Constraining the Gluon Shadowing in (Lead) Nuclei with the(LHC) Heavy-Flavour Data

J.P. LansbergIPN Orsay – Paris-Sud U. –CNRS/IN2P3 – Universite Paris-Saclay

Exploring the Perfect Liquid, September 6-8, 2018Max Planck Institute for Astrophysics, Garching, Germany

in collaboration with A. Kusina, I. Schienbein and H.S. Shaobased on PRL 121 (2018) 052004 & EPJC 77 (2017) 1

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 1 / 27

Introduction

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 2 / 27

Introduction

Cross-sections in nuclear collisions are modi�ed

Shadowing

Anti-shadowing

EMC-e�ect

Fermi-motion

We translate these modi�cations into universal quantities: nuclear PDFs (nPDFs)J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 3 / 27

Introduction

Nuclear PDFs (nPDFs) allow one to get information on:

the structure of the nuclei in terms of quarks and gluons;the initial state of heavy-ion collisions at the LHC and RHIC

esp. to use perturbative probes of the QGP to study its properties.

� nPDFs can not be computed and similarly to the proton PDFs are�t to experimental data. Only the evolution is perturbative

� Can heavy- avour data help us better constrain the nPDFs ?

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 4 / 27

Introduction

Nuclear PDFs (nPDFs) allow one to get information on:

the structure of the nuclei in terms of quarks and gluons;the initial state of heavy-ion collisions at the LHC and RHIC

esp. to use perturbative probes of the QGP to study its properties.

� nPDFs can not be computed and similarly to the proton PDFs are�t to experimental data. Only the evolution is perturbative

� Can heavy- avour data help us better constrain the nPDFs ?

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 4 / 27

Introduction

Nuclear PDFs (nPDFs) allow one to get information on:

the structure of the nuclei in terms of quarks and gluons;the initial state of heavy-ion collisions at the LHC and RHIC

esp. to use perturbative probes of the QGP to study its properties.

� nPDFs can not be computed and similarly to the proton PDFs are�t to experimental data. Only the evolution is perturbative

� Can heavy- avour data help us better constrain the nPDFs ?

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 4 / 27

Introduction

Nuclear PDFs (nPDFs) allow one to get information on:

the structure of the nuclei in terms of quarks and gluons;the initial state of heavy-ion collisions at the LHC and RHIC

esp. to use perturbative probes of the QGP to study its properties.

� nPDFs can not be computed and similarly to the proton PDFs are�t to experimental data. Only the evolution is perturbative

� Can heavy- avour data help us better constrain the nPDFs ?

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 4 / 27

Introduction

Nuclear PDFs (nPDFs) allow one to get information on:

the structure of the nuclei in terms of quarks and gluons;the initial state of heavy-ion collisions at the LHC and RHIC

esp. to use perturbative probes of the QGP to study its properties.

� nPDFs can not be computed and similarly to the proton PDFs are�t to experimental data. Only the evolution is perturbative

� Can heavy- avour data help us better constrain the nPDFs ?

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 4 / 27

Di�erences with the free-proton PDFs

�eoretical status of FactorisationParametrisation – more parameters to model the A-dependenceDi�erent data sets – much less data:

1

10

102

103

104

105

10-4

10-3

10-2

10-1

1

fixed target DIS and DY

LHC dijets

LHC W & Z

CHORUS neutrino data

PHENIX π0

x

Q2[GeV

2 ]

EPPS16 dataset

Less data� less constraining power�more assumptions (�xing) about �ttingparameters

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 5 / 27

Di�erences with the free-proton PDFs

�eoretical status of FactorisationParametrisation – more parameters to model the A-dependenceDi�erent data sets – much less data:

1

10

102

103

104

105

10-4

10-3

10-2

10-1

1

fixed target DIS and DY

LHC dijets

LHC W & Z

CHORUS neutrino data

PHENIX π0

x

Q2[GeV

2 ]

EPPS16 dataset

Less data� less constraining power�more assumptions (�xing) about �ttingparameters

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 5 / 27

Di�erences with the free-proton PDFs

�eoretical status of FactorisationParametrisation – more parameters to model the A-dependenceDi�erent data sets – much less data:

1

10

102

103

104

105

10-4

10-3

10-2

10-1

1

fixed target DIS and DY

LHC dijets

LHC W & Z

CHORUS neutrino data

PHENIX π0

x

Q2[GeV

2 ]

EPPS16 dataset

Less data� less constraining power�more assumptions (�xing) about �ttingparameters

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 5 / 27

nPDF framework

ParametrisationPDF of nucleus (A - mass, Z - charge)

f �A,Z�i �x,Q� �ZAf p~Ai �x,Q� �

A � ZA

f n~Ai �x,Q�

bound proton PDFs

nCTEQ15 [arXiv:1509.00792]

xf p~Ai �x, Q0� � c0xc1 �1 � x�c2 ec3x�1 � ec4 x�c5

ck � ck�A� � ck,0 � ck,1 �1 � A�ck,2 �

EPPS16 [arXiv:1612.05741]

f p~Ai �x, Q� � RAi �x, Q� f pi �x, Q�,

RAi �x, Q0� �

¢¨¦¨¤

a0 � a1�x � xa�2 x B xab0 � b1xα � b2x2α � b3x3α xa B x B xec0 � �c1 � c2x��1 � x��β xe B x B 1

di � di�A� � di�Aref�� AAref

�γi�di�Aref ��1� ,with di � ai , bi , . . . and Aref � 12

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 6 / 27

Current nPDFs

10-5 10-4 10-3 10-2 10-1 100

x

0.0

0.1

0.2

0.3

0.4

0.5

xu

Pb

val

Q=2 GeV

nCTEQ15

alternative nCTEQ fits

EPPS16

10-5 10-4 10-3 10-2 10-1 100

x

0.0

0.1

0.2

0.3

0.4

0.5

0.6

xd

Pb

val

10-5 10-4 10-3 10-2 10-1 100

x

0

1

2

3

4

5

6

7

8

9

xgP

b

10-5 10-4 10-3 10-2 10-1 100

x

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

Ru Pb

Q=2 GeV

nCTEQ15

alternative nCTEQ fits

EPPS16

10-5 10-4 10-3 10-2 10-1 100

x

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

Rd Pb

10-5 10-4 10-3 10-2 10-1 100

x

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

Rg Pb

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 7 / 27

[arXiv:1509.00792, arXiv:1012.1178]

[arXiv:1612.05741]

�����������������

������������������������������������������������������������� ��[ N. Armesto, 2015 LHeC Workshop,Chavannes-de-Bogis, June 26th 2015]

Heavy- avour LHC data

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 9 / 27

Available heavy- avour pPb LHC data�

D0 J~ψ B� J~ψ Υ�1S�µ0

¼4M2

D0 � P2T ,D0

¼M2

J~ψ � P2T ,J~ψ

½4M2

B � � MBMJ~ψ

PT ,J~ψ�2 ¼M2

Υ�1S� � P2T ,Υ�1S�

p+p data LHCb [1] LHCb [2,3] LHCb [2,3] ALICE [4], ATLAS [5],CMS [6], LHCb [7,8]

RpPb data ALICE [9], ALICE [10,11], LHCb [12] ALICE [13], ATLAS [14],LHCb [15] LHCb [16,12] LHCb [17]

[1] LHCb, R. Aaij et al., JHEP 06, 147 (2017), 1610.02230.[2] LHCb, R. Aaij et al., Eur. Phys. J. C71, 1645 (2011), 1103.0423.[3] LHCb, R. Aaij et al., JHEP 06, 064 (2013), 1304.6977.[4] ALICE, B. B. Abelev et al., Eur. Phys. J. C74, 2974 (2014), 1403.3648.[5] ATLAS, G. Aad et al., Phys. Rev. D87, 052004 (2013), 1211.7255.[6] CMS, S. Chatrchyan et al., Phys. Lett. B727, 101 (2013), 1303.5900.[7] LHCb, R. Aaij et al., Eur. Phys. J. C72, 2025 (2012), 1202.6579.[8] LHCb, R. Aaij et al., JHEP 11, 103 (2015), 1509.02372.[9] ALICE, B. B. Abelev et al., Phys. Rev. Lett. 113, 232301 (2014), 1405.3452.[10] ALICE, J. Adam et al., JHEP 06, 055 (2015), 1503.07179.[11] ALICE, B. B. Abelev et al., JHEP 02, 073 (2014), 1308.6726.[12] LHCb, R. Aaij et al., (2017), 1706.07122.[13] ALICE, B. B. Abelev et al., Phys. Lett. B740, 105 (2015), 1410.2234.[14] �e ATLAS collaboration, (2015), ATLAS-CONF-2015-050.[15] LHCb, R. Aaij et al., JHEP 1710 (2017) 090, 1707.02750.[16] LHCb, R. Aaij et al., JHEP 02, 072 (2014), 1308.6729.[17] LHCb, R. Aaij et al., JHEP 07, 094 (2014), 1405.5152.

� : At the time we performed our study (� we can add more in the future)

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 10 / 27

Expected nuclear e�ects on heavy quark(onium) productionin pA collisions

Nuclear modi�cation of PDFs: initial-state e�ectEnergy loss (w.r.t. pp collisions): initial-state or �nal-state e�ectBreak up of the quarkonium in the nuclear matter: �nal-state e�ectBreak up by comoving particles: �nal-state e�ectColour �ltering of intrinsic QQ pairs: initial-state e�ect...

We assume that the leading twist factorisation is valid – ONLYmodi�cations ofPDFs are present� “shadowing-only” hypothesis.

� we de facto do not consider excited quarkonium data.

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 11 / 27

Expected nuclear e�ects on heavy quark(onium) productionin pA collisions

Nuclear modi�cation of PDFs: initial-state e�ectEnergy loss (w.r.t. pp collisions): initial-state or �nal-state e�ectBreak up of the quarkonium in the nuclear matter: �nal-state e�ectBreak up by comoving particles: �nal-state e�ectColour �ltering of intrinsic QQ pairs: initial-state e�ect...

We assume that the leading twist factorisation is valid – ONLYmodi�cations ofPDFs are present� “shadowing-only” hypothesis.

� we de facto do not consider excited quarkonium data.

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 11 / 27

Automating the computation of nuclear PDFe�ects on HF production

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 12 / 27

An automated code to evaluate the impact of nuclear PDF IJPL, H.S. Shao Eur.Phys.J. C77 (2017) 1

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 13 / 27

Partonic scattering cross section �t from pp data with a Crystal Ball functionparametrising SAgg�HX S2 C.H. Kom, A. Kulesza, W.J. Stirling PRL 107 (2011) 082002

A way to evade the quarkonium-production-mechanism controversy ?To some extent, I would say ”yes”.

Applied to J~ψ, Υ, D and B: it can be extended to all the probes produced in 2� 2partonic processes with a single partonic contribution�e key point to compute nPDF e�ect is to have a partonic cross sectionValidated (for D0) by the state-of-the-art pQCD code FONLLAny nPDF set available in LHAPDF5 or 6 can be usedCurrently limited to processes dominated by a single partonic channel

(gg or qq, ...)Not yet interfaced to a Glauber model

[no centrality and no combination with other nuclear e�ects]

An automated code to evaluate the impact of nuclear PDF IJPL, H.S. Shao Eur.Phys.J. C77 (2017) 1

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 13 / 27

Partonic scattering cross section �t from pp data with a Crystal Ball functionparametrising SAgg�HX S2 C.H. Kom, A. Kulesza, W.J. Stirling PRL 107 (2011) 082002

A way to evade the quarkonium-production-mechanism controversy ?To some extent, I would say ”yes”.

Applied to J~ψ, Υ, D and B: it can be extended to all the probes produced in 2� 2partonic processes with a single partonic contribution�e key point to compute nPDF e�ect is to have a partonic cross sectionValidated (for D0) by the state-of-the-art pQCD code FONLLAny nPDF set available in LHAPDF5 or 6 can be usedCurrently limited to processes dominated by a single partonic channel

(gg or qq, ...)Not yet interfaced to a Glauber model

[no centrality and no combination with other nuclear e�ects]

An automated code to evaluate the impact of nuclear PDF IJPL, H.S. Shao Eur.Phys.J. C77 (2017) 1

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 13 / 27

Partonic scattering cross section �t from pp data with a Crystal Ball functionparametrising SAgg�HX S2 C.H. Kom, A. Kulesza, W.J. Stirling PRL 107 (2011) 082002

A way to evade the quarkonium-production-mechanism controversy ?To some extent, I would say ”yes”.

Applied to J~ψ, Υ, D and B: it can be extended to all the probes produced in 2� 2partonic processes with a single partonic contribution�e key point to compute nPDF e�ect is to have a partonic cross sectionValidated (for D0) by the state-of-the-art pQCD code FONLLAny nPDF set available in LHAPDF5 or 6 can be usedCurrently limited to processes dominated by a single partonic channel

(gg or qq, ...)Not yet interfaced to a Glauber model

[no centrality and no combination with other nuclear e�ects]

An automated code to evaluate the impact of nuclear PDF IJPL, H.S. Shao Eur.Phys.J. C77 (2017) 1

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 13 / 27

Partonic scattering cross section �t from pp data with a Crystal Ball functionparametrising SAgg�HX S2 C.H. Kom, A. Kulesza, W.J. Stirling PRL 107 (2011) 082002

A way to evade the quarkonium-production-mechanism controversy ?To some extent, I would say ”yes”.

Applied to J~ψ, Υ, D and B: it can be extended to all the probes produced in 2� 2partonic processes with a single partonic contribution�e key point to compute nPDF e�ect is to have a partonic cross sectionValidated (for D0) by the state-of-the-art pQCD code FONLLAny nPDF set available in LHAPDF5 or 6 can be usedCurrently limited to processes dominated by a single partonic channel

(gg or qq, ...)Not yet interfaced to a Glauber model

[no centrality and no combination with other nuclear e�ects]

An automated code to evaluate the impact of nuclear PDF IJPL, H.S. Shao Eur.Phys.J. C77 (2017) 1

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 13 / 27

Partonic scattering cross section �t from pp data with a Crystal Ball functionparametrising SAgg�HX S2 C.H. Kom, A. Kulesza, W.J. Stirling PRL 107 (2011) 082002

A way to evade the quarkonium-production-mechanism controversy ?To some extent, I would say ”yes”.

Applied to J~ψ, Υ, D and B: it can be extended to all the probes produced in 2� 2partonic processes with a single partonic contribution�e key point to compute nPDF e�ect is to have a partonic cross sectionValidated (for D0) by the state-of-the-art pQCD code FONLLAny nPDF set available in LHAPDF5 or 6 can be usedCurrently limited to processes dominated by a single partonic channel

(gg or qq, ...)Not yet interfaced to a Glauber model

[no centrality and no combination with other nuclear e�ects]

An automated code to evaluate the impact of nuclear PDF IJPL, H.S. Shao Eur.Phys.J. C77 (2017) 1

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 13 / 27

Partonic scattering cross section �t from pp data with a Crystal Ball functionparametrising SAgg�HX S2 C.H. Kom, A. Kulesza, W.J. Stirling PRL 107 (2011) 082002

A way to evade the quarkonium-production-mechanism controversy ?To some extent, I would say ”yes”.

Applied to J~ψ, Υ, D and B: it can be extended to all the probes produced in 2� 2partonic processes with a single partonic contribution�e key point to compute nPDF e�ect is to have a partonic cross sectionValidated (for D0) by the state-of-the-art pQCD code FONLLAny nPDF set available in LHAPDF5 or 6 can be usedCurrently limited to processes dominated by a single partonic channel

(gg or qq, ...)Not yet interfaced to a Glauber model

[no centrality and no combination with other nuclear e�ects]

An automated code to evaluate the impact of nuclear PDF IJPL, H.S. Shao Eur.Phys.J. C77 (2017) 1

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 13 / 27

Partonic scattering cross section �t from pp data with a Crystal Ball functionparametrising SAgg�HX S2 C.H. Kom, A. Kulesza, W.J. Stirling PRL 107 (2011) 082002

A way to evade the quarkonium-production-mechanism controversy ?To some extent, I would say ”yes”.

Applied to J~ψ, Υ, D and B: it can be extended to all the probes produced in 2� 2partonic processes with a single partonic contribution�e key point to compute nPDF e�ect is to have a partonic cross sectionValidated (for D0) by the state-of-the-art pQCD code FONLLAny nPDF set available in LHAPDF5 or 6 can be usedCurrently limited to processes dominated by a single partonic channel

(gg or qq, ...)Not yet interfaced to a Glauber model

[no centrality and no combination with other nuclear e�ects]

An automated code to evaluate the impact of nuclear PDF IJPL, H.S. Shao Eur.Phys.J. C77 (2017) 1

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 13 / 27

Partonic scattering cross section �t from pp data with a Crystal Ball functionparametrising SAgg�HX S2 C.H. Kom, A. Kulesza, W.J. Stirling PRL 107 (2011) 082002

A way to evade the quarkonium-production-mechanism controversy ?To some extent, I would say ”yes”.

Applied to J~ψ, Υ, D and B: it can be extended to all the probes produced in 2� 2partonic processes with a single partonic contribution�e key point to compute nPDF e�ect is to have a partonic cross sectionValidated (for D0) by the state-of-the-art pQCD code FONLLAny nPDF set available in LHAPDF5 or 6 can be usedCurrently limited to processes dominated by a single partonic channel

(gg or qq, ...)Not yet interfaced to a Glauber model

[no centrality and no combination with other nuclear e�ects]

An automated code to evaluate the impact of nuclear PDF IJPL, H.S. Shao Eur.Phys.J. C77 (2017) 1

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 13 / 27

Partonic scattering cross section �t from pp data with a Crystal Ball functionparametrising SAgg�HX S2 C.H. Kom, A. Kulesza, W.J. Stirling PRL 107 (2011) 082002

A way to evade the quarkonium-production-mechanism controversy ?To some extent, I would say ”yes”.

Applied to J~ψ, Υ, D and B: it can be extended to all the probes produced in 2� 2partonic processes with a single partonic contribution�e key point to compute nPDF e�ect is to have a partonic cross sectionValidated (for D0) by the state-of-the-art pQCD code FONLLAny nPDF set available in LHAPDF5 or 6 can be usedCurrently limited to processes dominated by a single partonic channel

(gg or qq, ...)Not yet interfaced to a Glauber model

[no centrality and no combination with other nuclear e�ects]

An automated code to evaluate the impact of nuclear PDF IIJPL, H.S. Shao Eur.Phys.J. C77 (2017) 1

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 14 / 27

Extensive comparisons directly with data,which make sense if nPDFs are the only nuclear e�ect

Conversely, one can test this hypothesis by comparing our curves to dataGlobal agreement

?� only nPDFs matter

One can go further in the data comparison with reweighting (see later)and then HF-data inclusion in nPDF �ts

Bonus: since the pp yields are �t, the procedure sometimes hints at normalisationissues (absent in RFB) which could otherwise be misinterpreted as nuclearsuppressions/enhancementsLast but not least: the automation of the evaluation allows one to study di�erent nPDFsets AND the scale uncertainties: better control of the theory uncertainties

An automated code to evaluate the impact of nuclear PDF IIJPL, H.S. Shao Eur.Phys.J. C77 (2017) 1

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 14 / 27

Extensive comparisons directly with data,which make sense if nPDFs are the only nuclear e�ect

Conversely, one can test this hypothesis by comparing our curves to dataGlobal agreement

?� only nPDFs matter

One can go further in the data comparison with reweighting (see later)and then HF-data inclusion in nPDF �ts

Bonus: since the pp yields are �t, the procedure sometimes hints at normalisationissues (absent in RFB) which could otherwise be misinterpreted as nuclearsuppressions/enhancementsLast but not least: the automation of the evaluation allows one to study di�erent nPDFsets AND the scale uncertainties: better control of the theory uncertainties

An automated code to evaluate the impact of nuclear PDF IIJPL, H.S. Shao Eur.Phys.J. C77 (2017) 1

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 14 / 27

Extensive comparisons directly with data,which make sense if nPDFs are the only nuclear e�ect

Conversely, one can test this hypothesis by comparing our curves to dataGlobal agreement

?� only nPDFs matter

One can go further in the data comparison with reweighting (see later)and then HF-data inclusion in nPDF �ts

Bonus: since the pp yields are �t, the procedure sometimes hints at normalisationissues (absent in RFB) which could otherwise be misinterpreted as nuclearsuppressions/enhancementsLast but not least: the automation of the evaluation allows one to study di�erent nPDFsets AND the scale uncertainties: better control of the theory uncertainties

An automated code to evaluate the impact of nuclear PDF IIJPL, H.S. Shao Eur.Phys.J. C77 (2017) 1

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 14 / 27

Extensive comparisons directly with data,which make sense if nPDFs are the only nuclear e�ect

Conversely, one can test this hypothesis by comparing our curves to dataGlobal agreement

?� only nPDFs matter

One can go further in the data comparison with reweighting (see later)and then HF-data inclusion in nPDF �ts

Bonus: since the pp yields are �t, the procedure sometimes hints at normalisationissues (absent in RFB) which could otherwise be misinterpreted as nuclearsuppressions/enhancementsLast but not least: the automation of the evaluation allows one to study di�erent nPDFsets AND the scale uncertainties: better control of the theory uncertainties

An automated code to evaluate the impact of nuclear PDF IIJPL, H.S. Shao Eur.Phys.J. C77 (2017) 1

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 14 / 27

Extensive comparisons directly with data,which make sense if nPDFs are the only nuclear e�ect

Conversely, one can test this hypothesis by comparing our curves to dataGlobal agreement

?� only nPDFs matter

One can go further in the data comparison with reweighting (see later)and then HF-data inclusion in nPDF �ts

Bonus: since the pp yields are �t, the procedure sometimes hints at normalisationissues (absent in RFB) which could otherwise be misinterpreted as nuclearsuppressions/enhancementsLast but not least: the automation of the evaluation allows one to study di�erent nPDFsets AND the scale uncertainties: better control of the theory uncertainties

An automated code to evaluate the impact of nuclear PDF IIJPL, H.S. Shao Eur.Phys.J. C77 (2017) 1

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 14 / 27

Extensive comparisons directly with data,which make sense if nPDFs are the only nuclear e�ect

Conversely, one can test this hypothesis by comparing our curves to dataGlobal agreement

?� only nPDFs matter

One can go further in the data comparison with reweighting (see later)and then HF-data inclusion in nPDF �ts

Bonus: since the pp yields are �t, the procedure sometimes hints at normalisationissues (absent in RFB) which could otherwise be misinterpreted as nuclearsuppressions/enhancementsLast but not least: the automation of the evaluation allows one to study di�erent nPDFsets AND the scale uncertainties: better control of the theory uncertainties

Results for pA collisions using nCTEQ15 &EPPS16 out-of-the-box

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 15 / 27

Some J~ψ comparisonsR

pPb

ycms(J/ψ)

Prompt J/ψ production at √ sNN=5.02 TeV LHC

LHCb dataEPPS16NLO

EPS09NLOnCTEQ15

PT<15

0.4

0.6

0.8

1

1.2

1.4

1.6

-5 -4 -3 -2 -1 0 1 2 3 4 5

HELAC-Onia 2.0

RpP

b

ycms(J/ψ)

Prompt J/ψ production at √ sNN=5.02 TeV LHC

ALICE inclusive data

EPPS16NLOEPS09NLO

nCTEQ15PT<15

0.4

0.6

0.8

1

1.2

1.4

1.6

-5 -4 -3 -2 -1 0 1 2 3 4 5

HELAC-Onia 2.0

RF

B

|ycms(J/ψ)|

Prompt J/ψ production at √ sNN=5.02 TeV LHC

LHCb dataEPPS16NLO

EPS09NLOnCTEQ15

PT<14

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

0 1 2 3 4 5

HELAC-Onia 2.0

RF

B

|ycms(J/ψ)|

Prompt J/ψ production at √ sNN=5.02 TeV LHC

ATLAS dataEPS09NLO

nCTEQ158<PT<30

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

0 0.5 1 1.5 2HELAC-Onia 2.0

RF

B

|ycms(J/ψ)|

Prompt J/ψ production at √ sNN=5.02 TeV LHC

LHCb dataEPPS16NLO

EPS09NLOnCTEQ15

PT<14

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

0 1 2 3 4 5

HELAC-Onia 2.0

Prompt J/ψ production at √ sNN=5.02 TeV LHC

-4.46<ycms<-2.96

EPPS16NLOEPS09NLO

nCTEQ15ALICE inclusive data 0.8

1

1.2

1.4

1.6

RpP

b

-1.37<ycms<0.43

0.4

0.6

0.8

1

1.2

PT(J/ψ) [GeV]

2.03<ycms<3.53

0.4

0.6

0.8

1

1.2

0 1 2 3 4 5 6 7 8 9 10

HELAC-Onia 2.0

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 16 / 27

More results: Υ�1S� and ... ηcJPL, H.S. Shao Eur.Phys.J. C77 (2017) 1

Rp

Pb

ycms(Υ(1S))

Inclusive Υ(1S) production at √sNN=5.02 TeV LHC

LHCb data

EPS09LOEPS09NLO

nCTEQ15PT<15

0.4

0.6

0.8

1

1.2

1.4

1.6

-5 -4 -3 -2 -1 0 1 2 3 4 5HELAC-Onia 2.0

Rp

Pb

ycms(Υ(1S))

Inclusive Υ(1S) production at √sNN=5.02 TeV LHC

ALICE data

EPS09LOEPS09NLO

nCTEQ15PT>0

0.4

0.6

0.8

1

1.2

1.4

1.6

-5 -4 -3 -2 -1 0 1 2 3 4 5

HELAC-Onia 2.0

RF

B

|ycms(Υ(1S))|

Inclusive Υ(1S) production at √sNN=5.02 TeV LHC

LHCb data

EPS09LOEPS09NLO

nCTEQ15PT<15

0.4

0.6

0.8

1

1.2

1.4

1.6

0 1 2 3 4 5

HELAC-Onia 2.0

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 17 / 27

Our 8 TeV predictionsin J.L. Albacete, et al..NPA 972 (2018) 18

More results: Υ�1S� and ... ηcJPL, H.S. Shao Eur.Phys.J. C77 (2017) 1

Rp

Pb

ycms(Υ(1S))

Inclusive Υ(1S) production at √sNN=5.02 TeV LHC

LHCb data

EPS09LOEPS09NLO

nCTEQ15PT<15

0.4

0.6

0.8

1

1.2

1.4

1.6

-5 -4 -3 -2 -1 0 1 2 3 4 5HELAC-Onia 2.0

Rp

Pb

ycms(Υ(1S))

Inclusive Υ(1S) production at √sNN=5.02 TeV LHC

ALICE data

EPS09LOEPS09NLO

nCTEQ15PT>0

0.4

0.6

0.8

1

1.2

1.4

1.6

-5 -4 -3 -2 -1 0 1 2 3 4 5

HELAC-Onia 2.0

RF

B

|ycms(Υ(1S))|

Inclusive Υ(1S) production at √sNN=5.02 TeV LHC

LHCb data

EPS09LOEPS09NLO

nCTEQ15PT<15

0.4

0.6

0.8

1

1.2

1.4

1.6

0 1 2 3 4 5

HELAC-Onia 2.0

Rp

Pb

ycms(ηc(1S))

Prompt ηc(1S) production at √sNN=5.02 TeV LHC

PT>6.5 GeV EPS09LO

EPS09NLO

nCTEQ15

0.4

0.6

0.8

1

1.2

1.4

-5 -4 -3 -2 -1 0 1 2 3 4 5

HELAC-Onia 2.0

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 17 / 27

Our 8 TeV predictionsin J.L. Albacete, et al..NPA 972 (2018) 18

Impact on nPDFs: reweighting analysis1

1From now on, all nPDF uncertainties are 68%CLJ.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 18 / 27

Reweighting: the principle behind

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 19 / 27

* N is the # of eigensets, Nrep is the # of constructedreplicas

* f0 is the ”central-value” of the nPDF vector (i.e. offunctions of x) in Nflavour dimension

* f���i (i > �1 � N�) is the ”upper/lower value” function ofa given eigenset i

* Rki is a number randomly choosen for each set of �k, i�(thus fixed for all Nflavour ) according to a standardNormal distribution

* fk is the constructed vector

* Tkj is the predicted theory for data j and replica k

* T is the tolerance factor (for 68% CL: 13 for nCTEQ15;19 for EPPS16)

Global data (or theory) uncertainties can be dealt with adjusting Tkj

When a replica k describes well the data, it gets a higher weight wkthanks to a smaller χ2kThe nPDF are then modified –reweighted– since the initial set ofreplicas is altered. If replicas closer to (further from) the central valueare favoured, the nPDF uncertainty is reduced (enlarged). nPDFuncertainties for any flavour can easily be redrawnAny other observables can also be redrawn (pA dσ , RpA , RFB , ...)

Reweighting: the principle behind

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 19 / 27

* N is the # of eigensets, Nrep is the # of constructedreplicas

* f0 is the ”central-value” of the nPDF vector (i.e. offunctions of x) in Nflavour dimension

* f���i (i > �1 � N�) is the ”upper/lower value” function ofa given eigenset i

* Rki is a number randomly choosen for each set of �k, i�(thus fixed for all Nflavour ) according to a standardNormal distribution

* fk is the constructed vector

* Tkj is the predicted theory for data j and replica k

* T is the tolerance factor (for 68% CL: 13 for nCTEQ15;19 for EPPS16)

Global data (or theory) uncertainties can be dealt with adjusting Tkj

When a replica k describes well the data, it gets a higher weight wkthanks to a smaller χ2kThe nPDF are then modified –reweighted– since the initial set ofreplicas is altered. If replicas closer to (further from) the central valueare favoured, the nPDF uncertainty is reduced (enlarged). nPDFuncertainties for any flavour can easily be redrawnAny other observables can also be redrawn (pA dσ , RpA , RFB , ...)

Reweighting: the principle behind

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 19 / 27

* N is the # of eigensets, Nrep is the # of constructedreplicas

* f0 is the ”central-value” of the nPDF vector (i.e. offunctions of x) in Nflavour dimension

* f���i (i > �1 � N�) is the ”upper/lower value” function ofa given eigenset i

* Rki is a number randomly choosen for each set of �k, i�(thus fixed for all Nflavour ) according to a standardNormal distribution

* fk is the constructed vector

* Tkj is the predicted theory for data j and replica k

* T is the tolerance factor (for 68% CL: 13 for nCTEQ15;19 for EPPS16)

Global data (or theory) uncertainties can be dealt with adjusting Tkj

When a replica k describes well the data, it gets a higher weight wkthanks to a smaller χ2kThe nPDF are then modified –reweighted– since the initial set ofreplicas is altered. If replicas closer to (further from) the central valueare favoured, the nPDF uncertainty is reduced (enlarged). nPDFuncertainties for any flavour can easily be redrawnAny other observables can also be redrawn (pA dσ , RpA , RFB , ...)

Reweighting: the principle behind

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 19 / 27

* N is the # of eigensets, Nrep is the # of constructedreplicas

* f0 is the ”central-value” of the nPDF vector (i.e. offunctions of x) in Nflavour dimension

* f���i (i > �1 � N�) is the ”upper/lower value” function ofa given eigenset i

* Rki is a number randomly choosen for each set of �k, i�(thus fixed for all Nflavour ) according to a standardNormal distribution

* fk is the constructed vector

* Tkj is the predicted theory for data j and replica k

* T is the tolerance factor (for 68% CL: 13 for nCTEQ15;19 for EPPS16)

Global data (or theory) uncertainties can be dealt with adjusting Tkj

When a replica k describes well the data, it gets a higher weight wkthanks to a smaller χ2kThe nPDF are then modified –reweighted– since the initial set ofreplicas is altered. If replicas closer to (further from) the central valueare favoured, the nPDF uncertainty is reduced (enlarged). nPDFuncertainties for any flavour can easily be redrawnAny other observables can also be redrawn (pA dσ , RpA , RFB , ...)

Reweighting with D0 data

nCTEQ15

µF=µ0

Original Reweighted LHCb data ALICE data

0.6

0.8

1

1.2

µF=2.0µ0

0.6

0.8

1

1.2

ycms(D0)

µF=0.5µ0

RpPb

0.6

0.8

1

1.2

-3 -2 -1 0 1 2 3

EPPS16

µF=µ0

µF=2.0µ0

ycms(D0)

µF=0.5µ0

-3 -2 -1 0 1 2 3

HELAC-Onia 2.0

LHCb [JHEP 1710 (2017) 090, 1707.02750]ALICE [PRL113, 232301 (2014), 1405.3452]

Initial description of data is good for bothnCTEQ15 and EPPS16.

Substantial reduction of uncertainty especiallyfor EPPS16.

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 20 / 27

Reweighting with D0 data

nCTEQ15

µF=µ0

Original Reweighted LHCb data ALICE data

0.6

0.8

1

1.2

µF=2.0µ0

0.6

0.8

1

1.2

ycms(D0)

µF=0.5µ0

RpPb

0.6

0.8

1

1.2

-3 -2 -1 0 1 2 3

EPPS16

µF=µ0

µF=2.0µ0

ycms(D0)

µF=0.5µ0

-3 -2 -1 0 1 2 3

HELAC-Onia 2.0

LHCb [JHEP 1710 (2017) 090, 1707.02750]ALICE [PRL113, 232301 (2014), 1405.3452]

Initial description of data is good for bothnCTEQ15 and EPPS16.

Substantial reduction of uncertainty especiallyfor EPPS16.

If we include factorisation scale uncertaintyerrors increase and it can become thedominant uncertainty.

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 20 / 27

Reweighting with J~ψ data

nCTEQ15

µF=µ0

Original Reweighted LHCb data

0.6

0.8

1

1.2

µF=2.0µ0

0.6

0.8

1

1.2

ycms(J/ψ)

µF=0.5µ0

RpPb

0.6

0.8

1

1.2

-3 -2 -1 0 1 2 3

EPPS16

µF=µ0

µF=2.0µ0

ycms(J/ψ)

µF=0.5µ0

-3 -2 -1 0 1 2 3

HELAC-Onia 2.0

LHCb [JHEP 02, 072 (2014), 1308.6729; PLB 774 (2017) 159,1706.07122 ]ALICE [JHEP 06, 055 (2015), 1503.07179; JHEP 02, 073(2014), 1308.6726]

Again we observe a good agreement with thedata; the scale uncertainty becomesimportant.

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 20 / 27

Reweighting with B� J~ψ data

nCTEQ15

µF=µ0

Original Reweighted LHCb data

0.6

0.8

1

1.2

µF=2.0µ0

0.6

0.8

1

1.2

ycms(J/ψ)

µF=0.5µ0

RpPb

0.6

0.8

1

1.2

-3 -2 -1 0 1 2 3

EPPS16

µF=µ0

µF=2.0µ0

ycms(J/ψ)

µF=0.5µ0

-3 -2 -1 0 1 2 3

HELAC-Onia 2.0

LHCb [PLB 774 (2017) 159, 1706.07122 ]

Scale uncertainty is reduced compared to theD0 and J~ψ case.

Data are not yet precise enough to givesubstantial constraints on nPDFs (but if theprecision rises there is big potential).

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 20 / 27

Reweighting with Υ�1S� data

nCTEQ15

µF=µ0

Original Reweighted ALICE data ATLAS data

0.6

0.8

1

1.2

µF=2.0µ0

0.6

0.8

1

1.2

ycms(ϒ(1S))

µF=0.5µ0

RpPb

0.6

0.8

1

1.2

-4 -3 -2 -1 0 1 2 3

EPPS16

µF=µ0

µF=2.0µ0

ycms(ϒ(1S))

µF=0.5µ0

-4 -3 -2 -1 0 1 2 3

HELAC-Onia 2.0

ALICE [PLB 740, 105 (2015), 1410.2234]ATLAS [ATLAS-CONF-2015-050 (updated in: 1709.03089)]

With the current precision we don’t get anyadditional constraints on the nPDFs.

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 20 / 27

Results of the reweighting processnCTEQ15

Q2=4 GeV2

D0

Original µF=µ0 µF=2.0µ0 µF=0.5µ0

0.4

0.6

0.8

1

1.2

1.4

1.6

x

J/ψ

RPbg

0.4

0.6

0.8

1

1.2

1.4

1.6

10-510-410-310-210-1

B→J/ψ

x

ϒ(1S)

10-510-410-310-210-1

HELAC-Onia 2.0

EPPS16

Q2=4 GeV2

D0

Original µF=µ0 µF=2.0µ0 µF=0.5µ0

0.4

0.6

0.8

1

1.2

1.4

1.6

x

J/ψ

RPbg

0.4

0.6

0.8

1

1.2

1.4

1.6

10-510-410-310-210-1

B→J/ψ

x

ϒ(1S)

10-510-410-310-210-1

HELAC-Onia 2.0

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 21 / 27

X Global coherence of the data constraints: necessary condition to assume a shadowing-only approachX First clear experimental observation on gluon SHADOWING at low x; Visible reduction of theEPPS16 uncertainties; con�rmation of the extrapolation done in nCTEQ15

X �e scale ambiguity for D and J~ψ production is now the dominant uncertaintyX Non-prompt J~ψ are really promising if improved data can be obtained

X Con�rmation of the existence of a gluon anti-shadowing : Rg�0.05 ß x @ß 0.1� A 1

Results of the reweighting processnCTEQ15

Q2=4 GeV2

D0

Original µF=µ0 µF=2.0µ0 µF=0.5µ0

0.4

0.6

0.8

1

1.2

1.4

1.6

x

J/ψ

RPbg

0.4

0.6

0.8

1

1.2

1.4

1.6

10-510-410-310-210-1

B→J/ψ

x

ϒ(1S)

10-510-410-310-210-1

HELAC-Onia 2.0

EPPS16

Q2=4 GeV2

D0

Original µF=µ0 µF=2.0µ0 µF=0.5µ0

0.4

0.6

0.8

1

1.2

1.4

1.6

x

J/ψ

RPbg

0.4

0.6

0.8

1

1.2

1.4

1.6

10-510-410-310-210-1

B→J/ψ

x

ϒ(1S)

10-510-410-310-210-1

HELAC-Onia 2.0

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 21 / 27

X Global coherence of the data constraints: necessary condition to assume a shadowing-only approachX First clear experimental observation on gluon SHADOWING at low x; Visible reduction of theEPPS16 uncertainties; con�rmation of the extrapolation done in nCTEQ15

X �e scale ambiguity for D and J~ψ production is now the dominant uncertaintyX Non-prompt J~ψ are really promising if improved data can be obtained

X Con�rmation of the existence of a gluon anti-shadowing : Rg�0.05 ß x @ß 0.1� A 1

Results of the reweighting processnCTEQ15

Q2=4 GeV2

D0

Original µF=µ0 µF=2.0µ0 µF=0.5µ0

0.4

0.6

0.8

1

1.2

1.4

1.6

x

J/ψ

RPbg

0.4

0.6

0.8

1

1.2

1.4

1.6

10-510-410-310-210-1

B→J/ψ

x

ϒ(1S)

10-510-410-310-210-1

HELAC-Onia 2.0

EPPS16

Q2=4 GeV2

D0

Original µF=µ0 µF=2.0µ0 µF=0.5µ0

0.4

0.6

0.8

1

1.2

1.4

1.6

x

J/ψ

RPbg

0.4

0.6

0.8

1

1.2

1.4

1.6

10-510-410-310-210-1

B→J/ψ

x

ϒ(1S)

10-510-410-310-210-1

HELAC-Onia 2.0

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 21 / 27

X Global coherence of the data constraints: necessary condition to assume a shadowing-only approachX First clear experimental observation on gluon SHADOWING at low x; Visible reduction of theEPPS16 uncertainties; con�rmation of the extrapolation done in nCTEQ15

X �e scale ambiguity for D and J~ψ production is now the dominant uncertaintyX Non-prompt J~ψ are really promising if improved data can be obtained

X Con�rmation of the existence of a gluon anti-shadowing : Rg�0.05 ß x @ß 0.1� A 1

Results of the reweighting processnCTEQ15

Q2=4 GeV2

D0

Original µF=µ0 µF=2.0µ0 µF=0.5µ0

0.4

0.6

0.8

1

1.2

1.4

1.6

x

J/ψ

RPbg

0.4

0.6

0.8

1

1.2

1.4

1.6

10-510-410-310-210-1

B→J/ψ

x

ϒ(1S)

10-510-410-310-210-1

HELAC-Onia 2.0

EPPS16

Q2=4 GeV2

D0

Original µF=µ0 µF=2.0µ0 µF=0.5µ0

0.4

0.6

0.8

1

1.2

1.4

1.6

x

J/ψ

RPbg

0.4

0.6

0.8

1

1.2

1.4

1.6

10-510-410-310-210-1

B→J/ψ

x

ϒ(1S)

10-510-410-310-210-1

HELAC-Onia 2.0

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 21 / 27

X Global coherence of the data constraints: necessary condition to assume a shadowing-only approachX First clear experimental observation on gluon SHADOWING at low x; Visible reduction of theEPPS16 uncertainties; con�rmation of the extrapolation done in nCTEQ15

X �e scale ambiguity for D and J~ψ production is now the dominant uncertaintyX Non-prompt J~ψ are really promising if improved data can be obtained

X Con�rmation of the existence of a gluon anti-shadowing : Rg�0.05 ß x @ß 0.1� A 1

Results of the reweighting processnCTEQ15

Q2=4 GeV2

D0

Original µF=µ0 µF=2.0µ0 µF=0.5µ0

0.4

0.6

0.8

1

1.2

1.4

1.6

x

J/ψ

RPbg

0.4

0.6

0.8

1

1.2

1.4

1.6

10-510-410-310-210-1

B→J/ψ

x

ϒ(1S)

10-510-410-310-210-1

HELAC-Onia 2.0

EPPS16

Q2=4 GeV2

D0

Original µF=µ0 µF=2.0µ0 µF=0.5µ0

0.4

0.6

0.8

1

1.2

1.4

1.6

x

J/ψ

RPbg

0.4

0.6

0.8

1

1.2

1.4

1.6

10-510-410-310-210-1

B→J/ψ

x

ϒ(1S)

10-510-410-310-210-1

HELAC-Onia 2.0

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 21 / 27

X Global coherence of the data constraints: necessary condition to assume a shadowing-only approachX First clear experimental observation on gluon SHADOWING at low x; Visible reduction of theEPPS16 uncertainties; con�rmation of the extrapolation done in nCTEQ15

X �e scale ambiguity for D and J~ψ production is now the dominant uncertaintyX Non-prompt J~ψ are really promising if improved data can be obtained

X Con�rmation of the existence of a gluon anti-shadowing : Rg�0.05 ß x @ß 0.1� A 1

Results of the reweighting processnCTEQ15

Q2=4 GeV2

D0

Original µF=µ0 µF=2.0µ0 µF=0.5µ0

0.4

0.6

0.8

1

1.2

1.4

1.6

x

J/ψ

RPbg

0.4

0.6

0.8

1

1.2

1.4

1.6

10-510-410-310-210-1

B→J/ψ

x

ϒ(1S)

10-510-410-310-210-1

HELAC-Onia 2.0

EPPS16

Q2=4 GeV2

D0

Original µF=µ0 µF=2.0µ0 µF=0.5µ0

0.4

0.6

0.8

1

1.2

1.4

1.6

x

J/ψ

RPbg

0.4

0.6

0.8

1

1.2

1.4

1.6

10-510-410-310-210-1

B→J/ψ

x

ϒ(1S)

10-510-410-310-210-1

HELAC-Onia 2.0

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 21 / 27

X Global coherence of the data constraints: necessary condition to assume a shadowing-only approachX First clear experimental observation on gluon SHADOWING at low x; Visible reduction of theEPPS16 uncertainties; con�rmation of the extrapolation done in nCTEQ15

X �e scale ambiguity for D and J~ψ production is now the dominant uncertaintyX Non-prompt J~ψ are really promising if improved data can be obtained

X Con�rmation of the existence of a gluon anti-shadowing : Rg�0.05 ß x @ß 0.1� A 1

X X X X

X X X X

Consistency with other data

We checked the consistency of the reweighted (nCTEQ15) nPDFs with other data setsentering global analyses:

DIS data (the most precise set NMC Sn/C [NPB 481 (1996) 23]).

LHCW~Z boson production data [EPJC 77, (2017) 488].

PHENIX forward J~ψ RdAu data [PRL 107 (2011) 142301; PRC 87, (2013) 034904].

�is is very non-trivial and further con�rms the “shadowing-only” hypothesis of theleading-twist factorisation is valid within the current data precision!Table: χ2~Ndata values for nCTEQ15 nPDFs before and a�er reweighting using B, D, J~ψ datafor di�erent scale choices.

after reweightingnCTEQ15 B µ0 B up B down D µ0 D up D down J~ψ µ0 J~ψ up J~ψ down

W~Z LHC (102) 2.43 2.10 2.15 2.08 2.49 3.11 2.14 2.66 3.25 2.25NMC FSn2 ~FC2 (111) 0.58 0.71 0.64 0.77 0.59 0.56 0.84 0.60 0.56 0.88

NMC FPb2 ~FC2 (14) 0.55 0.51 0.53 0.56 0.50 0.63 0.59 0.44 0.56 0.59PHENIX J~ψ RdAu 1.99 1.93 0.43 3.35

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 22 / 27

Consistency with other data

We checked the consistency of the reweighted (nCTEQ15) nPDFs with other data setsentering global analyses:

DIS data (the most precise set NMC Sn/C [NPB 481 (1996) 23]).

LHCW~Z boson production data [EPJC 77, (2017) 488].

PHENIX forward J~ψ RdAu data [PRL 107 (2011) 142301; PRC 87, (2013) 034904].

�is is very non-trivial and further con�rms the “shadowing-only” hypothesis of theleading-twist factorisation is valid within the current data precision!Table: χ2~Ndata values for nCTEQ15 nPDFs before and a�er reweighting using B, D, J~ψ datafor di�erent scale choices.

after reweightingnCTEQ15 B µ0 B up B down D µ0 D up D down J~ψ µ0 J~ψ up J~ψ down

W~Z LHC (102) 2.43 2.10 2.15 2.08 2.49 3.11 2.14 2.66 3.25 2.25NMC FSn2 ~FC2 (111) 0.58 0.71 0.64 0.77 0.59 0.56 0.84 0.60 0.56 0.88

NMC FPb2 ~FC2 (14) 0.55 0.51 0.53 0.56 0.50 0.63 0.59 0.44 0.56 0.59PHENIX J~ψ RdAu 1.99 1.93 0.43 3.35

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 22 / 27

Consistency with other data

We checked the consistency of the reweighted (nCTEQ15) nPDFs with other data setsentering global analyses:

DIS data (the most precise set NMC Sn/C [NPB 481 (1996) 23]).

LHCW~Z boson production data [EPJC 77, (2017) 488].

PHENIX forward J~ψ RdAu data [PRL 107 (2011) 142301; PRC 87, (2013) 034904].

�is is very non-trivial and further con�rms the “shadowing-only” hypothesis of theleading-twist factorisation is valid within the current data precision!Table: χ2~Ndata values for nCTEQ15 nPDFs before and a�er reweighting using B, D, J~ψ datafor di�erent scale choices.

after reweightingnCTEQ15 B µ0 B up B down D µ0 D up D down J~ψ µ0 J~ψ up J~ψ down

W~Z LHC (102) 2.43 2.10 2.15 2.08 2.49 3.11 2.14 2.66 3.25 2.25NMC FSn2 ~FC2 (111) 0.58 0.71 0.64 0.77 0.59 0.56 0.84 0.60 0.56 0.88

NMC FPb2 ~FC2 (14) 0.55 0.51 0.53 0.56 0.50 0.63 0.59 0.44 0.56 0.59PHENIX J~ψ RdAu 1.99 1.93 0.43 3.35

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 22 / 27

Consistency with other data

We checked the consistency of the reweighted (nCTEQ15) nPDFs with other data setsentering global analyses:

DIS data (the most precise set NMC Sn/C [NPB 481 (1996) 23]).

LHCW~Z boson production data [EPJC 77, (2017) 488].

PHENIX forward J~ψ RdAu data [PRL 107 (2011) 142301; PRC 87, (2013) 034904].

�is is very non-trivial and further con�rms the “shadowing-only” hypothesis of theleading-twist factorisation is valid within the current data precision!Table: χ2~Ndata values for nCTEQ15 nPDFs before and a�er reweighting using B, D, J~ψ datafor di�erent scale choices.

after reweightingnCTEQ15 B µ0 B up B down D µ0 D up D down J~ψ µ0 J~ψ up J~ψ down

W~Z LHC (102) 2.43 2.10 2.15 2.08 2.49 3.11 2.14 2.66 3.25 2.25NMC FSn2 ~FC2 (111) 0.58 0.71 0.64 0.77 0.59 0.56 0.84 0.60 0.56 0.88

NMC FPb2 ~FC2 (14) 0.55 0.51 0.53 0.56 0.50 0.63 0.59 0.44 0.56 0.59PHENIX J~ψ RdAu 1.99 1.93 0.43 3.35

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 22 / 27

Consistency with other data

We checked the consistency of the reweighted (nCTEQ15) nPDFs with other data setsentering global analyses:

DIS data (the most precise set NMC Sn/C [NPB 481 (1996) 23]).

LHCW~Z boson production data [EPJC 77, (2017) 488].

PHENIX forward J~ψ RdAu data [PRL 107 (2011) 142301; PRC 87, (2013) 034904].

�is is very non-trivial and further con�rms the “shadowing-only” hypothesis of theleading-twist factorisation is valid within the current data precision!Table: χ2~Ndata values for nCTEQ15 nPDFs before and a�er reweighting using B, D, J~ψ datafor di�erent scale choices.

after reweightingnCTEQ15 B µ0 B up B down D µ0 D up D down J~ψ µ0 J~ψ up J~ψ down

W~Z LHC (102) 2.43 2.10 2.15 2.08 2.49 3.11 2.14 2.66 3.25 2.25NMC FSn2 ~FC2 (111) 0.58 0.71 0.64 0.77 0.59 0.56 0.84 0.60 0.56 0.88

NMC FPb2 ~FC2 (14) 0.55 0.51 0.53 0.56 0.50 0.63 0.59 0.44 0.56 0.59PHENIX J~ψ RdAu 1.99 1.93 0.43 3.35

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 22 / 27

Consistency with other data

We checked the consistency of the reweighted (nCTEQ15) nPDFs with other data setsentering global analyses:

DIS data (the most precise set NMC Sn/C [NPB 481 (1996) 23]).

LHCW~Z boson production data [EPJC 77, (2017) 488].

PHENIX forward J~ψ RdAu data [PRL 107 (2011) 142301; PRC 87, (2013) 034904].

�is is very non-trivial and further con�rms the “shadowing-only” hypothesis of theleading-twist factorisation is valid within the current data precision!Table: χ2~Ndata values for nCTEQ15 nPDFs before and a�er reweighting using B, D, J~ψ datafor di�erent scale choices.

after reweightingnCTEQ15 B µ0 B up B down D µ0 D up D down J~ψ µ0 J~ψ up J~ψ down

W~Z LHC (102) 2.43 2.10 2.15 2.08 2.49 3.11 2.14 2.66 3.25 2.25NMC FSn2 ~FC2 (111) 0.58 0.71 0.64 0.77 0.59 0.56 0.84 0.60 0.56 0.88

NMC FPb2 ~FC2 (14) 0.55 0.51 0.53 0.56 0.50 0.63 0.59 0.44 0.56 0.59PHENIX J~ψ RdAu 1.99 1.93 0.43 3.35

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 22 / 27

Application to the LHC in the �xed-targetmode (AFTER@LHC)

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 23 / 27

HF kinematical coverage of LHCb and ALICE in the�xed-target mode [arXiv:1807.00603]

2x2−10 1−10 1

)2 c (

GeV

/T

m

0

5

10

15

20

25

= 115 GeV, AFTER@LHCbNNs pp

c < 15 GeV/T

bottomonium, p

c < 18 GeV/T

charmonium, p

c < 20 GeV/T

D meson, p

c < 16 GeV/T

B meson, p

<5Lab2<y

2x2−10 1−10 1

)2 c (

GeV

/T

m0

5

10

15

20

25

(z=0)µ = 115 GeV, AFTER@ALICENNs pp

c < 8 GeV/T

bottomonium, p

c < 12 GeV/T

charmonium, p

c < 12 GeV/µ

T from charm, 4 < pµ

<4Lab2.5<y

ALICE could cover η � 1 � 2 for quarkonium into dileptons with one muon in themuon arm and another in the central barrel

[done for UPCs in the collider mode]

NB:�e coverage depends on the target position

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 24 / 27

HF kinematical coverage of LHCb and ALICE in the�xed-target mode [arXiv:1807.00603]

2x2−10 1−10 1

)2 c (

GeV

/T

m

0

5

10

15

20

25

= 115 GeV, AFTER@LHCbNNs pp

c < 15 GeV/T

bottomonium, p

c < 18 GeV/T

charmonium, p

c < 20 GeV/T

D meson, p

c < 16 GeV/T

B meson, p

<5Lab2<y

2x2−10 1−10 1

)2 c (

GeV

/T

m0

5

10

15

20

25

(z=0)µ = 115 GeV, AFTER@ALICENNs pp

c < 8 GeV/T

bottomonium, p

c < 12 GeV/T

charmonium, p

c < 12 GeV/µ

T from charm, 4 < pµ

<4Lab2.5<y

OXD meson with the Central Barrel at z=0

ALICE could cover η � 1 � 2 for quarkonium into dileptons with one muon in themuon arm and another in the central barrel

[done for UPCs in the collider mode]

NB:�e coverage depends on the target position

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 24 / 27

Potential constraints on the gluon nPDFs

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 25 / 27

EMC gluon e�ect totally unknown�is is the realm of the FT LHCexperimentsFirst projections are extremely promising[NB: initial nPDF uncertainties for x A 0.1 are underestimated;

� simply no data exist there]

�is assumes that we control otherpotential CNM e�ects: global studyrequired !Similar studies for the proton PDFs are yetto be done along the lines of the studiescarried out for low-x gluon at the LHCPROSA Coll. Eur.Phys.J. C75 (2015) 396; R. Gauld, J. Rojo

PRL 118 (2017) 072001

In the pp case, contrary to nPDF studiesbearing on nuclear modi�cation factors,one needs ways to reduce the systematicaluncertainties

Reward: unique constraints on gluon PDFs at high x and low scales

Potential constraints on the gluon nPDFs

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 25 / 27

EMC gluon e�ect totally unknown�is is the realm of the FT LHCexperimentsFirst projections are extremely promising[NB: initial nPDF uncertainties for x A 0.1 are underestimated;

� simply no data exist there]

�is assumes that we control otherpotential CNM e�ects: global studyrequired !Similar studies for the proton PDFs are yetto be done along the lines of the studiescarried out for low-x gluon at the LHCPROSA Coll. Eur.Phys.J. C75 (2015) 396; R. Gauld, J. Rojo

PRL 118 (2017) 072001

In the pp case, contrary to nPDF studiesbearing on nuclear modi�cation factors,one needs ways to reduce the systematicaluncertainties

Reward: unique constraints on gluon PDFs at high x and low scales

Potential constraints on the gluon nPDFs

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 25 / 27

EMC gluon e�ect totally unknown�is is the realm of the FT LHCexperimentsFirst projections are extremely promising[NB: initial nPDF uncertainties for x A 0.1 are underestimated;

� simply no data exist there]

�is assumes that we control otherpotential CNM e�ects: global studyrequired !Similar studies for the proton PDFs are yetto be done along the lines of the studiescarried out for low-x gluon at the LHCPROSA Coll. Eur.Phys.J. C75 (2015) 396; R. Gauld, J. Rojo

PRL 118 (2017) 072001

In the pp case, contrary to nPDF studiesbearing on nuclear modi�cation factors,one needs ways to reduce the systematicaluncertainties

Reward: unique constraints on gluon PDFs at high x and low scales

Potential constraints on the gluon nPDFs

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 25 / 27

EMC gluon e�ect totally unknown�is is the realm of the FT LHCexperimentsFirst projections are extremely promising[NB: initial nPDF uncertainties for x A 0.1 are underestimated;

� simply no data exist there]

�is assumes that we control otherpotential CNM e�ects: global studyrequired !Similar studies for the proton PDFs are yetto be done along the lines of the studiescarried out for low-x gluon at the LHCPROSA Coll. Eur.Phys.J. C75 (2015) 396; R. Gauld, J. Rojo

PRL 118 (2017) 072001

In the pp case, contrary to nPDF studiesbearing on nuclear modi�cation factors,one needs ways to reduce the systematicaluncertainties

Reward: unique constraints on gluon PDFs at high x and low scales

RX

eg

x

D0

Lpp=10 fb-1

LpXe=100 pb-1

p+Xe √ sNN=115 GeV

OriginalµF=µ0

µF=2.0µ0µF=0.5µ0

0.4

0.6

0.8

1

1.2

1.4

1.6

0 0.2 0.4 0.6 0.8

HELAC-Onia 2.0

RX

eg

x

B±

Lpp=10 fb-1

LpXe=100 pb-1

p+Xe √ sNN=115 GeV

OriginalµF=µ0

µF=2.0µ0µF=0.5µ0

0.4

0.6

0.8

1

1.2

1.4

1.6

0 0.2 0.4 0.6 0.8

HELAC-Onia 2.0

RX

eg

x

J/ψ

Lpp=10 fb-1

LpXe=100 pb-1

p+Xe √ sNN=115 GeV

OriginalµF=µ0

µF=2.0µ0µF=0.5µ0

0.4

0.6

0.8

1

1.2

1.4

1.6

0 0.2 0.4 0.6 0.8

HELAC-Onia 2.0

RX

eg

x

Υ(1S)

Lpp=10 fb-1

LpXe=100 pb-1

p+Xe √ sNN=115 GeV

OriginalµF=µ0

µF=2.0µ0µF=0.5µ0

0.4

0.6

0.8

1

1.2

1.4

1.6

0 0.2 0.4 0.6 0.8

HELAC-Onia 2.0

Potential constraints on the gluon nPDFs

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 25 / 27

EMC gluon e�ect totally unknown�is is the realm of the FT LHCexperimentsFirst projections are extremely promising[NB: initial nPDF uncertainties for x A 0.1 are underestimated;

� simply no data exist there]

�is assumes that we control otherpotential CNM e�ects: global studyrequired !Similar studies for the proton PDFs are yetto be done along the lines of the studiescarried out for low-x gluon at the LHCPROSA Coll. Eur.Phys.J. C75 (2015) 396; R. Gauld, J. Rojo

PRL 118 (2017) 072001

In the pp case, contrary to nPDF studiesbearing on nuclear modi�cation factors,one needs ways to reduce the systematicaluncertainties

Reward: unique constraints on gluon PDFs at high x and low scales

RX

eg

x

D0

Lpp=10 fb-1

LpXe=100 pb-1

p+Xe √ sNN=115 GeV

OriginalµF=µ0

µF=2.0µ0µF=0.5µ0

0.4

0.6

0.8

1

1.2

1.4

1.6

0 0.2 0.4 0.6 0.8

HELAC-Onia 2.0

RX

eg

x

B±

Lpp=10 fb-1

LpXe=100 pb-1

p+Xe √ sNN=115 GeV

OriginalµF=µ0

µF=2.0µ0µF=0.5µ0

0.4

0.6

0.8

1

1.2

1.4

1.6

0 0.2 0.4 0.6 0.8

HELAC-Onia 2.0

RX

eg

x

J/ψ

Lpp=10 fb-1

LpXe=100 pb-1

p+Xe √ sNN=115 GeV

OriginalµF=µ0

µF=2.0µ0µF=0.5µ0

0.4

0.6

0.8

1

1.2

1.4

1.6

0 0.2 0.4 0.6 0.8

HELAC-Onia 2.0

RX

eg

x

Υ(1S)

Lpp=10 fb-1

LpXe=100 pb-1

p+Xe √ sNN=115 GeV

OriginalµF=µ0

µF=2.0µ0µF=0.5µ0

0.4

0.6

0.8

1

1.2

1.4

1.6

0 0.2 0.4 0.6 0.8

HELAC-Onia 2.0

Potential constraints on the gluon nPDFs

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 25 / 27

EMC gluon e�ect totally unknown�is is the realm of the FT LHCexperimentsFirst projections are extremely promising[NB: initial nPDF uncertainties for x A 0.1 are underestimated;

� simply no data exist there]

�is assumes that we control otherpotential CNM e�ects: global studyrequired !Similar studies for the proton PDFs are yetto be done along the lines of the studiescarried out for low-x gluon at the LHCPROSA Coll. Eur.Phys.J. C75 (2015) 396; R. Gauld, J. Rojo

PRL 118 (2017) 072001

In the pp case, contrary to nPDF studiesbearing on nuclear modi�cation factors,one needs ways to reduce the systematicaluncertainties

Reward: unique constraints on gluon PDFs at high x and low scales

RX

eg

x

D0

Lpp=10 fb-1

LpXe=100 pb-1

p+Xe √ sNN=115 GeV

OriginalµF=µ0

µF=2.0µ0µF=0.5µ0

0.4

0.6

0.8

1

1.2

1.4

1.6

0 0.2 0.4 0.6 0.8

HELAC-Onia 2.0

RX

eg

x

B±

Lpp=10 fb-1

LpXe=100 pb-1

p+Xe √ sNN=115 GeV

OriginalµF=µ0

µF=2.0µ0µF=0.5µ0

0.4

0.6

0.8

1

1.2

1.4

1.6

0 0.2 0.4 0.6 0.8

HELAC-Onia 2.0

RX

eg

x

J/ψ

Lpp=10 fb-1

LpXe=100 pb-1

p+Xe √ sNN=115 GeV

OriginalµF=µ0

µF=2.0µ0µF=0.5µ0

0.4

0.6

0.8

1

1.2

1.4

1.6

0 0.2 0.4 0.6 0.8

HELAC-Onia 2.0

RX

eg

x

Υ(1S)

Lpp=10 fb-1

LpXe=100 pb-1

p+Xe √ sNN=115 GeV

OriginalµF=µ0

µF=2.0µ0µF=0.5µ0

0.4

0.6

0.8

1

1.2

1.4

1.6

0 0.2 0.4 0.6 0.8

HELAC-Onia 2.0

Conclusions

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 26 / 27

Conclusions

Gluon nPDFs at low x are extrapolated (no low x data used in �ts)� need for new constraints at x ß 10�3

We have proposed a quick and robust method to evaluate e�ects of heavy-quark data onnPDFs – complementary to full (but time consuming) pQCD computations

� we used it to analyse D0 , J~ψ, B� J~ψ and Υ�1S� data.

We proposed a “shadowing-only” hypothesis based on the leading twist factorisation

� suggested by the good agreement with the LHC data

Under this hypothesis, we made, using a Bayesian reweighting, the �rst direct observationof gluon shadowing. Additionally, our analysis corroborates the existance of gluonantishadowing.

Overall, the heavy-quark data can improve our knowledge of the nuclear gluondistribution and it should be included in future nPDF �ts.

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 27 / 27

Conclusions

Gluon nPDFs at low x are extrapolated (no low x data used in �ts)� need for new constraints at x ß 10�3

We have proposed a quick and robust method to evaluate e�ects of heavy-quark data onnPDFs – complementary to full (but time consuming) pQCD computations

� we used it to analyse D0 , J~ψ, B� J~ψ and Υ�1S� data.

We proposed a “shadowing-only” hypothesis based on the leading twist factorisation

� suggested by the good agreement with the LHC data

Under this hypothesis, we made, using a Bayesian reweighting, the �rst direct observationof gluon shadowing. Additionally, our analysis corroborates the existance of gluonantishadowing.

Overall, the heavy-quark data can improve our knowledge of the nuclear gluondistribution and it should be included in future nPDF �ts.

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 27 / 27

Conclusions

Gluon nPDFs at low x are extrapolated (no low x data used in �ts)� need for new constraints at x ß 10�3

We have proposed a quick and robust method to evaluate e�ects of heavy-quark data onnPDFs – complementary to full (but time consuming) pQCD computations

� we used it to analyse D0 , J~ψ, B� J~ψ and Υ�1S� data.

We proposed a “shadowing-only” hypothesis based on the leading twist factorisation

� suggested by the good agreement with the LHC data

Under this hypothesis, we made, using a Bayesian reweighting, the �rst direct observationof gluon shadowing. Additionally, our analysis corroborates the existance of gluonantishadowing.

Overall, the heavy-quark data can improve our knowledge of the nuclear gluondistribution and it should be included in future nPDF �ts.

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 27 / 27

Conclusions

Gluon nPDFs at low x are extrapolated (no low x data used in �ts)� need for new constraints at x ß 10�3

We have proposed a quick and robust method to evaluate e�ects of heavy-quark data onnPDFs – complementary to full (but time consuming) pQCD computations

� we used it to analyse D0 , J~ψ, B� J~ψ and Υ�1S� data.

We proposed a “shadowing-only” hypothesis based on the leading twist factorisation

� suggested by the good agreement with the LHC data

Under this hypothesis, we made, using a Bayesian reweighting, the �rst direct observationof gluon shadowing. Additionally, our analysis corroborates the existance of gluonantishadowing.

Overall, the heavy-quark data can improve our knowledge of the nuclear gluondistribution and it should be included in future nPDF �ts.

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 27 / 27

Conclusions

Gluon nPDFs at low x are extrapolated (no low x data used in �ts)� need for new constraints at x ß 10�3

We have proposed a quick and robust method to evaluate e�ects of heavy-quark data onnPDFs – complementary to full (but time consuming) pQCD computations

� we used it to analyse D0 , J~ψ, B� J~ψ and Υ�1S� data.

We proposed a “shadowing-only” hypothesis based on the leading twist factorisation

� suggested by the good agreement with the LHC data

Under this hypothesis, we made, using a Bayesian reweighting, the �rst direct observationof gluon shadowing. Additionally, our analysis corroborates the existance of gluonantishadowing.

Overall, the heavy-quark data can improve our knowledge of the nuclear gluondistribution and it should be included in future nPDF �ts.

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 27 / 27

Conclusions

Gluon nPDFs at low x are extrapolated (no low x data used in �ts)� need for new constraints at x ß 10�3

We have proposed a quick and robust method to evaluate e�ects of heavy-quark data onnPDFs – complementary to full (but time consuming) pQCD computations

� we used it to analyse D0 , J~ψ, B� J~ψ and Υ�1S� data.

We proposed a “shadowing-only” hypothesis based on the leading twist factorisation

� suggested by the good agreement with the LHC data

Under this hypothesis, we made, using a Bayesian reweighting, the �rst direct observationof gluon shadowing. Additionally, our analysis corroborates the existance of gluonantishadowing.

Overall, the heavy-quark data can improve our knowledge of the nuclear gluondistribution and it should be included in future nPDF �ts.

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 27 / 27

Conclusions

Gluon nPDFs at low x are extrapolated (no low x data used in �ts)� need for new constraints at x ß 10�3

We have proposed a quick and robust method to evaluate e�ects of heavy-quark data onnPDFs – complementary to full (but time consuming) pQCD computations

� we used it to analyse D0 , J~ψ, B� J~ψ and Υ�1S� data.

We proposed a “shadowing-only” hypothesis based on the leading twist factorisation

� suggested by the good agreement with the LHC data

Under this hypothesis, we made, using a Bayesian reweighting, the �rst direct observationof gluon shadowing. Additionally, our analysis corroborates the existance of gluonantishadowing.

Overall, the heavy-quark data can improve our knowledge of the nuclear gluondistribution and it should be included in future nPDF �ts.

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 27 / 27

Conclusions

Gluon nPDFs at low x are extrapolated (no low x data used in �ts)� need for new constraints at x ß 10�3

We have proposed a quick and robust method to evaluate e�ects of heavy-quark data onnPDFs – complementary to full (but time consuming) pQCD computations

� we used it to analyse D0 , J~ψ, B� J~ψ and Υ�1S� data.

We proposed a “shadowing-only” hypothesis based on the leading twist factorisation

� suggested by the good agreement with the LHC data

Under this hypothesis, we made, using a Bayesian reweighting, the �rst direct observationof gluon shadowing. Additionally, our analysis corroborates the existance of gluonantishadowing.

Overall, the heavy-quark data can improve our knowledge of the nuclear gluondistribution and it should be included in future nPDF �ts.

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 27 / 27

BACKUP SLIDES

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 28 / 27

Data-driven vs. FONLLnCTEQ15

Q2=4 GeV2

Data-driven D0, µF

0.4

0.6

0.8

1

1.2

1.4

1.6

x

FONLL D0, µR

RPbg

0.4

0.6

0.8

1

1.2

1.4

1.6

10-510-410-310-210-1

FONLL D0, µF

x

FONLL D0, mc

10-510-410-310-210-1

HELAC-Onia 2.0

nCTEQ15

Q2=4 GeV2

Data-driven B→J/ψ, µF

0.4

0.6

0.8

1

1.2

1.4

1.6

x

FONLL B→J/ψ, µR

RPbg

0.4

0.6

0.8

1

1.2

1.4

1.6

10-510-410-310-210-1

FONLL B→J/ψ, µF

x

FONLL B→J/ψ, mc

10-510-410-310-210-1

HELAC-Onia 2.0

J.P. Lansberg (IPNO) Shadowing and LHC heavy-flavour data September 7, 2018 29 / 27