Abstract
Quantum jump codes are quantum codes which correct errors caused by quantum jumps. A t-spontaneous emission error design (t-SEED) was introduced by Beth et al. in 2003 to construct quantum jump codes. The number of designs (dimension) in a t-SEED corresponds to the number of orthogonal basis states in a quantum jump code. Denote by \(\overline{M}(t,k,v)\) the largest possible dimension m for which a nondegenerate t-(v, k; m) SEED exists. \(\overline{M}(2,3,v)\) has been determined completely, which is based on a great deal of research on large sets of Steiner triple systems and large sets of pairwise disjoint compatible 2-(v, 3, 1) packings. For \(k=4\), the upper bounds on dimensions of 2-(v, 4; m) SEEDs were also demonstrated and the corresponding leave graphs were investigated in Zhou and Chang (J Combin Des 24:439–460, 2016). In this paper we turn our attention to the case \(t=2\) and \(k=5\). We present general upper bounds on the dimensions of 2-(v, 5; m) SEEDs and describe the concrete leave graphs of the 2-SEEDs attaining the stated upper bounds.







Similar content being viewed by others
References
Alber G., Beth T., Charnes C., Delgado A., Grassl M., Mussinger M.: Stabilizing distinguishable qubits against spontaneous decay by detected-jump correcting quantum codes. Phys. Rev. Lett. 86, 4402–4405 (2001).
Beth T., Charnes C., Grassl M., Alber G., Delgado A., Mussinger M.: A new class of designs which protect against quantum jumps. Des. Codes Cryptogr. 29, 51–70 (2003).
Calderbank A.R., Shor P.W.: Good quantum error-correcting codes exist. Phys. Rev. A 54, 1098–1105 (1996).
Calderbank A.R., Rains E.M., Shor P.W., Sloane N.J.A.: Quantum error correction and orthogonal geometry. Phys. Rev. Lett. 78, 405–408 (1997).
Cao H., Ji L., Zhu L.: Large sets of disjoint packings on \(6k + 5\) points. J. Combin. Theory (A) 108, 169–183 (2004).
Chang Y., Ji L., Zheng H.: A completion of LS\((2^n4^1)\). Discret. Math. 340, 1080–1085 (2017).
Ji L.: Existence of large sets of disjoint group-divisible designs with block size three and type \(2^n4^1\). J. Combin. Des. 13, 302–312 (2005).
Ji L.: More results on perfect \((3, 3, 6k + 4; 6k-2)\)-threshold schemes. J. Combin. Des. 15, 151–166 (2007).
Lei J.: Completing the spectrum for LGDD\((m^v)\). J. Combin. Des. 5, 1–11 (1997).
Lin Y., Jimbo M.: Extremal properties of \(t\)-SEEDs and recursive constructions. Des. Codes Cryptogr. 73, 805–823 (2014).
Lu J.: On large sets of disjoint Steiner triple systems, I–III. J. Combin. Theory (A) 34, 140–182 (1983).
Lu J.: On large sets of disjoint Steiner triple systems, IV–V. J. Combin. Theory (A) 37, 136–192 (1984).
Shor P.W.: Scheme for reducing decoherence in quantum computer memory. Phys. Rev. A 52, R2493–R2496 (1995).
Steane A.: Multiple particle interference and quantum error correction. Proc. R. Soc. Lond. Ser. A 452, 2551–2577 (1996).
Steane A.: Error correcting codes in quantum theory. Phys. Rev. Lett. 77, 793–797 (1996).
Teirlinck L.: A completion of Lu’s determination of the spectrum for large sets of disjoint Steiner triple systems. J. Combin. Theory (A) 57, 302–305 (1991).
Zhou J., Chang Y.: Bounds on the dimensions of \(2\)-spontaneous emission error designs. J. Combin. Des. 24, 439–460 (2016).
Author information
Authors and Affiliations
Corresponding author
Additional information
Communicated by L. Teirlinck.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supported by NSFC Grants 11571034 and 11971053.
Appendix: continued proof of Lemma 3.8
Appendix: continued proof of Lemma 3.8
\(\underline{(4) |E(G[Y])|=3.}\) Then G[Y] has degree sequence (2, 2, 1, 1) or (2, 2, 2, 0) by (i), corresponding to a path or a cycle of length 3. Let \(E_Y^1=\{12,23,34\}\) and \(E_Y^2=\{12,23,13\}\). Denote the degree sum \(D_Z=\sum _{z\in Z}d_{G[Z]}(z)\). So \(D_Z=12\) by (i). Since \(|Z|=6\), G[Z] has all vertices of degree 2 or it contains at least one vertex of degree 1. We deal with these two possibilities.
Firstly let G[Z] be 2-regular. Note that this is impossible by (ii) if G[Y] has a vertex of degree 1. So we let \(E(G[Y])=E_Y^2\). It is easy to know that G[Z] consists of two cycles of length 3 or a cycle of length 6. Let (a) \(E(G[Z])=E_Z^1=\{ab,ac,bc,de,df,ef\}\), (b) \(E(G[Z])=E_Z^2=\{ab,af,bc,cd,de,ef\}\). Note that \(4z\in E\) for any \(z\in Z\).
- (a)
\(E(G[Z])=E_Z^1\). We assume \(N_{G}(a)\cap Y=\{1,2,4\}\). Then \(Y_{a}=\{3,d,e,f\}\) and \(de,df,ef\in E\). Moreover, the vertex 3 must be adjacent to at least one of d, e and f, that is, \(G[Y_{a}]\) contains a vertex of degree 3, a contradiction to (i).
- (b)
\(E(G[Z])=E_Z^2\). For \(i\in \{1,2,3\}\), let \(Y_i=\{0,4,\alpha _i,\beta _i\}\) where \(\alpha _i,\beta _i\in Z\). Note \(0\alpha _i,0\beta _i,4\alpha _i,\)\(4\beta _i\in E\). We have \(\alpha _i\beta _i\notin E\) by (i). Due to the symmetry of vertices in G[Z], we assume that (b1) \(N_{G}(1)\cap Z=\{a,c,d,e\}\) or (b2) \(N_{G}(1)\cap Z=\{a,c,d,f\}\). For (b1), \(Y_{b}=\{1,d,e,f\}\) and \(1e,de,ef\in E\), so \(d_{G[Y_b]}(e)=3\), a contradiction to (i). Thus case (b2) holds. Then \(2b,2e,3b,3e\in E\) by (vi). Now each of a, c, d, f has to be adjacent to exactly one of 2 and 3, two symmetric vertices. Three possibilities need to be considered: \(\{2a,2c,3d,3f\}\subset E\), \(\{2a,2d,3c,3f\}\subset E\), or \(\{2a,2f,3c,3d\}\subset E\). By the very first reasoning of this case, \(\{2a,2f,3c,3d\}\subset E\) is impossible by considering \(Y_2,Y_3\) because \(cd,af\in E\). Also note that if \(\{2a,2c,3d,3f\}\subset E\) then \(N_{G}(2)\cap Z=\{a,b,c,e\}\) is isomorphic to (b1) case. So we let \(\{2a,2d,3c,3f\}\subset E\). Therefore, G[X] is isomorphic to \(H_{17}\) and we check that it is a qualified subgraph.
Secondly we let G[Z] contain a vertex of degree 1, say f. Suppose that \(d_{G[Z]}(e)=d_0\) and \(ef\in E\). Based on the cases (1)–(3), we assume that the subgraph \(G[Y_x]\) has degree sum 6 or 8 with degree sequence (2, 2, 1, 1), (2, 2, 2, 0), or (2, 2, 2, 2) for any vertex \(x\in X\). Note that \(G[Y_f]\) is the induced subgraph of G[Z] by \(Z\setminus \{e,f\}\). We have the degree sum of \(G[Y_f]\)\(D_{Y_f}=D_Z-d_0-1-(d_0-1)=12-2d_0\). Hence \(D_{Y_f}=12-2d_0=6,8\), yielding \(d_0=3,2\), respectively.
When \(d_0=2\), \(D_{Y_f}=8\), meaning that \(G[Y_f]\) forms a 4-cycle by (i), say \((a\ b\ c\ d)\). Further let \(ae\in E\). Then the edge set \(E(G[Z])=\{ab,ad,ae,bc,cd,ef\}\). When G[Y] is a cycle of length 3, let \(E(G[Y])=E_Y^2\). W.l.o.g. assume that \(1a\in E\), then \(Y_{a}=\{2,3,c,f\}\). Note that \(23,2f,3f\in E\) and c must be adjacent to at least one of 2 and 3. So \(d_{G[Y_a]}(2)= 3\) or \(d_{G[Y_a]}(3)= 3\), contradicting to (i). When G[Y] is a path of length 3, let \(E(G[Y])=E_Y^1\). In this case, \(N_{G}(f)\cap Y=\{1,2,3,4\}\), \(N_{G}(1)\cap Z=N_{G}(4)\cap Z=\{b,c,d,e,f\}\), and \(2a,3a\in E\) by (ii) and (vi). Now each of vertices 2 and 3 has two neighbors undetermined and each of b, c, d, e has one. Let \(Y_2=\{0,4,\alpha ,\beta \}\) where \(\alpha ,\beta \in \{b,c,d,e\}\). Since \(0\alpha ,0\beta ,4\alpha ,4\beta \in E\), we have \(\alpha \beta \notin E\) by (i). Similarly, let \(Y_3=\{0,4,\alpha ',\beta '\}\) where \(\alpha ',\beta '\in \{b,c,d,e\}\). Then \(\alpha '\beta '\notin E\). So we assume \(2b,2d,3c,3e\in E\) and form a graph \(H_{18}\), a qualified graph.
When \(d_0=3\), \(D_{Y_f}=6\), meaning that \(G[Y_f]\) has degree sequence (2, 2, 2, 0) or (2, 2, 1, 1). So we let \(E(G[Y_f])=E_f^1=\{ab,ac,bc\}\) or \(E(G[Y_f])=E_f^2=\{ab,bc,cd\}\).
Firstly consider the case \(E(G[Y_f])=E_f^1\). Then \(de\in E\) by (i). Hence we assume \(E(G[Z])=\{ab,ac,ae,bc,de,ef\}\). Then d and f are adjacent to every vertex of Y by (vi). If E(G[Y] is a cycle of length 3, then let \(E(G[Y])=E_Y^2\). Obviously \(Y_{a}=\{i,j,d,f\}\) where \(\{i,j\}\subseteq \{1,2,3\}\) by (vi). Note that \(id,jd,if,jf\in E\) and \(ij\in E\), contradicting to (i). So we let G[Y] be a path of length 3 with edge set \(E_Y^1\). The nonadjacent vertex of 1 and 4 in Z is a or e by (ii). We consider whether or not \(N_{G}(1)\cap Z=N_{G}(4)\cap Z\). If \(N_{G}(1)\cap Z=\{b,c,d,e,f\}\) and \(N_{G}(4)\cap Z=\{a,b,c,d,f\}\). Then \(2a,3e\in E\) by (ii) and hence \(N_G(e)=\{0,1,3,a,d,f\}\) so that \(Y_{e}=\{2,4,b,c\}\). Note that \(4b,4c,bc\in E\) and 2 must be adjacent to one of b and c. This contradicts to (i). If \(N_{G}(1)\cap Z=N_{G}(4)\cap Z=\{a,b,c,d,f\}\), then \(Y_a=\{2,3,d,f\}\). But \(23,2d,2f,3d,3f\in E\), contradicting to (i). Finally, if \(N_{G}(1)\cap Z=N_{G}(4)\cap Z=\{b,c,d,e,f\}\), then \(2a,3a\in E\). Now we form the edges 2b and 3c to form a 6-regular graph (noting the symmetry of b and c). However, considering the vertex 1 gives \(E(G[Z_1])=\{2b,2d,2f,bc,de,ef\}\). Thus \(d_{G[Z_1]}(c)=1\), \(d_{G[Z_1]}(b)=2\), and \(bc\in E\). An analogue of this has been considered in the previous subcase \(d_0=2\).
Secondly we let \(E(G[Y_f])=E_f^2\). We consider three possibilities of the two neighbors of e in Z other than f: (a) \(ae,de\in E\), (b) \(be,ce\in E\), (c) \(ae,ce\in E\) or \(ae,be\in E\).
- (a)
\(ae,de\in E\). If \(E(G[Y])=E_Y^1\), by (ii), \(1e,4e\notin E\) and \(2e,3e\in E\). But then \(Y_e=\{1,4,b,c\}\) and \(1b,4b,bc\in E\), contradicting to (i). If \(E(G[Y])=E_Y^2\), then \(N_G(f)=Y\cup \{0,e\}\) and \(N_G(4)=Z\) by (vi). W.l.o.g. suppose that \(1b,2b\in E\), then \(N_G(b)=\{0,1,2,4,a,c\}\) and \(Y_{b}=\{3,d,e,f\}\). Note that \(3f,de,ef\in E\) in \(G[Y_{b}]\), so \(3e\notin E\) by (i) and thus \(N_G(3)\cap Z=\{a,c,d,f\}\). Now we only need to determine the last two neighbors in \(\{a,c,d,e\}\) for each of 1 and 2. Note that for \(i=1,2\) we have \(Y_i=\{0,4,\alpha _i,\beta _i\}\) where \(\alpha _i,\beta _i\in Z\) and \(\{0\alpha _i,0\beta _i,4\alpha _i,4\beta _i\}\subset E\). Hence \(\alpha _i\beta _i\notin E\) by (i). It follows that we may assume \(Y_1=\{0,4,a,d\}\) and \(Y_2=\{0,4,c,e\}\). Now we have that \(Y_a=\{1,c,d,f\}\), \(E(Y_a)=\{1c,1f,cd\}\), \(Z_a=\{0,2,3,4,b,e\}\), and \(G[Z_a]=\{0b,0e,23,2b,4b,4e\}\). Thus this is reduced to the parallel case \(|E(Y)|=3\) and \(d_0=2\), which has been considered previously.
- (b)
\(be,ce\in E\). In either case of \(E(G[Y])=E_Y^1\) or \(E(G[Y])=E_Y^2\), we may assume \(Y_b=\{2,3,d,f\}\). This is obvious if \(E(G[Y])=E_Y^2\). If \(E(G[Y])=E_Y^1\), then \(N_G(a)\cap Y=N_G(d)\cap Y=N_G(f)\cap Y=Y\) by (vi) and thus \(|N_G(i)\cap \{b,c,e\}|=1\) for \(i=2,3\). So at least one of \(\{b,c,e\}\) is adjacent to neither 2 nor 3 and thence we may also assume \(Y_b=\{2,3,d,f\}\). Clearly \(23,2d,2f\in E\), contradicting to (i).
- (c)
\(ae,ce\in E\) or \(ae,be\in E\). For the latter case, \(Y_d=\{a,b,e,f\}\) and \(d_{G[Y_d]}(e)=3\), contradicting to (i). Hence we let \(ae,ce\in E\). When \(E(G[Y])=E_Y^2\), we have \(Y_{e}=\{i,j,b,d\}\) where \(\{i,j\}\subseteq \{1,2,3\}\) and \(id,jd,ij\in E\), but b must be adjacent to at least one of i and j, which contradicts to (i). When \(E(G[Y])=E_Y^1\), the (only) nonadjacent vertex of 1 or 4 in G[Z] is c or e by (ii). Note that c and e are symmetric in G[Z]. If \(N_{G}(1)\cap Z=N_{G}(4)\cap Z=\{a,b,c,d,f\}\), then \(2e,3e\in E\) by (ii). Now \(Y_c=\{2,3,a,f\}\). Note that \(23,2f,3f\in E\) and a must be adjacent to one vertex of 2 and 3, which contradicts to (i). Thus we assume w.l.o.g. \(N_{G}(1)\cap Z=\{a,b,d,e,f\}\) and \(N_{G}(4)\cap Z=\{a,b,c,d,f\}\). Then \(2c,3e\in E\), \(N_G(d)\cap Y=N_G(f)\cap Y=Y\) by (ii) and (vi). Finally using (vi) leaves each vertex 2, 3, a, b only one neighbor to be determined. If \(2b,3a\in E\), then \(Y_2=\{0,4,a,e\}\) and we have a contradiction \(d_{G[Y_2]}(a)=3\). So we must have \(2a,3b\in E\). Therefore, G[X] is isomorphic to \(H_{19}\), a qualified subgraph.
\(\underline{(5) |E(G[Y])|=4.}\) Then G[Y] forms a 4-cycle, say \((1\ 2\ 3\ 4)\), by (i). Based on the cases (1)–(4), we assume that the subgraph \(G[Y_x]\) has degree sequence (2, 2, 2, 2) for any vertex \(x\in X\). Since \(|Z|=6\), the minimum degree in G[Z] is 1 or 2. We deal with these two possibilities.
Firstly let G[Z] contain a vertex of degree 1, say f. Suppose that \(d_{G[Z]}(e)=d_0\) and \(ef\in E\). In particular we let \(G[Y_f]\) has edge set \(E'=\{ab,bc,cd,ad\}\). We have \(8=\sum _{y\in Y_f}d_{G[Y_f]}(y)=\sum _{z\in Z}d_{G[Z]}(z)-d_0-1-(d_0-1)=14-2d_0\), yielding \(d_0=3\). So we have two forms of G[Z] by adding two edges to \(E'\cup \{ef\}\): (a) \(ae,ce\in E\) or (b) \(ce,de\in E\). For (a), it is easy to see that \(Y_b=\{y,d,e,f\}\) for some \(y\in \{1,2,3,4\}\) and \(d_{G[Y_b]}(d)\le 1\), contradicting to our assumption. For (b), w.l.o.g. assume that \(N_{G}(d)\cap Y=\{1,2\}\) or \(N_{G}(d)\cap Y=\{1,3\}\). If \(N_{G}(d)\cap Y=\{1,2\}\), then \(Y_{d}=\{3,4,b,f\}\) and \(34,3f,4f\in E\), so \(G[Y_{d}]\) cannot be a 4-cycle, a contradiction. So we let \(N_{G}(d)\cap Y=\{1,3\}\). Then \(Y_{d}=\{2,4,b,f\}\). Note that \(G[Y_{d}]\) is a 4-cycle and \(2f,4f\in E\), then \(2b,4b\in E\). Now b has to be adjacent to one of 1 and 3, two symmetric vertices, so w.l.o.g. assume \(1b\in E\). Hence \(Y_{b}=\{3,d,e,f\}\) and \(G[Y_b]\) contains de, ef, 3d, 3f. Thus \(3e\not \in E\) by (i) and the last two undetermined neighbors of 3 are a and c. According to \(Y_{3}=\{0,1,b,e\}\) and \(0b,0e,1b\in E\), we have \(1e\in E\). Now \(N_G(1)=\{2,4,b,d,e,f\}\). So \(N_{G}(a)\cap Y=\{2,3,4\}\) as \(1a\notin E\). Therefore, \(Y_{a}=\{1,c,e,f\}\), yet \(d_{G[Y_{a}]}(e)=3\), contradicting to (i).
Secondly we let the minimum degree of G[Z] be 2. It is easy to know that the degree sequence of G[Z] is (4, 2, 2, 2, 2, 2) or (3, 3, 2, 2, 2, 2).
If G[Z] has degree sequence (4, 2, 2, 2, 2, 2), we suppose \(E(G[Z])=\{ab,ac,ad,ae,bc,df,ef\}\). Considering the symmetry of vertices in G[Y], w.l.o.g. suppose that \(N_{G}(f)\cap Y=\{1,2,3\}\). Then \(Y_{f}=\{4,a,b,c\}\) and \(ab,ac,bc\in E\), so \(G[Y_{f}]\) cannot be a 4-cycle, a contradiction.
If G[Z] has degree sequence (3, 3, 2, 2, 2, 2), then it is easy to show that there are four non-isomorphic G[Z]. The edge set of G[Z] is: (a) \(\{ac,ad,af,bc,bd,be,ef\}\), (b) \(\{ab,ac,ad,bc,be,df,ef\}\), (c) \(\{ab,ac,ad,be,bf,cd,ef\}\), or (d) \(\{ab,ac,ad,be,bf,cf,de\}\).
For (a), assume that \(N_{G}(c)\cap Y=\{1,2,3\}\). Then \(Y_{c}=\{4,d,e,f\}\). Note that \(de,df\notin E\). Thus \(d_{G[Y_c]}(d)\le 1\), a contradiction. For (b), assume \(N_{G}(f)\cap Y=\{1,2,3\}\), then \(Y_{f}=\{4,a,b,c\}\). Because \(ab,ac,bc\in E\), \(G[Y_{f}]\) cannot be a 4-cycle. Similarly to (b), we have that (c) is impossible. For (d), since \(d_{G[Z]}(a)=3\), a has exactly two neighbors and two non-neighbors in Y. Then we may suppose (d1) \(1a,2a\notin E\) or (d2) \(1a,3a\notin E\). So \(Y_1=\{0,3,a,z\}\) where \(z\in Z\). Since \(G[Y_1]\) is a 4-cycle and has edges 0a, 0z, we must have \(3a,3z\in E\), meaning that (d2) does not hold. Hence we let \(1a,2a\notin E\). But then \(Y_a=\{1,2,e,f\}\) and \(G[Y_a]\) cannot be a 4-cycle as \(12\in E\) and \(ef\notin E\), a contradiction.
Rights and permissions
About this article
Cite this article
Zhu, B., Zhou, J. & Chang, Y. 2-(v, 5; m) spontaneous emission error designs. Des. Codes Cryptogr. 88, 951–970 (2020). https://doi.org/10.1007/s10623-020-00722-1
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10623-020-00722-1