The Effect of Penguins in the B_d to J/psi K^0 CP asymmetry

The useful thing here: Ciuchini, Pierini and Silvestrini replace the QCD-factorization estimate of penguin pollution in B0 → J/ψ K0 (below 10⁻³) with a data-driven one, extracting the range of |P_GIM2 − P2| from the SU(3) partner B0 → J/ψ π0 and feeding it back into the J/ψ K0 fit. The answer, ΔS ≈ 0.000 ± 0.017, is comparable to the experimental systematic (±0.020). That is worth knowing and worth citing; it changes how much weight J/ψ K0 should carry against the penguin modes in a UT fit. The method is sensible and the cross-checks are real: |E2 − P2| from J/ψ π0 agrees with the J/ψ K0 determination, and the 100%-SU(3)-error variant (±0.015) lands in the same place. The CKM inputs come from UTfit with the J/ψ K0 bound removed, so the comparison isn't circular. Credit for that; it would have been easy to get wrong. Now the soft spots. The one that matters: the abstract says ±0.012 with 95% range [−0.025, 0.024]; Eq. (3) says ±0.017 with [−0.035, 0.033]. No fit in the text produces ±0.012. Since the paper's entire product is the width of that distribution, this is not cosmetic — a reader folding the abstract number into a world average would understate the theory error by roughly 30%. It needs reconciling before anyone quotes it. Second, the width is prior-dominated, and the authors say so plainly: the J/ψ π0 bound cuts out a long tail, and without it ΔS would be O(1). Fine, but they never show how the width moves if you take the 95% rather than 99.9% range, or drop the |P_GIM2 − P2| < 2|E2 − P2| cut. A short sensitivity table would settle whether ±0.017 is robust or an artifact of where the boundary sits. The two-solution ambiguity in the J/ψ π0 fit is resolved by hand, but with the |E2 − P2| agreement as justification, which I find acceptable. Who should read it: anyone extracting sin 2β or doing UT fits with J/ψ K0, and anyone still repeating the 'theoretically clean' line. It deserves a referee — the analysis is honest and the point is real, and a referee is exactly who would catch the abstract/body mismatch.

The paper's only quantitative product is the width of the Delta S distribution, so any instability in that width is load-bearing by construction. Three numbers appear. Abstract [abstract:s2]: Delta S = 0.000 +- 0.012, 95% range [-0.025, 0.024]. Body Eq. (3): Delta S = 0.000 +- 0.017, 95% range [-0.035, 0.033]. The 100%-SU(3)-error variant: Delta S = 0.001 +- 0.015, 95% range [-0.025, 0.026]. The abstract's lower endpoint (-0.025) coincides with the SU(3)-error variant's lower endpoint, but its upper endpoint (0.024) does not (0.026), and its +-0.012 matches neither quoted fit. No procedure in the text produces +-0.012. This is not cosmetic: the abstract number is what a reader would fold into a sin 2 beta world average, and a ~30-40% understatement of the theory error changes the relative weight of J/psi K0 versus penguin modes in a UT fit. The discrepancy also points at a deeper sensitivity the authors themselves flag. They state that the J/psi pi0 bound on |P_GIM2 - P2| is 'extremely effective in cutting out a long tail', and that without it 'we would have obtained values of Delta S of order one'. That bound is a 99.9% upper limit taken from a fit with two solutions, the second (|P_GIM2 - P2| = 2.87 +- 0.43) discarded by hand on SU(3) grounds, and further tightened by a cut |P_GIM2 - P2| < 2|E2 - P2| imposed without a first-principles argument. The final +-0.017 is therefore prior-dominated: it is set by a hard boundary transferred from another channel, not by the J/psi K0 data themselves. A number that moves between +-0.012 and +-0.017 across the paper's own presentation is consistent with exactly this kind of prior sensitivity, and the authors do not reconcile it.

The reader's verdict of CONDITIONAL is defensible and I would not move it. Where I differ is in ranking. The reader names the SU(3) transfer of |P_GIM2 - P2| from J/psi pi0 to J/psi K0 as the weakest assumption, and the authors do defend that transfer: they cross-check |E2 - P2| between the two channels (1.22 +- 0.15 versus 1.44 +- 0.05, consistent), and they quote a 100%-SU(3)-error variant that agrees with the main result. An acknowledged and cross-checked assumption is a weaker target than an unacknowledged numerical inconsistency. The abstract/body mismatch is the sharper concern because it is unacknowledged anywhere in the text, it concerns the one number the paper exists to deliver, and it is checkable without re-deriving the physics. The reader lists it as a 'mechanical' point; I would promote it, because the size of the theory error is the paper's entire claim, and a 40% spread in that size across the paper's own pages is a correctness risk for anyone using the number. The prior-sensitivity argument reinforces this: the authors state plainly that without the J/psi pi0 bound Delta S could be of order one, which means the quoted uncertainty is a conditional statement about a transferred hard cutoff, not a measurement. That is fine physics, but it must be stated in the abstract with the number that the body actually supports. On the positive side, the three-step method is clearly laid out, the two-solution ambiguity is disclosed rather than hidden, the |E2 - P2| cross-check is a genuine independent consistency test, and the 100%-SU(3)-error variant is exactly the right robustness check to run. Nothing here suggests the physics conclusion is wrong; the qualitative claim that the penguin uncertainty is order 10^-2 rather than 10^-3 survives either number. The defect is in reporting, and it is fixable by a single re-run plus an abstract correction. Hence UNCHANGED: CONDITIONAL, with the numerical reconciliation added as an explicit condition.

If the quoted uncertainty holds, it changes the bookkeeping of the sin(2 beta) world average: a theory error of order 0.017 is no longer negligible against the experimental systematic and would reduce the weight of the J/psi K0 modes relative to penguin-dominated channels in global fits. The methodological idea is the valuable part: using the SU(3) partner channel J/psi pi0 to bound a doubly Cabibbo-suppressed quantity that J/psi K0 alone cannot constrain, rather than relying on the BSS/QCD-factorization estimate that the authors correctly note is only formally valid here. The paper also deserves credit for two things that are easy to omit: the explicit SU(3)-breaking stress test (the 100% error variant, which reproduces the main result), and the frank labelling of the shape-transferred result Delta S = 0.020 +- 0.007 as illustrative only and not for phenomenology. The comparison is not circular, since the target sin(2 beta) is an external UTfit input with the J/psi K0 bound removed and the penguin range comes from an independent channel. The result is a null result with a quantified uncertainty, which is a legitimate and useful output, and it is falsifiable in the concrete sense,

The Standard Model predicts that the CP coefficient $S_{CP}$ measured in $B^0$–$J/ψK^0$ equals $sin 2β$, the phase of $B_d$–$B̄_d$ mixing, and that equality is the benchmark against which every claimed hint of new physics in the quark sector is judged. The paper asks how far penguin contractions — loop-level, long-distance contributions carrying the weak phase of $V_{ub}$ rather than that of the dominant tree amplitude — can push the measured $S_{CP}$ away from $sin 2β$. The offending combination is doubly Cabibbo suppressed in $J/ψK^0$ and cannot be pinned down from that channel alone, so the authors take its allowed size from the SU(3) partner decay $B^0$–$J/ψπ^0$, where the same combination is not suppressed, and propagate that range back. The result is $ΔS equiv S_{CP}(J/ψK) − sin 2β = 0.000 ± 0.017$, with 95% range $[−0.035, 0.033]$ — an order of magnitude larger than earlier estimates built on the BSS mechanism and QCD factorization, and comparable to the systematic error the B-factories now quote. For completeness: the abstract of this version states a tighter $±0.012$ ($[−0.025, 0.024]$ at 95%), while Eq. (3) of the text gives $±0.017$; the text value is the one the fit produces.

What carries the argument. The argument is carried by an operator-product-expansion bookkeeping in which all contractions of the effective weak operators are grouped into renormalization-group-invariant hadronic parameters, so that $A(B^0$–$J/ψK^0) = V_{cb}^{*}V_{cs}(E_2 − P_2) + V_{ub}^{*}V_{us}(P_2^{GIM} − P_2)$. The first combination is the dominant tree amplitude and is fixed from the measured branching ratio; the second is the penguin combination with a different weak phase, and it is the only object that can rotate $S_{CP}$ away from $sin 2β$. Because it is doubly Cabibbo suppressed in $J/ψK^0$, its magnitude is instead extracted from $B^0$–$J/ψπ^0$, where the same parameters appear with CKM elements of comparable size, and the resulting 99.9% range $|P_2^{GIM} − P_2| < 2.3$ (with the cut $|P_2^{GIM} − P_2| < 2|E_2 − P_2|$ removing the swapped solution) is then imposed as a flat prior in the $J/ψK^0$ fit. The statistical treatment throughout is the Bayesian flat-prior scheme used in the unitarity-triangle fits, with the CKM input taken from a fit that excludes the $sin 2β$ measurement itself.

Core claim. Using a fit to the available $B$-factory data, the authors claim that the deviation of the measured $S_{CP}$ in $B^0$–$J/ψK^0$ from $sin 2β$ is not negligible at the $10^{-3}$ level, as previous treatments concluded, but is instead $ΔS = 0.000 ± 0.017$ with a 95% probability range of $[−0.035, 0.033]$. The claim rests on a chain of three fits: the branching ratio of $B^0$–$J/ψK^0$ fixes $|E_2 − P_2|$; the branching ratio, $C_{CP}$ and $S_{CP}$ of $B^0$–$J/ψπ^0$ fix $|E_2 − P_2|$, $|P_2^{GIM} − P_2|$ and the relative strong phase $δ_P$, with the two-fold ambiguity resolved by requiring agreement with the $J/ψK^0$ determination of $|E_2 − P_2|$; and finally $J/ψK^0$ is refitted with $|P_2^{GIM} − P_2|$ scanned over the range that the $J/ψπ^0$ fit allows. The conclusion the authors draw is that this uncertainty must be folded into the value and error of $sin 2β$ quoted from $S_{CP}$, and that it can be reduced by better measurements of $B^0$–$J/ψπ^0$ rather than by further theory input.

The premise it rests on. The allowed range of the penguin combination $|P_2^{GIM} − P_2|$, measured in $B^0$–$J/ψπ^0$, is assumed to hold equally in $B^0$–$J/ψK^0$; the authors state this directly, adopt the 99.9% range $[0, 2.3]$ as a flat prior, and acknowledge that SU(3) breaking and the neglected emission-annihilation term could invalidate the transfer.

The $sin 2β$ value extracted from $S_{CP}(J/ψK^0)$ should carry an additional theoretical error of about $0.017$, comparable to the current experimental systematic, before it is compared with unitarity-triangle fits or with penguin-dominated modes such as $B$–$φK_S$ and $B$–$η'K_S$.

Global fits that treat the $J/ψK^0$ asymmetry as an exactly clean measurement of $sin 2β$ are overstating its precision, and any tension they report with other determinations should be re-evaluated against the widened error bar.

Improving the measurements of $B^0$–$J/ψπ^0$ — its branching ratio, $C_{CP}$ and $S_{CP}$ — tightens the allowed range of $|P_2^{GIM} − P_2|$ and hence shrinks $ΔS$ without requiring new theoretical input.

If the shape, not merely the range, of the $|P_2^{GIM} − P_2|$ distribution from $J/ψπ^0$ were trusted, the deviation would sharpen to $ΔS = 0.020 ± 0.007$; the authors present this only as an illustration because of SU(3)-breaking and emission-annihilation uncertainties.

Adding a 100% error to the SU(3) relation between the hadronic parameters of the two channels gives $ΔS = 0.001 ± 0.015$ ($[−0.025, 0.026]$ at 95%), which the authors take as evidence that their procedure does not sizeably overestimate the uncertainty.

The same logic applies to every mode used as a $sin 2β$ cross-check: a penguin amplitude with a different weak phase can only be bounded through a flavour-partner channel, so the precision of these cross-checks is set by how well the partner is measured, not by how well the golden mode itself is measured.

If future $B^0$–$J/ψπ^0$ data push the allowed $|P_2^{GIM} − P_2|$ range downward, the widened error bar claimed here would shrink toward the older $10^{-3}$ estimates, making the size of $ΔS$ a direct diagnostic of whether the SU(3) transfer was conservative or merely loose.

A discrepancy between $sin 2β$ from $J/ψK^0$ and from a penguin-dominated mode such as $φK_S$ would be easier to interpret as new physics if the $J/ψK^0$ side were quoted with this $±0.017$ rather than with a negligible theory error, since the two sides would then be compared on more equal footing.

The internal mismatch between the abstract's $±0.012$ and the text's $±0.017$ suggests the quoted uncertainty is sensitive to the precise treatment of the $|P_2^{GIM} − P_2|$ prior, which is worth checking before the number is adopted in global fits.

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