2
angle as chiral TBG. By combining ideality, which allows
the construction of trial wavefunctions for topologically
ordered states, with higher Chern number in an exper-
imentally feasible system, these models provide a novel
and unexplored platform to study and realize interacting
topological phases. Ref. [80] numerically found model
FCI states with particle entanglement suggestive of par-
ticular “color-singlet” [40] Halperin states, though a gen-
eral analytic understanding of the range of possible states
in such a system remains lacking.
One approach to understanding correlated states in a
C > 1 band is to employ the hybrid Wannier functions
to decompose the band into CChern 1 bands by enlarg-
ing the unit cell [93]. However, this approach does not
respect the ideality condition: an ideal Chern C > 1
band generally decomposes into a set of C= 1 non-ideal
bands such that analytic techniques cannot be applied
to the decomposed basis. Furthermore, this decomposi-
tion assumes broken translation symmetry in a specific
direction from the outset, making it difficult to under-
stand translation-unbroken states like the FCIs observed
in Ref. [80]. This naturally leads to the question: is it
possible to decompose an ideal C > 1 band into ideal
C= 1 bands and in a way that captures both transla-
tionally symmetric and translation-breaking states? A
partial answer to this question was provided in Ref. [30]
which showed that it is impossible to decompose a generic
ideal C > 1 bands in terms of orthogonal and ideal C= 1
bands.
Here, we show that by lifting the orthogonality con-
straint, it is possible to decompose a generic ideal C > 1
into Cideal Chern 1 bands. From this decomposition,
we reveal a hidden non-unitary SU(C) action among the
decomposed bands combining real space and momentum
space translations. We note that although this SU(C)
action is generally non-unitary and is not a symmetry of
the Hamiltonian, it is often a symmetry of the ground
state manifold for short-range repulsive interactions, en-
abling us to make sharp predictions regarding the many-
body ground state at partial filling. Furthermore, the
SU(C) structure is reminiscent of the structure of multi-
component Landau levels [40], allowing us to interpret
these ground states in terms of more familiar correlated
states in multi-component quantum Hall systems. It
should be emphasized that the states we obtain are phys-
ically distinct and have non-trivial translation symmetry
breaking patterns visible through real-space charge den-
sity, compared to multi-component Landau levels which
have uniform charge density.
By employing this decomposition, we analytically con-
struct ground states at a variety of fillings of ideal Chern
C > 1 bands and make concrete predictions on their re-
alization. First, we identify a ground state manifold of
charge density waves (CDWs) at filling 1/C with emer-
gent SU(C) symmetry. Such states can be interpreted
as generalized quantum Hall ferromagnets in the decom-
posed basis. An immediate consequence of this iden-
tification is the existence of charged skyrmion textures
which correspond to a characteristic winding pattern of
the CDW order parameter (cf. Fig. 1). Such a wind-
ing pattern can be readily observed with local charge
probes such as STM, which has already been employed
in these systems [71, 72, 94–98]. Second, we characterize
the structure of translation symmetric fractional Chern
insulator states at fillings 1/(2Cs + 1), for each positive
integer s, by establishing a direct analogy with flavor-
singlet Halperin states. Finally, we discuss a manifold of
translation-breaking Laughlin states that appear at low
fillings 1/C(2s+ 1) where fractionalization and topolog-
ical order coexists with CDW order. All our results are
verified by numerical exact diagonalization (ED) on the
chiral model for twisted mono-bilayer graphene. These
results enable us to make experimental predictions for
graphene multilayers.
The paper is organized as follows. We begin with
an overview of the central theoretical results as well as
some consequences for experiment in Sec. II. Next, we re-
view the physical model for chiral graphene multilayers
in Sec. III and quantum geometry techniques to under-
stand the single particle physics in Sec. IV. We present
the central technical result, the decomposition of an ideal
higher Chern band, in Sec. V for C= 2. We study quan-
tum ferromagnets, which manifest as topological CDWs,
and their associated skyrmion description at half filling
for a spinless C= 2 band in Sec. VI and a spinful C= 2
band in Sec. VII. We discuss the possible fractional Chern
insulators, both translation symmetric and translation
broken, in Sec. VIII. We generalize the physical implica-
tions to ideal bands with C > 1 in Sec. IX. We conclude
with some future questions in Sec. X.
II. SUMMARY OF RESULTS
We begin by briefly summarizing our results and dis-
cussing their main implications. By now, a full under-
standing of ideal C= 1 bands has been achieved: they
have been related to the LLL of a Dirac particle in a mag-
netic field [8, 29], which enables the construction of exact
many body Laughlin-like ground states for short range
interactions [31]. Our main technical achievement in this
work is the generalization of such understanding for any
ideal higher Chern band. In particular, we provide a gen-
eral explicit construction to decompose any ideal higher
Chern band into ideal but non-orthogonal Chern 1 bands
with a non-unitary SU(C) action. Compared to the hy-
brid Wannier decomposition [93], our procedure has the
distinct advantages of preserving the ideal band geome-
try and allowing access to translation broken states in all
possible directions in the same basis. The preservation
of ideal band geometry means we leverage the analytic
knowledge of ideal C= 1 bands to analytically study in-
teracting phases – in particular, FCIs – in C > 1 bands.
The decomposition provides analytic many-body
states that are the exact ground states at certain frac-