WorksPhysics and Astronomy
Analytic Two-Loop Classical Bremsstrahlung Amplitude in Electrodynamics
Abstract
We compute the leading long-wavelength term of the two-loop five-point amplitude for photon emission in the scattering of two massive spinless charges. Classical consistency suggests that two Coulomb exchanges dress the radiative tree amplitude. We show that this iteration accounts for the entire leading term and evaluate the resulting convolution. At fixed external transfer, the exchanges shift the momentum in the tree numerator, thereby changing the relative weights of its two polarization structures. We determine both coefficients at nonzero photon frequency and general relative velocity, with exact dependence on the spacetime dimension.
Introduction
Introduction. Gravitational-wave observations have established merging black holes as a testing ground for general relativity in its strong-field, dynamical regime [1, 2]. Inferring masses and spins from these signals requires accurate models of the binary’s motion and radiation [3, 4]. During a quasicircular inspiral, the frequency evolution is governed by the radiated power and the change in binding energy per unit frequency [5]. At the high signal-to-noise ratios attainable with next-generation detectors, even small waveform errors can bias these estimates by more than their statistical uncertainties [4].
Quantum field theory offers a systematic route from relativistic scattering to binary dynamics. The post-Minkowskian (PM) expansion keeps the relative velocity fixed at each order in Newton’s constant. For conservative motion, analytic continuation relates the scattering angle to the periastron advance of a bound orbit [6, 7]. Combining effective field theory [8, 9] with -matrix techniques [10, 11, 12], several groups have determined the conservative dynamics of nonspinning bodies at 3PM [13, 14, 15] and 4PM [16, 17, 18, 19, 20]. The 5PM potential contribution has been extracted from four-loop elastic scattering for general mass ratios [21]. The conservative tail contribution to the deflection, due to radiation scattered by the binary’s gravitational field, has also been computed in a mass-ratio expansion [22, 23, 24]. For quasicircular, nonspinning inspirals, complementary post-Newtonian (PN) calculations yield the flux through 4PN and the phase through 4.5PN [5, 25].
In scattering, momentum conservation relates the total radiated four-momentum to the two bodies’ impulses, as in the 4PM calculation of radiation reaction [26, 27]. To resolve the frequency, direction, and polarization in an amplitude calculation, we keep the emitted graviton as a fifth external particle alongside the two incoming and two outgoing bodies. Fourier transforming the tree amplitude in the momentum transfer gives the leading scattering waveform [28, 29, 30]. The next correction combines the one-loop amplitude with products of tree amplitudes [31, 32, 33] and has been evaluated analytically at general relative velocity [34]. For nonspinning bodies with a general mass ratio, the two-loop waveform is known through relative 2PN order [35], and its even-parity quadrupole and octupole through 3.5PN [36, 37]. Since even small deflections can occur at relativistic speeds, a key input for extending this prediction to general velocities is the two-loop five-point amplitude.
In electrodynamics the corresponding five-point amplitude describes bremsstrahlung from two scattering charges [30, 32]. The spectrum and polarization are sensitive to how the charges accelerate throughout the encounter [38, 39]. A charge can exchange momentum with its partner both before and after emitting a photon. Summing these orderings can put intermediate matter lines on shell in the long-wavelength expansion. Although graviton self-interactions generate additional diagrams, the same identities apply to massive lines in gravity [40]. In electrodynamics, classical consistency relates one-loop emission to elastic and radiative tree amplitudes and suggests an iteration with two Coulomb exchanges at two loops [41].
In this work, we compute the classical two-loop five-point bremsstrahlung amplitude for two massive spinless charges at leading long-wavelength order. When expressed in terms of the physical masses and charges, the full diagram sum reduces to a convolution of the leading radiative tree amplitude with two Coulomb exchanges. Each exchange shifts the momentum transfer in both the tree propagator and its polarization numerator. We evaluate the convolution at finite photon frequency and general relative velocity, obtaining both polarization coefficients with exact dependence on the spacetime dimension. For opposite charges, we exhibit an exact tree zero where the double infrared pole vanishes, but the convolution samples other transfers and the single-pole residues from both emitting lines reinforce each other.
Long-wavelength expansion
We study bremsstrahlung in scalar electrodynamics with two complex fields of masses and charges (including the coupling). The Lagrangian is
In the charges are , with denoting the scale of dimensional regularization. We strip off the tree-level factors for emission and for elastic scattering, respectively. In the loop integrals we suppress per loop and restore it when extracting Laurent coefficients. We use Feynman gauge, the mostly-minus metric and an implicit for uncut propagators.
In the long-wavelength regime the exchanged and emitted momenta are small compared with the particle masses, even at relativistic speeds. We describe each massive line by its average momentum and transfer ,
Momentum conservation implies . Each transfer is spacelike, with . With , the mass-shell conditions imply and for a pole mass . The relative speed is , where and . The photon energy in the rest frame of is . We write , and .
In the soft region we scale together at fixed [42, 43]. For soft , the inverse scalar propagator has a linear leading term. Expanding about this term yields the eikonal propagator and its first correction.
Individual one- and two-loop integrands can be as singular as and , respectively. At positive photon energy these terms cancel in the sum or integrate to zero. We denote the leading surviving term of the amplitude by . Its integrand degree is , , at tree level, one loop, and two loops, respectively. Including the loop measures gives a degree of . With soft momenta proportional to and , the amplitude scales as in four dimensions [12, 41]. The two-loop term computed here is therefore enhanced by relative to the radiative tree amplitude.
At tree level, the two orderings of emission and exchange cancel at order on each line. At order the two-photon contact (seagull) vertex also contributes. The sum
can be written in terms of two contractions of the emitted photon’s field strength,
where and momentum conservation gives . At fixed , only depends on the momentum transfer. The leading tree amplitude is
Throughout, exchanges all particle labels, with following from .
From diagrams to transverse exchange
We integrate over the full soft region at fixed relative velocity. The transverse constraints on the two additional exchanges follow from summing their attachment orders on the massive lines. Summing two opposite orderings puts the intermediate massive line on its linearized mass shell through the identity
where is a linearized inverse propagator. The six orderings in Figure 1 give two such delta functions, or matter cuts [44, 45, 46]. Higher propagator powers lead to derivatives of these cuts.
Figure 1. A representative ordering sum on line 2. The upper line emits the photon at a contact vertex, and the three exchanged photons carry momenta with , so satisfy . For a fixed order on line 1, summing the six orderings on line 2 gives at eikonal order, up to common factors. The dashed lines on the right mark the two cuts, which define sector 22.
Of the 3322 two-loop diagrams generated with QGRAF [47], 1288 contain closed massive loops. When both loop momenta are soft, these diagrams have integrand degree at least because each massive propagator on a closed loop is of order one and each cubic vertex on it supplies a soft momentum. In hard subgraphs, charge conjugation cancels odd-photon currents, the definition of the charge in the Thomson limit absorbs the constant part of the vacuum polarization, and higher local photon interactions carry extra soft momenta [48]. Of the remaining 2034 diagrams, 280 have every internal photon connecting the two lines. In 32 of these diagrams, two photons join the same pair of seagulls, giving an integrand degree above . We sum the other 248 over attachment orders. With physical masses and charges and the external-line normalization specified in the Appendix, diagrams in which a photon returns to the same line give no net contribution at this order. These include the mixed sectors and , with one matter cut on each line. After scaleless terms are dropped, cut-preserving integration by parts leaves two terms in each mixed sector with identical denominators, equal coefficients and opposite numerators. Only 11 and 22 survive, with both cuts on line 1 or line 2 and charge factors or , respectively. Exchanging particle labels interchanges the sectors.
In sector 22 the two cuts set the loop energies to zero in the rest frame of . To combine the remaining linear denominators on the emitting line, we sum its attachment orders with their causal prescriptions and relabel exchanges under the integral. At this yields (Appendix). The two extra exchanges thus carry no energy in either rest frame and lie in the dimensions transverse to both velocities. The remaining exchange, , carries the longitudinal transfer into the radiative tree, leaving only the three photon denominators.
The two-loop amplitude. The two extra exchanges change the relative weights of and through a scalar transverse integral and its first moment along the transverse transfer. At fixed , the frequency dependence of both moments in sector 22 enters through . In the rest frame of , , so is the cosine of the angle between and . Positive photon energy restricts to . For we find
The full amplitude is obtained by adding the term, with angular variable . Replacing by sets in (3), so each charge sector separately satisfies the photon Ward identity. Before evaluating the master integrals, we also verified that all nine reduced coefficients in (A2) vanish individually under .
In (6), multiplies the tree combination . Only alters the relative weights.
For both moments converge absolutely in the strip . We evaluate them there and continue analytically in (Appendix). The exact results at this long-wavelength order are
where . At both hypergeometric functions equal one, so and are finite at generic . Here is parallel to , implying . The term in (6) therefore vanishes.
With factored out, the dimensionless combinations through order are
To obtain the finite Laurent coefficient of the amplitude, we expand through . The scale appears as in sector 22 and in sector 11.
A particular zero of the tree amplitude isolates the contribution of . At , a polarization satisfying with makes both charge terms vanish even in the unexpanded tree amplitude (A5). The bracket in (6) then reduces to . The single-pole coefficient of is for . For the opposite charges and real polarization specified after (A5), the two residues have the same sign, so a simple pole survives in their sum.
We independently check (8) through order by reducing the amplitude to master integrals using integration by parts [49, 50] with reverse unitarity to preserve both cuts [51]. We solve the differential equations [52, 53, 54, 55] with boundary values from auxiliary mass flow [56] computed using AMFlow [57, 58], and then test the solution at a second kinematic point (Appendix).
Coulomb dressing. The two extra exchanges shift the transfer entering the radiative tree to , where . Write for the transverse components of and . The longitudinal component of contributes to its inverse propagator. The transverse projection of has squared magnitude , giving the common integral
with and .
In the tree numerator, shifts by and leaves fixed. Summing the cubic and contact diagrams yields the entire term of (4) at the shifted transfer, while sector 11 accounts for emission from the other line. Each extra exchange contributes the elastic kernel , including the Jacobian from the two longitudinal constraints. With fixed and , we obtain
where shifts oppositely and accounts for the two identical exchanges. Both relations hold before expanding in (see Appendix for the one-loop derivation).
A Fourier transform in the transverse transfer at fixed converts each convolution into a product. At impact parameter the elastic kernel becomes
Equivalently, is the transverse Fourier transform of the elastic tree amplitude in (A7) divided by . At leading long-wavelength order, the one- and two-loop amplitudes in impact-parameter space are the linear and quadratic terms in the expansion of , respectively. The pole of this phase fixes the two-loop double pole. The finite parts of both polarization coefficients require the full convolution in (10), including exchange momenta comparable to the external transfers, where the tree numerator varies.
Infrared structure. The infrared poles in (9) arise when one or both of the additional exchange momenta become soft at fixed external transfer. When both exchanges become soft, the radiative tree is evaluated at the external transfer and has a double pole. The numerator of vanishes in this limit, eliminating its double pole. When only one exchange becomes soft, the other still shifts the tree numerator. For the third denominator is bounded below by , and the integral is ultraviolet convergent near .
Scaleless self-interaction integrals can contain mutually canceling ultraviolet and infrared poles. For external scalars with unit pole residues, the Ward identity fixes the relative sign of the logarithmic vertex and external-line corrections. These cancel, including their two-loop overlap (Appendix). The use of pole masses and the Thomson definition of the charges fixes the remaining local terms at this order. All poles at in the leading amplitude at nonzero photon frequency are therefore infrared.
The residue of the pole in is independent of and factors out of the inverse Fourier transform. With stripped off, the tree amplitude (4) is independent of at fixed external kinematics and polarization contractions. Restoring and denoting the coefficient of by , we obtain
At a tree zero, the finite part of the one-loop amplitude thus fixes the two-loop single pole. Multiplying by the inverse pole part of the virtual Coulomb phase, , removes the infrared poles from the leading long-wavelength amplitude through two loops [59, 60].
Soft-photon theorem. The infrared poles above come from virtual exchanges at fixed photon momentum. When the emitted photon itself becomes soft, the leading soft-photon theorem relates the five-point amplitude to elastic scattering [59, 61]. We take in a fixed direction, with and . Since and , is transverse to both velocities. Its magnitude remains small compared with both masses as . Expanding the incoming and outgoing soft factors to first order in the transfers gives
The bracket remains finite; since , the soft factor has at most a simple pole in . We compute the elastic amplitude independently at the same long-wavelength order by summing all orderings of three exchanged photons. Adding the two radiative charge sectors, we find
with the elastic amplitude evaluated at . We can take the soft limit in the convergent strip of (7) or after expanding about . Both procedures yield the same soft residue order by order (Appendix).
In sector 22 the soft residue depends on . Permutation symmetry among the three exchanges in (9) fixes its limit. The explicit logarithms in (8) cancel in this combination. Subtracting its zero-frequency limit at fixed removes the remaining poles, leaving as . In four dimensions, logarithms of the photon energy can enter at subleading order without affecting the residue in (14) [62].
Conclusions and Discussion. The Coulomb dressing in (10) accounts for the full leading long-wavelength term of the two-loop bremsstrahlung amplitude. At fixed external transfer the exchanges shift the tree numerator, thereby changing the relative polarization weights. We have evaluated both coefficients at nonzero photon frequency, including the finite terms left undetermined by the infrared poles.
In gravity one can likewise sum over attachment orders on a massive line. If all orderings share a numerator and the net energy transfer vanishes in that line’s rest frame, the eikonal denominators sum to delta functions. Explicit one-loop calculations of graviton emission [63, 64] recover the iteration of elastic and radiative trees predicted by the gravitational eikonal description [65]. At two loops the challenge is to identify complete ordering sums in the presence of momentum-dependent tensor numerators and graviton self-interactions. In electrodynamics, a higher-loop calculation could test whether repeated elastic exchanges continue to exhaust the leading amplitude, as classical consistency suggests [41].
Acknowledgments. We thank Yu Jiao Zhu and Yongqun Xu for useful discussions. This work was supported in part by the Start-up Research Fund of Southeast University (No. RF1028624160). YZ is supported by the National Natural Science Foundation of China (No. 12575078 and 12247103). YW is supported by ANR-24-CE31-7996. Large language models (primarily GPT-6 Astra and Kimi K3.1) were used under the authors’ supervision to assist with diagram analysis and manuscript polishing. The authors take full responsibility for the final manuscript and the results it reports.
References
Appendix
Two-loop integral families. The three families in sector 22 differ only in their linear denominators. For , and , define
For , the cut factor is
The integral vanishes if either cut index is nonpositive. An uncut propagator of index zero is omitted.
With , the photon denominators are , and . The linear denominators are
The amplitude is , with
Including the cut Jacobians, the coefficients are
where
The poles in require through first order in . The cuts reduce to a massless sunset integral in one fewer dimension. Bubble integration in leaves a linear propagator with prescription .
The two elliptic masters in each of families 1 and 3 have vanishing coefficients. An independent sum over crossline exchange diagrams reproduces all nine coefficients in sector 22. We obtain identical reductions in a common basis with Kira [66] and with FIRE [67] using LiteRed symmetries [68]. We verify the reduction and differential-equation matrices against exact IBP identities. Boundary data at fix the solution. At , we check 17 nonelliptic master-integral entries across the three families and both moments through against independent AMFlow evaluations at 216-digit working precision. Of 90 absolute differences from reference values , each normalized by , the largest is .
Ordering sum on the emitting line. The line-2 cuts and photon propagators are symmetric in the three exchanges. Terms containing scalar products of distinct exchanged momenta have support only at zero frequency after summing over the four cubic emission positions. The ordered sum for a seagull joining two virtual photons has the same support. At positive frequency, we order emission relative to the exchange and sum over the attachments, leaving their times unrestricted. Together with (5), Fourier transformation yields . The six orderings with emission and at a common seagull yield the same factor.
For independent real , the denominators and have the same negative imaginary regulator, so their sum has twice that regulator. Taking the real part as the regulator vanishes in the sense of distributions in gives
Relabeling the photons eliminates the principal-value products, leaving the delta functions in (A3).
By rotational symmetry, the vector first moment is parallel to . For nonzero transverse transfer, its coefficient is the moment with numerator divided by . Using and the measure conversion , we obtain (6) from (10).
Evaluation in dimensions. At fixed , the integral in (9) gives the massless bubble
Writing , we introduce a Feynman parameter and shift to obtain
Only the even part of the shifted first-moment numerator survives integration. The substitution yields Euler’s hypergeometric form. The original factor raises one beta-function argument by one, producing the prefactor ratio in (7).
Tree amplitude. Summing the six tree diagrams before the long-wavelength expansion, we obtain
where denotes with the particle labels exchanged. The leading term is (4). For , , and , with , the transfers satisfy . The unnormalized polarization then yields . For equal masses and charges , the two-loop single-pole residues in 22 and 11, in units of , are and , respectively. At , . The charge factors also have opposite signs, so the residues add.
One-loop Coulomb dressing. Summing both orientations in (5) cuts the line-1 propagator. Together, the two matter cuts leave the measure . With , and , the amplitude is
Since , twice the first moment is . The subtraction removes the ultraviolet tail. At small momentum, tends to , so the bubble cancels the subtracted massless propagator. The first moment remains finite at . Only , multiplying the unshifted tree numerator, carries the infrared pole in (12). The scaleless integral of vanishes in dimensional regularization, yielding (A6). Particle exchange gives the other sector. For , we find, through order , and .
Expanding and reproduces (12), including its single-pole terms.
Elastic amplitude and soft limit. We independently sum all photon orderings on both massive lines through two loops. Seagulls are suppressed at this order, and the transverse integrals give
with . The one-loop pole has residue times the elastic tree amplitude. At zero frequency, the three exchange momenta in (9) enter symmetrically and sum to the transverse transfer. Accordingly, equals two thirds of in the convergence strip. With , continuation in gives
For , the numerator cancels one power of the bracket in (A4) with a minus sign. Euler’s integral yields . The first nonanalytic correction scales as around zero, so expanding about preserves the residue in (14).
Self-interactions and hard regions. The logarithmic self-interaction integral on line is
At zero transfer, the logarithmic vertex correction is . Including seagulls, one such self-interaction on each line multiplies the tree by . With unit external-line residue, the Ward identity yields , canceling the overlap , including its mixed UV–IR pole. The pole-mass condition fixes the local mass term. Defining the charge in the Thomson limit subtracts the vacuum polarization at zero momentum.
In all 32 diagrams with two photons joining the same seagull pair, the three photon propagators scale as as the degree of the soft integrand is at least for the 24 diagrams with three internal scalar propagators and or higher for the eight with two.
When a hard subgraph joins both lines and one loop remains soft, one photon and at most three matter propagators produce an integrated degree of or higher in four dimensions. A fully hard scattering subgraph acquires at most one soft pole from external emission. Both are subleading to degree .
In (9) at , the ultraviolet degrees are for the scalar integral and at most for its first moment when either momentum grows large or both scale together in generic directions. At a fixed sum, the numerator remains unchanged, resulting in degrees at most for both. Thus (9) is ultraviolet finite near . At positive frequency, its poles at are infrared.
References
- [1]LIGO Scientific, Virgo collaboration, Observation of Gravitational Waves from a Binary Black Hole Merger, Phys. Rev. Lett. 116 (2016) 061102 [1602.03837].
- [2]LIGO Scientific, VIRGO, KAGRA collaboration, GWTC-4.0: Tests of General Relativity. I. Overview and General Tests, 2603.19019.
- [3]P. C. Peters, Gravitational Radiation and the Motion of Two Point Masses, Phys. Rev. 136 (1964) B1224.DOI
- [4]M. Pürrer and C.-J. Haster, Gravitational waveform accuracy requirements for future ground-based detectors, Phys. Rev. Res. 2 (2020) 023151 [1912.10055].DOI
- [5]L. Blanchet, G. Faye, Q. Henry, F. Larrouturou and D. Trestini, Gravitational-Wave Phasing of Quasicircular Compact Binary Systems to the Fourth-and-a-Half Post-Newtonian Order, Phys. Rev. Lett. 131 (2023) 121402 [2304.11185].DOI
- [6]G. Kälin and R. A. Porto, From Boundary Data to Bound States, JHEP 01 (2020) 072 [1910.03008].
- [7]G. Kälin and R. A. Porto, From boundary data to bound states, Part II. Scattering angle to dynamical invariants (with twist), JHEP 02 (2020) 120 [1911.09130].
- [8]W. D. Goldberger and I. Z. Rothstein, An effective field theory of gravity for extended objects, Phys. Rev. D 73 (2006) 104029 [hep-th/0409156].DOI
- [9]G. Kälin and R. A. Porto, Post-Minkowskian Effective Field Theory for Conservative Binary Dynamics, JHEP 11 (2020) 106 [2006.01184].DOI
- [10]N. E. J. Bjerrum-Bohr, P. H. Damgaard, G. Festuccia, L. Planté and P. Vanhove, General Relativity from Scattering Amplitudes, Phys. Rev. Lett. 121 (2018) 171601 [1806.04920].
- [11]C. Cheung, I. Z. Rothstein and M. P. Solon, From Scattering Amplitudes to Classical Potentials in the Post-Minkowskian Expansion, Phys. Rev. Lett. 121 (2018) 251101 [1808.02489].DOI
- [12]D. A. Kosower, B. Maybee and D. O’Connell, Amplitudes, Observables, and Classical Scattering, JHEP 02 (2019) 137 [1811.10950].DOI
- [13]Z. Bern, C. Cheung, R. Roiban, C.-H. Shen, M. P. Solon and M. Zeng, Scattering Amplitudes and the Conservative Hamiltonian for Binary Systems at Third Post-Minkowskian Order, Phys. Rev. Lett. 122 (2019) 201603 [1901.04424].DOI
- [14]Z. Bern, C. Cheung, R. Roiban, C.-H. Shen, M. P. Solon and M. Zeng, Black Hole Binary Dynamics from the Double Copy and Effective Theory, JHEP 10 (2019) 206 [1908.01493].DOI
- [15]G. Kälin, Z. Liu and R. A. Porto, Conservative Dynamics of Binary Systems to Third Post-Minkowskian Order from the Effective Field Theory Approach, Phys. Rev. Lett. 125 (2020) 261103 [2007.04977].DOI
- [16]Z. Bern, J. Parra-Martinez, R. Roiban, M. S. Ruf, C.-H. Shen, M. P. Solon et al., Scattering Amplitudes and Conservative Binary Dynamics at O(G⁴), Phys. Rev. Lett. 126 (2021) 171601 [2101.07254].DOI
- [17]C. Dlapa, G. Kälin, Z. Liu and R. A. Porto, Dynamics of binary systems to fourth Post-Minkowskian order from the effective field theory approach, Phys. Lett. B 831 (2022) 137203 [2106.08276].DOI
- [18]Z. Bern, J. Parra-Martinez, R. Roiban, M. S. Ruf, C.-H. Shen, M. P. Solon et al., Scattering Amplitudes, the Tail Effect, and Conservative Binary Dynamics at O(G⁴), Phys. Rev. Lett. 128 (2022) 161103 [2112.10750].DOI
- [19]C. Dlapa, G. Kälin, Z. Liu and R. A. Porto, Conservative Dynamics of Binary Systems at Fourth Post-Minkowskian Order in the Large-Eccentricity Expansion, Phys. Rev. Lett. 128 (2022) 161104 [2112.11296].
- [20]C. Dlapa, G. Kälin, Z. Liu and R. A. Porto, Local in Time Conservative Binary Dynamics at Fourth Post-Minkowskian Order, Phys. Rev. Lett. 132 (2024) 221401 [2403.04853].arxiv.org/abs/2403.04853
- [21]Z. Bern, E. Herrmann, R. Roiban, M. S. Ruf, A. V. Smirnov, S. Smith et al., Scattering Amplitudes and Conservative Binary Dynamics at O(G⁵) without Self-Force Truncation, Phys. Rev. Lett. 137 (2026) 021402 [2512.26354].arxiv.org/abs/2512.23654
- [22]M. Driesse, G. U. Jakobsen, G. Mogull, J. Plefka, B. Sauer and J. Usovitsch, Conservative Black Hole Scattering at Fifth Post-Minkowskian and First Self-Force Order, Phys. Rev. Lett. 132 (2024) 241402 [2403.07781].arxiv.org/abs/2403.07781
- [23]C. Dlapa, G. Kälin, Z. Liu and R. A. Porto, Local-in-Time Conservative Binary Dynamics at Fifth Post-Minkowskian and First Self-Force Orders, Phys. Rev. Lett. 135 (2025) 251401 [2506.20665].arxiv.org/abs/2506.20665
- [24]C. Dlapa, G. Kälin, Z. Liu and R. A. Porto, Nonlocal-in-time tail effects in gravitational scattering to fifth post-Minkowskian and tenth self-force orders, Phys. Rev. D 114 (2026) 024029 [2604.25916].arxiv.org/abs/2604.25916
- [25]L. Blanchet, G. Faye, Q. Henry, F. Larrouturou and D. Trestini, Gravitational-wave flux and quadrupole modes from quasicircular nonspinning compact binaries to the fourth post-Newtonian order, Phys. Rev. D 108 (2023) 064041 [2304.11186].
- [26]C. Dlapa, G. Kälin, Z. Liu, J. Neef and R. A. Porto, Radiation Reaction and Gravitational Waves at Fourth Post-Minkowskian Order, Phys. Rev. Lett. 130 (2023) 101401 [2210.05541].
- [27]C. Dlapa, G. Kälin, Z. Liu and R. A. Porto, Bootstrapping the relativistic two-body problem, JHEP 08 (2023) 109 [2304.01275].arxiv.org/abs/2304.01275
- [28]S. J. Kovacs and K. S. Thorne, The Generation of Gravitational Waves. 4. Bremsstrahlung, Astrophys. J. 224 (1978) 62.
- [29]G. U. Jakobsen, G. Mogull, J. Plefka and J. Steinhoff, Classical Gravitational Bremsstrahlung from a Worldline Quantum Field Theory, Phys. Rev. Lett. 126 (2021) 201103 [2101.12688].DOI
- [30]A. Cristofoli, R. Gonzo, D. A. Kosower and D. O’Connell, Waveforms from amplitudes, Phys. Rev. D 106 (2022) 056007 [2107.10193].DOI
- [31]A. Herderschee, R. Roiban and F. Teng, The sub-leading scattering waveform from amplitudes, JHEP 06 (2023) 004 [2303.06112].
- [32]A. ElkhdIr, D. O’Connell, M. Sergola and I. A. Vazquez-Holm, Radiation and reaction at one loop, JHEP 07 (2024) 272 [2303.06211].
- [33]S. Caron-Huot, M. Giroux, H. S. Hannessdottir and S. Mizera, What can be measured asymptotically?, JHEP 01 (2024) 139 [2308.02125].
- [34]G. Brunello, S. De Angelis and D. A. Kosower, Analytic one-loop scattering waveform in general relativity, Phys. Rev. D 114 (2026) 045004 [2511.05412].
- [35]D. Bini, T. Damour and A. Geralico, Gravitational bremsstrahlung waveform at the fourth post-Minkowskian order and the second post-Newtonian level, Phys. Rev. D 110 (2024) 064035 [2407.02076].
- [36]D. Bini, T. Damour and A. Geralico, Quadrupolar bremsstrahlung waveform at the third-and-a-half post-Newtonian accuracy, Phys. Rev. D 114 (2026) 064094 [2601.21522].arxiv.org/abs/2604.21522
- [37]D. Bini, T. Damour and A. Geralico, Octupolar bremsstrahlung waveform up to the two-loop level and the third-and-a-half post-Newtonian accuracy, 2609.11835.arxiv.org/abs/2609.11835
- [38]J. D. Jackson, Classical Electrodynamics. Wiley, 3rd ed., 1998.
- [39]M. V. S. Saketh, J. Vines, J. Steinhoff and A. Buonanno, Conservative and radiative dynamics in classical relativistic scattering and bound systems, Phys. Rev. Res. 4 (2022) 013127 [2109.05994].arxiv.org/abs/2109.05994
- [40]A. Brandhuber, G. R. Brown, G. Chen, S. De Angelis, J. Gowdy and G. Travaglini, One-loop gravitational bremsstrahlung and waveforms from a heavy-mass effective field theory, JHEP 06 (2023) 048 [2303.06111].DOI
- [41]A. Cristofoli, R. Gonzo, N. Moynihan, D. O’Connell, A. Ross, M. Sergola et al., The unitarity principle and classical amplitudes, JHEP 06 (2024) 181 [2112.07556].arxiv.org/abs/2112.07556
- [42]M. Beneke and V. A. Smirnov, Asymptotic expansion of Feynman integrals near threshold, Nucl. Phys. B 522 (1998) 321 [hep-ph/9711391].DOI
- [43]J. Parra-Martinez, M. S. Ruf and M. Zeng, Extremal black hole scattering at O(G^3): graviton dominance, modes from quasicircular nonspinning compact binaries to the fourth post-Newtonian order, Phys. Rev. D 108 (2023) 064041 [2304.11186].
- [44]M. Levy and J. Sucher, Eikonal approximation in quantum field theory, Phys. Rev. 186 (1969) 1656.
- [45]H. D. I. Abarbanel and C. Itzykson, Relativistic eikonal expansion, Phys. Rev. Lett. 23 (1969) 53.
- [46]H. Cheng and T. T. Wu, High-energy elastic scattering in quantum electrodynamics, Phys. Rev. Lett. 22 (1969) 666.DOI
- [47]P. Nogueira, Automatic Feynman Graph Generation, J. Comput. Phys. 105 (1993) 279.DOI
- [48]A. G. Grozin, Introduction to effective field theories. 1. Heisenberg-Euler effective theory, decoupling of heavy flavours, in Helmholtz International School - Workshop on Calculations for Modern and Future Colliders, 8, 2009, 0908.4392.arxiv.org/abs/0908.4392
- [49]K. G. Chetyrkin and F. V. Tkachov, Integration by parts: The algorithm to calculate β-functions in 4 loops, Nucl. Phys. B 192 (1981) 159.DOI
- [50]S. Laporta, High-precision calculation of multiloop Feynman integrals by difference equations, Int. J. Mod. Phys. A 15 (2000) 5087 [hep-ph/0102033].DOI
- [51]C. Anastasiou and K. Melnikov, Higgs boson production at hadron colliders in NNLO QCD, Nucl. Phys. B 646 (2002) 220 [hep-ph/0207004].DOI
- [52]A. V. Kotikov, Differential equations method: New technique for massive Feynman diagrams calculation, Phys. Lett. B 254 (1991) 158.DOI
- [53]E. Remiddi, Differential equations for Feynman graph amplitudes, Nuovo Cim. A 110 (1997) 1435 [hep-th/9711188].DOI
- [54]T. Gehrmann and E. Remiddi, Differential equations for two-loop four-point functions, Nucl. Phys. B 580 (2000) 485 [hep-ph/9912329].DOI
- [55]J. M. Henn, Multiloop integrals in dimensional regularization made simple, Phys. Rev. Lett. 110 (2013) 251601 [1304.1806].DOI
- [56]X. Liu, Y.-Q. Ma and C.-Y. Wang, A Systematic and Efficient Method to Compute Multi-loop Master Integrals, Phys. Lett. B 779 (2018) 353 [1711.09572].DOI
- [57]X. Liu and Y.-Q. Ma, AMFlow: A Mathematica package for Feynman integrals computation via auxiliary mass flow, Comput. Phys. Commun. 283 (2023) 108565 [2201.11669].DOI
- [58]R.-J. Huang, X. Liu and Y.-Q. Ma, AMFlow 2.0: significant algorithmic and software improvements for Feynman integral evaluation, 2607.08477.
- [59]S. Weinberg, Infrared photons and gravitons, Phys. Rev. 140 (1965) B516.DOI
- [60]D. R. Yennie, S. C. Frautschi and H. Suura, The infrared divergence phenomena and high-energy processes, Annals Phys. 13 (1961) 379.DOI
- [61]F. E. Low, Bremsstrahlung of very low-energy quanta in elementary particle collisions, Phys. Rev. 110 (1958) 974.DOI
- [62]B. Sahoo and A. Sen, Classical and Quantum Results on Logarithmic Terms in the Soft Theorem in Four Dimensions, JHEP 02 (2019) 086 [1808.03288].DOI
- [63]A. Georgoudis, C. Heissenberg and I. Vazquez-Holm, Inelastic exponentiation and classical gravitational scattering at one loop, JHEP 06 (2023) 126 [2303.07006].DOI
- [64]A. Georgoudis, C. Heissenberg and I. Vazquez-Holm, Addendum to Inelastic exponential and classical gravitational scattering at one loop, JHEP 02 (2024) 161 [2312.14710].DOI
- [65]P. Di Vecchia, C. Heissenberg, R. Russo and G. Veneziano, Classical gravitational observables from the Eikonal operator, Phys. Lett. B 843 (2023) 138049 [2210.12118].
- [66]F. Lange, J. Usovitsch and Z. Wu, Kira 3: integral reduction with efficient seeding and optimized equation selection, Comput. Phys. Commun. 322 (2026) 109999 [2505.20197].arxiv.org/abs/2505.20197
- [67]A. V. Smirnov and M. Zeng, FIRE 6.5: Feynman integral reduction with new simplification library, Comput. Phys. Commun. 302 (2024) 109261 [2311.02370].DOI
- [68]R. N. Lee, LiteRed 1.4: a powerful tool for reduction of multiloop integrals, J. Phys. Conf. Ser. 523 (2014) 012059 [1310.1145].
