Arxiv:1807.07401V2 [Hep-Th] 27 Jul 2018 This Type of Approach Has Been Advocated at Least Tion of Dirac’S String Prescription from the Dual Poten- Since [5]

Variations on the Dirac string

Brad Cownden∗ Department of Physics & Astronomy, University of Manitoba, Winnipeg, Manitoba R3T 2N2, Canada

Andrew R. Frey† Department of Physics, University of Winnipeg, 515 Portage Ave, Winnipeg, Manitoba R3B 2E9, Canada and Department of Physics & Astronomy, University of Manitoba Winnipeg, Manitoba R3T 2N2, Canada (Dated: July 31, 2018) Dirac’s original solution of the nontrivial Bianchi identity for magnetic monopoles [1], which re- deﬁnes the ﬁeldstrength along the Dirac string, diagonalizes the gauge and monopole degrees of freedom. We provide a variant of the Dirac string, which we motivate through a formal expansion of the Bianchi identity. We show how to use our variant prescription to study monopole electrodynamics without reference to a dual potential and provide some applications.

I. INTRODUCTION mathematically rigorous, treating the potential in this manner mixes the gauge and monopole degrees Magnetic monopoles, while experimentally elusive [2], of freedom because the overlap region moves with hold a special place in particle physics; they complete the monopole. This formalism also obscures the electric-magnetic duality, provide a reason for charge coupling between the potential and monopole (for quantization, and arise in the spontaneous breaking of example, [7] switch to a dual potential formalism to many grand unification models. At the same time, there derive the monopole’s classical equation of motion, is a long tradition of reformulating the description of while [8] following [9] combined Dirac’s formalism the monopole’s interaction with the electromagnetic field below with the rigorous gauge patching procedure). since Dirac’s original work [1, 3]: • The potential in a single coordinate patch extends • Monopoles arise as solitons of broken gauge the- to all of spacetime except for a half-line singular- ories; the microscopic description as a semiclassi- ity extending from the monopole (Dirac’s famous cal field configuration describes the monopole’s in- string singularity). In [1], Dirac noted that the elec- teractions but requires tracking the full degrees of tromagnetic fieldstrength can be written in terms of freedom of that gauge theory. While of course nec- a globally defined potential and an extra term sup- essary for processes like monopole creation and an- ported on the string singularity. This approach sep- nihilation (see eg [4]), it may be computationally arates the electromagnetic and monopole degrees of excessive when an effective description is valid. freedom and elucidates the monopole-fieldstrength coupling. However, it suffers some conceptual diffi- • In the absence of electric charges, the fieldstrength culties, including a singular fieldstrength along the can be defined in terms of a dual potential ?F ≡ (arbitrarily chosen) string and a constraint that dA˜, which couples to monopoles in the same way electric charges cannot intersect the Dirac string the vector potential couples to charges. Of course, worldsheet. Pragmatically, the semi-infinite string if the fields of both charges and monopoles are of can be awkward. interest, the fieldstrengths from both the potential In this paper, we present a new hybridization of Dirac’s and dual potential must be superposed in a “demo- string formalism with the rigorous gauge patching pre- cratic” formalism (which makes it ill-suited to the scription by allowing the Dirac string to end on an un- quantum mechanics of charges and monopoles). physical reference monopole, and we review the deriva- arXiv:1807.07401v2 [hep-th] 27 Jul 2018 This type of approach has been advocated at least tion of Dirac’s string prescription from the dual poten- since [5]. tial formalism with an emphasis on the origin of the • As pointed out originally in [6, 7], the vector poten- string. We also advocate for a particular configuration tial is not a globally defined function but a section for the string and give a simple physical picture for its of a fiber bundle. When the fieldstrength has a origin. The fixed prescription for the string means that nontrivial Bianchi identity, the potential is defined the string worldsheet embedding coordinated depend on in at least two coordinate patches; the potentials the monopole position along the entire worldsheet, unlike in the two patches are related by a gauge transfor- Dirac’s original formalism; our approach presents compu- mation in the overlap of the two patches. While tational advantages in several circumstances. The exten- sion of our results to higher-dimensional monopole-like branes appears in the companion paper [10] by one of us. Throughout this paper, we work in 4-dimensional ∗ [email protected] Minkowski spacetime for simplicity. We list our conven- † [email protected] tions, particularly signs, in an appendix. 2

II. DIRAC’S STRING VARIABLES Dirac furthermore locates the worldline M∗ at spatial infinity, so the string becomes semi-infinite. However, we The Maxwell equations and Bianchi identities with are free to specify the configuration of the string; a simple magnetic sources are choice in a given reference frame is to choose the reference monopole to be stationary (M∗ at constant position d ? F = − ? j , dF = − ? ˜ , (1) X~ ∗) and the string to be the straight line from X~ ∗ to the where jµ is the electric current and ˜µ the magnetic cur- monopole position X~ (t) at each time. We can build an ar- rent. For a single monopole, the current is bitrary worldsheet in this way by then extending a Dirac Z string from M∗ to another reference worldline M∗∗, and µ µ 4 µ 3 ˜ = g dτ∂τ X (τ)δ (x−X) = gu (t)δ (~x−X~ ) , (2) so on, and then letting the reference worldlines approach M each other. On the other hand, with a single line seg- where g is the monopole charge, M is the monopole’s ment, the worldsheet embedding coordinates Y µ depend worldline, τ the worldline time, and Xµ the monopole on the monopole position at all worldsheet at all world- position. The second equality follows in a fixed reference sheet points τ, σ unlike in Dirac’s formalism. We moti- frame with a worldline static gauge τ = t; uµ is the vate this prescription for the Dirac string configuration monopole’s 4-velocity. in section IV. Dirac’s key observation stems from the fact that any It is also instructive to consider the Dirac quantization conserved current (in Minkowski spacetime) can be writ- of charge for different formalisms. With distinct gauge ten as the divergence of a two-form, or ?˜ = d?H for some patches, quantization enforces the requirement that a H. In fact, the field of an isolated monopole provides single-value wavefunction in one coordinate patch re- such an H subject to the additional constraint of the mains single-valued after the gauge transformation in sourceless Maxwell equations. Instead, following Dirac passing to another patch. In Dirac’s original formalism, [1], we define there is only a single gauge, but the wavefunction picks Z up a phase under motion of the string equal to g times the Gµν (x) ≡ dY¯ µ ∧ dY¯ ν δ4(x − Y ) , (3) electric flux through the surface swept out by the world- N sheet — the surface is noncontractible since the charge where N is the worldsheet of a string with bound- cannot intersect the worldsheet, as we see below. Since the surface swept out by the string is closed (including ary ∂N = M − M∗ (M∗ is a so-far arbitrary worldline), Y µ(τ, σ) is the string position at worldsheet co- the point at infinity), charge quantization is given by the condition that the wavefunction remain single-valued if ordinates τ, σ with σ increasing from M∗ to M, and ¯ we sweep the string around the charge. When the refer- d = dτ∂τ + dσ∂σ is the worldsheet exterior derivative. This has divergence ence monopole is at a finite position, both mechanisms Z for charge quantization are possible. In particular, if the µ ¯ ν ¯ µ 4 (? d ? G) = dY ∧ dY ∂ν δ (x − Y ) string worldsheet is arbitrary but M∗ is fixed, we can N sweep the string in a closed surface around a charge, as Z per Dirac. With a set prescription for the worldsheet like = − d¯δ4(x − Y )dY¯ µ N discussed above, however, the worldsheet only moves if µ µ we move the reference position X~ , so the string sweeps = (1/g)˜ − (1/g)˜∗ (4) ∗ out open surfaces, removing the quantization condition. 1 in terms of the currents of monopoles along M and M∗. On the other hand, the potential must be defined in As a result, the field strength defined as F = dA − g ? G patches around the reference point, so single-valuedness automatically solves the Bianchi identity for the dynami- of the wavefunction still leads to charge quantization. cal monopole along M as long as A includes the potential The Dirac string formalism also provides a direct for a monopole with reference worldline M∗ (with appro- means of finding the monopole equation of motion, which priate gauge patching). is otherwise carried out indirectly through the dual po- It is worth pausing now to discuss the configuration of tential formalism. Including the string coupling G in the the string. In Dirac’s formalism, the string is a dynam- fieldstrength, the variation of the Maxwell action with ical object with no kinetic term (in modern parlance, respect to the monopole and worldsheet positions is a tensionless string), so the worldsheet N is completely arbitrary except for the specification of its boundary.2 g Z δS = d4xF µν (x)δGλρ(x) (5) Max 4 µνλρ Z Z g 4 µν λ ρ α 4 = d x F (x) dY¯ ∧ dY¯ δY ∂ α δ (x − Y ) 1 An additional sign enters at the last equality from the worldsheet µνλρ Y 4 N orientation as described in the appendix. ¯ λ ¯ ρ ¯ λ ¯ ρ 4 2 This arbitrariness reflects how a gauge transformation modifies + dδY ∧ dY + dY ∧ dδY δ (x − Y ) . where the potential of a monopole becomes singular (or more precisely the patching required for the potential), so the string We can now restrict to a worldsheet integral, first con- can never develop a tension, even quantum mechanically, for an verting the derivative on the delta function to one with unbroken gauge symmetry. respect to xα and integrating by parts. Then, with all 3 partial derivatives now with respect to Y , potential, simplified from versions presented in [11, 12]. Z In particular, we start with the dual fieldstrength and po- g ¯ µν λ ¯ ρ δSMax = µνλρ d F (Y )δY dY (6) tential and a magnetic current in the absence of electric 2 N charges (which lead to a Dirac string for the dual field- 1 strength) in order to emphasize that the Poincaréduality + ∂ F µν (Y ) dY¯ α ∧ dY¯ λδY ρ 2 α itself leads to the Dirac string for the fieldstrength in the presence of monopoles. ¯ λ ¯ ρ α ¯ ¯ α λ +dY ∧ dY δY + dY ρ ∧ dY δY . In form notation, the action for the dual electromagnetism with a single monopole is The latter lines of (6) reorganize to Z 1 Z S = − F˜ ∧ ?F˜ + A˜ ∧ ?˜ , (9) g ¯ ν ¯ λ ρ 2 (d ? F )νλρdY ∧ dY δY , (7) 2 N where A,˜ F˜ are the dual potential and fieldstrength and which would yield an interaction between the (unphys- ˜ is the monopole current (2). To dualize back to the ical) string — and therefore the monopole — with the “usual” potential in the absence of monopoles, we treat electric current when the gauge fields are on shell. To F˜ as the independent variable and add a Lagrange mul- avoid this unphysical result, Dirac imposed the additional tiplier term AdF˜ to enforce the Bianchi identity for F˜ condition that charges not intersect the string. The first since A˜ does not appear. Solving the equation of motion term, on the other hand, gives an integral over M (as- for F˜ and substituting gives the usual Maxwell action. 3 suming M∗ fixed), so it combines with the variation of With the monopole current, we must first find a way the monopole’s kinetic term to give the magnetic Lorentz to eliminate the dual potential from (9). We proceed force equation by recalling that any conserved current can be written

ν as the divergence of a 2-form (in other words, any co- ∂τ pµ = −g(?F )µν ∂τ X . (8) closed form in Minkowski spacetime is co-exact). In fact, ˜ Meanwhile, the electric charge couples to the redefined the Maxwell equation d ? F = ?˜ shows that ˜ can be potential A, so its equation of motion turns out as usual written in terms of the monopole’s field strength (which ˜ except for contact terms with the Dirac string, which are also satisfies dF = 0). We have also seen above that the forbidden due to Dirac’s condition. Dirac string coupling G is another such two form up to It is important to note that our derivation of the the subtraction of a current along the reference worldline monopole equation of motion treated the potential A M∗, so we can write as an independent degree of freedom in contrast to the Z 1 ˜ ˜ ˜ ˜ case without a Dirac string and with gauge patches. In S = − F ∧ ?F + A ∧ d ? (gG + F∗) 2 that case, there is no coupling between the monopole and Z gauge field, only a hidden dependence of A on Xµ which 1 0 = − F˜ ∧ ?F˜ + F˜ ∧ ?(gG + F˜∗) − A ∧ dF˜ (10), should be treated as an explicit dependence. The Dirac 2 string removes this explicit dependence from the vector ˜ potential, but A does still have a “hidden” explicit depen- where F∗ is the dual fieldstrength of the reference dence on the arbitrary reference position. The equation monopole and we have added the Lagrange multiplier ˜ 0 of motion for the reference position is F µν δF /δXλ = 0, term. This last term can be rewritten as F dA up to µν ∗ ˜ which enforces the condition that the fieldstrength is in- a total derivative, so the equation of motion is ?F = ˜ 0 dependent of the reference position. This condition de- ?(gG + F∗) − dA . The action is therefore classically µ equivalent to termines the explicit dependence of the potential on X∗ . Z 1 S = − dA0 − ?F˜ − g ? G ∧ ? dA0 − ?F˜ − g ? G . (11) 2 ∗ ∗ III. THE STRING FROM THE DUAL POTENTIAL We recognize − ? F˜∗ = F∗ as the reference monopole’s 0 field strength, so we define dA + F∗ = dA in terms of a The dual potential formalism is useful in many ap- potential A with the appropriate gauge patches for the plications since it translates the electrodynamics of reference monopole. We are left with F = dA − g ? G monopoles into the more familiar electrodynamics of and the usual Maxwell action − R F?F/2 including the charges (and can even allow for the interaction of charges Dirac string. and monopoles since electromagnetism is linear). Here we review a derivation of the Dirac string from the dual IV. AN INTERPRETATION OF THE STRING

We can gain new insight into the Dirac string through 3 with sign determined per footnote 1. the Bianchi identity. For clarity, we pick a reference 4 frame and work in static gauge for the monopole (ie, choose a static gauge5 t = Y 0(τ, σ) = τ and Y~ (τ, σ) = 0 worldline time τ = Y ). In non-relativistic notation, the X~ ∗ + σδX~ (t) with 0 ≤ σ ≤ 1. Then the Dirac string Bianchi identity is coupling becomes

3 3 Z 1 ∇~ · B~ = gδ (~x − X~ ) , ∇~ × E~ + ∂tB~ = −g∂tXδ~ (~x − X~ ) . 0i i 3 G = dσ δX δ (~x − X~ ∗ − σδX~ ) , (16) (12) 0 Our goal is to diagonalize the degrees of freedom, remov- ∞ 1 Z 1 i X n,i1···in ∗ n 3 n ing the explicit dependence of Aµ on the monopole posi- = δX (δX~ ) (∇~ ) δ (~x − X~ ∗) dσ σ n! i1···in tion. We carry out a formal expansion of the monopole n=0 0 current around a static reference monopole at fixed posi- and similarly tion X~ ∗; therefore, we end up with a static gauge patch- ~ ∞ ing procedure around X∗ and will see that the remaining X 1 Gij = (∂ XiδXj − ∂ XjδXi) (δX~ )n,i1···in terms can be organized as a contribution to the field- t t n! strength.4 n=0 We begin by writing Z 1 × (∇~ ∗)n δ3(~x − X~ ) dσ σn+1 (17) i1···in ∗ 3 3 0 δ (~x − X~ ) = δ (~x − X~ ∗) (13) ∞ after expanding the delta function. Carrying out the 1 X i1 in ∗ ∗ 3 + δX ··· δX ∇ · · · ∇ δ (~x − X~ ∗) , integral and using the usual relation between the field- n! i1 in n=1 strength Fµν and fields E,~ B~ , we find (14,15). So the particular solution of the Bianchi identity that ~ ∗ ~ ~ where ∇ is the gradient with respect to X∗ and δX = we found by expanding the delta function around X~ is a ~ ~ ∗ X − X∗. Converting one of the derivatives to one with particular realization of the Dirac string. As a result, we respect to ~x, the terms in the sum can be written as a see that the expansion of the delta function separates the divergence, so defining explicit dependence of the fieldstrength on the monopole ~ ~ ~ position from the potential. This also tells us that the B = ∇ × A (14) Dirac string is a way of treating a monopole’s motion ∞ 1 as a fluctuation (even a large one) around a fixed posi- X n,i1···in ∗ n 3 − gδX~ (δX~ ) (∇~ ) δ (~x − X~ ∗) (n + 1)! i1···in tion. The key difference with Dirac’s arbitrary string is n=0 that this interpretation suggests treating the embedding where A~ includes the potential of a fixed monopole at coordinates of the string as dependent on the monopole position along the entire worldsheet. In fact, we have X~ solves the scalar equation of (12). Subtracting this ∗ done this explicitly in deriving (16,17). contribution from the vector part of the Bianchi identity For future reference, it is useful to give the exact ex- similarly extracts a curl from the expansion of the delta pressions for the fields with the linear string configura- function. We can therefore define tion: ~ ~ ~ ~ ~ E = −∇Φ − ∂tA + g(∂tδX × δX) (15) Z 1 ∞ E~ = −∇~ Φ − ∂ A~ + g ∂ X~ × δX~ dσ σδ3 ~x − X~ − σδX~ n + 1 t t ∗ X n,i1···in ∗ n 3 0 × (δX~ ) (∇~ ) δ (~x − X~ ∗) (n + 2)! i1···in Z 1 n=0 3 B~ = ∇~ × A~ − gδX~ dσ δ ~x − X~ ∗ − σδX~ . (18) to solve the Bianchi identity. 0 To confirm that this approach separates the gauge and The displacement D~ and field H~ are determined as usual monopole degrees of freedom, we can vary the Maxwell ~ ~ ~ from E, B in linear media, including the Dirac string con- action with respect to X and, with some manipulation, tribution. find the magnetic Lorentz force equation by treating the potentials as independent variables. As with the Dirac string, the variation of the action also includes terms pro- V. THE STRING AS A SOURCE portional to the sourceless Maxwell equations and deriva- 3 ~ tives of δ (~x − X∗). The Dirac string in the fieldstrength automatically As an alternate approach, we can consider the Dirac solves the Bianchi identity, so the Bianchi identity no ~ string with M∗ a static worldline at X∗. Then, as we longer determines the fieldstrength associated with the suggested in the section above, take N at any fixed time dynamic monopole. Instead, as Dirac realized, the string to be the line segment from X~ ∗ to X~ . Then we can

5 The argument goes through with little change for any function 4 This is the approach taken in [13] for D3-branes. of τ. 5 coupling G leads to an effective electric current as a Specifying a string configuration that depends explic- source for dA, the fieldstrength away from the Dirac itly on the monopole position along the worldsheet gives string. In particular, d ? dA = −gdG, so the effective a well-defined effective current from the string. In prin- current is jeff = g ? dG. ciple, we can now solve for the electromagnetic fields For practical use, we again choose a line segment con- for arbitrary monopole motion as a superposition of the figuration as in section IV. The effective charge is magnetic field from the fixed reference monopole and the effective current. The linear string configuration seems 0 ρeff = jeff (19) particularly well suited to this type of calculation. Z 1 3 = −g ∂tX~ × δX~ · ∇~ dσ σδ (~x − X~ ∗ − σδX~ ) , 0 VI. APPLICATIONS where the gradient is with respect to ~x. Evaluating the effective current in the static gauge is slightly subtler. As we have noted previously, there are technical diffi- We have culties in the theory of electrodynamics with monopoles. Z Z 1 Here we present several possible applications in which the j k jeff,i = −gi0jk dτ dσ (σ∂τ X δX ) linear Dirac string configuration yields simplifications. 0 We indicated above that one such application is a di- h 3 i ×∂t δ(t − τ)δ (~x − X~ ∗ − σδX~ ) rect determination of radiation from moving monopoles, including energy loss in dielectric materials (such as Z 1 6 k 3 Cherenkov radiation), which may be useful for monopole +gij0k dσ δX ∂jδ (~x − X~ ∗ − σδX~ ) .(20) 0 search experiments. While an infinite, arbitrarily moving Dirac string is unwieldy, the linear Dirac string config- We must take care to convert the t derivative to a τ uration gives a well-defined current, which is simply a derivative and integrate by parts where the delta function growing or shrinking solenoid in many contexts. In the in time is differentiated. We are left with presence of materials, as noted, it is necessary to write Z 1 the fields nonrelativistically as E,~ B~ and include the per- ~ ~ 3 ~ ~ ~eff = g∂t ∂tX × δX dσ σδ (~x − X∗ − σδX) mittivity and permeability in defining D,~ H~ . 0 Z 1 We have emphasized that the Dirac string formulation 3 −gδX~ × ∇~ dσ δ (~x − X~ ∗ − σδX~ ) . (21) separates the gauge and monopole degrees of freedom; as 0 a result, it provides a basis for Hamiltonian and there- This is worth two comments. First, the current for fore quantum treatments. In this form, the introduction of extra unphysical degrees of freedom for the Dirac arbitrary linear motion of the monopole, with X~ chosen ∗ string leads to constraints [16]. On the other hand, when to lie on the line, is the Dirac string takes the linear configuration, the en- Z 1 tire string depends on the monopole position. In the 3 ~eff = −gδX~ × ∇~ dσ δ (~x − X~ ∗ − σδX~ ) . (22) variation of the action, these appear through terms pro- 0 portional to the sourceless Maxwell equations (see equation (7)) and are trivial on-shell. On the other hand, ~ Taking δX to lie along the ±z axis, (22) is the surface these terms can contribute off-shell, for example in the current in the ±φ direction of an infinitely tightly wound, path integral. Understanding how these contribute to the ~ ~ infinitesimally thin solenoid stretching from X∗ to X. quantum mechanics of monopoles is an interesting ques- The direction and magnitude of the current are precisely tion. Meanwhile, the analogous terms for D3-branes (see ~ such that the flux into the solenoid at X∗ (which is spher- the discussion below) also play an important role in the ically symmetric since the solenoid is thin) precisely can- 4-dimensional effective action of type IIB string theory cels the magnetic flux due to the reference monopole at [10, 13]. X~ ∗. Similarly, the flux out of the solenoid at X~ pre- The Dirac string formalism straightforwardly extends cisely reproduces the flux of the dynamical monopole. to curved spacetimes and higher dimensions, and the lin- Second, the current is conserved, as it must be, if we ear configuration for the string becomes a worldsheet allow for singular distributions. The solenoidal current with geodesics as constant-time slices. A second endpoint (22) is clearly conserved, and the other terms take the to the string on M∗ therefore allows us to use the Dirac ~ ~ form ρeff = −∇~ · k,~eff = ∂tk, which is manifestly con- string for monopoles on compact manifolds. While the served. magnetic Gauss law constraint (net magnetic charge on a In media, the effective charge and current follow from compact manifold vanishes) means that any Dirac string ∇~ · D~ and ∇~ × H~ , so derivatives of the permittivity and can end on oppositely charged physical monopoles, an permeability will appear in general. For piecewise constant dielectric constant and permeability, it is natural to define the effective charge and current in a piecewise manner as well. 6 See [14, 15] for the energy loss rate in different approximations. 6 arbitrary reference endpoint allows for a cleaner separa- Theoretical Physics. Research at Perimeter Institute is tion of the dynamics of the different monopoles. Further- supported by the Government of Canada through the more, it allows us to avoid having multiple Dirac strings Department of Innovation, Science and Economic Devel- end on one monopole in the case that the monopoles on opment and by the Province of Ontario through the Min- a compact manifold carry different numbers of magnetic istry of Research and Innovation. quanta (for example, there are two monopoles of charge +1 and one of charge −2). Again, in the case of higher- dimensional monopole-like branes, monopole charge can Appendix: Conventions dissolve into the flux of other fields, so the Dirac string from a monopole may not even have another monopole Here we briefly lay out our conventions, includ- on which to end, necessitating the reference endpoint. ing signs. To start, we take the mostly plus met- Finally, higher-dimensional branes of string theory are ric convention with 0123 = +1. The Hodge star magnetic sources for various rank form fields. As in elec- for a differential p-form is given by (?F )µ1···µ4−p = ν1···νp p(4−p)+1 tromagnetism, a mathematically rigorous treatment con- (1/p!)µ1···µ4−p Fν1···νp , so ??F = (−1) . siders the potentials as sections, while a Dirac-like for- With standard conventions (see [14, 17]), the Maxwell malism allows the separation of the gauge and brane de- equations with magnetic currents included are grees of freedom (see [13]; one of us will detail this for- ~ ~ ~ ~ ~ malism in the forthcoming [10]). The formalism therefore ∇ · E = ρ , ∇ × B − ∂tE = ~ provides an alternative to the democratic (dual potential) ∇~ · B~ =ρ ˜ , ∇~ × E~ + ∂tB~ = −~˜ , (A.1) formalism for the derivation of brane equations of motion. It is particularly helpful for a careful accounting of where ρ,~ are the electric charge and current andρ, ˜ ~˜ degrees of freedom, as needed in dimensional reduction. are the magnetic. In relativistic notation, we take Aµ = Determining the lower-dimensional effective action is also (Φ, A~), jµ = (ρ,~) (and likewise for the magnetic cur- an off-shell calculation, and extra terms from the gener- rent), so F0i = −Ei,Fij = ijkBk. The Maxwell equa- alized Dirac string worldvolume (which vanish on-shell) tions become are critical to account for all the kinetic terms required µν ν ρ by supergravity [10, 13]. ∂µF = −j , ∂µFνλ + ∂ν Fλµ + ∂λFµν = −µνλρ˜ , (A.2) or d ? F = − ? j and dF = − ? ˜ in terms of forms. The dual field strength F˜ ≡ ?F therefore satisfies the dual Maxwell equations dF˜ = − ? j and d ? F˜ = + ? ˜. As a ACKNOWLEDGMENTS result, the dual electric current is usually defined as −˜; for simplicity of comparison, we do not introduce this AF would like to thank K. Dasgupta, N. Afshordi, and sign. L. Boyle for interesting discussions. BC and AF are sup- Finally, we define the Dirac string coupling G as a ported by the Natural Sciences and Engineering Research form integral over the string worldsheet coordinates τ, σ. Council of Canada Discovery Grant program. Part of We choose the orientation by taking integration measure this work was supported by the Perimeter Institute for d2σ = dτ ∧ dσ = −dσ ∧ dτ.

[1] Paul A. M. Dirac, “The theory of magnetic poles,” Phys. [7] Tai Tsun Wu and Chen Ning Yang, “Dirac’s monopole Rev. 74, 817–830 (1948). without strings: classical Lagrangian theory,” Phys. Rev. [2] B. Acharya et al. (MoEDAL), “Search for magnetic D14, 437–445 (1976), [509 (1976)]. monopoles with the MoEDAL forward trapping de- [8] K. Hirata, “Classical Lagrangian theory of Dirac’s tector in 2.11 fb−1 of 13 TeV proton-proton colli- monopole: avoiding Dirac’s veto,” Phys. Lett. 81B, 169– sions at the LHC,” Phys. Lett. B782, 510–516 (2018), 172 (1979). arXiv:1712.09849 [hep-ex]. [9] Richard A. Brandt and Joel R. Primack, “Avoiding [3] Paul A. M. Dirac, “Quantized singularities in the elec- Dirac’s veto in monopole theory,” Phys. Rev. D15, 1798– tromagnetic field,” Proc. Roy. Soc. Lond. A133, 60–72 1802 (1977). (1931), [,278(1931)]. [10] Andrew R. Frey, “Dirac branes for Dirichlet branes,” [4] Tanmay Vachaspati, “Creation of magnetic monopoles (2018), in preparation. in classical scattering,” Phys. Rev. Lett. 117, 181601 [11] Stanley Deser, A. Gomberoff, M. Henneaux, and C. Teit- (2016), arXiv:1607.07460 [hep-th]. elboim, “P-brane dyons and electric magnetic dual- [5] F. Rohrlich, “Classical theory of magnetic monopoles,” ity,” Nucl. Phys. B520, 179–204 (1998), arXiv:hep- Phys. Rev. 150, 1104–1111 (1966). th/9712189 [hep-th]. [6] Tai Tsun Wu and Chen Ning Yang, “Dirac monopole [12] K. Lechner and P. A. Marchetti, “Duality invariant quan- without strings: monopole harmonics,” Nucl. Phys. tum field theories of charges and monopoles,” Nucl. Phys. B107, 365 (1976). B569, 529–576 (2000), arXiv:hep-th/9906079 [hep-th]. 7

[13] Brad Cownden, Andrew R. Frey, M. C. David Marsh, Cim. 37, 145 (1983). and Bret Underwood, “Dimensional reduction for D3- [16] A. P. Balachandran, R. Ramachandran, J. Schechter, brane moduli,” JHEP 12, 139 (2016), arXiv:1609.05904 Kameshwar C. Wali, and Heinz Rupertsberger, “Hamil- [hep-th]. tonian formulation of monopole theories with strings,” [14] John David Jackson, Classical electrodynamics (Wiley, Phys. Rev. D13, 354 (1976). 1975 (2nd ed), 1998 (3rd ed)). [17] David J. Griﬃths, Introduction to electrodynamics (Pear- [15] J. Gea-Banacloche, Kevin E. Cahill, and D. Rossbach, son, 2012 (4th ed)). “Energy loss by slow magnetic monopoles,” Lett. Nuovo