Revisiting the Coupling of Thermodynamics and Electromagnetics

Henning Struchtrup; Manuel Torrilhon; Stefanie Braun

REVIEW 1 major objections 5 minor 14 references

Revisiting the Coupling of Thermodynamics and Electromagnetics

T0 review · 1 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Two independent derivations of the equations for moving polarisable and magnetisable matter produce the same structure once polarisation and magnetisation are redefined.

desk verdict Completing Mazur's route, the structural-equivalence claim is solid but rests on a mass-correction estimate only half-proved; referee time, not a desk reject. read the letter →

arxiv 2608.07595 v1 pith:NKVEC3AM submitted 2026-08-06 cond-mat.stat-mech

Stefanie Braun , Henning Struchtrup , Manuel Torrilhon This is my paper · ORCID

classification cond-mat.stat-mech

keywords thermodynamicselectromagneticsconservationlawspolarisationmagnetisationentropyprinciplestatisticalmechanicsmovingmedia

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper tries to settle how the balance laws of thermodynamics should be coupled to Maxwell's equations when matter moves, polarises, and magnetises. It does this by taking two routes that start from opposite ends—an axiomatic bulk theory closed by an entropy principle, and an ensemble-averaging derivation from individual charge carriers—and showing that, after a redefinition of polarisation and magnetisation, the two sets of equations agree structurally. Along the way the authors establish that the source of the internal energy balance must be the invariant pairing of the non-convective current with the electromotive intensity, and that a single identity fixes both the admitted entropy variables and the signs of the bound-current ansatz. If correct, this means the apparent disagreements between these frameworks are mostly bookkeeping, with one genuine remainder: a momentum contribution from microscopic field fluctuations that no purely macroscopic theory can produce.

What carries the argument

The central object is the vector identity (26), which the paper derives from the definition of the bound current together with the Maxwell equation for $B$; in the paper's notation it reads $\rho D_t(M\cdot\hat B)+\nabla\cdot(\mathcal{E}\times M)=-\sigma_{PM}:\nabla v+\rho\mathcal{E}\cdot D_t\hat P+\rho\hat B\cdot D_t M-\mathcal{E}\cdot J_P$. The identity is a pure consequence of definitions and carries no extra physical input, yet it reorganises the energy balance, selects the entropy variables, and converts the sign ambiguity in the bound-current ansatz into a question that the entropy principle can answer. On the statistical side, the carrying method is Mazur's ensemble average over the phase space of atoms with internal charge carriers, which yields exact microscopic definitions of polarisation $P$, quadrupole moment $Q$, and magnetisation $M$; the paper extends this averaging to mass, momentum, and energy, and estimates the second-order mass corrections via the ratio $|1/(\Omega_{c,ki}t)|$ between the thermodynamic time scale and the cyclotron gyration period.

What would settle it

Compute the ratio $|1/(\Omega_{c,ki}t)|$ for a concrete system, for example electrons in a $1\,\text{T}$ field undergoing a nanosecond thermodynamic transient, and check whether it is genuinely small; if it is not small, the neglected mass corrections are comparable to the magnetisation terms and the claimed structural agreement fails in that regime. Alternatively, evaluate the ensemble average $\langle\epsilon_0 e'\times b' f\rangle$ for a simple model such as a dilute gas of harmonic oscillators in a uniform magnetic field and see whether the surviving fluctuation momentum is negligible; if it is large, the remaining difference between the two theories is physically relevant.

Watch

Extended reading notes

Core claim

Stated on the paper's own terms, the discovery is that the Dreyer–Guhlke–Müller bulk theory and the theory obtained by completing Mazur's statistical-mechanical route describe the same physical system. With the redefinitions $P_{\text{Dreyer}}=P_{\text{Mazur}}-\nabla\cdot Q_{\text{Mazur}}$ and $M_{\text{Dreyer}}=M_{\text{Mazur}}-\langle\sum_k v'_k\times(\mu^{\text{el}}_k-\nabla\cdot Q_k)\delta(R_k-R)f\rangle$, the continuity, momentum, energy, and Maxwell equations of the two routes coincide up to second-order mass corrections that the paper estimates as negligible. The single structural difference that cannot be removed is the microscopic field-fluctuation momentum $\langle\epsilon_0 e'\times b' f\rangle$, which a macroscopic theory cannot reproduce. The paper further claims that the electromotive intensity $\mathcal{E}$ and the Lorentz magnetisation $\mathcal{M}$ are not modelling choices: they appear necessarily in the energy balance and in the entropy flux, and the identity $\rho D_t(M\cdot\hat B)+\nabla\cdot(\mathcal{E}\times M)=-\sigma_{PM}:\nabla v+\rho\mathcal{E}\cdot D_t\hat P+\rho\hat B\cdot D_t M-\mathcal{E}\cdot J_P$ determines which entropy variables are admissible and forces the bound-current signs to $\lambda_P=\lambda_M=+1$ through the second law, concavity, and positive susceptibility.

Load-bearing premise

The comparison rests on the estimate that thermodynamic processes are slow enough that the mass-correction terms, which scale like $1/(\Omega_{c,ki}t)$ relative to the magnetisation terms, can be dropped; if that time-scale ordering fails—under very strong fields or very fast transients—the two sets of conservation laws no longer coincide.

Editorial extensions

If this is right

The bound-current signs are fixed by the second law: with a concave entropy and positive susceptibility, only $\lambda_P=\lambda_M=+1$ is thermodynamically admissible, so a wrong sign in the ansatz is ruled out rather than being a free choice.
The energy source is pinned down: only the pairing of the non-convective current with the electromotive intensity, $J_e\cdot\mathcal{E}$, is observer-invariant, and it is the one that survives after the mechanical power of the Lorentz force is subtracted.
The electromotive intensity and the Lorentz magnetisation are outputs, not inputs: they appear in the energy balance, in the entropy flux $\phi=(q+\mathcal{E}\times M)/T$, and in the entropy production without being inserted by hand.
The two routes agree: after redefining polarisation and magnetisation, the Dreyer–Guhlke–Müller equations and the Mazur-based equations have the same structure, with only negligible second-order mass corrections and the irreducible field-fluctuation momentum left over.
The asymmetry of the entropy is bookkeeping: choosing $u+M\cdot\hat B$ as the energy variable makes the entropy look asymmetric in electric and magnetic variables, but changing the variable moves the asymmetry without altering the physics.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

Editorial inference: the identification $P_{\text{Dreyer}}=P_{\text{Mazur}}-\nabla\cdot Q_{\text{Mazur}}$ implies that macroscopic theories keeping only the dipole polarisation are silently dropping quadrupole contributions of the same formal order as the magnetisation; field-gradient and dielectrophoretic experiments would be a natural place to test whether that term matters.
Editorial inference: the sign-fixing argument could be probed by constructing a microscopic model with negative effective susceptibility and checking whether the Dreyer closure then violates the entropy inequality; the paper's logic predicts it must, but such a model would reveal whether concavity plus positive susceptibility is the right criterion.
Editorial inference: the surviving fluctuation momentum $\langle\epsilon_0 e'\times b' f\rangle$ is the continuum analogue of the term that drives the Abraham–Minkowski momentum debate, so estimating its size in a simple many-atom system would connect this structural difference to a long-standing experimental question about electromagnetic momentum in matter.
Editorial inference: because Mazur's exact polarisation balance contains unresolved fluxes involving $v'_k$, the Dreyer-style closure relations can be read as models of those fluctuations; this suggests a systematic programme of deriving macroscopic relaxation closures by computing $v'$-correlations from kinetic or molecular-dynamics models.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Braun-Struchtrup-Torrilhon is worth a serious referee. The paper earns its keep: it derives the conservation laws Mazur left out, writes down identity (26) that Dreyer et al. never wrote, and shows the DGM and Mazur routes match after the P/M redefinition (66)-(67). The redefinition is not hand-waving; it follows from the two theories' genuinely different definitions of P and M, and the paper says the equivalence is structural, not a claim that the microscopic quantities are identical. The sign-fixing from the entropy principle plus positive susceptibility is a genuine result, though it depends on the extra positive-susceptibility assumption, which the authors state explicitly.

The soft spots are in proportion. The Sec. 4.7 estimate (Eq. 63) is load-bearing for the structural-equivalence claim, and the stress-test note has a real point: Eq. (63) estimates only the first mass-correction term, using the centre-of-mass velocity R_k dot on the thermodynamic time scale L/t. The second term contains the internal charge-carrier velocity r_dot, which is not on that scale. For an electron, |r_dot|/r is of order 10^16 s^-1, while omega_c at 1 T is about 10^11 s^-1, so the ratio is omega_atomic/omega_c, not 1/(omega_c t). The paper says "a similar estimate with the same result holds for the second one" without demonstrating it. That is a gap, but not a fatal one: the r_dot terms appear in the expansion of the flux and sit next to terms that are absorbed into t, q, and u_int in Sec. 5.1. Whether the internal-velocity term is exactly absorbable or a pure divergence is not shown. A referee should ask for a rigorous bound or an absorption proof. The unestimated fluctuation term is honestly flagged as open; the impossibility claim is asserted, and it is plausible because the macroscopic theory has no microscopic fluctuation degrees of freedom, but it would be nice to see the size estimate the authors promise as open work.

The paper is honest about its limits and does not oversell. The algebra is not machine-checked, but the structure is clean and the identities are checkable by hand. I'd take the reader's CONDITIONAL verdict as roughly right, with the stress-test note downgraded from load-bearing to a moderate gap. Who benefits: anyone working on electrolyte models, moving-media electrodynamics, or continuum thermodynamics. I would bring it to a reading group. I would not cite it yet, because the mass-correction estimate is load-bearing in my view. A serious editor should send this to peer review.

Referee Report

1 major / 5 minor

Summary. This paper revisits two routes to the coupling of thermodynamics and electromagnetism: the macroscopic axiomatic theory of Dreyer, Guhlke and Müller and the statistical-mechanical ensemble-averaging route of Mazur. The authors show that in the Dreyer theory the internal-energy source must be built with the non-convective current paired with the electromotive intensity, derive the identity (26) that links the polarisation current, Lorentz magnetisation and entropy variables, and use it to argue that the signs of the bound-current ansatz are fixed by the entropy principle. In the Mazur part, they derive the missing mass, momentum and energy conservation laws from the microscopic phase-space setup and estimate the second-order mass-correction terms. The central claim is that, after the redefinitions (66) and (67) of polarisation and magnetisation, the two sets of equations agree structurally and that the only irreducible difference is the momentum contribution from microscopic field fluctuations, ⟨ε0 e′×b′ f⟩. The paper also discusses the status of the Lorentz magnetisation, the field-energy versus matter-energy representation, and the different nature of the polarisation balances in the two theories.

Significance. If the central structural-equivalence claim is established, the paper would provide a valuable bridge between a phenomenological entropy-based continuum theory and a microscopic statistical-mechanical derivation, and it would settle a sign ambiguity in the bound-current ansatz without introducing fitted parameters. The derivations in Sections 3 and 4 and in Appendix A are mostly careful, and the paper is explicit about which of its steps are identities and which are closures; the absence of free parameters and the transparency about modelling choices are clear strengths. However, the advertised equivalence rests on the order-of-magnitude estimate in Section 4.7, which is not fully justified for one of the two mass-correction terms. Because that estimate is load-bearing for the comparison in Section 5.1 and Table 1, the manuscript needs revision before the central claim can be accepted as stated.

major comments (1)

[Section 4.7, Eq. (63)] The estimate is carried out only for the first term in the mass-correction bracket. For the second term, −m_ki \dot r_ki (r_ki · ∇)δ, the natural ratio to the retained magnetisation term is m_ki |\dot r_ki| / (e_ki |b| r_ki) ≈ ω_atomic / Ω_c,ki, not 1/(Ω_c,ki t), because |\dot r_ki|/r_ki is an internal atomic frequency rather than the thermodynamic rate 1/t. For an electron in an atom, |\dot r_ki|/r_ki is of order 10^16 s^-1 while the cyclotron frequency at 1 T is of order 10^11 s^-1, so this ratio can be much larger than unity. The sentence 'a similar estimate with the same result holds for the second one' is therefore not demonstrated. Since Section 5.1 and Table 1 explicitly drop these terms to obtain structural agreement, the central claim is not established. The authors should provide a bound for the second term that accounts for the time derivative in Eq. (61) and for the ensemble average, or prove that the term is absorbable into the stress tensor or into a redefined momentum; otherwise the comparison should be weakened.

minor comments (5)

[Section 3.2, after Eq. (19)] The sentence 'The naive product J_e·E is not the correct source' is confusing because Eq. (19) states that J_e·𝓔 is the source; the intended contrast is presumably with j_e·E or with the pairing using the lab electric field rather than the electromotive intensity. Please rephrase to avoid the apparent contradiction.
[Section 5.2, Eqs. (67)–(68)] The displayed mappings are not algebraically equivalent: Eq. (68) omits the term −v×(P_Mazur − ∇·Q_Mazur) that appears in Eq. (67). If Eq. (68) is intended to define Dreyer's Lorentz magnetisation rather than the lab-frame magnetisation, this should be stated explicitly; as written, the 'or equivalently' is incorrect.
[Appendix A.4] The reference to Feynman appears as a placeholder '[?]' in the sentence about the relation between energy flow and momentum density; the missing citation should be supplied.
[Abstract and general formatting] Several phrases have missing spaces, for example 'electromotive intensityE' and 'Lorentz magnetisationM' in the abstract; these should be corrected in the final version.
[Section 4.7, Eq. (63)] The notation in the first numerator estimate appears to be missing the dot over R_k: the displayed expression reads 'mki |Rk|r^2/(t L^2)' where |\dot R_k| is evidently meant. Please fix the notation.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the Dreyer–Guhlke–Müller/Mazur equivalence is an openly declared redefinition of free potentials, with the structural content (convective magnetisation, bound-current signs, quadrupole offset) independently derived from the entropy principle and the microscopic averages.

full rationale

The paper's two central claims—structural equivalence of the Dreyer–Guhlke–Müller and Mazur-based theories, and entropy-principle fixing of the bound-current signs—are derived, not assumed. The comparison (Sec. 5) is an explicit equivalence transformation: Dreyer's polarisation and magnetisation are admitted free potentials (Sec. 5.4: '(22) is a kinematic identity ... P and M are potentials for n_P and J_P'), so the dictionary (66)–(68) declares their values in terms of Mazur's exact moments rather than fitting them to the conclusion. The load-bearing content is the forced structure: the convective Lorentz magnetisation M+v×P emerges from Mazur's ensemble averaging (Eq. 58) with the sign that Dreyer postulates, the quadrupole offset −∇·Q appears at the same expansion order as M, and the one unmatched term (⟨ε_0 e'×b'⟩) is retained as an irreducible difference rather than absorbed. The sign argument (Sec. 3.6) walks λ_P, λ_M ∈ {±1} through to an entropy production whose equilibrium conditions λ_P E = E_eq, λ_M B̂ = B̂_eq, concavity, and positive susceptibility select λ=+1; this is a conditional derivation, not an input renamed as output, although the abstract's 'fixes ... the signs' overstates the body's conditional conclusion. The mass-correction estimate (Eqs. 63–64) uses computed physical constants (cyclotron frequency), not fitted parameters; the body admits the second numerator term is only asserted to behave the same, and Sec. 6.2 lists the unquantified fluctuation term as open—these are rigor gaps in the equivalence claim, but a missing bound is not a circular step. No self-citations occur: all cited works ([1]–[14]) are external (Dreyer, Mazur, Jackson, Müller, Steigmann). The derivation chain is therefore self-contained: inputs are the microscopic model and the entropy axiom; outputs (identity (26), sign fixing, structural agreement) are not presupposed by those inputs.

Assumptions & free parameters 0 free parameters · 6 assumptions · 0 invented entities

The analysis introduces no new physical entities and fits no data. It relies on standard statistical-mechanical averaging, the Dreyer et al. entropy principle, and two physical assumptions: positive susceptibility and a time-scale separation that justifies dropping mass corrections. These assumptions are flagged in the paper but are not empirically calibrated.

assumptions (6)

standard math The microscopic Maxwell equations (14) are exact in SI units and the ensemble average in (47) commutes with spatial and temporal derivatives.
Used throughout Section 4 and Appendix A; without this the macroscopic Maxwell equations and the averaged conservation laws do not follow.
standard math The probability distribution f satisfies the Liouville equation (48).
Basis for the prototype conservation equation (72) from which mass, momentum and energy balances are derived.
domain assumption The delta function delta(Rki-R) is expanded to second order in |rki|/|Rk|, with first-order terms vanishing by (45) and higher orders neglected.
Used in Appendix A.2-A.4 to obtain equations (60)-(62); also produces the quadrupole correction in D and H in Section 4.4.
domain assumption Galilean invariance is postulated for the macroscopic fields, as in Dreyer et al.
Invoked in Section 3.1 to constrain transformation properties and to define the electromotive intensity (20) and Lorentz magnetisation (23).
domain assumption A positive susceptibility is assumed when using the entropy principle to fix the signs of the bound-current ansatz.
Section 3.6 requires 'together with a positive susceptibility' to select lambda_P = lambda_M = 1; this is an additional physical assumption beyond the entropy principle alone.
domain assumption The thermodynamic time scale t is much longer than the cyclotron period 1/omega_c,ki, so second-order mass corrections can be neglected while magnetisation terms of the same formal order are kept.
Section 4.7, estimate (63)-(65); this is needed for the continuity and momentum equations to coincide in Section 5.1.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Revisiting the Coupling of Thermodynamics and Electromagnetics." pith.science (2026). https://pith.science/paper/NKVEC3AM

@misc{pith2026260807595,
  author       = {Pith},
  title        = {Pith review of: Revisiting the Coupling of Thermodynamics and Electromagnetics},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NKVEC3AM}},
  note         = {Machine review of arXiv:2608.07595}
}

abstract

We revisit the coupling of continuum thermodynamics and electromagnetic theory for polarisable and magnetisable matter in motion. Two routes are followed and then compared. The first route is the axiomatic bulk theory of Dreyer, Guhlke and M\"uller, in which universal balance laws are closed by an entropy principle. We show that the source of the internal energy balance must be built with the non-convective electric current, that the polarisation current and the Lorentz magnetisation enter through one single identity, which Dreyer et al.\ do not write down, and that this identity fixes both the admissible entropy variables and the signs of the bound-current ansatz. The second route is the statistical-mechanical one of Mazur, in which the macroscopic Maxwell equations are obtained by ensemble averaging over a system of atoms with internal charge carriers. Mazur stops before the conservation laws, so we derive them, and we estimate the size of the mass-correction terms that appear. The comparison shows that after a redefinition of polarisation and magnetisation the two sets of equations agree structurally. The only irreducible difference is a momentum contribution from microscopic field fluctuations, which can not be reproduced in a purely macroscopic theory. We further show that the electromotive intensity $\mathcal{E}$ and the Lorentz magnetisation $\mathcal{M}$ are not modelling choices but appear by themselves, and that the asymmetric look of the entropy function is a consequence of the chosen energy variable and not a defect of the theory.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

14 extracted references · 14 canonical work pages

[1]

S. M. Barnett. Resolution of the Abraham–Minkowski dilemma.Physical Review Letters, 104(7):070401, 2010. [DOI]

work page 2010
[2]

Chapman and T

S. Chapman and T. G. Cowling.The Mathematical Theory of Non-Uniform Gases. Cam- bridge University Press, Cambridge, 2 edition, 1952

work page 1952
[3]

S. R. de Groot and P. Mazur.Non-Equilibrium Thermodynamics. North-Holland, Amster- dam, 1962. ISBN 0-486-64741-2. Reprinted by Dover, New York, 1984

work page 1962
[4]

Dreyer, C

W. Dreyer, C. Guhlke, and R. Müller. Bulk-surface electrothermodynamics and applications to electrochemistry.Entropy, 20(12):939, 2018. [DOI]

work page 2018
[5]

Dreyer, P.-E

W. Dreyer, P.-E. Druet, P. Gajewski, and C. Guhlke. Analysis of improved Nernst–Planck– Poisson models of compressible isothermal electrolytes.Zeitschrift für angewandte Mathe- matik und Physik (ZAMP), 71(4):119, 2020. [DOI]

work page 2020
[6]

Overcoming the shortcomings of the Nernst–Planck model.Physical Chemistry Chemical Physics, 15(19):7075–7086, 2013

Wolfgang Dreyer, Clemens Guhlke, and Rüdiger Müller. Overcoming the shortcomings of the Nernst–Planck model.Physical Chemistry Chemical Physics, 15(19):7075–7086, 2013. [DOI]

work page 2013
[7]

J. D. Jackson.Classical Electrodynamics. John Wiley & Sons, New York, 3 edition, 1999. ISBN 978-0-471-30932-1

work page 1999
[8]

Kovetz.Electromagnetic Theory

A. Kovetz.Electromagnetic Theory. Oxford University Press, Oxford, 2006. ISBN 978-0- 19-850604-1. [DOI]. First edition 2000

work page 2006

Show all 14 references

[9]

H. A. Lorentz.The Theory of Electrons. Teubner, Leipzig, 1909. 2nd ed. 1916, reprinted by Dover, New York, 1952

1909
[10]

G. A. Maugin.Continuum Mechanics of Electromagnetic Solids. North-Holland, Amster- dam, 1988

1988
[11]

P. Mazur. On statistical mechanics and electromagnetic properties of matter. In I. Prigogine, editor,Advances in Chemical Physics, volume 1, pages 309–360. John Wiley and Sons, Ltd, Hoboken, 1958. [DOI]

1958
[12]

Mazur and B

P. Mazur and B. R. A. Nijboer. On the statistical mechanics of matter in an electromagnetic field. I.Physica, 19:971–986, 1953. [DOI]. 20

1953
[13]

Müller, Elena N

Wolfgang H. Müller, Elena N. Vilchevskaya, and Victor A. Eremeyev. Electrodynamics from the viewpoint of modern continuum theory—a review.ZAMM – Journal of Applied Math- ematics and Mechanics, 103(4):e202200179, 2023. [DOI]. Published online 28 September 2022

2023
[14]

Steigmann

David J. Steigmann. On the formulation of balance laws for electromagnetic continua. Mathematics and Mechanics of Solids, 14(4):390–402, 2009. [DOI]. 21

2009

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.

[1] [1]

S. M. Barnett. Resolution of the Abraham–Minkowski dilemma.Physical Review Letters, 104(7):070401, 2010. [DOI]

work page 2010

[2] [2]

Chapman and T

S. Chapman and T. G. Cowling.The Mathematical Theory of Non-Uniform Gases. Cam- bridge University Press, Cambridge, 2 edition, 1952

work page 1952

[3] [3]

S. R. de Groot and P. Mazur.Non-Equilibrium Thermodynamics. North-Holland, Amster- dam, 1962. ISBN 0-486-64741-2. Reprinted by Dover, New York, 1984

work page 1962

[4] [4]

Dreyer, C

W. Dreyer, C. Guhlke, and R. Müller. Bulk-surface electrothermodynamics and applications to electrochemistry.Entropy, 20(12):939, 2018. [DOI]

work page 2018

[5] [5]

Dreyer, P.-E

W. Dreyer, P.-E. Druet, P. Gajewski, and C. Guhlke. Analysis of improved Nernst–Planck– Poisson models of compressible isothermal electrolytes.Zeitschrift für angewandte Mathe- matik und Physik (ZAMP), 71(4):119, 2020. [DOI]

work page 2020

[6] [6]

Overcoming the shortcomings of the Nernst–Planck model.Physical Chemistry Chemical Physics, 15(19):7075–7086, 2013

Wolfgang Dreyer, Clemens Guhlke, and Rüdiger Müller. Overcoming the shortcomings of the Nernst–Planck model.Physical Chemistry Chemical Physics, 15(19):7075–7086, 2013. [DOI]

work page 2013

[7] [7]

J. D. Jackson.Classical Electrodynamics. John Wiley & Sons, New York, 3 edition, 1999. ISBN 978-0-471-30932-1

work page 1999

[8] [8]

Kovetz.Electromagnetic Theory

A. Kovetz.Electromagnetic Theory. Oxford University Press, Oxford, 2006. ISBN 978-0- 19-850604-1. [DOI]. First edition 2000

work page 2006