Three Quantum Latin Squares of Order 6 with Cardinalities 13, 15, and 17

Zhipeng Xu · 2026

Read Three Quantum Latin Squares of Order 6 with Cardinalities 13, 15, and 17 by Zhipeng Xu (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-15540.txt.

Introduction

Zhipeng Xu builds three explicit quantum Latin squares of order six. Their sizes are thirteen, fifteen, and seventeen. Vectors that differ just by a global phase count as one. The size-thirteen build splits six-dimensional space into four- and two-dimensional parts. The two larger builds use two-dimensional Hadamard pairs on coordinate planes.

Catalog record

CREATORZhipeng Xu
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE16738
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.15540

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Three Quantum Latin Squares of Order 6 with Cardinalities 13, 15, and 17 Zhipeng Xu Abstract We give three explicit quantum Latin squares of order 6 6 , with cardinalities 13 13 , 15 15 , and 17 17 . Throughout, vectors differing only by a global phase are counted as identical. The cardinality- 13 13 construction is based on an orthogonal direct-sum decomposition ℂ 6 = ℂ 4 ⊕ ℂ 2 \mathbb{C}^{6}=\mathbb{C}^{4}\oplus\mathbb{C}^{2} . The cardinality- 15 15 and cardinality- 17 17 constructions are based on two-dimensional Hadamard pairs supported on coordinate planes. Keywords: Quantum Latin square; cardinality; order six; orthogonal basis; quantum combinatorics 1 Definitions and background Classical Latin squares go back at least to Euler’s work on Graeco-Latin squares [ 1 ] ; standard background can be found in the monographs and surveys on Latin squares and combinatorial designs [ 2 , 3 ] . The classical existence theory for mutually orthogonal Latin squares includes the disproof of Euler’s conjecture by Bose, Shrikhande, and Parker [ 4 ] . Quantum Latin squares were introduced by Musto and Vicary [ 7 ] as a quantum analogue of Latin squares, motivated in part by constructions of unitary error bases and related quantum-information protocols [ 5 , 6 ] . Subsequent work developed notions of orthogonality and quantum combinatorial designs [ 8 , 9 ] , including order-six quantum constructions related to Euler’s 36 officers problem [ 10 , 11 ] . The cardinality convention used below follows recent work on the cardinalities of quantum Latin squares [ 12 , 13 , 14 ] ; closely related maximal-cardinality questions are studied in [ 15 ] . Definition 1 (Quantum Latin square) . A quantum Latin square of order n , denoted by \mathrm{QLS}(n) , is an n\times n array \Phi=(\lvert\phi_{ij}\rangle)_{0\leq i,j\leq n-1} whose entries are unit vectors in \mathbb{C}^{n} , such that every row and every column of \Phi forms an orthonormal basis of \mathbb{C}^{n} . Two unit vectors are identified if they differ only by a global phase. That is, \lvert u\rangle\sim\lvert v\rangle\quad\Longleftrightarrow\quad\lvert u\rangle=e^{\mathrm{i}\theta}\lvert v\rangle\quad\text{for some }\theta\in\mathbb{R}. The cardinality of a quantum Latin square \Phi is the number of distinct vectors appearing in the array, modulo this equivalence relation: \operatorname{card}(\Phi)=\#\{[\lvert\phi_{ij}\rangle]:0\leq i,j\leq n-1\}. 2 A QLS(6) of cardinality 13 Let \lvert 0\rangle,\lvert 1\rangle,\lvert 2\rangle,\lvert 3\rangle,\lvert 4\rangle,\lvert 5\rangle be the computational basis of \mathbb{C}^{6} . Decompose \mathbb{C}^{6}=U\oplus V, where U=\operatorname{span}\{\lvert 0\rangle,\lvert 1\rangle,\lvert 2\rangle,\lvert 3\rangle\},\qquad V=\operatorname{span}\{\lvert 4\rangle,\lvert 5\rangle\}. In U , define a=\lvert 0\rangle,\qquad b=\lvert 1\rangle,\qquad c=\lvert 2\rangle,\qquad d=\lvert 3\rangle. Set e=\frac{c+d}{\sqrt{2}},\qquad v=\frac{c-d}{\sqrt{2}}. Now define f=\frac{b+v}{\sqrt{2}}=\frac{\lvert 1\rangle}{\sqrt{2}}+\frac{\lvert 2\rangle-\lvert 3\rangle}{2},\qquad g=\frac{b-v}{\sqrt{2}}=\frac{\lvert 1\rangle}{\sqrt{2}}-\frac{\lvert 2\rangle-\lvert 3\rangle}{2}, and h=\frac{a+v}{\sqrt{2}}=\frac{\lvert 0\rangle}{\sqrt{2}}+\frac{\lvert 2\rangle-\lvert 3\rangle}{2},\qquad r=\frac{a-v}{\sqrt{2}}=\frac{\lvert 0\rangle}{\sqrt{2}}-\frac{\lvert 2\rangle-\lvert 3\rangle}{2}. In V , define x=\lvert 4\rangle,\qquad y=\lvert 5\rangle, and z=\frac{x+y}{\sqrt{2}},\qquad w=\frac{x-y}{\sqrt{2}}. For reference, the entries used in this construction have the following coordinate representations with respect to the ordered computational basis (\lvert 0\rangle,\lvert 1\rangle,\lvert 2\rangle,\lvert 3\rangle,\lvert 4\rangle,\lvert 5\rangle) : \small\begin{array}[]{lll}a=(1,0,0,0,0,0)^{\mathsf{T}},&b=(0,1,0,0,0,0)^{\mathsf{T}},&c=(0,0,1,0,0,0)^{\mathsf{T}},\\[5.69054pt] d=(0,0,0,1,0,0)^{\mathsf{T}},&e=(0,0,\frac{1}{\sqrt{2}},\frac{1}{\sqrt{2}},0,0)^{\mathsf{T}},&f=(0,\frac{1}{\sqrt{2}},\frac{1}{2},-\frac{1}{2},0,0)^{\mathsf{T}},\\[5.69054pt] g=(0,\frac{1}{\sqrt{2}},-\frac{1}{2},\frac{1}{2},0,0)^{\mathsf{T}},&h=(\frac{1}{\sqrt{2}},0,\frac{1}{2},-\frac{1}{2},0,0)^{\mathsf{T}},&r=(\frac{1}{\sqrt{2}},0,-\frac{1}{2},\frac{1}{2},0,0)^{\mathsf{T}},\\[5.69054pt] x=(0,0,0,0,1,0)^{\mathsf{T}},&y=(0,0,0,0,0,1)^{\mathsf{T}},&z=(0,0,0,0,\frac{1}{\sqrt{2}},\frac{1}{\sqrt{2}})^{\mathsf{T}},\\[5.69054pt] w=(0,0,0,0,\frac{1}{\sqrt{2}},-\frac{1}{\sqrt{2}})^{\mathsf{T}}.&&\end{array} Consider the following 6\times 6 array: \Phi_{13}=\begin{pmatrix}c&a&x&y&b&d\\ d&w&a&b&z&c\\ y&f&g&x&e&a\\ a&g&f&e&w&z\\ x&e&y&h&r&b\\ b&z&e&r&h&w\end{pmatrix}. Proposition 1 . The array \Phi_{13} is a \mathrm{QLS}(6) with cardinality 13 . Proof. Define three orthonormal bases of U by B_{0}=\{a,b,c,d\},\qquad B_{1}=\{a,e,f,g\},\qquad B_{2}=\{b,e,h,r\}. The set B_{0} is the computational basis of U . Since e=\frac{c+d}{\sqrt{2}},\qquad v=\frac{c-d}{\sqrt{2}}, the pair \{e,v\} is an orthonormal basis of \operatorname{span}\{c,d\} . The pair \{f,g\} is obtained from \{b,v\} by the Hadamard rotation f=\frac{b+v}{\sqrt{2}},\qquad g=\frac{b-v}{\sqrt{2}}. Therefore \{f,g\} is an orthonormal basis of \operatorname{span}\{b,v\} , and it is orthogonal to both a and e . Hence B_{1} is an orthonormal basis of U . Similarly, the pair \{h,r\} is obtained from \{a,v\} by the Hadamard rotation h=\frac{a+v}{\sqrt{2}},\qquad r=\frac{a-v}{\sqrt{2}}. Thus \{h,r\} is an orthonormal basis of \operatorname{span}\{a,v\} , and it is orthogonal to both b and e . Hence B_{2} is also an orthonormal basis of U . Define two orthonormal bases of V by S_{0}=\{x,y\},\qquad S_{1}=\{z,w\}. The set S_{0} is the computational basis of V , and S_{1} is obtained from S_{0} by a two-dimensional Hadamard rotation. Since U\perp V , every union B_{i}\cup S_{j} is an orthonormal basis of \mathbb{C}^{6}=U\oplus V . The six rows of \Phi_{13} are of the following types: \begin{array}[]{c|c}\text{row}&\text{orthonormal basis}\\ \hline\cr R_{1}&B_{0}\cup S_{0}\\ R_{2}&B_{0}\cup S_{1}\\ R_{3}&B_{1}\cup S_{0}\\ R_{4}&B_{1}\cup S_{1}\\ R_{5}&B_{2}\cup S_{0}\\ R_{6}&B_{2}\cup S_{1}\end{array} and the six columns are of the following types: \begin{array}[]{c|c}\text{column}&\text{orthonormal basis}\\ \hline\cr C_{1}&B_{0}\cup S_{0}\\ C_{2}&B_{1}\cup S_{1}\\ C_{3}&B_{1}\cup S_{0}\\ C_{4}&B_{2}\cup S_{0}\\ C_{5}&B_{2}\cup S_{1}\\ C_{6}&B_{0}\cup S_{1}\end{array} Thus every row and every column is an orthonormal basis of \mathbb{C}^{6} , so \Phi_{13} is a \mathrm{QLS}(6) . The vectors appearing in \Phi_{13} are exactly \mathcal{V}(\Phi_{13})=\{a,b,c,d,e,f,g,h,r,x,y,z,w\}. This set has 13 elements. No two of them differ only by a global phase: vectors from U and vectors from V have orthogonal supports; the four vectors x,y,z,w in V are pairwise not scalar multiples; and the nine vectors a,b,c,d,e,f,g,h,r in U have different supports or different relative signs on a common support. Therefore \operatorname{card}(\Phi_{13})=13. ∎ 3 A QLS(6) of cardinality 15 For later comparison with the cardinality- 17 example, we first give a coordinate-plane construction of cardinality 15 . Let e_{i}=\lvert i\rangle,\qquad 0\leq i\leq 5. For 0\leq i<j\leq 5 , write p_{ij}=\frac{e_{i}+e_{j}}{\sqrt{2}},\qquad q_{ij}=\frac{e_{i}-e_{j}}{\sqrt{2}}. Thus \{p_{ij},q_{ij}\} is an orthonormal basis of the coordinate plane \operatorname{span}\{e_{i},e_{j}\} . We use the five coordinate pairs (0,1),\qquad(0,4),\qquad(1,2),\qquad(2,3),\qquad(2,4), and the five computational basis vectors e_{0},\qquad e_{1},\qquad e_{3},\qquad e_{4},\qquad e_{5}. The complete list of vectors used in the construction is therefore \begin{gathered}p_{01},q_{01},\quad p_{04},q_{04},\quad p_{12},q_{12},\quad p_{23},q_{23},\quad p_{24},q_{24},\\ e_{0},e_{1},e_{3},e_{4},e_{5}.\end{gathered} Equivalently, the fifteen vectors used in \Phi_{15} have the following coordinate forms: \small\begin{array}[]{lll}p_{01}=(\frac{1}{\sqrt{2}},\frac{1}{\sqrt{2}},0,0,0,0)^{\mathsf{T}},&q_{01}=(\frac{1}{\sqrt{2}},-\frac{1}{\sqrt{2}},0,0,0,0)^{\mathsf{T}},&p_{04}=(\frac{1}{\sqrt{2}},0,0,0,\frac{1}{\sqrt{2}},0)^{\mathsf{T}},\\[5.69054pt] q_{04}=(\frac{1}{\sqrt{2}},0,0,0,-\frac{1}{\sqrt{2}},0)^{\mathsf{T}},&p_{12}=(0,\frac{1}{\sqrt{2}},\frac{1}{\sqrt{2}},0,0,0)^{\mathsf{T}},&q_{12}=(0,\frac{1}{\sqrt{2}},-\frac{1}{\sqrt{2}},0,0,0)^{\mathsf{T}},\\[5.69054pt] p_{23}=(0,0,\frac{1}{\sqrt{2}},\frac{1}{\sqrt{2}},0,0)^{\mathsf{T}},&q_{23}=(0,0,\frac{1}{\sqrt{2}},-\frac{1}{\sqrt{2}},0,0)^{\mathsf{T}},&p_{24}=(0,0,\frac{1}{\sqrt{2}},0,\frac{1}{\sqrt{2}},0)^{\mathsf{T}},\\[5.69054pt] q_{24}=(0,0,\frac{1}{\sqrt{2}},0,-\frac{1}{\sqrt{2}},0)^{\mathsf{T}},&e_{0}=(1,0,0,0,0,0)^{\mathsf{T}},&e_{1}=(0,1,0,0,0,0)^{\mathsf{T}},\\[5.69054pt] e_{3}=(0,0,0,1,0,0)^{\mathsf{T}},&e_{4}=(0,0,0,0,1,0)^{\mathsf{T}},&e_{5}=(0,0,0,0,0,1)^{\mathsf{T}}.\end{array} Consider the following 6\times 6 array: \Phi_{15}=\begin{pmatrix}p_{04}&p_{12}&q_{12}&e_{3}&q_{04}&e_{5}\\ p_{23}&e_{5}&e_{4}&p_{01}&q_{23}&q_{01}\\ e_{1}&e_{0}&e_{3}&p_{24}&e_{5}&q_{24}\\ q_{04}&q_{12}&p_{12}&e_{5}&p_{04}&e_{3}\\ q_{23}&e_{4}&e_{5}&q_{01}&p_{23}&p_{01}\\ e_{5}&e_{3}&e_{0}&q_{24}&e_{1}&p_{24}\end{pmatrix}. Proposition 2 . The array \Phi_{15} is a \mathrm{QLS}(6) with cardinality 15 . Proof. Each row and each column is a disjoint union of coordinate-plane Hadamard pairs and computational basis vectors. In the following table, (ij) denotes the pair \{p_{ij},q_{ij}\} and \{k\} denotes the singleton \{e_{k}\} . The six rows have the decompositions \begin{array}[]{c|c}\text{row}&\text{coordinate decomposition}\\ \hline\cr R_{1}&(04)\cup(12)\cup\{3\}\cup\{5\}\\ R_{2}&(23)\cup(01)\cup\{4\}\cup\{5\}\\ R_{3}&(24)\cup\{0\}\cup\{1\}\cup\{3\}\cup\{5\}\\ R_{4}&(04)\cup(12)\cup\{3\}\cup\{5\}\\ R_{5}&(23)\cup(01)\cup\{4\}\cup\{5\}\\ R_{6}&(24)\cup\{0\}\cup\{1\}\cup\{3\}\cup\{5\}\end{array} and the six columns have the decompositions \begin{array}[]{c|c}\text{column}&\text{coordinate decomposition}\\ \hline\cr C_{1}&(04)\cup(23)\cup\{1\}\cup\{5\}\\ C_{2}&(12)\cup\{0\}\cup\{3\}\cup\{4\}\cup\{5\}\\ C_{3}&(12)\cup\{0\}\cup\{3\}\cup\{4\}\cup\{5\}\\ C_{4}&(01)\cup(24)\cup\{3\}\cup\{5\}\\ C_{5}&(04)\cup(23)\cup\{1\}\cup\{5\}\\ C_{6}&(01)\cup(24)\cup\{3\}\cup\{5\}.\end{array} In each row and in each column, the displayed coordinate pieces are mutually orthogonal and together cover the coordinate set \{0,1,2,3,4,5\} . Hence each row and each column is an orthonormal basis of \mathbb{C}^{6} . Thus \Phi_{15} is a \mathrm{QLS}(6) . The vectors appearing in \Phi_{15} are exactly \begin{gathered}p_{01},q_{01},\quad p_{04},q_{04},\quad p_{12},q_{12},\quad p_{23},q_{23},\quad p_{24},q_{24},\\ e_{0},e_{1},e_{3},e_{4},e_{5}.\end{gathered} The five coordinate pairs contribute 10 Hadamard-pair vectors, and the five listed computational basis vectors also occur. No Hadamard-pair vector is a scalar multiple of a computational basis vector, and two Hadamard-pair vectors from different coordinate pairs have different supports. For a fixed coordinate pair (i,j) , the vectors p_{ij} and q_{ij} have different relative signs, so they are not scalar multiples. Therefore the number of phase-classes is 10+5=15. Hence \operatorname{card}(\Phi_{15})=15 . ∎ 4 A QLS(6) of cardinality 17 For 0\leq i<j\leq 5 , define the two Hadamard-pair vectors p_{ij}=\frac{\lvert i\rangle+\lvert j\rangle}{\sqrt{2}},\qquad q_{ij}=\frac{\lvert i\rangle-\lvert j\rangle}{\sqrt{2}}. The pair \{p_{ij},q_{ij}\} is an orthonormal basis of the coordinate plane \operatorname{span}\{\lvert i\rangle,\lvert j\rangle\} . We use the seven coordinate pairs (0,1),\quad(0,3),\quad(0,4),\quad(1,2),\quad(1,5),\quad(2,3),\quad(2,4). For completeness, the seventeen vectors used in \Phi_{17} are listed below in coordinates: \small\begin{array}[]{lll}p_{01}=(\frac{1}{\sqrt{2}},\frac{1}{\sqrt{2}},0,0,0,0)^{\mathsf{T}},&q_{01}=(\frac{1}{\sqrt{2}},-\frac{1}{\sqrt{2}},0,0,0,0)^{\mathsf{T}},&p_{03}=(\frac{1}{\sqrt{2}},0,0,\frac{1}{\sqrt{2}},0,0)^{\mathsf{T}},\\[5.69054pt] q_{03}=(\frac{1}{\sqrt{2}},0,0,-\frac{1}{\sqrt{2}},0,0)^{\mathsf{T}},&p_{04}=(\frac{1}{\sqrt{2}},0,0,0,\frac{1}{\sqrt{2}},0)^{\mathsf{T}},&q_{04}=(\frac{1}{\sqrt{2}},0,0,0,-\frac{1}{\sqrt{2}},0)^{\mathsf{T}},\\[5.69054pt] p_{12}=(0,\frac{1}{\sqrt{2}},\frac{1}{\sqrt{2}},0,0,0)^{\mathsf{T}},&q_{12}=(0,\frac{1}{\sqrt{2}},-\frac{1}{\sqrt{2}},0,0,0)^{\mathsf{T}},&p_{15}=(0,\frac{1}{\sqrt{2}},0,0,0,\frac{1}{\sqrt{2}})^{\mathsf{T}},\\[5.69054pt] q_{15}=(0,\frac{1}{\sqrt{2}},0,0,0,-\frac{1}{\sqrt{2}})^{\mathsf{T}},&p_{23}=(0,0,\frac{1}{\sqrt{2}},\frac{1}{\sqrt{2}},0,0)^{\mathsf{T}},&q_{23}=(0,0,\frac{1}{\sqrt{2}},-\frac{1}{\sqrt{2}},0,0)^{\mathsf{T}},\\[5.69054pt] p_{24}=(0,0,\frac{1}{\sqrt{2}},0,\frac{1}{\sqrt{2}},0)^{\mathsf{T}},&q_{24}=(0,0,\frac{1}{\sqrt{2}},0,-\frac{1}{\sqrt{2}},0)^{\mathsf{T}},&e_{3}=(0,0,0,1,0,0)^{\mathsf{T}},\\[5.69054pt] e_{4}=(0,0,0,0,1,0)^{\mathsf{T}},&e_{5}=(0,0,0,0,0,1)^{\mathsf{T}}.&\end{array} Consider the following 6\times 6 array: \Phi_{17}=\begin{pmatrix}p_{04}&p_{12}&q_{12}&\lvert 3\rangle&q_{04}&\lvert 5\rangle\\ p_{23}&\lvert 5\rangle&\lvert 4\rangle&p_{01}&q_{23}&q_{01}\\ p_{15}&p_{03}&q_{03}&p_{24}&q_{15}&q_{24}\\

Read the complete text