systematics in nucleon matrix element calculations€¦ · motivation three reasons for studying...
TRANSCRIPT
Systematics in nucleon matrix element calculations
Jeremy Green
NIC, DESY, Zeuthen
The 36th International Symposium on Laice Field TheoryEast Lansing, MI, USA
July 22–28, 2018
Motivation
Three reasons for studying structure of protons and neutrons:I Understanding the quark and gluon substructure of a hadron.I Understanding nucleons as tools in experiments.I Validation of laice QCD using “benchmark” observables.
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 2
Neutron beta decay
W−
n
pνee Coupling toW boson via axial current depends on nucleonaxial charge,
〈p (P , s ′) |uγ µγ5d |n(P , s )〉 = дAup (P , s ′)γ µγ5un (P , s ).
Interpreted as isovector quark spin contribution.→ Yi-Bo Yang plenary
PDG 2018: дA = 1.2724(23).
History plots 1
1960 1970 1980 1990 2000 2010850
900
950
1000
1050
1100
1960 1970 1980 1990 2000 201080
85
90
95
100
105
1990 2000 20100.8
1.0
1.2
1.4
1.6
1.8
2.0
1970 1980 1990 2000 2010-1.28
-1.26
-1.24
-1.22
-1.20
-1.18
-1.16
1970 1980 1990 2000 20106
8
10
12
14
16
1990 2000 20100.30
0.35
0.40
0.45
0.50
0.55
0.60
0.65
1970 1980 1990 2000 2010547.0
547.5
548.0
548.5
549.0
549.5
550.0
1970 1980 1990 2000 20100.52
0.53
0.54
0.55
0.56
0.57
0.58
0.59
1990 2000 201077
78
79
80
81
82
83
84
1960 1970 1980 1990 2000 20101115.0
1115.2
1115.4
1115.6
1115.8
1980 1990 2000 20101770
1780
1790
1800
1810
1820
1830
1970 1980 1990 2000 2010-20
0
20
40
60
80
100
Figure 1: A historical perspective of values of a few particle properties tabulated in this Review as a function of date of publication of the
Review. A full error bar indicates the quoted error; a thick-lined portion indicates the same but without the “scale factor.”
Citation: M. Tanabashi et al. (Particle Data Group), Phys. Rev. D 98, 030001 (2018)
1DARIUS 17 calculates this value from the measurement of the a parameter (see below).2MENDENHALL 13 gets A = −0.11954 ± 0.00055 ± 0.00098 and λ = −1.2756 ±0.0030. We quote the nearly identical values that include the earlier UCNA measurement(PLASTER 12), with a correction to that result.
3This MUND 13 value includes earlier PERKEO II measurements (ABELE 02 andABELE 97D).
4MOSTOVOI 01 measures the two P-odd correlations A and B, or rather SA and SB,where S is the n polarization, in free neutron decay.
5YEROZOLIMSKY 97 makes a correction to the EROZOLIMSKII 91 value.6This PLASTER 12 value is identical with that given in LIU 10, but the experiment isnow described in detail.
7 This is the combined result of ABELE 02 and ABELE 97D.8 These experiments measure the absolute value of gA/gV only.9KROHN 75 includes events of CHRISTENSEN 70.
10KROPF 74 reviews all data through 1972.
WEIGHTED AVERAGE-1.2724±0.0023 (Error scaled by 2.2)
BOPP 86 SPEC 4.3YEROZLIM... 97 CNTR 11.7LIAUD 97 TPC 2.5MOSTOVOI 01 CNTR 0.7SCHUMANN 08 CNTRMUND 13 SPEC 3.6MENDENHALL 13 UCNA 1.1DARIUS 17 SPEC
χ2
23.8(Confidence Level = 0.0002)
-1.29 -1.28 -1.27 -1.26 -1.25 -1.24
λ ≡ gA / gV
e− ASYMMETRY PARAMETER Ae− ASYMMETRY PARAMETER Ae− ASYMMETRY PARAMETER Ae− ASYMMETRY PARAMETER AThis is the neutron-spin electron-momentum correlation coefficient. Unless otherwisenoted, the values are corrected for radiative effects and weak magnetism. In theStandard Model, A is related to λ ≡ gA/gV by A = −2 λ (λ + 1) / (1 + 3λ2); thisassumes that gA and gV are real.
VALUE DOCUMENT ID TECN COMMENT
−0.1184 ±0.0010 OUR AVERAGE−0.1184 ±0.0010 OUR AVERAGE−0.1184 ±0.0010 OUR AVERAGE−0.1184 ±0.0010 OUR AVERAGE Error includes scale factor of 2.4. See the ideogrambelow.−0.11952±0.00110 1 MENDENHALL13 UCNA Ultracold n, polarized
−0.11926±0.00031+0.00036−0.00042
2 MUND 13 SPEC Cold n, polarized
−0.1160 ±0.0009 ±0.0012 LIAUD 97 TPC Cold n, polarized
−0.1135 ±0.0014 3 YEROZLIM... 97 CNTR Cold n, polarized
−0.1146 ±0.0019 BOPP 86 SPEC Cold n, polarized
HTTP://PDG.LBL.GOV Page 11 Created: 6/5/2018 19:00
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 3
Other observables
Precision β-decay experiments may be sensitive to BSM physics; leadingcontributions are controlled by the scalar and tensor charges:T. Bhaacharya et al., Phys. Rev. D 85, 054512 (2012) [1110.6448]
〈p (P , s ′) |ud |n(P , s )〉 = дS up (P , s ′)un (P , s ),〈p (P , s ′) |uσ µνd |n(P , s )〉 = дT up (P , s ′)σ µνun (P , s ).
Sigma terms control the sensitivity of direct-detection dark maer searchesto WIMPs that interact via Higgs exchange.
σπN =mud 〈N |uu + dd |N 〉 σq =mq〈N |qq |N 〉
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 4
Toward precision nucleon structure
Trustworthy laice calculations need control over all systematics:I Physical quark massesI Isolation of ground stateI Infinite volumeI Continuum limit
Want to transition from this . . .
0.9
1.0
1.1
1.2
1.3
1.4
1.5
1.6
0.10 0.15 0.20 0.25 0.30
gA
mπ (GeV)
Nf=2+1+1 TM (ETMC)Nf=2 TM+clover (ETMC)Nf=2 TM (ETMC)Nf=2+1+1 mixed (PNDME)Nf=2+1+1 mixed (CalLat)Nf=2+1 DW (RBC/UKQCD)Nf=2+1 clover (CSSM)
Nf=2+1 clover (LHPC)Nf=2+1 clover (NME)Nf=2 clover (Mainz)Nf=2 clover (RQCD)Nf=2 clover (QCDSF)PDG
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 5
Toward precision nucleon structure
Trustworthy laice calculations need control over all systematics:I Physical quark massesI Isolation of ground stateI Infinite volumeI Continuum limit
Want to transition from this . . . to something like this.
22 24 26 28 30 32 34
=+
+=
+=
pheno.
Weinberg 77Leutwyler 96Kaiser 98Narison 06Oller 07PDG
QCDSF/UKQCD 06ETM 07RBC 07ETM 10BETM 14D
FLAG average for =
MILC 04, HPQCD/MILC/UKQCD 04RBC/UKQCD 08PACS-CS 08MILC 09MILC 09APACS-CS 09Blum 10RBC/UKQCD 10ABMW 10A, 10BLaiho 11PACS-CS 12RBC/UKQCD 12RBC/UKQCD 14B
FLAG average for = +
ETM 14FNAL/MILC 14A
FLAG average for = + +
/
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 5
Outline
1. Methodologies for nucleon matrix elements
2. Excited-state eects: theory and practice
3. Other systematics: L, (a),mq
4. Outlook
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 6
Hadron correlation functions
Compute two-point and three-point functions, using interpolator χ andoperator insertion O. In simplest case:
t0
τ0 T
C2pt (t ) ≡ 〈χ (t )χ† (0)〉=
∑n
|Zn |2e−Ent
→ |Z0 |2e−E0t(1 +O (e−∆Et )
),
where Zn = 〈Ω |χ |n〉,
C3pt (τ ,T ) ≡ 〈χ (T )O (τ )χ† (0)〉=
∑n,n′
Zn′Z∗n〈n′ |O|n〉e−Enτ e−En′ (T−τ )
→ |Z0 |2〈0|O|0〉e−E0T(1 +O (e−∆Eτ ) +O (e−∆E (T−τ ) )
)Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 7
Hadron matrix elements
Ratio method
R (τ ,T ) ≡ C3pt (τ ,T )
C2pt (T )= 〈0|O|0〉 +O (e−∆Eτ ) +O (e−∆E (T−τ ) )
Midpoint yields R (T2 ,T ) = 〈0|O|0〉 +O (e−∆ET /2).
Summation methodL. Maiani et al., Nucl. Phys. B 293, 420 (1987); дA in S. Güsken et al., Phys. Le. B 227, 266 (1989)
S (T ) ≡∑τ
R (τ ,T ),d
dTS (T ) = 〈0|O|0〉 +O (Te−∆ET )
Sum can be over all timeslices or from τ0 to T − τ0.Improved asymptotic behaviour noted in talks at Laice 2010.S. Capitani et al., PoS LATTICE2010 147 [1011.1358]; J. Bulava et al., ibid. 303 [1011.4393]
In practice noisier than ratio method at same T .
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 8
Ratio and summation method
1.15
1.20
1.25
1.30
2 4 6 8 10 12
g A(bare)
T/a
R(T/2, T )
S(T + a)− S(T )
physicalmπ , a = 0.116 fmN. Hasan, JG, et al., in preparation
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 9
Feynman-Hellmann approach
If we add a term to the Lagrangian L (λ) ≡ L + λO, then
∂
∂λEn (λ)
λ=0= 〈n |O|n〉.
Discrete derivatives sometimes used, particularly for sigma terms:
σq ≡mq〈N |qq |N 〉 =mq∂mN
∂mq.
Exact derivatives lead directly to summation method (neglecting 〈Ω |O|Ω〉):
− ∂∂λ
logC2pt (t )λ=0= S (t ).
L. Maiani et al., Nucl. Phys. B 293, 420 (1987)A. J. Chambers et al. (CSSM and QCDSF/UKQCD), Phys. Rev. D 90, 014510 (2014) [1405.3019]C. Bouchard et al., Phys. Rev. D 96, 014504 (2017) [1612.06963]
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 10
Controlling excited states
Three approaches:I Go to large T . Need ∆ET 1.
But signal-to-noise decays as e−(E0− 32mπ )T .
I Improve the interpolating operator.I Fit correlators to remove excited states.
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 11
Excited-state energies
First approximation: noninteracting stable hadrons in a box. Nπ , Nππ , . . .
3.0 3.5 4.0 4.5 5.0mπL
0.0
0.2
0.4
0.6
0.8
1.0∆E
(GeV
)
Nπ
Physicalmπ .
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 12
Excited-state energies
First approximation: noninteracting stable hadrons in a box. Nπ , Nππ , . . .
3.0 3.5 4.0 4.5 5.0mπL
0.0
0.2
0.4
0.6
0.8
1.0∆E
(GeV
)
Nπ Nππ
Physicalmπ .
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 12
Excited-state energies
First approximation: noninteracting stable hadrons in a box. Nπ , Nππ , . . .
3.0 3.5 4.0 4.5 5.0mπL
0.0
0.2
0.4
0.6
0.8
1.0∆E
(GeV
)
Nπ Nππ Nπππ
Physicalmπ .
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 12
Excited-state energies
First approximation: noninteracting stable hadrons in a box. Nπ , Nππ , . . .
3.0 3.5 4.0 4.5 5.0mπL
0.0
0.2
0.4
0.6
0.8
1.0∆E
(GeV
)
Nπ Nππ Nπππ Nππππ
Physicalmπ .
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 12
Excited-state energies
First approximation: noninteracting stable hadrons in a box. Nπ , Nππ , . . .
0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50mπ (GeV)
0.0
0.2
0.4
0.6
0.8
1.0∆E
(GeV
)
Nπ Nππ Nπππ Nππππ
mπL = 4.
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 12
Excited-state energies
M. T. Hansen and H. B. Meyer, Nucl. Phys. B 923, 558 (2017) [1610.03843]
Focus on Nπ sector and apply finite-volume quantization usingscaering phase shi from experiment.
3.0 3.5 4.0 4.5 5.0 5.5 6.0MπL
1000
1200
1400
1600
1800
E (
MeV
)I(JP ) =1/2(1/2+ )
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 13
Going to largeT
Statistical errors:
δstat ∼ N −1/2e (E0− 32mπ )T
Excited state systematics:
δexc ∼ e−∆ET /2 (ratio method)
Suppose we want these to be equal, i.e. δstat = δexc ≡ δ .
Required statistics are given by
N ∝ δ−(2+ 4E0−6mπ
∆E
).
At the physical point with ∆E = 2mπ , the exponent is ≈ −13.
Multi-level methods could potentially improve this.
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 14
Predicting excited-state contributions
Need to know:
1. energies En2. matrix elements 〈n′ |O|n〉3. overlaps Zn
Have been studied in ChPT. B. C. Tiburzi; O. Bär
Key insight: at leading order a single LEC controls coupling of localinterpolator to N and Nπ states.
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 15
ChPT prediction of excited-state eectsm
e(T)/m
0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.01.00
1.01
1.02
1.03
1.04
T (fm)
physical pointmπL = 4
truncated to fiveNπ states
need T & 2 fm forChPT to be reliable
O. Bär, Int. J. Mod. Phys. A 32 no. 15, 1730011 (2017) [1705.02806]
generalization to axial FFs→ Oliver Bär parallel, Wed. 14:00
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 16
ChPT prediction of excited-state eectsRX(T
2,T
)/X
Results: Midpoint estimate for all charges and moments
hxiud
hxiud
hxiud
gS
gA
gT
1.0 1.5 2.0 2.51.00
1.05
1.10
1.15
1.20
1.25
T (fm)
physical pointmπL = 4
truncated to fiveNπ states
need T & 2 fm forChPT to be reliable
O. Bär, Int. J. Mod. Phys. A 32 no. 15, 1730011 (2017) [1705.02806]
generalization to axial FFs→ Oliver Bär parallel, Wed. 14:00
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 16
Going beyond ChPT
M. T. Hansen and H. B. Meyer, Nucl. Phys. B 923, 558 (2017) [1610.03843]
Deviations of 〈Ω |χ |Nπ 〉 and 〈Nπ |Aµ |N 〉 from ChPT in the resonanceregime modeled with parameters α and γ .
↵ = 0, = 0↵ = 1, = 0↵ = 1, = 4
ML = 4ML = 6
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 17
Numerical studies
Available data depends onhow C3pt is computed.
τ0 T
For connected diagrams:I Fixed sink: most common. Get data at all τ and
any operator insertion. Cost increases with eachvalue of T .
I Fixed operator and τ . Get data at all T and anyset of interpolators. variational studies by CSSM
I Fixed operator and summed τ . Get summationdata at all T and any set of interpolators.used by CalLat, NPLQCD
−6 −4 −2 0 2 4 6(τ−T/2)/a
1.30
1.35
1.40
1.45
g A(b
are)
T/a =
14121086
ratiosummation
JG et al., PRD 95, 114502 [1703.06703]
21 22 23 24 25 26 27 28 29 30
1.2
1.4
1.6
tS
gA
B. J. Owen et al., PLB 723, 217 [1212.4668]
0 5 10 15t/a
1.15
1.20
1.25
1.30
1.35
1.40
geff
A
a09m220
C. C. Chang et al., Nature 558, 91[1805.12130]
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 18
Improving the nucleon interpolator
I Standard operator: χ = ϵabc (uTaCγ5db )uc , with smeared quark fields.I Smearing width typically tuned so thatme has early plateau.I Variational approach: χ =
∑i ci χi with optimized ci .
I Simple basis: use varying smearing widths for χi .I Can be extended with more local structures (derivatives and Gµν ).I More expensive: include nonlocal operators for beer isolation of Nπ
and Nππ states.
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 19
Variational approach: dierent smearings — eective mass
3 5 7 9 11 13 15 17 19 21 23
t
0.40
0.42
0.44
0.46
0.48
0.50
0.52
0.54
0.56
0.58
aM
N
Mass 2exp Variational Comparison
sm32
sm64
sm128
t0 =2 ∆t=2
mπ = 460 MeV, a = 0.074 fmJ. Dragos et al., PRD 94, 074505 (2016) [1606.03195]
0.42
0.44
0.46
0.48
0.50
0.52
0.54
0.56
4 6 8 10 12 14 16 18 20
MN
t
S3S3
S5S5
S7S7
V357
mπ = 312 MeV, a = 0.081 fmB. Yoon et al. (NME), PRD 93, 114506 (2016)
[1602.07737]
Small improvement over widest smearing.
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 20
Variational approach: dierent smearings — eective mass
3 5 7 9 11 13 15 17 19 21 23
t
0.40
0.42
0.44
0.46
0.48
0.50
0.52
0.54
0.56
0.58
aM
N
Mass 2exp Variational Comparison
sm32
sm64
sm128
t0 =2 ∆t=2
mπ = 460 MeV, a = 0.074 fmJ. Dragos et al., PRD 94, 074505 (2016) [1606.03195]
0.42
0.44
0.46
0.48
0.50
0.52
0.54
0.56
4 6 8 10 12 14 16 18 20
MN
t
S5S5
S7S7
S9S9
V579
mπ = 312 MeV, a = 0.081 fmB. Yoon et al. (NME), PRD 93, 114506 (2016)
[1602.07737]
Small improvement over widest smearing. . . so use even wider smearings.
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 20
Variational approach: dierent smearings — axial charge
8 6 4 2 0 2 4 6 8
τ−t/20.95
1.00
1.05
1.10
1.15
1.20
R( τ,t
)
gA Variational Comparison
sm32
sm64
sm128
t0 =2 ∆t=2
mπ = 460 MeV, a = 0.074 fmT = 0.96 fmJ. Dragos et al., PRD 94, 074505 (2016) [1606.03195]
1.20
1.25
1.30
1.35
1.40
1.45
1.50
-6 -4 -2 0 2 4 6
gA
τ - tsep/2
tsep=10tsep=12tsep=14
tsep=16
V357V579
mπ = 312 MeV, a = 0.081 fmvariational at T = 0.97 fm versus S5B. Yoon et al. (NME), PRD 93, 114506 (2016)
[1602.07737]
Improvement over best tuned single smearing seems small.
See also: including negative parity operators in basis when ~p , 0.F. Stokes et al., in preparation
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 21
Variational approach: dierent smearings — axial charge
8 6 4 2 0 2 4 6 8
τ−t/20.95
1.00
1.05
1.10
1.15
1.20
R( τ,t
)
gA Variational Comparison
sm32
sm64
sm128
t0 =2 ∆t=2
mπ = 460 MeV, a = 0.074 fmT = 0.96 fmJ. Dragos et al., PRD 94, 074505 (2016) [1606.03195]
1.20
1.25
1.30
1.35
1.40
1.45
1.50
-6 -4 -2 0 2 4 6
gA
τ - tsep/2
tsep=10tsep=12tsep=14
tsep=16
V357V579
mπ = 312 MeV, a = 0.081 fmvariational at T = 0.97 fm versus S9B. Yoon et al. (NME), PRD 93, 114506 (2016)
[1602.07737]
Improvement over best tuned single smearing seems small.
See also: including negative parity operators in basis when ~p , 0.F. Stokes et al., in preparation
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 21
Variational approach: dierent smearings — axial charge
8 6 4 2 0 2 4 6 8
τ−t/20.95
1.00
1.05
1.10
1.15
1.20
R( τ,t
)
gA Variational Comparison
sm32
sm64
sm128
t0 =2 ∆t=2
mπ = 460 MeV, a = 0.074 fmT = 0.96 fmJ. Dragos et al., PRD 94, 074505 (2016) [1606.03195]
1.20
1.25
1.30
1.35
1.40
1.45
1.50
-6 -4 -2 0 2 4 6
gA
τ - tsep/2
tsep=10tsep=12tsep=14
tsep=16
V357V579
mπ = 312 MeV, a = 0.081 fmvariational at T = 0.97 fm versus S9B. Yoon et al. (NME), PRD 93, 114506 (2016)
[1602.07737]
Improvement over best tuned single smearing seems small.
See also: including negative parity operators in basis when ~p , 0.F. Stokes et al., in preparation
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 21
Variational approach: larger basis — eective mass
Calculation performed using distillation.Operators with derivatives and hybrid operators with Gµν included in basis.→ C. Egerer parallel, Wed. 14:40
How does this compare with a standard calculation with tuned smearing?
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 22
Variational approach: larger basis — tensor charge
Standard operator VariationalCalculation performed using distillation.Operators with derivatives and hybrid operators with Gµν included in basis.→ C. Egerer parallel, Wed. 14:40
How does this compare with a standard calculation with tuned smearing?
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 22
Variational approach: larger basis — tensor charge
Standard operator VariationalCalculation performed using distillation.Operators with derivatives and hybrid operators with Gµν included in basis.→ C. Egerer parallel, Wed. 14:40
How does this compare with a standard calculation with tuned smearing?
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 22
Fiing approaches
Fit correlators using ansatz based on N -state model.I En and Zn generally determined from C2pt.I 〈n′ |O|n〉 determined from C3pt, either in combined or separate fit.
Typical energy gap:
0.5 GeV < E1 − E0 < 1 GeV,
usually greater than expected 2mπ .I Each model state approximates contribution from several states.I Fit ansatz should be considered a somewhat uncontrolled model.
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 23
Fits for axial charge
Fixed sink
1.1
1.2
1.3
1.4
−10 −5 0 5 10
τ :∞ 12 14 16
a09m220
R. Gupta et al. (PNDME), 1806.09006
constrained 3-state fit: τ ,T − τ ≥ 3a
Note: more than 10 states with∆E < 1 GeV =⇒ e−3a∆E > 0.25.
Summed current insertion
0 5 10 15t/a
1.15
1.20
1.25
1.30
1.35
1.40
geff
A
a09m220
C. C. Chang et al., Nature 558, 91 [1805.12130]
data with smeared and point sinksunconstrained 2-state fit: T ≥ 3a
1.22
1.26
g A/
g V
a09m220
meff tmin geffA tmin geff
V tmin
8 9 10 11 2 3 4 5 5 6 7 810−2
10−1
100
P
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 24
Fits for axial charge
Fixed sink
1.1
1.2
1.3
1.4
−10 −5 0 5 10
τ :∞ 12 14 16
a09m220
R. Gupta et al. (PNDME), 1806.09006
constrained 3-state fit: τ ,T − τ ≥ 3a
Note: more than 10 states with∆E < 1 GeV =⇒ e−3a∆E > 0.25.
Summed current insertion
0 5 10 15t/a
1.15
1.20
1.25
1.30
1.35
1.40
geff
A
a09m220
C. C. Chang et al., Nature 558, 91 [1805.12130]
data with smeared and point sinksunconstrained 2-state fit: T ≥ 3a
1.22
1.26
g A/
g V
a09m220
meff tmin geffA tmin geff
V tmin
8 9 10 11 2 3 4 5 5 6 7 810−2
10−1
100
P
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 24
Simultaneous fits
data at tsep = 1.0fmfit model at tsep = 1.0fm
fit result
tsep/fm
gu−d
A
10.50-0.5-1
1.3
1.25
1.2
1.15
1.1
1.05
1
0.95
0.9data at tsep = 1.0fm
fit model at tsep = 1.0fmfit result
tsep/fm
gu−d
S
10.50-0.5-1
2
1.5
1
0.5
0data at tsep = 1.0fm
fit model at tsep = 1.0fmfit result
tsep/fm
gu−d
T
10.50-0.5-1
1.3
1.2
1.1
1
0.9
0.8
data at tsep = 1.0fmfit model at tsep = 1.0fm
fit result
tsep/fm
〈x〉 u
−d
10.50-0.5-1
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0data at tsep = 1.0fm
fit model at tsep = 1.0fmfit result
tsep/fm
〈x〉 ∆
u−∆d
10.50-0.5-1
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0data at tsep = 1.0fm
fit model at tsep = 1.0fmfit result
tsep/fm
〈x〉 δ
u−δd
10.50-0.5-1
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0
mπ = 347 MeV, a = 0.064 fm, T = 16a(all T included in fit).
Two-state fit to ratios for six observables in range τ ,T − τ ≥ tstart withtstartMπ = 0.4 fixed globally for all ensembles.∆E determined only from ratios, not C2pt.→ K. Onad parallel, Thu 12:00
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 25
Simultaneous fits
data at tsep = 1.2fmfit model at tsep = 1.2fm
fit result
tsep/fm
gu−d
A
10.50-0.5-1
1.3
1.25
1.2
1.15
1.1
1.05
1
0.95
0.9data at tsep = 1.2fm
fit model at tsep = 1.2fmfit result
tsep/fm
gu−d
S
10.50-0.5-1
2
1.5
1
0.5
0data at tsep = 1.2fm
fit model at tsep = 1.2fmfit result
tsep/fm
gu−d
T
10.50-0.5-1
1.3
1.2
1.1
1
0.9
0.8
data at tsep = 1.2fmfit model at tsep = 1.2fm
fit result
tsep/fm
〈x〉 u
−d
10.50-0.5-1
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0data at tsep = 1.2fm
fit model at tsep = 1.2fmfit result
tsep/fm
〈x〉 ∆
u−∆d
10.50-0.5-1
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0data at tsep = 1.2fm
fit model at tsep = 1.2fmfit result
tsep/fm
〈x〉 δ
u−δd
10.50-0.5-1
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0
mπ = 347 MeV, a = 0.064 fm, T = 18a(all T included in fit).
Two-state fit to ratios for six observables in range τ ,T − τ ≥ tstart withtstartMπ = 0.4 fixed globally for all ensembles.∆E determined only from ratios, not C2pt.→ K. Onad parallel, Thu 12:00
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 25
Simultaneous fits
data at tsep = 1.3fmfit model at tsep = 1.3fm
fit result
tsep/fm
gu−d
A
10.50-0.5-1
1.3
1.25
1.2
1.15
1.1
1.05
1
0.95
0.9data at tsep = 1.3fm
fit model at tsep = 1.3fmfit result
tsep/fm
gu−d
S
10.50-0.5-1
2
1.5
1
0.5
0data at tsep = 1.3fm
fit model at tsep = 1.3fmfit result
tsep/fm
gu−d
T
10.50-0.5-1
1.3
1.2
1.1
1
0.9
0.8
data at tsep = 1.3fmfit model at tsep = 1.3fm
fit result
tsep/fm
〈x〉 u
−d
10.50-0.5-1
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0data at tsep = 1.3fm
fit model at tsep = 1.3fmfit result
tsep/fm
〈x〉 ∆
u−∆d
10.50-0.5-1
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0data at tsep = 1.3fm
fit model at tsep = 1.3fmfit result
tsep/fm
〈x〉 δ
u−δd
10.50-0.5-1
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0
mπ = 347 MeV, a = 0.064 fm, T = 20a(all T included in fit).
Two-state fit to ratios for six observables in range τ ,T − τ ≥ tstart withtstartMπ = 0.4 fixed globally for all ensembles.∆E determined only from ratios, not C2pt.→ K. Onad parallel, Thu 12:00
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 25
Simultaneous fits
data at tsep = 1.4fmfit model at tsep = 1.4fm
fit result
tsep/fm
gu−d
A
10.50-0.5-1
1.3
1.25
1.2
1.15
1.1
1.05
1
0.95
0.9data at tsep = 1.4fm
fit model at tsep = 1.4fmfit result
tsep/fm
gu−d
S
10.50-0.5-1
2
1.5
1
0.5
0data at tsep = 1.4fm
fit model at tsep = 1.4fmfit result
tsep/fm
gu−d
T
10.50-0.5-1
1.3
1.2
1.1
1
0.9
0.8
data at tsep = 1.4fmfit model at tsep = 1.4fm
fit result
tsep/fm
〈x〉 u
−d
10.50-0.5-1
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0data at tsep = 1.4fm
fit model at tsep = 1.4fmfit result
tsep/fm
〈x〉 ∆
u−∆d
10.50-0.5-1
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0data at tsep = 1.4fm
fit model at tsep = 1.4fmfit result
tsep/fm
〈x〉 δ
u−δd
10.50-0.5-1
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0
mπ = 347 MeV, a = 0.064 fm, T = 22a(all T included in fit).
Two-state fit to ratios for six observables in range τ ,T − τ ≥ tstart withtstartMπ = 0.4 fixed globally for all ensembles.∆E determined only from ratios, not C2pt.→ K. Onad parallel, Thu 12:00
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 25
Simultaneous fits
data at tsep = 1.5fmfit model at tsep = 1.5fm
fit result
tsep/fm
gu−d
A
10.50-0.5-1
1.3
1.25
1.2
1.15
1.1
1.05
1
0.95
0.9data at tsep = 1.5fm
fit model at tsep = 1.5fmfit result
tsep/fm
gu−d
S
10.50-0.5-1
2
1.5
1
0.5
0data at tsep = 1.5fm
fit model at tsep = 1.5fmfit result
tsep/fm
gu−d
T
10.50-0.5-1
1.3
1.2
1.1
1
0.9
0.8
data at tsep = 1.5fmfit model at tsep = 1.5fm
fit result
tsep/fm
〈x〉 u
−d
10.50-0.5-1
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0data at tsep = 1.5fm
fit model at tsep = 1.5fmfit result
tsep/fm
〈x〉 ∆
u−∆d
10.50-0.5-1
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0data at tsep = 1.5fm
fit model at tsep = 1.5fmfit result
tsep/fm
〈x〉 δ
u−δd
10.50-0.5-1
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0
mπ = 347 MeV, a = 0.064 fm, T = 24a(all T included in fit).
Two-state fit to ratios for six observables in range τ ,T − τ ≥ tstart withtstartMπ = 0.4 fixed globally for all ensembles.∆E determined only from ratios, not C2pt.→ K. Onad parallel, Thu 12:00
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 25
Simultaneous fits
∆Nπ
2Mπ
Fitted ∆0
Mπtstart
∆0/[G
eV]
0.80.70.60.50.40.30.20.1
3
2.5
2
1.5
1
0.5
0
mπ = 347 MeV, a = 0.064 fm
Two-state fit to ratios for six observables in range τ ,T − τ ≥ tstart withtstartMπ = 0.4 fixed globally for all ensembles.∆E determined only from ratios, not C2pt.→ K. Onad parallel, Thu 12:00
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 25
Fit qualityCan we trust these fits? Ideally we would require that they have good fitquality. (Not a suicient condition!)
0.0 0.2 0.4 0.6 0.8 1.0p
0.0
0.2
0.4
0.6
0.8
1.0
Cum
ulat
ive
dist
ribu
tion
func
tion
uniformCalLatPNDME
Should expect uniformly distributed p-values.Following
Sz. Borsanyi et al., Science 347, 1452 (2015) [1406.4088]
can use Kolmogorov-Smirnov test.
PNDME: p = 0.26CalLat: p = 0.00077
p-values not provided by Mainz,but reduced χ 2 are large.
0.0 0.5 1.0 1.5 2.0 2.5 3.0χ2 per degree of freedom
0.0
0.2
0.4
0.6
0.8
1.0
Cum
ulat
ive
dist
ribu
tion
func
tion
Mainz
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 26
Fit qualityCan we trust these fits? Ideally we would require that they have good fitquality. (Not a suicient condition!)
0.0 0.2 0.4 0.6 0.8 1.0p
0.0
0.2
0.4
0.6
0.8
1.0
Cum
ulat
ive
dist
ribu
tion
func
tion
uniformCalLatPNDME
Should expect uniformly distributed p-values.Following
Sz. Borsanyi et al., Science 347, 1452 (2015) [1406.4088]
can use Kolmogorov-Smirnov test.
PNDME: p = 0.26CalLat: p = 0.00077
p-values not provided by Mainz,but reduced χ 2 are large.
0.0 0.5 1.0 1.5 2.0 2.5 3.0χ2 per degree of freedom
0.0
0.2
0.4
0.6
0.8
1.0
Cum
ulat
ive
dist
ribu
tion
func
tion
Mainz
May be diicult to estimate χ 2 due to diiculty inverting large covariancematrix.
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 26
What about Nπ states?
Expected Nπ states not seen in variational or fiing results. Why?I Volume suppression? Contamination from each state ∝ L−3.
But density of states ∝ L3 compensates.I At short T , higher states dominate?=⇒ not yet in asymptotic regime.
Familiar from mesonspectroscopy.
Must include nonlocaloperators in variationalbasis to identify completespectrum.
0.10
0.15
0.20
D. J. Wilson et al. (HadSpec), Phys. Rev. D 92, 094502 (2015)[1507.02599]
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 27
Comparison of several approachessm
32sm
64sm
128
t=
10t=
13t=
16t=
19t=
22
δt=
0δt
=1
δt=
2δt
=3
δt=
0δt
=1
δt=
2δt
=3
δt=
0δt
=1
δt=
2δt
=3
δt=
2δt
=3
δt=
4δt
=2
δt=
3δt
=4
δt=
2δt
=3
δt=
4
δt=
2δt
=3
δt=
4
δt=
2δt
=3
δt=
4
t=
13t=
16
Methods
0.90
0.95
1.00
1.05
1.10
1.15
1.20
1.25
Valu
e
Fits Fits Summation 2exp 2exp 2exp Var
Smears Sink Times Method sm32 t=13 All t 10,13 t0 =2
t=13 sm32 All t 10 t 10,13 sm32 sm64 sm128 sm32sm32 sm32sm32 ∆t=2
gA Summary
mπ = 460 MeV, a = 0.074 fmJ. Dragos et al., PRD 94, 074505 (2016) [1606.03195]
1.05
1.15
1.25
1.35
g A
Plateau Summation Two-state
0.9 1.0 1.1 1.2 1.3 1.4ts[fm]
0.50
0.55
0.60
0.65
0.70
gu+
dA
Connected
0.9 1.0 1.1 1.2 1.3tlow
s [fm]
1.05
1.15
1.25
1.35
g A
Plateau Summation Two-state
0.9 1.0 1.1 1.2 1.3 1.4ts[fm]
0.50
0.55
0.60
0.65
0.70
gu+
dA
Connected
0.9 1.0 1.1 1.2 1.3tlow
s [fm]
mπ = 130 MeV, a = 0.093 fmC. Alexandrou et al. (ETMC), PRD 96, 054507 (2017)
[1705.03399]
May be best to estimate systematic uncertainty using multiple dierentapproaches.
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 28
Finite-volume eects
In general suppressed as e−mπ L .For дA, computed in heavy baryon ChPT including ∆ degrees of freedom:S. R. Beane and M. J. Savage, Phys. Rev. D 70 074029 (2004) [hep-ph/0404131]
дA (L) − дAдA
=m2π
3π 2 f 2π
[д2AF1 (mπL) + д
2∆N (1 + 25д∆∆
81дA )F2 (mπ ,∆,L)
+ F3 (mπL) + д2∆N F4 (mπ ,∆,L)
],
Neglecting loops with ∆ baryons, the leading contribution is
дA (L) − дAдA
∼ m2πд
2A
3π 2 f 2π
√π
2mπLe−mπ L
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 29
Finite-volume eects: controlled studies
0.80
0.85
0.90
0.95
1.00
1.05
2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0
gA
(L)/gA
(Lmax
)
mπL
RQCD 150 MeVRQCD 290 MeVRQCD 420 MeV
PNDME 220 MeVCalLat 220 MeVETMC 130 MeV
QCDSF 261 MeVQCDSF 290 MeVLHPC 356 MeVLHPC 254 MeVRBC 420 MeV
Reference data at Lmax indicated with black dot.
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 30
Finite-volume eects: controlled studies
0.80
0.85
0.90
0.95
1.00
1.05
0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20
gA
(L)/gA
(Lmax
)
(mπ/mphysπ )2e−mπL/
√mπL
RQCD 150 MeVRQCD 290 MeVRQCD 420 MeVPNDME 220 MeVCalLat 220 MeVETMC 130 MeV
QCDSF 261 MeVQCDSF 290 MeVLHPC 356 MeVLHPC 254 MeVRBC 420 MeV
Reference data at Lmax indicated with black dot.
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 30
Finite-volume eects: global fitsgu−d
A
MπL
1.2
1.3
3 4 5 6 7
R. Gupta et al. (PNDME), 1806.09006
дA (mπ ,L,a) = f (mπ ,a) + cm2πe−mπ L
atmphysπ ,mπL = 4: −0.9(5)% eect
0.000 0.005 0.010 0.015 0.020 0.025
e−mπL/(mπL)1/2
1.23
1.25
1.27
1.29
g A
NNLO+ct χPT
NLO χPT predictionNLO χPT prediction
C. C. Chang et al., Nature 558, 91 [1805.12130]
leading HBChPT expression (no ∆)
+c (mπ /fπ )3F1 (mπL)
Larger (few %) eect seen by Mainz group! K. Onad parallel, Thu 12:00
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 31
Finite-volume eects: global fitsgu−d
A
MπL
1.2
1.3
3 4 5 6 7
R. Gupta et al. (PNDME), 1806.09006
дA (mπ ,L,a) = f (mπ ,a) + cm2πe−mπ L
atmphysπ ,mπL = 4: −0.9(5)% eect
0.000 0.005 0.010 0.015 0.020 0.025
e−mπL/(mπL)1/2
1.23
1.25
1.27
1.29
g A
NNLO+ct χPT
NNLO χPT estimateNNLO χPT estimate
C. C. Chang et al., Nature 558, 91 [1805.12130]
leading HBChPT expression (no ∆)+c (mπ /fπ )
3F1 (mπL)
Larger (few %) eect seen by Mainz group! K. Onad parallel, Thu 12:00
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 31
Finite-volume eects: global fitsgu−d
A
MπL
1.2
1.3
3 4 5 6 7
R. Gupta et al. (PNDME), 1806.09006
дA (mπ ,L,a) = f (mπ ,a) + cm2πe−mπ L
atmphysπ ,mπL = 4: −0.9(5)% eect
0.000 0.005 0.010 0.015 0.020 0.025
e−mπL/(mπL)1/2
1.23
1.25
1.27
1.29
g A
model average
NNLO χPT estimateNNLO χPT estimate
C. C. Chang et al., Nature 558, 91 [1805.12130]
leading HBChPT expression (no ∆)+c (mπ /fπ )
3F1 (mπL)
Larger (few %) eect seen by Mainz group! K. Onad parallel, Thu 12:00
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 31
Chiral extrapolation
Increasing number of calculations at physicalmπ .Otherwise need extrapolation.
Heavy baryon ChPT for дA:
дA (mπ ) = д0 − (д0 + 2д30 )
( mπ
4πF
)2log
m2π
µ2 + c1m2π + c2m
3π + . . .
or use simple polynomials inmπ orm2π .
Range of convergence is a priori unknown.
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 32
Axial charge: extrapolated and physical-point results
0.8 0.9 1.0 1.1 1.2 1.3 1.4gA
2 clover (RQCD ’15)
2+1+1 clover on HISQ (PNDME ’16)
2 twisted mass clover (ETMC ’17)
2 clover (Mainz ’17)
2+1 overlap (JLQCD ’18)
2+1+1 domain wall on HISQ (CalLat ’18)
2+1 overlap on domain wall (χQCD ’18)
2+1+1 clover on HISQ (PNDME ’18)
2+1 clover (PACS ’18)
2+1 clover (PACS ’18b)
2+1 clover (Mainz ’18)
2+1+1 twisted mass clover (ETMC ’18)
PDG
filled symbols: published
Selection based on quality criteriastill missing!
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 33
Axial charge: extrapolated and physical-point results
0.8 0.9 1.0 1.1 1.2 1.3 1.4gA
2 clover (RQCD ’15)
2+1+1 clover on HISQ (PNDME ’16)
2 twisted mass clover (ETMC ’17)
2 clover (Mainz ’17)
2+1 overlap (JLQCD ’18)
2+1+1 domain wall on HISQ (CalLat ’18)
2+1 overlap on domain wall (χQCD ’18)
2+1+1 clover on HISQ (PNDME ’18)
2+1 clover (PACS ’18)
2+1 clover (PACS ’18b)
2+1 clover (Mainz ’18)
2+1+1 twisted mass clover (ETMC ’18)
PDG
filled symbols: published
Selection based on quality criteriastill missing!
1.05
1.15
1.25
1.35
g A
Plateau Summation Two-state
0.9 1.0 1.1 1.2 1.3 1.4ts[fm]
0.50
0.55
0.60
0.65
0.70
gu+
dA
Connected
0.9 1.0 1.1 1.2 1.3tlow
s [fm]
1.05
1.15
1.25
1.35
g A
Plateau Summation Two-state
0.9 1.0 1.1 1.2 1.3 1.4ts[fm]
0.50
0.55
0.60
0.65
0.70
gu+
dA
Connected
0.9 1.0 1.1 1.2 1.3tlow
s [fm]
One ensemble: mπ = 130 MeV,a = 0.093 fm,mπL = 3.0C. Alexandrou et al. (ETMC),Phys. Rev. D 96, 054507 (2017) [1705.03399]
1.1
1.15
1.2
1.25
1.3
1.35
0.8 1 1.2 1.4 1.6 1.8
gA
Time separation [fm]
L=48, a~0.094, Nf=2L=64, a~0.094, Nf=2 (preliminary)L=64, a~0.081, Nf=2+1+1 (preliminary)
New Nf = 2 ensemble: larger volume.C. Lauer poster
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 33
Axial charge: extrapolated and physical-point results
0.8 0.9 1.0 1.1 1.2 1.3 1.4gA
2 clover (RQCD ’15)
2+1+1 clover on HISQ (PNDME ’16)
2 twisted mass clover (ETMC ’17)
2 clover (Mainz ’17)
2+1 overlap (JLQCD ’18)
2+1+1 domain wall on HISQ (CalLat ’18)
2+1 overlap on domain wall (χQCD ’18)
2+1+1 clover on HISQ (PNDME ’18)
2+1 clover (PACS ’18)
2+1 clover (PACS ’18b)
2+1 clover (Mainz ’18)
2+1+1 twisted mass clover (ETMC ’18)
PDG
filled symbols: published
Selection based on quality criteriastill missing!
0.00 0.05 0.10 0.15 0.20 0.25 0.30ǫπ = mπ/(4πFπ)
1.10
1.15
1.20
1.25
1.30
1.35
g A
model average gLQCDA (ǫπ , a = 0)
gPDGA = 1.2723(23)
gA(ǫπ , a ≃ 0.15 fm)gA(ǫπ , a ≃ 0.12 fm)gA(ǫπ , a ≃ 0.09 fm)
a ≃ 0.15 fma ≃ 0.12 fma ≃ 0.09 fm
Bayesian average over fit models forдA (mπ /Fπ ,mπL,a
2/w20 ).
C. C. Chang et al., Nature 558, 91 (2018)[1805.12130]
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 33
Axial charge: extrapolated and physical-point results
0.8 0.9 1.0 1.1 1.2 1.3 1.4gA
2 clover (RQCD ’15)
2+1+1 clover on HISQ (PNDME ’16)
2 twisted mass clover (ETMC ’17)
2 clover (Mainz ’17)
2+1 overlap (JLQCD ’18)
2+1+1 domain wall on HISQ (CalLat ’18)
2+1 overlap on domain wall (χQCD ’18)
2+1+1 clover on HISQ (PNDME ’18)
2+1 clover (PACS ’18)
2+1 clover (PACS ’18b)
2+1 clover (Mainz ’18)
2+1+1 twisted mass clover (ETMC ’18)
PDG
filled symbols: published
Selection based on quality criteriastill missing!
1:24 1:28 1:32
gA
model avg
NNLO fflPT
NNLO+ct fflPT
NLO Taylor ›2ı
NNLO Taylor ›2ı
NLO Taylor ›ı
NNLO Taylor ›ı
+O(¸sa2) disc.
+O(a) disc.
omit FV
NLO FV
2×LO width
2× all widths
mı ≤ 350 MeV
mı ≤ 310 MeV
mı ≥ 220 MeV
a ≤ 0:12 fm
a ≥ 0:12 fm
N3LO fflPT
NLO fflPT(∆)
0:0 0:5 1:0
ffl2aug=dof
0:0 0:5 1:0
Bayes factor
C. C. Chang et al., Nature 558, 91 (2018)[1805.12130]
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 33
Axial charge: extrapolated and physical-point results
0.8 0.9 1.0 1.1 1.2 1.3 1.4gA
2 clover (RQCD ’15)
2+1+1 clover on HISQ (PNDME ’16)
2 twisted mass clover (ETMC ’17)
2 clover (Mainz ’17)
2+1 overlap (JLQCD ’18)
2+1+1 domain wall on HISQ (CalLat ’18)
2+1 overlap on domain wall (χQCD ’18)
2+1+1 clover on HISQ (PNDME ’18)
2+1 clover (PACS ’18)
2+1 clover (PACS ’18b)
2+1 clover (Mainz ’18)
2+1+1 twisted mass clover (ETMC ’18)
PDG
filled symbols: published
Selection based on quality criteriastill missing!
gu−d
A
a [fm]
a15m310a12m310a12m220La12m220a12m220S
a09m310a09m220a09m130a06m310a06m220
a06m135extrap.1-extrap.
1.2
1.3
0 0.03 0.06 0.09 0.12 0.15
gu−d
A
M2π [GeV2]
1.2
1.3
0.02 0.04 0.06 0.08 0.1 0.12
gu−d
A
MπL
1.2
1.3
3 4 5 6 7
Main ansatz:
дA (mπ ,L,a) = c1 + c2a + c3m2π
+ c4m2πe−mπ L
R. Gupta et al. (PNDME), 1806.09006
R. Gupta parallel, Thu 12:40
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 33
Axial charge: extrapolated and physical-point results
0.8 0.9 1.0 1.1 1.2 1.3 1.4gA
2 clover (RQCD ’15)
2+1+1 clover on HISQ (PNDME ’16)
2 twisted mass clover (ETMC ’17)
2 clover (Mainz ’17)
2+1 overlap (JLQCD ’18)
2+1+1 domain wall on HISQ (CalLat ’18)
2+1 overlap on domain wall (χQCD ’18)
2+1+1 clover on HISQ (PNDME ’18)
2+1 clover (PACS ’18)
2+1 clover (PACS ’18b)
2+1 clover (Mainz ’18)
2+1+1 twisted mass clover (ETMC ’18)
PDG
filled symbols: published
Selection based on quality criteriastill missing!
0 5 10 15t/a
0.8
1
1.2
1.4
1.6
PDG16m
π=0.146 GeV
gA
ren
One ensemble: mπ = 146 MeV,a = 0.085 fm,mπL = 6.0, T = 1.3 fm.K.-I. Ishikawa et al. (PACS), 1807.03974
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 33
Axial charge: extrapolated and physical-point results
0.8 0.9 1.0 1.1 1.2 1.3 1.4gA
2 clover (RQCD ’15)
2+1+1 clover on HISQ (PNDME ’16)
2 twisted mass clover (ETMC ’17)
2 clover (Mainz ’17)
2+1 overlap (JLQCD ’18)
2+1+1 domain wall on HISQ (CalLat ’18)
2+1 overlap on domain wall (χQCD ’18)
2+1+1 clover on HISQ (PNDME ’18)
2+1 clover (PACS ’18)
2+1 clover (PACS ’18b)
2+1 clover (Mainz ’18)
2+1+1 twisted mass clover (ETMC ’18)
PDG
filled symbols: published
Selection based on quality criteriastill missing!
0 5 10 15t/a
0.8
1
1.2
1.4
1.6
PDG16m
π=0.146 GeV
gA
ren
One ensemble: mπ = 146 MeV,a = 0.085 fm,mπL = 6.0, T = 1.3 fm.K.-I. Ishikawa et al. (PACS), 1807.03974
New ensemble: mπ = 135 MeV,L = 10.8 fm. Y. Kuramashi parallel, Thu 9:50
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 33
Axial charge: extrapolated and physical-point results
0.8 0.9 1.0 1.1 1.2 1.3 1.4gA
2 clover (RQCD ’15)
2+1+1 clover on HISQ (PNDME ’16)
2 twisted mass clover (ETMC ’17)
2 clover (Mainz ’17)
2+1 overlap (JLQCD ’18)
2+1+1 domain wall on HISQ (CalLat ’18)
2+1 overlap on domain wall (χQCD ’18)
2+1+1 clover on HISQ (PNDME ’18)
2+1 clover (PACS ’18)
2+1 clover (PACS ’18b)
2+1 clover (Mainz ’18)
2+1+1 twisted mass clover (ETMC ’18)
PDG
filled symbols: published
Selection based on quality criteriastill missing!
lattice data (corrected)value at Mπ,phys
chiral + continuum + FS extrapolation
Mπ/GeV
t0M2π
gu−d
A
0.350.300.250.200.150.100
0.070.060.050.040.030.020.010
1.4
1.3
1.2
1.1
1
0.9
0.8
Ensembles havems varied such thatms +mu +md = const.Fit ansatz:
дA (mπ ,L,a) = c1 + c2a2 + c3m
2π
+ c4m2πe−mπ L
K. Onad parallel, Thu 12:00
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 33
Axial charge: extrapolated and physical-point results
0.8 0.9 1.0 1.1 1.2 1.3 1.4gA
2 clover (RQCD ’15)
2+1+1 clover on HISQ (PNDME ’16)
2 twisted mass clover (ETMC ’17)
2 clover (Mainz ’17)
2+1 overlap (JLQCD ’18)
2+1+1 domain wall on HISQ (CalLat ’18)
2+1 overlap on domain wall (χQCD ’18)
2+1+1 clover on HISQ (PNDME ’18)
2+1 clover (PACS ’18)
2+1 clover (PACS ’18b)
2+1 clover (Mainz ’18)
2+1+1 twisted mass clover (ETMC ’18)
PDG
filled symbols: published
Selection based on quality criteriastill missing!
1.1
1.15
1.2
1.25
1.3
1.35
0.8 1 1.2 1.4 1.6 1.8
gA
Time separation [fm]
Two state fitL=64, a~0.081, Nf=2+1+1 (preliminary)
One ensemble: physicalmπ ,a = 0.081 fm,mπL = 3.6C. Lauer posterM. Constantinou parallel, Fri 17:50
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 33
Axial charge: extrapolated and physical-point results
0.8 0.9 1.0 1.1 1.2 1.3 1.4gA
2 clover (RQCD ’15)
2+1+1 clover on HISQ (PNDME ’16)
2 twisted mass clover (ETMC ’17)
2 clover (Mainz ’17)
2+1 overlap (JLQCD ’18)
2+1+1 domain wall on HISQ (CalLat ’18)
2+1 overlap on domain wall (χQCD ’18)
2+1+1 clover on HISQ (PNDME ’18)
2+1 clover (PACS ’18)
2+1 clover (PACS ’18b)
2+1 clover (Mainz ’18)
2+1+1 twisted mass clover (ETMC ’18)
PDG
filled symbols: published
Selection based on quality criteriastill missing!
Other preliminary results:
1
1.1
1.2
1.3
1.4
1.5
8 9 10 11 12
gA/gV
T
2:-2 ft3:-3 ft4:-4 ft
Experiment
Nf = 2 + 1 domain wall (RBC+LHPC)One ensemble: mπ = 139 MeV,a = 0.11 fm,mπL = 3.9S. Ohta parallel, Thu 11:40
Nf = 2 + 1 + 1 staggered(Fermilab-MILC)Blinded analysis.Y. Lin parallel, Thu 9:30
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 33
Tensor and scalar charges
0.7 0.8 0.9 1.0 1.1 1.2gT (MS, 2 GeV)
2+1 clover (LHPC ’12)
2 clover (RQCD ’15)
2+1+1 clover on HISQ (PNDME ’16)
2 twisted mass clover (ETMC ’17)
2+1 overlap (JLQCD ’18)
2+1+1 clover on HISQ (PNDME ’18)
2+1 clover (Mainz ’18)
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4gS (MS, 2 GeV)
2+1 clover (LHPC ’12)
2 clover (RQCD ’15)
2+1+1 clover on HISQ (PNDME ’16)
2 TM-clover (ETMC ’17)
2+1 overlap (JLQCD ’18)
2+1+1 clover on HISQ (PNDME ’18)
2+1 clover (Mainz ’18)
LQCD average (∆mN/∆mq)QCD(Gonzalez-Alonso ’14)
Selection based on quality criteria still missing!
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 34
Sigma terms
25 30 35 40 45 50 55 60 65
RQCD
ETMC
χQCD
BMWc
M. Hoferichter et al., PRL 115, 092301
0 20 40 60 80 100 120 140 160σπN (MeV) σs (MeV)
phenomenologyLQCD Feynman-HellmannLQCD direct
Published results with (near-)physicalmπ .Update on BMWc Nf =1+1+1+1 results: L. Varnhorst and C. Hoelbling, Wed. 16:10 and 16:30
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 35
Outlook
I Excited-state contamination is a major focus.A full variational study including nonlocal Nπ interpolators would helpdefinitively resolve the issue.
I No sign of large finite-volume eect in fully-controlled studiesat low pion mass.
I For LQCD averages, individual calculations must be selected usingquality criteria. Fiing approaches for excited states vary; seingstandards may be diicult.I One aempt: H.-W. Lin et al., “Parton distributions and laice QCD calculations: A
community white paper” Prog. Part. Nucl. Phys. 100, 107 (2018) [1711.07916]I Some nucleon structure to appear in next FLAG review.
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 36
Thanks to all who sent results and replied to my questions:
Constantia AlexandrouChia Cheng Chang
Colin EgererRajan Gupta
Jian LiangShigemi Ohta
Konstantin OnadFinn Stokes
André Walker-LoudTakeshi Yamazaki
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 37
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 38
Scalar and tensor charges
Precision β-decay experiments may be sensitive to BSM physics; leadingcontributions are controlled by the scalar and tensor charges:T. Bhaacharya et al., Phys. Rev. D 85, 054512 (2012) [1110.6448]
〈p (P , s ′) |ud |n(P , s )〉 = дS up (P , s ′)un (P , s ),〈p (P , s ′) |uσ µνd |n(P , s )〉 = дT up (P , s ′)σ µνun (P , s ).
Tensor charges also control contribution from quark electric dipole momentto neutron EDM:
LqEDM = − 12
∑q
dqqiσµνγ5qFµν =⇒ dn = duд
dT + ddд
uT + dsд
sT + · · · .
For scalar charge, conserved vector current relation implies
дS =(mn −mp )QCD
md −mu,
up to second-order isospin breaking.M. González-Alonso and J. Martin Camalich, Phys. Rev. Le. 112, 042501 (2014) [1309.4434]
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 39
Sigma terms
The contributions from quark masses to the nucleon mass.
σπN =mud 〈N |uu + dd |N 〉 σq =mq〈N |qq |N 〉Control the sensitivity of direct-detection dark maer searches to WIMPsthat interact via Higgs exchange.
Phenomenological value σπN = 59.1 ± 3.5 MeV determined based oncombining:I Cheng-Dashen low-energy theoremI Roy-Steiner equations to constrain πN scaering amplitudeI Precise πN scaering lengths from pionic atoms
M. Hoferichter et al., Phys. Rev. Le. 115, 092301 (2015) [1506.04142]
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 40
Variational approach: nonzero momentum
16 18 20 22 24 26 28 30 32 34
t/a
−1.0
−0.8
−0.6
−0.4
−0.2
0.0
0.2
GM(Q
2)/µN
dp (PEVA) dp (Conv.)
Include operators with negative parity in variational basis.F. Stokes et al., in preparation Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 41
Discretization eects: controlled studies
C. Alexandrou et al. (ETMC),Phys. Rev. D 83, 045010 (2011) [1012.0857]
Nf = 2 Wilson twisted mass.No significant eect seen.
1.05
1.1
1.15
1.2
1.25
1.3
1.35
1.4
0 0.001 0.002 0.003 0.004 0.005 0.006 0.007 0.008
PRELIMINARY
g A
a2 [fm2]
Nf = 3, mπ ≈ 420 MeVExpt
G. S. Bali et al. (RQCD), PoS LATTICE2016, 106[1702.01035]
Nf = 3 clover.Linear in a2/t0 extrapolations.
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 42
Discretization eects: global fitsgu−d
A
a [fm]
a15m310a12m310a12m220La12m220a12m220S
a09m310a09m220a09m130a06m310a06m220
a06m135extrap.1-extrap.
1.2
1.3
0 0.03 0.06 0.09 0.12 0.15
R. Gupta et al. (PNDME), 1806.09006
Nf = 2 + 1 + 1 clover on HISQ.pink: дA (mπ ,L,a) = f (mπ ,L) + ca/r1gray: дA (mπ ,L,a) = дA + ca/r1
0.00 0.01 0.02 0.03 0.04 0.05 0.06ǫ2
a = a2/(4πw20 )
1.10
1.15
1.20
1.25
1.30
1.35
g A
model average gLQCDA (ǫphys.
π , ǫa)
gPDGA = 1.2723(23)
gA(ǫ(130)π , ǫa)
gA(ǫ(220)π , ǫa)
gA(ǫ(310)π , ǫa)
gA(ǫ(350)π , ǫa)
gA(ǫ(400)π , ǫa)
C. C. Chang et al., Nature 558, 91 [1805.12130]
Nf = 2 + 1 + 1 domain wall on HISQ.дA (mπ ,L,a) = f (mπ ,L) + ca
2/w20
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 43
Chiral log?
Preferred fit: дA (mπ ) = c1 + c2m2π
ChPT: дA (mπ ) = д0 − (д0 + 2д30 )
( mπ
4πF
)2logm2
π + cm2π + . . .
0.00 0.05 0.10 0.15 0.20 0.25 0.30ǫπ = mπ/(4πFπ)
1.10
1.15
1.20
1.25
1.30
1.35
g A
model average gLQCDA (ǫπ , a = 0)
gPDGA = 1.2723(23)
gA(ǫπ , a ≃ 0.15 fm)gA(ǫπ , a ≃ 0.12 fm)gA(ǫπ , a ≃ 0.09 fm)
a ≃ 0.15 fma ≃ 0.12 fma ≃ 0.09 fm
0.00 0.05 0.10 0.15 0.20 0.25 0.30ǫπ = mπ/(4πFπ)
1.10
1.15
1.20
1.25
1.30
1.35
g A
NNLO χPT gLQCDA (ǫπ , a = 0)
gPDGA = 1.2723(23)
gA(ǫπ , a ≃ 0.15 fm)gA(ǫπ , a ≃ 0.12 fm)gA(ǫπ , a ≃ 0.09 fm)
a ≃ 0.15 fma ≃ 0.12 fma ≃ 0.09 fm
C. C. Chang et al., Nature 558, 91 (2018) [1805.12130]
0.95 1.00 1.05 1.10 1.15 1.20 1.25 1.30 1.35Axial charge in chiral limit
2+1+1 domain wall on HISQ (CalLat ’18)
2+1 clover (Mainz ’18)
2+1+1 domain wall on HISQ (CalLat ’18)
2+1 clover (Mainz ’18)
Preferred fit
ChPT May need precise data belowmphysπ
to see chiral log.
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 44
Chiral log?
Preferred fit: дA (mπ ) = c1 + c2m2π
ChPT: дA (mπ ) = д0 − (д0 + 2д30 )
( mπ
4πF
)2logm2
π + cm2π + . . .
0.00 0.05 0.10 0.15 0.20 0.25 0.30ǫπ = mπ/(4πFπ)
1.10
1.15
1.20
1.25
1.30
1.35
g A
NNLO χPT
LONLO
NNLO
0.00 0.05 0.10 0.15 0.20 0.25 0.30ǫπ = mπ/(4πFπ)
1.10
1.15
1.20
1.25
1.30
1.35
g A
NNLO χPT gLQCDA (ǫπ , a = 0)
gPDGA = 1.2723(23)
gA(ǫπ , a ≃ 0.15 fm)gA(ǫπ , a ≃ 0.12 fm)gA(ǫπ , a ≃ 0.09 fm)
a ≃ 0.15 fma ≃ 0.12 fma ≃ 0.09 fm
C. C. Chang et al., Nature 558, 91 (2018) [1805.12130]
0.95 1.00 1.05 1.10 1.15 1.20 1.25 1.30 1.35Axial charge in chiral limit
2+1+1 domain wall on HISQ (CalLat ’18)
2+1 clover (Mainz ’18)
2+1+1 domain wall on HISQ (CalLat ’18)
2+1 clover (Mainz ’18)
Preferred fit
ChPT May need precise data belowmphysπ
to see chiral log.
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 44
Chiral log?
Preferred fit: дA (mπ ) = c1 + c2m2π
ChPT: дA (mπ ) = д0 − (д0 + 2д30 )
( mπ
4πF
)2logm2
π + cm2π + . . .
0.00 0.05 0.10 0.15 0.20 0.25 0.30ǫπ = mπ/(4πFπ)
1.10
1.15
1.20
1.25
1.30
1.35
g A
NNLO χPT
LONLO
NNLO
0.00 0.05 0.10 0.15 0.20 0.25 0.30ǫπ = mπ/(4πFπ)
1.10
1.15
1.20
1.25
1.30
1.35
g A
NNLO χPT gLQCDA (ǫπ , a = 0)
gPDGA = 1.2723(23)
gA(ǫπ , a ≃ 0.15 fm)gA(ǫπ , a ≃ 0.12 fm)gA(ǫπ , a ≃ 0.09 fm)
a ≃ 0.15 fma ≃ 0.12 fma ≃ 0.09 fm
C. C. Chang et al., Nature 558, 91 (2018) [1805.12130]
0.95 1.00 1.05 1.10 1.15 1.20 1.25 1.30 1.35Axial charge in chiral limit
2+1+1 domain wall on HISQ (CalLat ’18)
2+1 clover (Mainz ’18)
2+1+1 domain wall on HISQ (CalLat ’18)
2+1 clover (Mainz ’18)
Preferred fit
ChPT May need precise data belowmphysπ
to see chiral log.
Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 44
Calculations with near-physicalmπ
1.00
1.05
1.10
1.15
1.20
1.25
1.30
1.35
3.0 4.0 5.0 6.0 7.0
gA(near-ph
ysical
mπ)
mπL
Nf=2 TM-clover (ETMC)Nf=2+1+1 TM-clover (ETMC)Nf=2 clover (RQCD)Nf=2+1 clover (PACS)Nf=2+1+1 clover/HISQ (PNDME)Nf=2+1+1 DW/HISQ (CalLat)
see also: Nf = 2 + 1 domain wall (RBC+LHPC) → S. Ohta parallel, Thu 11:40
Nf = 2 + 1 + 1 HISQ staggered → Y. Lin parallel, Thu 9:30Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 45
Calculations with near-physicalmπ
1.00
1.05
1.10
1.15
1.20
1.25
1.30
1.35
0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16
gA(near-ph
ysical
mπ)
a (fm)
Nf=2 TM-clover (ETMC)Nf=2+1+1 TM-clover (ETMC)Nf=2 clover (RQCD)Nf=2+1 clover (PACS)Nf=2+1+1 clover/HISQ (PNDME)Nf=2+1+1 DW/HISQ (CalLat)
see also: Nf = 2 + 1 domain wall (RBC+LHPC) → S. Ohta parallel, Thu 11:40
Nf = 2 + 1 + 1 HISQ staggered → Y. Lin parallel, Thu 9:30Jeremy Green | DESY, Zeuthen | Laice 2018 | Page 45