betaFeedback

WorksPhysics and Astronomy

Exact flavor conservation at the quark level in causal clusters: strange particles, protons and light nuclei in small collision systems

Abstract

The suppression of strange hadrons in pp and p–Pb collisions at the LHC, including the ϕϕ meson, has recently been described by the exact conservation of strangeness imposed on the quarks of a causally connected cluster before hadronization. Here we extend this approach to the light flavors. The numbers of up and down quarks and antiquarks are conserved in the same cluster, and the suppression of a baryon or a nucleus depends only on how many light quarks of each flavor it contains. With the strangeness sector fixed by the previous analysis, a single constant, the ratio rr of light to strange quark abundances, determines the multiplicity dependence of protons, deuterons, He and hypertriton in pp and p–Pb collisions. Including the finite size of the nuclei, one obtains a very good fit of the data for r=3.76±0.38r=3.76\pm0.38, i.e.\ strange quarks about 0.27 times as abundant as light quarks per flavor, close to the value known from string fragmentation

Introduction

In central nucleus-nucleus collisions the abundances of hadrons and light nuclei are those of an ideal gas of resonances at a temperature T≃155T \simeq155 MeV [1]. In pp and p–Pb collisions the ALICE Collaboration has shown that the yields of strange hadrons, relative to pions, rise smoothly with the charged-particle multiplicity towards the values of central Pb–Pb collisions, more strongly for hadrons with more strange quarks [2, 3, 4]. Light nuclei show a similar, even stronger, rise: the deuteron-to-proton ratio increases by a factor four from the lowest to the highest multiplicities in pp collisions [5, 6], while the proton-to-pion ratio changes by little more than 10% [7].

The suppression of strange hadrons is usually described by the exact conservation of strangeness in a finite volume, imposed on hadrons [8, 9, 10, 11, 12]. In a previous letter [13] we showed that the conservation law should instead be imposed on the strange quarks of a causally connected cluster, before hadronization. For hadrons with open strangeness the two procedures give the same hierarchy of suppression factors. For the ϕ\phi meson, which has no net strangeness, at the hadron level there is no suppression, whereas at the quark level one can distinguish the combinations of a strange quark with the antiquark produced together with it from those with the antiquark of another pair, and the different quantum numbers of the two lead to a suppression of the ϕ\phi in agreement with the data.

The same distinction connects strangeness production to the spin correlation of ΛΛ‾\Lambda\overline{\Lambda} pairs recently observed by the STAR Collaboration [14], interpreted as the survival of the parallel spins of an ss‾s\overline{s} pair created from the vacuum. A Λ\Lambda and a Λ‾\overline{\Lambda} carry this correlation only if their strange quark and antiquark come from the same pair. The fraction of such combinations is fixed by the same pair distribution that determines the suppression of strange hadrons; it decreases with multiplicity, and the spin correlation is predicted to decrease in the same proportion [13]. Strangeness suppression, the ϕ\phi meson and the hyperon spin correlations are thus three consequences of a single mechanism.

In this paper we extend the same description to the light flavors. Up and down quarks are conserved in the same cluster, and the suppression of protons and light nuclei follows from the numbers of light quarks they contain. Since the strangeness sector is fixed by the previous analysis, only one new constant appears: the ratio of light to strange quark abundances. For nuclei we add the overlap between the bound state and the cluster, which is relevant because their size is comparable with that of the cluster.

In Sec. II we summarize the results for strangeness, including the spin correlation of hyperon pairs. In Sec. III we extend the formalism to light quarks, including the finite size of the nuclei. Section IV presents the comparison with the data on protons and light nuclei, Sec. V discusses the cluster size and the meaning of the new constant, and Sec. VI contains the discussion and the conclusions.

Strangeness at the quark level

We first recall the framework and the results of Ref. [13], to which we refer for the derivations.

Causal clusters and pair distribution

Consider a boost-invariant expansion in which partons materialize at the proper time τq\tau_q and hadrons appear at τh\tau_h. Two points can exchange a signal during this interval only if their proper longitudinal distance does not exceed [15]

d=τhτq(τh−τq).(1)d = \sqrt{\frac{\tau_h}{\tau_q}}(\tau_h-\tau_q). \tag*{(1)}

Since flavor is conserved locally, quarks appear together with their antiquarks, whether they come from string breaking or from gluon splitting, and the two members of a pair remain within the distance dd until hadronization. The causally connected region, which we approximate by a sphere of radius R=d/2R=d/2 and volume VV, therefore contains as many quarks as antiquarks of every flavor. We refer to it as a cluster.

We assume that the production in a cluster is statistical from the outset: every configuration compatible with the conservation law is populated according to its statistical weight. If each quark and each antiquark of flavor ff has the weight xf/2x_f/2, the probability of finding nn pairs is [13]

P(n)=1I0(xf)(xf/2)2n(n!)2,(2)P(n)=\frac{1}{I_0(x_f)}\frac{(x_f/2)^{2n}}{(n!)^2}, \tag*{(2)}

where IkI_k denotes the modified Bessel function of order kk. The parameter xfx_f is the average number of quarks plus antiquarks of flavor ff that a cluster of the same volume would contain if the constraint were absent; for a Boltzmann gas at temperature TT,

xf=2Vnf,nf=g2π2mf2TK2(mfT),(3)x_f=2Vn_f,\qquad n_f=\frac{g}{2\pi^2}m_f^2T K_2\left(\frac{m_f}{T}\right), \tag*{(3)}

with g=6g=6 spin and color states. The quantity that enters the hadron yields is the factorial moment

⟨n(n−1)⋯(n−k+1)⟩=(xf2)kIk(xf)I0(xf),(4)\left\langle n(n-1)\cdots(n-k+1)\right\rangle=\left(\frac{x_f}{2}\right)^k\frac{I_k(x_f)}{I_0(x_f)}, \tag*{(4)}

i.e. the number of ways of choosing kk distinct quarks of flavor ff in the cluster. Without the constraint this number would be (xf/2)k(x_f/2)^k. A hadron that needs kk quarks (or kk antiquarks) of flavor ff is therefore reduced, with respect to a very large cluster, by the factor Ik(xf)/I0(xf)I_k(x_f)/I_0(x_f).

Open and hidden strangeness

For strange quarks we write Sk≡Ik(xs)/I0(xs)S_k\equiv I_k(x_s)/I_0(x_s) and then Kaons and Λ\Lambda are reduced by S1S_1, the Ξ\Xi by S2S_2 and the Ω\Omega by S3S_3. The ϕ\phi meson requires one strange quark and one strange antiquark, and the relevant number is ⟨n2⟩\langle n^2\rangle, which is exactly (xs/2)2(x_s/2)^2: the conservation law by itself leaves hidden strangeness unsuppressed. However, the combinations of a quark with an antiquark are of two kinds,

⟨n2⟩=⟨n(n−1)⟩+⟨n⟩=(xs2)2[S2+(1−S2)],(5)\langle n^2\rangle=\langle n(n-1)\rangle+\langle n\rangle=\left(\frac{x_s}{2}\right)^2\left[S_2+(1-S_2)\right], \tag*{(5)}

the first term counting quarks and antiquarks of different pairs, the second a quark with its own partner. The two

kinds differ in their quantum numbers. A pair produced from the vacuum is in the state JPC=0++J^{PC}=0^{++} and cannot become a ϕ\phi, whose quantum numbers are 1−−1^{--}; a pair produced by a gluon has the right JPCJ^{PC} but carries color, and must neutralize it before it can form a ϕ\phi. Denoting by ww the fraction of same-pair combinations that end up in a ϕ\phi, the suppression factor of the ϕ\phi becomes

Cϕ=S2+w(1−S2).(6)C_\phi=S_2+w(1-S_2). \tag*{(6)}

Fit to the strange hadrons

In Ref. [13] these factors were compared with ALICE data on K/πK/\pi, ϕ/π\phi/\pi, Ξ/π\Xi/\pi and Ω/π\Omega/\pi, in pp collisions at s=7\sqrt{s}=7 TeV [7] and in p–Pb collisions at sNN=5.02\sqrt{s_{NN}}=5.02 TeV [3, 16]. The ratios expected for a very large cluster were taken from central Pb–Pb collisions at sNN=2.76\sqrt{s_{NN}}=2.76 TeV, and each measurement was given a normalization factor accounting for the part of its systematic uncertainty that is common to all multiplicity classes. The dependence of the strange-quark content on the charged-particle multiplicity was parametrized as

xs=aNb,N=⟨dNch/dη⟩,(7)x_s=aN^b,\qquad N=\langle dN_{\mathrm{ch}}/d\eta\rangle, \tag*{(7)}

and the fit gave

a=1.54±0.13,b=0.63±0.06,w=0.43±0.09,(8)a=1.54\pm0.13,\qquad b=0.63\pm0.06,\qquad w=0.43\pm0.09, \tag*{(8)}

with χ2=37.9\chi^2=37.9 for 59 degrees of freedom. The pp and p–Pb data lie on the same curves (Figure 2). The two extreme assumptions for the ϕ\phi, no suppression (w=1w=1) and suppression as a doubly strange hadron (w=0w=0), are excluded by 5.7 and 4.8 standard deviations. The average number of strange quarks and antiquarks per cluster grows from 2.6 at N=2.3N=2.3 to 17 at N=45N=45. As an independent check, the ϕ\phi yield in p–Pb collisions at sNN=8.16\sqrt{s_{NN}}=8.16 TeV [17], which was not included in the fit, follows the predicted multiplicity dependence. Throughout this paper aa and bb are kept at the values of Eq. (8).

Figure 2. Strange hadron to pion ratios in p–Pb collisions at sNN=5.02\sqrt{s_{NN}}=5.02 TeV [3, 16] and in pp collisions at s=7\sqrt{s}=7 TeV [7], divided by the large-cluster limits and by the fitted normalization factors, versus charged-particle multiplicity, from [13]. Solid lines: S1S_1, CϕC_{\phi}, S2S_2 and S3S_3 with Eq. (8). Dashed line in (b): w=0w=0. Shaded bands: central Pb–Pb values with their uncertainties.

Spin correlation of Lambda anti-Lambda pairs

The STAR Collaboration has observed a relative polarization of 0.181±0.035±0.0220.181\pm0.035\pm0.022 for ΛΛˉ\Lambda\bar{\Lambda} pairs that are close in rapidity and azimuthal angle, in pp collisions at s=200\sqrt{s}=200 GeV, while no correlation is found for distant pairs or for ΛΛ\Lambda\Lambda pairs [14]. The effect is attributed to the parallel spins of an ssˉs\bar{s} pair produced from the vacuum, which survive hadronization.

Within a cluster, the spin correlation can be carried only by a Λ\Lambda and a Λˉ\bar{\Lambda} whose strange quark and antiquark were produced together; this holds also for pairs produced by a gluon, whose spins are parallel as well. The fraction of such combinations is the second term of Eq. (5) divided by the total,

D=⟨n⟩⟨n2⟩=2xsS1=1−S2,(9)D=\frac{\langle n\rangle}{\langle n^2\rangle}=\frac{2}{x_s}S_1=1-S_2, \tag*{(9)}

Figure 1. Fraction DD of ΛΛˉ\Lambda\bar{\Lambda} combinations whose strange quark and antiquark were produced in the same pair, Eq. (9), with its 1σ1\sigma band (solid line), and suppression of the ϕ\phi, 1−Cϕ1-C_{\phi}, Eq. (10) (dashed line), as functions of multiplicity. From [13].

where we used I0(x)−I2(x)=(2/x)I1(x)I_0(x)-I_2(x)=(2/x)I_1(x). Comparing with Eq. (6) gives

1−Cϕ=(1−w)D.(10)1-C_{\phi}=(1-w)D. \tag*{(10)}

The combinations that preserve the spin correlation are the same ones that, in a fraction $1-w$, fail to produce a ϕ\phi. The dilution of the spin correlation and the suppression of the ϕ\phi are therefore proportional at every multiplicity. With the parameters of Eq. (8), DD decreases from 0.44±0.020.44\pm0.02 at N=4.3N=4.3 to 0.21±0.020.21\pm0.02 at N=16N=16 and 0.11±0.020.11\pm0.02 at $N=45 ({xref}Fig. 1`).

The model thus predicts that the ΛΛˉ\Lambda\bar{\Lambda} spin correlation falls with multiplicity, and disappears in central nucleus-nucleus collisions. The transfer of polarization from the strange quark to the Λ\Lambda, and the dilution due to feed-down, multiply the correlation by a common factor that largely cancels when two multiplicities are compared; the dependence on multiplicity is a cleaner test than the absolute magnitude.

Indeed, as for the magnitude, pp collisions at s=200\sqrt{s}=200 GeV correspond to N≃3N\simeq3 [18], where D=0.53D=0.53 if xsx_s depends on multiplicity only. A sample of pairs entirely in the spin-triplet state, with complete transfer of the spin to the Λ\Lambda and the Λˉ\bar{\Lambda}, would give a relative polarization of 1/31/3; using the SU(6) wave function and the feed-down fractions of PYTHIA, Ref. [14] obtains a maximum of 0.096 instead. With this maximum the dilution leads to 0.05, about three standard deviations below the data. The measured value, however, already exceeds this maximum by two standard deviations, so that even an undiluted correlation (D=1D=1) would not reproduce it. With the factor 1/31/3 and no feed-down correction the dilution gives 0.18, close to the measurement. The comparison of absolute values is thus limited by the polarization transfer rather than by DD.

Light quarks

The argument of Sec. II applies to every flavor. Up and down quarks are also created in pairs, and their numbers are conserved separately in each cluster. We take the two light flavors to be equally abundant, with the same parameter xlx_l, and write

Lk≡Ik(xl)I0(xl).(11)L_k\equiv\frac{I_k(x_l)}{I_0(x_l)}. \tag*{(11)}

A hadron or nucleus that contains kuk_u up, kdk_d down and ksk_s strange quarks (or the same numbers of antiquarks) is suppressed, relative to the limit of a large cluster, by the product of the factors of the three flavors,

F=Iku(xl)I0(xl)Ikd(xl)I0(xl)Iks(xs)I0(xs).(12)F=\frac{I_{k_u}(x_l)}{I_0(x_l)}\frac{I_{k_d}(x_l)}{I_0(x_l)}\frac{I_{k_s}(x_s)}{I_0(x_s)}. \tag*{(12)}

The pion π+\pi^+ has F=L12F=L_1^2 and the ratios to pions considered below follow from Table I. The light-quark factors are close to one for hadrons, since xlx_l is large, but not for nuclei, which contain many light quarks of the same flavor (the deuteron needs three up and three down quarks from the same cluster).

Exact conservation of the light flavors requires that the cluster carry no net baryon number. At LHC energies the antiproton-to-proton ratio at midrapidity is compatible with unity [19], so that the valence quarks of the colliding hadrons do not contribute appreciably to the clusters at midrapidity.

(ku,kd,ks)(k_u,k_d,k_s)FFratio
π+\pi^+(1,1,0)(1,1,0)L22L_2^2—
pp(2,1,0)(2,1,0)L2L1L_2L_1p/πp/\pi: L2/L1L_2/L_1
dd(3,3,0)(3,3,0)L32L_3^2d/pd/p: L32/(L2L1)L_3^2/(L_2L_1)
3He{}^3\mathrm{He}(5,4,0)(5,4,0)L5L4L_5L_43He/p{}^3\mathrm{He}/p: L5L4/(L2L1)L_5L_4/(L_2L_1)
Λ3H{}^3_{\Lambda}\mathrm{H}(4,4,1)(4,4,1)L42S1L_4^2S_1Λ3H/π{}^3_{\Lambda}\mathrm{H}/\pi: L42S1/L22L_4^2S_1/L_2^2
4He{}^4\mathrm{He}(6,6,0)(6,6,0)L62L_6^24He/p{}^4\mathrm{He}/p: L62/(L2L1)L_6^2/(L_2L_1)

Table 1. Quark content (ku,kd,ks)(k_u,k_d,k_s) and suppression factor FF of Eq. (12) for the particles considered; the last column is the factor of the ratio used in the analysis.

Light and strange quarks are produced in the same clusters, and their abundances are proportional. We write

xl=rxs,(13)x_l = r x_s, \tag*{(13)}

where rr is the ratio of the mean number of up quarks (or down quarks) to that of strange quarks that a cluster would contain without the constraint. It is the only new constant: with xsx_s fixed by Eqs. (7) and (8), xlx_l is determined for every multiplicity. In a thermal picture r=nu/nsr = n_u/n_s is fixed by the quark masses through Eq. (3); in string fragmentation 1/r1/r is the strangeness suppression factor λs\lambda_s of the string-breaking probabilities [20, 21]. We determine rr from the data and return to its meaning in § 2.4.

The factors LkL_k and SkS_k refer to a single cluster. An event consists of several clusters along the rapidity axis, each extending over about one unit of rapidity. A baryon or a nucleus is formed from the quarks of one cluster, and its suppression is determined by the number of light quarks available in that cluster, i.e. by xlx_l. The clusters of an event are formed under the same conditions and have the same quark content; the number of clusters per unit of rapidity cancels in the ratios p/πp/\pi, d/pd/p and 3He/p{}^3\mathrm{He}/p, which therefore depend only on the content of a single cluster. This is the same mechanism that applies to strangeness. It differs from the canonical description at the hadron level, in which baryon number is conserved in a correlation volume that has to be chosen separately: the volume required by the deuteron data suppresses protons in pp collisions much more than observed, while at the quark level a single volume describes both.

Finally, for light nuclei in small systems, the cluster size is comparable with their radii. Equation (12) counts the quarks available in the cluster, but not whether the bound state fits in it. For a nucleus of AA nucleons with a Gaussian wave function of size parameter rAr_A, formed in a Gaussian source of radius RR, the probability is reduced by [22, 23]

GA(R)=[1+(rA2R)2]−3(A−1)/2.(14)G_A(R) = \left[1+\left(\frac{r_A}{2R}\right)^2\right]^{-3(A-1)/2}. \tag*{(14)}

where RR is the radius of the cluster, obtained from xlx_l through Eq. (3) with a light quark mass of 300 MeV.

We use rd=3.2r_d=3.2 fm and r3He=2.48r_{{}^3\mathrm{He}}=2.48 fm [23]; for 4He{}^4\mathrm{He} we use rA≃2.0r_A\simeq2.0 fm, obtained from its measured charge radius [29] with the same relation between rAr_A and the root-mean-square radius as for 3He{}^3\mathrm{He} in Ref. [23]. The hypertriton is treated as a deuteron and a Λ\Lambda, with factor G2(R;rd)G2(R;rΛ)G_2(R;r_d)G_2(R;r_\Lambda) and rΛ=8.7r_\Lambda=8.7 fm, which corresponds to a root-mean-square separation of about 10 fm between the Λ\Lambda and the deuteron [25, 26] and lies within the range 6.8–14.1 fm given in Ref. [23].

Comparison with data

Data and fit

The proton-to-pion ratio has been measured in pp collisions at s=7\sqrt{s}=7 TeV in ten multiplicity classes, with NN from 2.3 to 21 [7], and in p–Pb collisions at sNN=5.02\sqrt{s_{NN}}=5.02 TeV in seven classes, with NN from 4.3 to 45 [16]. Proton and deuteron yields are available in pp collisions at s=5.02\sqrt{s}=5.02 TeV in nine classes, with NN from 2.4 to 18.5 [5], and at 13 TeV in three high-multiplicity classes, with NN from 30 to 36 [6]; from these we use the proton per unit of charged-particle density and the ratio d/pd/p.

The same two analyses provide the 3He{}^3\mathrm{He}-to-proton ratio, in four points with NN from 4.9 to 31.5. Finally, the hypertriton yield has been measured in the 40% highest-multiplicity p–Pb collisions at 5.02 TeV [24], at N=29.4N=29.4, and we use it per unit of charged-particle density.

The large-cluster limits of d/pd/p and 3He/p{}^3\mathrm{He}/p are the values in central Pb–Pb collisions at 2.76 TeV [27], 3.6×10−33.6\times10^{-3} and 9.3×10−69.3\times10^{-6}. For the hypertriton we use the yield measured in the 10% most central Pb–Pb collisions at 2.76 TeV [28], with the branching ratio of 25% for the decay into 3Heπ−{}^3\mathrm{He}\pi^- adopted in Ref. [24], per unit of charged-particle density: 1.07×10−71.07\times10^{-7}.

As in Ref. [13], each data set has a normalization factor, constrained to unity within the uncertainty of its reference value (16, 25 and 30%).

The absolute proton yields are not compared with Pb–Pb collisions, where they are affected by annihilation in the hadronic phase; their normalizations are free, and only their dependence on multiplicity is used. The strangeness parameters are fixed to Eq. (8), and the only parameter of the model is rr.

The fit gives

r=3.76±0.38,(15)r = 3.76 \pm0.38, \tag*{(15)}

with χ2=27.4\chi^2=27.4 for 42 degrees of freedom. The details of the procedure are given in the Supplemental Material.

The size factor of Eq. (14) is required. Without it, the suppression of the nuclei must come entirely from quark counting: the fit then gives r≃2.1r\simeq2.1, which suppresses the protons too much, and χ2\chi^2 increases from 27.4 to 50.6, mostly from the proton data. The effect of the size factor at fixed rr is shown in Figure 4. The result is stable with respect to the size attributed to the hypertriton: with rΛ=6.8r_{\Lambda}=6.8 and 10.3 fm, χ2=26.3\chi^2=26.3 and 28.8, with the same rr.

Figure 4. (a) Deuteron-to-proton ratio, (b) 3He{}^{3}\mathrm{He}-to-proton ratio, (c) hypertriton yield per unit of charged-particle density, versus charged-particle multiplicity. Data: ALICE [5, 6, 24]. Solid lines: model with the size factor of Eq. (14); dashed lines: same rr and normalization, without the size factor. Shaded bands: central Pb–Pb values with their uncertainties [27, 28].

Protons

The model predicts p/π∝L2/L1p/\pi\propto L_2/L_1. Since xlx_l is between 10 and 64 in the range of the data, the suppression is small: 15% at the lowest multiplicity in pp collisions and 2% at the highest in p–Pb collisions. The proton-to-pion ratio is constant within a few percent above N≃4N \simeq4, and decreases by about 12% in the lowest multiplicity class of pp collisions at 7 TeV Figure 3(a). The 17 points are described with χ2=1.5\chi^2=1.5.

Figure 3. (a) Proton-to-pion ratio in p–Pb collisions at sNN=5.02\sqrt{s_{NN}} = 5.02 TeV [16] and in pp collisions at s=7\sqrt{s} = 7 TeV [7], divided by the fitted normalization, versus charged-particle multiplicity. Solid line: L2/L1L_2/L_1 with Eqs. (8) and (15). (b) Protons per unit of charged-particle density in pp collisions at 5.02 TeV [5] and 13 TeV [6], with the model.

The number of protons per unit of charged-particle density in pp collisions at 5.02 and 13 TeV is depicted in Figure 3(b) (χ2=15.6\chi^2=15.6 for 12 points). In fact, this ratio, unlike p/πp/\pi, includes the dependence on multiplicity of the number of pions per charged particle, which in p–Pb collisions decreases by about 6% over the measured range. The three points at 13 TeV lie below the extrapolation of those at 5.02 TeV, and give 8.1 of the total.

Most protons come from resonance decays, which are not included in Table I. Their effect can be estimated roughly. About 45% of the protons come from Δ\Delta decays; the Δ++\Delta^{++} and Δ−\Delta^{-} need three light quarks of one flavor, and for this fraction the factor L2L1L_2L_1 is replaced by (L3+L2L1)/2(L_3+L_2L_1)/2. About 30% of the charged pions come from ρ0\rho^0, ω\omega and η\eta, which contain a light quark and antiquark of the same flavor and, like the ϕ\phi, have the factor L2+w(1−L2)L_2+w(1-L_2) instead of L12L_1^2. With these fractions the ratio p/πp/\pi changes from 0.894 to 0.868 at N=4N=4, from 0.940 to 0.924 at N=10N=10 and from 0.973 to 0.966 at N=36N=36. Resonance decays change the multiplicity dependence of the proton-to-pion ratio by less than 3%.

Deuterons

Relative to the proton, the deuteron suppression factor is L32/(L2L1)L_3^2/(L_2L_1), multiplied by the size factor G2G_2. The two effects are comparable: at N=4.3N=4.3 quark counting gives 0.63 and the size factor 0.54, at N=36N=36 they give 0.89 and 0.75. The total suppression goes from 0.24 at N=2.4N=2.4 to 0.67 at N=36N=36. The twelve d/pd/p points in pp collisions at 5.02 and 13 TeV are described with χ2=0.5\chi^2=0.5 Figure 4(a), with normalization 1.02, i.e. the data extrapolate to the central Pb–Pb value.

The rise of d/pd/p with multiplicity, by a factor four, is thus explained with no parameter specific to the deuteron.

3He and hypertriton

For 3He{}^{3}\mathrm{He}, with content (5, 4, 0), quark counting gives 0.31 and the size factor 0.47 at N=5N=5, and 0.70 and 0.69 at N=31.5N=31.5; the total suppression relative to protons is 0.15 and 0.48. The four points are described with χ2=4.8\chi^2=4.8 Figure 4(b); the point at N=13.6N=13.6 lies 1.8σ1.8\sigma below the curve. The normalization is 0.61, 2σ2\sigma below the central Pb–Pb value.

The hypertriton has content (4, 4, 1). At N=29.4N=29.4 quark counting gives 0.70, but the size factor is 0.17, because the bound state is much larger than the cluster. With the large-cluster limit the model gives a yield per unit of charged-particle density of 1.3×10−81.3\times10^{-8}, against the measured (2.1±0.7)×10−8(2.1\pm0.7)\times10^{-8} Figure 4(c).

Predictions for nuclei

The model makes two predictions for nuclei that have not yet been measured as functions of multiplicity.

The first concerns the double ratio (Λ3H/3He)/(Λ/p)({}^{3}_{\Lambda}\mathrm{H}/{}^{3}\mathrm{He})/(\Lambda/p), which compares the hypertriton with a nucleus of the same mass number. Quark counting alone would increase it in small clusters, since the hypertriton needs one light quark fewer than 3He{}^{3}\mathrm{He}; the size factor, however, acts in the opposite direction and dominates, because the hypertriton is much larger than 3He{}^{3}\mathrm{He}. The double ratio is therefore reduced with respect to its large-cluster value, by a factor 0.14 at N=4N=4 and 0.26 at N=30N=30.

The second concerns 4He{}^{4}\mathrm{He}, which needs six light quarks of each flavor and has a size parameter rA≃2.0r_A\simeq2.0 fm: relative to protons it is suppressed by a factor 0.14 at N=10N=10 and 0.34 at N=30N=30.

Cluster size and the light-to-strange ratio

Table II gives, for several multiplicities, the parameters of the cluster and the main suppression factors.

The cluster radius, obtained from xlx_l with a light constituent quark mass of 300 MeV, grows from 2.0 fm at N=2.3N=2.3 to 3.7 fm at N=45N=45, and the hadronization time obtained from Eq. (1) with τq=1\tau_q=1 fm from 3.2 to 4.5 fm. This determination relies on the nuclei: the size factor of Eq. (14) compares the radius of the cluster with the size of the bound state, which is fixed by its wave function, whereas the strange hadrons depend only on xsx_s in Eq. (3).

The corresponding volume depends on the mass of the strange quark. With T=155T=155 MeV and the constituent mass ms=460m_s=460 MeV, as suggested by the ratio ms/mu≃1.5m_s/m_u\simeq1.5 obtained from the baryon magnetic moments [31] and by ms/mu≃1.5m_s/m_u\simeq1.5–1.6, with ms−mu≃173m_s-m_u\simeq173 MeV for both mesons and baryons, from a fit to the hadron spectrum [32], the values of xsx_s of Eq. (8) give a radius from 1.5 fm at N=2.3N=2.3 to 2.9 fm at N=45N=45 and a hadronization time from 2.8 to 3.9 fm.

To check whether this is compatible with the nuclei, we have repeated the fit of Sec. IV with the size factor computed with this radius. The fit is as good as the main one, χ2=25.6\chi^2=25.6 for 42 degrees of freedom, with $r=5.8\pm1.3 and the normalization of d/pd/p equal to 1.14. The nuclei fix the radius only together with rr, and a radius about 20% smaller is compensated by a weaker suppression from the light quarks. The χ2\chi^2 is nearly independent of the strange quark mass between 300 and 500 MeV.

The size of the cluster obtained from strangeness with the constituent strange quark mass is therefore compatible with the one required by the nuclei.

The value r=3.76r = 3.76 corresponds to 1/r=0.271/r = 0.27: a cluster contains about 0.27 strange quarks per up (or down) quark; with the radius obtained from strangeness one finds 1/r=0.17±0.041/r = 0.17 \pm0.04. Both values are close to the strangeness suppression factor of string fragmentation, λs\lambda_s, which fixes the relative probability of creating an ssˉs\bar{s} pair rather than a uuˉu\bar{u} or ddˉd\bar{d} pair when a string breaks [20]. The standard tunes of PYTHIA use λs≃0.2\lambda_s \simeq0.2 [21, 30], and larger effective values are obtained in high-multiplicity events when the interaction of overlapping strings is taken into account [33].

Discussion and conclusions

The proposed description has a simple structure. Indeed, one function of multiplicity, the mean number of strange quarks per cluster, determined by the strange hadrons, and one constant, the ratio of light to strange quark abundances, determined by the nuclei, fix the multiplicity dependence of all the hadrons and nuclei considered.

Protons are nearly unsuppressed because the light-quark parameter xlx_l is large; nuclei are strongly suppressed because they need many light quarks of the same flavor in the same cluster, and because they are as large as the cluster.

For the deuteron, the size factor of Eq. (14) is the one of coalescence models [22, 23]. The difference is that here the source radius is not a free function of multiplicity but the radius of the causal cluster.

Two points require further study. (i) The normalization of the multiplicity dependence of the yields is fixed by the data. tion of 3He^{3}\mathrm{He} is two standard deviations below the central Pb–Pb value, and the protons per charged particle at 13 TeV lie below those at 5.02 TeV; both could indicate a dependence on the collision energy that is not included here.

NNxsx_sxlx_lRR (fm)τh\tau_h (fm)L1L_1L2L_2GdG_dp/πp/\pid/pd/p3He/p^{3}\mathrm{He}/p
2.32.69.81.963.190.950.810.470.850.230.055
4.33.914.52.243.420.960.870.540.900.340.13
9.86.524.42.663.750.980.920.630.940.480.26
16.18.933.32.953.960.980.940.680.960.560.36
23.211.241.93.194.130.990.950.710.960.610.42
30.513.349.83.374.270.990.960.740.970.650.47
45.016.963.73.664.460.990.970.770.980.690.54

Table 2. Cluster parameters and suppression factors versus multiplicity, with Eqs. (8) and (15). RR is the cluster radius and τh\tau_h the hadronization time (τq=1 fm\tau_q = 1\ \mathrm{fm}); GdG_d is the size factor of the deuteron; the last three columns are the suppression of p/πp/\pi, d/pd/p and 3He/p^{3}\mathrm{He}/p relative to the large-cluster limit.

(ii) Resonance decays are included only roughly; their effect on protons is small, and a comprehensive assessment of feed-down would yield more precise estimates.

In conclusion, exact flavor conservation in a causal cluster, imposed on the quarks before hadronization, extends from strangeness to the light flavors with one new constant, the ratio of light to strange quark abundances. Together with the finite size of the nuclei, it describes the multiplicity dependence of protons, deuterons, 3He^{3}\mathrm{He} and hypertriton in pp and p–Pb collisions at the LHC.

The model predicts the multiplicity dependence of 4He^{4}\mathrm{He} and of the double ratio (Λ3H/3He)/(Λ/p)(^{3}_{\Lambda}\mathrm{H}/^{3}\mathrm{He})/(\Lambda/p), and, from the strangeness sector, a decrease of the ΛΛˉ\Lambda\bar{\Lambda} spin correlation with multiplicity proportional to the suppression of the ϕ\phi.

References

JHEP 03, 148 (2015), arXiv:1412.6259.

References

  1. [1]A. Andronic, P. Braun-Munzinger, K. Redlich, and J. Stachel, Nature 561, 321 (2018), arXiv:1710.09425.DOI
  2. [2]J. Adam et al. (ALICE Collaboration), Nature Phys. 13, 535 (2017), arXiv:1606.07424.
  3. [3]J. Adam et al. (ALICE Collaboration), Phys. Lett. B 758, 389 (2016), arXiv:1512.07227.
  4. [4]S. Acharya et al. (ALICE Collaboration), Eur. Phys. J. C 80, 167 (2020), arXiv:1908.01861.
  5. [5]S. Acharya et al. (ALICE Collaboration), Eur. Phys. J. C 82, 289 (2022), arXiv:2112.00610.
  6. [6]S. Acharya et al. (ALICE Collaboration), JHEP 01, 106 (2022), arXiv:2109.13026.
  7. [7]S. Acharya et al. (ALICE Collaboration), Phys. Rev. C 99, 024906 (2019), arXiv:1807.11321.
  8. [8]J. Rafelski and M. Danos, Phys. Lett. B 97, 279 (1980).
  9. [9]R. Hagedorn and K. Redlich, Z. Phys. C 27, 541 (1985).
  10. [10]S. Hamieh, K. Redlich, and A. Tounsi, Phys. Lett. B 486, 61 (2000).
  11. [11]I. Kraus, J. Cleymans, H. Oeschler, K. Redlich, and S. Wheaton, Phys. Rev. C 76, 064903 (2007).
  12. [12]I. Kraus, J. Cleymans, H. Oeschler, and K. Redlich, Phys. Rev. C 79, 014901 (2009).
  13. [13]P. Castorina and H. Satz, Strangeness suppression, the ϕ meson and ΛΛ̄ spin correlations from exact strangeness conservation at the quark level, arXiv:2610.20449 [hep-ph], submitted to Phys. Lett. B.
  14. [14]B. E. Aboona et al. (STAR Collaboration), Nature 650, 65 (2026), arXiv:2506.05499.
  15. [15]P. Castorina and H. Satz, Int. J. Mod. Phys. E 23, 1450019 (2014), arXiv:1310.6932.DOI
  16. [16]J. Adam et al. (ALICE Collaboration), Eur. Phys. J. C 76, 245 (2016), arXiv:1601.07868.
  17. [17]S. Acharya et al. (ALICE Collaboration), Phys. Rev. C 107, 055201 (2023), arXiv:2110.10042.
  18. [18]B. I. Abelev et al. (STAR Collaboration), Phys. Rev. C 79, 034909 (2009), arXiv:0808.2041.DOI
  19. [19]K. Aamodt et al. (ALICE Collaboration), Phys. Rev. Lett. 105, 072002 (2010), arXiv:1006.5432.
  20. [20]B. Andersson, G. Gustafson, G. Ingelman, and T. Sjöstrand, Phys. Rep. 97, 31 (1983).
  21. [21]C. Bierlich et al., SciPost Phys. Codeb. 2022, 8 (2022), arXiv:2203.11601.arxiv.org/abs/2203.11601
  22. [22]R. Scheibl and U. Heinz, Phys. Rev. C 59, 1585 (1999).
  23. [23]F. Bellini and A. P. Kalweit, Phys. Rev. C 99, 054905 (2019).
  24. [24]S. Acharya et al. (ALICE Collaboration), Phys. Rev. Lett. 128, 252003 (2022), arXiv:2107.10627.
  25. [25]H. Nemura, Y. Suzuki, Y. Fujiwara, and C. Nakamoto, Prog. Theor. Phys. 103, 929 (2000), arXiv:nucl-th/9912065.DOI
  26. [26]F. Hildenbrand and H. W. Hammer, Phys. Rev. C 100, 034002 (2019); 102, 039901(E) (2020), arXiv:1904.05818.DOI
  27. [27]J. Adam et al. (ALICE Collaboration), Phys. Rev. C 93, 024917 (2016), arXiv:1506.08951.
  28. [28]J. Adam et al. (ALICE Collaboration), Phys. Lett. B 754, 360 (2016), arXiv:1506.08453.
  29. [29]J. J. Krauth et al. (CREMA Collaboration), Nature 589, 527 (2021).
  30. [30]P. Skands, S. Carrazza, and J. Rojo, Eur. Phys. J. C 74, 3024 (2014), arXiv:1404.5630.DOI
  31. [31]S. Navas et al. (Particle Data Group), Phys. Rev. D 110, 030001 (2024), review “Quark model” by C. Amsler, V. Crede, and T. DeGrand.
  32. [32]M. Karliner, S. Nussinov, and J. L. Rosner, Phys. Rev. D 95, 034011 (2017), arXiv:1611.00348.DOI
  33. [33]C. Bierlich, G. Gustafson, L. Lönnblad, and A. Tarasov, JHEP 03, 148 (2015), arXiv:1412.6259.

Paper details

vlnc:W_v9xyrjpg
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents