New Territories, Hong Kong, China
Torus algebra and logical operators at low energy
Abstract
Given a modular tensor category , we construct an associative algebra , which we call the torus algebra. We prove that the torus algebra is semisimple by explicitly constructing all the simple modules. Suppose that a topological ordered phase described by is put on a torus. Physically, each simple module over consists of the low energy states on the torus with one anyon excitation, or equivalently, the ground states on a punctured torus where the anyon is enclosed by the puncture. Elements in can be physically interpreted as anyon hopping processes on the torus. We give the precise formula how an arbitrary logical operator on the low energy states on a torus can be realized by moving anyons on the torus. Our work thus provides a theoretical proposal that the low energy states on a torus can serve as topological qudits and one can arbitrarily manipulate them by moving anyons around.
1 Introduction
In Ma Ma et al. proposed an operator algebra approach to calculate the ground state degeneracy of the Ising cage-net model Prem_2019 . In this paper, we restrict our interest to the operator algebra corresponding to 2+1D topological orders Wen89 ; WN90 ; Wen90 ; Wen91 ; Wen9506066 , which is constructed based on the data of a modular tensor category (MTC) and physically can be thought of as the anyon hopping processes on the torus. We refer to such operator algebra as the torus algebra, denoted by .
Ref. Ma conjectured a method to decompose the torus algebra, by quotienting out certain ``loop" elements. Unfortunately, this method does not work in general cases. In this work we give a serious treatment of the torus algebra. The semisimplicity is rigorously proved by constructing all the simple modules. Moreover, we find that the simple modules have clear physical meanings: they are nothing but the low energy subspace on the torus with one anyon excitation. More precisely, given a punctured torus where anyon is enclosed in the puncture, the corresponding low energy subspace, denoted by , forms a simple module over . The algebra action can be intuitively depicted by the following topological quantum field theory (TQFT) picture in Figure 1. This picture agrees with the universal skein theory stated in bartlett2015modular : the extended 3d TQFT associated with MTC assigns the punctured torus to the space of ``internal string diagrams’’ inside the torus, which can be identified with the simple modules .
The main result of this work is the isomorphism (Theorem 5.7). Based on this result, we know that any logical operator in can be realized by elements in , i.e., certain patterns of anyon hopping on the torus. Therefore, our work paves the way for exploiting the low energy states on the torus as topological qudits, by manipulating them via moving anyons around the torus.
The paper is organized as follows. In Section 2, we introduce the necessary preliminaries and fix our notations. In Section 3, we give the precise definition of torus algebra. In Section 4, we discuss special elements in the torus algebra, dubbed the centrifugal loops, which lead to a set of central orthogonal idempotents. In Section 5, we construct the simple modules and prove the main theorem. In Section 6, we discuss the extra properties of when is a unitary MTC. In Section 7, we discuss the modular transformation on the punctured torus as an application of our main result. In Section 8, we give some simple but nontrivial examples.
2 Preliminaries
We assume that the reader has necessary knowledge on tensor categories and graphical calculus (see e.g., etingof2016tensor ; Kitaev_2006 ; Bojko ). In this section we fix our notations and list important results to be used later.
Given a MTC , we use to denote the set of (isomorphism classes) of simple objects in . For , we depict the basis vectors in and by
respectively. The morphisms in our graphs are to be composed from bottom to top.
The dimension of is denoted by , called the fusion rules. The dual object of is denoted by and depicted as
We will add arrow to the lines only when necessary; when there is no arrow it is understood as a line with an upwards arrow. The quantum dimension of is denoted by , graphically
We take the normalization such that the evaluation is
(1) |
We adopt the convention of F-move as follows.
(2) |
The colored nodes will not be labeled by letters unless necessary. The braiding is depicted by and the convention for -move is
(3) |
The -matrix is given by
(4) |
where is the total quantum dimension. For a MTC, the matrix is unitary and one has the Verlinde formula:
(5) |
where denotes the complex conjugate.
Also, a twist is
(6) |
where is a phase factor. The multiplicative chiral central charge is defined as
(7) |
where is the additive chiral central charge.
The following results will be used extensively in this paper:
Lemma 2.1.
-
1.
(8) -
2.
One can apply an F-move and the corresponding inverse F-move together on two matching patches of the graph,
(9) -
3.
Likewise for R-moves,
(10) In the above graph, we follow the nice Einstein-like convention of henriques2016bicommutant ; green2023enriched , that the pair of nodes with the same color is automatically summed over and the summation symbol is omitted. From now on, we will keep using this convention.
-
4.
In the Einstein-like convention, we have
(11)
3 Torus algebra
In this section, we will construct the torus algebra, motivated by stacking anyon hopping operators on a torus. We cut the torus along two non-contractible loops, and study the resulting graph on a square:
This allows us to intuitively identify the basis vector as {tikzpicture} . We then define the stacking of anyon hopping operators as the multiplication,
We now give the rigorous definition:
Definition 3.1.
Given a MTC , the torus algebra is the following vector space
equipped with the multiplication map
(12) |
and unit .
Remark 3.2.
We give a graphical proof that the multiplication in the above definition is associative, i.e. ,
(13) |
Proof.
Remark 3.3.
One should not confuse the torus algebra with the tube algebra (see e.g., ocneanu1994chirality ; hardiman2020graphical ; LW1311.1784 . They have different background manifolds and the multiplications are given by stacking versus gluing.
4 Central idempotent decomposition
In this section, we will introduce the central elements in and the construction of central idempotents to decompose the torus algebra systematically.
Firstly, consider the following graph
Since the multiplication is stacking graphs, and using the naturality of braiding, the purple loop can be made not overlap with a general element , we know the result is the same no matter the purple loop is above or below . In other words, the purple loop gives rise to a central element. More rigorously, we introduce
Definition 4.1 (Centrifugal loop).
We define the centrifugal loop elements in , for each , to be the elements of the following special form
(15) |
and represent them graphically as loops surrounding the four corners of the square. Gluing parallel ends of the square makes them contractible loops on the torus; however, such deformation is forbidden in and they are nontrivial elements.
Proposition 4.2.
Centrifugal loops of any are in the center of , i.e. ,
(16) |
Proof.
Proposition 4.3.
The muliplication of centrifugal loop elements obey the fusion rule of the MTC, i.e.,
(18) |
Proof.
(19) |
∎
By the Verlinde formula, the fusion rule can be diagonalized by the -matrix. Thus, for simple , we denote
(20) |
Then,
Proposition 4.4.
are mutually orthogonal idempotents, i.e.,
(21) |
They are also central and sum to the unit of the algebra.
Proof.
Corollary 4.5.
admits the direct sum decomposition as algebras.
We will further prove that each direct summand is simple (a matrix algebra) by explicitly constructing the corresponding simple modules.
5 Simple modules and punctured torus
Definition 5.1.
For any simple object , we define the left -module, denoted by , as the following vector space
(24) |
equipped with the action of ,
(25) | ||||
(26) |
Alternatively, the action can be expressed as the algebra homomorphism by .
Remark 5.2.
We give a graphical proof that the action in the above definition is associative.
(27) |
Proof.
The proof uses the F-move convention to allow the direct fusion of to .
(28) |
∎
Proposition 5.3.
The action of centrifugal loop on vector is the same vector with a -loop revolving around :
(29) |
Proof.
Under the action of the centrifugal loop on vector , summing over produces separated components of and . Note that is a number times , it commutes with . Further summing over gives us the same vector with a -loop revolving around .
(30) |
∎
Proposition 5.4.
The image of the idempotents under the algebra homomorphism is .
Proof.
Theorem 5.5.
is surjective. The element in that produces a given linear map in can be explicitly constructed as the following.
Proof.
Pick a basis , and the corresponding dual basis in , such that the normalization follows equation (1).
For any , with matrix elements , we can construct by
(32) |
We now check that :
(33) |
∎
Corollary 5.6.
is a simple -module.
Since , is also surjective. Taking the direct sum over on both sides we get an algebra homomorphism
Theorem 5.7.
is an algebra isomorphism.
Proof.
is surjective. Computing the dimension of and , which are both equal to , we know it is bijective. ∎
Remark 5.8.
Note that the element constructed in the proof of Theorem 5.5 has the property for . We know the inverse to is given by .
Corollary 5.9.
-
1.
is an algebra isomorphism.
-
2.
is a primitive central idempotent.
-
3.
span the center of .
6 Unitary structures: involution, inner product on module, unitary module
In physical application, we are usually interested in the case that the MTC is unitary. Denote the unitary structure of by . We can further equipped with a positive involution and the modules are moreover unitary.
Definition 6.1.
has an involution , defined by
(34) |
Remark 6.2.
We give a graphical proof that is indeed an involution. It is anti-linear by the anti-linearity of . It squares to identity, i.e.,
(35) |
It is also an algebra antihomomorphism, i.e.,
(36) |
graphically,
(37) |
Proposition 6.3.
The anitilinear map is a positive involution map, i.e.,
(38) |
Proof.
(39) |
In particular, the term in the above sum is 0, i.e.
(40) |
∎
Remark 6.4.
The positive involution also implies that the algebra is semisimple. This is the approach of Ma to argue the semisimplicity of the torus algebra. Nonetheless, we have provided a proof for semisimplicity independent from unitary structure and positive involution.
Definition 6.5.
The inner product on is defined by
(41) |
Proposition 6.6.
is an unitary module, i.e.,
(42) |
equivalently
Proof.
(43) |
∎
7 Punctured modular transformation
We consider the modular transformation on the punctured torus in this section. To begin with, we define a linear map for each simple via
(44) |
It has the inverse
(45) |
which can be checked explicitly
(46) |
To see that such linear map physically corresponds to ``rotating" the (square presentation of) torus by , we note that on the vector space one can associate a different action, by ``rotating" the elements in by . We denote such -module by , with the same underlying vector space as , and the following action
Theorem 7.1.
is a -module map from to , i.e., for any and ,
(47) |
Proof.
(48) |
∎
Invoking Theorem 5.5 we can construct an element
(49) |
such that , and further an element . More precisely,
(50) |
such that for any , one has . is the preimage of under the isomorphism .
Since are invertible, we conclude that is also invertible in with . The rotated action of is in fact induced by conjugation by :
Therefore, we can say that realizes the S-transformation on the punctured torus.
The linear map of charge conjugation operator for each simple is defined as
(51) |
Whereas the T-transformation on the punctured torus is defined via
(52) |
The corresponding elements of can be found similarly as the construction of above. We now have a complete set of operators , , and . One can easily show that , , and .
Remark 7.2.
In Kitaev_2006 , Kitaev also introduced punctured transformations. Denoting Kitaev's by and , they are related to our definition by . Kitaev's convention for punctured transformations is also adopted in wen2019distinguish ; Lan_2020 .
8 Examples
In this section, we demonstrate our systematic approach to identify all the central idempotents. We will decompose the torus algebras in two of the most commonly known anyon models, the chiral Ising anyon model and the Fibonacci anyon model, whose explicit data could be found in e.g., Bonderson_2008 . We will use the following basis vectors of torus algebras
(53) |
Since the two examples are all multiplicity-free where , there are no extra degrees of freedom on the vertices, and thus no vertex labels.
8.1 The Chiral Ising Anyon Model
In Ma Ma et al. decomposed the torus algebra of the chiral Ising anyon model by taking a ``quotient'' of . First, they listed all the basis vectors in terms of in . Second, by noticing the dimension of is , they recognized that consists of one block with dimension denoted as and another with dimension denoted as . Lastly, they find the projector of such that . They used the intuition that the centrifugal loop of a non-abelian anyon (which is a non-trivial element in ) should be set to its quantum dimension, and by taking a quotient forcing such conditions, they obtain the projector corresponding to the ground state subspace on the torus. Based on our general discussion on , indeed by setting one has
(54) |
Thus, this method is good enough to obtain the subspace where and other . However, it fails to produce the entire decomposition of unless there are only two blocks.
In our approach, we first obtain centrifugal loops of all anyon charges, then we can calculate all the idempotents using equation (20) to decompose completely.
With equation (20), the idempotents are
(56) | ||||
(57) | ||||
(58) |
We recovered the projector from Ma in a systematical way, which corresponds to the ground state subspace. It is easy to see restricting to the subspace of with and is equivalent to setting setting , i.e., forcing the centrifugal loop to be its quantum dimension .
8.2 Fibonacci anyon model
The Fibonacci model has anyon charges . First, we evaluate the centrifugal loops of all anyon charges in this model, and it is obvious that . The quantum dimension of is . The nontrivial entries of the -symbol are
The evaluation of centrifugal loop is
(59) |
Similar to the previous example, we define
The idempotents are
(60) | ||||
(61) |
Again, corresponds to the ground state subspace, and the subspace of with and is obtained by setting setting , i.e., forcing the centrifugal loop to be its quantum dimension .
Acknowledgements.
We would like to thank Holiverse Yang for the discussions. TL is supported by start-up funding from The Chinese University of Hong Kong, and by funding from Research Grants Council, University Grants Committee of Hong Kong (ECS No. 24304722).References
- (1) X. Ma, A. Malladi, Z. Wang, Z. Wang and X. Chen, Ground state degeneracy of the ising cage-net model, Phys. Rev. B 107 (2023) 085123.
- (2) A. Prem, S.-J. Huang, H. Song and M. Hermele, Cage-net fracton models, Physical Review X 9 (2019) .
- (3) X.G. Wen, Vacuum degeneracy of chiral spin states in compactified space, Physical Review B 40 (1989) 7387.
- (4) X.G. Wen and Q. Niu, Ground-state degeneracy of the fractional quantum Hall states in the presence of a random potential and on high-genus Riemann surfaces, Physical Review B 41 (1990) 9377.
- (5) X.G. Wen, Topological orders in rigid states, International Journal of Modern Physics B 04 (1990) 239.
- (6) X.G. Wen, Non-Abelian statistics in the fractional quantum hall states, Physical Review Letters 66 (1991) 802.
- (7) X.-G. Wen, Topological orders and edge excitations in FQH states, Advances in Physics 44 (1995) 405 [cond-mat/9506066].
- (8) B. Bartlett, C.L. Douglas, C.J. Schommer-Pries and J. Vicary, Modular categories as representations of the 3-dimensional bordism 2-category, 1509.06811.
- (9) P. Etingof, S. Gelaki, D. Nikshych and V. Ostrik, Tensor Categories, Mathematical Surveys and Monographs, American Mathematical Society (2016).
- (10) A. Kitaev, Anyons in an exactly solved model and beyond, Annals of Physics 321 (2006) 2.
- (11) B. Bakalov and A. Kirillov, Lectures on tensor categories and modular functors, Amer. Math. Soc. Univ. Lect. Ser. 21 (2001) .
- (12) A. Henriques and D. Penneys, Bicommutant categories from fusion categories, 1511.05226.
- (13) D. Green, P. Huston, K. Kawagoe, D. Penneys, A. Poudel and S. Sanford, Enriched string-net models and their excitations, 2305.14068.
- (14) A. Ocneanu, Chirality for operator algebras, Subfactors (Kyuzeso, 1993) (1994) 39.
- (15) L. Hardiman, A graphical approach to the drinfeld centre, 1911.07271.
- (16) T. Lan and X.-G. Wen, Topological quasiparticles and the holographic bulk-edge relation in (2+1)-dimensional string-net models, Physical Review B 90 (2014) 115119 [1311.1784].
- (17) X. Wen and X.-G. Wen, Distinguish modular categories and 2+1d topological orders beyond modular data: Mapping class group of higher genus manifold, 1908.10381.
- (18) T. Lan, X. Wen, L. Kong and X.-G. Wen, Gapped domain walls between 2+1d topologically ordered states, Physical Review Research 2 (2020) .
- (19) P. Bonderson, K. Shtengel and J. Slingerland, Interferometry of non-abelian anyons, Annals of Physics 323 (2008) 2709–2755.