PhysRevD 109 094516

PHYSICAL REVIEW D 109, 094516 (2024)
Multiparticle interpolating operators in quantum field theories

with cubic symmetry
William Detmold ,1,2 William I. Jay ,1,* Gurtej Kanwar ,3,2,†
Phiala E. Shanahan ,1,2 and Michael L. Wagman 4,‡
1
Center for Theoretical Physics, Massachusetts Institute of Technology,
Cambridge, Massachusetts 02139, USA
2
The NSF AI Institute for Artificial Intelligence and Fundamental Interactions
3
Albert Einstein Center, Institute for Theoretical Physics, University of Bern, 3012 Bern, Switzerland
4
Fermi National Accelerator Laboratory, Batavia, Illinois 60510, USA
(Received 20 March 2024; accepted 6 May 2024; published 31 May 2024)
Numerical studies of lattice quantum field theories are conducted in finite spatial volumes, typically with
cubic symmetry in the spatial coordinates. Motivated by these studies, this work presents a general
algorithm to construct multiparticle interpolating operators for quantum field theories with cubic symmetry.
The algorithm automates the block diagonalization required to combine multiple operators of definite linear
momentum into irreducible representations of the appropriate little group. Examples are given for
distinguishable and indistinguishable particles including cases with both zero and nonzero spin.
DOI: 10.1103/PhysRevD.109.094516
I. INTRODUCTION depends sensitively on the choice of interpolating operators

used in the calculation. To determine the energies of the
The determination of energy spectra is a central task in
eigenstates most effectively, it is advantageous to project
numerical studies of lattice quantum field theories (QFTs)
these interpolating operators to particular symmetry sec-
and is the precursor to more complex studies of the
tors, thereby reducing the number of states that contribute
properties and interactions of the states in the theory. In
to the corresponding correlation functions.
strongly-coupled field theories, very little is known about
Numerical studies in lattice QFTs are usually performed
the spectra a priori. However, analysis of the Euclidean-
in a finite cubic spatial volume. In this setting, continuous
time dependence of two-point correlation functions
rotational symmetry is broken down to a discrete subgroup,
between operators with the quantum numbers of the states
which (combined with spatial inversions) is the cubic
of interest provides an avenue for first-principles determi-
group, Oh. Moreover, provided translational symmetry is
nations of spectra. This approach has been used to explore
preserved (e.g., via the use of periodic boundary condi-
many different field theories, most notably to determine the
tions), the total momentum of an energy eigenstate is a
low-energy excitations in the hadronic and nuclear spectra
conserved quantity which further reduces the spatial
in quantum chromodynamics (QCD), the theory of the
symmetry. Specifically, the little group of rotations that
strong interactions. The wide-reaching goals and achieve-
leave a given total momentum P invariant is the subgroup
ments of lattice QCD are summarized, e.g., in Refs. [1–6].
GP ≤ Oh . Understanding the transformation properties of
To construct correlation functions sensitive to the eigen-
states—and interpolating operators—under the cubic group
states of a strongly-interacting theory such as QCD requires
and its subgroups is thus critical in analysis of lattice QFT
the use of interpolating operators, composite objects built
calculations [7,8].
from products of the elementary fields of the theory. In
Multiparticle interpolating operators constructed in coor-
numerical studies, the behavior of correlation functions
dinate space or through momentum projection typically
transform reducibly under these finite groups, obscuring the
*
[email protected] transformation properties of energy eigenstates. However, it
†
[email protected] is possible to decompose the operators into linear subspaces
‡
[email protected] that do not mix with each other and only transform
internally. A complete decomposition into these irreducible
Published by the American Physical Society under the terms of representations (irreps) is possible for any set of operators
the Creative Commons Attribution 4.0 International license.
Further distribution of this work must maintain attribution to that is closed under the little group [9]. For example, suppose
the author(s) and the published article’s title, journal citation, an interpolating operator transforms under a reducible
and DOI. Funded by SCOAP3. representation labeled by s, with representation matrices
2470-0010=2024=109(9)=094516(28) 094516-1 Published by the American Physical Society

DETMOLD, JAY, KANWAR, SHANAHAN, and WAGMAN PHYS. REV. D 109, 094516 (2024)
DðsÞ ðRÞ for R ∈ GP. Decomposing the reducible represen- V ¼ L3 , lattice spacing a, and with associated Hilbert
tation into a direct sum of irreps ΓðsÞ ¼ Γa ⊕ Γb ⊕ , there space H and vacuum state jΩiH ∈ H.1 The spatial sym-
exist block-diagonalization (or change-of-basis) matrices metry group is therefore the cubic group, Oh . In the
UðsÞ specific to ΓðsÞ that enact the decomposition into irrep following, all quantities will be given in lattice units with
matrices DðΓÞ ðRÞ according to a ¼ 1, where Λ ∼ ðZL Þ3 .
Multiparticle states can be created by acting on the
½UðsÞ † DðsÞ ðRÞUðsÞ ¼ ⨁DðΓi Þ ðRÞ: ð1Þ vacuum with interpolating operators with the quantum
i numbers of the desired states. Often, products of N local
operators serve as useful interpolating operators for multi-
After changing to the irrep basis, the transformation proper- particle states. Such a product may generically be written as
ties of multiparticle operators—and the states that they O1 ðx1 Þ ON ðxN Þ, where the labels x1 ; …; xN ∈ Λ indi-
create—are then simpler to understand. Previous work has cate three-vector coordinates of spatial lattice sites. The
addressed the construction of lattice interpolating operators operators Oi are built from the fundamental fields in the
for single baryons [10,11], single mesons [12], two-hadron theory and need not be distinguishable. For simplicity,
systems [13,14] with arbitrary spin and momenta [15,16], the initial discussion will focus on distinguishable oper-
and three-boson systems [17]. These results cover the ators transforming as scalars under spatial rotations, in
construction of local and extended operators with definite which case only the coordinates x1 ; …; xN of the operators
cubic transformation properties, as well as their combination transform under the spatial symmetry group. The extension
into irreps of the relevant little group in cases of up to three to operators with nonzero spin or composite operators with
local operators. nontrivial Oh -transformation properties is addressed in
The present work provides a concrete algorithm and a Sec. III. The extension to indistinguishable operators is
numerical implementation [18] that carries out the block addressed in Sec. IV.
diagonalization for any product of N operators with definite For a fixed list of N local operators, O1 ; …; ON , this
momentum, spin, and permutation properties. Each oper- work investigates the transformation properties of linear
ator can be a point-like operator, a smeared or extended combinations of products constructed via
operator, or an even more general construction. It is further
shown that the block-diagonalization matrices can be X X
determined for all N by enumerating a small set of cðx1 ; …; xN ÞO1 ðx1 Þ ON ðxN Þ; ð2Þ
examples. For spin-zero operators, only examples from x1 ∈ Λ xN ∈ Λ
N ¼ 1 and N ¼ 2 operators are required to specify the
decomposition for general N. Additional internal sym-
metries such as flavor, as well as any combination of where the cðx1 ; …; xN Þ ∈ C are V N arbitrary coefficients.2
fermionic and bosonic operator-exchange symmetries, can Acting on the vacuum with any such linear combination
be incorporated with a simple extension of the formalism yields a state
that is also described herein.
The remainder of this article is organized as follows. jψiH
Section II lays out the formalism for building the X X
block-diagonalization matrices for the simple case of N ≡ cðx1 ; …; xN ÞO1 ðx1 Þ ON ðxN ÞjΩiH : ð3Þ
distinguishable, spin-zero operators. Section III addresses x1 ∈ Λ xN ∈ Λ
distinguishable operators with spin, including a generic
method for calculating block-diagonalization matrices; con- Invariance of the vacuum state under spatial rotations
crete results are given for several examples. Section IV means that any such jψiH inherits the transformation
presents the generalization of this construction to operators properties of the multiparticle operator itself.
involving internal symmetries or identical particles. The space defined by the coefficients cðx1 ; …; xN Þ forms
Section V collects the formalism of the preceding sections
a V N -dimensional vector space which will be denoted V.
to give the complete block-diagonalization algorithm.
A symbolic (position-space) basis for V consists of the
Finally, Sec. VI provides an outlook. The appendices specify
vectors jx1 ; …; xN i for each choice of fxi ∈ Λg. These
group-theoretical conventions, discuss the method of polari-
zation tensors for evaluating irrep matrices, and give explicit 1
examples of the block-diagonalization matrices appearing The temporal geometry (discrete or continuous) and the
in Eq. (1). spacetime metric (Euclidean or Minkowski) are left unspecified
because the classification of operator representations under
2 Pis insensitive to theseRchoices.
spatial rotations/translations Q 3
II. DISTINGUISHABLE SPIN-ZERO OPERATORS In a continuum theory, xi ∈ Λ is replaced with i d xi and
cðx1 ; …; xN Þ should be replaced by a continuous field over N
Consider a lattice QFT defined on a geometry whose coordinates, but otherwise the formalism developed in this work
spatial structure is a periodic cubic lattice Λ with volume applies.
094516-2
MULTIPARTICLE INTERPOLATING OPERATORS IN QUANTUM … PHYS. REV. D 109, 094516 (2024)
basis vectors are defined to be orthonormal with respect to based on their definition in Eq. (5). Through the map L
the inner product on V, [defined following Eq. (4)], this defines the action of
rotations on the corresponding interpolating operators.
hx01 ; …; x0N jx1 ; …; xN i ¼ δx01 ;x1 δx0N ;xN : ð4Þ The remainder of this section proceeds as follows.
First, Sec. II A constructs the reducible representations
To avoid ambiguity, elements of the abstract space V are associated with interpolating operators in the form of
written with unadorned kets, while quantum states such as Eq. (6) for distinguishable spin-zero operators. Second,
jΩiH or jψiH carry a subscript. Since the list of operators is Sec. II B reviews the irreps of the cubic group and its
fixed, each vector in V is associated uniquely with an subgroups. Section II C presents the algorithm for con-
interpolating operator by the linear map L∶ jx1 ; …; xN i ↦ structing the block-diagonalization matrices that change
O1 ðx1 Þ ON ðxN Þ. basis from the interpolating operators in Eq. (6) into
A useful starting point for the decomposition of a state operators that transform irreducibly. Section II D discusses
jψiH into subspaces that transform irreducibly under spatial a classification of cases with identical block-diagonaliza-
symmetries is the plane-wave basis of V, defined by vectors tion matrices using the stabilizer groups of specific
momenta, and finally Sec. II E presents examples of the
jn1 ; …; nN i block-diagonalization that are sufficient to implement this
X X 2π 2π
process for any number of spin-zero operators.
≡ ei L n1 ·x1 ei L nN ·xN jx1 ; …; xN i; ð5Þ
x1 ∈ Λ xN ∈ Λ
A. Momentum orbits and their representations
for each possible (ordered) set of wave vectors To construct the reducible representations associated
fni ∈ ðZL Þ3 g.3 Here, the distinction between the posi- with plane-wave interpolating operators, the space V is
tion-space and plane-wave bases is made by the use of systematically decomposed into subspaces associated with
the letters x and n, respectively. Since there is a one-to-one momentum orbits, i.e., sets of plane-wave momenta closed
mapping between wave vectors ni and momenta ð2π=LÞni , under little-group transformations. These subspaces,
these terms will be used interchangeably for the wave labeled by s, each correspond to a definite total momentum
vectors ni . Each vector jn1 ; …; nN i is associated with a P and carry a reducible representation ΓðsÞ of the little
plane-wave interpolating operator, group GP . The representation matrices DðsÞ ðRÞ take a
simple form and are explicitly constructed below.
Õ1 ðn1 Þ ÕN ðnN Þ
X X 2π
Given a plane Pwave jn1 ; …; nN i ∈ V, the total momen-
2π tum is P ¼ 2π i ni . The ordered list of wave vectors
≡ ei L n1 ·x1 O1 ðx1 Þ ei L nN ·xN ON ðxN Þ: ð6Þ L
defining the basis state jn1 ; …; nN i is denoted by
x1 ∈ Λ xN ∈ Λ
½n1 ; …; nN and will be used to define momentum orbits
The position-space basis vectors jx1 ; …; xN i and plane- below. The set of such lists of wave vectors can be
wave basis vectors jn1 ; …; nN i have simple transformation partitioned into disjoint subsets K P of fixed total momen-
properties under Oh and its subgroups. Each group element tum P,
R ∈ Oh can be specified by its action on arbitrary three-
vectors x ∈ Λ, i.e., with each group element defined as a 2π X
K P ≡ ½n1 ; …; nN ∶ n ¼P : ð9Þ
unique orthogonal 3 × 3 matrix, R ∈ Oh ≤ Oð3Þ. A con- L i i
crete specification of these matrices is presented in
Appendix A. When acting on vectors in V, the abstract For each total momentum P, the associated little group is
operator associated with rotation R will be denoted defined as the subgroup GP ≤ Oh that leaves P invariant,
DðRÞ. Its action on the position-space basis vectors is
given by GP ≡ fR ∈ Oh ∶RP ¼ Pg: ð10Þ
DðRÞjx1 ; …; xN i ¼ jRx1 ; …; RxN i ð7Þ The possible little groups, up to equivalence under rotation
of the total momentum, are denoted C4v , C3v , C2v , CR2 , CP2 ,
and extends linearly to the rest of V, including to the plane- and C1 following the notation of Ref. [9], where Cnv is the
wave basis vectors, which transform as symmetry group of the n-gon with 2n elements and Cn is
the cyclic group with n elements. The little groups CR2 and
DðRÞjn1 ; …; nN i ¼ jRn1 ; …; RnN i ð8Þ CP2 both have the group structure of C2 , but appear as little
groups for inequivalent choices of total momenta. Details
of these groups are presented in Appendix A; note that to
3
In a continuum theory, ZL is replaced with Z. match these definitions the little group may need to be
094516-3
rotated (by conjugation) so that the total momentum aligns These subspaces are linearly independent and jointly
with the present conventions. compose the full space,
By construction, the space K P is closed under the action
ðsÞ
of the little group GP . However, it can be further partitioned V ¼ ⨁V P ; ð16Þ
into minimal invariant subsets that are closed under the P;s
ðsÞ
action of GP . These momentum orbits, denoted by K P , are where the sum ranges over all possible total momenta P and
constructed by computing the sets (for a given P) over all distinct orbits labeled by s.
ðsÞ
The vectors in V P inherit their transformation properties
ðsÞ
K P ≡ fR · ½n1 ; …; nN ∶R ∈ GP g ⊆ K P ; ð11Þ from the plane waves,
where R · ½n1 ; …; nN ¼ ½Rn1 ; …; RnN . Here, s is an ðsÞ

DðRÞjs; mi ¼ jR · ½n1 ; …; nN m i ¼ js; m0 i; ð17Þ
abstract label distinguishing inequivalent subsets; the label
s can also be defined concretely with one representative or which extends by linearity to the entire space. Each element
ðsÞ ðsÞ
fiducial list of wave vectors per orbit. Two examples of K P R ∈ GP permutes the basis vectors of V P due to the
are given by the orbits of the wave vectors ½n1 ; n2 ¼ invariance property Eq. (15) and therefore acts as a linear
½ð0; 0; 1Þ; ð0; 0; −1Þ and ½n1 ; n2 ¼ ½ð0; 2; 1Þ; ð0; −2; −1Þ ðsÞ ðsÞ
operator V P → V P with matrix elements
with GP ¼ Oh as these cases both have vanishing total
momentum. ðsÞ
Dm0 m ðRÞ ≡ hs; m0 jDðRÞjs; mi ∈ f0; 1g: ð18Þ
It is useful to provide a conventional way of ordering the
elements of a given momentum orbit. First, select a fiducial
These matrices are precisely the representation matrices
arrangement of momenta ½n1 ; …; nN . With P being the
DðsÞ appearing in Eq. (1). The superscript on DðsÞ empha-
total momentum defined by the fiducial momenta, let
sizes the block-diagonal nature of this representation within
½R1 ; …; Rn , with n ¼ jGP j, denote the fixed ordering of
the larger space V, since distinct momentum orbits do not
the elements of little group GP given in Appendix A.4 The
ðsÞ mix under the cubic group. Each such space therefore
orbit K P then inherits its ordering from the little group via forms a (generally reducible) representation which will be
Eq. (11), with the group elements acting in order on the denoted by ΓðsÞ.
ðsÞ
fiducial arrangement of momenta. The mth element of K P
in this conventional ordering will be denoted by
ðsÞ
B. Irreducible representations
½n1 ; …; nN m so that the set can be equivalently written as
The irreps Γ of Oh and the irreps of its subgroups have
ðsÞ ðsÞ ðsÞ been previously cataloged in many places. Following
KP ¼ f½n1 ; …; nN m ∶ m ∈ f1; 2; …; jK P jgg: ð12Þ
Ref. [9], such representations are completely specified
ðΓÞ
The partitioning into momentum orbits gives rise to a by basis vectors jBμ i, where μ ∈ f1; …; jΓjg labels the
corresponding decomposition of the space of interpolating rows of the irrep Γ. The basis vectors can be written in an
operators. For convenience, the plane-wave vector in V abstract coordinate space as basis functions hrjBμ i ¼
ðΓÞ
associated with the mth list of wave vectors within the orbit ðΓÞ
Bμ ðrÞ on the unit sphere, in terms of the normalized
labeled by s is denoted by
three-vector coordinate r ≡ ðx; y; zÞT ∈ R3 , with jrj ¼ 1.
ðsÞ The basis functions used in this work, for the irreps of Oh
js; mi ≡ j½n1 ; …; nN m i: ð13Þ
and its subgroups, are specified in Table I. Additional
ðsÞ details related to the choice of basis functions are described
It bears emphasis that each momentum orbit K P is defined
in Appendix A.
in terms of a particular total momentum P with little group
For the present work, the primary utility of the basis
GP ; this dependence is left implicit in the notation js; mi.
functions is in defining the irrep matrices DðΓÞ ðRÞ appear-
Each orbit defines an invariant subspace
ing in the block diagonalization of Eq. (1). For instance,
ðsÞ ðsÞ
given the basis functions, the matrix representation asso-
V P ≡ spanfjs; mi∶m ∈ f1; 2; …; jK P jgg ⊂ V ð14Þ ciated with R ∈ GP in irrep Γ can be computed as
satisfying ðΓÞ 1 ðΓÞ ðΓÞ

Dμ0 μ ðRÞ ¼ hB 0 jDðRÞjBμ i
N ðΓÞ Z μ
ðsÞ
DðRÞV P ¼ V P ;
ðsÞ
for R ∈ GP : ð15Þ 1 ðΓÞ ðΓÞ
¼ ðΓÞ dΩhBμ0 jrihrjDðRÞjBμ i
N
Z
1 ðΓÞ ðΓÞ
Through the present work, jXj indicates the number of ¼ ðΓÞ dΩBμ0 ðrÞBμ ðR−1 rÞ; ð19Þ
4
elements in the set X. N
094516-4
ðΓÞ ðΓÞ
TABLE I. Basis functions hrjBμ i ¼ Bμ ðrÞ used in this work for the irreps of the cubic group Oh and its
subgroups. As indicated, basis functions for the irreps of the subgroups C4v , C3v , C2v , CR2 , CP2 , and C1 can be written
in terms of the basis functions of irreps of Oh . The final column indicates the z component of the SOð3Þ angular
momentum operator modulo 4, denoted by lz.
ðΓÞ
Group Irrep μ Basis function Bμ ðrÞ Notes
Oh Aþ
1 1 p1ffiffi ðx2 þ y2
3
þz Þ 2 lz ¼ 0
Oh Aþ
2 1 p1ffiffi ½x4 ðy2
6
2 4 2 2
− z Þ þ y ðz − x Þ þ z ðx − y Þ 4 2 2 lz ¼ 2
Oh þ 1 p1ffiffi ð2z2 − x2 − y2 Þ lz ¼ 0
E
6
Oh þ 2 p1ffiffi ðx2 − y2 Þ lz ¼ 2
E
2
Oh Tþ
1 1 p1ffiffi xyðx2 − y2 Þ
2
lz ¼ 0
Oh Tþ
1 2 1 2 2
2 ½−yzðy − z Þ þ izxðz
2
− x2 Þ lz ¼ 1
Oh Tþ
1 3 1 2 2
2 ½yzðy − z Þ þ izxðz
2 − x2 Þ lz ¼ 3
Oh Tþ
2 1 p1ffiffi ð−zx þ iyzÞ
2
lz ¼ 1
Oh Tþ
2 2 −ixy lz ¼ 2
Oh Tþ
2 3 p1ffiffi ðzx þ iyzÞ
2
lz ¼ 3
Oh A−1 1 p1ffiffi xyz½x4 ðy2 − z Þ þ y ðz − x Þ þ z4 ðx2 − y2 Þ
2 4 2 2 lz ¼ 0
6
Oh A−2 1 xyz lz ¼ 2
Oh E− 1 p1ffiffi xyzðx2 − y2 Þ lz ¼ 0
2
Oh E− 2 − p1ffiffi xyz½2z2 − x2 − y2 lz ¼ 2
6
Oh T −1 1 z lz ¼ 0
Oh T −1 2 p1ffiffi ð−x þ iyÞ lz ¼ 1
2
Oh T −1 3 p1ffiffi ðx þ iyÞ lz ¼ 3
2
Oh T −2 1 1 2 2 2
2 ½−yðz − x Þ þ ixðy − z Þ
2 lz ¼ 1
Oh T −2 2 − piffiffi2 zðx2 − y2 Þ lz ¼ 2
Oh T −2 3 1 2 2 2
2 ½yðz − x Þ þ ixðy − z Þ
2 lz ¼ 3
C4v A1 1 ðOh ;Aþ Þ
B1 1
ðrÞ
C4v A2 1 ðO ;A− Þ
B1 h 1 ðrÞ
C4v B1 1 ðO ;Eþ Þ
B2 h ðrÞ
C4v B2 1 ðO ;T þ Þ
B2 h 2 ðrÞ
C4v E 1 ðO ;T þ Þ ðO ;T þ Þ
−xz ¼ p1ffiffi ðB h 2 ðrÞ − B h 2 ðrÞÞ
2 1 3
C4v E 2 ðO ;T þ Þ ðO ;T þ Þ
−yz ¼ piffiffi ðB h 2 ðrÞ þ B h 2 ðrÞÞ
2 1 3

B1 1
ðrÞ
C3v A2 1 ðO ;Aþ Þ
B1 h 2 ðrÞ
C3v E 1 ðO ;Eþ Þ
B1 h ðrÞ
C3v E 2 ðO ;Eþ Þ
B2 h ðrÞ
B1 1
ðrÞ
C2v A2 1 ðO ;A− Þ
B1 h 1 ðrÞ
C2v B1 1 ðO ;A− Þ
B1 h 2 ðrÞ
C2v B2 1 ðO ;Aþ Þ
B1 h 2 ðrÞ
CR2 A 1 ðOh ;Aþ Þ
B1 1
ðrÞ
CR2 B 1 ðOh ;A−1 Þ
B1 ðrÞ
CP2 A 1 ðOh ;Aþ Þ
B1 1
ðrÞ
CP2 B 1 ðO ;Aþ Þ
B1 h 2 ðrÞ
C1 A 1 ðOh ;Aþ Þ
B1 1
ðrÞ
094516-5
where the integration is over the solid-angle measure on the To separate the ki degenerate copies of each irrep Γi , it is
unit sphere, i.e., dΩ ≡ drδðjrj − 1Þ, and the normalization helpful first to decompose each space V ðΓi ;sÞ into sectors
constant is associated with the individual rows μ of the irrep Γi as
ðΓ ;sÞ
Z V ðΓi ;sÞ ¼ ⨁μ V μ i . Any set of ki linearly independent
ðΓÞ ðΓÞ
N ðΓÞ ¼ dΩ Bμ ðrÞBμ ðrÞ; ð20Þ vectors in V μ
ðΓi ;sÞ
, which can be labeled as
with no summation over μ (by construction, all rows of a fjs; Γi ; κ; μi∶κ ∈ f1; …; ki gg; ð24Þ
given irrep are identically normalized so N ðΓÞ does not
carry a row index). The action of a little group trans- provide a basis for this space,
formation therefore corresponds to right multiplication of ðΓi ;sÞ
basis vectors, Vμ ¼ spanfjs; Γi ; κ; μi∶κ ∈ f1; …; ki gg: ð25Þ
ðΓÞ
X ðΓÞ ðΓÞ If a consistent set of ki basis vectors is chosen for all μ, the
DðRÞjBμ i ¼ jBμ0 iDμ0 μ ðRÞ: ð21Þ
transformation properties of the basis vectors follow from
μ0
Eq. (21),
In explicit calculations, it is convenient to recast the X ðΓ Þ
integral in Eq. (19) algebraically using polarization tensors DðRÞjs; Γi ; κ; μi ¼ js; Γi ; κ; μ0 iDμ0 μi ðRÞ; ð26Þ
as discussed in Appendix B. μ0
ðΓ Þ
where the irrep matrices Dμ0 μi ðRÞ for R ∈ GP are defined as
C. Block diagonalization
ðsÞ described in Sec. II B. The explicit construction of such a
Elements of the momentum orbit K P transform in the basis is detailed below. Note that the labeling of degenerate
reducible representation ΓðsÞ , which can be decomposed copies of irreps by κ is not unambiguously specified. A
into a direct sum of irreps particular choice will be made in the following construction.
The unitary matrix UðsÞ that transforms from the reduc-
ΓðsÞ ¼ Γ1 ⊕ … ⊕ Γ1 ⊕ Γ2 ⊕ … ⊕ Γ2 ⊕ : ð22Þ ible momentum-orbit basis fjs; mig to the irrep basis
|fflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflffl} |fflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflffl}
k1 copies k2 copies fjs; Γi ; κ; μig is defined by its elements
ðΓ ;κ;sÞ
Above, Γ1 ; Γ2 ; … label the distinct irreps of the group GP , Umμi ≡ hs; mjs; Γi ; κ; μi; ð27Þ
with each Γi appearing with multiplicity ki ≥ 0 in the
decomposition of the momentum-orbit representation. This where κ ∈ f1; …; ki g, μ ∈ f1; …; jΓi jg, and the total
decomposition into irreps induces an associated decom- momentum P is left implicit. In particular, these matrices
ðsÞ ðsÞ
position of the vector space V P associated with K P into enact the block diagonalization anticipated by Eq. (1),
subspaces that each transform according to a given irrep:
ðΓ Þ
X ðΓ ;κ;sÞ ðsÞ ðΓ ;κ;sÞ
Dμμi0 ðRÞ ¼ ½Umμi Dmm0 ðRÞUm0iμ0 ð28Þ
ðsÞ m;m0
VP ¼ V ðΓ1 ;sÞ ⊕ V ðΓ2 ;sÞ ⊕ ; ð23Þ
where dependence on P is left implicit on the right-hand ¼ ð½UðsÞ † DðsÞ ðRÞUðsÞ Þμμ0 : ð29Þ
side. Due to the potential for degenerate copies of irreps,5
Although the labels Γi and κ appear as superscripts labeling
each summand V ðΓi ;sÞ has dimension ki · jΓi j. Methods ðsÞ ðsÞ
for constructing operators that project this space into the choice of irrep, U ðsÞ is indeed a unitary jK P j × jK P j
particular irrep rows have been used in many previous matrix if m is taken as a row index and the indices Γi , κ, and
studies [10–17]. However, only cases where irreps have μ are enumerated jointly as a column index.
relatively small multiplicities (1–2) have been considered, The change-of-basis matrix UðsÞ for each momentum
and general methods for consistently decomposing spaces orbit can be constructed explicitly using Schur’s lemma,
where irreps have arbitrary multiplicities have not been which says6
discussed.
jΓj X ðΓÞ ðΓ0 Þ
Dμν ðRÞ Dμ0 ν0 ðRÞ ¼ δΓΓ0 δμμ0 δνν0 ð30Þ
5
Here and below, “degenerate” is used to describe situations jGP j R ∈ G
P
where multiple copies of an irrep appear in a given momentum-
orbit representation. In Hilbert space, the physical states asso-
6
ciated with these operators are not necessarily degenerate in the Written in this form, Schur’s lemma is sometimes referred to
sense of having the same energy. as the wonderful orthogonality theorem [9].
094516-6
for arbitrary irreps Γ and Γ0 . For each irrep Γi of the ðΓ ;sÞ jΓi j X ðΓi Þ
T μνi ≡ Dμν ðRÞ DðRÞ; ð37Þ
little group GP , it is convenient to define a projection jGP j R ∈ G
P
operator
ðΓ ;sÞ jΓi j X ðΓi Þ ðsÞ
ðΓi ;sÞ
Πμ
ðsÞ
∶ VP → Vμ
ðΓi ;sÞ
; ð31Þ ðT μνi Þm0 m ≡ Dμν ðRÞ Dm0 m ðRÞ; ð38Þ
jGP j R ∈ G
P
ðΓi ;sÞ jΓi j X ðΓi Þ

Πμ ≡ Dμμ ðRÞ DðRÞ; ð32Þ X
ki
jGP j R ∈ G ¼
ðΓ ;κ;sÞ
Um0iμ
ðΓ ;κ;sÞ
½Umνi : ð39Þ
P
κ¼1
with no sum on μ implied and with matrix elements
ðΓ ;sÞ ðΓ ;sÞ
By definition, T μμi ≡ Πμ i . The transition operators are
ðΓ ;sÞ 0 ðΓ ;sÞ
ðΠμ i Þm0 m ¼ hs; m jΠμ i js; mi ð33Þ ðΓ ;sÞ ðΓ ;sÞ
Hermitian, ½T μνi † ¼ T νμi , and satisfy
X ðΓ ;sÞ
jΓ j X ðΓi Þ ðsÞ ðΓ ;sÞ ðΓ ;sÞ
ðT μνi Þmm0 ðT ρσi Þm0 n ¼ δνρ ðT μσi Þmn ; ð40Þ
¼ i Dμμ ðRÞ Dm0 m ðRÞ: ð34Þ
jGP j R ∈ G m0
P
which follows from unitarity of the change-of-basis matri-

This projector onto a given row μ of the irrep Γi is closely ces. Likewise, it follows from Eq. (39) and unitarity of the
related to the desired change-of-basis matrices: change-of-basis matrices that the transition operators relate
the different rows μ and ν,
ðΓi ;sÞ
X
ki
ðΠμ Þm0 m ¼ hs; m0 js; Γi ; κ; μihs; Γi ; κ; μjs; mi ð35Þ ðΓ ;κ;sÞ
X ðΓ Þ;s ðΓ ;κ;sÞ
κ¼1 Umμi ¼ ðT μνi Þmn Unνi ; ð41Þ
n
X
ki
ðΓ ;κ;sÞ ðΓ ;κ;sÞ
¼ Um0iμ ½Umμi : ð36Þ for fixed μ, ν, which allows all elements of the change-of-
κ¼1 basis matrices to be determined once those with μ ¼ 1
are known.
The first equality follows from P Schur’s lemma after
insertions of the identity, 1 ¼ Γ;κ;μ js; Γ; κ; μihs; Γ; κ; μj.
D. Stabilizer subgroups
By unitarity of the change-of-basis matrices, the projection
ðΓ ;sÞ The block-diagonalization matrices for many different
operators Πμ i are idempotent.
ðΓ ;κ;sÞ momentum orbits can be demonstrated to be identical using
The change-of-basis matrix elements U mμi can be
general arguments from group theory. The construction and
extracted by suitable orthogonalization of the rows of ðsÞ
ðΓ ;sÞ ordering of a momentum orbit K P are always performed in
ðΠμ i Þm0 m . The choice of an orthogonalization scheme,
e.g., a Gram-Schmidt procedure, fixes the otherwise terms of a fiducial list of plane waves, with the associated
ambiguous κ labeling of degenerate copies of each irrep. set of wave vectors denoted by i ≡ ½n1 ; …; nN . The
Degenerate copies must be orthogonalized consistently stabilizer subgroup of the little group GP can then be
ðsÞ
for all rows μ within each irrep in order to achieve the defined as the subgroup of rotations HP ≤ GP that leave i
simple transformation rule in Eq. (26).7 A natural pre- invariant. For any particular i, the stabilizer is easily
scription is first to orthonormalize the μ ¼ 1 rows within identified by acting with all elements of the (finite) little
each irrep and then to use transition operators to move group GP . By construction, each stabilizer subgroup is one
between the remaining rows. Schur’s lemma applied to the of the finite groups listed after Eq. (10).
off-diagonal elements of the rotational irrep matrices The orbit-stabilizer theorem then implies that elements
furnishes the transition operators, of the momentum orbit are in one-to-one-correspondence
ðsÞ
with the left cosets of HP [19],
ðsÞ
K P ¼ fR · i∶R ∈ GP g
7
For example, applying the Gram–Schmidt procedure to each ðsÞ ðsÞ
ðΓ ;sÞ ↔ fRj HP ∶j ∈ f1; 2; …jK P jgg; ð42Þ
row of ðΠμ i Þm0 m independently would lead to a more cumber-
some transformation rule in which states js; Γ; κ; μi are mapped
by the action of R ∈ GP to linear combinations of states because the rotations in each coset map the fiducial arrange-
ðsÞ
js; Γ; κ 0 ; μ0 i including those with κ 0 ≠ κ. The coefficients of these ment i to a single, unique element in K P . In Eq. (42), the jth
linear combinations would need to be computed separately for ðsÞ ðsÞ
each momentum orbit and would generically depend upon s, κ, coset is defined as Rj HP ≡ fRj h∶h ∈ HP g where Rj ∈ GP
and κ0 as well as Γ, μ, and μ0 . is a representative element of the coset. Acting from the left
094516-7
with a group element R0 permutes the cosets on the right of problem. Solution of the two-operator problem is divided
Eq. (42) in the same way as the states on the left of Eq. (42). into two steps:
ðsÞ (1) Enumerate the little groups and stabilizer groups
Once HP is determined from i, the full structure of the
momentum orbit is encoded in the right-hand side and the associated with all possible two-body momenta.
details of the fiducial state i no longer matter. This means that (2) Compute the block-diagonalization matrices UðsÞ in
if two distinct momentum orbits, labeled by s and s0 , share the each case [cf. Eq. (1)].
ðsÞ ðs0 Þ The remainder of this section classifies possible two-body
same little group GP and stabilizer group HP ¼ HP , they momentum configurations and their associated little groups
transform identically under GP, i.e., and stabilizers. The results of this classification are summarized
in Table II. The block-diagonalization matrices UðsÞ can be
ðsÞ ðs0 Þ computed using the method described in Sec. II C. Subsets of
Dm0 m ðRÞ ¼ Dm0 m ðRÞ ð43Þ
these results have previously been presented in Refs. [13,14].
ðsÞ
for all m; m0 ∈ f1; …; jK P jg and each element R ∈ GP . As 1. Rest-frame systems
such, the block diagonalization to irreps of the cubic group
can also be achieved with the same set of matrices. In the rest frame, all seven conjugacy classes of little
ðsÞ
ðsÞ ðsÞ groups can act as stabilizer subgroups HP within the total
Conjugating H P by some element R ∈ GP , HP →
ðsÞ symmetry group Oh . The stabilizer subgroups and repre-
R−1 HP R, is equivalent to rotating the basis states to pick sentative choices of rest-frame momenta ½n or ½n1 ; n2
a different fiducial state in the same orbit, which amounts to corresponding to N ∈ f1; 2g distinguishable particles
reordering the basis states. This means that irrep decom- stabilized by the group are given by:
positions need only be performed for one subgroup within ðsÞ
(i) H P ¼ Oh : ½n ¼ ½ð0; 0; 0Þ
each conjugacy class of subgroups of GP . The block- ðsÞ
(ii) H P ¼ C4v : ½n1 ; n2 ¼ ½ð0; 0; nÞ; ð0; 0; −nÞ
diagonalization matrices for other subgroups in the same ðsÞ
class are related by reordering the columns. (iii) H P ¼ C3v : ½n1 ; n2 ¼ ½ðn; n; nÞ; ð−n; −n; −nÞ
ðsÞ
When all operators are distinguishable, the structure of (iv) H P ¼ C2v : ½n1 ; n2 ¼ ½ð0; n; nÞ; ð0; −n; −nÞ
ðsÞ ðsÞ
the stabilizer group H P is also severely restricted. The (v) H P ¼ CR2 : ½n1 ; n2 ¼ ½ð0; m; nÞ; ð0; −m; −nÞ
ðsÞ ðsÞ
stabilizer group HP associated with products of N dis- (vi) H P ¼ CP2 : ½n1 ; n2 ¼ ½ðm; n; nÞ; ð−m; −n; −nÞ
tinguishable operators must be the intersection of a set of ðsÞ
(vii) H P ¼ C1 : ½n1 ; n2 ¼ ½ðm; l; nÞ; ð−m; −l; −nÞ.
individual little groups GP1 ; …GPN ⊂ Oh , where each GPj Examples of explicit block-diagonalization matrices for
is the little group of the wave vector of the jth operator. rest-frame systems are tabulated in Appendix C, following
These intersections can be shown to give other—identical the conventions described in Appendix A.
or smaller—finite groups in every case. Cataloging all two-
operator cases, for which all of the finite groups listed after
Eq. (10) appear as stabilizer groups, thus already deter- 2. Boosted systems
mines the wave functions for all possible distinguishable- Boosted systems can be analyzed similarly to rest-frame
particle operators. For cases with N > 2 operators, iden- systems. First, one identifies for each little group GP a
ðsÞ representative total momentum P. Second, one identifies all
tifying the stabilizer group HP allows one to select the
ðsÞ
appropriate block-diagonalization matrices already con- (conjugacy classes of) subgroups H P of the little group GP
structed from specific examples in the two-operator case. which are compatible with stabilizing a set of plane waves.
The fact that two-operator cases already give rise to all Representative examples for all valid choices of GP and
possible change-of-basis matrices is specific to the case of ðsÞ
HP can be chosen as follows:
distinguishable spin-zero operators and does not hold in the (1) GP ¼ C4v : P ¼ 2πL ð0; 0; nÞ
cases of nonzero spin; however, analogous stabilizer group ðsÞ
(a) HP ¼ C4v : ½n ¼ ½ð0; 0; nÞ
considerations still restrict the distinct change-of-basis ðsÞ
matrices to a finite number once the spin and permutation (b) HP ¼ CR2 : ½n1 ; n2 ¼ ½ð0; m; nÞ; ð0; −m; 0Þ
properties of all operators are specified, as described in Note that CR2 appears in noncanonical form here
Secs. III and IV below. as the set of reflections of the x axis.
ðsÞ
(c) HP ¼ CP2 : ½n1 ; n2 ¼ ½ðm; m; nÞ; ð−m; −m; 0Þ
ðsÞ
(d) HP ¼ C1 : ½n1 ; n2 ¼ ½ðm; l; nÞ; ð−m; −l; 0Þ,
E. Complete classification for N distinguishable
(2) GP ¼ C3v : P ¼ 2πL ðn; n; nÞ
spin-zero particles ðsÞ
(a) HP ¼ C3v : ½n ¼ ½ðn; n; nÞ
As discussed in Sec. II D, for distinguishable spin-zero ðsÞ
particles, solving the block-diagonalization problem for (b) HP ¼ CP2 : ½n1 ; n2 ¼ ½ðn; n; 0Þ; ð0; 0; nÞ
ðsÞ
two operators in fact solves the generic N-operator (c) HP ¼ C1 : ½n1 ; n2 ¼ ½ðn þ m; n; 0Þ; ð−m; 0; nÞ,
094516-8
TABLE II. The complete solution to the N-body block diagonalization problem for distinguishable spin-zero operators. The solution
ðsÞ
follows from classifying the possible combinations of little groups GP and stabilizers HP arising in the two-body case. Example states
are denoted by jπðn1 Þ; Kðn2 Þi, as πK operator construction provides a simple example of two distinguishable, spin-zero operators.
ðsÞ
GP HP Example state Orbit dim Irrep decomposition
Oh Oh jπð0; 0; 0Þ; Kð0; 0; 0Þi 1 Aþ
1
Oh C4v jπð0; 0; 1Þ; Kð0; 0; −1Þi 6 Aþ
1 ⊕ Eþ ⊕ T −1
Oh C2v jπð0; 1; 1Þ; Kð0; −1; −1Þi 12 Aþ þ þ
1 ⊕ E ⊕ T2 ⊕ T1 ⊕ T2
− −
Oh CR2 jπð2; 1; 0Þ; Kð−2; −1; 0Þi 24 A1 ⊕ A2 ⊕ 2Eþ ⊕ T 1 ⊕ T þ

þ þ þ
2 ⊕ 2T − −
1 ⊕ 2T 2
Oh C3v jπð1; 1; 1Þ; Kð−1; −1; −1Þi 8 Aþ
1 ⊕ T þ
2 ⊕ A −
2 ⊕ T −
1
Oh CP2 jπð1; 1; 2Þ; Kð−1; −1; −2Þi 24 Aþ þ þ þ − −
1 ⊕ E ⊕ T 1 ⊕ 2T 2 ⊕ A2 ⊕ E ⊕ 2T 1 ⊕ T 2
− −
þ þ þ þ
Oh C1 jπð3; 2; 1Þ; Kð−3; −2; −1Þi 48 A1 ⊕ A2 ⊕ 2E ⊕ 3T 1 ⊕ 3T 2 ⊕ A1 ⊕ A2 ⊕ 2E ⊕ 3T −1 ⊕ 3T −2
þ − − −
C4v C4v jπð0; 0; 1Þ; Kð0; 0; 0Þi 1 A1

C4v CR2 jπð1; 0; 1Þ; Kð−1; 0; 0Þi 4 A 1 ⊕ B1 ⊕ E
C4v CP2 jπð1; 1; 1Þ; Kð−1; −1; 0Þi 4 A 1 ⊕ B2 ⊕ E
C4v C1 jπð2; 1; 1Þ; Kð−2; −1; 0Þi 8 A1 ⊕ A2 ⊕ B1 ⊕ B2 ⊕ 2E
C3v C3v jπð1; 1; 1Þ; Kð0; 0; 0Þi 1 A1
C3v CP2 jπð1; 1; 0Þ; Kð0; 0; 1Þi 3 A1 ⊕ E
C3v C1 jπð1; 0; −1Þ; Kð0; 1; 2Þi 6 A1 ⊕ A2 ⊕ 2E
C2v C2v jπð0; 1; 1Þ; Kð0; 0; 0Þi 1 A1
C2v CR2 jπð0; 0; 1Þ; Kð0; 1; 0Þi 2 A 1 ⊕ B2
C2v CP2 jπð2; 1; 1Þ; Kð−2; 0; 0Þi 2 A 1 ⊕ B1
C2v C1 jπð−2; 0; 1Þ; Kð2; 1; 0Þi 4 A1 ⊕ A2 ⊕ B1 ⊕ B2
CR2 CR2 jπð1; 2; 0Þ; Kð0; 0; 0Þi 1 A
CR2 C1 jπð1; 0; 1Þ; Kð0; 2; −1Þi 2 A⊕B
CP2 CP2 jπð1; 2; 2Þ; Kð0; 0; 0Þi 1 A
CP2 C1 jπð1; 0; 0Þ; Kð0; 1; 2Þi 2 A⊕B
C1 C1 jπð1; 2; 3Þ; Kð0; 0; 0Þi 1 A1
(3) GP ¼ C2v : P ¼ 2π
L ð0; n; nÞ rather than Oh . Consequently, the procedures discussed
ðsÞ
(a) HP ¼ C2v : ½n ¼ ½ð0; n; nÞ above must be generalized. The approach taken here is to
ðsÞ decompose operators into irreps of the cubic group (and its
(b) HP ¼ CR2 : ½n1 ; n2 ¼ ½ð0; 0; nÞ; ð0; n; 0Þ
ðsÞ double cover) using the projection method applied to the
(c) HP ¼ CP2 : ½n1 ; n2 ¼ ½ð−m; n; nÞ; ðm; 0; 0Þ full momentum-spin space.8
ðsÞ
(d) HP ¼ C1 : ½n1 ; n2 ¼ ½ð−m; 0; nÞ; ðm; n; 0Þ, Extended operators involving fields evaluated at multiple
(4) GP ¼ CR2 : P ¼ 2π
L ðn; m; 0Þ
lattice sites can also have nontrivial “internal” cubic
ðsÞ transformation properties in addition to the transformation
(a) HP ¼ CR2 : ½n1 ¼ ½ðn; m; 0Þ
ðsÞ of the coordinate xi of each operator. The same formalism
(b) HP ¼ C1 : ½n1 ; n2 ¼ ½ðn; 0; lÞ; ð0; m; −lÞ,
presented in this section for particles with spin can be
(5) GP ¼ CP2 : P ¼ 2π
L ðn; n; mÞ applied to the case of extended operators with such
ðsÞ
(a) HP ¼ CP2 : ½n1 ¼ ½ðn; n; mÞ properties by replacing spinor representation matrices with
ðsÞ
(b) HP ¼ C1 : ½n1 ; n2 ¼ ½ðn; 0; 0Þ; ð0; n; mÞ,
¼ C1 : P ¼ 2π
L ðn; m; pÞ
8
(6) GP An alternative approach for incorporating spin, not pursued in
ðsÞ this work, would treat the group representation as a tensor
(a) HP ¼ C1 : ½ðn; m; pÞ. product of the representation under transformation of the spatial
coordinates xi and the internal spin representations. The spatial
representation associated with N coordinates can be decomposed
III. OPERATORS WITH SPIN as above. The problem then reduces to decomposing tensor
products of the rotational-symmetry irreps with the spin repre-
This section extends the discussion of the previous sentations of each operator. Computation of the Clebsch–Gordan
section to operators with nonzero spin. Since operators coefficients required for this strategy is straightforward [20], and
with half-integer spin transform in representations of SUð2Þ this approach has been used in practice for constructing two-
(the double cover of SOð3Þ) instead of SOð3Þ in an nucleon operators [14]. However, a drawback is that the number
of tensor products grows rapidly with the number of operators
unbounded, continuous three-dimensional space, nonzero included. The number of terms in, and complexity of, the
spin introduces the complication that the relevant symmetry resulting block-diagonalization matrices also grow rapidly with
group in a cubic lattice is OD h (the double cover of Oh ), the number of operators.
094516-9
the appropriate representation matrices for such extended positive-parity irrep Gþ 1 and a two-dimensional negative-
operators. − D
parity irrep G1 of Oh . By convention, the Dirac spinor
The remainder of this section is organized as follows. representation is defined in the parity eigenbasis, known as the
Section III A discusses the irreps of the double-cover group Dirac-Pauli basis, with explicit basis states given by
ODh . Section III B discusses the transformation properties of
typical spinor operators. Section III C describes the con- j1=2; þ1=2; þi ¼ ð 1 0 0 0 ÞT ;
struction of representation matrices associated with
momentum-spin orbits. Section III D presents several j1=2; −1=2; þi ¼ ð 0 1 0 0 ÞT ;
examples of irrep decompositions for distinguishable oper- j1=2; −1=2; −i ¼ ð 0 0 1 0 ÞT ;
ators with nonzero spin.
j1=2; þ1=2; −i ¼ ð 0 0 0 1 ÞT : ð44Þ
A. Double-cover irreps and basis vectors where the basis states are labeled as jJ; Jz ; i in terms of
For bosonic irreps, the irreps of Oh immediately furnish their eigenvalues of total spin J, the spin z-projection J z ,
irreps of ODh (see Appendix A for a concrete specification and parity. These basic irreps can be used to construct the
of ODh ). The key observation is that 2π rotations act trivially full set of fermionic irreps. Concrete basis vectors for the
on states in bosonic irreps. Therefore, the ODh -irrep matrices
irreps of OD h and relevant subgroups are given in terms of
for group elements differing by rotations of 2π can be these spin–1=2 and higher-spin basis states in Table III.
identified with the relevant Oh -irrep matrix. Higher-spin vectors appearing in Table III are constructed
For fermionic irreps, note first that the Dirac spinor in the usual way, e.g., j32 ; 32 ; i follows from the tensor
representation used to define the group OD h is a reducible product of three spin-12 vectors. Additional details related to
representation that can be decomposed into a two-dimensional the choice of basis vectors are given in Appendix A.
TABLE III. Basis vectors used in this work for the fermionic irreps of the double-cover group OD h and its
subgroups. As discussed in the main text, bosonic irreps of OD h follow immediately from those of O h given in
Table I. The basis vectors are identical to those used in Ref. [11], and they lead to identical representation matrices to
those in Ref. [15] for all fermionic irreps.
Group Irrep μ Basis vector

OD
h G
1 1 j1=2; þ1=2; i
OD
h G
1 2 j1=2; −1=2; i
OD
h H 1 j3=2; þ3=2; i
OD
h H 2 j3=2; þ1=2; i
OD
h H 3 j3=2; −1=2; i
OD H 4 j3=2; −3=2; i
h pffiffiffiffiffiffiffiffi pffiffiffiffiffiffiffiffi
OD G
2 1 1=6j5=2; −5=2; i − 5=6j5=2; þ3=2; i
h
pffiffiffiffiffiffiffiffi pffiffiffiffiffiffiffiffi
OD
h G
2 2 1=6j5=2; þ5=2; i − 5=6j5=2; −3=2; i
Dic4 G1 1 j1=2; þ1=2; þi
Dic4 G1 2 pffiffiffiffiffiffiffiffi j1=2; −1=2; þiffi
pffiffiffiffiffiffiffi
Dic4 G2 1 1=6j5=2; −5=2; þi − 5=6j5=2; þ3=2; þi
pffiffiffiffiffiffiffiffi pffiffiffiffiffiffiffiffi
Dic4 G2 2 1=6j5=2; þ5=2; þi − 5=6j5=2; −3=2; þi
Dic3 G 1 j1=2; þ1=2; þi
Dic3 G 2 j1=2; −1=2; þi
Dic3 F1 1 Eq. (A12)
Dic3 F2 2 Eq. (A13)
Dic2 G 1 j1=2; þ1=2; þi
Dic2 G 2 j1=2; −1=2; þi
CR4 F1 1 j1=2; þ1=2; þi
CR4 F2 1 j1=2; −1=2; þi
pffiffiffiffiffiffiffiffi
CP4 F1 1 1=2j1=2; 1=2; þi þ ð1 − iÞ=2j1=2; −1=2; þi
pffiffiffiffiffiffiffiffi
CP4 F2 1 1=2j1=2; 1=2; þi − ð1 − iÞ=2j1=2; −1=2; þi
CD
1 F 1 j1=2; 1=2; þi
094516-10
B. Operator representations of the little group instead of those defined by continuum

Having defined the double-cover group structure and spin J. Though most of the OD h irreps coincide with the
irreps for OD jJ; Jz ; i basis states, for large-J states and extended
h and its little groups, it is useful to record the
transformation properties of typical spinor interpolating operators, the OD h representation is typically reducible.
operators.
Creation and annihilation operators for spin-1=2 par- C. Momentum-spin orbits and their representations
ticles, denoted ψ̄ðxÞ and ψðxÞ respectively, transform as the For operators with spin, the notion of a momentum orbit
spinor representation and its conjugate: from Sec. II A is extended to a combined momentum-spin
X orbit. Because the spatial coordinates and spin degrees of
ÛðRD Þψ̄ðxÞα Û † ðRD Þ ¼ ψ̄ β ðRxÞSβα ðRD Þ; freedom transform in distinct spaces, the combined orbit
β belongs to their tensor product. The spin portion of the orbit
X
†
ÛðR ÞψðxÞα Û ðR Þ ¼
D
ψ β ðRxÞSβα ðRD Þ :
D
ð45Þ can itself be understood as the tensor product of the
β
individual spin-J1 through spin-JN basis states. In other
words, the extended momentum-spin orbit can be written as
Here ÛðRD Þ indicates the quantum operator that implements ðsÞ
K̃ P ≡ fRD · ½n1 ; α1 …; nN ;αN ∶RD ∈ GD
P gg
the ODh transformation R and SðR Þ is the spinor repre-
D D
sentation matrix associated with the group element RD , ðsÞ ðsÞ

¼ f½n1 ;α1 …;nN ; αN m ∶ m∈f1; 2; …; jK̃ P jgg; ð48Þ
SðRD Þ ≡ RD ∈ OD
h; ð46Þ where the αi contain the spin and parity quantum numbers.
As above, elements within this orbit are indexed by
since this is the defining representation. the integer label m. The value for αi of the mth element
Since products of ψ̄ α ðxÞ operators act on the vacuum to ½m
in the orbit will be denoted αi . The dimension of the
create single- and multiparticle states, the same trans- ðsÞ
formation rule for ψ̄ α ðxÞ operators is chosen as for irrep momentum-spin orbit, jK̃ p j, is given by the product of the
basis vectors in Eq. (21). Spinor operators with larger spin dimension of the momentum orbit with the dimensions of
(e.g., spin 3=2) can be constructed using tensor products of the individual spin representations jJi j ¼ 2J i þ 1, i.e.,
spin-1=2 operators.
ðsÞ ðsÞ
Y
N
With these conventions, one readily confirms the jK̃ p j ¼ jK p j × jJi j: ð49Þ
invariance of the usual kinetic terms of the Hamiltonian
R 3 i¼1
for
P free relativistic spin-1=2 fermions, H ψ ¼ d x×
j ψ̄ðxÞγ ∂j ψðxÞ. This invariance also applies on a dis-
j Once the momentum-spin orbit has been constructed, the
ðsÞ
crete spatial lattice under the subgroup Oh of all rotations, representation matrices D̃m0 m ðRD Þ follow using the analog
meaning these conventions are compatible with familiar of Eq. (18) as matrix elements, labeled by m0 and m,
Hamiltonians (or actions) appearing in practical calcula- between states of the momentum-spin operators. For the
tions in lattice gauge theory. case of N spin-1=2 operators, these matrix elements are
These transformation properties lead to a natural gener- explicitly given by
alization of the space of operators introduced in Sec. II.
Operators with spin can be specified in various ways; the ðsÞ
D̃m0 m ðRD Þ ≡ h½n1 ; α1 ; …; nN ; αN m0
generalized algorithm defined in the following is insensi-
tive to this choice. Particularly for products of local × jDðRD Þj½n1 ; α1 ; …; nN ; αN m i:
operators, one useful representation arises from construct- ¼ h½n1 ; …; nN m0 jDðRÞj½n1 ; …; nN m i
ing operators with definite total spin J under SUð2Þ. In this ½m0 ½m ½m0 ½m
case, besides wave numbers ni specifying the momenta, × hα1 jDðRD Þjα1 i…hαN jDðRD ÞjαN i
the extended space is defined by the total spin J, spin ðsÞ
¼ Dm0 m ðRÞSα½m0 α½m ðRD Þ Sα½m0 α½m ðRD Þ; ð50Þ
component Jz , and intrinsic parity . Given the operator 1 1 N N
transformation rules above, these states transform as

where SðRD Þ denotes the spinor representation matrices
X
J defined in Eq. (46) and R denotes the restriction of RD to
½J
DðRÞjn; J; Jz ; i ¼ jRn; J; J 0z ; iDJ0z ;Jz ðRD Þ; ð47Þ Oh . Expressions involving a mixture of spin-0 and spin-1=2
J0z ¼−J operators are obtained simply by removing spin labels αi
from spin-0 states and removing the corresponding trans-
where D½J ðRD Þ is the appropriate representation matrix formation factors of SðRD Þ.
for the spin-J operator. In general, one could choose For operators (e.g., spin-J operators or operators describ-
operators in any reducible or irreducible representation ing spatially extended objects) in a generic representation Γ̃
094516-11
of the little group, the associated representation matrices can be tabulated analogously to the case of N spin-zero
½Γ̃ plane-wave operators. This section collects several
are denoted Dαβ ðRD Þ, with α; β ∈ f1; 2; …jΓ̃jg. Given these
explicit examples of phenomenologically relevant systems
representations for the individual operator transformations,
described by products of operators with nonzero spin and
products of N such plane-wave operators transform with
presents the irrep decompositions that arise during their
ðsÞ ðsÞ ½Γ̃1 ½Γ̃N block diagonalization. Explicit change-of-basis matrices
D̃m0 m ðRD Þ ¼ Dm0 m ðRÞD ½m0 ½m
ðRD Þ D ½m0 ½m
ðRD Þ: ð51Þ are constructed in the accompanying code package [18].
α1 α1 αN αN
ðsÞ
Once the representation matrices D̃m0 m ðRD Þ are deter- 1. np
mined, precisely the same steps described in Sec. II are
used to determine block-diagonalization matrices of The np system provides an example of a system with the
ðΓÞ quantum numbers of two distinguishable spin-12 particles.
Eq. (1): project against the irrep matrices Dμ0 μ ðRD Þ using
The cubic irreps of the n and p operators both correspond
Schur’s lemma, orthogonalize degenerate irreps, and fill out
ðΓ;sÞ
to Gþ 1 . Consideration of isospin will be deferred until
the remaining rows using the transition operators T μν . discussion of internal symmetry groups in Sec. IV.
Table IV classifies the different orbit patterns in terms of
D. Examples little groups and stabilizer groups and shows the associated
Unlike the spin-zero case, the irrep decompositions for irrep decompositions of np, extending the results presented
products of N plane-wave operators with nonzero spin in Ref. [14]. Several general features arising in decom-
depend on the spins of the operators and cannot be positions for operators with nonzero spin are illustrated by
cataloged completely in terms of two-body results. For a this example. In particular, the orbit dimensions are four
given set of spins, however, the irrep decompositions that times larger than in the spin-zero case because the two-
can arise are severely restricted and can be fully classified nucleon spin space ð12 ⊗ 12 ¼ 0 ⊕ 1Þ has dimension four.
DðsÞ
for each of the little groups GDP and stabilizer groups H P ,
Different patterns of irreps arise for np than for the case of
DðsÞ distinguishable spin-zero particles.
where GD P ≤ Oh and H P ≤ GD
D
P indicate the little group
and stabilizer group constructed from the momenta as
before, now within the double cover group OD h . The 2. pπ +
DðsÞ
D
possible GP and HP can then be enumerated, and their The pπ þ system is an example of a fermionic system
irrep decompositions and block-diagonalization matrices with distinguishable operators transforming in different
TABLE IV. Combinations of irreps arising in decompositions of np operator orbits. Details are as in Table II.
ðsÞ
Oh Oh jnð0; 0; 0Þ; pð0; 0; 0Þi 4 Aþ
1 ⊕ T1
þ
þ þ þ
Oh C4v jnð0; 0; 1Þ; pð0; 0; −1Þi 24 A1 ⊕ E ⊕ 2T 1 ⊕ T 2 ⊕ A−1 ⊕ E− ⊕ 2T −1 ⊕ T −2
þ
Oh C2v jnð0; 1; 1Þ; pð0; −1; −1Þi 48 A1 ⊕ A2 ⊕ 2Eþ ⊕ 3T þ

þ þ þ − − − −
1 ⊕ 3T 2 ⊕ A1 ⊕ A2 ⊕ 2E ⊕ 3T 1 ⊕ 3T 2
−
jnð1; 1; 1Þ; pð−1; −1; −1Þi þ þ þ þ þ − − − − −

Oh C3v 32 A1 ⊕ A2 ⊕ E ⊕ 2T 1 ⊕ 2T 2 ⊕ A1 ⊕ A2 ⊕ E ⊕ 2T 1 ⊕ 2T 2
Oh CR2 jnð2; 1; 0Þ; pð−2; −1; 0Þi 96 2Aþ þ þ þ þ − − − −
1 ⊕ 2A2 ⊕ 4E ⊕ 6T 1 ⊕ 6T 2 ⊕ 2A1 ⊕ 2A2 ⊕ 4E ⊕ 6T 1 ⊕ 6T 2
−
þ þ þ þ
Oh CP2 jnð2; 1; 1Þ; pð−2; −1; −1Þi 96 2A1 ⊕ 2A2 ⊕ 4E ⊕ 6T 1 ⊕ 6T 2 ⊕ 2A1 ⊕ 2A2 ⊕ 4E ⊕ 6T 1 ⊕ 6T −2
þ − − − −
Oh C1 jnð3; 2; 1Þ; pð−3; −2; −1Þi 192 4Aþ þ þ þ þ − − − −

1 ⊕ 4A2 ⊕ 8E ⊕ 12T 1 ⊕ 12T 2 ⊕ 4A1 ⊕ 4A2 ⊕ 8E ⊕ 12T 1 ⊕ 12T 2
−
C4v C4v jnð0; 0; 1Þ; pð0; 0; 0Þi 4 A1 ⊕ A2 ⊕ E

C4v CR2 jnð1; 0; 1Þ; pð−1; 0; 0Þi 16 2A1 ⊕ 2A2 ⊕ 2B1 ⊕ 2B2 ⊕ 4E
C4v CP2 jnð1; 1; 1Þ; pð−1; −1; 0Þi 16 2A1 ⊕ 2A2 ⊕ 2B1 ⊕ 2B2 ⊕ 4E
C4v C1 jnð2; 1; 1Þ; pð−2; −1; 0Þi 32 4A1 ⊕ 4A2 ⊕ 4B1 ⊕ 4B2 ⊕ 8E
C3v C3v jnð1; 1; 1Þ; pð0; 0; 0Þi 4 A1 ⊕ A2 ⊕ E
C3v CP2 jnð1; 1; 0Þ; pð0; 0; 1Þi 12 2A1 ⊕ 2A2 ⊕ 4E
C3v C1 jnð1; 0; −1Þ; pð0; 1; 2Þi 24 4A1 ⊕ 4A2 ⊕ 8E
C2v C2v jnð0; 1; 1Þ; pð0; 0; 0Þi 4 A 1 ⊕ A 2 ⊕ B1 ⊕ B2
C2v CR2 jnð0; 0; 1Þ; pð0; 1; 0Þi 8 2A1 ⊕ 2A2 ⊕ 2B1 ⊕ 2B2
C2v CP2 jnð2; 1; 1Þ; pð−2; 0; 0Þi 8 2A1 ⊕ 2A2 ⊕ 2B1 ⊕ 2B2
C2v C1 jnð−2; 0; 1Þ; pð2; 1; 0Þi 16 4A1 ⊕ 4A2 ⊕ 4B1 ⊕ 4B2
CR2 CR2 jnð1; 2; 0Þ; pð0; 0; 0Þi 4 2A ⊕ 2B
CR2 C1 jnð1; 0; 1Þ; pð0; 2; −1Þi 8 4A ⊕ 4B
C1 C1 jnð1; 2; 3Þ; pð0; 0; 0Þi 4 4A
094516-12
TABLE V. Combinations of irreps arising in decompositions of pπ þ operator orbits. Details are as in Table II.
ðsÞ
Oh Oh jpð0; 0; 0Þ; π þ ð0; 0; 0Þi 2 G−1
þ
Oh C4v jpð0; 0; 1Þ; π ð0; 0; −1Þi 12 Gþ
1 ⊕H
þ ⊕ G−1 ⊕ H−
Oh C2v jpð0; 1; 1Þ; π þ ð0; −1; −1Þi 24 þ þ
G1 ⊕ G2 ⊕ 2H þ ⊕ G−1 ⊕ G−2 ⊕ 2H−
Oh C3v jpð1; 1; 1Þ; π þ ð−1; −1; −1Þi 16 Gþ þ
1 ⊕ G2 ⊕ H
þ
⊕ G−1 ⊕ G−2 ⊕ H−
Oh CR2 jpð2; 1; 0Þ; π þ ð−2; −1; 0Þi 48 þ þ
2G1 ⊕ 2G2 ⊕ 4H þ ⊕ 2G−1 ⊕ 2G−2 ⊕ 4H−
Oh CP2 jpð2; 1; 1Þ; π þ ð−2; −1; −1Þi 48 2Gþ þ
1 ⊕ 2G2 ⊕ 4H
þ
⊕ 2G−1 ⊕ 2G−2 ⊕ 4H−
Oh C1 jpð3; 2; 1Þ; π þ ð−3; −2; −1Þi 96 þ þ
4G1 ⊕ 4G2 ⊕ 8H þ ⊕ 4G−1 ⊕ 4G−2 ⊕ 8H−
C4v C4v jpð0; 0; 1Þ; π þ ð0; 0; 0Þi 2 G1
C4v CR2 jpð1; 0; 1Þ; π þ ð−1; 0; 0Þi 8 2G1 ⊕ 2G2
C4v CP2 jpð1; 1; 1Þ; π þ ð−1; −1; 0Þi 8 2G1 ⊕ 2G2
C4v C1 jpð2; 1; 1Þ; π þ ð−2; −1; 0Þi 16 4G1 ⊕ 4G2
C3v C3v jpð1; 1; 1Þ; π þ ð0; 0; 0Þi 2 G
C3v CP2 jpð1; 1; 0Þ; π þ ð0; 0; 1Þi 6 F1 ⊕ F2 ⊕ 2G
C3v C1 jpð1; 0; −1Þ; π þ ð0; 1; 2Þi 12 2F1 ⊕ 2F2 ⊕ 4G
C2v C2v jpð0; 1; 1Þ; π þ ð0; 0; 0Þi 2 G
C2v CR2 jpð0; 0; 1Þ; π þ ð0; 1; 0Þi 4 2G
C2v CP2 jpð2; 1; 1Þ; π þ ð−2; 0; 0Þi 4 2G
C2v C1 jpð−2; 0; 1Þ; π þ ð2; 1; 0Þi 8 4G
CR2 CR2 jpð1; 2; 0Þ; π þ ð0; 0; 0Þi 2 F1 ⊕ F2
CR2 C1 jpð1; 0; 1Þ; π þ ð0; 2; −1Þi 4 2F1 ⊕ 2F2
C1 C1 jpð1; 2; 3Þ; π þ ð0; 0; 0Þi 2 2F
TABLE VI. Combinations of irreps arising in decompositions of pπ þ π 0 operator orbits. Details are as in Table II.
ðsÞ
Oh Oh jpð0; 0; 0Þ; π þ ð0; 0; 0Þ; π 0 ð0; 0; 0Þi 2 Gþ1
Oh C4v jpð0; 0; 1Þ; π þ ð0; 0; −1Þ; π 0 ð0; 0; 0Þi 12 Gþ 1 ⊕H
þ
⊕ G−1 ⊕ H−
þ þ
Oh C2v jpð0; 1; 1Þ; π þ ð0; −1; −1Þ; π 0 ð0; 0; 0Þi 24 G1 ⊕ G2 ⊕ 2H þ ⊕ G−1 ⊕ G−2 ⊕ 2H−
Oh C3v jpð1; 1; 1Þ; π þ ð−1; −1; −1Þ; π 0 ð0; 0; 0Þi 16 Gþ þ
1 ⊕ G2 ⊕ H
þ ⊕ G−1 ⊕ G−2 ⊕ H−
þ þ
Oh CR2 jpð2; 1; 0Þ; π þ ð−2; −1; 0Þ; π 0 ð0; 0; 0Þi 48 2G1 ⊕ 2G2 ⊕ 4H þ ⊕ 2G−1 ⊕ 2G−2 ⊕ 4H−
Oh CP2 jpð2; 1; 1Þ; π þ ð−2; −1; −1Þ; π 0 ð0; 0; 0Þi 48 2Gþ þ
1 ⊕ 2G2 ⊕ 4H
þ ⊕ 2G−1 ⊕ 2G−2 ⊕ 4H−
Oh C1 jpð3; 2; 1Þ; π þ ð−3; −2; −1Þ; π 0 ð0; 0; 0Þi 96 4Gþ
1 ⊕ 4G þ
2 ⊕ 8H þ
⊕ 4G−1 ⊕ 4G−2 ⊕ 8H−
C4v C4v jpð0; 0; 1Þ; π þ ð0; 0; 0Þ; π 0 ð0; 0; 0Þi 2 G1
C4v CR2 jpð1; 0; 1Þ; π þ ð−1; 0; 0Þ; π 0 ð0; 0; 0Þi 8 2G1 ⊕ 2G2
C4v CP2 jpð1; 1; 1Þ; π þ ð−1; −1; 0Þ; π 0 ð0; 0; 0Þi 8 2G1 ⊕ 2G2
C4v C1 jpð2; 1; 1Þ; π þ ð−2; −1; 0Þ; π 0 ð0; 0; 0Þi 16 4G1 ⊕ 4G2
C3v C3v jpð1; 1; 1Þ; π þ ð0; 0; 0Þ; π 0 ð0; 0; 0Þi 2 G
C3v CP2 jpð1; 1; 0Þ; π þ ð0; 0; 1Þ; π 0 ð0; 0; 0Þi 6 F1 ⊕ F2 ⊕ 2G
C3v C1 jpð1; 0; −1Þ; π þ ð0; 1; 2Þ; π 0 ð0; 0; 0Þi 12 2F1 ⊕ 2F2 ⊕ 4G
C2v C2v jpð0; 1; 1Þ; π þ ð0; 0; 0Þ; π 0 ð0; 0; 0Þi 2 G
C2v CR2 jpð0; 0; 1Þ; π þ ð0; 1; 0Þ; π 0 ð0; 0; 0Þi 4 2G
C2v CP2 jpð2; 1; 1Þ; π þ ð−2; 0; 0Þ; π 0 ð0; 0; 0Þi 4 2G
C2v C1 jpð−2; 0; 1Þ; π þ ð2; 1; 0Þ; π 0 ð0; 0; 0Þi 8 4G
C2 C2 jpð1; 2; 0Þ; π þ ð0; 0; 0Þ; π 0 ð0; 0; 0Þi 2 F1 ⊕ F2
C2 C1 jpð1; 0; 1Þ; π þ ð0; 2; −1Þ; π 0 ð0; 0; 0Þi 4 2F1 ⊕ 2F2
C1 C1 jpð1; 2; 3Þ; π þ ð0; 0; 0Þ; π 0 ð0; 0; 0Þi 2 2F
irreps. The operator spins in this example correspond to the groups and stabilizer groups. This extends the results pre-
Gþ −
1 irrep for the proton and the A1 irrep for the pion. sented in Ref. [16]. The orbit dimensions are twice as large as
Table V shows the distinct irrep decompositions of pπ þ in the spin-zero particle case because of the nucleon spin and
orbits, which are again classified by the corresponding little can be decomposed into direct sums of fermionic irreps.
094516-13
3. pπ + π0 S ≡ SN 1 × SN 2 × ≤ SN ð52Þ
The pπ þ π 0 system provides an example with more than
two distinguishable operators and nonzero spin. States with corresponding to exchanges among subsets of identically
definite isospin (I ∈ f12 ; 32g) involve linear combinations labeled particles of size N 1 ; N 2 ; with N 1 þN 2 þ… ¼ N.
In this case, the categorization of rotational irreps can be
with nπ þ π þ operators and can be treated using the methods
given by the rotational representation and the individual
of Sec. IV below.
SN 1 ; SN 2 ; … irreps as ðΓ; λ1 ; λ2 ; …Þ.
Table VI shows the irrep decomposition of pπ þ π 0 orbits
classified by the corresponding little groups and stabilizer
groups. The patterns of little groups and stabilizer groups B. Extended orbits
occurring are identical to those in Table V. The only differ- Orbits constructed as in Secs. II A and III C may not
ence in the irreps appearing in the pπ þ and pπ þ π 0 decom- necessarily be closed under the exchange group S. To
positions is for the case of all operators at rest, which for pπ þ ensure states can always be constructed with definite
corresponds to G−1 but for pπ þ π 0 corresponds to Gþ 1.
permutation properties, the orbit must be extended to
include all states generated by applying permutations in
S. The labeling s of orbits and the indices m of basis states
IV. INTERNAL SYMMETRIES will continue to be used for these extended orbits, which
AND IDENTICAL PARTICLES ðsÞ
will be denoted as K̂ P :
In addition to the rotational transformation properties
ðsÞ
discussed so far, physical states also carry quantum K̂ P ≡ fg · ½n1 ; α1 ; ε1 ; …; nN ; αN ; εN ∶g ∈ GD
P × Sg
numbers such as charge and flavor. Moreover, for states ðsÞ
including identical particles, exchanging such particles ¼ f½n1 ; α1 ; ε1 ; …; nN ; αN ; εN m ∶
leaves the state unchanged up to a possible sign. More ðsÞ
m ∈ f1; 2; …; jK̂ P jgg; ð53Þ
precisely, as already exploited in Ref. [17], little-group
irreps must be paired appropriately with irreps under other
where αi label the spin and parity of the ith particle and
quantum numbers such that their combined transformations where εi labels its internal quantum numbers. Orbits then
under particle exchanges are symmetric (antisymmetric) furnish representations of exchange-group elements σ ∈ S
with respect to all possible exchanges of identical bosons in the usual way
(fermions), and are otherwise unconstrained. By identify-
ing the definite operator-exchange properties of cubic- ðsÞ
Dm0 m ðσÞ ¼ hs; m0 jDðσÞjs; mi: ð54Þ
group irreps, the framework described above can thus be
readily extended to be compatible with other quantum
Since σ corresponds to identical-operator exchange,
numbers. ðsÞ
Dm0 m ðσÞ ∈ f0; 1g.
A. Labeling by exchange-group irreps
C. Projection to exchange-group irreps
Since the same rotations are applied to all operators in an
Just as Schur’s lemma was applied in Eq. (36) to project
N-operator basis, rotation and permutation operations com-
into a basis with specific rotational-group irrep Γ, it can be
mute, and Schur–Weyl duality guarantees that the rotational
applied to project simultaneously into a basis with a
group irreps can be simultaneously labeled by specific irreps
specific permutation-group irrep. The representation theory
of the symmetric group SN [21]. This naturally divides any of the permutation group has been well studied (see
space that is closed under rotations and permutations into Ref. [23] for guidance on standard textbooks). Each irrep
blocks described by the pair ðΓ; λÞ of a rotational-group irrep of Sn is labeled by a Young diagram consisting of n boxes
Γ and an SN irrep λ. In the following, Young diagrams will be in left-justified rows with row lengths in nonincreasing
used to identify particular choices of λ [22]. Note that order. Closely related is the notion of a Young tableau, in
multiple blocks may have the same Γ or the same λ, i.e., there which the n boxes in a given Young diagram are filled
is no one-to-one correspondence between the rotational with the numbers f1; …; ng distributed such that each
irreps and permutation irreps. row and each column is strictly ascending. Let Y n denote
In some cases, it is not necessary to consider definite the set of Young tableaux with n boxes. For example,
representations under the full space of permutations. For
example, operators may be distinguishable by having .
different total isospin or other flavor quantum number. The connection to the internal symmetry group follows
To make this identification concrete and automatic, it is from the fact that products of operators can be taken to
assumed that the operators O1 ; …; ON can be respectively transform as (a row of) an irrep of the exchange group,
associated with internal labels ε1 ; …; εN . In this case, the labeled by a Young tableau Θ ∈ Y n . For each Young tableau
exchange group is taken to be the subgroup of permutations Θ, there exists a projection operator in the group algebra
094516-14
ΠΘ which projects onto the relevant row [23–25]. In fact,

mirroring the construction in Sec. II C, the group algebra ð60cÞ
can be decomposed completely into a basis of idempotents
ΠΘ and Hermitian (T †ΘΦ ¼ T ΦΘ ) transition operators [24]
T ΘΦ such that, for fixed Θ; Θ0 ; Φ; Φ0 ,
ΠΘ ≡ T ΘΘ ; ð55Þ ð60dÞ
T ΘΦ T Φ0 Θ0 ¼ δΦΦ0 T ΘΘ0 ; ð56Þ

X Here and below, cycle notation ði; j; k; …; zÞ is used to
ΠΘ ¼ 1: ð57Þ indicate permutations mapping the elements cyclically
Θ ∈ Yn
i → j → k… → z → i. The transition operators between
Eq. (56) is analogous to the product rule for the transition the two rows of the two-dimensional standard irrep are [24]
operators of the cubic group in Eq. (40).
When multiple instances of the same irrep row Θ appear
in the decomposition of a space acted on by the exchange
group, they can often be distinguished by letting T ΘΦ act on ð61aÞ
a fiducial vector for different Φ. All such products are left
invariant by ΠΘ. In this way, the transition operators acting
on fiducial vectors give a way to construct relevant multi-
particle operators. Concrete examples are discussed below.
ð61bÞ
For arbitrary Sn, recursive formulas for Hermitian
projection are given in Refs. [23–25] and for Hermitian
transition operators in Ref. [24]. Generic projection and where the normalization factors follow from Eq. (56).
transition operators are elements of the group algebra, The projection matrix acting on the extended orbit then
follows from linearity, with components:
X ðΘÞ
ΠΘ ¼ cσ σ; ð58Þ ðsÞ
X ðΘÞ ðsÞ
σ ∈ Sn Dm0 m ðΠΘ Þ ¼ cσ Dm0 m ðσÞ: ð62Þ
σ∈S
X ðΘΦÞ
T ΘΦ ¼ cσ σ; ð59Þ
σ ∈ Sn
The symmetric-group projection matrices are applied to the
orbit representation matrices in Eq. (51) to construct
ðΘΘÞ ðΘÞ ðΘΦÞ projected orbit representation matrices
with cσ ≡ cσ . As indicated, the coefficients cσ
depend explicitly on Θ; Φ ∈ Y n . A concrete example illus-
D̂ðs;ΘÞ ðRD Þ ≡ DðsÞ ðΠΘ Þ · D̃ðsÞ ðRD Þ · DðsÞ ðΠΘ Þ
trates the important features of the general case. Consider the
permutation group S3 , for which Table VII summarizes the ¼ D̃ðsÞ ðRD Þ · DðsÞ ðΠΘ Þ: ð63Þ
irreps. The projectors ΠΘ , with Θ ∈ Y 3 , are [24]
The equality in the second line follows from the fact that the
permutations commute with rotations and that DðsÞ ðΠΘ Þ is
ð60aÞ idempotent. Subsequent application of block diagonaliza-
tion using Schur’s lemma, orthogonalization, and rota-
tional transition operators—needed to construct the
block-diagonalization matrices in the analog of Eq. (1)—
remains unchanged.
ð60bÞ
D. Examples
TABLE VII. Irreps of the symmetric group S3 . 1. Identical fermions: nn and nnn
Name Dimension Young diagram Young tableaux

Identical fermions provide a first example with internal
symmetry. Operators constructed from products of identical
Trivial 1 fermions such as neutrons, e.g., nn or nnn, must be totally
Sign 1 antisymmetric under simultaneous exchange of the spins
Standard 2 , and momenta of any two particles. For nn, the exchange
group is S2 . Fermion antisymmetry implies nn operators
094516-15
transform in the sign irrep ( ) of S2 . Irrep decomposition little-group transformations affects the resulting irrep
proceeds as in the distinguishable np case summarized decomposition. These results illustrate the general fact that
Table IV, with the additional projection step of Eq. (63) the irrep decomposition of a multiparticle operator depends
ðsÞ
using the projection operator for the sign irrep of S2 , on the little group GP , the stabilizer group HP , and on the
irrep of the exchange group S. It is also noteworthy that in
ð64Þ this case the application of the exchange group projects
away the orbit with n1 ¼ n2 ¼ n3 ¼ 0; fermion antisym-
metry dictates that this orbit vanishes from the irrep
Table VIII summarizes the irrep decompositions for nn decomposition.
operators.
Operators constructed from products of more neutrons 2. Three pions with isospin
can be decomposed analogously. For nnn, the exchange
The three-pion system provides an example of the
group is S3 . Fermion antisymmetry requires that nnn
interplay between the cubic group and nontrivial inter-
operators transform in the sign irrep of S3 , corresponding
nal symmetries. Since each pion transforms as an isotrip-
to the projection operator in Eq. (60b). Table IX summa-
let with I ¼ 1, the three-pion system has the isospin
rizes the irrep decompositions for nnn operators. Note that,
decomposition
in this case, specification of the little group and stabilizer
group does not suffice to specify the irrep decomposition
uniquely—whether certain permutations correspond to 1 ⊗ 1 ⊗ 1 ¼ 3 ⊕ 2 ⊕ 2 ⊕ 1 ⊕ 1 ⊕ 1 ⊕ 0: ð65Þ
TABLE VIII. Combinations of irreps arising in decompositions of nn operator orbits. Details are as in Table II.
ðsÞ
GP HP S2 irrep Example state Orbit dim Irrep decomposition
Oh Oh jnð0; 0; 0Þ; nð0; 0; 0Þi 1 Aþ
1
þ
Oh C4v jnð0; 0; 1Þ; nð0; 0; −1Þi 12 A1 ⊕ E ⊕ A1 ⊕ E− ⊕ T −1 ⊕ T −2
þ −
Oh C2v jnð0; 1; 1Þ; nð0; −1; −1Þi 24 Aþ þ þ − − −

1 ⊕ E ⊕ T 2 ⊕ A1 ⊕ A2 ⊕ 2E ⊕ 2T 1 ⊕ 2T 2
− −
jnð1; 1; 1Þ; nð−1; −1; −1Þi þ þ − − − −

Oh C3v 16 A1 ⊕ T 2 ⊕ A1 ⊕ E ⊕ T 1 ⊕ 2T 2
Oh CR2 jnð2; 1; 0Þ; nð−2; −1; 0Þi 48 Aþ
1 ⊕ A þ
2 ⊕ 2E þ ⊕ T þ ⊕ T þ ⊕ 2A− ⊕ 2A− ⊕ 4E− ⊕ 4T − ⊕ 4T −
1 2 1 2 1 2
Oh CP2 jnð2; 1; 1Þ; nð−2; −1; −1Þi 48 Aþ þ þ þ − − −
1 ⊕ E ⊕ T 1 ⊕ 2T 2 ⊕ 2A1 ⊕ A2 ⊕ 3E ⊕ 4T 1 ⊕ 5T 2
− −
Oh C1 jnð3; 2; 1Þ; nð−3; −2; −1Þi 96 Aþ þ þ þ þ − − −

1 ⊕ A2 ⊕ 2E ⊕ 3T 1 ⊕ 3T 2 ⊕ 3A1 ⊕ 3A2 ⊕ 6E ⊕ 9T 1 ⊕ 9T 2
− −
C4v C4v jnð0; 0; 1Þ; nð0; 0; 1Þi 1 A1

C4v C4v jnð0; 0; 1Þ; nð0; 0; 0Þi 4 A1 ⊕ A2 ⊕ E
C4v CR2 jnð1; 0; 1Þ; nð−1; 0; 1Þi 8 2A1 ⊕ A2 ⊕ 2B1 ⊕ B2 ⊕ E
C4v CR2 jnð1; 0; 1Þ; nð−1; 0; 0Þi 16 2A1 ⊕ 2A2 ⊕ 2B1 ⊕ 2B2 ⊕ 4E
C4v CP2 jnð1; 1; 1Þ; nð−1; −1; 1Þi 8 2A1 ⊕ A2 ⊕ B1 ⊕ 2B2 ⊕ E
C4v CP2 jnð1; 1; 1Þ; nð−1; −1; 0Þi 16 2A1 ⊕ 2A2 ⊕ 2B1 ⊕ 2B2 ⊕ 4E
C4v C1 jnð2; 1; 1Þ; nð−2; −1; 1Þi 16 3A1 ⊕ 3A2 ⊕ 3B1 ⊕ 3B2 ⊕ 2E
C4v C1 jnð2; 1; 1Þ; nð−2; −1; 0Þi 32 4A1 ⊕ 4A2 ⊕ 4B1 ⊕ 4B2 ⊕ 8E
C3v C3v jnð1; 1; 1Þ; nð1; 1; 1Þi 1 A1
C3v C3v jnð1; 1; 1Þ; nð0; 0; 0Þi 4 A1 ⊕ A2 ⊕ E
C3v CP2 jnð1; 1; 0Þ; nð0; 0; 1Þi 12 2A1 ⊕ 2A2 ⊕ 4E
C3v C1 jnð1; 0; −1Þ; nð0; 1; 2Þi 24 4A1 ⊕ 4A2 ⊕ 8E
C2v C2v jnð0; 1; 1Þ; nð0; 1; 1Þi 1 A1
C2v C2v jnð0; 1; 1Þ; nð0; 0; 0Þi 4 A1 ⊕ A2 ⊕ B1 ⊕ B2
C2v CR2 jnð0; 0; 1Þ; nð0; 1; 0Þi 4 2A1 ⊕ A2 ⊕ B1
C2v C2v jnð0; 2; 2Þ; nð0; −1; −1Þi 4 A1 ⊕ A2 ⊕ B1 ⊕ B2
C2v CP2 jnð2; 1; 1Þ; nð−2; 1; 1Þi 4 2A1 ⊕ A2 ⊕ B2
C2v CP2 jnð2; 1; 1Þ; nð−2; 0; 0Þi 8 2A1 ⊕ 2A2 ⊕ 2B1 ⊕ 2B2
C2v C1 jnð−2; 0; 1Þ; nð2; 1; 0Þi 8 3A1 ⊕ 3A2 ⊕ B1 ⊕ B2
C2v C1 jnð1; 2; 1Þ; nð−1; −1; 0Þi 16 4A1 ⊕ 4A2 ⊕ 4B1 ⊕ 4B2
CR2 CR2 jnð1; 2; 0Þ; nð1; 2; 0Þi 1 A
CR2 CR2 jnð1; 2; 0Þ; nð0; 0; 0Þi 4 2A ⊕ 2B
CR2 C1 jnð1; 0; 1Þ; nð0; 2; −1Þi 8 4A ⊕ 4B
C1 C1 jnð1; 2; 3Þ; nð1; 2; 3Þi 1 A
C1 C1 jnð1; 2; 3Þ; nð0; 0; 0Þi 4 4A
094516-16
TABLE IX. Combinations of irreps arising in decompositions of nnn operator orbits. Details are as in Table II.
ðsÞ
GP HP S3 irrep Example state Orbit dim Irrep decomposition
Oh C4v jnð0; 0; 1Þ; nð0; 0; −1Þ; nð0; 0; 0Þi 24 Gþ
⊕ Hþ ⊕ 2G−1 ⊕ G−2 ⊕ 3H−
1
Oh C4v jnð0; 0; 1Þ; nð0; 0; 2Þ; nð0; 0; −3Þi 48 3G1 ⊕ Gþ
þ þ − −
2 ⊕ 4H ⊕ 3G1 ⊕ G2 ⊕ 4H
−
þ þ
Oh C2v jnð0; 1; 1Þ; nð0; −1; −1Þ; nð0; 0; 0Þi 48 G1 ⊕ G2 ⊕ 2H ⊕ 3G1 ⊕ 3G2 ⊕ 6H−
þ − −
Oh C2v jnð0; −1; −1Þ; nð0; −2; −2Þ; nð0; 3; 3Þi 96 4Gþ þ þ − −
1 ⊕ 4G2 ⊕ 8H ⊕ 4G1 ⊕ 4G2 ⊕ 8H
−
jnð1; 1; 1Þ; nð−1; −1; −1Þ; nð0; 0; 0Þi þ þ þ − − −

Oh C3v 32 G1 ⊕ G2 ⊕ H ⊕ 2G1 ⊕ 2G2 ⊕ 4H
Oh C3v jnð1; 1; 1Þ; nð2; 2; 2Þ; nð−3; −3; −3Þi 64 3Gþ þ þ − −
1 ⊕ 3G2 ⊕ 5H ⊕ 3G1 ⊕ 3G2 ⊕ 5H
−
þ þ
Oh CR2 jnð2; 1; 0Þ; nð−2; −1; 0Þ; nð0; 0; 0Þi 96 2G1 ⊕ 2G2 ⊕ 4H ⊕ 6G1 ⊕ 6G2 ⊕ 12H−
þ − −
Oh CR2 jnð0; 1; 1Þ; nð0; −1; 0Þ; nð0; 0; −1Þi 96 4Gþ þ þ − −

1 ⊕ 4G2 ⊕ 8H ⊕ 4G1 ⊕ 4G2 ⊕ 8H
−
þ þ
Oh CR2 jnð2; 1; 0Þ; nð−2; 0; 0Þ; nð0; −1; 0Þi 192 8G1 ⊕ 8G2 ⊕ 16H ⊕ 8G1 ⊕ 8G2 ⊕ 16H−
þ − −
Oh CP2 jnð2; 1; 1Þ; nð−2; −1; −1Þ; nð0; 0; 0Þi 96 2Gþ þ þ − −

1 ⊕ 2G2 ⊕ 4H ⊕ 6G1 ⊕ 6G2 ⊕ 12H
−
þ þ
Oh CP2 jnð1; 1; 1Þ; nð−1; −1; 0Þ; nð0; 0; −1Þi 192 8G1 ⊕ 8G2 ⊕ 16H ⊕ 8G1 ⊕ 8G2 ⊕ 16H−
þ − −
Oh C1 jnð0; −1; 1Þ; nð2; 1; 0Þ; nð−2; 0; −1Þi 192 8Gþ þ þ − −

1 ⊕ 8G2 ⊕ 16H ⊕ 8G1 ⊕ 8G2 ⊕ 16H
−
þ þ
Oh C1 jnð2; 1; 1Þ; nð−2; −1; 0Þ; nð0; 0; −1Þi 384 16G1 ⊕ 16G2 ⊕ 32H ⊕ 16G1 ⊕ 16G2 ⊕ 32H−
þ − −
The relevant exchange group is S3 , for which the irreps and Finally, the isosinglet transforms in the sign representation
projectors have been summarized above. As in Sec. II, of S3 , which is totally antisymmetric. Explicit expressions
Schur’s lemma provides the means to decompose the prod- for the associated states have been given in Ref. [17].
uct-state isospin representation into irreps of S3 . Since I z is Combining these results with block diagonalization for
conserved by permutations, it suffices to work at fixed I z . the cubic group amounts to applying Eq. (62), where ΠΘ is
Permutations of the states with a given I z furnish the selected to project onto the rows of Table X with the desired
ðI Þ isospin. Applying these projectors in Eq. (63) reproduces
reducible representation matrices Dm0zm ðσÞ. For example,
the seven states with I z ¼ 0 yield 7 × 7 representation the cubic-group irrep decompositions given in Appendix D
ðI ¼0Þ of Ref. [17] for the rest frame.
matrices Dm0zm ðσÞ. Irrep matrices DðλÞ for S3 are also readily As discussed in Sec. IV C, the transition operators T ΘΦ
obtained. For the trivial irrep they are simply unity, while for can be used to construct multiparticle operators with definite
the sign representation they are equal to the permutation Sn transformation properties. For the case at hand, consider
signature. Irrep matrices for the two-dimensional standard three-pion operators with I z ¼ 0 and I ∈ f0; 1; 2; 3g.
representation of S3 are given in Ref. [17]. These operators can be built from permutations of the
Table X shows the result of applying Schur’s lemma, i.e., fiducial ordering of single-pion operators, say, jπ þ π − π 0 i.
using Eq. (30) in terms of the DðλÞ and DðIz Þ matrices, to Letting permutations act in the natural way, e.g.,
extract the overlap of the reducible I z representation onto
ð1; 2; 3Þjπ þ π − π 0 i ¼ jπ 0 π þ π − i, gives the seven three-pion
the irrep λ. As expected, I ¼ 3 corresponds to the trivial
states with I z ¼ 0 constructed in Ref. [17]:
representation of S3 for each I z ∈ f−3; …; 3g, which is
totally symmetric. The doublets of states with I ¼ 2 for ð66aÞ
each I z fall in the standard representation of S3 . The three
copies with I ¼ 1 split into a doublet from the standard
representation and a trivial representation for each I z . ð66bÞ
TABLE X. Decomposition of the πππ system into irreps of S3 given in Table VII. For each column of fixed I z , the
number of check marks equals the sum of the dimensions in the irrep decomposition.
Iz
I þ3 þ2 þ1 0 −1 −2 −3 S3 irrep
3 ✓ ✓ ✓ ✓ ✓ ✓ ✓
2 ✓ ✓ ✓ ✓ ✓ o
2 ✓ ✓ ✓ ✓ ✓ o
1 ✓ ✓ ✓
1 ✓ ✓ ✓
1 ✓ ✓ ✓
0 ✓
094516-17
group irrep A1 . Both I ¼ 1 and I ¼ 2 may be combined

ð66cÞ with the cubic group irrep E, but the correct degenerate
copy of this irrep must be chosen.
ð66dÞ
3. DDπ
Decay channels with resonances are also categorized by
ð66eÞ isospin and provide examples of internal symmetry where
not all particles are identical. For instance, the doubly
charmed tetraquark T cc ð3875Þþ has been observed just
below threshold for D⋆þ D0 in the decay mode D0 D0 π þ
ð66fÞ
with charmness C ¼ þ2 and charge Q ¼ þ1, correspond-
ing to I z ¼ 0 [26]. The isospin decomposition of the DDπ
system is
ð66gÞ
1 1
⊗ ⊗ 1 ¼ 2 ⊕ 1 ⊕ 1 ⊕ 0; ð69Þ
2 2
The totally symmetric cases (Eqs. (66a) and (66b)) contain a
component proportional to jπ 0 π 0 π 0 i, which vanishes in all where the D-meson isodoublet is ðDþ ; D0 ÞT . Similar to the
other representations; the relative coefficient between the preceding example, for each fixed I z the direct sum is
jπ þ π − π 0 i and jπ 0 π 0 π 0 i terms is related to the isospin decomposed into irreps of S2 (permutations of the two
Clebsch–Gordan decomposition. As in Ref. [17], the two- D-meson operators) using projectors (defined in
dimensional standard irrep is spanned by a basis denoted by Eq. (64) above) and
jχ 1 i and jχ 2 i, with a final subscript outside the ket giving the
total isospin. As expected, Eqs. (66c) and (66d) are invariant ð70Þ
under Eq. (60c) and associated with degenerate copies of the
first row jχ 1 i of the irrep . Similarly, Eqs. (60e) and (66f)
are invariant under Eq. (60d) and associated with degenerate The four states with I z ¼ 0 are constructed from linear
copies of the second row jχ 2 i. combinations of the states jDþ Dþ π − i, jDþ D0 π þ i,
Three pions in boosted frames provide examples where jD0 Dþ π þ i, and jD0 D0 π þ i. In terms of fiducial orderings,
the full S3 exchange group influences the cubic-group irrep
decomposition. A minimal example is the system with jψ 00 i ≡ jD0 D0 π þ i;
momenta ½n1 ; n2 ; n3 ¼ ½ð1; 0; 0Þ; ð0; 1; 0Þ; ð0; 0; 1Þ, for jψ þþ i ≡ jDþ Dþ π − i;
which the momenta can all be permuted by little group
operations. The little group of the total momentum is C3v , jψ þ0 i ≡ jDþ D0 π þ i; ð71Þ
while the stabilizer is C1 . Without projecting under the
exchange group, the resulting irreps can be seen from the states of definite isospin with I z ¼ 0 are given by
relevant row of Table II to be
A1 ⊕ A2 ⊕ 2E: ð67Þ
Applying the various S3 projectors ΠΘ with Θ ∈ Y 3 restricts

to the following irreps:
ð72Þ
ð68Þ where the left-hand side uses the notation of Ref. [27]. In
the state jðDDÞIDD πiI , I is the total isospin, and I DD is the
isospin of the DD subsystem. As expected, the states with
I ¼ 2 and I ¼ 0 transform in the symmetric representation
Comparison with Table X determines which combinations of S2 . Of the two copies of I ¼ 1, one is symmetric, while
of cubic group and isospin irreps are compatible with the other is antisymmetric. Results for irrep decompositions
bosonic statistics for this set of momenta. For example, of DDπ operators are summarized in Table XI, where the
total I ¼ 0 must be combined with the cubic group irrep A2 , orbit dimension refers to the rank of the projected orbit-
while total I ¼ 1 or I ¼ 3 must be combined with the cubic representation matrices D̂ðs;ΘÞ . The trivial and sign irreps of
094516-18
TABLE XI. Combinations of irreps arising in decompositions of DDπ operator orbits. In the example states, the label D refers
collectively to D0 and Dþ and the pion is either π 0 or π þ. Details are as in Table II.
ðsÞ
GP HP S2 Irrep Example state Orbit dim Irrep decomposition
Oh Oh jDð0; 0; 0Þ; Dð0; 0; 0Þ; πð0; 0; 0Þi 1 A−1
Oh Cv4 jDð0; 0; 1Þ; Dð0; 0; −1Þ; πð0; 0; 0Þi 3 A−1 ⊕ E−
Oh Cv4 jDð0; 0; 1Þ; Dð0; 0; −1Þ; πð0; 0; 0Þi 3 Tþ
1
Oh Cv2 jDð0; 1; 1Þ; Dð0; −1; −1Þ; πð0; 0; 0Þi 6 A1 ⊕ E− ⊕ T −2
−
Oh Cv2 jDð0; 1; 1Þ; Dð0; −1; −1Þ; πð0; 0; 0Þi 6 Tþ

1 ⊕ T2
þ
Oh Cv3 jDð1; 1; 1Þ; Dð−1; −1; −1Þ; πð0; 0; 0Þi 4 A−1 ⊕ T −2

Oh Cv3 jDð1; 1; 1Þ; Dð−1; −1; −1Þ; πð0; 0; 0Þi 4 Aþ
2 ⊕ T1
þ
Oh CR2 jDð2; 1; 0Þ; Dð−2; −1; 0Þ; πð0; 0; 0Þi 12 A1 ⊕ A2 ⊕ 2E ⊕ T −1 ⊕ T −2

− − −
Oh CR2 jDð2; 1; 0Þ; Dð−2; −1; 0Þ; πð0; 0; 0Þi 12 2T þ

1 ⊕ 2T 2
þ
Oh CP2 jDð2; 1; 1Þ; Dð−2; −1; −1Þ; πð0; 0; 0Þi 12 A1 ⊕ E ⊕ T 1 ⊕ 2T −2

− − −
Oh CP2 jDð2; 1; 1Þ; Dð−2; −1; −1Þ; πð0; 0; 0Þi 12 Aþ þ

2 ⊕ E ⊕ 2T 1 ⊕ T 2
þ þ
Oh C1 jDð3; 2; 1Þ; Dð−3; −2; −1Þ; πð0; 0; 0Þi 24 A1 ⊕ A2 ⊕ 2E ⊕ 3T 1 ⊕ 3T −2

− − − −
Oh C1 jDð3; 2; 1Þ; Dð−3; −2; −1Þ; πð0; 0; 0Þi 24 Aþ þ þ

1 ⊕ A2 ⊕ 2E ⊕ 3T 1 ⊕ 3T 2
þ þ
S2 correspond to odd- and even-parity irreps in the Operators with the same quantum numbers can be
decomposition of the cubic group, respectively, such constructed from products of two octet baryon operators
that the overall exchange of two identical D mesons is and one pseudoscalar octet meson operator. This corre-
symmetric. sponds to the product of SUð3Þ irreps 8 ⊗ 8 ⊗ 8, which
includes two copies of the 1 irrep relevant for the
H-dibaryon. Tensor operators describing these products
4. H-dibaryon
are given by
The H-dibaryon provides an example where flavor
and operator-exchange symmetry come together to sat- dabc Ba Bb M c ; f abc Ba Bb Mc ; ð74Þ
isfy bosonic symmetry or fermionic antisymmetry. The
H-dibaryon is a hypothetical two-baryon bound state with where a, b, c are SUð3Þ adjoint indices, Ma is a pseudo-
strangeness S ¼ −2 that corresponds to an SUð3Þ-flavor scalar octet meson operator, and dabc (f abc ) are totally
singlet when the up, down, and strange quark masses are symmetric (antisymmetric) structure constants. Although
equal. Interpolating operators with these quantum numbers these operators involve linear combinations of several
can be constructed from two flavor-octet baryon interpolat- different products of meson and baryon flavors, for
ing operators using the SUð3Þ-flavor irrep decomposition example ΛΛπ 0 , ΛpK − , and Σþ Σ0 π − , their cubic irrep
decompositions and block-diagonalization matrices only
8 ⊗ 8 ¼ 1 ⊕ 8A ⊕ 8S ⊕ 10 ⊕ 10 ⊕ 27; ð73Þ depend on the permutation transformation properties of the
flavor tensors dabc and f abc as well as the fermionic nature
where 8A (8S ) denotes an SUð3Þ octet irrep where the two- of the baryon fields.
baryon flavor state is antisymmetric (symmetric) under Because Ba and M a fields represent distinct types of
exchange. The singlet irrep 1 corresponds to a symmetric SUð3Þ-octet particles while Ba and Bb are identical
flavor state associated with the operator Ba Bb δab where Ba besides their SUð3Þ flavor indices, the relevant exchange
is a baryon octet field with SUð3Þ adjoint index a. Writing group for this case is S2 . Writing explicitly the baryon spin
explicitly the baryon spin representation indices α1 , α2 representation indices α1 , α2 and the momentum labels
and the momentum labels n1 ; … for each field, fermion n1 ; …; n3 for each field, fermion antisymmetry implies
antisymmetry implies Baα1 ðn1 ÞBbα2 ðn2 Þ ¼ −Bbα2 ðn2 ÞBaα1 ðn1 Þ. Baα1 ðn1 ÞBbα2 ðn2 ÞMc ðn3 Þ ¼ −Bbα2 ðn2 ÞBaα1 ðn1 ÞM c ðn3 Þ. When
This implies that the momentum-spin states jn1 ; α1 ; n2 ; α2 i contracted with the totally symmetric tensor dabc, momen-
associated with δab Baα1 ðn1 ÞBbα2 ðn2 Þ will be antisymmetric tum-spin states jn1 ; α1 ; n2 ; α2 ; n3 i associated with these
under the exchange of momentum-spin pairs ðn1 ; α1 Þ ↔ operators will be antisymmetric under the exchange of
ðn2 ; α2 Þ. These states therefore transform in the sign irrep momentum-spin pairs ðn1 ; α1 Þ ↔ ðn2 ; α2 Þ. These states
of S2 , and the permutation projector defined in Eq. (64) therefore transform in the sign irrep of S2 , and the
should be applied. These two-baryon operators, which are permutation projector defined in Eq. (64) should be
linear combinations of ΛΛ, ΣΣ, and NΞ, therefore have applied. Conversely, when contracted with the totally
identical irrep decompositions to the case of nn operators antisymmetric tensor f abc, momentum-spin states created
summarized in Table VIII. by these operators will be symmetric under the exchange of
094516-19
TABLE XII. Permutation irreps of S2 or S3 for SUð3Þ-singlet invariant under exchange of both the first two momenta and
operators arising from products of three meson octet M a and the last two momenta and therefore transform in the trivial
baryon octet Ba operators. irrep of both S2 factors. The appropriate projector is a
product of the S2 trivial irrep projectors given in Eq. (70),
Flavor tensor Ma Mb Mc M a M b Bc Ba B b M c Ba Bb Bc
dabc
ð75Þ
fabc
In this case, the dimensionality of the extended orbit, i.e.,

momentum-spin pairs ðn1 ; α1 Þ ↔ ðn2 ; α2 Þ. These states the rank of the projected orbit-representation matrices
therefore transform in the trivial irrep of S2 , and the D̂ðs;ΘÞ , and the irrep decomposition depend on the momen-
permutation projector defined in Eq. (70) should be tum configuration.
applied. With four identical bosons such as π þ π þ π þ π þ , states
Similar considerations apply to other SUð3Þ-singlet transform in the trivial representation of the exchange
operators built from different combinations of three meson- group S4 . The appropriate permutation projector is there-
and baryon-octet operators, M a M b Bc , M a Mb Mc , and fore obtained from the normalized sum of all 4! ¼ 24
Ba Bb Bc , which provide further examples of the interplay elements of S4 , that is
between identical-particle labels and flavor-transformation
properties. Table XII summarizes the possible combina- ð76Þ
tions with irreps of the particle-exchange group S2 or S3.
The irrep decompositions for each of these cases are
5. ππKK shown in Table XIII. For a fixed exchange group, the size of
the irrep is related to the size of the orbit. In the second row,
An illustrative example of how the exchange group and
identical particles are moving back-to-back, which reduces
extended orbit dimension depend on the configuration of
the size of the decomposition compared to the third row,
the momentum orbit is provided by the ππKK and ππππ
where exchange group elements only affect operators with
systems with two pairs of particles moving back to back.
momenta that cannot be related by cubic transformations.
This also provides an example where the exchange group is
a direct product of nontrivial subgroups. Table XIII shows
the irrep decomposition for four spin-zero particles moving V. ALGORITHM SUMMARY
pairwise back-to-back with momenta n1 ¼ ð0; 0; 1Þ, To collect details spread across several sections, the
n2 ¼ ð0; 2; 0Þ, −n1 , and −n2 . Depending on how many steps of the full algorithm are reproduced here, including
operators correspond to identical bosons, different permu- for the case of identical particles with nonzero spin.
tation projectors are used and lead to different cubic-group The algorithm begins by selecting a fiducial state i ¼
irrep decompositions. ½n1 ; α1 ; ε1 ; …; nN ; αN ; εN (in terms of the momentum ni ,
The simplest case is π þ π − K þ K − , which has no identical spin and parity αi, and internal quantum numbers εi of the
particles. The (extended) orbits for all momentum con- ith particle) and specifying its permutation properties as
figurations have the standard irrep decomposition for corresponding to a row Θ of an exchange-group irrep.
distinguishable spin-zero operator products with stabilizer Given the fiducial state i and desired exchange-group irrep
group CR2 . row Θ, the algorithm proceeds as follows:
When there are two pairs of identical bosons such as (1) Compute the little group GP , defined in Eq. (10), of
π þ π þ K þ K þ , the exchange group is S2 × S2 . States are the total momentum.
TABLE XIII. Combinations of irreps arising in the decomposition of momentum orbits with four spin-zero operators, where n1 ¼
ð0; 0; 1Þ and n2 ¼ ð0; 2; 0Þ. Projection from symmetrization over identical operators generically reduces the number of irreps appearing
in the decomposition.
ðsÞ
GP HP Exchange Example state Orbit dim Irrep decomposition
Oh CR2 − jπ þ ðn − þ −
1 Þ; π ð−n1 Þ; K ðn2 Þ; K ð−n2 Þi 24 Aþ
1 ⊕ Aþ
⊕ 2Eþ ⊕ T þ
2
þ − −
1 ⊕ T 2 ⊕ 2T 1 ⊕ 2T 2
Oh CR2 jπ ðn1 Þ; π ð−n1 Þ; K ðn2 Þ; K þ ð−n2 Þi
þ þ þ 6 þ þ
A1 ⊕ A2 ⊕ 2E þ
Oh CR2 jπ þ ðn1 Þ; π þ ðn2 Þ; K þ ð−n1 Þ; K þ ð−n2 Þi 24 Aþ þ þ þ þ −

1 ⊕ A2 ⊕ 2E ⊕ T 1 ⊕ T 2 ⊕ 2T 1 ⊕ 2T 2
−
Oh CR2 jπ þ ðn1 Þ; π þ ð−n1 Þ; π þ ðn2 Þ; π þ ð−n2 Þi 6 þ þ

A1 ⊕ A2 ⊕ 2E þ
094516-20
ðΓÞ
(2) Compute irrep matrices of the little group Dμ0 μ ðRÞ supported in part by the U.S. Department of Energy,
via Eq. (B5). Office of Science under grant Contract No. DE-
ðsÞ
(3) Compute the extended orbit K̂ P of the fiducial state SC0011090 and by the SciDAC5 Award No. DE-
under the action of the little group and the exchange SC0023116. This work is supported by the National
group via Eq. (53). Science Foundation under Cooperative Agreement
(4) Compute the momentum-orbit representation ma- No. PHY-2019786 (The NSF AI Institute for Artificial
ðsÞ
trices Dm0 m ðRÞ, Eq. (18). Intelligence and Fundamental Interactions, [39]). P. E. S.
½J
(5) Compute the spin-representation matrices DJ0z Jz ðRD Þ is additionally supported by Early Career Award No. DE-
associated with each interpolating operator, defined SC0021006 and by Simons Foundation Grant No. 994314
generically in Eq. (47) or specifically for spin-half (Simons Collaboration on Confinement and QCD
(Gþ1 ) operators in Eq. (46).
Strings). G. K. is supported by funding from the Swiss
(6) Construct the combined momentum-spin-orbit rep- National Science Foundation (SNSF) through grant
resentation matrices via Eq. (51). agreement No. 200020_200424. This manuscript has
(7) Compute the exchange-group projector ΠΘ, Eq. (58), been authored by Fermi Research Alliance, LLC under
as described in Refs. [23–25]. Contract No. DE-AC02-07CH11359 with the U.S.
(8) Construct projected orbit-representation matrices Department of Energy, Office of Science, Office of
via Eq. (62). High Energy Physics. The reference implementation of
(9) Apply Schur’s lemma in the form of Eq. (34) to the this work makes use of NUMPY [40,41], SCIPY [42], and
projected orbit representation matrices in order to SYMPY [43]. Preliminary work and internal verifications
compute the first row for each irrep in the block- were performed with Wolfram Mathematica [44].
diagonalization matrices. Orthogonalize any degen-
erate copies which appear. APPENDIX A: GROUP CONVENTIONS
(10) Construct transition operators via Eq. (38) and use
Eq. (41) to fill the remaining rows of each irrep. The This appendix describes the conventions for the groups
result is the complete set of block-diagonalization Oh , ODh , and their subgroups, giving the concrete forms
ðΓ;κ;sÞ used in the numerical implementation of the algorithm
matrices Umμ .
presented in this work, available in Ref. [18]. Connections
VI. OUTLOOK to conventions in the literature are also discussed.
This work presents a general algorithm with which to 1. The cubic group Oh
construct multiparticle interpolating operators for quantum
field theories with cubic symmetry, including both lattice For the cubic group Oh , any element R ∈ Oh may be
and continuum theories. The algorithm, together with the written as a product of a reflection r and a permutation p [8],
implementation in Ref. [18], automates the block diago-
R ¼ rp; ðA1Þ
nalization to build multiparticle interpolating operators
transforming under irreps of the relevant little group.
with r and p given by
Automating this technical component allows the focus of
interpolating-operator construction to shift to the design of
r ∈ fe; rz ; ry ; ry rz ; rx ; rx rz ; rx ry ; rx ry rz g
local and extended operators to access multiparticle states
of interest. It also helps facilitate construction of large p ∈ fe; pxy ; pyz ; pxz ; pxyz ; pxzy g; ðA2Þ
operator sets in variational calculations aiming to constrain
finite-volume spectra precisely. where e is the identity matrix, rk acts on 3-vectors by
These or similar methods can be expected to play an multiplying the kth component by −1, pij acts on 3-vectors
increasingly important role in lattice QCD studies of multi- by permuting their ith and jth components, and the cyclic
particle systems. Especially for systems with multiple permutations are defined by pxyz ¼ pxy pyz and pxzy ¼
baryons, the field has developed rapidly over the past several pyz pxy . For example,
years as algorithmic advances (e.g., Refs. [28–34]) have
0 1 0 1
rendered variational studies a practical reality [14,35–37]. 0 1 0 −1 0 0
B C B C
pxy ¼ @ 1 0 0 A and rx ¼ @ 0 1 0 A: ðA3Þ
An implementation of the algorithm is publicly available
at [38]. 0 0 1 0 0 1
An ordering for the 48 elements of Oh can be established by

ACKNOWLEDGMENTS
labeling permutations and reflections as ra and pb with
The authors gratefully acknowledge useful discussions a ∈ f1; …; 8g and b ∈ f1; …; 6g ordered as shown in
with Fernando Romero-López. W. D., W. J. and P. E. S. are Eq. (A2) and Rc ¼ ra pb labeled by c ¼ 6ða − 1Þ þ b with
094516-21
c ∈ f1; …; 48g. For completeness, the elements are enumer- basis vectors was used for Oh irreps in Ref. [15], and the
ated as: explicit representation matrices obtained in the present
work therefore differ from those in Ref. [15] by a change
R1 ¼ e; R2 ¼ pxy ; of basis.
The subgroups of Oh are summarized in Table XIV. The
R3 ¼ pyz ; R4 ¼ pxz ;
present work follows the naming scheme of Ref. [9] which
R5 ¼ pxyz ; R6 ¼ pxzy ; labels one-dimensional irreps as variants of A or B and two-
R7 ¼ rz ; R8 ¼ rz pxy ; dimensional irreps as variants of E. It bears emphasizing
that irreps of different groups may have identical names,
R9 ¼ rz pyz ; R10 ¼ rz pxz ; but should be distinguished. The Clebsch-Gordan coeffi-
R11 ¼ rz pxyz ; R12 ¼ rz pxzy ; cients for little-group irreps below can be deduced from the
corresponding Clebsch-Gordan results in Ref. [11] for the
R13 ¼ ry ; R14 ¼ ry pxy ; Oh basis vectors identified with the little-group basis
R15 ¼ ry pyz ; R16 ¼ ry pxz ; vectors, or they can be calculated directly from the little-
group irrep matrices as described for example in Ref. [20].
R17 ¼ ry pxyz ; R18 ¼ ry pxzy ; As shown in Table I, basis functions for all irreps follow
R19 ¼ ry rz ; R20 ¼ ry rz pxy ; from the irreps of Oh .
It is useful to make several notes regarding the con-
R21 ¼ ry rz pyz ; R22 ¼ ry rz pxz ; ventions in Table I.
R23 ¼ ry rz pxyz ; R24 ¼ ry rz pxzy ; (i) For C4v, the irrep names and the coefficients appear-
ing in the E irrep definition are chosen so that identical
R25 ¼ rx ; R26 ¼ rx pxy ;
representation matrices for little group transforma-
R27 ¼ rx pyz ; R28 ¼ rx pxz ; tions are obtained as those presented in Ref. [15].
(ii) For C4v, the basis functions for irreps of Oh in
R29 ¼ rx pxyz ; R30 ¼ rx pxzy ;
Table I are eigenstates of Lz , which singles out the
R31 ¼ rx rz ; R32 ¼ rx rz pxy ; êz –axis. Other choices for the reference momentum,
R33 ¼ rx rz pyz ; R34 ¼ rx rz pxz ; e.g., P0 ¼ 2πL ðn; 0; 0Þ, remain valid but less conven-
ient, since the associated basis functions for C4v
R35 ¼ rx rz pxyz ; R36 ¼ rx rz pxzy ; must then be permuted.
R37 ¼ rx ry ; R38 ¼ rx ry pxy ; (iii) For C3v, the A1 and A2 representation matrices built
from these basis vectors using Eq. (19) match those
R39 ¼ rx ry pyz ; R40 ¼ rx ry pxz ; explicitly presented in Ref. [15]. The E representa-
R41 ¼ rx ry pxyz ; R42 ¼ rx ry pxzy ; tion matrices corresponding to this definition differ
from those of Ref. [15] by interchange of the rows/
R43 ¼ rx ry rz ; R44 ¼ rx ry rz pxy ; ðC ;EÞ ðC ;EÞ
columns [equivalent to B1 3v ðrÞ ↔ B2 3v ðrÞ] for
R45 ¼ rx ry rz pyz ; R46 ¼ rx ry rz pxz ; consistency with the convention of increasing lz
R47 ¼ rx ry rz pxyz ; R48 ¼ rx ry rz pxzy : ðA4Þ with μ applied here to irreps of Oh . The conventions
adopted here permit the Clebsch-Gordan coefficients
The basis functions for the irreps of Oh used in this work for the A1 , A2 , and E irreps of Ref. [11] to be applied
are specified in Table I. Irreps are classified by their to the corresponding irreps of C3v .
dimension and eigenvalue (1) under the parity operation (iv) For C2v, the irrep names are chosen so that the
representation matrices match those explicitly pre-
R43 ∶ r ↦ −r. The basis functions for the irreps Aþ −
1 , T1 ,
þ þ − sented in Ref. [15]. Clebsch-Gordan coefficients
T 2 , E , A2 are chosen to match those used in Ref. [11] and
for this and other little groups with only one-
correspond to linear combinations of spherical harmonics dimensional irreps are equal to the Clebsch-Gordan
with lz equal to 0, 1, 2, 2, and 3, respectively. The basis coefficients in Ref. [11] for the irreps corresponding
vectors for the remaining irreps A−1 , T þ − − þ
1 , T 2 , E , and A2 are to the same basis vectors.
taken to be linear combinations of the corresponding basis
vectors in Ref. [9]. The linear combinations are chosen so
that the same Clebsch-Gordan coefficients presented in 2. The double-cover group OD
h
Ref. [11] can be used for positive and negative parity irreps In the present work, the group OD h is defined using the
in all cases. Note however that the rows of the T 1 irreps are Dirac spinor representation consisting of the direct sum of a
ordered differently here than in Ref. [11] and Clebsch- positive-parity and a negative-parity spin-1=2 state, which
Gordan coefficient results must be transposed accordingly provides a faithful representation of the full group of spatial
(cf. conventions in Appendix A 3 below). A different set of transformations.
094516-22
TABLE XIV. Explicit forms for groups appearing in this work for Oh, OD h , and their subgroups. Basis functions for irreps of Oh (and
its subgroups) are given in Table I. Basis functions for the fermionic irreps of OD h (and its subgroups) are given in Table III. Group
parametrizations for the subgroups of OD h follow from those of O h by the replacements ri → rD
i , pij → pij and the inclusion of
D
inversions.
Momentum GP Order Irreps Group parametrization rp Group elements

2π
L ð0; 0; 0Þ
Oh 48 fA
1 ; A2 ; E ; T 1 ; T 2 g Eq. (A2) Eq. (A4)
2π
L ð0; 0; nÞ
C4v 8 fA1 ; A2 ; B1 ; B2 ; Eg fe; rx ; ry ; rx ry g × fe; pxy g f1; 2; 13; 14; 25; 26; 37; 38g
2π
L ðn; n; nÞ
C3v 6 fA1 ; A2 ; Eg feg × fe; pxy ; pyz ; pzx ; pxyz ; pxzy g f1; 2; 3; 4; 5; 6g
2π fA1 ; A2 ; B1 ; B2 g fe; rx g × fe; pyz g f1; 3; 25; 27g
L ð0; n; nÞ
C2v 4
2π fA; Bg fe; rz g × feg f1; 7g
L ðn; m; 0Þ
CR2 2
2π
L ðn; n; mÞ
CP2 2 fA; Bg feg × fe; pxy g f1; 2g
2π
L ðn; m; pÞ
C1 1 fAg feg × feg f1g
2π
L ð0; 0; 0Þ
ODh 96 fG
1 ; G2 ; H g 2π g × ðdoubled Eq: ðA2ÞÞ
feD ; RD f1; 2; …; 96g
2π fG1 ; G2 g feD ; RD
L ð0; 0; nÞ
Dic4 16 2π g × ðdoubled C4v Þ (doubled C4v )
2π fG; F1 ; F2 g feD ; RD
L ðn; n; nÞ
Dic3 12 2π g × ðdoubled C3v Þ (doubled C3v )
2π
L ð0; n; nÞ
Dic2 8 fGg feD ; RD 2π g × ðdoubled C2v Þ (doubled C2v )
2π
L ðn; m; 0Þ
CR4 4 fF1 ; F2 g feD ; RD 2π g × ðdoubled C2 Þ
R (doubled CR2 )
2π fF1 ; F2 g fe ; R2π g × ðdoubled CP2 Þ
L ðn; n; mÞ
CP4 4 D D (doubled CP2 )
2π fFg feD ; RD
L ðn; m; pÞ
CD1 2 2π g × ðdoubled C1 Þ (doubled C1 )
The group elements of Oh can be mapped to (half of) the the Euclidean and mostly-positive Minkowski gamma
group elements of OD h by replacing rotation operators in the matrices). The superscript D is used to denote double-
defining representations of SOð3Þ with the corresponding cover group elements here and below. The definition in
rotation operators in the Dirac spinor representation. To do Eq. (A7) implies the transformation property
so, first note that the explicit matrix representation of Oh in X
terms of permutations and reflections in Eqs. (A2) and (A3) RD γ j ðRD Þ† ¼ γ i Rij : ðA8Þ
can be (nonuniquely) related to a matrix representation i
in terms of rotations and the parity operator P ¼ The Dirac spinor representation of the parity element PD is
diagð−1; −1; −1Þ by given by the temporal gamma matrix up to an overall phase.
The Euclidean γ 4 and Minkowski γ 0 are equivalent up to a
rk ¼ P · Rðπ êk Þ; phase choice. The present work takes PD ¼ γ 4 , which

π satisfies fγ i ; γ 4 g ¼ 0, γ †4 ¼ γ 4 , and γ 24 ¼ 1.
pij ¼ P · Rðπ êi Þ · R êi × êj : ðA5Þ
2 The double-cover permutations and reflections are then
defined in the Dirac spinor representation by
Above, thePêi are unit vectors in the ith direction and
k ¼ P · R ðπ êk Þ ¼ γ 5 γ k ;
rD D D
Rð ⃗ωÞ ≡ Rð k ωk êk Þ describes a rotation by angle j ⃗ωj

about the ω̂ axis,
D π
pD ¼ PD
· R D
ðπ êi Þ · R ê × ê
X
ij
2 i j
Rð ⃗ωÞ ¼ exp − ωk tk ðA6Þ 1

¼ pffiffiffi γ 5 ðγ i − γ j Þ; ðA9Þ
k 2
in terms of the soð3Þ generators ½tk ij ¼ εijk . where γ 5 ¼ γ 1 γ 2 γ 3 γ 4 is the fifth Euclidean gamma matrix.
The Dirac spinor representation of the corresponding The set of products of rD D
k and pij analogous to Eq. (A4)
element of the double cover is given by provides an explicit matrix representation of the first 48
elements of OD h . Note that this set is not closed under group
1X multiplication; Oh is not a subgroup of OD
RD ð ⃗ωÞ ¼ exp − ω ε ½γ ; γ ; ðA7Þ h . The remaining
8 i;j;k k ijk i j 48 elements can be obtained by multiplying these elements
by a 2π rotation,
where the γ i are the spatial gamma matrices satisfying
fγ i ; γ j g ¼ 2δij and γ †i ¼ γ i (this choice coincides with both 2π ¼ diagð−1; −1; −1; −1Þ;
RD ðA10Þ
094516-23
pffiffiffi
where the form above holds for a 2π rotation about ðDic ;F Þ 13 3 ð1 þ iÞð2 − i 2Þ3 1
any axis. B1 3 2 ¼ ; ;þ þ pffiffiffi 2 ; 2 ;þ
22 2 4 3
The numerical implementation in Ref. [18] uses the pffiffiffi
Dirac-Pauli basis, in which the γ-matrices are represented 2 − i3 1 1 þ i3 3
− pffiffiffi ;− ;þ þ pffiffiffi ;− ;þ : ðA13Þ
in 2 × 2 block form as 2 3 2 2 2 22 2

0 −iσ k I 0
γk ¼ ; γ4 ¼ ;
iσ k 0 0 −I 3. Phase conventions

0 −I The block-diagonalization matrices in Eq. (1) are only
γ5 ¼ ; ðA11Þ defined up to an overall phase. This appendix records the
−I 0
phase conventions used in Ref. [18].
where the σ k are the usual Pauli matrices and I is the 2 × 2 For each irrep except T 2 of the cubic group, the overall
identity matrix; for more details and relations to other phase within each irrep is selected such that the first
ðΓi ;κ;sÞ
common bases see Ref. [11].9 nonzero entry of U m;μ¼1 is real and positive. For T
2 of
To distinguish the notion of abstract group elements the cubic group, the overall phase is selected such that
RD ∈ OD h from the specific spinor representation in this
ðΓi ;κ;sÞ
Um;μ¼2 is purely imaginary with negative imaginary part.
basis, the action of the group on these spinors is denoted by This choice matches the basis-vector conventions of
SðRD Þ, following the convention of Ref. [15]. Because this Ref. [11], where the combination of spherical harmonics
is the defining representation, the matrix representations are Y 22 − Y −22 is used as the μ ¼ 2 basis vector for T 2. This
simply given by SðRD Þ ¼ RD using the definitions in choice ensures that the Clebsch-Gordan coefficients pre-
Eq. (A9) and the Dirac-Pauli basis above. sented in Ref. [11] are applicable to operators constructed
Basis vectors for irreps of OD h and its subgroups are using the methods of the present work (noting the different
summarized in Table III. The naming convention in the ordering of the rows of the T 1 irreps discussed above).
present work follows Ref. [15] in denoting 1, 2, and 4
dimensional fermionic irreps by F, G, and H, respectively.
For alternative strategies involving subduction of OD APPENDIX B: POLARIZATION TENSORS
h irreps
into little-group irreps, see Refs. [15,45]. This appendix recasts the computation of irrep matrices
The only nontrivial basis vectors are associated with ðΓÞ
Dμ0 μ ðRÞ(see Eq. (19)) algebraically using polarization
1-dimensional irreps F1 and F2 of Dic3 , which serves as the
tensors. For a given irrep Γ, the basis functions are
little group for the momentum P ¼ 2π L ðn; n; nÞ for any homogeneous polynomials of fixed degree d, i.e.,
n ∈ Znf0g. Basis vectors for F1 and F2 may be obtained ðΓÞ ðΓÞ
by projecting to the linear combinations of fj32 ; 32 ; þi; Bμ ðλrÞ ¼ λd Bμ ðrÞ for λ ∈ R. Homogenous polynomials
ðΓÞ
j32 ; 12 ; þi; j32 ; − 12 ; þi; j32 ; − 32 ; þig transforming in these Bμ ðrÞ can be expressed in terms of so-called polarizations
ðΓÞ
irreps using Schur’s lemma (see Eq. (30) and the characters PBμ , symmetric rank-d tensors defined via [46]
of F1 and F2 given in Ref. [15]. It is also possible to
identify these irreps from first principles by determining the ðΓÞ
½PBμ ðrð1Þ ; …; rðdÞ Þ
two-dimensional orthogonal complement of the G irrep
contained in the J ¼ 3=2 representation using Gram- ðΓÞ
≡ ðPBμ Þa1 a2 …ad rað1Þ rð2Þ …raðdÞ
1 a2 d
Schmidt orthonormalization, then solving for linear com-

1 ∂ ∂ ðΓÞ

binations that diagonalize the representation matrices for ≡ Bμ ðλ1 rð1Þ þ þ λd rðdÞ Þ ; ðB1Þ
this orthogonal compliment. The basis vectors used in this d! ∂λ1 ∂λd λi ¼0
work are:
pffiffiffi where frðiÞ ∈ R3 ; 1 ≤ i ≤ dg are arbitrary auxiliary vec-
ðDic3 ;F1 Þ 13 3 ð1 − iÞð 2 − 2iÞ3 1 tors.10 In the second line, summation is implied over the
B1 ¼ ; ;þ − pffiffiffi 2 ; 2 ;þ
22 2 4 3 repeated indices ai ∈ f1; 2; 3g. The polarization is sym-
pffiffiffi metric and tensorial due to the symmetry and linearity of
2 þ i3 1 1 þ i3 3
þ pffiffiffi ;− ;þ − pffiffiffi ;− ;þ ; ðA12Þ the derivatives. Evaluated diagonally (i.e., contracted with
2 3 2 2 2 22 2
10
The identification between basis functions and symmetric
tensors amounts to a map between the polynomial ring
9
Kðfx; y; zgÞ and the symmetric tensor space SðR3 Þ. Such a
Note that the change-of-basis matrix relating the Dirac-Pauli relationship is quite general. In fact, for any vector space V with
and DeGrand-Rossi basespdenoted
ffiffiffi U ðDRÞ in Ref. [11] should be basis B and dual space V , the two spaces are canonically
ðDRÞ
U ¼ ð−iγ 2 þ iγ 1 γ 3 Þ= 2 in terms of Dirac-Pauli matrices. isomorphic: KðBÞ ≃ SðV Þ [46].
094516-24
the same vector r along all d indices), the polarization ðT þ

2Þ ðT þ
2Þ
1
ðB̄1 Þ13 ¼ ðB̄1 Þ31 ¼ − ;
returns the original homogeneous polynomial: 2
ðT þ
2Þ ðT þ
2Þ
i
ðΓÞ ðΓÞ ðB̄1 Þ23 ¼ ðB̄1 Þ32 ¼ : ðB7Þ
½PBμ ðr; …; rÞ ¼ Bμ ðrÞ: ðB2Þ 2
This relationship can provide a useful consistency check in As an example of the algebraic setup, consider the
explicit calculations. In this work, polarizations are nor- transformation about the y-axis sending x → z and z → −x:
malized with respect to the tensor inner product 0 1
0 0 1
⟪X; Y⟫ ≡ ðX Þa1 a2 …ad ðYÞa1 a2 …ad : ðB3Þ B C
R¼@ 0 1 0 A: ðB8Þ
−1 0 0
Polarizations normalized with respect to this inner product
ðΓÞ
are denoted by B̄μ . ðT þ
2Þ
ðΓÞ The matrix element of this rotation between B̄1 and itself
The transformation of B̄μ under rotations follows
immediately from the definition in Eq. (B1), is given by
ðT þ Þ ðT þ Þ ðT þ Þ
ðΓÞ
B̄μ ðR−1 r; …; R−1 rÞ D112 ðRÞ ¼ ⟪B̄1 2
; R ∘ B̄1 2
⟫ ðB9Þ
ðΓÞ 20 10 13
¼ ðB̄μ Þa1 …ad Rb1 a1 Rbd ad rb1 rbd 0 0 −1=2 0 i=2 1=2
ðΓÞ 6B CB C7
≡ ðR ∘ B̄μ Þb1 …bd rb1 rbd ; ðB4Þ ¼ tr4@ 0 0 −i=2 A@ i=2 0 0 A5
−1=2 −i=2 0 1=2 0 0
where the second line uses R−1 ¼ RT and the final line
ðΓÞ ðB10Þ
defines R ∘ B̄μ . Given this transformation property, the
inner product in Eq. (19) reduces to ¼ −1=2: ðB11Þ
ðΓÞ ðΓÞ ðΓÞ
Dμ0 μ ðRÞ ¼ ⟪B̄μ0 ; R ∘ B̄μ ⟫: ðB5Þ The equivalent calculation in the integral setup is
Z
The tensorial method for computing the matrix elements ðT þ Þ 8 −zx þ iyz xz þ iyx
via Eq. (B5) generalizes easily to particles with spin (see D112 ðRÞ ¼ 2 dΩ pffiffiffi pffiffiffi ðB12Þ
π 2 2
Section III), since spin vectors can be viewed as basis
Z
vectors for the double-cover of the relevant little group with −4
transformation properties analogous to Eq. (21). ¼ 2 dΩeiφ sin2 θcosθcosφðcosθ þisinθsinφÞ
π
An explicit example of polarization tensors in this context
follows. Table I gives the basis function for lz ¼1 row of the ¼ −1=2: ðB13Þ
ðT þ
2Þ
pffiffiffi
irrep T þ2 of Oh : B1 ðrÞ ¼ ð−zx þ iyzÞ= 2. This function Although the integral in the second line is elementary,
is a homogeneous degree-2 polynomial, so the polarization evaluating many such integrals is cumbersome when
will be a rank-2 tensor. Using the coordinates rðiÞ ¼ compared to tensor algebra.
ðxðiÞ ; yðiÞ ; zðiÞ Þ for auxiliary vectors, the polarization is
ðT þ Þ
½P B̄1 2
ðrð1Þ ; rð2Þ Þ APPENDIX C: EXPLICIT BLOCK-
DIAGONALIZATION MATRICES
1
¼ pffiffiffi ½−zð1Þ xð2Þ − xð1Þ zð2Þ þ iyð1Þ zð2Þ þ izð1Þ yð2Þ : ðB6Þ
2 2 This appendix presents examples of block-diagonalization
matrices in the rest frame in a few illustrative cases. The
The nonzero components of the normalized basis results in Tables XV–XVII employ the phase conventions
tensor are in Appendix A 3.
094516-25
ðΓ;κ;sÞ
TABLE XV. Table of U mμ for the orbit of ½n1 ; n2 ¼ ½ð0; 0; 1Þ; ð0; 0; −1Þ. Columns of the tables are listed in
order of increasing m ∈ f1; …; 6g. Only the momentum of the first operator is shown in the column header.
Γ, κ μ (0, 0, 1) (0, 1, 0) (1, 0, 0) (0, 0, −1) (0, −1, 0) (−1, 0, 0)

Aþ
1 ;1 1 p1ffiffi
6
p1ffiffi
6
p1ffiffi
6
p1ffiffi
6
p1ffiffi
6
p1ffiffi
6
Eþ ; 1 1 p1ffiffi − 2p1 ffiffi3 − 2p1 ffiffi3 p1ffiffi − 2p1 ffiffi3 − 2p1 ffiffi3
3 3
2 0 − 12 1
2
0 − 12 1
2
T −1 ; 1 1 p1ffiffi 0 0 − p1ffiffi2 0 0
2
2 0 i
2 − 12 0 − 2i 1
2
1
3 0 i
2 2
0 − 2i − 12
ðΓ;κ;sÞ
TABLE XVI. Table of U mμ for the orbit of ½n1 ; n2 ¼ ½ð0; 1; 1Þ; ð0; −1; −1Þ. Columns of the tables are listed in order of increasing
m ∈ f1; …; 12g. Only the momentum of the first operator is shown in the column header.
Γ, κ μ (1, 1, 0) (1, 0, 1) (0, 1, 1) (1, 0, −1) (0, 1, −1) (1, −1, 0) (0, −1, 1) (0, −1; −1) (−1, 1, 0) (−1, 0, 1) ð−1; 0; −1Þ ð−1; −1; 0Þ
Aþ
1 ;1 1
1 ffiffi
p
2 3
1 ffiffi
p
2 3
1 ffiffi
p
2 3
1 ffiffi
p
2 3
1 ffiffi
p
2 3
1 ffiffi
p
2 3
1 ffiffi
p
2 3
1 ffiffi
p
2 3
1 ffiffi
p
2 3
1 ffiffi
p
2 3
1 ffiffi
p
2 3
1 ffiffi
p
2 3
Eþ ; 1 1 p1ffiffi − 2p1 ffiffi6 − 2p1 ffiffi6 − 2p1 ffiffi6 − 2p1 ffiffi6 p1ffiffi − 2p1 ffiffi6 − 2p1 ffiffi6 p1ffiffi − 2p1 ffiffi6 − 2p1 ffiffi6 p1ffiffi
6 6 6 6
2 0 − p1 ffiffi
2 2
1 ffiffi
p
2 2
− p1 ffiffi
2 2
1 ffiffi
p
2 2
0 1 ffiffi
p
2 2
1 ffiffi
p
2 2
0 − p1 ffiffi
2 2
− p1 ffiffi
2 2
0
i ffiffi
Tþ
2 ;1 1 0 − 2p1 ffiffi2 p
2 2
1 ffiffi
p
2 2
− 2pi ffiffi2 0 − 2pi ffiffi2 i ffiffi
p
2 2
0 1 ffiffi
p
2 2
− 2p1 ffiffi2 0
2 − 2i 0 0 0 0 i
2
0 0 i
2
0 0 − 2i
i ffiffi
3 0 1 ffiffi
p
2 2
p
2 2
− 2p1 ffiffi2 − pi ffiffi
2 2
0 − pi ffiffi
2 2
i ffiffi
p
2 2
0 − 2p1 ffiffi2 1 ffiffi
p
2 2
0
T −1 ; 1 1 0 1 ffiffi
p 1 ffiffi
p − 2p1 ffiffi2 − 2p1 ffiffi2 0 1 ffiffi
p − 2p1 ffiffi2 0 1 ffiffi
p − 2p1 ffiffi2 0
2 2 2 2 2 2 2 2
2 − 14 þ 4i − 14 i
4 − 14 i
4 − 14 − 4i − 4i − 4i 1
4 þ i
4
1
4
1
4
1
4 − 4i
3 14 þ 4i 1
4
i
4
1
4
i
4
1
4−4
i
− 4i − 4i − 14 þ 4i − 14 − 14 − 14 − 4i
T −2 ; 1 1 1
4 þ 4i − 4i − 14 − 4i − 14 − 14 þ 4i 1
4
1
4
1
4 − 4i i
4
i
4 − 14 − 4i
2 0 − 2pi ffiffi2 pi ffiffi
2 2
pi ffiffi
2 2
− 2pi ffiffi2 0 i ffiffi
p
2 2
− 2pi ffiffi2 0 − 2pi ffiffi2 i ffiffi
p
2 2
0
3 − 14 þ 4i − 4i 1
4
− 4i 1
4
1
4 þ 4i − 14 − 14 − 14 − 4i i
4
i
4
1
4 − 4i
ðΓ;κ;sÞ
TABLE XVII. Table of U mμ for the orbit of ½n1 ; n2 ¼ ½ð1; 1; 1Þ; ð−1; −1; −1Þ. Columns of the tables are listed in order of
increasing m ∈ f1; …; 8g. Only the momentum of the first operator is shown in the column header.
Γ, κ μ (1, 1, 1) ð1; 1; −1Þ ð1; −1; 1Þ ð1; −1; −1Þ ð−1; 1; 1Þ ð−1; 1; −1Þ ð−1; −1; 1Þ ð−1; −1; −1Þ
Aþ
1 ;1 1 1 ffiffi
p
2 2
1 ffiffi
p
2 2
1 ffiffi
p
2 2
1 ffiffi
p
2 2
1 ffiffi
p
2 2
1 ffiffi
p
2 2
1 ffiffi
p
2 2
1 ffiffi
p
2 2
Tþ
2 ;1 1 − 14 þ 4i 1
4−4
i
− 14 − 4i 1
4 þ 4i 1
4 þ 4i − 14 − 4i 1
4−4
i
− 14 þ 4i
2 − 2pi ffiffi2 − 2pi ffiffi2 i ffiffi
p
2 2
i ffiffi
p
2 2 2 2
i ffiffi
p i ffiffi
p
2 2
− 2pi ffiffi2 − 2pi ffiffi2
1
3 4 þ 4i − 14 − 4i 1
4−4
i
− 14 þ 4i − 14 þ 4i 1
4−4
i
− 14 − 4i 1
4 þ 4i
A−2 ; 1 1 1 ffiffi
p − 2p1 ffiffi2 − 2p1 ffiffi2 1 ffiffi
p − 2p1 ffiffi2 1 ffiffi
p 1 ffiffi
p − 2p1 ffiffi2
2 2 2 2 2 2 2 2
T −1 ; 1 1 p1 ffiffi
− 2p1 ffiffi2 1 ffiffi
p − 2p1 ffiffi2 p1 ffiffi
− 2p1 ffiffi2 1 ffiffi
p − 2p1 ffiffi2
2 2 2 2 2 2 2 2
1
2 − 4 þ 4i − 14 þ 4i 1
− 4 − 4i − 14 − 4i 1
4þ4
i 1
4 þ i
4
1
4−4
i 1
4 − 4i
1 1 1 1 1
3 4þ4
i
4þ4
i
4−4
i
4−4
i
− 4 þ 4i − 14 þ 4i 1
− 4 − 4i − 14 − 4i
094516-26
[1] W. Detmold, R. G. Edwards, J. J. Dudek, M. Engelhardt, [17] M. T. Hansen, F. Romero-López, and S. R. Sharpe, General-
H.-W. Lin, S. Meinel, K. Orginos, and P. Shanahan izing the relativistic quantization condition to include all
(USQCD Collaboration), Hadrons and nuclei, Eur. Phys. three-pion isospin channels, J. High Energy Phys. 07 (2020)
J. A 55, 193 (2019). 047; 02 (2021) 14.
[2] A. Bazavov, F. Karsch, S. Mukherjee, and P. Petreczky [18] W. Detmold, W. I. Jay, G. Kanwar, P. E. Shanahan, and M.
(USQCD Collaboration), Hot-dense Lattice QCD: USQCD Wagman, github.com/LatticeQCDTools/mhi.
whitepaper 2018, Eur. Phys. J. A 55, 194 (2019). [19] M. Artin, Algebra (Pearson Education, Boston, MA, 2011).
[3] A. S. Kronfeld, D. G. Richards, W. Detmold, R. Gupta, [20] K. Rykhlinskaya and S. Fritzsche, Generation of Clebsch–
H.-W. Lin, K.-F. Liu, A. S. Meyer, R. Sufian, and S. Gordan coefficients for the point and double groups,
Syritsyn (USQCD Collaboration), Lattice QCD and neu- Comput. Phys. Commun. 174, 903 (2006).
trino-nucleus scattering, Eur. Phys. J. A 55, 196 (2019). [21] W. Fulton and J. Harris, Representation Theory, Graduate
[4] C. Lehner et al. (USQCD Collaboration), Opportunities for Texts in Mathematics (Springer, New York, 2004), 10.1007/
lattice QCD in quark and lepton flavor physics, Eur. Phys. 978-1-4612-0979-9.
J. A 55, 195 (2019). [22] H. Georgi, Lie Algebras in Particle Physics: From Isospin
[5] V. Cirigliano, Z. Davoudi, T. Bhattacharya, T. Izubuchi, To Unified Theories (Taylor & Francis, Boca Raton,
P. E. Shanahan, S. Syritsyn, and M. L. Wagman (USQCD 2000), 10.1201/9780429499210.
Collaboration), The role of lattice QCD in searches for [23] S. Keppeler and M. Sjödahl, Hermitian Young Operators,
violations of fundamental symmetries and signals for new J. Math. Phys. (N.Y.) 55, 021702 (2014).
physics, Eur. Phys. J. A 55, 197 (2019). [24] J. Alcock-Zeilinger and H. Weigert, Transition operators,
[6] R. C. Brower, A. Hasenfratz, E. T. Neil, S. Catterall, G. J. Math. Phys. (N.Y.) 58, 051703 (2017).
Fleming, J. Giedt, E. Rinaldi, D. Schaich, E. Weinberg, and [25] J. Alcock-Zeilinger and H. Weigert, Compact Hermitian
O. Witzel (USQCD Collaboration), Lattice gauge theory for Young projection operators, J. Math. Phys. (N.Y.) 58,
physics beyond the standard model, Eur. Phys. J. A 55, 198 051702 (2017).
(2019). [26] R. Aaij et al. (LHCb Collaboration), Observation of an
[7] J. E. Mandula, G. Zweig, and J. Govaerts, Covariant lattice exotic narrow doubly charmed tetraquark, Nat. Phys. 18,
glueball fields, Nucl. Phys. B228, 109 (1983). 751 (2022).
[8] J. E. Mandula, G. Zweig, and J. Govaerts, Representations [27] M. T. Hansen, F. Romero-López, and S. R. Sharpe, Incor-
of the rotation reflection symmetry group of the four- porating DDπ effects and left-hand cuts in lattice QCD
dimensional cubic lattice, Nucl. Phys. B228, 91 (1983). studies of the T cc ð3875Þþ , arXiv:2401.06609.
[9] M. S. Dresselhaus, G. Dresselhaus, and A. Jorio, Group [28] M. Peardon, J. Bulava, J. Foley, C. Morningstar, J. Dudek,
Theory: Application to the Physics of Condensed Matter R. G. Edwards, B. Joó, H.-W. Lin, D. G. Richards, and K. J.
(Springer, New York, 2008), 10.1007/978-3-540-32899-5. Juge (Hadron Spectrum Collaboration), A Novel quark-field
[10] S. Basak, R. G. Edwards, G. T. Fleming, U. M. Heller, C. creation operator construction for hadronic physics in lattice
Morningstar, D. Richards, I. Sato, and S. Wallace, Group- QCD, Phys. Rev. D 80, 054506 (2009).
theoretical construction of extended baryon operators in [29] C. Morningstar, J. Bulava, J. Foley, K. J. Juge, D. Lenkner,
lattice QCD, Phys. Rev. D 72, 094506 (2005). M. Peardon, and C. H. Wong, Improved stochastic estima-
[11] S. Basak, R. Edwards, G. T. Fleming, U. M. Heller, C. tion of quark propagation with Laplacian Heaviside smear-
Morningstar, D. Richards, I. Sato, and S. J. Wallace (LHPC ing in lattice QCD, Phys. Rev. D 83, 114505 (2011).
Collaboration), Clebsch-Gordan construction of lattice in- [30] W. Detmold, D. J. Murphy, A. V. Pochinsky, M. J. Savage,
terpolating fields for excited baryons, Phys. Rev. D 72, P. E. Shanahan, and M. L. Wagman, Sparsening algorithm
074501 (2005). for multihadron lattice QCD correlation functions, Phys.
[12] C. E. Thomas, R. G. Edwards, and J. J. Dudek, Helicity Rev. D 104, 034502 (2021).
operators for mesons in flight on the lattice, Phys. Rev. D [31] Y. Li, S.-C. Xia, X. Feng, L.-C. Jin, and C. Liu, Field
85, 014507 (2012). sparsening for the construction of the correlation functions
[13] T. Luu and M. J. Savage, Extracting scattering phase-shifts in lattice QCD, Phys. Rev. D 103, 014514 (2021).
in higher partial-waves from lattice QCD calculations, Phys. [32] J. Chen, R. G. Edwards, and W. Mao, Graph contractions
Rev. D 83, 114508 (2011). for calculating correlation functions in lattice QCD, in
[14] S. Amarasinghe, R. Baghdadi, Z. Davoudi, W. Detmold, M. Proceedings of the Platform for Advanced Scientific Com-
Illa, A. Parreno, A. V. Pochinsky, P. E. Shanahan, and M. L. puting Conference, PASC ’23 (Association for Computing
Wagman, Variational study of two-nucleon systems with Machinery, New York, NY, USA, 2023), 10.1145/
lattice QCD, Phys. Rev. D 107, 094508 (2023). 3592979.3593409.
[15] C. Morningstar, J. Bulava, B. Fahy, J. Foley, Y. C. Jhang, [33] N. Humphrey, W. Detmold, R. D. Young, and J. M. Zanotti,
K. J. Juge, D. Lenkner, and C. H. Wong, Extended hadron Novel algorithms for computing correlation functions of nuclei,
and two-hadron operators of definite momentum for spec- Proc. Sci., LATTICE2021 (2022) 431 [arXiv:2201.04269].
trum calculations in lattice QCD, Phys. Rev. D 88, 014511 [34] Q. Wang, B. Ren, J. Chen, and R. G. Edwards, Micco: An
(2013). enhanced multi-GPU scheduling framework for many-body
[16] S. Prelovsek, U. Skerbis, and C. B. Lang, Lattice operators correlation functions, in 2022 IEEE International Parallel
for scattering of particles with spin, J. High Energy Phys. 01 and Distributed Processing Symposium (IPDPS) (IEEE,
(2017) 129. New York, 2022), pp. 135–145.
094516-27
[35] A. Francis, J. R. Green, P. M. Junnarkar, C. Miao, T. D. Rae, [40] S. van der Walt, S. C. Colbert, and G. Varoquaux, The
and H. Wittig, Lattice QCD study of the H dibaryon using NUMPY Array: A structure for efficient numerical compu-
hexaquark and two-baryon interpolators, Phys. Rev. D 99, tation, Comput. Sci. Eng. 13, 22 (2011).
074505 (2019). [41] C. R. Harris et al., Array programming with NUMPY, Nature
[36] B. Hörz et al., Two-nucleon S-wave interactions at (London) 585, 357 (2020).
the SUð3Þ flavor-symmetric point with mud ≃ mphys s : [42] P. Virtanen et al., SCIPY1.0–Fundamental Algorithms for
A first lattice QCD calculation with the stochastic Scientific Computing in Python, Nat. Methods 17, 261 (2020).
Laplacian Heaviside method, Phys. Rev. C 103, [43] A. Meurer et al., SYMPY: Symbolic computing in Python,
014003 (2021). PeerJ Comput. Sci. 3, e103 (2017).
[37] J. R. Green, A. D. Hanlon, P. M. Junnarkar, and H. [44] W. R. Inc., Mathematica, Version 14.0, champaign, IL, 2024,
Wittig, Weakly bound H dibaryon from SU(3)- https://www.wolfram.com/mathematica.
flavor-symmetric QCD, Phys. Rev. Lett. 127, 242003 [45] D. C. Moore and G. T. Fleming, Angular momentum on the
(2021). lattice: The case of non-zero linear momentum, Phys. Rev.
[38] W. Detmold, W. I. Jay, G. Kanwar, P. E. Shanahan, and D 73, 014504 (2006); 74, 079905(E) (2006).
M. L. Wagman, mhi, Zenodo, 10.5281/zenodo.11267518 [46] C. Procesi, Lie Groups: An Approach Through Invariants
(2024). and Representations (Springer, New York, 2007), 10.1007/
[39] http://iaifi.org/. 978-0-387-28929-8.
094516-28

PhysRevD 109 094516

Uploaded by

Copyright:

Available Formats

PhysRevD 109 094516

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

PhysRevD 109 094516

Uploaded by

Copyright:

Available Formats

PHYSICAL REVIEW D 109, 094516 (2024)

Multiparticle interpolating operators in quantum field theories

(Received 20 March 2024; accepted 6 May 2024; published 31 May 2024)

I. INTRODUCTION depends sensitively on the choice of interpolating operators

2470-0010=2024=109(9)=094516(28) 094516-1 Published by the American Physical Society

where R · ½n1 ; …; nN ¼ ½Rn1 ; …; RnN . Here, s is an ðsÞ

satisfying ðΓÞ 1 ðΓÞ ðΓÞ

elements in the set X. N

C3v A1 1 ðOh ;Aþ Þ

ðΓi ;sÞ jΓi j X ðΓi Þ

which follows from unitarity of the change-of-basis matri-

Oh CR2 jπð2; 1; 0Þ; Kð−2; −1; 0Þi 24 A1 ⊕ A2 ⊕ 2Eþ ⊕ T 1 ⊕ T þ

C4v C4v jπð0; 0; 1Þ; Kð0; 0; 0Þi 1 A1

Group Irrep μ Basis vector

B. Operator representations of the little group instead of those defined by continuum

sentation matrix associated with the group element RD , ðsÞ ðsÞ

transformation rules above, these states transform as

Oh C2v jnð0; 1; 1Þ; pð0; −1; −1Þi 48 A1 ⊕ A2 ⊕ 2Eþ ⊕ 3T þ

jnð1; 1; 1Þ; pð−1; −1; −1Þi þ þ þ þ þ − − − − −

Oh C1 jnð3; 2; 1Þ; pð−3; −2; −1Þi 192 4Aþ þ þ þ þ − − − −

C4v C4v jnð0; 0; 1Þ; pð0; 0; 0Þi 4 A1 ⊕ A2 ⊕ E

ΠΘ which projects onto the relevant row [23–25]. In fact,

T ΘΦ T Φ0 Θ0 ¼ δΦΦ0 T ΘΘ0 ; ð56Þ

Name Dimension Young diagram Young tableaux

Oh C2v jnð0; 1; 1Þ; nð0; −1; −1Þi 24 Aþ þ þ − − −

jnð1; 1; 1Þ; nð−1; −1; −1Þi þ þ − − − −

Oh C1 jnð3; 2; 1Þ; nð−3; −2; −1Þi 96 Aþ þ þ þ þ − − −

C4v C4v jnð0; 0; 1Þ; nð0; 0; 1Þi 1 A1

jnð1; 1; 1Þ; nð−1; −1; −1Þ; nð0; 0; 0Þi þ þ þ − − −

Oh CR2 jnð0; 1; 1Þ; nð0; −1; 0Þ; nð0; 0; −1Þi 96 4Gþ þ þ − −

Oh CP2 jnð2; 1; 1Þ; nð−2; −1; −1Þ; nð0; 0; 0Þi 96 2Gþ þ þ − −

Oh C1 jnð0; −1; 1Þ; nð2; 1; 0Þ; nð−2; 0; −1Þi 192 8Gþ þ þ − −

group irrep A1 . Both I ¼ 1 and I ¼ 2 may be combined

Applying the various S3 projectors ΠΘ with Θ ∈ Y 3 restricts

Oh Cv2 jDð0; 1; 1Þ; Dð0; −1; −1Þ; πð0; 0; 0Þi 6 Tþ

Oh Cv3 jDð1; 1; 1Þ; Dð−1; −1; −1Þ; πð0; 0; 0Þi 4 A−1 ⊕ T −2

Oh CR2 jDð2; 1; 0Þ; Dð−2; −1; 0Þ; πð0; 0; 0Þi 12 A1 ⊕ A2 ⊕ 2E ⊕ T −1 ⊕ T −2

Oh CR2 jDð2; 1; 0Þ; Dð−2; −1; 0Þ; πð0; 0; 0Þi 12 2T þ

Oh CP2 jDð2; 1; 1Þ; Dð−2; −1; −1Þ; πð0; 0; 0Þi 12 A1 ⊕ E ⊕ T 1 ⊕ 2T −2

Oh CP2 jDð2; 1; 1Þ; Dð−2; −1; −1Þ; πð0; 0; 0Þi 12 Aþ þ

Oh C1 jDð3; 2; 1Þ; Dð−3; −2; −1Þ; πð0; 0; 0Þi 24 A1 ⊕ A2 ⊕ 2E ⊕ 3T 1 ⊕ 3T −2

Oh C1 jDð3; 2; 1Þ; Dð−3; −2; −1Þ; πð0; 0; 0Þi 24 Aþ þ þ

In this case, the dimensionality of the extended orbit, i.e.,

Oh CR2 jπ þ ðn1 Þ; π þ ðn2 Þ; K þ ð−n1 Þ; K þ ð−n2 Þi 24 Aþ þ þ þ þ −

Oh CR2 jπ þ ðn1 Þ; π þ ð−n1 Þ; π þ ðn2 Þ; π þ ð−n2 Þi 6 þ þ

An ordering for the 48 elements of Oh can be established by

Momentum GP Order Irreps Group parametrization rp Group elements

Rð ⃗ωÞ ¼ exp − ωk tk ðA6Þ 1

Schmidt orthonormalization, then solving for linear com-

the same vector r along all d indices), the polarization ðT þ

Γ, κ μ (0, 0, 1) (0, 1, 0) (1, 0, 0) (0, 0, −1) (0, −1, 0) (−1, 0, 0)

You might also like