Determining weak phases from B->J/Psi P decays

Worth your time, with one caveat about how the result is sold. The genuinely new piece is the linear SU(3)-breaking decomposition in Sec. 3 and App. A: three additional reduced matrix elements from the (8) mass-breaking tensor, the observation that the coefficient matrix is rank-deficient so only two combinations A_ε1, A_ε2 actually enter, and the resulting parametrization in Eq. 3.7. The SU(3) limit reproduces Zeppenfeld, which the author says plainly. The algebra in Tabs. 5 and 6 is checkable by hand and I found no slips. The fit machinery is transparent — NLopt with augmented Lagrangian, cross-checked in Mathematica, all inputs tabulated — and the negative result on γ in Sec. 4.3 is a useful corrective to the more optimistic claim in De Bruyn-Fleischer-Koppenburg. That deflation is the part I'd actually cite, more than the ΔS number. On the soft spot the stress-test flags: it lands, but it's a presentation problem, not a physics error. The |ΔS| ≲ 0.01 of abstract:s5 is a fit output whose width comes from r_SU(3) ≤ 40%, r_pen ≤ 50%, δ = 0, and hand-set caps on Im(R_ε1,2). The author is candid about all of this — Fig. 6 shows the 95% CL range widening by roughly a factor of three at 200% priors, and the constrained fits are labelled as such. What the abstract does not say is that the bound is prior-driven; 'model-independent' there means model-independent within the SU(3) truncation and the assumed ranges, which is a narrower claim than a casual reader will take away. Minor: δ = 0 is justified by a χ² comparison (18.2/3 vs 22.3/5) that is suggestive rather than decisive, and the future-scenario γ numbers in App. B rest on hand-chosen central values, so treat those as illustrations. Neither is load-bearing. Who should read it: anyone doing penguin pollution in the golden mode, or fitting the UT with B→J/ψK_S in the input set. The framework is the contribution; the number is provisional until the CP asymmetries in B_s→J/ψK_S and the B→J/ψK rate ratio sharpen. Send it out — the referee should push on the abstract's phrasing of the bound, not on the derivation.

The central claim is a bound, not a point measurement: |ΔS| ≲ 0.01 (abstract:s5; Sec. 4.2, Eqs. 4.14–4.15). That bound is obtained with r_SU(3) ≤ 40%, r_pen ≤ 50%, δ = 0, and hand-set caps on Im(Rε1,2) (Sec. 3; App. A after Eq. A.7). The paper is transparent that these are theoretical priors, and Fig. 6 shows that relaxing them to 200% widens the ΔS range by roughly a factor of three. The concern is therefore not an algebraic error but a mismatch between the abstract's unqualified '|ΔS| <~ 0.01' and the fact that the quoted width is set by the assumed ranges. The same data with a weaker prior do not support the same precision. This is the single most load-bearing soft spot because the paper's novelty is precisely the quantified smallness of the penguin shift; if the width is prior-driven, the 'model-independent extraction' is model-independent only within the SU(3) truncation and prior choices. The body acknowledges this; the abstract does not.

I agree with the reader's weakest_assumption. The paper is a careful SU(3) analysis with a checkable decomposition (App. A) and tabulated inputs; I found no internal inconsistency in Eq. 3.7 or the fit logic. The real soft spot is that the headline number is a fit output whose width is controlled by theoretical priors. The paper itself shows this in Fig. 6 and discusses it in Sec. 4.3, so the authors are not hiding it. But the abstract states |ΔS| <~ 0.01 without the prior caveat, and the 68% CL upper edge for dataset 2 is already 0.011 (Eq. 4.15), while the 95% CL edge is 0.012. The gamma conclusion in Sec. 4.3 is weaker still because it depends on hand-chosen future central values (App. B). None of this is a fatal error; it is a precision claim whose precision is partly assumed. The concrete test is to quantify the relaxed-prior interval. If it exceeds 0.01, the abstract needs a one-sentence qualification. That is a minor revision, so I would keep the verdict at CONDITIONAL rather than reject. The reader's MODERATE confidence is appropriate given the unverified Clebsch-Gordan coefficients and NLopt fits.

The SU(3) decomposition including linear breaking terms (App. A, Tabs. 5–6) is a genuinely new ingredient, and the fit framework is reproducible: the minimisation uses NLopt with an augmented-Lagrangian algorithm and the best-fit points are cross-checked with Mathematica (Sec. 4). The result is falsifiable: the predicted sign and size of ΔS, and the predicted CP asymmetries in B_s → J/ψ K_S (Fig. 4), can be tested with coming LHCb data. The critical re-examination of the γ sensitivity (Sec. 4.3) is a useful corrective to the more optimistic estimate of Ref. [8]. The paper is also candid about its limitations, noting that the present data do not strictly require the penguin contributions (Sec. 4.2) and that a fully general model-independent analysis of NP in the decay amplitudes is impossible (Sec. 5).

The decay $B_d^0$ → $J/ψ K_S$ is the standard candle for the $B_d$ mixing phase, which in the Standard Model is the CKM angle $β$. Its cleanliness rests on dropping doubly Cabibbo-suppressed terms; at collider precision those terms are no longer free. Martin Jung asks whether the neglected pieces can be constrained from data instead of calculated. He writes the six amplitudes for $B_{u,d,s}$ → $J/ψ P$ with $P = π, K$ in terms of SU(3) reduced matrix elements, adds the leading symmetry-breaking terms from the strange-quark mass, and fits the resulting eleven hadronic parameters plus $γ$ to the measured rates and CP asymmetries. The answer he reports is that the penguin-induced shift $ΔS$ in the indirect CP asymmetry is at most about $0.01$, with a slight preference for a positive sign. That matters because it converts an unquantified theoretical uncertainty on $sin 2β$ into a number with an error bar, and because the sign and size bear on the long-standing mild tension among different determinations of $β$.

What carries the argument. The machinery is an SU(3) tensor decomposition of the $b → s$ and $b → d$ effective Hamiltonians, combined with linear SU(3) breaking. The penguin-dominated part transforms as a triplet, the tree and electroweak-penguin parts as $15$, $6̄$ and $3$; multiplying by the isospin-conserving octet mass term generates three further reduced matrix elements. The resulting coefficient matrix is rank-deficient, so the seven amplitudes collapse onto six independent combinations, described by eleven hadronic parameters plus $γ$. These are the leading amplitude $N$, two symmetry-breaking combinations $R_{ε1,2}$, two dominant subleading combinations $R_{u1,2}$, and a remainder $δ$. A power counting $ξ ≈ 0.2$ sets expectations $|R_{ε1,2}|, |R_{u1,2}| ∼ ξ$ and $|δ| ∼ ξ^3$, and the fit imposes the priors $r_{SU(3)} ≤ 40%$ and $r_{pen} ≤ 50%$ on the underlying matrix-element ratios. The same framework lets the author check whether the data prefer non-zero penguin terms at all, and whether the observed isospin asymmetry $A_K^I ≃ -0.037 ± 0.025$ is a hint of new physics or of underestimated hadronic contributions.

Core claim. The central claim is that $ΔS_{J/ψK_S}$, the penguin-induced correction to the indirect CP asymmetry of the golden mode, is small: $|ΔS| ≲ 0.01$. Fitting the two data sets gives $ΔS = [0.001, 0.005]$ and $[0.004, 0.011]$ at 68% CL, with the 95% ranges extending slightly below zero. The extracted mixing phase is $φ = 0.74 ± 0.03$, to be compared with $0.73 ± 0.03$ from the naive relation that sets $ΔS = 0$; the precision is unchanged, the central value moves slightly. Including SU(3) breaking is essential to this result: in the exact-symmetry limit the fit fails badly ($χ^2/d.o.f. = 22.3/5$ for the first data set), and the branching-ratio ratios $R_{KK}$ and $R_{πK}$ can be absorbed by symmetry-breaking parameters of about 20% rather than by large penguin terms. The paper also reports a negative result on $γ$: even in a far-future scenario with $50$ ab$^{-1}$ of Super Flavour Factory data, the allowed range on $γ$ stays near $±35°$, so these modes do not deliver the competitive $γ$ determination that earlier work anticipated.

The premise it rests on. The quoted width of the bound depends on hand-imposed ceilings on how large the symmetry-breaking and penguin matrix elements are allowed to be, namely 40% and 50% of the leading amplitude, together with setting the smallest remainder parameter to zero because including it made the fit worse.

Values of $sin 2β$ quoted from $B_d^0$ → $J/ψ K_S$ can carry a quantified penguin shift of order $10^{-2}$ rather than an open theoretical error, which is the size relevant to LHCb and Super Flavour Factory precision.

The preferred positive sign of $ΔS$ slightly reduces, but does not remove, the tension between the golden-mode determination of $β$ and fits that include $B → τν$.

Improved measurements of the direct CP asymmetries and of the branching-ratio ratios $R_{KK}$ and $R_{πK}$ translate directly into a tighter bound on $ΔS$; the paper tabulates three luminosity scenarios with errors shrinking from $0.030$ to $0.015$ on $φ$ and from $0.015$ to $0.005$ on the width of the $ΔS$ range.

A measurement of the CP asymmetries in $B_s^0$ → $J/ψ K_S$, not yet available, would remove part of the allowed region and is the single most useful near-term addition to the fit.

Because the same expressions hold under minimal flavour violation, an unexpectedly large subleading amplitude could be reinterpreted as a new-physics contribution, with correlated effects expected in decays such as $B → D D$.

The reported bound is a statement about the assumed ranges as much as about the data: the paper's own Fig. 6 shows that relaxing the priors to 200% widens the $ΔS$ range roughly threefold, so an independent lattice or QCD-based estimate of the SU(3)-breaking matrix elements would be worth more here than another factor of two in luminosity.

Since the fit already prefers the newer LHCb value of $R_{πK}$ over the older world average, the analysis is a worked example of how a single discrepant branching ratio can masquerade as a hadronic or new-physics effect; a remeasurement of that ratio is a cheap discriminant.

The failure of the exact-SU(3) fit and the success of a fit with only $O(20\%)$ breaking terms suggests that similar 'golden mode' analyses in other channels should test the symmetry limit explicitly before attributing residual discrepancies to penguins.

The persistent $±35°$ range on $γ$ implies that the proposed Super Flavour Factory programme should not budget these modes as a $γ$ measurement, but rather as a cross-check whose value lies in constraining the subleading amplitudes that feed $β$.

This review was written by a machine. No person has reviewed this paper on Pith yet. A scientist listed on Pith can review it for a fee, and their review is published here whatever it concludes.

Request a human review