Truncation uncertainties for accurate quantum

simulations of lattice gauge theories

Anthony N. Ciavarella 1 , Siddharth Hariprakash 1,2
, Jad C. Halimeh 3,4,5
, and
Christian W. Bauer 1,2

1 Physics Division, Lawrence Berkeley National Laboratory, Berkeley, California 94720, USA
arXiv:2508.00061v1 [quant-ph] 31 Jul 2025

2 Leinweber Institute for Theoretical Physics and Department of Physics, University of California, Berkeley, California 94720,
USA
3 Max Planck Institute of Quantum Optics, 85748 Garching, Germany
4 Department of Physics and Arnold Sommerfeld Center for Theoretical Physics (ASC), Ludwig Maximilian University of
Munich, 80333 Munich, Germany
5 MunichCenter for Quantum Science and Technology (MCQST), 80799 Munich, Germany
August 4, 2025

The encoding of lattice gauge theories onto quantum computers requires a discretization
of the gauge field’s Hilbert space on each link, which presents errors with respect to the
Kogut–Susskind limit. In the electric basis, Hilbert space fragmentation has recently been
shown to limit the excitation of large electric fields. Here, we leverage this to develop a
formalism for estimating the size of truncation errors in the electric basis. Generically, the
truncation error falls off as a factorial of the field truncation. Examples of this formalism
are applied to the Schwinger model and a pure U(1) lattice gauge theory. For reasonable
choices of parameters, we improve on previous error estimates by a factor of 10306 .

1 Introduction
Quantum computers offer the ability to directly simulate the real-time dynamics of quantum field
theories [1]. This is anticipated to enable the prediction of dynamical non-perturbative quantities
from quantum chromodynamics (QCD) [2–4]. As a particular example, quantum simulations of lattice
QCD will enable the computation of soft functions, relevant to understanding the production of QCD
jets from high-energy particle collisions [5, 6]. It will also enable computations of the quark-gluon
plasma viscosity [7, 8] and inelastic scattering amplitudes for hadron collisions [9–15]. In addition to
their relevance to high-energy physics, quantum simulations of lattice gauge theories are anticipated
to give insights into topological phases and quantum spin liquids, where they emerge as an effective
description of the dynamics [16–19]. They have also been shown to demonstrate multiple forms of
ergodicity-breaking behavior [20–23].
The large potential for quantum simulation and the recent development of noisy quantum computers
has led to a large amount of work exploring how to use these devices to perform simulations. Lattice
gauge theories with a number of different gauge groups in 1 + 1D have been simulated on quantum
computers [24–56]. Limited simulations have been performed in higher spatial dimensions [8, 57–68].
This is due to gauge fields having continuous degrees of freedom and quantum computers having finite,
Anthony N. Ciavarella : [email protected]
Siddharth Hariprakash : [email protected]
Jad C. Halimeh : [email protected]
Christian W. Bauer : [email protected]

1
discrete degrees of freedom. This challenge is not present in 1 + 1D since Gauss’s law can be used to
integrate out gauge fields, leaving only fermions, which have a finite Hilbert space. This leads to long-
range interactions; however, the range of this interaction can be truncated due to the exponential decay
of correlations in low-energy states [43]. Mapping the gauge fields in higher dimensions onto quantum
computers requires a truncation of the gauge fields. For the standard Kogut–Susskind Hamiltonian,
a number of approaches have been developed, including electric basis truncations [57, 59, 68–89],
discrete subgroups [90–99], and hybrid electric/magnetic bases [100–106]. Alternatives to the usual
Kogut–Susskind Hamiltonian have also been developed, such as quantum link models [107–112] and
orbifold lattice gauge theory [113–115], which are conjectured to have QCD as their continuum limit.
Ultimately, simulations capable of making predictions with precision will need to have controllable
uncertainties. This will have contributions from both the hardware and the theoretical formulation of
the simulation. The effects of a finite lattice spacing can be estimated using the same Symanzik action
approach used in traditional lattice QCD [116]. This formalism can also be used to understand the
error coming from discrete (possibly Trotterized) time steps [117, 118]. Recent work has developed
the formalism necessary for estimating finite volume errors in a generic correlation function computed
on a quantum computer [15]. The truncation of the gauge fields will also contribute to the theoretical
uncertainty. Several works have shown a fast convergence of energy eigenstates with field truncation
for a number of different theories [59, 70, 79, 119–126]. For scalar field theories, the exponential con-
vergence is a reflection of the Nyquist-Shannon sampling theorem [122], however a similar mechanism
for non-Abelian gauge theories has not yet been identified. Previous work has shown that truncation
errors in dynamics converge exponentially fast, provided one uses a truncation that grows linearly with
time [127]. However, it is essential to have tight estimates of the truncation error, as overestimating
errors will unnecessarily delay quantum simulations of scientific importance. In this work, the presence
of Hilbert space fragmentation in the Kogut–Susskind Hamiltonian is used to obtain estimates of the
truncation error for truncations in the electric basis. These error estimates do not require the trunca-
tion to grow with time to maintain accuracy and go to zero as a factorial in the size of the truncated
link Hilbert space. Numerical simulations are performed for the Schwinger model and a pure U(1)
lattice gauge theory on a plaquette ladder. These simulations show that the error estimates in this
work correctly capture the dynamics of lattice gauge theories.

2 Truncation Errors in Pure Gauge Theories

2.1 Eigenstate Truncation Errors
To motivate why improvements in truncation error are expected to be possible, traditional perturbation
theory will be applied to eigenstates of truncated Hamiltonians. We will consider a Hamiltonian, Ĥ,
defined on a single bosonic mode, (with basis states |0⟩ , |1⟩ , |2⟩ , · · · ) and assume that the terms in the
Hamiltonian do not increase the boson number by more than 1, i.e. ⟨n + k| Ĥ |n⟩ = 0 for k > 1. For
a lattice gauge theory, this would be analogous to a theory with a single plaquette and no matter. A
truncated Hamiltonian, ĤΛ , can be defined by removing the off-diagonal elements of Ĥ above some
truncation Λ. Explicitly, the relation between the truncated and untruncated Hamiltonian is given by
Ĥ = ĤΛ + V̂Λ , (1)
where ⟨n| Ĥ |n⟩ = ⟨n| ĤΛ |n⟩ ∀n, ⟨n| Ĥ |k⟩ = ⟨n| ĤΛ |k⟩ if both n, k ≤ Λ, and ⟨n| ĤΛ |k⟩ = 0 otherwise.
Denoting the eigenstates and energies of the full Hamiltonian by |n⟩ and En , and the eigenstates and
energies of the truncated Hamiltonian by ψnΛ and EnΛ , we have

|ψn ⟩ = ψnΛ + δψnΛ

  1
δψnΛ = 1 − ψnΛ ψnΛ V̂Λ |ψn ⟩
En − ĤΛ
En − EnΛ = ψnΛ V̂Λ δψnΛ , (2)

2
where ψnΛ δψnΛ = 0. Note that |ψn ⟩ is not normalized to 1, but when expanding perturbatively,
correcting this will only contribute at higher orders in perturbation theory. The leading order correction
to |ψn ⟩ is given by
⟨Λ + 1| V̂Λ |Λ⟩ Λ ψnΛ
δψnΛ = |Λ + 1⟩ . (3)
EnΛ − ⟨Λ + 1| ĤΛ |Λ + 1⟩
Note that this expression depends on the overlap of the truncated eigenstate with the largest boson
number kept in the truncated Hilbert space. To determine the scaling as the truncation is raised, this
overlap will be determined by expanding perturbatively around a lower truncation Λ0 . Keeping only
the leading terms in perturbation theory, the overlap is given by
Λ−1
Y ⟨k + 1| V̂Λ0 |k⟩
Λ ψnΛ = Λ0 ψnΛ0 . (4)
k=Λ0
EnΛ0 − ⟨k + 1| ĤΛ0 |k + 1⟩

Therefore, the leading corrections to the states and energies at a truncation of Λ are
Λ−1
⟨Λ + 1| V̂Λ |Λ⟩ Y ⟨k + 1| V̂Λ0 |k⟩
δψnΛ = |Λ + 1⟩ Λ0 ψnΛ0 (5)
EnΛ − ⟨Λ + 1| ĤΛ |Λ + 1⟩ k=Λ0
EnΛ0 − ⟨k + 1| ĤΛ0 |k + 1⟩
2
Λ−1
1 Y ⟨k + 1| V̂Λ0 |k⟩
En − EnΛ = Λ0 ψnΛ0 ⟨Λ| V̂Λ0 |Λ + 1⟩ .
EnΛ − ⟨Λ + 1| ĤΛ0 |Λ + 1⟩ k=Λ0
EnΛ0 − ⟨k + 1| ĤΛ0 |k + 1⟩

For these corrections to systematically decrease as the truncation is raised, it is necessary for the energy
of the eigenstate
 Q to be below ⟨Λ0| ĤΛ0 |Λ0 ⟩. For such low energy states, the size of the corrections will
⟨n+1|V̂Λ0 |n⟩
scale as O n<Λ ⟨n+1|Ĥ |n+1⟩ .
Λ0

For a U (1) lattice gauge theory on a single plaquette, V̂Λ will contain matrix elements of the
plaquette operator, and the diagonal matrix elements of Ĥ are given  by the electric energy. Therefore,
the truncation effects for low-lying states in this theory scale as O (Λ!)
1
2 . For a theory with a boson
√
creation operator and number operator in the Hamiltonian, V̂Λ matrix elements  as Λ, and the
 scale
diagonal entries of ĤΛ scale as Λ. This gives truncation effects that are O √1Λ! . Note that these
expressions were derived for a theory with a single bosonic mode, but it is expected that the general
scaling of the truncation corrections should hold for larger systems.

2.2 Time Dependent Truncation Errors with Fragmented States

2.2.1 Single Plaquettes
The results of the previous section show that truncation errors in eigenstates go to zero as a factorial
of the field truncation, but do not provide quantitative estimates of the truncation error. In this
section, it will be shown how time-dependent perturbation theory can be used to estimate truncation
errors quantitatively. In the authors’ previous work, it was demonstrated that the Kogut–Susskind
Hamiltonian generically demonstrates Hilbert space fragmentation [20]. This is due to the quadratic
nature of the electric energy terms, leading to large gaps in the spectrum for states with large electric
fields. This will allow us to apply a strong coupling expansion to the dynamics of states with these
large electric fields. In the following discussion, a single plaquette will be considered, allowing the
entire state of the system to be specified by the electric field on a single link. Note that while errors
in the time evolution will be estimated here, the results in this section can be used to estimate
contributions to eigenstates by using an adiabatic switching procedure, such as the one used in the
proof of the Gellmann-Low theorem [128]. The results of estimating truncation errors in such a manner

3
are consistent with the previous section. Explicitly, to study the evolution of a state |ϕ(t)⟩, define the
interaction picture state |ϕI (t)⟩ by

|ϕ(t)⟩ = e−iĤΛ t |ϕI (t)⟩ . (6)

The state |ϕI (t)⟩ satisfies

Z t
|ϕI (t)⟩ = |ϕ(0)⟩ − i ds eiĤΛ s V̂Λ e−iĤΛ s |ϕI (s)⟩ . (7)
0

By repeatedly substituting |ϕI (t)⟩ into this equation, one can generate a series for |ϕI (t)⟩ in powers of
V̂Λ . Using this result, the difference between the time evolution generated by the exact and truncated
Hamiltonian is given to leading order by
Z t
−iĤt −iĤΛ t −iĤΛ t
(e −e ) |ϕ(0)⟩ = −ie ds eiĤΛ s V̂Λ e−iĤΛ s |ϕ(0)⟩ (8)
0
Z t
= −ie−iĤΛ t |Λ + 1⟩ ds ei⟨Λ+1|ĤΛ |Λ+1⟩s ⟨Λ + 1| V̂Λ |Λ⟩ ⟨Λ| e−iĤΛ s |ϕ(0)⟩ .
0

To determine the size of this expression, the behavior of ⟨Λ| e−iĤΛ s |ϕ(0)⟩ needs to be understood. For
Λ large enough for fragmentation to occur, the phase of this term will be dominated by the diagonal
piece of the Hamiltonian so we can approximate ⟨Λ| e−iĤΛ s |ϕ(0)⟩ ≈ c e−i⟨Λ|ĤΛ |Λ⟩s where c is a slowly
2
varying function with |c| ≤ 1. With this approximation, we can evaluate this integral and find
2
(e−iĤt − e−iĤΛ t ) |ϕ(0)⟩ ≤ ⟨Λ + 1| V̂Λ |Λ⟩ . (9)
⟨Λ + 1| ĤΛ |Λ + 1⟩ − ⟨Λ| ĤΛ |Λ⟩

The above calculation is valid provided that we truncate above the first electric field strength, Λ0 ,
where HSF occurs. However, in practice, we will likely truncate at some Λ > Λ0 . In this case, time-
dependent perturbation theory can be used to compute the leading contribution to ⟨Λ| e−iĤΛ s |ϕ(0)⟩.
Using this, the leading correction to the truncated time evolution is
Z t Z t0
(e−iĤt − e−iĤΛ t ) |ϕ(0)⟩ = −ie−iĤΛ t dt0 eiĤΛ t0 V̂Λ e−iĤΛ t0 (−i)Λ−Λ0 dt1 eiĤΛ0 t1 V̂Λ0 e−iĤΛ0 t1
0 0
Z t1 Z tΛ−Λ +1
0
× dt2 eiĤΛ0 t2 V̂Λ0 e−iĤΛ0 t2 · · · dtΛ−Λ0 eiĤΛ0 tΛ−Λ0 V̂Λ0 e−iĤΛ0 tΛ−Λ0 |ϕ(0)⟩ . (10)
0 0

This expression consists of integrals of exponentials with phases set by the change in energy from
applying V̂Λ or V̂Λ0 . These integrals can be explicitly evaluated as a function of t, using the fact that
the operator V̂Λ0 only acts on bosonic states |n⟩ with n ≥ Λ0 and for those states the Hamiltonian
ĤΛ0 is diagonal. One finds
 
Λ  Λ+1
X  −i⟨Λ+1|Ĥ |Λ+1⟩t Y 1
(e−iĤt − e−iĤΛ t ) |ϕ(0)⟩ = i −i⟨k|HΛ |k⟩t

 e Λ
− e 
 ⟨k|HΛ |k⟩ − ⟨l|HΛ |l⟩ 
k=Λ0 l=Λ0
l̸=k
 
Λ
Y
× ⟨k + 1|V̂Λ |k⟩ . (11)
k=Λ0

These bounds will be independent of what the initial state was, provided that the initial state only
has support on basis states with electric fields below Λ0 . In practice, one needs to be able to estimate

4
Λ0 . A rough way of doing this is to pick Λ0 so that the leading order contribution to the leakage error
is smaller than 1.
To make this discussion concrete, we will apply these results to estimate errors in the truncated
simulation of a single plaquette in a U (1) lattice gauge theory. The electric basis states are given by
|n⟩ where n is an integer, and the Hamiltonian is given by
1  ˆ†
ˆ −□

Ĥ = 2g 2 Ê 2 + 2 − □ (12)
2g 2
∞
X
Ê = n |n⟩ ⟨n| (13)
n=−∞
X∞
ˆ =
□ |n⟩ ⟨n + 1| . (14)
n=−∞

We can split this up as

Ĥ = ĤΛ + V̂Λ , (15)

with
1  ˆ†
ˆΛ − □

ĤΛ = 2g 2 Ê 2 + 2
2−□ Λ
2g
Λ
X
ˆΛ =
□ |n⟩ ⟨n + 1|
n=−Λ
1 X
V̂Λ = − |n⟩ ⟨n + 1| + h.c. . (16)
2g 2
|n|≥Λ

This implies that for k ≥ Λ0 we have

1
⟨k + 1|V̂Λ0 |k⟩ = −
2g 2
⟨k|ĤΛ0 |k⟩ = 2g 2 k 2 . (17)

We therefore find
 
 Λ
2(Λ+1−Λ0 ) X  Λ+1
1  −i2g2 (Λ+1)2 t −i2g 2 k2 t
Y 1 
(e−iĤt − e−iĤΛ t ) |ϕ(0)⟩ =i(−1)Λ+1−Λ0 2
 e − e 
2g  k − l2 
2
k=Λ0 l=Λ0
l̸=k

× (cΛ0 |Λ + 1⟩ + c−Λ0 |−Λ − 1⟩) , (18)

ˆ
where c±Λ0 are the slowly varying norms of ⟨±Λ| e−iHΛ tΛ0 |ϕ(0)⟩. Defining the leakage amplitude by
 
Λ  Λ+1
X  −i2g2 (Λ+1)2 t −i2g 2 k2 t
Y 1 
L(g, Λ, Λ0 , T ) = maxt<T  e − e  , (19)
 k 2 − l2 
k=Λ0 l=Λ0
l̸=k

the leading error in the truncated state is at most

 Λ−Λ0
−iĤt −iĤΛ t 1
(e −e ) |ϕ(0)⟩ ≤ 2L(g, Λ, Λ0 , T ) . (20)
2g 2

5
 101

Max Error in E2
10 1

10 3

10 5

10 7
|0
10 9
|1
|2
10 11
Strong Coupling Estimate
2 4 6 8 10

Figure 1: Maximum error in the expectation of Ê 2 as a function of time on a single plaquette with g = 0.5 and a
max evolution time of T = 30. The blue, green, and purple lines correspond to using different electric basis states
as the initial state. The red curve is the error bound in Eq. (22), computed using Λ0 = 4.

To get an idea of how this scales, we can place an upper bound on L(g, Λ, Λ0 , T ). The integral in the
definition of this function can be exactly evaluated to be a sum of exponentials multiplied by different
 Λ−Λ0
(2Λ0 −1)!!
prefactors. There are at most 2Λ−Λ0 such exponentials and the largest prefactor is 2g12 (2Λ−1)!!
where n!! = n × (n − 2) × (n − 4) × · · · × 1. This lets us upper bound L(g, Λ, Λ0 , T ) by
 Λ−Λ0
1 (2Λ0 − 1)!!
L(g, Λ, Λ0 , T ) ≤ . (21)
g2 (2Λ − 1)!!

Note that this is a loose bound as it neglects destructive interference between many fast oscillating
phases in the actual value of the integral. Regardless, this predicts that the leading error due to
truncation converges as a factorial and is time-independent.
To understand the performance of these error estimates, we apply these techniques to estimate the
error in the expectation of the electric energy Ê 2 . Using the leading correction to the states, it can be
seen that the leading error in the expectation of the electric energy is
 Λ−Λ0
1
⟨ϕ| eiĤt Ê 2 e−iĤt |ϕ⟩ − ⟨ϕ| eiĤΛ t Ê 2 e−iĤΛ t |ϕ⟩ ≤ 2(Λ + 1)2 L(g, Λ, Λ0 , T )2 (22)
4g 4

We can now compare this bound against the numerical calculation of this expectation values. Ê 2
was measured as a function of time for a single plaquette with g = 0.5 and maximum evolution time
of T = 30. The evolution with Λ = 20 was treated as the exact evolution and was compared to
the evolution with lower truncations. The maximum error in the expectation of Ê 2 achieved when
beginning in different electric basis states, |n⟩, is shown in Fig. 1. As this figure shows, the bound on
the error from Eq. (22) correctly upper bounds the actual error for all of these initial states. Note that
the bound becomes tighter for initial states with larger electric energies.

6
2.2.2 Comparison to Previous Work
Previous work has derived resource estimates for bosonic theories through a combination of rigorous
upper bounds on the Dyson series expansion in an interaction picture and constraints from energy
conservation [127]. To understand how the current work compares to these results, we will study the
analytic bounds of this previous work for the evolution of the electric vacuum state on a single plaquette
with a U (1) gauge field. Note that these techniques apply to larger system sizes, but for comparison,
we will restrict to a single plaquette. The previous bounds on truncation errors were derived by placing
upper bounds on the expectation of Π̂n = |n⟩ ⟨n| as a function of time. These leakage bounds were
then turned into a bound on error in the time evolution operator due to truncation.
One starts from the expectation value of the full Hamiltonian in the electric vacuum |0⟩, which by
energy conservation has to be time invariant. Working again with a U(1) pure gauge theory, the total
energy is always greater than the electric energy; one therefore obtains
1
⟨0| Ĥ |0⟩ = ⟨0| eiĤt Ĥe−iĤt |0⟩ = . (23)
2g 2
Since both the electric and magnetic term in the Hamiltonian are positive definite, we can bound the
total energy from above by the electric energy

⟨0| eiĤt Ĥe−iĤt |0⟩ ≥ ⟨0| eiĤt ĤE e−iĤt |0⟩ = 2g 2 ⟨0| eiĤt Ê 2 e−iĤt |0⟩ , (24)

Using the expression of the electric operator in Eq. (13), we can bound the electric energy from above
by the contributions of a single bosonic state |Λ⟩ such giving

⟨0| eiĤt Ê 2 e−iĤt |0⟩ ≥ Λ2 ⟨0| eiĤt Π̂Λ e−iĤt |0⟩ . (25)

Combining everything together, we obtain the bound

1
≥ 2g 2 Λ2 ⟨0| eiĤt Π̂Λ e−iĤt |0⟩ , (26)
2g 2
or
1
⟨0| eiĤt Π̂Λ e−iĤt |0⟩ ≤ , (27)
g 4 Λ2
for all times t. Note that for a larger lattice, similar volume-independent bounds can be derived for
translationally invariant states.
For short times, a time-dependent bound can be derived by upper-bounding the error in a Dyson
series expansion. Explicitly, one works in the interaction picture where the free part of the Hamiltonian
is given by 2g 2 Ê 2 . The Hamiltonian in the interaction picture is then given by
1 X i2g2 t(2n+1)
ĤI (t) = − e |n + 1⟩ ⟨n| + h.c. . (28)
2g 2 n

Denoting the electric vacuum evolved in the interaction picture by |ϕI (t)⟩, the expectation of the
projector Π̂Λ is given by
2
2
⟨0| eiĤt Π̂Λ e−iĤt |0⟩ = ⟨Λ| e−iĤt |0⟩ = |⟨Λ|ϕI (t)⟩| . (29)

The Dyson series expansion is generated by inserting |ϕI (t)⟩ recursively into the equation
Z T
|ϕI (T )⟩ = |ϕI (0)⟩ − i dt ĤI (t) |ϕI (t)⟩ . (30)
0

7
The quantity ⟨Λ|ϕI (t)⟩ only receives a non-zero contribution from the Λ-th order which is given by
Z t Z t1 Z tΛ−1
⟨Λ|ϕI (t)⟩ = (−i)Λ dt1 dt2 · · · dtΛ ĤI (t1 )ĤI (t2 ) · · · ĤI (tΛ ) |ϕI (tΛ )⟩ . (31)
0 0 0

Since the spectral norm of HI (t) is bounded by

1
||HI (t)|| ≤ , (32)
g2
the magnitude of this overlap is upper bounded by
Z t t1 tΛ−1
tΛ
Z Z
Λ
|⟨Λ|ϕI (t)⟩| ≤ ||HI (t)|| dt1 dt2 · · · dtΛ = . (33)
0 0 0 g 2Λ Λ!
We therefore find
t2Λ
⟨0| eiĤt Π̂Λ e−iĤt |0⟩ ≤ . (34)
g 4Λ Λ!2
Similar techniques can be used to show

t∆
Π̂Λ e−iĤt Π̂Λ−∆ ≤ . (35)
g 2∆ ∆!
These results are referred to as short-time bounds, as at long times they will exceed the bound derived
from energy conservation.
Time dependent bounds that are valid for longer times were also derived by inserting intermediate
projectors and bounding individual terms in the sum. For example, one can consider adding one
intermediate projector, defining
X
Π̂>Λ = |n⟩ ⟨n|
|n|>Λ
X
Π̂≤Λ = |n⟩ ⟨n| . (36)
|n|≤Λ

This allows to derive

 
Π̂Λ e−iĤt Π̂Λ−∆1 −∆2 = Π̂Λ e−iĤ(t−t1 ) Π̂>Λ−∆1 + Π̂≤Λ−∆1 e−iĤt1 Π̂Λ−∆1 −∆2

≤ Π̂>Λ−∆1 e−iĤt1 Π̂Λ−∆1 −∆2 + Π̂Λ e−iĤ(t−t1 ) Π̂≤Λ−∆1

t∆
1
2
(t − t1 )∆1
≤ + , (37)
g 2∆2 ∆2 ! g 2∆1 ∆1 !
and t1 can be chosen to minimize this expression. To obtain a long-term bound for the expectation
value ⟨0| eiĤt Π̂Λ e−iĤt |0⟩ one can insert up to Λ projectors and optimize over the intermediate times.
These are referred to as long-time bounds as they take longer to saturate the bound from energy
conservation.
In summary, the work of [127] derived three types of bounds, namely time-independent bounds
based on energy conservation, as well as time-dependent bounds valid at short and longer times.
These can be combined into a more optimal time-dependent bound, by taking their minimum. To
determine the tightness of the resulting bound, the time evolution of the electric vacuum state will be
simulated numerically. Note that
2
⟨0| eiĤt Π̂Λ e−iĤt |0⟩ ≤ Π̂Λ e−iĤt Π̂0 , (38)

8
 10 1
10 1
iHt|0
10 3

10 3
10 5
0|eiHt e

10 7
10 5
10 9

=1
10 7 =2 10 11
Max of
=3 Energy Bound
=4 10 13
Max of Perturbative Ev
0 5 10 15 20 25 30 2 4 6 8 10
t
Figure 2: Evolution of the electric vacuum state on a single plaquette. Numerical simulations were performed with
a maximum electric field of 20 and g = 0.5. The left panel shows the expectation of Π̂Λ for various values of Λ as
a function of time t. The solid curves are the exact time evolution, and the dashed lines are the rigorous long-time
bounds. The right panel shows the maximum of the expectation of Π̂Λ for the simulated time evolution. The blue
curve is the exact result, the green curve is the bound from energy conservation, and the red curve is the leading
contribution to the expectation obtained by calculating L(g, Λ, Λ0 , T ) with Λ0 = 4.

so these bounds directly translate to bounds on the expectation of operators. The electric vacuum was
evolved under a Hamiltonian with a truncation of Λ = 20 and g = 0.5. The solid lines in the left panel
of Fig. 2 show the exact time evolution of the expectation of Π̂Λ and the dashed lines show the bounds
as derived above (the minimum of the three approaches discussed). As this figure shows, the long-time
bounds quickly saturate and overestimate the expectation of Π̂Λ . The figure on the right shows the
behavior of the maximum of the expectation value of the projection operator taken over large times,
as a function of the truncation Λ. In blue we show the result from the numerical simulation, which
shows that the maximum expectation for Λ = 10 is below 10−13 . The green curve is the result of [127]
derived in this section, which clearly overestimates the true value by many orders of magnitude. We
also show in red the new results of this work, which were derived in the previous section. One can see
that these results provide a much tighter bound.
As discussed in [127], the bounds derived in that work imply that to guarantee ⟨k + Λ| e−iĤt |k⟩ <
√
0.01 for g = 3 for a maximum evolution time of t = 8 would require a cutoff satisfying Λ > 100. On
the other hand, using the upper bound derived in this work for L(g, Λ, Λ0 , T ), given in Eq. (21), one
estimates that ⟨k + Λ| e−iĤt |k⟩ < 6 × 10−308 for these parameters. This shows in a pretty dramatic
fashion how much tighter the bounds derived in this work are. To verify this claim
√ numerically, the
evolution of the electric vacuum on a single plaquette was simulated for g = 3 and a maximum
evolution time of t = 8. Fig. 3 shows the maximum value reached by the expectation of Π̂Λ during this
time evolution and the predicted maximum for the perturbative calculation done in this work. As this
figure shows, the predicted maximum is in good agreement with the numerical simulation, indicating
the tightness of the bounds derived in this work.
The bounds obtained in Ref [127] are extremely loose due to the bounds being computed by
applying the triangle inequality to upper-bound the Dyson series expansion. The Dyson series consists
of a sum over many oscillating phases, and the actual physics of the system comes from how these
phases constructively or destructively interfere. As the triangle inequality removes these phases, it
should not be surprising that these bounds fail to come remotely close to the qualitative behavior
of the system. The estimate for the maximum of Π̂Λ obtained using the strong coupling expansion
appears to fall off at the same rate as the actual system, but is still a loose overestimate. This is

9
 10 3
Max of
10 10 Max of Perturbative Ev

10 17

10 24

10 31

10 38

10 45

10 52

2 4 6 8 10

Figure 3: Evolution of the electric vacuum

√ state on a single plaquette. Numerical simulations were performed with
a maximum electric field of 20, g = 3, and a maximum evolution time of t = 8. The blue curve is the maximum
of the expectation of Π̂Λ during this evolution. The green points are the leading contribution to the expectation
obtained by calculating L(g, Λ, Λ0 , T ) with Λ0 = 0.

because the strong coupling expansion cannot be applied to states with small electric fields, and the
amplitude for leakage into the fragmented sector of the theory was upper-bounded by 1.

2.3 Extension to Larger Systems

To extract new predictions from simulations of lattice gauge theories on quantum computers, it is
necessary to simulate lattices with more than one plaquette and have estimates of the errors due to
truncations of the gauge fields. This can be done using similar techniques to the single plaquette case.
We will write the Hamiltonian for the untruncated theory, Ĥ, as a sum of the truncated Hamiltonian,
ĤΛ , plus a sum over V̂Λ⃗x which contains all off-diagonal couplings for electric fields above the truncation
Λ for the plaquette at position ⃗x, explicitly
X
Ĥ = ĤΛ + V̂Λ⃗x . (39)
x

Following the same approach as Section 2.2.1, the leading correction to the truncated evolution is given
by Z t
X
(e−iĤt − e−iĤΛ t ) |ϕ(0)⟩ = −ie−iĤΛ t ds eiĤΛ s V̂Λ⃗x e−iĤΛ s |ϕ(0)⟩ . (40)
⃗
x 0

The quantity V̂Λ⃗x e−iĤΛ s |ϕ(0)⟩ can be estimated to leading order by applying time dependent pertur-
bation theory in the fragmented subspace as before to give at leading order
XZ t Z t0
(e−iĤt − e−iĤΛ t ) |ϕ(0)⟩ = −ie−iĤΛ t dt0 eiĤΛ t0 V̂Λ⃗x e−iĤΛ t0 (−i)Λ−Λ0 dt1 eiĤΛ0 t1 V̂Λ⃗x0 e−iĤΛ0 t1
⃗
x 0 0
Z t1 Z tΛ−Λ0 +1
× dt2 eiĤΛ0 t2 V̂Λ⃗x0 e−iĤΛ0 t2 · · · dtΛ−Λ0 eiĤΛ0 tΛ−Λ0 V̂Λ⃗x0 e−iĤΛ0 tΛ−Λ0 |ϕ(0)⟩ . (41)
0 0

10
As before, the integrand in this expression consists of phases given by changes in energy from applying
V̂Λ⃗x0 or V̂Λ⃗x . These changes in energy can be approximated by the change in electric energy. This
gives the leading corrections to the state vector and can be used to evaluate the leading error in the
expectation of observables. Note that at leading order, local observables will only have errors coming
from the V̂Λ⃗x terms that share support with the observable.
To make this discussion concrete, the truncation error on the expectation value of an operator will
be estimated for a pure U (1) lattice gauge theory on a plaquette chain. The Hamiltonian is given by
X g2 X 1  
Ĥ = Êl2 + 2 − □ ˆ †p ,
ˆp − □ (42)
2 p
2g 2
l

where the sum over l corresponds to summing over all links on the lattice and the sum over p cor-
responds to summing over all plaquettes on the lattice. This Hamiltonian can be gauge-fixed so
gauge-invariant states are given by specifying the electric field on a single plaquette per link [20]. The
gauge-fixed Hamiltonian is given by
X  1 2  1  
2 2 ˆ ˆ †
Ĥ = g Êp + Êp − Êp+1 + 2 2 − □p − □p
p
2 2g
∞
X
Êp = np |np ⟩ ⟨np |
np =−∞
∞
X
ˆp =
□ |np ⟩ ⟨np + 1| , (43)
np =−∞

and for a chain of length L the electric basis states are given by |n1 , n2 , · · · nL ⟩. The operator we
will choose for this example will be the electric energy on the four links of a given plaquette, and
for concreteness we choose the time-dependent expectation value of Êp2 . This operator will receive
contributions from plaquette p, as well as the neighboring plaquettes p ± 1. This implies that the
truncation errors can be estimated using a 3 plaquette chain, with a generic state given by |n1 , n2 , n3 ⟩.
To simplify the notation in the calculation, only the error from the truncation on positive electric fields
will be estimated (the contribution from negative electric fields will only contribute a factor of 2).
To start, we write as before
X
|ϕ(0)⟩ = cn1 ,n2 ,n3 |n1 , n2 , n3 ⟩ . (44)
n1 ,n2 ,n3 ≤Λ0

Following similar steps as before, the time-evolved perturbation can be written as

(2) −1 X 2
eiĤΛ0 t V̂Λ0 e−iĤΛ0 t |ϕ(0)⟩ = cn ,Λ ,n eitg (4Λ0 −n1 −n3 +2) |n1 , Λ0 + 1, n3 ⟩ . (45)
2g 2 n ,n 1 0 3
1 3

Defining the function

 Λ Z
Λ−Λ0 Y tn+1
−i 2
(4Λ0 −n−m+2)tn
An,m (g, Λ, Λ0 , T ) = dtn eig , (46)
2g 2 0
n=Λ0

with T ≡ tΛ+1 the correction to the state vector is therefore given by

X
(e−iĤt − e−iĤΛ t ) |ϕ(0)⟩ = An,m (g, Λ, Λ0 , t)cn,Λ0 ,m |n, Λ + 1, m⟩ . (47)
n,m

11
 100
0.8
10 2

Error in E2
0.6
10 4
E2

0.4
=5 10 6
0.2 =4
=3
=2 10 8

0.0 =1
0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 6 7 8
t t
Figure 4: Simulation of an infinite plaquette chain with g = 0.8. The left panel shows the electric energy as a
function of time for different truncations of the electric field. The solid curves in the right panel show the difference
in electric energy between the Λ = 5 truncation and the lower truncations. The dashed lines show the predicted
error in the electric energy from truncating the Hamiltonian, computed using Eq. (50) with Λ0 = 1. See Appendix A
for details regarding the iMPS implementation.

Combining this, we obtain the correction to the expectation of the electric energy on the center
plaquette (Ê22 ),

⟨ϕ| eiĤt Ê22 e−iĤt |ϕ⟩ − ⟨ϕ| eiĤΛ t Ê22 e−iĤΛ t |ϕ⟩
X h  i
2 2 2 2
= |An,m (g, Λ, Λ0 , t)| (Λ + 1)2 |cn,Λ0 ,m | + m2 |cΛ0 ,m,n | + |cn,m,Λ0 | . (48)
n,m≤Λ0

2
Using the fact that |cn1 ,n2 ,n3 | = 1, the error can be upper bounded by
P
n1 ,n2 ,n3 ≤Λ0
n o
2
⟨ϕ| eiĤt Ê22 e−iĤt |ϕ⟩ − ⟨ϕ| eiĤΛ t Ê22 e−iĤΛ t |ϕ⟩ ≤maxτ <t (Λ + 1)2 |An,m (g, Λ, Λ0 , τ )| : n, m ≤ Λ0
n o
2
+ 2maxτ <t n2 |An,m (g, Λ, Λ0 , τ )| : n, m ≤ Λ0 .
(49)

The above expression is only for truncating the positive electric fields with strength above Λ. When
truncating the positive and negative electric fields with strength above Λ, the resulting expression for
the leading error in the electric field evolution is given by
n o
2
⟨ϕ| eiĤt Ê22 e−iĤt |ϕ⟩ − ⟨ϕ| eiĤΛ t Ê22 e−iĤΛ t |ϕ⟩ ≤2maxτ <t (Λ + 1)2 |An,m (g, Λ, Λ0 , τ )| : |n|, |m| ≤ Λ0
n o
2
+ 4maxτ <t n2 |An,m (g, Λ, Λ0 , τ )| : |n|, |m| ≤ Λ0 .
(50)

Note that while the expression was derived using a three-plaquette lattice, the result will generalize to
a plaquette chain of arbitrary length. The only information used about the initial state was that it did
not have support on electric fields above Λ0 . This bound could potentially be made tighter by using
energy conservation arguments to upper bound the probability of having a state with an electric field of
Λ0 present. For translationally invariant states, these energy bounds will have no volume dependence.
To demonstrate the performance of these error estimates, an infinite MPS was used to simulate
the time evolution of an infinite plaquette chain with g = 0.8 [129]. The system was initialized in the

12
electric vacuum and evolved until t = 8. The details of the implementation are in Appendix A. Fig. 4
shows the expectation of Ê 2 for a single plaquette on the lattice as a function of time. The truncation
error in the time evolution is consistent with the estimate of the error computed in Eq. (50). This
agreement suggests that truncation errors in local observables do not have a system size dependence
as predicted from the perturbative calculation.
The probability of exciting a given link on the lattice to a large electric field on a plaquette chain
can be estimated using this framework. Denoting Π̂Λ,⃗x = |Λ⃗x ⟩ ⟨Λ⃗x |, the probability of exciting a link
to electric field Λ when the initial state, |ϕ⟩ has all electric fields below Λ0 is upper bounded by
n o
2
⟨ϕ| eiĤt Π̂Λ,⃗x e−iĤt |ϕ⟩ ≤ P (Λ, Λ0 , g, t) = maxτ <t |An,m (g, Λ, Λ0 , τ )| : |n|, |m| ≤ Λ0 (51)

provided that HSF is present for states with electric fields above Λ0 . From Eq. (46), it follows that
P (Λ, Λ0 , g, t) is upper bounded by
1
P (Λ, Λ0 , g, t) ≤ × f (Λ, Λ0 ) , (52)
g 8(Λ−Λ0 )

where f (Λ, Λ0 ) is a g independent function. Fig. 5 shows a numerical demonstration of the validity of
this upper bound. Using an infinite MPS we again simulate the time evolution of an infinite plaquette
chain with four different values of g = 1.0, 0.9, 0.8, and 0.7. Starting from the electric vacuum, we
perform time evolution until t = 8 and compute the expectation value ⟨Π̂Λ,⃗x ⟩ as a function of time.
We observe that our numerical results are consistent with Eq. (52). For further implementation details
pertaining to the time evolution see Appendix A. Fig. 5 also shows a comparison between the bound
shown in Eq. (52) and the time independent energy conservation based bound shown in Eq. (27). In
particular, we show that for the largest value of Λ considered for each value of g, the bound based on
energy conservation is between 5 and 8 orders of magnitude larger than the bound based on Eq. (52).

13
 100
g=1.0 g=0.9
10 1

10 2
10 3
, xe iHt|0

10 4
10 5

10 6 10 7

10 9
10 8

10 11
0|eiHt

10 10
=1 10 13 =2
=2 =3
10 12
=3 10 15 =4
=4 =5
10 14 10 17
0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 6 7 8

100
g=0.8 g=0.7
10 1
10 2
10 3
, xe iHt|0

10 4
10 5

10 6
10 7

10 8
10 9
0|eiHt

10 10
10 11
=2 =3
=3 =4
10 12
=4 10 13
=5
=5 =6
10 14 10 15
0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 6 7 8
t t

Figure 5: Simulation of an infinite plaquette chain with four different values of g = 1.0, 0.9, 0.8, and 0.7 and
maximum electric fields of 5, 6, 6, and 7 respectively. The solid curves in each panel show the expectation value of
the projection operator Π̂Λ,⃗x with respect to the states obtained by time evolving the electric vacuum for various
choices of Λ. See Appendix A for details regarding the iMPS time evolution. The dashed lines in each panel
correspond to the upper bound on the probability given by Eq. (52) with Λ0 = 0, 1, 1, and 2 respectively for each
choice of g. The dash-dotted lines in each panel show the upper bound based on energy conservation, computed
using Eq. (27) for the largest value of Λ considered for each value of g.

14
3 Truncated Schwinger Model
Simulations of lattice QCD will require including quarks in addition to the gauge theory degrees of
freedom. To understand how the inclusion of matter affects the truncation error, we will apply our
framework to the Schwinger model. The Hamiltonian is given by

Ĥ = ĤK + Ĥm + ĤE

1X †
ĤK = ψ̂ Ûx ψ̂x + h.c.
2 x x+1
X
Ĥm = m (−1)x ψ̂x† ψ̂x
x
g2 X 2
ĤE = Ê , (53)
2 x x

where ψ̂x is the fermion field at site x, Ûx is the parallel transporter between sites x and x + 1, and
Êx is the corresponding electric field. Note that in this theory, all gauge fields are sourced by fermions
somewhere on the lattice. To have an electric field of strength Λ, there needs to be at least Λ positively
charged particles and Λ negatively charged particles. This means that to raise the maximum electric
field in a state by ∆, ∆ pairs of particles and anti-particles must be created. Furthermore, one can
have at most one fermion on each lattice site, such that all Λ fermions need to sit on one side of the link
with electric field Λ, while the anti-fermions sit on the other side. Starting from the electric vacuum,
creating a single electric field requires a single hopping term. Creating an additional electric flux on
this link requires first moving the fermion and anti-fermion created in the first step by two lattice units
(because of staggering) to the left and right, requiring 2 × 2 = 4 more hopping terms, and then adding
an additional fermion anti-fermion pair. This requires a total of 6 hopping terms. Creating 3 units of
electric flux requires moving the two fermion anti-fermion pairs once more, requiring 4 × 2 = 8 before
adding the extra pair, giving 15 hopping terms. In general, to create n units of flux requires at least
n(2n − 1) hopping terms. This complicates the perturbative calculation of the leakage probabilities
compared to the pure gauge case, but should also lead to a stronger suppression of states with large
electric fields. To perform the perturbative estimate, the truncated Hamiltonian ĤΛ is related to the
untruncated Hamiltonian by
X
Ĥ = ĤΛ + V̂Λx
x
1 †
V̂Λx = ψ̂x+1 ÛxΛ ψ̂x + h.c.
2
X X
ÛxΛ = |n − 1⟩ ⟨n| + |n − 1⟩ ⟨n| . (54)
n>Λ n≤−Λ

Following the same approach as the previous sections, the leading correction to the truncated time
evolution is given by
X Z t
−iĤt −iĤΛ t −iĤΛ t
(e −e ) |ϕ(0)⟩ = −ie ds eiĤΛ s V̂Λx e−iĤΛ s |ϕ(0)⟩ . (55)
x 0

Assuming that Λ is chosen to be the electric field strength where HSF begins, the leakage amplitude
to go from an electric field of Λ to Λ + 1 can be estimated using a lattice with two staggered sites
(one physical site), open boundary conditions, and an external electric field. For a system of this size,
the gauge invariant states can be specified by basis states |E, f0 , f1 ⟩ where E is an external electric
field and fi are the fermion occupation numbers on each site. f0 = 1 corresponds to a quark being
present on site 0 and f1 = 0 corresponds to an anti-quark being present on site 1. The leading order

15
contribution to the truncation error is
2
Z t
−iĤt −iĤΛ t −it( g2 (Λ+1)2 +2m) ′ ′
(e −e ) |ϕ(0)⟩ = −ie |Λ + 1, 1, 0⟩ ⟨Λ + 1, 1, 0| dt′ eiĤΛ t V̂Λ e−iĤΛ t |ϕ(0)⟩
0
2
−it( g2 (Λ+1)2 +2m) t
−ie
Z
g2
= |Λ + 1, 1, 0⟩ dt′ cΛ,0,1 eit(2m+ 2 (2Λ+1))
(56)
2 0

where cΛ,0,1 is defined analogously to the previous sections. This quantity is upper bounded by

1
(e−iĤt − e−iĤΛ t ) |ϕ(0)⟩ ≤ . (57)
2m + g 2 (Λ + 1/2)

Note that the amplitude to increase the electric field is suppressed when either m is large or g 2 Λ is
large. Increasing the electric field by two requires 6 hopping terms acting on 6 staggered sites. As
before, the state of a lattice with 6 staggered sites is given by an external electric field and the fermion
occupation number on 6 sites. Therefore, if HSF begins at Λ − 1, the leading error in the state on a
lattice with 6 staggered sites is given by the expression

(e−iĤt −e−iĤΛ t ) |ϕ(0)⟩

6 Z
!
Y tk−1
= 14(−i)6 e−iĤΛ t dtk V̂Λ3 (t1 )V̂Λ−1
4 2
(t2 )V̂Λ−1 5
(t3 )V̂Λ−1 3
(t4 )V̂Λ−1 1
(t5 )V̂Λ−1 (t6 ) |ϕ(0)⟩
k=1 0
6 Z
!
Y tk−1
= −14 |Λ − 1, 1, 1, 1, 0, 0, 0⟩ dtk
k=1 0

× ⟨Λ − 1, 1, 1, 1, 0, 0, 0| V̂Λ3 (t1 )V̂Λ−1

4 2
(t2 )V̂Λ−1 5
(t3 )V̂Λ−1 3
(t4 )V̂Λ−1 1
(t5 )V̂Λ−1 (t6 ) |ϕ(0)⟩
6
!
Z t
Y k−1
1 it1 (2m+2g2 (Λ+1)) it2 (−2m+g2 (Λ+ 1 ))
= −14 |Λ − 1, 1, 1, 1, 0, 0, 0⟩ dtk e e 2

0 26
k=1
it3 (−2m+g 2 (Λ+ 12 )) it4 (2m+g 2 (Λ− 12 )) it5 (2m+g 2 (Λ− 12 )) it6 (2m+g 2 (Λ− 12 ))
×e e e e cΛ−1,0,1,0,1,0,1
= A(Λ, g, m, t)cΛ−1,0,1,0,1,0,1 |Λ − 1, 1, 1, 1, 0, 0, 0⟩ (58)

where V̂Λx (t) = eiĤΛ t V̂Λx0 e−iĤΛ t , t0 = t, and A(Λ, g, m, t) has been defined to be the integral in the
above expression. The factor of 14 comes from counting the number of possible orderings of operators
that can be applied to raise the electric field on the center link by two. One possible ordering of
operators to raise the field truncation is shown in Fig. 6. From this expression, it can be seen that
the leading contribution to the error scales as 1/(g 2 Λ)6 . This is a much more rapid falloff than in
the equivalent scenario in the pure gauge theory. In high spatial dimensions with both gauge fields
and matter, the electric field on a link can be raised by applying either plaquette or link operators.
The leading terms in the truncation error will contain at most one hopping operator, and the rest
will be plaquette operators. Consequently, the truncation errors in the 1 + 1D Schwinger model will
not be representative of the behavior of truncation errors in higher spatial dimensions. However, the
more rapid convergence of truncation errors in 1 + 1D suggests that non-trivial simulations of 1 + 1D
dynamics with controlled uncertainties may be within near-term reach.
As in the pure gauge case, these estimates of leakage errors can be used to determine the truncation
error of local observables. The error in a P local observable will be given by the expectation of the
observable in the state created by applying x V̂Λx to the truncated state weighted by the probability
of reaching that state. If the expression in Eq. (57) is used to estimate the leakage probability, the

16
Figure 6: One possible way of increasing electric field strength by 2. The green circles correspond to positive charges,
and the red circles correspond to negative charges. First, the hopping operators are applied to every other link to
create three q q̄ pairs. The other hopping operators are then applied to remove the q q̄ pairs in the center of the
lattice. In the last step, a q q̄ pair is excited in the center of the lattice.

error in the electric energy (Êl2 ) and chiral condensate (χ̂x = (−1)x ψ̂x† ψ̂x ) is given by
 2
2 2 2 1
δE = (2Λ + (Λ + 1) ) × 2
2m + g (Λ + 1/2)
 2
1
δχ = 2 × . (59)
2m + g 2 (Λ + 1/2)

Note that the electric energy receives a contribution from the link itself being excited to Λ + 1 and
when its neighbor is excited to Λ + 1. For higher truncations, HSF will occur at electric field strengths
below Λ, and the expression in Eq. (58) can be used to estimate the error. Defining

M (Λ, g, m, t) = maxτ <t |A(Λ, g, m, τ )| , (60)

the error in the electric energy and chiral condensate is given by

δE 2 = 2(Λ − 1)2 + 4Λ2 + (Λ + 1)2 M 2 (Λ, g, m, t)

δχ = 4M 2 (Λ, g, m, t) , (61)

where A(Λ, g, m, τ ) is defined in Eq. (58).

To demonstrate the performance of these error estimates, the dynamics of the Schwinger model
were simulated using infinite MPS with g = 0.8 and m = 0.1. The details of the implementation
are in Appendix A. The system was initialized in the electric vacuum state and evolved in time until
t = 8. The left panel of Fig. 7 shows the expectation of the electric energy and chiral condensate
as a function of time for different truncations of the electric field. The truncation error from these
numerical simulations is consistent with the analytical estimates in Eq. (59) and Eq. (61).

17
 101
0.4
10 2

0.3
10 5

Error in E2
E2

0.2 10 8

10 11
0.1
=3
=2 10 14
=1
0.0 =4
0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 6 7 8
0.8
10 1

0.6
10 3

0.4 10 5

10 7
Error in

0.2
10 9

10 11
0.0
10 13

0.2 10 15

0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 6 7 8
t t
Figure 7: Evolution of the electric vacuum in the Schwinger model with g = 0.8 and m = 0.1. The left panel shows
the electric energy per link and the chiral condensate per site as a function of time for different truncations. The
right panel shows the difference in the expectation of the observable between the Λ = 4 and lower truncations. The
dashed lines in the right panel show the predicted truncation error. For Λ = 1, the truncation error was computed
using Eq. (59) and for larger Λ, Eq. (61) was used.

18
4 Discussion
In this work, a formalism has been developed for estimating the errors due to electric basis truncations
in the simulation of lattice gauge theories. This formalism is based on the presence of Hilbert space
fragmentation in the Kogut–Susskind Hamiltonian. This enables the use of time-dependent pertur-
bation theory to describe the dynamics of states with large electric fields. Consequently, the leading
contributions to the truncation error fall off factorially with the field truncation. Numerical simulations
were performed to demonstrate the performance of these error estimates. The details of the truncation
error depend on the change in electric energy from applying the plaquette operator. This suggests
lattices with more links per plaquette, such as a honeycomb lattice [83, 88, 116, 126] or a triamond
lattice [67, 68], will have a stronger suppression of truncation errors.
The truncation errors estimated in this work are significantly smaller than in previous work. This
is due to previous work being based on a loose estimate of the remainder term in the Dyson series
expansion, while this work was able to derive the form of the dominant error due to truncation. As a
result, for some parameter choices, our results reduce the error estimate by an astronomical factor of
10306 . Our formalism does not apply to generic bosonic simulations; however, preliminary simulations
in Appendix B suggest that there may be other mechanisms that restrict the growth of boson numbers
more generally. While the formalism was only applied to Abelian gauge theories in one dimension
in this work, the generalization to non-Abelian gauge groups and higher spatial dimensions is clear.
This formalism also applies to scalar field theories with a compact scalar field, such as the O(3) non-
linear σ model. The calculations performed in this work show that different observables have different
dependencies on the truncation errors. One could potentially use this to construct observables with
less sensitivity to truncation errors and potentially extract useful information from simulations with
low truncations. This formalism can also be extended to improvement schemes using the similarity
renormalization group to reduce the effects of truncation errors [49]. These extensions of this work will
enable quantum simulations of lattice QCD with complete theoretical control over the uncertainties
due to truncating the gauge fields.

Acknowledgments
We would like to acknowledge useful conversations with Jesse Stryker, Ivan Burbano, Irian D’Andrea, Neel Modi, Chris
Kane, and helpful feedback from John Preskill. A.N.C, S.H. and C.W.B. acknowledge funding through the U.S. Depart-
ment of Energy (DOE), Office of Science under contract DE-AC02-05CH11231, partially through Quantum Information
Science Enabled Discovery (QuantISED) for High Energy Physics (KA2401032). Additional support is acknowledged
from the U.S. Department of Energy, Office of Science, National Quantum Information Science Research Centers, Quan-
tum Systems Accelerator. J.C.H. acknowledges funding by the Max Planck Society, the Deutsche Forschungsgemeinschaft
(DFG, German Research Foundation) under Germany’s Excellence Strategy – EXC-2111 – 390814868, and the Euro-
pean Research Council (ERC) under the European Union’s Horizon Europe research and innovation program (Grant
Agreement No. 101165667)—ERC Starting Grant QuSiGauge. Views and opinions expressed are however those of the
author(s) only and do not necessarily reflect those of the European Union or the European Research Council Executive
Agency. Neither the European Union nor the granting authority can be held responsible for them. This work is part of
the Quantum Computing for High-Energy Physics (QC4HEP) working group.

A Infinite MPS implementation

The simulations of the infinite plaquette chain were performed using the time-evolved block decimation
algorithm (TEBD) [129]. A Trotter step size of ∆t = 0.01 was used. For the results shown in Fig. 4, the
time evolution was performed with a fixed bond dimension. The bond dimension was then increased
by 10, and the full simulation was rerun. This process was repeated until the difference in the electric
energy between two consecutive runs was an order of magnitude smaller than the predicted truncation
error. The resulting bond dimensions are shown in Table 1. The results shown in Fig. 5 were obtained

19
 Λ Bond Dimension
1 60
2 90
3 100
4 170
5 170

Table 1: Bond dimension required for convergence of the electric energy for an infinite plaquette chain with g = 0.8
and truncation Λ shown in Fig. 4.

Λ Bond Dimension
1 90
2 90
3 110
4 110

Table 2: Bond dimension required for convergence of the electric energy for the Schwinger model on an infinite
lattice with g = 0.8, m = 0.1, and truncation Λ.

by similarly performing the evolution with a fixed bond dimension and then again after increasing the
bond dimension by 25. We repeated this process until the difference in the results from consecutive
runs was 2 orders of magnitude below the upper bound computed using Eq. (52) (corresponding to
the dashed horizontal lines shown in Fig. 5). We note that the simulation for all values of g and Λ
converged within a bond dimension of 180.
The simulations of the Schwinger model were also performed using TEBD. The Hilbert space for
the fermion on each site and the right neighboring link were combined into a single site, represented by
a single tensor index in the matrix product state representation. A Trotter step size of ∆t = 0.01 was
used. It was found that performing simulations with a large fixed bond dimension led to numerical
instabilities. For this reason, the simulation was performed using a bond dimension of 50 until t = 1.5.
The bond dimension was then increased, and the same criteria was used as in the plaquette chain
simulation to determine when the simulation had converged to sufficient precision. The resulting
maximum bond dimensions are shown in Table 2.

B Hubbard-Holstein Model
Previous work studying gauge field truncation errors for quantum simulation of lattice gauge theories
also applied their results to bosonic systems [127]. The Hubbard-Holstein model was used as an
example system. This theory is a model of electron-phonon interactions. The Hamiltonian is given by

Ĥ = ĤF + g ĤBF + ω ĤB

1X †
ĤF = ψ̂ ψ̂i + ψ̂i† ψ̂i+1
2 i i+1
X †
ĤBF = ψ̂i ψ̂i (âi + â†i )
i

â†i âi
X
ĤF =
i
(62)

where ψ̂i is a fermion annihilation operator at site i, âi is a boson annihilation operator at site i, ω is
the phonon frequency and g is the electron-phonon coupling. Note that this theory does not exhibit

20
 n=2
n=4
0.20 n=6
n=8 10 1

max(P(n))
0.15
P(n)

0.10
10 2

0.05

0.00
0 2 4 6 8 10 2 4 6 8 10
t t
Figure 8: Simulation of the Hubbard-Holstein model on two sites with g = 1 and ω = 1 for a time t = 10. The left
panel shows the probability of the boson mode on the left site having an occupation number of n as a function of
time. The right panel shows the maximum probability of occupying each boson number during this time evolution.

the same Hilbert space fragmentation that was used to derive truncation error estimates for lattice
gauge theories. It was estimated that to simulate the dynamics of this model with g = 1 for a time
t = 1, a boson truncation of 100 was necessary to keep leakage errors below ϵ = 0.2. As a probe of this
estimate, a simulation was performed of this model on two sites with g = 1, ω = 1, and a single fermion
present. In this simulation, the boson number at each site was truncated at Λ = 100. The system was
initialized with the boson number at both sites equal to zero, and the fermion was placed on the left
site. The system was evolved for a time t = 10. The left panel of Figure 8 shows the probability of the
boson mode on the left site having an occupation number of n as a function of time. The right panel
shows the maximum probability of occupying each boson number during this time evolution. As this
figure shows, the leakage probability is significantly smaller than previous work estimated.

References
[1] Richard P. Feynman. “Simulating physics with computers”. Int. J. Theor. Phys. 21, 467–
488 (1982).
[2] Travis S. Humble et al. “Snowmass White Paper: Quantum Computing Systems and Software
for High-energy Physics Research”. In Snowmass 2021. (2022). arXiv:2203.07091.
[3] Christian W. Bauer et al. “Quantum Simulation for High-Energy Physics”. PRX Quantum 4,
027001 (2023). arXiv:2204.03381.
[4] Douglas Beck et al. “Quantum Information Science and Technology for Nuclear Physics. Input
into U.S. Long-Range Planning, 2023” (2023). arXiv:2303.00113.
[5] Christian W. Bauer, Marat Freytsis, and Benjamin Nachman. “Simulating Collider Physics on
Quantum Computers Using Effective Field Theories”. Phys. Rev. Lett. 127, 212001 (2021).
arXiv:2102.05044.
[6] Christian W. Bauer. “Efficient use of quantum computers for collider physics” (2025).
arXiv:2503.16602.
[7] Thomas D. Cohen, Henry Lamm, Scott Lawrence, and Yukari Yamauchi. “Quantum algorithms
for transport coefficients in gauge theories”. Phys. Rev. D 104, 094514 (2021). arXiv:2104.02024.
[8] Francesco Turro, Anthony Ciavarella, and Xiaojun Yao. “Classical and quantum comput-
ing of shear viscosity for (2+1)D SU(2) gauge theory”. Phys. Rev. D 109, 114511 (2024).
arXiv:2402.04221.

21
 [9] M. Luscher. “Volume Dependence of the Energy Spectrum in Massive Quantum Field Theories.
1. Stable Particle States”. Commun. Math. Phys. 104, 177 (1986).
[10] M. Luscher. “Volume Dependence of the Energy Spectrum in Massive Quantum Field Theories.
2. Scattering States”. Commun. Math. Phys. 105, 153–188 (1986).
[11] Anthony Ciavarella. “Algorithm for quantum computation of particle decays”. Phys. Rev. D
102, 094505 (2020). arXiv:2007.04447.
[12] Raúl A. Briceño, Juan V. Guerrero, Maxwell T. Hansen, and Alexandru M. Sturzu. “Role of
boundary conditions in quantum computations of scattering observables”. Phys. Rev. D 103,
014506 (2021). arXiv:2007.01155.
[13] Raul A. Briceño, Marco A. Carrillo, Juan V. Guerrero, Maxwell T. Hansen, and Alexandru M.
Sturzu. “Accessing scattering amplitudes using quantum computers”. PoS LATTICE2021,
315 (2022). arXiv:2112.01968.
[14] Marco A. Carrillo, Raúl A. Briceño, and Alexandru M. Sturzu. “Inclusive reactions from finite
Minkowski spacetime correlation functions”. Phys. Rev. D 110, 054503 (2024). arXiv:2406.06877.
[15] Ivan M. Burbano, Marco A. Carrillo, Rana Urek, Anthony N. Ciavarella, and Raúl A. Briceño.
“Real-time Estimators for Scattering Observables: A full account of finite volume errors for
quantum simulation” (2025). arXiv:2506.06511.
[16] Michael A. Levin and Xiao-Gang Wen. “String net condensation: A Physical mechanism for
topological phases”. Phys. Rev. B 71, 045110 (2005). arXiv:cond-mat/0404617.
[17] Lucile Savary and Leon Balents. “Quantum spin liquids: a review”. Rept. Prog. Phys. 80,
016502 (2017). arXiv:1601.03742.
[18] Simon Trebst and Ciarán Hickey. “Kitaev materials”. Phys. Rept. 950, 1–37 (2022).
[19] Yukitoshi Motome et al. “Hunting Majorana Fermions in Kitaev Magnets”. J. Phys. Soc. Jap.
89, 012002 (2020). arXiv:1909.02234.
[20] Anthony N. Ciavarella, Christian W. Bauer, and Jad C. Halimeh. “Generic Hilbert Space Frag-
mentation in Kogut–Susskind Lattice Gauge Theories” (2025). arXiv:2502.03533.
[21] Jean-Yves Desaules, Guo-Xian Su, Ian P. McCulloch, Bing Yang, Zlatko Papić, and Jad C. Hal-
imeh. “Ergodicity Breaking Under Confinement in Cold-Atom Quantum Simulators”. Quantum
8, 1274 (2024). arXiv:2301.07717.
[22] Jean-Yves Desaules, Thomas Iadecola, and Jad C. Halimeh. “Mass-assisted local deconfinement
in a confined Z2 lattice gauge theory”. Phys. Rev. B 112, 014301 (2025). arXiv:2404.11645.
[23] Jared Jeyaretnam, Tanmay Bhore, Jesse J. Osborne, Jad C. Halimeh, and Zlatko Papić. “Hilbert
space fragmentation at the origin of disorder-free localization in the lattice Schwinger model”.
Commun. Phys. 8, 172 (2025). arXiv:2409.08320.
[24] E. A. Martinez et al. “Real-time dynamics of lattice gauge theories with a few-qubit quantum
computer”. Nature 534, 516–519 (2016). arXiv:1605.04570.
[25] N. Klco, E. F. Dumitrescu, A. J. McCaskey, T. D. Morris, R. C. Pooser, M. Sanz, E. Solano,
P. Lougovski, and M. J. Savage. “Quantum-classical computation of Schwinger model dynamics
using quantum computers”. Phys. Rev. A 98, 032331 (2018). arXiv:1803.03326.
[26] Frederik Görg, Kilian Sandholzer, Joaquín Minguzzi, Rémi Desbuquois, Michael Messer, and
Tilman Esslinger. “Realization of density-dependent Peierls phases to engineer quantized gauge
fields coupled to ultracold matter”. Nature Phys. 15, 1161–1167 (2019). arXiv:1812.05895.
[27] Alexander Mil, Torsten V. Zache, Apoorva Hegde, Andy Xia, Rohit P. Bhatt, Markus K.
Oberthaler, Philipp Hauke, Jürgen Berges, and Fred Jendrzejewski. “A scalable realization
of local U(1) gauge invariance in cold atomic mixtures”. Science 367, 1128–1130 (2020).
arXiv:1909.07641.
[28] Bing Yang, Hui Sun, Robert Ott, Han-Yi Wang, Torsten V. Zache, Jad C. Halimeh, Zhen-
Sheng Yuan, Philipp Hauke, and Jian-Wei Pan. “Observation of gauge invariance in a 71-site
Bose–Hubbard quantum simulator”. Nature 587, 392–396 (2020). arXiv:2003.08945.
[29] Zhao-Yu Zhou, Guo-Xian Su, Jad C. Halimeh, Robert Ott, Hui Sun, Philipp Hauke, Bing Yang,
Zhen-Sheng Yuan, Jürgen Berges, and Jian-Wei Pan. “Thermalization dynamics of a gauge
theory on a quantum simulator”. Science 377, abl6277 (2022). arXiv:2107.13563.

22
[30] Zhan Wang et al. “Observation of emergent Z2 gauge invariance in a superconducting circuit”.
Phys. Rev. Res. 4, L022060 (2022). arXiv:2111.05048.
[31] Nhung H. Nguyen, Minh C. Tran, Yingyue Zhu, Alaina M. Green, C. Huerta Alderete, Zohreh
Davoudi, and Norbert M. Linke. “Digital Quantum Simulation of the Schwinger Model and
Symmetry Protection with Trapped Ions”. PRX Quantum 3, 020324 (2022). arXiv:2112.14262.
[32] Guo-Xian Su, Hui Sun, Ana Hudomal, Jean-Yves Desaules, Zhao-Yu Zhou, Bing Yang, Jad C.
Halimeh, Zhen-Sheng Yuan, Zlatko Papić, and Jian-Wei Pan. “Observation of many-body scar-
ring in a Bose-Hubbard quantum simulator”. Phys. Rev. Res. 5, 023010 (2023). arXiv:2201.00821.
[33] Han-Yi Wang, Wei-Yong Zhang, Zhi-Yuan Yao, Ying Liu, Zi-Hang Zhu, Yong-Guang Zheng,
Xuan-Kai Wang, Hui Zhai, Zhen-Sheng Yuan, and Jian-Wei Pan. “Interrelated Thermalization
and Quantum Criticality in a Lattice Gauge Simulator”. Phys. Rev. Lett. 131, 050401 (2023).
arXiv:2210.17032.
[34] Wei-Yong Zhang et al. “Observation of microscopic confinement dynamics by a tunable topolog-
ical θ-angle”. Nature Phys. 21, 155–160 (2025). arXiv:2306.11794.
[35] Takis Angelides, Pranay Naredi, Arianna Crippa, Karl Jansen, Stefan Kühn, Ivano Tavernelli,
and Derek S. Wang. “First-order phase transition of the Schwinger model with a quantum
computer”. npj Quantum Inf. 11, 6 (2025). arXiv:2312.12831.
[36] Julius Mildenberger, Wojciech Mruczkiewicz, Jad C. Halimeh, Zhang Jiang, and Philipp Hauke.
“Confinement in a Z2 lattice gauge theory on a quantum computer”. Nature Phys. 21, 312–
317 (2025). arXiv:2203.08905.
[37] Clement Charles, Erik J. Gustafson, Elizabeth Hardt, Florian Herren, Norman Hogan, Henry
Lamm, Sara Starecheski, Ruth S. Van de Water, and Michael L. Wagman. “Simulating Z2 lattice
gauge theory on a quantum computer”. Phys. Rev. E 109, 015307 (2024). arXiv:2305.02361.
[38] Niklas Mueller, Joseph A. Carolan, Andrew Connelly, Zohreh Davoudi, Eugene F. Dumitrescu,
and Kübra Yeter-Aydeniz. “Quantum Computation of Dynamical Quantum Phase Transitions
and Entanglement Tomography in a Lattice Gauge Theory”. PRX Quantum 4, 030323 (2023).
arXiv:2210.03089.
[39] Arinjoy De et al. “Observation of string-breaking dynamics in a quantum simulator” (2024).
arXiv:2410.13815.
[40] Zohreh Davoudi, Chung-Chun Hsieh, and Saurabh V. Kadam. “Scattering wave packets of
hadrons in gauge theories: Preparation on a quantum computer”. Quantum 8, 1520 (2024).
arXiv:2402.00840.
[41] Yibin Guo, Takis Angelides, Karl Jansen, and Stefan Kühn. “Concurrent VQE for Simulating
Excited States of the Schwinger Model” (2024). arXiv:2407.15629.
[42] Roland C. Farrell, Marc Illa, Anthony N. Ciavarella, and Martin J. Savage. “Scalable Circuits
for Preparing Ground States on Digital Quantum Computers: The Schwinger Model Vacuum on
100 Qubits”. PRX Quantum 5, 020315 (2024). arXiv:2308.04481.
[43] Roland C. Farrell, Marc Illa, Anthony N. Ciavarella, and Martin J. Savage. “Quantum simulations
of hadron dynamics in the Schwinger model using 112 qubits”. Phys. Rev. D 109, 114510 (2024).
arXiv:2401.08044.
[44] Yasar Y. Atas, Jinglei Zhang, Randy Lewis, Amin Jahanpour, Jan F. Haase, and Christine A.
Muschik. “SU(2) hadrons on a quantum computer via a variational approach”. Nature Commun.
12, 6499 (2021). arXiv:2102.08920.
[45] Yasar Y. Atas, Jan F. Haase, Jinglei Zhang, Victor Wei, Sieglinde M. L. Pfaendler, Randy Lewis,
and Christine A. Muschik. “Simulating one-dimensional quantum chromodynamics on a quantum
computer: Real-time evolutions of tetra- and pentaquarks”. Phys. Rev. Res. 5, 033184 (2023).
arXiv:2207.03473.
[46] Anton T. Than et al. “The phase diagram of quantum chromodynamics in one dimension on a
quantum computer” (2024). arXiv:2501.00579.
[47] Roland C. Farrell, Ivan A. Chernyshev, Sarah J. M. Powell, Nikita A. Zemlevskiy, Marc Illa, and
Martin J. Savage. “Preparations for quantum simulations of quantum chromodynamics in 1+1
dimensions. I. Axial gauge”. Phys. Rev. D 107, 054512 (2023). arXiv:2207.01731.

23
[48] Roland C. Farrell, Ivan A. Chernyshev, Sarah J. M. Powell, Nikita A. Zemlevskiy, Marc Illa,
and Martin J. Savage. “Preparations for quantum simulations of quantum chromodynamics in
1+1 dimensions. II. Single-baryon β-decay in real time”. Phys. Rev. D 107, 054513 (2023).
arXiv:2209.10781.
[49] Anthony N. Ciavarella. “Quantum simulation of lattice QCD with improved Hamiltonians”.
Phys. Rev. D 108, 094513 (2023). arXiv:2307.05593.
[50] Anthony N. Ciavarella. “String breaking in the heavy quark limit with scalable circuits”. Phys.
Rev. D 111, 054501 (2025). arXiv:2411.05915.
[51] Ivan A. Chernyshev et al. “Pathfinding Quantum Simulations of Neutrinoless Double-β De-
cay” (2025). arXiv:2506.05757.
[52] Jiunn-Wei Chen, Yu-Ting Chen, and Ghanashyam Meher. “Parton Distributions on a Quantum
Computer” (2025). arXiv:2506.16829.
[53] Zi-Hang Zhu et al. “Probing false vacuum decay on a cold-atom gauge-theory quantum simula-
tor” (2024). arXiv:2411.12565.
[54] Gaurav Gyawali et al. “Observation of disorder-free localization using a (2+1)D lattice gauge
theory on a quantum processor” (2024). arXiv:2410.06557.
[55] Julian Schuhmacher, Guo-Xian Su, Jesse J. Osborne, Anthony Gandon, Jad C. Halimeh, and
Ivano Tavernelli. “Observation of hadron scattering in a lattice gauge theory on a quantum
computer” (2025). arXiv:2505.20387.
[56] Zohreh Davoudi, Chung-Chun Hsieh, and Saurabh V. Kadam. “Quantum computation of hadron
scattering in a lattice gauge theory” (2025). arXiv:2505.20408.
[57] Natalie Klco, Jesse R. Stryker, and Martin J. Savage. “SU(2) non-Abelian gauge field the-
ory in one dimension on digital quantum computers”. Phys. Rev. D 101, 074512 (2020).
arXiv:1908.06935.
[58] Danny Paulson et al. “Simulating 2D Effects in Lattice Gauge Theories on a Quantum Computer”.
PRX Quantum 2, 030334 (2021). arXiv:2008.09252.
[59] Anthony Ciavarella, Natalie Klco, and Martin J. Savage. “Trailhead for quantum simulation
of SU(3) Yang-Mills lattice gauge theory in the local multiplet basis”. Phys. Rev. D 103,
094501 (2021). arXiv:2101.10227.
[60] Anthony N. Ciavarella and Ivan A. Chernyshev. “Preparation of the SU(3) lattice Yang-Mills
vacuum with variational quantum methods”. Phys. Rev. D 105, 074504 (2022). arXiv:2112.09083.
[61] Sarmed A Rahman, Randy Lewis, Emanuele Mendicelli, and Sarah Powell. “Self-mitigating
Trotter circuits for SU(2) lattice gauge theory on a quantum computer”. Phys. Rev. D 106,
074502 (2022). arXiv:2205.09247.
[62] Emanuele Mendicelli, Randy Lewis, Sarmed A. Rahman, and Sarah Powell. “Real time evo-
lution and a traveling excitation in SU(2) pure gauge theory on a quantum computer.”. PoS
LATTICE2022, 025 (2023). arXiv:2210.11606.
[63] Zhiyao Li, Dorota M. Grabowska, and Martin J. Savage. “Sequency Hierarchy Truncation (Se-
qHT) for Adiabatic State Preparation and Time Evolution in Quantum Simulations” (2024).
arXiv:2407.13835.
[64] Jad C. Halimeh, Uliana E. Khodaeva, Dmitry L. Kovrizhin, Roderich Moessner, and Jo-
hannes Knolle. “Disorder-Free Localization for Benchmarking Quantum Computers” (2024).
arXiv:2410.08268.
[65] Sabhyata Gupta, Younes Javanmard, Tobias J. Osborne, and Luis Santos. “Simulation of a
Rohksar–Kivelson ladder on a NISQ device”. Sci. Rep. 14, 29276 (2024). arXiv:2401.16326.
[66] Arianna Crippa, Karl Jansen, and Enrico Rinaldi. “Analysis of the confinement string in (2
+ 1)-dimensional Quantum Electrodynamics with a trapped-ion quantum computer” (2024).
arXiv:2411.05628.
[67] Ali H. Z. Kavaki and Randy Lewis. “False vacuum decay in triamond lattice gauge theory” (2025).
arXiv:2503.01119.
[68] Ali H. Z. Kavaki and Randy Lewis. “From square plaquettes to triamond lattices for SU(2) gauge
theory”. Commun. Phys. 7, 208 (2024). arXiv:2401.14570.

24
[69] Tim Byrnes and Yoshihisa Yamamoto. “Simulating lattice gauge theories on a quantum com-
puter”. Phys. Rev. A 73, 022328 (2006). arXiv:quant-ph/0510027.
[70] Mari Carmen Bañuls, Krzysztof Cichy, J. Ignacio Cirac, Karl Jansen, and Stefan Kühn. “Efficient
basis formulation for 1+1 dimensional SU(2) lattice gauge theory: Spectral calculations with
matrix product states”. Phys. Rev. X 7, 041046 (2017). arXiv:1707.06434.
[71] Erez Zohar and J. Ignacio Cirac. “Removing Staggered Fermionic Matter in U (N ) and SU (N )
Lattice Gauge Theories”. Phys. Rev. D 99, 114511 (2019). arXiv:1905.00652.
[72] Randy Lewis and R. M. Woloshyn. “A qubit model for U(1) lattice gauge theory” (2019).
arXiv:1905.09789.
[73] Indrakshi Raychowdhury and Jesse R. Stryker. “Solving Gauss’s Law on Digital Quan-
tum Computers with Loop-String-Hadron Digitization”. Phys. Rev. Res. 2, 033039 (2020).
arXiv:1812.07554.
[74] Indrakshi Raychowdhury and Jesse R. Stryker. “Loop, string, and hadron dynamics in SU(2)
Hamiltonian lattice gauge theories”. Phys. Rev. D 101, 114502 (2020). arXiv:1912.06133.
[75] Jesse Stryker and Indrakshi Raychowdhury. “Tailoring Non-Abelian Gauge Theory for Digital
Quantum Simulation”. PoS LATTICE2019, 144 (2020).
[76] Saurabh V. Kadam, Indrakshi Raychowdhury, and Jesse R. Stryker. “Loop-string-hadron for-
mulation of an SU(3) gauge theory with dynamical quarks”. Phys. Rev. D 107, 094513 (2023).
arXiv:2212.04490.
[77] Saurabh Vasant Kadam, Indrakshi Raychowdhury, and Jesse R. Stryker. “Loop-string-hadron
formulation of an SU(3) gauge theory with dynamical quarks”. PoS LATTICE2022, 373 (2023).
[78] Saurabh V. Kadam, Aahiri Naskar, Indrakshi Raychowdhury, and Jesse R. Stryker. “Loop-string-
hadron approach to SU(3) lattice Yang-Mills theory: Gauge invariant Hilbert space of a trivalent
vertex” (2024). arXiv:2407.19181.
[79] Torsten V. Zache, Daniel González-Cuadra, and Peter Zoller. “Quantum and Classical Spin-
Network Algorithms for q-Deformed Kogut-Susskind Gauge Theories”. Phys. Rev. Lett. 131,
171902 (2023). arXiv:2304.02527.
[80] Angus Kan, Lena Funcke, Stefan Kühn, Luca Dellantonio, Jinglei Zhang, Jan F. Haase, Chris-
tine A. Muschik, and Karl Jansen. “Investigating a (3+1)D topological θ-term in the Hamiltonian
formulation of lattice gauge theories for quantum and classical simulations”. Phys. Rev. D 104,
034504 (2021). arXiv:2105.06019.
[81] Anthony Ciavarella, Natalie Klco, and Martin J. Savage. “Some Conceptual Aspects of
Operator Design for Quantum Simulations of Non-Abelian Lattice Gauge Theories” (2022).
arXiv:2203.11988.
[82] Anthony N. Ciavarella and Christian W. Bauer. “Quantum Simulation of SU(3) Lattice Yang-
Mills Theory at Leading Order in Large-Nc Expansion”. Phys. Rev. Lett. 133, 111901 (2024).
arXiv:2402.10265.
[83] Berndt Müller and Xiaojun Yao. “Simple Hamiltonian for quantum simulation of strongly coupled
(2+1)D SU(2) lattice gauge theory on a honeycomb lattice”. Phys. Rev. D 108, 094505 (2023).
arXiv:2307.00045.
[84] Tomoya Hayata and Yoshimasa Hidaka. “q deformed formulation of Hamiltonian SU(3) Yang-
Mills theory”. JHEP 09, 123 (2023). arXiv:2306.12324.
[85] Marco Rigobello, Giuseppe Magnifico, Pietro Silvi, and Simone Montangero. “Hadrons in (1+1)D
Hamiltonian hardcore lattice QCD” (2023). arXiv:2308.04488.
[86] Pierpaolo Fontana, Marc Miranda Riaza, and Alessio Celi. “An efficient finite-resource formula-
tion of non-Abelian lattice gauge theories beyond one dimension” (2024). arXiv:2409.04441.
[87] Praveen Balaji, Cianán Conefrey-Shinozaki, Patrick Draper, Jason K. Elhaderi, Drishti Gupta,
Luis Hidalgo, Andrew Lytle, and Enrico Rinaldi. “Quantum Circuits for SU(3) Lattice Gauge
Theory” (2025). arXiv:2503.08866.
[88] Marc Illa, Martin J. Savage, and Xiaojun Yao. “Improved Honeycomb and Hyper-Honeycomb
Lattice Hamiltonians for Quantum Simulations of Non-Abelian Gauge Theories” (2025).
arXiv:2503.09688.

25
 [89] Anthony N. Ciavarella, I. M. Burbano, and Christian W. Bauer. “Efficient Truncations of SU(Nc )
Lattice Gauge Theory for Quantum Simulation” (2025). arXiv:2503.11888.
[90] Henry Lamm, Scott Lawrence, and Yukari Yamauchi. “General Methods for Digital Quantum
Simulation of Gauge Theories”. Phys. Rev. D 100, 034518 (2019). arXiv:1903.08807.
[91] Andrei Alexandru, Paulo F. Bedaque, Siddhartha Harmalkar, Henry Lamm, Scott Lawrence,
and Neill C. Warrington. “Gluon Field Digitization for Quantum Computers”. Phys. Rev. D
100, 114501 (2019). arXiv:1906.11213.
[92] M. Sohaib Alam, Stuart Hadfield, Henry Lamm, and Andy C. Y. Li. “Primitive quantum gates
for dihedral gauge theories”. Phys. Rev. D 105, 114501 (2022). arXiv:2108.13305.
[93] Yao Ji, Henry Lamm, and Shuchen Zhu. “Gluon digitization via character expansion for quantum
computers”. Phys. Rev. D 107, 114503 (2023). arXiv:2203.02330.
[94] Erik J. Gustafson, Henry Lamm, Felicity Lovelace, and Damian Musk. “Primitive quantum
gates for an SU(2) discrete subgroup: Binary tetrahedral”. Phys. Rev. D 106, 114501 (2022).
arXiv:2208.12309.
[95] Erik J. Gustafson and Henry Lamm. “Robustness of Gauge Digitization to Quantum
Noise” (2023). arXiv:2301.10207.
[96] Erik J. Gustafson, Henry Lamm, and Felicity Lovelace. “Primitive quantum gates for an SU(2)
discrete subgroup: Binary octahedral”. Phys. Rev. D 109, 054503 (2024). arXiv:2312.10285.
[97] Erik J. Gustafson, Yao Ji, Henry Lamm, Edison M. Murairi, Sebastian Osorio Perez, and Shuchen
Zhu. “Primitive quantum gates for an SU(3) discrete subgroup: Σ(36×3)”. Phys. Rev. D 110,
034515 (2024). arXiv:2405.05973.
[98] Benoît Assi and Henry Lamm. “Digitization and subduction of SU(N) gauge theories”. Phys.
Rev. D 110, 074511 (2024). arXiv:2405.12204.
[99] Tobias Hartung, Timo Jakobs, Karl Jansen, Johann Ostmeyer, and Carsten Urbach. “Digitising
SU(2) gauge fields and the freezing transition”. Eur. Phys. J. C 82, 237 (2022). arXiv:2201.09625.
[100] Christian W. Bauer and Dorota M. Grabowska. “Efficient representation for simulating U(1)
gauge theories on digital quantum computers at all values of the coupling”. Phys. Rev. D 107,
L031503 (2023). arXiv:2111.08015.
[101] Irian D’Andrea, Christian W. Bauer, Dorota M. Grabowska, and Marat Freytsis. “New basis for
Hamiltonian SU(2) simulations”. Phys. Rev. D 109, 074501 (2024). arXiv:2307.11829.
[102] Timo Jakobs, Marco Garofalo, Tobias Hartung, Karl Jansen, Johann Ostmeyer, Dominik Rolfes,
Simone Romiti, and Carsten Urbach. “Canonical momenta in digitized Su(2) lattice gauge theory:
definition and free theory”. Eur. Phys. J. C 83, 669 (2023). arXiv:2304.02322.
[103] Marco Garofalo, Tobias Hartung, Timo Jakobs, Karl Jansen, Johann Ostmeyer, Dominik Rolfes,
Simone Romiti, and Carsten Urbach. “Testing the SU(2) lattice Hamiltonian built from S3
partitionings”. PoS LATTICE2023, 231 (2024). arXiv:2311.15926.
[104] Dorota M. Grabowska, Christopher F. Kane, and Christian W. Bauer. “A Fully Gauge-Fixed
SU(2) Hamiltonian for Quantum Simulations” (2024). arXiv:2409.10610.
[105] I. M. Burbano and Christian W. Bauer. “Gauge Loop-String-Hadron Formulation on General
Graphs and Applications to Fully Gauge Fixed Hamiltonian Lattice Gauge Theory” (2024).
arXiv:2409.13812.
[106] Timo Jakobs, Marco Garofalo, Tobias Hartung, Karl Jansen, Johann Ostmeyer, Simone Romiti,
and Carsten Urbach. “Dynamics in Hamiltonian Lattice Gauge Theory: Approaching the Con-
tinuum Limit with Partitionings of SU(2)” (2025). arXiv:2503.03397.
[107] R. Brower, S. Chandrasekharan, and U. J. Wiese. “QCD as a quantum link model”. Phys. Rev.
D 60, 094502 (1999). arXiv:hep-th/9704106.
[108] R. Brower, S. Chandrasekharan, S. Riederer, and U. J. Wiese. “D theory: Field quantization
by dimensional reduction of discrete variables”. Nucl. Phys. B 693, 149–175 (2004). arXiv:hep-
lat/0309182.
[109] Hanqing Liu, Tanmoy Bhattacharya, Shailesh Chandrasekharan, and Rajan Gupta. “Phases of
2d massless QCD with qubit regularization” (2023). arXiv:2312.17734.

26
[110] Shailesh Chandrasekharan. “Qubit Regularization of Quantum Field Theories”. In 41st Interna-
tional Symposium on Lattice Field Theory. (2025). arXiv:2502.05716.
[111] Shailesh Chandrasekharan, Rui Xian Siew, and Tanmoy Bhattacharya. “Monomer-dimer tensor-
network basis for qubit-regularized lattice gauge theories” (2025). arXiv:2502.14175.
[112] Andrei Alexandru, Paulo F. Bedaque, Andrea Carosso, Michael J. Cervia, Edison M. Murairi,
and Andy Sheng. “Fuzzy gauge theory for quantum computers”. Phys. Rev. D 109, 094502 (2024).
arXiv:2308.05253.
[113] Alexander J. Buser, Hrant Gharibyan, Masanori Hanada, Masazumi Honda, and Junyu
Liu. “Quantum simulation of gauge theory via orbifold lattice”. JHEP 09, 034 (2021).
arXiv:2011.06576.
[114] Georg Bergner, Masanori Hanada, Enrico Rinaldi, and Andreas Schafer. “Toward QCD on
quantum computer: orbifold lattice approach”. JHEP 05, 234 (2024). arXiv:2401.12045.
[115] Jad C. Halimeh, Masanori Hanada, Shunji Matsuura, Franco Nori, Enrico Rinaldi, and Andreas
Schäfer. “A universal framework for the quantum simulation of Yang-Mills theory” (2024).
arXiv:2411.13161.
[116] Marc Illa, Martin J. Savage, and Xiaojun Yao. “Dynamical Local Tadpole-Improvement in
Quantum Simulations of Gauge Theories” (2025). arXiv:2504.21575.
[117] Marcela Carena, Henry Lamm, Ying-Ying Li, and Wanqiang Liu. “Lattice renormalization of
quantum simulations”. Phys. Rev. D 104, 094519 (2021). arXiv:2107.01166.
[118] Christopher F. Kane, Siddharth Hariprakash, and Christian W. Bauer. “Obtaining con-
tinuum physics from dynamical simulations of Hamiltonian lattice gauge theories” (2025).
arXiv:2506.16559.
[119] Torsten V. Zache, Maarten Van Damme, Jad C. Halimeh, Philipp Hauke, and Debasish Banerjee.
“Toward the continuum limit of a ð1 + 1ÞD quantum link Schwinger model”. Phys. Rev. D 106,
L091502 (2022). arXiv:2104.00025.
[120] Jad C. Halimeh, Maarten Van Damme, Torsten V. Zache, Debasish Banerjee, and Philipp Hauke.
“Achieving the quantum field theory limit in far-from-equilibrium quantum link models”. Quan-
tum 6, 878 (2022). arXiv:2112.04501.
[121] Falk Bruckmann, Karl Jansen, and Stefan Kühn. “O(3) nonlinear sigma model in 1+1 dimensions
with matrix product states”. Phys. Rev. D 99, 074501 (2019). arXiv:1812.00944.
[122] Natalie Klco and Martin J. Savage. “Digitization of scalar fields for quantum computing”. Phys.
Rev. A 99, 052335 (2019). arXiv:1808.10378.
[123] Jack Y. Araz, Sebastian Schenk, and Michael Spannowsky. “Toward a quantum simula-
tion of nonlinear sigma models with a topological term”. Phys. Rev. A 107, 032619 (2023).
arXiv:2210.03679.
[124] Andrei Alexandru, Paulo F. Bedaque, Andrea Carosso, Michael J. Cervia, and Andy Sheng.
“Qubitization strategies for bosonic field theories”. Phys. Rev. D 107, 034503 (2023).
arXiv:2209.00098.
[125] Anthony N. Ciavarella, Stephan Caspar, Hersh Singh, and Martin J. Savage. “Preparation for
quantum simulation of the (1+1)-dimensional O(3) nonlinear σ model using cold atoms”. Phys.
Rev. A 107, 042404 (2023). arXiv:2211.07684.
[126] Xiaojun Yao, Lukas Ebner, Berndt Müller, Andreas Schäfer, and Clemens Seidl. “Testing
eigenstate thermalization hypothesis for non-Abelian gauge theories”. EPJ Web Conf. 296,
13008 (2024). arXiv:2312.13408.
[127] Yu Tong, Victor V. Albert, Jarrod R. McClean, John Preskill, and Yuan Su. “Provably accurate
simulation of gauge theories and bosonic systems”. Quantum 6, 816 (2022). arXiv:2110.06942.
[128] Murray Gell-Mann and Francis Low. “Bound states in quantum field theory”. Phys. Rev. 84,
350–354 (1951).
[129] G. Vidal. “Efficient simulation of one-dimensional quantum many-body systems”. Phys. Rev.
Lett. 93, 040502 (2004). arXiv:quant-ph/0310089.

27