FERMILAB-FN-1129-T

Notes on Lepton Gyromagnetic Ratios

Chris Quigg∗ Theoretical Department, Fermi National Accelerator Laboratory P.O. Box 500, Batavia, Illinois 60510 USA (Dated: June 2, 2021) A compendium for outsiders.

I. ORIGINS relativity and the general transformation theory leads to an explanation of all duplexity phenomena without fur- Dirac’s quantum theory of the tells us what ther assumption [4]. an electron is [1]: a particle that carries half a quan- tum (~/2) of spin angular momentum, −1 unit of elec- tric charge, and magnetic moment µe of minus one Bohr II. MAGNETIC MOMENT AS A DIAGNOSTIC magneton, −µB ≡ −~e/2me. Here ~ is the quantum of action, −e the electron charge, and me the electron mass. The Bohr magneton is defined as the magnitude A. Generalities of the magnetic dipole moment of a point electron orbit- ing an atom with one unit (~) of orbital angular momen- Within quantum electrodynamics (QED), a renormal- tum. [The numerical value is µB = 5.788 381 8060(17) × izable local relativistic quantum field theory of photons −11 −1 10 MeV T [2].] On the classical level, an orbiting and , the gyromagnetic ratio of a lepton, g`, point particle with electric charge e and mass m exhibits emerges unambiguously from the perturbation expan- a magnetic dipole moment given by sion. An extensive and interesting literature explores what makes g` = 2 the “natural value” for a structure- e~ less point particle of spin- 1 [5]. The requirement of good ~µL = L.~ (1) 2 2m high-energy behavior implies that g` − 2 must vanish at Dirac’s prediction for the electron magnetic moment is tree level for any well-behaved theory [6]. thus twice the value that would arise for a half unit of Quantum corrections induce a deviation from the Dirac orbital angular momentum, if that were possible. The moment that is traditionally expressed as the magnetic 1 moment anomaly, ratio ge ≡ µe/(− 2 µB) is called the gyromagnetic ratio of the electron; the Dirac equation predicts ge = 2. These properties are precisely what is required to ac- g` − 2 a` ≡ . (2) count for what Dirac calls “duplexity phenomena,” the 2 observed number of quantum states for an electron in an atom being twice the number given by the quantum the- The predicted value of a` can be confronted by experi- 1 ment. The model dependence of the magnetic anomaly ory of a spinless point particle. A spin- 2 electron with a magnetic moment of one Bohr magneton matches pre- makes it an incisive test of QED and a sensitive probe cisely what Uhlenbeck and Goudsmit [3] inferred from for new-physics contributions [7]. The interplay between their study of atomic spectra. theory and experiment played a decisive role in the devel- Dirac writes, The question remains as to why Nature opment and validation of quantum electrodynamics [8]. should have chosen this particular model for the elec- Beyond its role in challenging QED and probing for tron instead of being satisfied with the point-charge. One the virtual influence of undiscovered particles and forces, arXiv:2105.07866v2 [hep-ph] 31 May 2021 would like to find some incompleteness in the previous the anomalous magnetic moment provides stringent con- methods of applying quantum mechanics to the point- straints on lepton substructure [9]. For that purpose, we charge electron such that, when removed, the whole of the may examine the magnetic anomaly defect, defined as the duplexity phenomena follow without arbitrary assump- difference between theoretical prediction and experimen- tions. In the present paper it is shown that this is the tal determination, case, the incompleteness of the previous theories lying th exp in their disagreement with relativity, or, alternatetively, δa` ≡ a` − a` . (3) with the general transformation theory of quantum me- chanics. It appears that the simplest Hamiltonian for a A very simple working hypothesis is that no cancellations point-charge electron satisfying the requirements of both or suppression factors due to symmetries in the underly- ing dynamics account for the small mass m` of the com- posite lepton itself. Alternatively, so-called chiral models provide more modest constraints. In either case, it is in- ∗ [email protected]; ORCID: 0000-0002-2728-2445 ∗ formative to relate the compositeness scale M` and the 2

[4:µ] 2 2 radius R` of the lepton to the magnetic anomaly defect, so that ae ≈ 0.000 000 52 (α /π ). A form useful for evaluating fermion bubbles for all values of the mass ratio m` |δa`| = = m`R` no suppression, or (4a) is given in Ref. [15]. M ∗ ` Through heroic work over many decades, the calcula-  2 m` 2 2 tion of the electron’s magnetic anomaly has been carried |δa`| = ∗ = m` R` chiral model. (4b) 3 M` out through five loops in QED. The three-loop O(α ) contribution is given in closed form in [16]; evaluated nu- Searches for quark and lepton compositeness in high- merically, it gives energy collisions, reviewed in §92 of Ref. [2], reach above 10 TeV, again assuming no dynamical conspiracies. α3 a[6] = (1.181 241 456 ...) . (10) e π

B. The Electron Analytical calculations, reinforced by numerical evalua- tions, exist through four loops, i.e., up to O(α/π)4, or eighth order in the electron charge e. The coefficient of The electron was the focus of Dirac’s theory and the (α/π)5 is known from numerical integrations. In addi- test case for the developing theory of quantum electro- tion, the contributions of weak and hadronic interactions dynamics in the late nineteen-forties. It is stable on have been estimated. The predicted value as of 2019 the time scale of any conceivable experiment; the cur- 28 is [17] rent bound on the electron lifetime, τe > 6.6 × 10 yr at 90% C.L. [10], greatly exceeds the age of the uni- th[Cs] −14 ae = 115 965 218 160.6(11)(12)(229) × 10 , (11) verse. The electron mass is me = (0.510 998 946 1 ± 0.000 000 003 1) MeV. CPT invariance requires that the where the first two uncertainties are from the tenth- gyromagnetic ratios of electron and positron be identical: order QED term and the hadronic term, respectively. ge− = ge+ . The third and largest uncertainty comes from the value The anomalous magnetic moment of the electron was of the fine structure constant obtained from atom- discovered in 1947 by Kusch and Foley [11], who inferred interferometry measurements of the Cs atom, α−1(Cs) = 137.035 999 046(27) [18]. A more recent determination [KF] ae− = 0.001 19(5) (5) using matter-wave interferometry to measure the recoil velocity of a rubidium atom that absorbs a photon, leads from their study of hyperfine structure in gallium atoms. to α−1(Rb20) = 137.035 999 206(11) [19]. These two Julian Schwinger showed that the one-loop (O(e2)) quan- highly precise values differ by approximately 5.4 stan- tum correction to the electron’s magnetic moment con- dard deviations [20]. tributes [12] To compare theory and experiment, it is efficient to combine the theoretical uncertainties of Eqn. (11), thus: [2] α ae = ≈ 0.001 162, (6) 2π th[Cs] −14 ae = 115 965 218 161(023) × 10 . (12) where the numerical value reflects the value of the fine- This represents a prediction at the level of 0.23 parts per structure constant as then known, α = 1/137. trillion (ppt) for g . Adopting instead the 2020 value of The two-loop contribution to the gyromagnetic ratio e α−1(Rb) [19], the standard-model prediction is of the electron is also known analytically [13]. It is th[Rb20] −14 α2 197 π2 3 π2  ae = 115 965 218 025.2(95) × 10 , (13) a[4] = + + ζ(3) − ln 2 (7a) e π2 144 12 4 2 which carries an uncertainty of 0.1 ppt for ge and lies α2 [Cs] −1 = −0.328 ≈ −0.000 001 77. (7b) 5.5σ below ae . A smaller value of α , which is to say π a larger value of α, implies a larger calculated value of P∞ −3 ae. The difference is very closely given by the Schwinger Here ζ(3) = i=1 i = 1.202 056 903 ... is the Riemann zeta function of 3. The sum of Eqn. (6) and Eqn. (7) contribution, yields the theoretical prediction ca. 1957 (for a theory of α(Cs) − α(Rb20) photons and electrons only), ≈ 0.136(25) × 10−11, (14) 2π [2] [4] ae ≈ ae + ae = 0.001 159 6. (8) a shift to which modern measurements of ge are sensitive. An independent numerical evaluation of the O(e10) The contributions of heavier fermion (ff¯) bubbles are contribution of diagrams without lepton loops has been suppressed by mass ratios. The leading contribution carried out by Volkov [21], with a result that differs when the mass ratio (me/mf ) is small is [14] slightly from the result of Ref. [17]. Although the con- tending values differ by 4.8σ, the implications for ae are 2  2 [4:f] 1 α me not significant for current comparisons of theory and ex- ae ≈ 2 , (9) 45 π mf periment; a resolution will be needed in the near future. 3

Today, experimental determinations of ge have at- The uncertainties are from the tenth-order QED term, tained sub-ppt precision—a stunning achievement. The [H08] hadronic term, and ae− , respectively. The inferred evolution of experimental technique up to 1972 is re- −1 value, α (ae) = 137.035 999 150(33), lies 0.104(43) × viewed in [22], which also contains a thorough historical 10−6 (2.4σ) above α−1(Cs) [18] and 0.056(35) × 10−6 summary of theoretical developments. That chronology (1.6σ) below α−1(Rb) [19], as we anticipate from recounts the landmark experiments at the University of Eqns. (20, 21). Michigan that directly observed the spin precession of po- Planned improvements in technique by the Gabrielse larized electrons in a region of static magnetic field. This Research Group at (formerly technique culminated in a 3.5 parts-per-billion (ppb) de- Harvard) [26] aim at an order-of-magnitude improvement termination of ge [23], [H08] over the precision of ae− and a 150-fold improvement a[Mich] = 1 159 657.7 (3.5) × 10−9. (15) in the accuracy of ge− /ge+ compared with Eqn. (18). e− Should furher studies establish a discrepancy between Measurements on trapped single electrons led, over the standard-model prediction and the experimental de- time, to a 4.3-ppt determination of ge at the University termination of ge, it is of interest to ask whether a new of Washington (UW) [24]. The result for the anomalous light X boson, feebly coupled to the electron, might be moment is responsible. The NA64 experiment at CERN, which di- rects a 100-GeV electron beam onto stationary nuclei has a[UW] = 1 159 652 188.4 (4.3) × 10−12 (16) e− set new exclusion limits on a sub-GeV boson that decays The same experiment yields a positron anomaly of predominantly into invisible final states [27].

[UW] −12 ae+ = 1 159 652 187.9 (4.3) × 10 , (17) C. The Muon validating the prediction of CPT symmetry to a remark- able degree: The muon, the second-generation charged lepton, has −12 a mass of m = 105.658 374 5 (24) MeV and a mean life ge− /ge+ = 1 + (0.5 ± 2.1) × 10 . (18) µ −6 τµ = 2.196 981 1 (22) × 10 s. The parity-violating de- + + These stood as the definitive measurements for nearly cay µ → e νeν¯µ (and charge-conjugate) correlates the two decades. muon spin direction and the direction of the emitted The most precise published result, from the Harvard positron (or electron), as first demonstrated by Garwin, University team, is obtained by resolving the quantum Lederman, and Weinrich [28] at Columbia University’s cyclotron and spin levels of a single electron suspended Nevis Cyclotron. In leading approximation, the correla- for months in a cylindrical . This work 1 tion is given by (1 − 3 cos θ), where θ is the angle be- reaches an uncertainty of 0.28 ppt for ge− [25]. Written tween the muon spin and the emitted electron momen- in the same form as the UW results, it is tum [29]. Garwin, et al. were able to exploit this fact in their discovery experiment to constrain the muon’s a[H08] = 1 159 652 18 0.73(0.28) × 10−12. (19) e− gyromagnetic ratio to lie within 5% of the O(α) predic- Comparing the prediction Eqn. (11) and measurement tion, 2 (1 + α/2π) ≈ 2.001 162 (cf. Eqn. (6)). This was Eqn. (19), we find that the magnetic anomaly defect of quickly followed by an improved measurement in Liv- the electron is erpool, which reported gµ = 2.004 ± 0.014 and is no- table for an early example of the “wiggle plot” that is [Cs] th[Cs] [H08] −14 δae− ≡ ae − ae− = (88 ± 37) × 10 , (20) a feature of later storage-ring experiments [30]. Sub- sequent measurements by Garwin and collaborators led with the measurement 2.4σ below the prediction. The by 1960 to the determination a = 0.001 13+0.000 16, in difference between Eqn. (13) and Eqn. (19) yields µ −0.000 12 good agreement with the contemporaneous prediction, th:1960 δa[Rb20] ≡ ath[Rb20] − a[H08] = (−48 ± 30) × 10−14, (21) aµ = 0.001 16 [31]. e− e e− Over nearly two decades, three elegant storage-ring ex- with the measurement 1.6σ above the prediction. Taking periments at CERN greatly advanced both technique and −12 |δae| . 10 as a rough measure of a potential offset be- precision. They reported tween theory and experiment, we infer from Eqn. (4a) a 1965 −9 ∗ 5 aµ = 116 200 0 (5000) × 10 [32], (23a) compositeness scale Me & 5 × 10 TeV, or equivalently a −25 1972 −9 composite-electron radius Re . 4 × 10 m. The chiral- aµ = 116 616 0 (310) × 10 [33], (23b) invariant Ansatz gives, through Eqn. (4b), the more mod- 1979 −9 ∗ −19 aµ = 116 592 4 (8.5) × 10 [34]. (23c) est limits Me & 500 GeV and Re . 4 × 10 m. By equating their formal (five-loop) expression for the The final entry averages the measured anomalous mag- electron anomaly to Eqn. (19), Aoyama et al. [17] take netic moments of positive and negative muons, aµ+ = [H08] −9 −9 116 591 1(11) × 10 and a − = 116 593 7(12) × 10 . a − as a measure of the fine structure constant, µ e The progression of Eqn. (23) represents a nearly 600- −1 α (ae− ) = 137.035 999 149 6(13)(14)(330). (22) fold improvement through the series of experiments, 4 and the beginning of meaningful constraints on non- contending measurements of the fine structure constant standard contributions to the anomalous moment. The (cf. Eqn. (14)) is negligible on the scale of current exper- CERN experiments provided a baseline for later efforts imental capabilities.] at Brookhaven and Fermilab. The consensus relies on data-driven dispersion-relation The muon anomalous magnetic moment reported by evaluations of the hadronic vacuum polarization [40, 41]. Brookhaven experiment E821 value [35], Much current activity is devoted to a priori evaluations using lattice techniques. The path toward a purely the- BNL −11 aµ = 116 592 089 (63) × 10 , (24) oretical calculation was laid out by Blum [42] nearly two decades ago. An independent evaluation is of course de- set a new standard for precision: its uncertainty of 0.54 sirable; in addition, the lattice calculation has advantages ppm represented a 14-fold improvement compared to the for the separation of QED effects from hadronic correc- classic CERN measurements. The result was highly con- tions and the treatment of isospin breaking. Perhaps half sequential, because it suggested a mismatch between the- a dozen groups are pursuing this program, aiming to de- ory and experiment. In the words of the Abstract, termine the hadronic vacuum polarization contribution While the QED processes account for most with sub-percent precision. Their progress is reviewed in of the anomaly, the largest theoretical uncer- Refs. [39, 43]. tainty, ≈ 0.55 ppm, is associated with first- Fermilab E989 recently reported its first results [44]: order hadronic vacuum polarization. Present + − FNAL −11 evaluations, based on e e aµ = 116 592 040 (54) × 10 . (28) hadronic cross sections, lie 2.2–2.7 standard deviations below the experimental result. This 0.46-ppm determination of aµ, within 0.6σ of the BNL value, is 3.3σ above the theory consensus. Com- The E821 report stimulated not only a vigorous com- bining the BNL and FNAL measurements gives a 2021 merce in candidate interpretations of the putative break world average, in the standard model, but also intensive effort to make 2021 −11 the theoretical prediction more secure, largely by scruti- haµ i = 116 592 061(41) × 10 (0.35 ppm), (29) nizing the expectations for the hadronic vacuum polar- ization and the contribution—confirmed to be small—of 4.2σ above the theory consensus. The support for the light-by-light scattering. Planning for an improved ex- BNL central value and the modest strengthening of the perimental test led to the development of Fermilab Muon offset between theory and experiment is suggestive that g − 2 program using the relocated BNL muon storage new physics of some sort may be at play. Although mea- ring [36]. surements of aµ are less precise than those of ae, the A detailed review of theory and experiment through muon’s heavier mass confers greater sensitivity to new- 2009 appears in Ref. [14]. At two-loop order, the con- physics effects [45]. tribution of a muon loop in the photon vacuum polar- It is noteworthy that the observed offset, ization may be taken from Eqn. (7). The electron loop 2021 [WP:2020] −11 contributes [37] haµ i − aµ = 251 (59) × 10 (30) 2    [4:e] α 25 1 mµ me is not small. It is comparable in size to the electroweak aµ = 2 − + ln + O (25a) π 36 3 me mµ contribution to the anomaly [46], α2 ≈ 1.094 . (25b) [EW] −11 π2 aµ = 153.6 (1.0) × 10 . (31) Combining with Eqn. (6) and Eqn. (7b), we find A good story needs a wrinkle, and Muon g − 2 is no exception. The Budapest–Marseille–Wuppertal Collab- α α2 aµ = + 0.766 + · · · ≈ (0.000 116 6). (26) oration has presented a new lattice-QCD evaluation of 2π π2 the hadronic vacuum polarization, which enables them The QED contributions to the anomaly are now known to predict [47] through five loops (O(e10)) [38]. [BMW] −11 In support of the experimental program at Fermilab aµ = 116 591 954 (55) × 10 . (32) (E989), an international Muon g − 2 Theory Initiative is working to refine the standard-model prediction. Their The difference from the consensus value Eqn. (27) is sub- −11 consensus, reported in a 2020 White Paper (WP) [39], is tle but telling. The BMW value lies 144 (70) × 10 (2.1σ) above the consensus value and just 107 (69)×10−11 [WP:2020] −11 aµ = 116 591 810(43) × 10 . (27) (1.6σ) below the experimental world average. If it accu- rately reflects the standard-model prediction, there is no It differs from the BNL E821 result Eqn. (24) by 279(76) particular evidence for a discrepancy with experiment. in the three trailing places, so lies 3.7σ below the ex- Needless to say, theoretical and experimental work con- [2] perimental result. [The uncertainty in aµ due to the tinues, with a particular focus on the hadronic vacuum 5 polarization. The BABAR Collaboration [48], the CMD- that analyze its spin direction. However, experiments 3 [49] and SND [50] experiments at the Budker Institute similar to those carried out for the muon are made in Novosibirsk, and the Belle-II experiment [51] recently challenging by the short lifetime of the τ lepton, −15 commissioned at the KEK Super-B factory all aim at ττ = (290.3 ± 0.5) × 10 s. For that reason, conceiving + − refining the e e → hadrons data sets. The MUonE ex- a technique to measure aτ demands original thinking [54]. periment foreseen at CERN aims to gather information in the spacelike region by making precise measurements A recent theoretical prediction, within the standard- of µe elastic scattering [52]. model paradigm, is [41, 55] The (experimental average) magnetic anomaly defect th −8 with respect to the consensus prediction [39] is aτ = 117 717.1 (3.9) × 10 , (35) slightly greater than a and a . What information we δa = 251(59) × 10−11 (33) e µ µ have from experiment is indirect, derived from limits which reflects the 4.2σ mismatch between calculation and on anomalous σµν couplings of τ to electromagnetism. −9 The Review of Particle Physics [2] takes as the best cur- world average measurement. If we take |δaµ| . 3 × 10 , we estimate M ∗ 3.5×104 TeV and R 5.6×10−24 m rent constraint a limit from the DELPHI experiment at µ & µ . LEP [56] that derives from the measured cross section using Eqn. (4a) or M ∗ 1.9 TeV and R 10−19 m µ & µ . for the reaction e+e− → e+e−τ +τ −, which is to say using Eqn. (4b). γγ → τ +τ −: The analysis reported in Ref. [44] is based on 6% of the planned E989 data sample. The current uncertain- − 0.052 < aτ < 0.013 at 95% CL, (36) ties are 434 ppb statistical and 157 ppb systematic. By the summer of 2022, the collaboration expects to report or aτ = −0.018 ± 0.017. While this result is consistent on their Run 2 and Run 3 data sets, increasing to four with the Dirac value, gτ = 2, it does not meaningfully times the current sample (approximately 10 times the test either the standard-model prediction or the presence BNL E821 sample) and reducing the experimental error of unexpected quantum corrections. by a factor of two. At that point, they foresee a system- How could we do better? A reasonable near-term goal atic uncertainty at the 100 ppb level. The ultimate goal is might be to reach the level of the Schwinger contribu- to record and analyze 20 times the BNL sample, leading tion, Eqn. (6). A spin-precession experiment seems out to a further reduction of a factor of two in uncertainty. of the question, so improved measurements of the rates At the Japan Accelerator Research Complex in for two-photon production of tau pairs, or of hard-photon Tokai, J-PARC experiment E34 [53] will employ a very emission accompanying tau-pair production, merit con- different technique, using a 300 MeV/c reaccelerated ther- sideration. An example of original thinking is a pro- mal muon beam with 50% polarization. The beam will posal to study γγ → τ +τ − in ultraperipheral heavy- be injected vertically into a solenoid storage ring with 1 ion collisions at the Large Hadron Collider, aiming at a ppm local magnetic field uniformity for the muon stor- three-fold improvement on the DELPHI constraint [57]. age region with an orbit diameter of 66 cm. Compare The BELLE-II experiment [51] aims to record 45 bil- the Brookhaven and Fermilab experiments, with muon lion tau pairs, which will enable new searches for a (P- momentum of 3.09 GeV/c and orbit diameter 14.224 m. and T-violating) electric dipole moment and constraints The precision goal for aµ+ is a statistical uncertainty of on aτ [58]. What are the prospects for future electron– 450 ppb, similar to the statistical weight of the BNL and positron colliders: the International Linear Collider [59], FNAL-2021 samples, and a systematic uncertainty less FCC-ee [60], CEPC [61], CLIC [62], etc.? than 70 ppb.

D. The Tau Lepton III. g = 2 WITHOUT DIRAC? The third charged lepton, τ, has a mass mτ = 1 776.86 (12) MeV, so its anomalous magnetic moment Dirac’s statement of purpose and the elegant coherence should have a heightened sensitivity to quantum correc- of his results suggest that the gyromagnetic ratio g = 2 tions from heavy particles. In particular, aτ is more for the electron is a consequence of special relativity [63]. sensitive than aµ to heavy “new physics” by a ratio of And yet, following an argument given by Feynman [64], 2 (mτ /mµ) ≈ 280. By the same logic, aτ should also be modern quantum mechanics textbooks show that the more sensitive to strong-interaction contributions. canonical result can be recovered by simple manipula- In common with the muon, the tau lepton has parity- tions within the framework of nonrelativistic quantum violation decays, here mechanics, without explicitly invoking relativity [65]. − − Using the vector identity τ → µ ν¯µντ (17.39 ± 0.04)% (34a) − − τ → e ν¯eντ (17.82 ± 0.04)% (34b) σ · X σ · Y = X · Y + iσ · X × Y, (37) 6 where X and Y are arbitrary 3-vectors and σ is the Pauli ∇ × A = B as the magnetic field, we find spin matrix, we may write the free-electron Schr¨odinger ie Hamiltonian as Π × Π = ~B. (41) c

P 2 (σ · P)2 The second term of Eqn. (40) becomes H = = (38) 2me 2me e − ~ σ · B, (42) 2mec To include electromagnetic interactions we couple the so the gyromagnetic ratio g = 2 emerges from the vector potential A as prescribed by nonrelativistic me- Schr¨odingerequation, without specific mention of rela- chanics (P → Π = P − eA/c) and expand 1 tivity. We did have to identify the electron as a spin- 2 particle, whereas that emerges from Dirac’s construction, 1 and we all learned in school that spin- 2 particles corre- spond to particular representations of the Lorentz group, (σ · Π)(σ · Π) which seems to point to relativistic roots [66]. (39) 2me ACKNOWLEDGMENTS to obtain I thank Aida El-Khadra, Andreas Kronfeld, and Mo- gens Dam for helpful correspondence, Sam McDermott (P − eA/c)2 (P − eA/c) × (P − eA/c) for calling my attention to Ref. [57], and Florian Herren H= +iσ· . (40) 2me 2me for pointing me to Ref. [58]. Fermilab is operated by Fermi Research Alliance, LLC under Contract No. DE- AC02-07CH11359 with the U.S. Department of Energy, Then with the substitution P = −i~∇ and recognizing Office of Science, Office of High Energy Physics.

[1] P. A. M. Dirac, “The quantum theory of the electron,” 205 (2003); S. Ferrara, M. Porrati, and V. L. Telegdi, Proc. Roy. Soc. Lond. A 117, 610–624 (1928). See also “g = 2 as the natural value of the tree-level gyromag- The Principles of Quantum Mechanics (Oxford Univer- netic ratio of elementary particles,” Phys. Rev. D 46, sity Press, 1981) §70. 3529 (1992). [2] P. A. Zyla et al. (Particle Data Group), “Review of Par- [6] Steven Weinberg, “Dynamic and algebraic symmetries,” ticle Physics,” Prog. Theor. Exp. Phys. 2020, 083C01 in Lectures on Elementary Particles and Quantum Field (2020). This is the source for parameters not otherwise Theory, Vol. 1, edited by S. Deser, M. Grisaru, and attributed in this note. H. Pendleton (MIT Press, Cambridge, MA, 1970) 1970 [3] G. E. Uhlenbeck and S. Goudsmit, “Spinning elec- Brandeis University Summer Institute. trons and the structure of spectra,” Nature 117, 264– [7] For a helpful introduction to the subject, see G. F. 265 (1926). Their first publication on electron spin is Giudice, P. Paradisi, and M. Passera, “Testing new “Ersetzung der Hypothese vom unmechanischen Zwang physics with the electron g − 2,” JHEP 11, 113 (2012), durch eine Forderung bez¨uglich des inneren Verhaltens arXiv:1208.6583 [hep-ph]. jedes einzelnen Elektrons,” Die Naturwissenschaften 13, [8] Silvan S. Schweber, QED and the Men Who Made It 953–954 (1925). Spin is identified as the mysterious (Princeton University Press, Princeton, 1994). See in par- fourth quantum number posited by W. Pauli, “Uber¨ ticular Chapter 5, “The Lamb Shift and the Magnetic den Zusammenhang des Abschlusses der Elektronengrup- Moment of the Electron.” pen im Atom mit der Komplexstruktur der Spektren,” [9] Stanley J. Brodsky and S. D. Drell, “The Anomalous Zeitschrift fur Physik 31, 765–783 (1925). Magnetic Moment and Limits on Fermion Substructure,” [4] What he understands by “transformation theory” is ex- Phys. Rev. D 22, 2236 (1980). plained in P. A. M. Dirac, “The physical interpretation [10] M. Agostini et al. (Borexino), “A test of electric charge of the quantum dynamics,” Proceedings of the Royal So- conservation with Borexino,” Phys. Rev. Lett. 115, ciety of London A113, 621–641 (1927). We might simply 231802 (2015), arXiv:1509.01223 [hep-ex]. Borexino is an say, “quantum mechanics.” exquisitely radiopure liquid scintillation detector located [5] For a systematic presentation, see Barry R. Holstein, deep underground at the Gran Sasso Laboratory. The “How large is the “natural” magnetic moment?” Ameri- lifetime bound derives from a search for electron decay can Journal of Physics 74, 1104–1111 (2006). Other inter- into a neutrino and a single monoenergetic photon. esting considerations appear in H. Pfister and M. King, [11] P. Kusch and H. M. Foley, “Precision Measurement of 2 2 “The gyromagnetic factor in electrodynamics, quantum the Ratio of the Atomic ‘g Values’ in the P 3 and P 1 2 2 theory and general relativity,” Class. Quantum Grav. 20, 7

States of Gallium,” Phys. Rev. 72, 1256–1257 (1947); ture; Lowell S. Brown and Gerald Gabrielse, “Geonium “The magnetic moment of the electron,” Phys. Rev. 74, Theory: Physics of a Single Electron or Ion in a Penning 250–263 (1948). Trap,” Rev. Mod. Phys. 58, 233 (1986). [12] Julian S. Schwinger, “On Quantum electrodynamics and [25] D. Hanneke, S. Fogwell, and G. Gabrielse, “New Mea- the magnetic moment of the electron,” Phys. Rev. 73, surement of the Electron Magnetic Moment and the 416–417 (1948). Schwinger’s grave marker in the Mount Fine Structure Constant,” Phys. Rev. Lett. 100, 120801 Auburn Cemetery in Cambridge, Massachusetts, bears (2008), arXiv:0801.1134 [physics.atom-ph]. The value of α −1 the inscription, 2π . α inferred from the measurement, based on eighth- [13] The first calculation to include a fermion loop was given order theory, is superseded by the tenth-order result by Robert Karplus and Norman M. Kroll, “Fourth-Order quoted in Eqn. (22). Corrections in Quantum Electrodynamics and the Mag- [26] G. Gabrielse, S. E. Fayer, T. G. Myers, and netic Moment of the Electron,” Phys. Rev. 77, 536–549 X. Fan, “Towards an Improved Test of the Standard (1950). An important correction leading to the result Model’s Most Precise Prediction,” Atoms 7, 45 (2019), displayed here is due to A. Petermann, “Fourth order arXiv:1904.06174 [quant-ph]. For general information, magnetic moment of the electron,” Helv. Phys. Acta 30, see the Gabrielse Lab web site. 407–408 (1957) and to Charles M. Sommerfield, “Mag- [27] Yu. M. Andreev et al. (NA64), “Constraints on New netic Dipole Moment of the Electron,” Phys. Rev. 107, Physics in the Electron g − 2 from a Search for Invis- 328–329 (1957); “The magnetic moment of the electron,” ible Decays of a Scalar, Pseudoscalar, Vector, and Ax- Annals of Physics 5, 26–57 (1958). ial Vector,” (2021), Phys. Rev. Lett. (to be published), [14] Fred Jegerlehner and Andreas Nyffeler, “The Muon g−2,” arXiv:2102.01885 [hep-ex]. This article contains an ex- Phys. Rept. 477, 1–110 (2009), arXiv:0902.3360 [hep-ph]. tensive list of useful references. [15] M. Passera, “The Standard model prediction of the muon [28] R. L. Garwin, L. M. Lederman, and Marcel Weinrich, anomalous magnetic moment,” J. Phys. G 31, R75–R94 “Observations of the Failure of Conservation of Parity (2005), arXiv:hep-ph/0411168. and Charge Conjugation in Meson Decays: The Mag- [16] S. Laporta and E. Remiddi, “The Analytical value of the netic Moment of the Free Muon,” Phys. Rev. 105, 1415– electron (g − 2) at order α3 in QED,” Phys. Lett. B 379, 1417 (1957); see also J. I. Friedman and V. L. Telegdi, 283–291 (1996), arXiv:hep-ph/9602417. “Nuclear Emulsion Evidence for Parity Nonconservation [17] Tatsumi Aoyama, Toichiro Kinoshita, and Makiko Nio, in the Decay Chain π+ → µ+ → e+,” Phys. Rev. 106, “Theory of the anomalous magnetic moment of the elec- 1290–1293 (1957). tron,” Atoms 7, 28 (2019). [29] For a treatment of the muon spin asymmetry through [18] Richard H. Parker, Chenghui Yu, Weicheng Zhong, Brian O(α2) in QED, see Fabrizio Caola, Andrzej Czarnecki, Estey, and Holger M¨uller,“Measurement of the fine- Yi Liang, Kirill Melnikov, and Robert Szafron, “Muon structure constant as a test of the standard model,” Sci- decay spin asymmetry,” Phys. Rev. D 90, 053004 (2014), ence 360, 191–195 (2018). arXiv:1403.3386 [hep-ph]. [19] L´eoMorel, Zhibin Yao, Pierre Clad´e,and Sa¨ıdaGuellati- [30] J. M. Cassels et al., “Experiments with a polarized Kh´elifa,“Determination of the fine-structure constant muon beam,” Proc. Phys. Soc. A 70, 543–546 (1957). with an accuracy of 81 parts per trillion,” Nature 588, The (anomalous) precession frequency of a muon mov- 61–65 (2020); Sa¨ıda Guellati-Kh´elifa,“Measuring the ing through a magnetic field is imprinted on the time fine-structure constant to refine the Standard Model pre- distribution of high-energy positrons emitted along the dictions,” (CERN EP Seminar, May 4, 2021). For an as- µ+ spin direction in the form of oscillations on the expo- sessment, see Holger M¨uller,“Standard model of parti- nential decay curve characteristic of the (boosted) muon cle physics tested by the fine-structure constant,” Nature lifetime. 588, 37–38 (2020). [31] T. Coffin, R. L. Garwin, S. Penman, L. M. Lederman, [20] Ref. [17] quoted an earlier measurement of α−1(Rb): and A. M. Sachs, “Magnetic Moment of the Free Muon,” Rym Bouchendira, Pierre Clade, Saida Guellati-Khelifa, Phys. Rev. 109, 973–979 (1958); R. L. Garwin, D. P. Francois Nez, and Francois Biraben, “New determi- Hutchinson, S. Penman, and G. Shapiro, “Accurate De- nation of the fine structure constant and test of the termination of the µ+ Magnetic Moment,” Phys. Rev. quantum electrodynamics,” Phys. Rev. Lett. 106, 080801 118, 271–283 (1960). (2011), arXiv:1012.3627 [physics.atom-ph]. [32] G. Charpak, P. J. M. Farley, E. L. Garwin, T. Muller, [21] Sergey Volkov, “Calculating the five-loop QED contri- J. C. Sens, and A. Zichichi, “The anomalous magnetic bution to the electron anomalous magnetic moment: moment of the muon,” Nuovo Cim. 37, 1241–1363 (1965). Graphs without lepton loops,” Phys. Rev. D 100, 096004 [33] J. Bailey, W. Bartl, G. von Bochmann, R. C. A. Brown, (2019), arXiv:1909.08015 [hep-ph]. See §I for the conse- F. J. M. Farley, M. Giesch, H. Jostlein, S. van der Meer, −1 quences for ae and α . E. Picasso, and R. W. Williams, “Precise Measure- [22] A. Rich and J. C. Wesley, “The current status of the ment of the Anomalous Magnetic Moment of the Muon,” lepton g factors,” Rev. Mod. Phys. 44, 250–283 (1972). Nuovo Cim. A 9, 369–432 (1972). [23] John C. Wesley and Arthur Rich, “High-field electron [34] J. Bailey et al. (CERN-Mainz-Daresbury), “Final Report g − 2 measurement,” Phys. Rev. A 4, 1341–1363 (1971). on the CERN Muon Storage Ring Including the Anoma- [24] Robert S. Van Dyck, Paul B. Schwinberg, and Hans G. lous Magnetic Moment and the Electric Dipole Moment Dehmelt, “New high-precision comparison of electron of the Muon, and a Direct Test of Relativistic Time Di- and positron g factors,” Phys. Rev. Lett. 59, 26–29 lation,” Nucl. Phys. B 150, 1–75 (1979). (1987). See also Hans Dehmelt, “Experiments with an [35] G. W. Bennett et al., “Final Report of the Muon E821 isolated subatomic particle at rest,” Rev. Mod. Phys. Anomalous Magnetic Moment Measurement at BNL,” 62, 525–530 (1990), alternate source: 1989 Nobel Lec- Phys. Rev. D 73, 072003 (2006), arXiv:hep-ex/0602035. 8

Also see the E821 web site. Colloquium, April 13, 2021). [36] R. M. Carey et al., “The New (g −2) Experiment: A pro- [48] Vladimir Druzhinin (BaBar), “Recent results on hadronic posal to measure the muon anomalous magnetic moment cross sections measurements at BABAR for the g − 2 to ±0.14 ppm precision,” FERMILAB-PROPOSAL- calculation,” PoS EPS-HEP2019, 535 (2020). 0989 (2009). [49] A. Ryzhenenkov et al., “Overview of the CMD-3 recent [37] Hiroshi Suura and Eyvind H. Wichmann, “Magnetic Mo- results,” Journal of Physics: Conference Series 1526, ment of the Mu Meson,” Phys. Rev. 105, 1930–1931 012009 (2020). (1957); A. Petermann, “Magnetic moment of the mu [50] M. N. Achasov et al. (SND), “Measurement of the meson,” Phys. Rev. 105, 1931 (1957); and Sommerfield, e+e− → π+π− process cross section with the SND de- Ref. [13]. For a complete evaluation of the electron loop, tector at the√ VEPP-2000 collider in the energy region retaining me/mµ terms, see H. H. Elend, “On the anoma- 0.525 < s < 0.883 GeV,” JHEP 01, 113 (2021), lous magnetic moment of the muon,” Physics Letters 20, arXiv:2004.00263 [hep-ex]. 682–684 (1966), erratum: ibid. 21, 720 (1966). [51] E. Kou, P. Urquijo, W. Altmannshofer, et al., “The [38] Tatsumi Aoyama, Masashi Hayakawa, Toichiro Ki- Belle II Physics Book,” PTEP 2019, 123C01 (2019), noshita, and Makiko Nio, “Complete Tenth-Order QED arXiv:1808.10567 [hep-ex]. Contribution to the Muon g−2,” Phys. Rev. Lett. 109, [52] C. M. Carloni Calame, M. Passera, L. Trentadue, and 111808 (2012), arXiv:1205.5370 [hep-ph]. G. Venanzoni, “A new approach to evaluate the leading [39] T. Aoyama et al. (Muon g − 2 Theory Initiative), hadronic corrections to the muon g-2,” Phys. Lett. B 746, “The anomalous magnetic moment of the muon in 325–329 (2015), arXiv:1504.02228 [hep-ph]; G. Abbiendi the Standard Model,” Phys. Rept. 887, 1–166 (2020), et al., “Measuring the leading hadronic contribution to arXiv:2006.04822 [hep-ph]. the muon g-2 via µe scattering,” Eur. Phys. J. C 77, 139 [40] M. Davier, A. Hoecker, B. Malaescu, and Z. Zhang, (2017), arXiv:1609.08987 [hep-ex] For a recent update, “A new evaluation of the hadronic vacuum polarisa- see Giovanni Abbiendi, “Status of the MUonE experi- tion contributions to the muon anomalous magnetic mo- ment,” PoS ICHEP2020, 223 (2021), arXiv:2012.07016 2 ment and to α(mZ ),” Eur. Phys. J. C 80, 241 (2020), [hep-ex]. arXiv:1908.00921 [hep-ph]. [53] M. Abe et al., “A new approach for measuring the muon [41] Alexander Keshavarzi, Daisuke Nomura, and Thomas anomalous magnetic moment and electric dipole mo- 2 Teubner, “g − 2 of charged leptons, α(MZ ) , and the hy- ment,” Progress of Theoretical and Experimental Physics perfine splitting of muonium,” Phys. Rev. D 101, 014029 2019, 053C02 (2019). More information is available at (2020), arXiv:1911.00367 [hep-ph]. the experiment’s web page. [42] T. Blum, “Lattice calculation of the lowest-order [54] See Problem 1.6 of Chris Quigg, Gauge Theories of the hadronic contribution to the muon anomalous magnetic Strong, Weak, and Electromagnetic Interactions (Ben- moment,” Phys. Rev. Lett. 91, 052001 (2003). jamin / Cummings, Reading, Massachusetts, 1983); sec- [43] Harvey B. Meyer and Hartmut Wittig, “Lattice qcd and ond edition, (Princeton University Press, Princeton, the anomalous magnetic moment of the muon,” Progress 2013). in Particle and Nuclear Physics 104, 46–96 (2019). [55] For an earlier estimate, see S. Eidelman and M. Passera, [44] B. Abi et al. (Muon g − 2 Collaboration), “Measurement “Theory of the tau lepton anomalous magnetic moment,” of the positive muon anomalous magnetic moment to 0.46 Mod. Phys. Lett. A 22, 159–179 (2007), arXiv:hep- ppm,” Phys. Rev. Lett. 126, 141801 (2021). More infor- ph/0701260. mation is available at the E939 web site. [56] J. Abdallah et al. (DELPHI), “Study of tau-pair pro- [45] It is an interesting challenge to exploit the greater preci- duction in photon-photon collisions at LEP and limits sion of ae to test interpretations of the aµ deviation. See on the anomalous electromagnetic moments of the tau F. Terranova and G. M. Tino, “Testing the aµ anomaly lepton,” Eur. Phys. J. C 35, 159–170 (2004), arXiv:hep- in the electron sector through a precise measurement of ex/0406010. Compare the value aτ = 0.004±0.027±0.023 h/m,” Phys. Rev. A 89, 052118 (2014). from the rate of hard photons in e+e− → τ +τ −γ re- [46] Andrzej Czarnecki, William J. Marciano, and Arkady ported by M. Acciarri et al. (L3), “Measurement of the Vainshtein, “Refinements in electroweak contributions anomalous magnetic and electric dipole moments of the to the muon anomalous magnetic moment,” Phys. tau lepton,” Phys. Lett. B 434, 169–179 (1998). See also Rev. D 67, 073006 (2003), [Erratum: Phys.Rev.D 73, K. Ackerstaff et al. (OPAL), “An upper limit on the 119901 (2006)], arXiv:hep-ph/0212229; C. Gnendiger, anomalous magnetic moment of the tau lepton,” Phys. D. St¨ockinger, and H. St¨ockinger-Kim, “The elec- Lett. B 431, 188–198 (1998), arXiv:hep-ex/9803020 for troweak contributions to (g − 2)µ after the Higgs bo- the result −0.068 < aτ < 0.065. son mass measurement,” Phys. Rev. D 88, 053005 [57] Lydia Beresford and Jesse Liu, “New physics and tau (2013), arXiv:1306.5546 [hep-ph]. The corresponding g −2 using LHC heavy ion collisions,” Phys. Rev. D 102, [EW] contribution for the electron is ae = 3.053 (23) × 113008 (2020), arXiv:1908.05180 [hep-ph]. 10−14: Fred Jegerlehner, “Variations on Photon Vac- [58] Matteo Fael, Lorenzo Mercolli, and Massimo Passera, uum Polarization,” EPJ Web Conf. 218, 01003 (2019), “Towards a determination of the tau lepton dipole mo- arXiv:1711.06089 [hep-ph]. ments,” Nucl. Phys. B Proc. Suppl. 253-255, 103–106 [47] Sz. Borsanyi et al. (BMW Collaboration), “Leading (2014), arXiv:1301.5302 [hep-ph]; S. Eidelman, D. Epi- hadronic contribution to the muon magnetic mo- fanov, M. Fael, L. Mercolli, and M. Passera, “τ dipole ment from lattice QCD,” Nature 593, 51–55 (2021), moments via radiative leptonic τ decays,” JHEP 03, 140 arXiv:2002.12347 [hep-lat]. See also Zoltan Fodor, “(g − (2016), arXiv:1601.07987 [hep-ph]. 2)µ from lattice QCD and experiments: 4.2σ?” (DESY [59] “The International Linear Collider,” web site. [60] A. Abada et al. (FCC), “FCC Physics Opportunities: Fu- 9

ture Circular Collider Conceptual Design Report Volume Benjamin, 1961). 1,” Eur. Phys. J. C 79, 474 (2019). [65] J. J. Sakurai, Modern Quantum Mechanics (Addison- [61] Mingyi Dong et al. (CEPC Study Group), “CEPC Con- Wesley, 1967), pp. 78–79. R. Shankar, Principles of ceptual Design Report: Volume 2 - Physics & Detector,” Quantum Mechanics (Springer US, 1994) Chapter 20, (2018), arXiv:1811.10545 [hep-ex]. where Dirac’s derivation is recast in modern vector no- [62] “CLIC Detectors and Physics,” web site. tation. See also the survey by Ronald J. Alder and [63] This conviction is reinforced by Dirac’s debt to L. H. Robert A. Martin, “The electron g factor and factoriza- Thomas, “The motion of a spinning electron,” Nature tion of the Pauli equation,” American Journal of Physics 117, 514 (1926); “The kinematics of an electron with an 60, 837–839 (1992). axis,” Phil. Mag. Ser. 7 3, 1–21 (1927). For a reminis- [66] For an argument that spin emerges from (nonrelativistic) cence, see “Recollections of the discovery of the Thomas Galilean invariance, see Jean-Marc L´evy-Leblond,“Non- precessional frequency,” AIP Conf. Proc. 95, 4–12 (1983). relativistic particles and wave equations,” Communica- [64] Richard P. Feynman, Quantum Electrodynamics (W. A. tions in Mathematical Physics 6, 286–311 (1967).