz ∧ dx + dAx ∧ dy ∧ dz ∂Ay ∂Ax ∂Az dz ∧ dx ∧ dy + dy ∧ dz ∧ dx + dx ∧ dy ∧ dz ∂z ∂y ∂x   ∂Az ∂Ay ∂Ax + + dx ∧ dy ∧ dz ∂z ∂y ∂x (6) so that the exterior derivative can also reproduce the divergence. consider the effect of d on an arbitrary 1-form. ∗ . or star. i. Later we will see how these components relate to those in a general coordinate system. 1. We must be careful here. Finally. These differences will be resolved when we give a detailed discussion of the metric in Section (5.d2 f 1 2 0 = =  ∂2f ∂2f − ∂xj ∂xi ∂xi ∂xj  dxj ∧ dxi since partial derivatives commute.3 The Hodge dual To truly have the curl in eq. y. both 1-forms and 2-forms have three independent components. while both 0. Prove the Poincaré Lemma: d2 ω = 0 where ω is an arbitrary p-form. z) dx ∧ dy ∧ dz. Next.e.2) that the components Ai are distinct from the usual vector components Ai . In Cartesian coordinates. suppose we set ∗ ∗ (dx ∧ dy) = dz ∗ (dy ∧ dz) = dx ∗ (dz ∧ dx) = dy (dx ∧ dy ∧ dz) = 1 6 . Compute the exterior derivative of the arbitrary 3-form. This suggests that we can define an invertible mapping between these pairs. recall from Section (4. the action of d on a 1-form gives us a coordinate invariant way to calculate the curl.2. operator.(5). a 1-form and a way to turn a 3-form into a 0-form. Ultimately. We have dA  = d Ai dxi   ∂Ai j = dx ∧ dxi ∂xj   ∂Aj 1 ∂Ai − dxj ∧ dxi = 2 ∂xj ∂xi (5) We have the components of the curl of the vector A. A = f (x.6).(6) or the curl in eq. Fill in the missing steps in the derivation of eq. Notice that in 3-dim. we need a way to turn a 2-form into a vector. This leads us to introduce the Hodge dual..(6). because these are the components of the curl only in Cartesian coordinates.and 3-forms have one component. however.

In fact. We have now shown how three operations – the wedge product ∧. curl and gradient. no information is lost in restricting our attention to them. 2 2. To explore these generalizations. This will allow us to define the Hodge dual in arbitrary coordinates. but only orthogonal transformations preserve orthonormal frames. Given an orthonormal frame ea = em a dxm we can use the coefficient matrix em a and its inverse to transform back and forth between orthonormal and coordinate indices. Nonetheless. Write the Cartesian dot product of two 1-forms in terms of wedge products and duals. Arbitrary coordinate transformations – diffeomorphisms – preserve the class of frames. We can always return to general frames if we wish. Show that the dual of a general 1-form. But as long as we maintain the restriction. we must first explore properties of the metric and look at coordinate transformations. ∗∗ =1 With these rules we can find the Hodge dual of any form in 3-dim. The correspondence between general frame and orthonormal frame tensors is provided by the orthonormal frame itself. the exterior derivative d. given any vector in an arbitrary coordinate basis. they do much more – they extend all of these operations to arbitrary coordinates and arbitrary numbers of dimensions. Thus.1 Transformations and the volume form Transformations Since the use of orthonormal frames is simply a convenient choice of basis.and further require the star to be its own inverse. A = Ai dxi is the 2-form S = Az dx ∧ dy + Ay dz ∧ dx + Ax dy ∧ dz Show that for an arbitrary (Cartesian) 1-form A = Ai dxi that ∗ ∗ d A = divA Write the curl of A  curl (A) = ∂Az ∂Ay − ∂z ∂y   dx + ∂Az ∂Ax − ∂x ∂z   dy + ∂Ax ∂Ay − ∂y ∂x  dz in terms of the exterior derivative and the Hodge dual. the class of tensors is remains unchanged – there is a 1-1. v = vm ∂ ∂xm we may insert the identity in the form δnm = en a ea 7 m . we can work with a reduced form of the symmetry group. and the Hodge dual ∗ – together encompass the usual dot and cross products as well as the divergence. onto correspondence between tensors with diffeomorphism covariance and those with orthogonal covariance.

... for example T m1 .ns em1 . ebs ns Similar expressions may be written for tensors with their contravariant and covariant indices in other orders. .in ∂xi1 ∂xi2 ∂xin ..ar b1 .ns may be written in an orthonormal basis by using one factor of em a or ea T a1 .... since multiplying any vector by the identity should be independent of coordinate system.bs = T m1 . emnar eb1 n for each linear slot: n1 .. . however. 1). together with the transformation law for gij .. z) = z Notice that the identity matrix should exist in any coordinate system. .jn ∂y j1 ∂y j2 ∂y jn 8 . 2... where it is given by the totally antisymmetric symbol εi1 i2 . ϕ.. we have only defined the Levi-Civita tensor in Cartesian coordinates. ϕ. z) = ρ sin ϕ z = z (ρ. = Jεj1 j2 . The easiest way to do this is to use the equation gij g jk = δik and the result of exercise 2. defined to be the unit matrix in one coordinate system.. . is not quite a tensor because under a diffeomorphism it becomes εi1 i2 . 1. This symbol. Derive the form of the metric in cylindrical coordinates directly from the coordinate transformation.to write v ∂ ∂xm = v n δnm = v n en a ea = (v n en a ) ea = v a ea m ∂ ∂xm The mapping v a = v n en is invertible because en a a is invertible. z) = ρ cos ϕ y = y (ρ.mr n1 . x = x (ρ. has the same form in every other coordinate system. We showed in Section (3) that the components of the metric are related to the Cartesian components by gjk = ∂xm ∂xn ηmn ∂y j ∂y k where we have corrected the index positions and inserted the Cartesian form of the metric explicitly as ηmn = diag(1. ..2 The volume form So far. any tensor.in in n dimensions.mr a1 n1 . Show that the matrix δ i j ... Notice that the upper index will transform like a contravariant vector and the lower index like a covariant vector. Similarly.. Show that the inverse to the metric transforms as a contravariant second rank tensor.. Also note that δ i j = δj i . ϕ..

We can correct for this to form a tensor by dividing by another density.. A= 1 Ai .. Since the metric transforms as 0 gmn = ∂xi ∂xj gij ∂y m ∂y n the determinants are related by g0 0 det gmn  i  ∂x ∂xj = det g ij ∂y m ∂y n ∂xi ∂xj = det m deg tij det n ∂y ∂y 2 = J g = Therefore. . leaving e0i. g jn in εi1 i2 .where  J = det ∂xm ∂y n  is the Jacobian of the coordinate transformation. 3 Calculus using differential forms Define a p-form as a linear map from oriented p-dimensional volumes to the reals: Λp : Vp → R Linearity refers to both the forms and the volumes. aΛ1p + bΛ2p is also a p-form.. ..jn = √ ε g This is also a tensor. inserting a convenient normalization... Then. For p-forms.    Λp Vp1 + Vp2 = Λp Vp1 + Λp Vp2 In Section 3.. we showed for 1-forms that these conditions specify the differential of functions. we have √ j1 i1 j2 i2 ej1 j2 . The most convenient choice is the determinant of the metric.j is a tensor... If we raise all indices on ei1 i2 .. for any two p -forms.j so that ei. in the combination ei..jn gg ε = 1 j1 j2 ..in .. and any constants a and b.. ...jn = gg g . while for any two disjoint p-volumes.j the factors of J cancel. ∧ dxip p! 1 p 9 .. Vp1 and Vp2 . Thus. but J is a density not a scalar. using n copies of the inverse metric.i dxi1 ∧ . .j = p g 0 εi. Λ1p and Λ2p . Let A be a p-form in d-dimensions..j = √ gεi. The transformation is linear and homogeneous. they are equivalent to linear combinations of wedge products of p 1-forms..in √ −1 j1 j2 ....

∂f dxi ∂xi  Notice that the coefficients are components of a type.i e 1 p (d − p)!p! 1 p ip+1 . Differential calculus is defined in terms of these three operations.The action of the exterior derivative..01 tensor. . but now we can write its full definition. associative.1 Grad. we require the metric: ∂f i [∇f ] = g ij j ∂x df = 10 . they allow us to perform all standard calculus operations in any number of dimensions and in a way independent of any coordinate choice. In Cartesian coordinates. Curl and Laplacian It is straightforward to write down the familiar operations of gradient and curl and divergence. but it makes a difference when we look at other coordinate systems. so that if we want the gradient to be a vector. . (ρ. Together. on such a p-form is   1 ∂ dA = Ai .. and apply each in polar coordinates. d) . . the Hodge dual. these two forms represent the same array of numbers... .id dx ip+1 ∧ . θ. ∧ dxip p! ∂xk 1 p We also defined the wedge product as a distributive. d. (∧. z) .. was provisionally defined in Cartesian coordinates.i dxk ∧ dxi1 ∧ .∗ . Div. The dual of A is defined to be the (d − p)-form ∗ A= 1 i . We specify each.i Ai . Recall that the metric in polar coordinates is   1  ρ2 gmn =  1 its inverse is   1 1 ρ2 g mn =   1 and its determinant is g = det gmn = ρ2 Gradient The gradient of a function is given by the exterior derivative of a 0 -form. ∧ dxid Notice that we have written the first p indices of the Levi-Civita tensor in the superscript position to keep with our convention of always summing an up index with a down index. antisymmetric product on 1-forms:  adxi + bdxi ∧ dxj  dxi ∧ dxj ∧ dxk dxi ∧ dxj = adxi ∧ dxj + bdxi ∧ dxj  = dxi ∧ dxj ∧ dxk = −dxj ∧ dxi A third operation. 3....

the in the middle is also. components of the original form. so the expression reduces to the familiar form of the divergence and 1 ∂ √ m ∇·ω = √ ( gω ) g ∂xm In polar coordinates we have ∇·ω = = = 1 ∂ p 2 m  p ρ ω ρ2 ∂xm   1 ∂ p 2 ρ  ∂ p 2 ϕ  ∂ p 2 z  p ρ ω + ρ ω + ρ ω ∂ϕ ∂z ρ2 ∂ρ ∂ω ϕ ∂ω z 1 ∂ (ρω ρ ) + + ρ ∂ρ ∂ϕ ∂z 11 . we may replace ω i = g ij ωj so that 1 ∂ √ m d ω=√ ( gω ) = ∇ · ω g ∂xm ∗ ∗ Since the operations on the left are all coordinate invariant. Notice that in Cartesian coordinates the metric is just δij . Starting with a 1-form.For example. rather than form. the gradient in polar coordinates has components    ∂f   1 ∂ρ  ∂f   i 1  [∇f ] =  = 2  ρ ∂ϕ  ∂f 1 ∂z so ∇f = ∂f ∂ρ 1 ∂f ρ ∂ϕ ∂f ∂z    ∂f 1 ∂f ∂f ˆ ρˆ + ϕ ˆ+ k ∂ρ ρ ∂ϕ ∂z Divergence The use of differential forms leads to an extremely useful expression for the divergence – important enough that it goes by the name of the divergence theorem. with determinant 3. ω = ωi dxi . we compute ∗ ∗ d ω = = = = = = = = ∗ ∗ d ωi dxi   1 ∗ ωi ei jk dxj ∧ dxk d 2  1∗ √ d ωi gg in εnjk dxj ∧ dxk 2  1∗ ∂ √ ωi gg in εnjk dxm ∧ dxj ∧ dxk m 2 ∂x  1 ∂ √ ωi gg in εnjk emjk m 2 ∂x 1 1 ∂ √ in  ω gg εnjk εmjk √ i 2 g ∂xm  1 1 ∂ √ ωi gg in 2δnm √ m 2 g ∂x  1 ∂ √ ωi gg im √ m g ∂x In terms of the vector.

if ω = ωi dxi then ∗ dω = = = = = ∂ ωi dxj dxi ∂xj   ∂ ji e k ωi dxk ∂xj ∂ eji k gim g mn ωn dxk ∂xj   ∂ ∂ mn eji k gim (g mn ωn ) − ωn g dxk ∂xj ∂xj   ∂ m ∂ mn lj s elmk g ω − ω gsn g dxk ∂xj ∂xj ∗ Now observe that gsn ∂ mn g ∂xj ∂ ∂ (gsn g mn ) − g mn (gsn ) ∂xj ∂xj ∂ m ∂ = δ − g mn (gsn ) ∂xj s ∂xj ∂ gsn = −g mn ∂xj = so that ∗ dω = =   ∂ m ∂ elmk g lj ω + ω s g mn gsn dxk ∂xj ∂xj   ∂ ∂ m ω + ω s ejn k gsn dxk elmk g lj ∂xj ∂xj Next consider ejn k ∂ gsn ∂xj = = = = ejn k ∂j gsn 1 jn e k (∂j gsn − ∂n gsj ) 2 1 jn e k (∂j gsn − ∂n gsj + ∂s gjn ) 2 ejn k Γnsj This combines to ∗ dω = = = = = =   ∂ lj ∂ m s jn elmk g ω +ω e k gsn dxk ∂xj ∂xj   lj ∂ m s jn ω + ω e k Γnsj dxk elmk g ∂xj   ∂ m j nm s emk ω + g ω Γnsj dxk ∂xj  ejmk ∂j ω m + ω s Γm sj dxk ejmk Dj ω m dxk  ejmk Dj ω m dxk 12 .Curl The curl is the dual of the exterior derivative of a 1-form. Thus.

wz . start with the components of the vector. wz )  The corresponding form has components ωi = gij wj = wρ . ρ2 wϕ . let wi = (wρ . if we raise the free index. wϕ . Therefore. In our familiar example in polar coordinates.Therefore. the curl is [∇ × ω] i = g ik ejmk Dj ω m 1 = √ εijk Dj ωk g  Also consider d∗ ω = d ei = d j k jk ωi dx dx  eijk ω i dxj dxk  √  gεijk ω i dxj dxk  ∂ √ i = gω εijk dxm dxj dxk m ∂x   ∂ = eji k j ωi dxk ∂x = d The simplest form computationally uses this to write ∗ i dω = [∇ × ω] gik dxk To apply the formula. the exterior derivative is  dω = d wρ dρ + ρ2 wϕ dϕ + wz dz ∂wρ ∂wρ = dϕ ∧ dρ + dz ∧ dρ ∂ϕ ∂z   ∂ ∂ + ρ2 wϕ dρ ∧ dϕ + ρ2 wϕ dz ∧ dϕ ∂ρ ∂z ∂wz ∂wz + dρ ∧ dz + dϕ ∧ dz ∂ρ ∂ϕ     z  ∂wρ  ∂ ∂w ∂ 2 ϕ 2 ϕ = ρ w − − ρ w dϕ ∧ dz dρ ∧ dϕ + ∂ρ ∂ϕ ∂ϕ ∂z  ρ  ∂w ∂wz + − dz ∧ dρ ∂z ∂ρ Now the dual maps the basis as 1 dz ρ 1 ∗ dϕ ∧ dz = e231 g11 dρ = dρ ρ ∗ 312 dz ∧ dρ = e g22 dϕ = ρdϕ ∗ dρ ∧ dϕ = e123 g33 dz = so that ∗ dω     ∂wρ ∂ 1 ∂wz ∂ 2 ϕ ϕ ρ w − dz + −ρ (w ) dρ = ∂ρ ∂ϕ ρ ∂ϕ ∂z  ρ  ∂w ∂wz +ρ − dϕ ∂z ∂ρ 1 ρ  13 .

Express your results using a basis of unit vectors. If we restrict to 3-dim and integrate dη along any curve the result depends only on the value of η at the endpoints.Now. curl. divergence and laplacian in spherical coordinates. Consider how this relates to the familiar 3-dim form of Stokes’ theorem. if dω = 0 throughout a concave region. then there exists a (p − 1)-form η such that ω = dη This provides a converse because if ω = dη then the Poincaré lemma immediately implies dω = 0. ˆ η= ω C This integral gives rise to a (single valued) function if and only if it vanishes on all closed paths. ˛ dη = 0 C Now consider the integral along a curve C for an arbitrary 1-form ω. 4 The Poincaré lemma and Stokes’ theorem We have already seen the Poincaré lemma. ˆ dη = η (x2 ) − η (x1 ) C Therefore. since ∗ i dω = [∇ × ω] gik dxk we use the inverse metric on the components of ∗ dω to find ω i = g ij ωj so we have 1 = 2 = 3 = [∇ × ω] [∇ × ω] [∇ × ω] ∂ 1 ∂wz −ρ (wϕ ) ρ ∂ϕ ∂z   1 ∂ω ρ ∂ω z − ρ ∂z ∂ρ    ∂wρ 1 ∂ ρ2 w ϕ − ρ ∂ρ ∂ϕ Work out the form of the gradient. which states that for any p-form ω. ˛ ω=0 C 14 . d2 ω ≡ 0 for any p-form ω. around any closed curve the integral vanishes. The extremely important Stokes’ theorem is the converse.

β. ˛ 0 = ω C ¨ = dω S and the arbitrariness of the surface S allows us to conclude dω = 0 Show that if ω is a 2-form in 3-dim. 4π ρ c 0 ∇·E = ∇·B = 1 ∂B ∇×E+ = c ∂t 1 ∂E ∇×B− = c ∂t 0 4π J c may be written in terms of .Stokes’ theorem allows us to write this as a surface integral of dω. then the generalized Stokes’ theorem is the divergence theorem. κ and ρ as = 4π ρ c 0 = 0 = 4π κ c ∗ ∗ d  = dβ 1 ∂β dε + c ∂t 1 ∂ ∗ ∗ d β− c ∂t The third equation may be proved as follows: dε + 1 ∂β c ∂t =  1 ∂ 1 i d E i gij dxj + B eijk dxj ∧ dxk c ∂t 2 15 . 5 Exercises: Maxwell’s equations In an orthonormal vector basis the electric and magnetic fields and the current are E = E i ei B = B i ei = J i ei J Define equivalent forms in arbitrary coordinates by = E i gij dxj = Ej dxj 1 i B eijk dxj ∧ dxk β = 2 κ = J i gij dxj  Show that Maxwell’s equations.

Substitute this result into the remaining homogeneous equation. Start with the magnetic equation dβ = 0 Then the converse to the Poincaré lemma shows immediately that β = dA for some 1-form A. 1 ∂ dε + dA = 0 c ∂t   1 ∂ d ε+ A = 0 c ∂t A second use of the converse to the Poincaré lemma shows that there exist a 0-form −ϕ such that ε+ 1 ∂ A = −dϕ c ∂t and therefore ε = −dϕ − 16 1 ∂ A c ∂t . Using the homogeneous Maxwell equations dβ = 0 1 ∂β = 0 dε + c ∂t show that the electric and magnetic fields arise from a potential. d  = dβ 1 ∂β dε + c ∂t 1∂  ∗ ∗ d β− c ∂t = 4π ρ c 0 = 0 = 4π κ c ∗ ∗ show that 1 ∂ ρ +∗ d∗ κ = 0 c ∂t Show that this equation is the continuity equation by writing it in the usual vector notation.= = = = = = = 1 ∂ 1 i ∂Ej dxm ∧ dxj + B eijk dxj ∧ dxk m ∂x c ∂t2 ∂ ∂ 1 ∂ 1 i 1 Ej − Em dxm ∧ dxj + B eijk dxj ∧ dxk 2 ∂xm ∂xj c ∂t 2   1 ∂ 1 ∂ 1 i ∂ inl E − E eijk dxj ∧ dxk + B eijk dxj ∧ dxk l n e 4 ∂xn ∂xl c ∂t 2     ∂ 1 1 ∂ 1 ∂ i inl E − E e + B eijk dxj ∧ dxk l n 2 2 ∂xn ∂xl c ∂t   1 inl 1 ∂ i e ∂n El + B eijk dxj ∧ dxk 2 c ∂t  i 1 1 ∂ ∇×E+ B eijk dxj ∧ dxk 2 c ∂t 0 From Maxwell’s equations.