1 Introduction

Throughout this paper, we are only concerned with finite and simple graphs. For a plane graph G, let \(V(G), E(G), F(G), \Delta (G)\) and \(\delta (G)\) be the vertex set, edge set, face set, maximum degree and minimum degree of G, respectively. For an arbitrary \(x\in V(G)\cup F(G)\), let \(d_G(x)\) denote the degree of x in G. Let \(N_G(v)\) denote the set of neighbors of a vertex v in G. A vertex v satisfying \(d_G(v)=k\) (\(d_G(v)\ge k\), \(d_G(v)\le k\)) is a k-vertex (\(k^+\)-vertex, \(k^-\)-vertex). The k-face and \(k^+\)-face are defined similarly. For each \(v\in V(G)\), let \(d_G^k(v)\) denote the number of k-vertices adjacent to v in G. We call a 3-vertex \(v\in V(G)\) bad if \(d_G^3(v)=1\) and good if \(d_G^3(v)=0\). Let \(d_G^{3b}(v)\) and \(d_G^{3g}(v)\) denote the number of bad and good 3-vertices adjacent to v in G, respectively. A 3-face (or cycle) \(v_1v_2v_3\) is called a \((k_1,k_2,k_3)\)-face (or cycle) if \(v_i\) is a \(k_i\)-vertex for all \(1\le i\le 3\). A 3-cycle is bad if it is incident with two 3-vertices. Any undefined notation can refer to (Bondy and Murty 1976).

A proper k-edge coloring of a graph G is a mapping \(\varphi :E(G)\rightarrow \{1,2,\ldots ,k\}\) such that \(\varphi (e)\ne \varphi (e')\) for any two adjacent edges e and \(e'\) of G. For any \(v\in V(G)\), let \(C_\varphi (v)=\{\varphi (uv)| uv \in E(G)\}\) be the color set of v with respect to \(\varphi \). For two adjacent vertices u and v, we call u conflict with v respect to \(\varphi \) if \(C_\varphi (u)=C_\varphi (v)\). A proper k-edge coloring \(\varphi \) is a k-adjacent vertex distinguishing edge coloring (k-avd-coloring for short) provided that \(C_\varphi (u)\ne C_\varphi (v)\) for all \(uv\in E(G)\). The adjacent vertex distinguishing edge chromatic index of G, denoted by \(\chi '_a(G)\), is the smallest k such that G has a k-avd-coloring. A graph without isolated edges is normal. Clearly, only normal graph can have avd-colorings. Thus, for avd-coloring, we only consider normal graphs.

Zhang et al. (2002) first introduced the concept of avd-coloring and put forward the following conjecture.

Conjecture 1

Zhang et al. (2002) If G is a connected graph with order at least 6, then \(\chi '_{{a}}(G)\le \Delta (G)+2\).

Conjecture 1 was determined by Balister et al. (2007) for bipartite graphs and graphs with maximum degree 3. Horňák et al. (2014) showed that Conjecture 1 holds for planar graphs with maximum degree at least 12. Bonamy et al. (2013) verified that \(\chi '_{a}(G)\le \Delta (G)+1\) for any planar graph G with \(\Delta (G)\ge 12\). Wang and Huang (2015) proved that \(\chi '_{a}(G)\le \Delta (G)+1\) for any planar graph G with \(\Delta (G)\ge 16\) and \(\chi '_{a}(G)=\Delta (G)+1\) if and only if G contains two adjacent vertices of maximum degree.

A graph is 1-planar if it can be drawn in the plane such that each edge is crossed by at most one other edge. Albertson (2008) first introduced the definition of IC-planar graph. A graph is IC-planar if it admits a drawing in the plane where each edge is crossed at most once and no two crossings are incident with the same vertex. Clearly, each IC-planar graph is 1-planar. The associated plane graph \(G^\times \) of a 1-planar graph G is a plane graph obtained by turning all crossings of G into new 4-vertices. A vertex \(v\in V(G^\times )\) is false if v is not a vertex of G and real otherwise. A face is false if it is incident with at least one false vertex. Clearly, for an associated plane graph \(G^\times \) of an IC-planar graph G, each real vertex in \(G^\times \) is adjacent to at most one false vertex and incident with at most two false 3-faces in \(G^\times \). In the following, we always assume that every IC-planar graph is drawn in a plane such that the number of crossings is as few as possible.

Lemma 1

Zhang and Wu (2011) Let G be a 1-plane graph and \(G^\times \) be the associated plane graph of G. If \(d_G(u)=3\) and v is a false vertex of \(G^\times \), then either \(uv\notin E(G^\times )\) or uv is not incident with two 3-faces.

In this paper, we will prove that Conjecture 1 is true for any IC-planar graph with maximum degree at least 16, which can be expressed more concisely as follows:

Theorem 1

Let G be an IC-planar graph, then \(\chi '_{a}(G)\le \max \{\Delta (G)+2,18\}\).

2 The proof of Theorem 1

We will prove Theorem 1 by contradiction. Let G be a counterexample to Theorem 1 minimizing \(|V(G)|+|E(G)|\). Clearly, G is a connected graph. Let \(t_G=\max \{\Delta (G)+2,18\}\) and \(C=\{1,2,\ldots ,t_G\}\). Then \(\{1,2,\ldots ,18\}\subseteq C\). First we will prove the following claims.

Claim 1

There is no edge \(uv\in E(G)\) with \(d_G(u)=1\) and \(d_G(v)\le 9\).

Proof

Assume, to the contrary, that G contains an edge uv with \(d_G(u)=1\) and \(d_G(v)\le 9\). We have \({d_G}(v)\ge 2\) because G is normal. Let \(H=G-u\). If H contains only one edge, then we color this edge with 1 and color uv with 2 to obtain a \(t_G\)-avd-coloring of G, a contradiction. If H contains at least two edges, H has a \(t_G\)-avd-coloring \(\varphi \) with the color set C by the minimality of G. Note that v has at most eight conflict vertices. Hence we can color uv with a color in \(C\setminus {C_\varphi }(v)\) such that v does not conflict with its neighbors, which yields a \(t_G\)-avd-coloring of G, a contradiction. \(\square \)

Remark 1

Claim 1 implies that for an arbitrary \(e\in E(G)\), \(H=G-e\) is normal. Therefore \(\chi '_a(H)\le t_G\) by the minimality of G.

Remark 2

In the following, if \(d_G(v)=k\), set \(N_G(v) := \{v_1,v_2,\ldots ,v_k\}\).

Claim 2

Let v be a k-vertex of G with \(2\le k\le 6\), then \({d_G^{k}}(v)\le 1\).

Proof

Assume, to the contrary, that G contains a k-vertex v (\(2\le k\le 6\)) satisfying \({d_G^{k}}(v)\ge 2\). We prove the case that \(k=6\) (the proof can be given similarly and simply for \(2\le k\le 5\)). Assume that \(d_G(v_1)=d_G(v_2)=6\). Let \(N_G(v_1)=\{v,w_1,w_2,w_3,w_4,w_5\}\). Let \(H=G-vv_1\), by Remark 1, H has a \(t_G\)-avd-coloring \(\varphi \) with the color set C. Without loss of generality (W.l.o.g.), \(\varphi (vv_i)=i-1\) for \(2\le i\le 6\) and \(\varphi (v_1w_i)=a_i\) for \(1\le i\le 5\). We consider the next three cases.

Case 1: \(3\le |\{a_1,a_2,\ldots ,a_5\}\cap \{1,2,\ldots ,5\}|\le 5\). If \(|\{a_1,a_2,\ldots ,a_5\}\cap \{1,2,\ldots ,5\}|=5\), then we recolor \(vv_2\) with a color in \(C\setminus ({C_\varphi }(v)\cup {C_\varphi }(v_2))\) such that \(v_2\) does not conflict with its neighbors. So we may assume that \(3\le |\{a_1,a_2,\ldots ,a_5\}\cap \{1,2,\ldots ,5\}|\le 4\). Hence we can color \(vv_1\) with a color in \(C\setminus ({C_\varphi }(v)\cup {C_\varphi }(v_1))\) such that v and \(v_1\) do not conflict with their neighbors, which yields a \(t_G\)-avd-coloring of G, a contradiction.

Case 2: \(1\le |\{a_1,a_2,\ldots ,a_5\}\cap \{1,2,\ldots ,5\}|\le 2\). Set \(|\{a_1,a_2,\ldots ,a_5\}\cap \{1,2,\ldots ,5\}|=l\), then \(1\le l\le 2\). W.l.o.g., \(a_i=i\) for \(1\le i\le l\) and \(a_i=i-l+5\) for \(l+1\le i\le 5\). Suppose that \(vv_1\) cannot be colored without causing conflicts, say, \(C_\varphi (v_i) = \{1,2,3,4,5,i-l+9\}\) for \(2\le i\le 6\) and \(C_\varphi (w_i) = \{1,6,7,8,16-7l,i-l+15\}\) for \(1\le i\le l+3\). We recolor \(vv_2\) with a color in \(\{13,14,\ldots ,18\}\) such that \(v_2\) does not conflict with its neighbors, then we color \(vv_1\) with a color in \(\{11,12\}\) such that \(v_1\) does not conflict with its neighbors, which yields a \(t_G\)-avd-coloring of G, a contradiction.

Case 3: \(|\{a_1,a_2,\ldots ,a_5\}\cap \{1,2,\ldots ,5\}|=0\). W.l.o.g., \(a_i=i+5\) for \(1\le i\le 5\). Suppose that \(vv_1\) cannot be colored without causing conflicts, say, \(C_\varphi (v_i)=\{1,2,3,4,5,i+9\}\) for \(2\le i\le 6\) and \(C_\varphi (w_i)=\{6,7,8,9,10,i+15\}\) for \(1\le i\le 3\), or \(C_\varphi (v_i) = \{1,2,3,4,5,i+9\}\) for \(2\le i\le 5\) and \(C_\varphi (w_i)=\{6,7,8,9,10,i+14\}\) for \(1\le i\le 4\). If \(C_\varphi (v_i)=\{1,2,3,4,5,i+9\}\) for \(2\le i\le 6\) and \(C_\varphi (w_i)=\{6,7,8,9,10,i+15\}\) for \(1\le i\le 3\), then we recolor \(vv_2\) with a color in \(\{6,7,8,16,17,18\}\) such that \(v_2\) does not conflict with its neighbors, and color \(vv_1\) with a color in \(\{12,13,14\}\) such that \(v_1\) does not conflict with \(w_4\) and \(w_5\), which yields a \(t_G\)-avd-coloring of G, a contradiction. If \(C_\varphi (v_i)=\{1,2,3,4,5,i+9\}\) for \(2\le i\le 5\) and \(C_\varphi (w_i)=\{6,7,8,9,10,i+14\}\) for \(1\le i\le 4\), then we recolor \(vv_2\) with a color in \(\{6,7,8,9,10,18\}\) such that \(v_2\) does not conflict with its neighbors, and color \(vv_1\) with a color in \(\{12,13,14\}\) such that v and \(v_1\) do not conflict with their neighbors, which yields a \(t_G\)-avd-coloring of G, a contradiction. \(\square \)

Claim 3

There is no edge \(vv_1\in E(G)\) with \(2\le d_G(v_1)\le 6\) and \(d_G(v_1)+1\le d_G(v)\le 9\).

Proof

Assume, to the contrary, that G contains an edge \(vv_1\) with \(2\le d_G(v_1)\le 6\) and \(d_G(v_1)+1\le d_G(v)\le 9\). We prove the case that \(d_G(v_1)=6\) and \(d_G(v)=9\) (the proof can be given similarly and simply for other cases). Let \(H=G-vv_1\), by Remark 1, H has a \(t_G\)-avd-coloring \(\varphi \) with the color set C. W.l.o.g., \(\varphi (vv_i)=i-1\) for \(2\le i\le 9\) and \(C_\varphi (v_1)\subseteq \{1,2,\ldots ,13\}\). By Claim 2, every 6-vertex has at most one conflict vertex. Suppose that \(vv_1\) cannot be colored without causing conflicts, say, \(C_\varphi (v_i)=\{1,2,\ldots ,8,i+12\}\) for \(2\le i\le 5\) and \(C_\varphi (v_1)=\{9,10,...,13\}\). Without considering the conflict of v, for any given integer i (\(2\le i\le 5\)), we select \(\{b_i,d_i\}\) from \(\{9,10,\ldots ,18\}\setminus \{i+12\}\) to recolor \(vv_i\) and color \(vv_1\) such that \(v_i\) and \(v_1\) do not conflict with their neighbors. \(\{b_i,d_i\}\) has at least two selected ways. Since i has four possibilities, we have at least \(2\times 4=8\) ways such that \(v_1\) does not conflict with its neighbors and v does not conflict with \(v_2,v_3,v_4\) and \(v_5\), while v has at most four conflict vertices other than \(v_2,v_3,v_4\) and \(v_5\). So we can obtain a \(t_G\)-avd-coloring of G, a contradiction. \(\square \)

Claim 4

Let v be a k-vertex of G with \(10\le k\le 11\), then \({d_G^{(16-k)^-}}(v)\le 1\).

Proof

Assume, to the contrary, that G contains a k-vertex v (\(10\le k\le 11\)) satisfying \({d_G^{(16-k)^-}}(v)\ge 2\). Suppose that \(d_G(v_1)=d_G(v_2)=16-k\) (the proof can be given similarly and simply for other cases). Let \(H=G-vv_1\), by Remark 1, H has a \(t_G\)-avd-coloring \(\varphi \) with the color set C. W.l.o.g., \(\varphi (vv_i)=i-1\) for \(2\le i\le k\). Clearly, \(|C_\varphi (v_i)\cap \{k,k+1,\ldots ,18\}|\le 15-k\) for \(1\le i\le 2\). By Claim 2, every \(6^-\)-vertex has at most one conflict vertex. If \(v_i\) has a conflict vertex \(w_i\), and \(|C_\varphi (v_i)\cap \{k,k+1,\ldots ,18\}|=15-k\) for \(1\le i\le 2\), then we recolor \(v_iw_i\) with a color in \(\{2,3,\ldots ,9\}\setminus C_\varphi (w_i)\). Without considering the conflict of v, we have the following two types of proper colorings. (a): We color \(vv_1\) with a color in \(\{k,k+1,\ldots ,18\}\) such that \(v_1\) does not conflict with its neighbors. There are at least four available colors. (b): We select \(\{b_1,b_2\}\) from \(\{k,k+1,\ldots ,18\}\) to recolor \(vv_2\) and color \(vv_1\) such that \(v_2\) and \(v_1\) do not conflict with their neighbors. \(\{b_1,b_2\}\) has at least \(\frac{4\times 3}{2}=6\) selected ways. Hence we have at least \(4+6=10\) ways, while v has at most \(k-2\le 9\) conflict vertices. So we can obtain a \(t_G\)-avd-coloring of G, a contradiction. \(\square \)

Claim 5

Let v be a 12-vertex of G, then \({d_G^{3^-}}(v)\le 1\).

Proof

Assume, to the contrary, that G contains a 12-vertex v satisfying \({d_G^{3^-}}(v)\ge 2\). Suppose that \(d_G(v_1)=d_G(v_2)=3\) (the proof can be given similarly and simply for other cases). Let \(H=G-vv_1\), by Remark 1, H has a \(t_G\)-avd-coloring \(\varphi \) with the color set C. W.l.o.g., \(\varphi (vv_i)=i-1\) for \(2\le i\le 12\). Clearly, \(|C_\varphi (v_i)\cap \{12,13,\ldots ,18\}|\le 2\) for \(1\le i\le 2\). By Claim 2, each 3-vertex has at most one conflict vertex. If \(v_i\) has a conflict vertex \(w_i\) for \(1\le i\le 2\), we assume that \(\varphi (v_iw_i)\notin \{12,13,\ldots ,18\}\) (if \(\varphi (v_iw_i)\in \{12,13,\ldots ,18\}\), then we recolor \(v_iw_i\) with a color in \(\{2,3,\ldots ,11\}\setminus (C_\varphi (v_i)\cup C_\varphi (w_i))\) to satisfy this condition). Without considering the conflict of v, we have the following two types of proper colorings. (a): We color \(vv_1\) with a color in \(\{12,13,\ldots ,18\}\) such that \(v_1\) does not conflict with its neighbors. There are at least five available colors. (b): We select \(\{b_1,b_2\}\) from \(\{12,13,\ldots ,18\}\) to recolor \(vv_2\) and color \(vv_1\) such that \(v_2\) and \(v_1\) do not conflict with their neighbors. \(\{b_1,b_2\}\) has at least \(\frac{5\times 4}{2}=10\) selected ways. Hence we have at least \(5+10=15\) ways, while v has at most ten conflict vertices. So we can obtain a \(t_G\)-avd-coloring of G, a contradiction. \(\square \)

Claim 6

Let v be a k-vertex of G with \(11\le k\le 12\), then \({d_G^{6^-}}(v)\le 3k-31\).

Proof

Assume, to the contrary, that G contains a k-vertex v (\(11\le k\le 12\)) satisfying \({d_G^{6^-}}(v)\ge 3k-30\). Suppose that \(d_G(v_i)=6\) for \(1\le i\le 3k-30\) (the proof can be given similarly and simply for other cases). Let \(H=G-vv_1\), by Remark 1, H has a \(t_G\)-avd-coloring \(\varphi \) with the color set C. W.l.o.g., \(\varphi (vv_i)=i-1\) for \(2\le i\le k\). Clearly, \(|C_\varphi (v_i)\cap \{k,k+1,\ldots ,18\}|\le 5\) for \(1\le i\le 3k-30\). By Claim 2, each 6-vertex has at most one conflict vertex. If \(v_i\) has a conflict vertex \(w_i\), and \(|C_\varphi (v_i)\cap \{k,k+1,\ldots ,18\}|=5\) for \(1\le i\le 3k-30\), then we recolor \(v_iw_i\) with a color in \(\{3k-30,3k-29,\ldots ,k-1\}\setminus C_\varphi (w_i)\). Without considering the conflict of v, we have the following two types of proper colorings. (a): We color \(vv_1\) with a color in \(\{k,k+1,\ldots ,18\}\) such that \(v_1\) does not conflict with its neighbors. There are at least \(14-k\) available colors. (b): For any given integer i (\(2\le i\le 3k-30\)), we select \(\{b_i,d_i\}\) from \(\{k,k+1,\ldots ,18\}\) to recolor \(vv_i\) and color \(vv_1\) such that \(v_i\) and \(v_1\) do not conflict with their neighbors. \(\{b_i,d_i\}\) has at least \(\frac{(14-k)\times (13-k)}{2}\) selected ways. Since i has \(3k-31\) possibilities, we have at least \(\frac{(14-k)\times (13-k)}{2}\times (3k-31)=17-k\) different coloring ways. Hence we have at least \(14-k+17-k=31-2k\) ways, while v has at most \(k-(3k-30)=30-2k\) conflict vertices. So we can obtain a \(t_G\)-avd-coloring of G, a contradiction. \(\square \)

Claim 7

Let v be a k-vertex of G with \(13\le k\le 14\), then the following statements hold.

(1) \({d_G^{2^-}}(v)\le k-12\);

(2) If \({d_G^{m^-}}(v)\ge 1\) for \( m\le 18-k\), then \({d_G^{k}}(v)\ge (19-k-m){d_G^{{(19-k)}^-}}(v)+1\).

Proof

(1) Assume, to the contrary, that G contains a k-vertex v (\(13\le k\le 14\)) satisfying \({d_G^{2^-}}(v)\ge k-11\). Suppose that \(d_G(v_i)=2\) for \(1\le i\le k-11\) (the proof can be given similarly and simply for other cases). Let \(H=G-vv_1\), by Remark 1, H has a \(t_G\)-avd-coloring \(\varphi \) with the color set C. W.l.o.g., \(\varphi (vv_i)=i-1\) for \(2\le i\le k\). Clearly, \(|C_\varphi (v_i)\cap \{k,k+1,\ldots ,18\}|\le 1\) for \(1\le i\le k-11\). By Claim 2, each 2-vertex has at most one conflict vertex. If \(v_i\) has a conflict vertex \(w_i\) for \(1\le i\le k-11\), we assume that \(\varphi (v_iw_i)\notin \{k,k+1,\ldots ,18\}\) (if \(\varphi (v_iw_i)\in \{k,k+1,\ldots ,18\}\), then we recolor \(v_iw_i\) with a color in \(\{3,4,\ldots ,12\}\setminus (C_\varphi (v_i)\cup C_\varphi (w_i))\) to satisfy this condition). Without considering the conflict of v, we have the following two types of proper colorings. (a): We color \(vv_1\) with a color in \(\{k,k+1,\ldots ,18\}\) such that \(v_1\) does not conflict with its neighbors. There are at least \(18-k\ge 4\) available colors. (b): For any given integer i (\(2\le i\le k-11\)), we select \(\{b_i,d_i\}\) from \(\{k,k+1,\ldots ,18\}\) to recolor \(vv_i\) and color \(vv_1\) such that \(v_i\) and \(v_1\) do not conflict with their neighbors. \(\{b_i,d_i\}\) has at least \(\frac{(18-k)(17-k)}{2}\) selected ways. Since i has \(k-12\) possibilities, we have at least \(\frac{(18-k)(17-k)}{2}\times (k-12)\ge 10\) different coloring ways. Hence we have at least \(4+10=14\) ways, while v has at most eleven conflict vertices. So we can obtain a \(t_G\)-avd-coloring of G, a contradiction.

(2) Assume, to the contrary, that there is a k-vertex \(v\in V(G)\) (\(13\le k\le 14\)) and an integer m (\( m\le 18-k\)) satisfying \({d_G^{m^-}}(v)\ge 1\), where \({d_G^{k}}(v)\le (19-k-m){d_G^{{(19-k)}^-}}(v)\). Set \({d_G^{{(19-k)}^-}}(v)=l\). W.l.o.g., \({d_G}(v_1)=m\) and \({d_G}(v_i)\le 19-k\) for \(1\le i\le l\) (the proof can be given similarly and simply for other cases). Let \(H=G-vv_1\), by Remark 1, H has a \(t_G\)-avd-coloring \(\varphi \) with the color set C. Suppose that \(\varphi (vv_i)=i-1\) for \(2\le i\le k\). Clearly, \(|C_\varphi (v_i)\cap \{k,k+1,\ldots ,18\}|\le 18-k\) for \(1\le i\le l\). By Claim 2, each \(6^-\)-vertex has at most one conflict vertex. If \(v_i\) has a conflict vertex \(w_i\), and \(|C_\varphi (v_i)\cap \{k,k+1,\ldots ,18\}|= d_G(v_i)-1\) for \(1\le i\le l\), then we recolor \(v_iw_i\) with a color in \(\{7,8,\ldots ,12\}\setminus C_\varphi (w_i)\). Without considering the conflict of v, we have the following two types of proper colorings. (a): We color \(vv_1\) with a color in \(\{k,k+1,\ldots ,18\}\) such that \(v_1\) does not conflict with its neighbors. There are at least \(20-k-m\) available colors. (b): For any given integer i (\(2\le i\le l\)), we select \(\{b_i,d_i\}\) from \(\{k,k+1,\ldots ,18\}\) to recolor \(vv_i\) and color \(vv_1\) such that \(v_i\) and \(v_1\) do not conflict with their neighbors. \(\{b_i,d_i\}\) has at least \(19-k-m\) selected ways. Since i has \(l-1\) possibilities, we have at least \((19-k-m)(l-1)\) different coloring ways. Hence we have at least \((20-k-m)+(19-k-m)(l-1)=(19-k-m)l+1\) ways, while v has at most \((19-k-m)l\) conflict vertices. So we can obtain a \(t_G\)-avd-coloring of G, a contradiction. \(\square \)

Claim 8

Let v be a 15-vertex of G, then the following statements hold.

(1) \({d_G^{2^-}}(v)\le 3\);

(2) If \({d_G^{2^-}}(v)\ge 1\), then \({d_G^{3^-}}(v)\le 4\);

(3) If \({d_G^{m^-}}(v)\ge 1\) for \(m\le 3\), then \({d_G^{15}}(v)\ge (4-m){d_G^{4^-}}(v)+1\);

(4) If v is incident with a bad 3-cycle, then \({d_G^{15}}(v)\ge 9\).

Proof

(1) Assume, to the contrary, that G contains a 15-vertex v satisfying \({d_G^{2^-}}(v)\ge 4\). Suppose that \(d_G(v_i)=2\) for \(1\le i\le 4\) (the proof can be given similarly and simply for other cases). Let \(H=G-vv_1\), by Remark 1, H has a \(t_G\)-avd-coloring \(\varphi \) with the color set C. Suppose that \(\varphi (vv_i)=i-1\) for \(2\le i\le 15\). Clearly, \(|C_\varphi (v_i)\cap \{15,16,17,18\}|\le 1\) for \(1\le i\le 4\). By Claim 2, each 2-vertex has at most one conflict vertex. If \(v_i\) has a conflict vertex \(w_i\) for \(1\le i\le 4\), we assume that \(\varphi (v_iw_i)\notin \{15,16,17,18\}\) (if \(\varphi (v_iw_i)\in \{15,16,17,18\}\), then we recolor \(v_iw_i\) with a color in \(\{4,5,\ldots ,14\}\setminus (C_\varphi (v_i)\cup C_\varphi (w_i))\) to satisfy this condition). Without considering the conflict of v, we have the following two types of proper colorings. (a): We color \(vv_1\) with a color in \(\{15,16,17,18\}\) such that \(v_1\) does not conflict with its neighbors. There are at least three available colors. (b): For any given integer i (\(2\le i\le 4\)), we select \(\{b_i,d_i\}\) from \(\{15,16,17,18\}\) to recolor \(vv_i\) and color \(vv_1\) such that \(v_i\) and \(v_1\) do not conflict with their neighbors. \(\{b_i,d_i\}\) has at least three selected ways. Since i has three possibilities, we have at least \(3\times 3=9\) different coloring ways. Hence we have at least \(3+9=12\) ways, while v has at most eleven conflict vertices. So we can obtain a \(t_G\)-avd-coloring of G, a contradiction.

(2) Assume, to the contrary, that G contains a 15-vertex v satisfying \({d_G^{2^-}}(v)\ge 1\), where \({d_G^{3^-}}(v)\ge 5\). Suppose that \({d_G}(v_1)=2\) and \({d_G}(v_i)=3\) for \(2\le i\le 5\) (the proof can be given similarly and simply for other cases). Let \(H=G-vv_1\), by Remark 1, H has a \(t_G\)-avd-coloring \(\varphi \) with the color set C. W.l.o.g., \(\varphi (vv_i)=i-1\) for \(2\le i\le 15\). Clearly, \(|C_\varphi (v_i)\cap \{15,16,17,18\}|\le 2\) for \(1\le i\le 5\). By Claim 2, each \(3^-\)-vertex has at most one conflict vertex. If \(v_i\) has a conflict vertex \(w_i\) for \(1\le i\le 5\), we assume that \(\varphi (v_iw_i)\notin \{15,16,17,18\}\) (if \(\varphi (v_iw_i)\in \{15,16,17,18\}\), then we recolor \(v_iw_i\) with a color in \(\{8,9,\ldots ,14\}\setminus (C_\varphi (v_i)\cup C_\varphi (w_i))\) to satisfy this condition). Without considering the conflict of v, we have the following two types of proper colorings. (a): We color \(vv_1\) with a color in \(\{15,16,17,18\}\) such that \(v_1\) does not conflict with its neighbors. There are at least three available colors. (b): For any given integer i (\(2\le i\le 5\)), we select \(\{b_i,d_i\}\) from \(\{15,16,17,18\}\) to recolor \(vv_i\) and color \(vv_1\) such that \(v_i\) and \(v_1\) do not conflict with their neighbors. \(\{b_i,d_i\}\) has at least two selected ways. Since i has four possibilities, we have at least \(2\times 4=8\) different coloring ways. Hence we have at least \(3+8=11\) ways, while v has at most ten conflict vertices. So we can obtain a \(t_G\)-avd-coloring of G, a contradiction.

(3) Assume, to the contrary, that there is a 15-vertex \(v\in V(G)\) and an integer m (\(m\le 3\)) satisfying \({d_G^{m^-}}(v)\ge 1\), where \({d_G^{15}}(v)\le (4-m){d_G^{{4}^-}}(v)\). Set \({d_G^{{4}^-}}(v)=l\). Suppose that \({d_G}(v_1)=m\) and \({d_G}(v_i)\le 4\) for \(1\le i\le l\) (the proof can be given similarly and simply for other cases). Let \(H=G-vv_1\), by Remark 1, H has a \(t_G\)-avd-coloring \(\varphi \) with the color set C. Suppose that \(\varphi (vv_i)=i-1\) for \(2\le i\le 15\). Clearly, \(|C_\varphi (v_i)\cap \{15,16,17,18\}|\le 3\) for \(1\le i\le l\). By Claim 2, each \(4^-\)-vertex has at most one conflict vertex. If \(v_i\) has a conflict vertex \(w_i\) for \(1\le i\le l\), we assume that \(\varphi (v_iw_i)\notin \{15,16,17,18\}\) (if \(\varphi (v_iw_i)\in \{15,16,17,18\}\), then we recolor \(v_iw_i\) with a color in \(\{8,9,\ldots ,14\}\setminus (C_\varphi (v_i)\cup C_\varphi (w_i))\) to satisfy this condition). Without considering the conflict of v, we have the following two types of proper colorings. (a): We color \(vv_1\) with a color in \(\{15,16,17,18\}\) such that \(v_1\) does not conflict with its neighbors. There are at least \(5-m\) available colors. (b): For any given integer i (\(2\le i\le l\)), we select \(\{b_i,d_i\}\) from \(\{15,16,17,18\}\) to recolor \(vv_i\) and color \(vv_1\) such that \(v_i\) and \(v_1\) do not conflict with their neighbors. \(\{b_i,d_i\}\) has at least \(4-m\) selected ways. Since i has \(l-1\) possibilities, we have at least \((4-m)(l-1)\) different coloring ways. Hence we have at least \((5-m)+(4-m)(l-1)=(4-m)l+1\) ways, while v has at most \((4-m)l\) conflict vertices. So we can obtain a \(t_G\)-avd-coloring of G, a contradiction.

(4) Assume, to the contrary, that there exists a 15-vertex \(v\in V(G)\) incident with a bad 3-cycle \(vv_1v_2\) (\(d_G(v_1)=d_G(v_2)=3\)), where \({d_G^{15}}(v)\le 8\). Let \(w_i\) (\(1\le i\le 2\)) be the neighbor of \(v_i\) other than \(v, v_{3-i}\). Let \(H=G-v_1v_2\), by Remark 1, H has a \(t_G\)-avd-coloring \(\varphi \) with the color set C. By Claim 2, \(v_i\) (\(1\le i\le 2\)) has exactly one conflict vertex. If \(C_\varphi (v_1)\ne C_\varphi (v_2)\), then we color \(v_1v_2\) with a color in \(C\setminus (C_\varphi (v_1)\cup C_\varphi (v_2))\) to get a \(t_G\)-avd-coloring of G, a contradiction. If \(C_\varphi (v_1)=C_\varphi (v_2)\), w.l.o.g., \(\varphi (vv_1)=\varphi (v_2w_2)=1,\varphi (vv_2)=\varphi (v_1w_1)=2\) and \(\varphi (vv_i)=i\) for \(3\le i\le 15\). Without considering the conflict of v, we have the following two types of proper colorings. (a): For any given integer i (\(1\le i\le 2\)), we recolor \(vv_i\) with an arbitrary color in \(\{16,17,18\}\) and color \(v_1v_2\) with 3. Since i has two possibilities, we have \(3\times 2=6\) different coloring ways. (b): We select \(\{b_1,b_2\}\) from \(\{16,17,18\}\) to recolor \(vv_1\) and \(vv_2\), and color \(v_1v_2\) with 3. \(\{b_1,b_2\}\) has three selected ways. Hence we have \(6+3=9\) ways, while v has at most eight conflict vertices. So we can obtain a \(t_G\)-avd-coloring of G, a contradiction. \(\square \)

Claim 9

Let v be a k-vertex of G with \(k\ge 14\), then v is incident with at most one bad 3-cycle.

Proof

Assume, to the contrary, that there exists a k-vertex \(v\in V(G)\) (\(k\ge 14\)) incident with two bad 3-cycles \(vv_1v_2,vv_3v_4\), where \(d_G(v_i)=3\) for \(1\le i\le 4\). Let \(w_i\) be the neighbor of \(v_i\) for \(1\le i\le 4\). Let \(H=G-v_1v_2\), by Remark 1, H has a \(t_G\)-avd-coloring \(\varphi \) with the color set C. By Claim 2, each 3-vertex has at most one conflict vertex. If \(C_\varphi (v_1)\ne C_\varphi (v_2)\), then we color \(v_1v_2\) with an arbitrary color in \(C\setminus (C_\varphi (v_1)\cup C_\varphi (v_2))\) to yield a \(t_G\)-avd-coloring of G, a contradiction. If \(C_\varphi (v_1)=C_\varphi (v_2)\), w.l.o.g., \(\varphi (vv_1)=\varphi (v_2w_2)=1,\varphi (vv_2)=\varphi (v_1w_1)=2\) and \(\varphi (vv_i)=i\) for \(3\le i\le k\). Note that \(|\{\varphi (v_3w_3),\varphi (v_4w_4)\}\cap \{3,4\}|\le 1\), w.l.o.g., \(\varphi (v_4w_4)\ne 3\). Clearly, \(|\{\varphi (v_4w_4)\}\cap \{1,2\}|\le 1\), w.l.o.g., \(\varphi (v_4w_4)\ne 1\). We first delete the color of \(v_3v_4\), switch the colors of \(vv_1\) and \(vv_4\), then color \(v_1v_2,v_3v_4\) properly to yield a \(t_G\)-avd-coloring of G, a contradiction. \(\square \)

Claim 10

Let v be a k-vertex of G with \(k\ge 16\). If v is incident with a bad 3-cycle, then \({d_G^{k}}(v)\ge 2d_G^{4^-}(v)+1\).

Proof

Assume, to the contrary, that there exists a k-vertex \(v\in V(G)\) (\(k\ge 16\)) incident with a bad 3-cycle \(vv_1v_2\) (\(d_G(v_1)=d_G(v_2)=3\)), where \(d_G^{k}(v)\le 2d_G^{4^-}(v)\). Let \(w_i\) (\(1\le i\le 2\)) be the neighbor of \(v_i\) other than \(v,v_{3-i}\). Set \(d_G^{4^-}(v)=m\). Suppose that \(d_G(v_i)\le 4\) for \(1\le i\le m\). Let \(H=G-v_1v_2\), by Remark 1, H has a \(t_G\)-avd-coloring \(\varphi \) with the color set C. By Claim 2, each \(4^-\)-vertex has at most one conflict vertex. If \(C_\varphi (v_1)\ne C_\varphi (v_2)\), then we color \(v_1v_2\) with an arbitrary color in \(C\setminus (C_\varphi (v_1)\cup C_\varphi (v_2))\) to yield a \(t_G\)-avd-coloring of G, a contradiction. If \(C_\varphi (v_1)=C_\varphi (v_2)\), w.l.o.g., \(\varphi (vv_1)=\varphi (v_2w_2)=1,\varphi (vv_2)=\varphi (v_1w_1)=2\) and \(\varphi (vv_i)=i\) for \(3\le i\le k\). Clearly, \(|C_\varphi (v_i)\cap \{1,2,k+1,k+2\}|\le 3\) for \(1\le i\le m\). If \(v_i\) has a conflict vertex \(w_i\) for \(3\le i\le m\), we assume that \(\varphi (v_iw_i)\notin \{1,2,k+1,k+2\}\) (if \(\varphi (v_iw_i)\in \{1,2,k+1,k+2\}\), then we recolor \(v_iw_i\) with a color in \(\{k-6,k-5,\ldots ,k\}\setminus (C_\varphi (v_i)\cup C_\varphi (w_i))\) to satisfy this condition). Without considering the conflict of v, we have the following three types of proper colorings. (a): For any given integer i (\(1\le i\le 2\)), we recolor \(vv_i\) with an arbitrary color in \(\{k+1,k+2\}\) and color \(v_1v_2\) with 3. Since i has two possibilities, we have \(2\times 2=4\) different coloring ways. (b): We recolor \(vv_i\) with \(k+i\) for \(1\le i\le 2\) and color \(v_1v_2\) with 3. (c): For any given integer i (\(3\le i\le m\)), we recolor \(vv_i\) with \(b_i\) in \(\{1,2,k+1,k+2\}\) such that \(v_i\) does not conflict with its neighbors. If \(b_i\in \{1,2\}\), then we recolor \(vv_{b_i}\) with \(k+1\) or \(k+2\) and color \(v_1v_2\) with 3, so there are two coloring ways. If \(b_i\in \{k+1,k+2\}\), then we recolor \(vv_1\) or \(vv_2\) with a color in \(\{k+1,k+2\}\setminus \{b_i\}\) and color \(v_1v_2\) with 3, so there are two ways. Since i has \(m-2\) possibilities, we have \(2(m-2)\) ways. Hence we have \(4+1+2(m-2)=2m+1\) ways, while v has at most 2m conflict vertices. So we can obtain a \(t_G\)-avd-coloring of G, a contradiction. \(\square \)

Claim 11

Yan et al. (2012) Let v be a k-vertex of G with \(k\ge 16\). If \({d_G^{2^-}}(v)\ge 1\), then \({d_G^{3^-}}(v)\le \left\lceil {\frac{k}{2}}\right\rceil -1\) and \({d_G^{k}}(v)\ge {d_G^{3^-}}(v)+1\).

Let H be one of the connected component of the graph which is obtained from G by deleting all \(2^-\)-vertices. By Claims 135, 7811, the relation between \(d_G(v)\) and \(d_{H}(v)\) is as in Table 1.

Table 1 The relation between \(d_G(v)\) and \(d_{H}(v)\)

By Table 1, we deduce that \(\delta (H)\ge 3\), and for any \(v\in V(H)\), we have \(d_H^k(v)=d_G^k(v)\), where \(3\le k\le 6\). Let \(H^\times \) be the associated plane graph of H. By Claims 24, 11 and Table 1, every 3-face of \(H^\times \) is one of the following types:

Type I: (3, 3, 4)-faces, (4, 4, 4)-faces;

Type II: \((3,3,10^+)\)-faces, \((3,4,10^+)\)-faces, \((4,4,9^+)\)-faces, \((4,5,9^+)\)-faces;

Type III: \((3,10^+,10^+)\)-faces, (4, 5, 5)-faces, (4, 6, 6)-faces, \((4,6,9^+)\)-faces, \((4,7^+,7^+)\)-faces, \((5,5,9^+)\)-faces, \((5,9^+,9^+)\)-faces, \((6,6,9^+)\)-faces, \((6,9^+,9^+)\)-faces;

Type IV: \((7^+,7^+,7^+)\)-faces.

Let \(c_f\) be the false vertex incident with a false 3-face f, and \(N_{\bar{f}}(c_f)\) be the set of neighbors of \(c_f\) which are not incident with f. f is the corresponding face of the vertices in \(N_{\bar{f}}(c_f)\). By Claims 23, v has at most one corresponding 3-face of Type I. A vertex v is of Type I if it has a corresponding 3-face of Type I. Let \(n_i(v)\) be the number of 3-faces of Type i incident with v, \(i\in \{{\mathrm{II,III,IV}}\}\). Let \(n_{4^+}(v)\) be the number of \(4^+\)-faces incident with v in \(H^\times \).

By Euler’s formula \(|V(H^\times )|-|E(H^\times )|+|F(H^\times )|=2\), we have:

$$\begin{aligned} \sum \limits _{v\in V(H^\times )}{\left( {{d_{H^\times }}(v)-4}\right) }+\sum \limits _{f\in F({H^\times })}{\left( {{d_{H^\times }}(f)-4}\right) }=-8 \end{aligned}$$

Next, we will apply the discharging method to derive a contradiction. We define the initial charge function \(w(x)=d_{H^\times }(x)-4\) for \(x\in V({H^\times })\cup F({H^\times })\), and design discharging rules to redistribute charges. Let \(w'\) be the new charge after the discharging process, then we will show that \(w'(x)\ge 0\) for \(x\in V({H^\times })\cup F({H^\times })\), which leads to a contradiction.

The discharging rules are defined as follows. In the following rules, the degree of a vertex refers to its degree in H.

R1: Each 3-face f of Type I gets \({\frac{1}{2}}\) from every \(9^+\)-vertex in \(N_{\bar{f}}(c_f)\) (by Claims 23, f is false and \(N_{\bar{f}}(c_f)\) consists of two \(9^+\)-vertices);

R2: Each 3-face of Type II gets 1 from its incident \(9^+\)-vertex;

R3: Each of \((5,9^+,9^+)\)-faces and \((6,9^+,9^+)\)-faces gets \({\frac{1}{2}}\) from every incident \(9^+\)-vertex, and each other 3-face of Type III gets \({\frac{1}{2}}\) from every incident \(5^+\)-vertex;

R4: Each 3-face of Type IV gets \({\frac{1}{3}}\) from every incident \(7^+\)-vertex;

R5: Each good 3-vertex gets \({\frac{1}{3}}\) from every adjacent \(10^+\)-vertex in H, and each bad 3-vertex gets \({\frac{1}{2}}\) from every adjacent \(10^+\)-vertex in H.

We first verify the new charge of \(f\in F(H^\times )\).

\(\bullet \) \(d_{H^\times }(f)=3\). By R1–R4, \(w'(f)\ge 0\).

\(\bullet \) \(d_{H^\times }(f)\ge 4\). The charge remains unchanged, \(w'(f)=d_{H^\times }(f)-4\ge 0\).

Next, we verify the new charge of \(v\in V(H^\times )\). For each real vertex \(v\in V(H^\times )\), we have \(d_{H^\times }(v)=d_G(v)-d_G^{2^-}(v)\).

\(\bullet \) \(d_{H^\times }(v)=3\). By Claims 24 and Table 1, \(d_H^{9^-}(v)=d_H^3(v)\le 1\). If v is good, then \(d_H^{10^+}(v)=3\), otherwise \(d_H^{10^+}(v)=2\). By R5, \(w'(v)\ge 3-4+\min \{\frac{1}{3}\times 3,\frac{1}{2}\times 2\}=0\).

\(\bullet \) \(d_{H^\times }(v)=4\). No rule applies to v, then \(w'(v)=4-4=0\).

\(\bullet \) \(d_{H^\times }(v)=5\). By Claims 23 and Table 1, \(d_H^{8^-}(v)=d_H^5(v)\le 1\). By R3, only (4, 5, 5)-faces and \((5,5,9^+)\)-faces incident with v get charges from v. There are at most two such faces incident with v. By R3, \(w'(v)\ge 5-4-\frac{1}{2}\times 2=0\).

\(\bullet \) \(d_{H^\times }(v)=6\). By Claims 23 and Table 1, \(d_H^{8^-}(v)=d_H^6(v)\le 1\). By R3, only (4, 6, 6)-faces, \((4,6,9^+)\)-faces and \((6,6,9^+)\)-faces incident with v get charges from v. There are at most four such faces incident with v. By R3, \(w'(v)\ge 6-4-\frac{1}{2}\times 4=0\).

\(\bullet \) \(7\le d_{H^\times }(v)\le 8\). By Claim 3 and Table 1, \(d_H^{6^-}(v)=0\) and v is not of Type I. Thus we have \(n_{\mathrm{III}}(v)\le 2\). By R3–R4, \(w'(v)\ge d_{H^\times }(v)-4-\frac{1}{2}\times {2}-\frac{1}{3}\times {(d_{H^\times }(v)-2)}=\frac{2d_{H^\times }(v)-13}{3}>0\).

\(\bullet \) \(d_{H^\times }(v)=9\). We first give the following fact.

Fact 1

If \(d_{H^\times }(v)=9\), then \(d_H^{3}(v)=0\) and \(d_H^{6^-}(v)\le 1\).

Proof

By Table 1, we have \(d_G(v)\in \{9,10,16,17\}\). If \(d_G(v)=9\), by Claim 3, \(d_H^{6^-}(v)=0\). If \(d_G(v)=10\), then \(d_G^{2^-}(v)=1\). By Claim 4, \(d_H^{6^-}(v)=0\). If \(d_G(v)=k\) (\(16\le k\le 17\)), then \(d_G^{2^-}(v)=k-9\). By Claim 11, \(d_G^{3^-}(v)\le \lceil \frac{k}{2}\rceil -1=k-9\) and \(d_G^k(v)\ge d_G^{3^-}(v)+1\). Thus \(d_H^{3}(v)=0\) and \(d_H^{6^-}(v)\le k-d_G^{2^-}(v)-d_G^{k}(v)\le k-(k-9)-(k-8)\le 1\). \(\square \)

By Fact 1, if v is of Type I, then \(n_{\mathrm{II}}(v)=0\), otherwise \(n_{\mathrm{II}}(v)\le 1\). By R1–R4, \(w'(v)\ge 9-4-\max \{\frac{1}{2}+\frac{1}{2}\times 9,1+\frac{1}{2}\times {8}\}=0\).

\(\bullet \) \(d_{H^\times }(v)=10\). We first give the following fact.

Fact 2

If \(d_{H^\times }(v)=10\), then \(d_H^{3}(v)\le 1\) and \(d_H^{6^-}(v)\le 3\).

Proof

By Table 1, we have \(d_G(v)\in \{10,11\}\) or \(d_G(v)\ge 16\). If \(d_G(v)=10\), by Claim 4, \(d_H^{6^-}(v)\le 1\). If \(d_G(v)=11\), then \(d_G^{2^-}(v)=1\). By Claims 4 and 6, \(d_G^{5^-}(v)\le 1\) and \(d_G^{6^-}(v)\le 2\). Thus \(d_H^{3^-}(v)=0\) and \(d_H^{6^-}(v)\le 1\). If \(d_G(v)=k\) (\(k\ge 16\)), then \(d_G^{2^-}(v)=k-10\). By Claim 11, \(d_G^{3^-}(v)\le \lceil \frac{k}{2}\rceil -1\) and \(d_G^k(v)\ge d_G^{3^-}(v)+1\ge d_G^{2^-}(v)+1\). Thus \(d_H^{3}(v)\le \lceil \frac{k}{2}\rceil -1-(k-10)\le 1\) and \(d_H^{6^-}(v)\le k-d_G^{2^-}(v)-d_G^{k}(v)\le k-(k-10)-(k-9)\le 3\). \(\square \)

By Fact 2, if v is of Type I, then \(n_{\mathrm{II}}(v)\le 1\) and \(n_{\mathrm{III}}(v)\le 5\); otherwise we have either \(n_{\mathrm{II}}(v)\le 1\), or \(n_{\mathrm{II}}(v)=2\) and \(n_{\mathrm{III}}(v)\le 4\). Noting that \(d_H^{3}(v)\le 1\), by R1–R5, we have \(w'(v)\ge 10-4-\max \{\frac{1}{2}+1+\frac{1}{2}\times 5+\frac{1}{3}\times 4,1+\frac{1}{2}\times 9,1\times 2+\frac{1}{2}\times {4}+\frac{1}{3}\times 4\}-\frac{1}{2}=0\).

\(\bullet \) \(d_{H^\times }(v)=11\). We first give the following fact.

Fact 3

If \(d_{H^\times }(v)=11\), then \(d_H^{3}(v)\le 2\) and \(d_H^{6^-}(v)\le 5-d_H^{3}(v)\).

Proof

By Table 1, we have \(d_G(v)\in \{11,12\}\) or \(d_G(v)\ge 16\). If \(d_G(v)=11\), by Claims 4 and 6, \(d_H^{3^-}(v)\le 1\) and \(d_H^{6^-}(v)\le 2\). If \(d_G(v)=12\), then \(d_G^{2^-}(v)=1\). By Claims 56, \(d_G^{3^-}(v)\le 1\) and \(d_G^{6^-}(v)\le 5\). Thus \(d_H^{3}(v)=0\) and \(d_H^{6^-}(v)\le 4\). If \(d_G(v)=k\) (\(k\ge 16\)), then \(d_G^{2^-}(v)=k-11\). By Claim 11, \(d_G^{3^-}(v)\le \lceil \frac{k}{2}\rceil -1\) and \(d_G^k(v)\ge d_G^{3^-}(v)+1\). Thus \(d_H^{3}(v)=d_G^{3}(v)\le \lceil \frac{k}{2}\rceil -1-(k-11)\le 2\) and \(d_H^{6^-}(v)\le k-d_G^{2^-}(v)-d_G^{k}(v)\le k-(k-11)-(k-10+d_G^{3}(v))\le 5-d_H^{3}(v)\). \(\square \)

- \(d_H^{3}(v)\ne 0\). By Fact 3, if v is of Type I, then \(n_{\mathrm{II}}(v)\le 1\) and \(n_{\mathrm{III}}(v)\le 7\); otherwise we have either \(n_{\mathrm{II}}(v)\le 1\) or \(n_{\mathrm{II}}(v)=2\) and \(n_{\mathrm{III}}(v)\le 6\). Noting that \(d_H^{3}(v)\le 2\), by R1–R5, we have \(w'(v)\ge 11-4-\max \{\frac{1}{2}+1+\frac{1}{2}\times 7+\frac{1}{3}\times 3,1+\frac{1}{2}\times 10,1\times 2+\frac{1}{2}\times 6+\frac{1}{3}\times 3\}-\frac{1}{2}\times 2=0\).

- \(d_H^{3}(v)=0\). By Fact 3, if v is of Type I, then \(n_{\mathrm{II}}(v)\le 2\), otherwise \(n_{\mathrm{II}}(v)\le 3\). By R1–R4, \(w'(v)\ge 11-4-\max \{\frac{1}{2}+1\times 2+\frac{1}{2}\times 9,1\times 3+\frac{1}{2}\times 8\}=0\).

\(\bullet \) \(d_{H^\times }(v)=12\). We first give the following fact.

Fact 4

If \(d_{H^\times }(v)=12\), then either \(d_H^{3}(v)\le 1\) and \(d_H^{5^-}(v)\le 7-d_H^{3}(v)\), or \(2\le d_H^{3}(v)\le 3\) and \(d_H^{6^-}(v)\le 7-d_H^{3}(v)\).

Proof

By Table 1, we have \(d_G(v)\ge 12\). (a): \(d_G(v)=12\). By Claims 56, \(d_H^{3}(v)\le 1\) and \( d_H^{5^-}(v)\le 5\). So, in this case, Fact 4 holds. (b): \(d_G(v)=k\) (\(13\le k\le 14\)). Then \(d_G^{2^-}(v)=k-12>0\), by Claim 7(2), let \(m=2\), we have \(d_G^{{k}}(v)\ge (17-k)d_G^{{(19-k)}^-}(v)+1\). Noting that \(d_G^{{(19-k)}^-}(v)+d_G^k(v)\le k\), we get that \(d_H^{{(19-k)}^-}(v)=d_G^{{(19-k)}^-}(v)-d_G^{2^-}(v)\le \lfloor \frac{k-1}{18-k}\rfloor -(k-12)=1\). So, in this case, Fact 4 holds. (c): \(d_G(v)=15\), then \(d_G^{2^-}(v)=3\). By Claim 8(2), \(d_H^{3}(v)=d_G^{3^-}(v)-d_G^{2^-}(v)\le 1\). By Claim 8(3), let \(m=2\), we have \(d_G^{15}(v)\ge 2d_G^{{4}^-}(v)+1\). Thus \(d_H^{5^-}(v)\le d_G(v)-d_G^{2^-}(v)-d_G^{15}(v)\le 14-3d_G^{2^-}(v)=5\). So, in this case, Fact 4 holds. (d): \(d_G(v)=k\) (\(k\ge 16\)), then \(d_G^{2^-}(v)=k-12\). By Claim 11, \(d_G^{3^-}(v)\le \lceil \frac{k}{2}\rceil -1\) and \(d_G^k(v)\ge d_G^{3^-}(v)+1\). Thus \(d_H^{3}(v)= d_G^{3}(v)\le \lceil \frac{k}{2}\rceil -1-(k-12)\le 3\) and \(d_H^{6^-}(v)\le k-d_G^{2^-}(v)-d_G^{k}(v)\le k-(k-12)-(k-11+d_G^{3}(v))\le 7-d_H^{3}(v)\). So, in this case, Fact 4 holds. \(\square \)

- \(d_H^{3}(v)=3\) and \(d_H^{6^-}(v)\le 4\). If v is of Type I, by Lemma 1, we have \(n_{\mathrm{II}}(v)\le 1\) and \(n_{4^+}(v)\ge 1\); otherwise we have either \(n_{\mathrm{II}}(v)\le 1\), or \(n_{\mathrm{II}}(v)=2\) and \(n_{\mathrm{III}}(v)\le 6\). By R1–R5, \(w'(v)\ge 12-4-\max \{\frac{1}{2}+1+\frac{1}{2}\times 10,1+\frac{1}{2}\times {11},1\times 2+\frac{1}{2}\times 6+\frac{1}{3}\times 4\}-\frac{1}{2}\times 3=0\).

- \(d_H^{3}(v)=2\) and \(d_H^{6^-}(v)\le 5\). If v is of Type I, then either \(n_{\mathrm{II}}(v)\le 1\), or \(n_{\mathrm{II}}(v)=2\) and \(n_{\mathrm{III}}(v)\le 6\); otherwise we have either \(n_{\mathrm{II}}(v)\le 2\), or \(n_{\mathrm{II}}(v)=3\) and \(n_{\mathrm{III}}(v)\le 6\). By R1–R5, \(w'(v)\ge 12-4-\max \{\frac{1}{2}+1+\frac{1}{2}\times 11,\frac{1}{2}+1\times 2+\frac{1}{2}\times 6+\frac{1}{3}\times 4,1\times 2+\frac{1}{2}\times 10,1\times 3+\frac{1}{2}\times 6+\frac{1}{3}\times 3\}-\frac{1}{2}\times 2=0\).

- \(d_H^{3}(v)\le 1\). By Fact 4, if v is of Type I, then \(n_{\mathrm{II}}(v)\le 3-d_H^{3}(v)\), otherwise \(n_{\mathrm{II}}(v)\le 4-d_H^{3}(v)\). By R1–R5, \(w'(v)\ge 12-4-\max \{\frac{1}{2}+1\times (3-d_H^{3}(v))+\frac{1}{2}\times (12-(3-d_H^{3}(v))),1\times (4-d_H^{3}(v))+\frac{1}{2}\times (12-(4-d_H^{3}(v)))\}-\frac{1}{2}d_H^{3}(v)=0\).

\(\bullet \) \(d_{H^\times }(v)=13\). We first give the following fact.

Fact 5

If \(d_{H^\times }(v)=13\), then \(d_H^{3}(v)\le 4\) and \(d_H^{5^-}(v)\le 9-d_H^{3}(v)\). Furthermore, if \(2\le d_H^{3}(v)\le 4\) and \(d_H^{5^-}(v)\ge 7-d_H^{3}(v)\), then v is not incident with any bad 3-cycle.

Proof

By Table 1, we have \(d_G(v)\ge 13\). (a): \(d_G(v)=13\). If \(d_G^3(v)\ge 1\), by Claim 7(2), \(d_G^{13}(v)\ge 3d_G^{6^-}(v)+1\). Noting that \(d_G^{6^-}(v)+d_G^{13}(v)\le 13\), we have \(d_H^{5^-}(v)\le d_G^{6^-}(v)\le 3\). If \(d_G^3(v)=0\) and \(d_G^{5^-}(v)\ge 1\), by Claim 7(2), \(d_G^{13}(v)\ge d_G^{5^-}(v)+1\). Noting that \(d_G^{5^-}(v)+d_G^{13}(v)\le 13\), we have \(d_H^{5^-}(v)\le d_G^{5^-}(v)\le 6\). So, in this case, Fact 5 holds. (b): \(d_G(v)=14\), then \(d_G^{2^-}(v)=1\). By Claim 7(2), let \(m=2\), we have \(d_G^{14}(v)\ge 3d_G^{{5}^-}(v)+1\). Noting that \(d_G^{5^-}(v)+d_G^{14}(v)\le 14\), we get that \(d_H^{5^-}(v)=d_G^{{5}^-}(v)-d_G^{2^-}(v)\le 3-1=2\). So, in this case, Fact 5 holds. (c): \(d_G(v)=15\), then \(d_G^{2^-}(v)=2\). By Claim 8(2), we have \(d_H^{3}(v)=d_G^{3^-}(v)-d_G^{2^-}(v)\le 2\). By Claim 8(3), let \(m=2\), we have \(d_G^{15}(v)\ge 2d_G^{{4}^-}(v)+1\). Thus \(d_H^{5^-}(v)\le d_G(v)-d_G^{2^-}(v)-d_G^{15}(v)\le 14-3d_G^{2^-}(v)-2d_G^{3}(v)=8-2d_H^{3}(v)\). So, in this case, Fact 5 holds. (d): \(d_G(v)=k\) (\(k\ge 16\)), then \(d_G^{2^-}(v)=k-13\). By Claim 11, \(d_G^{3^-}(v)\le \lceil \frac{k}{2}\rceil -1\) and \(d_G^k(v)\ge d_G^{3^-}(v)+1\). Thus \(d_H^{3}(v)=d_G^{3}(v)\le \lceil \frac{k}{2}\rceil -1-(k-13)\le 4\) and \(d_H^{5^-}(v)\le k-d_G^{2^-}(v)-d_G^{k}(v)\le k-(k-13)-(k-12+d_G^{3}(v))\le 9-d_H^{3}(v)\). Furthermore, suppose that \(2\le d_H^{3}(v)\le 4\) and \(d_H^{5^-}(v)\ge 7-d_H^{3}(v)\). Assume that v is incident with a bad 3-cycle, by Claim 10, \(d_G^{k}(v)\ge 2d_G^{4^-}(v)+1\). Noting that \(d_G^{2^-}(v)+d_H^{5^-}(v)+d_G^{k}(v)-k\le 0\), while \(d_G^{2^-}(v)+d_H^{5^-}(v)+d_G^{k}(v)-k\ge k-13+ 7-d_H^{3}(v)+2(k-13) +2d_H^{3}(v)+1-k>2k-31>0\), a contradiction. So, in this case, Fact 5 holds. \(\square \)

- \(d_H^{3}(v)=4\) and \(d_H^{5^-}(v)\le 5\). By Fact 5, v is not incident with any bad 3-cycle. If v is of Type I, by Lemma 1, \(n_{\mathrm{II}}(v)=0\), otherwise \(n_{\mathrm{II}}(v)\le 1\). By R1–R5, \(w'(v)\ge 13-4-\max \{\frac{1}{2}+\frac{1}{2}\times 13,1+\frac{1}{2}\times 12\}-\frac{1}{2}\times 4=0\).

- \(d_H^{3}(v)=3\). By Fact 5, \(d_H^{5^-}(v)=3\), or \(4\le d_H^{5^-}(v)\le 6\) and v is not incident with any bad 3-cycle. If v is of Type I, then \(n_{\mathrm{II}}(v)\le 1\), otherwise \(n_{\mathrm{II}}(v)\le 2\). By R1–R5, \(w'(v)\ge 13-4-\max \{\frac{1}{2}+1+\frac{1}{2}\times 12,1\times 2+\frac{1}{2}\times 11\}-\frac{1}{2}\times 3=0\).

- \(d_H^{3}(v)=2\). By Fact 5, \(d_H^{5^-}(v)\le 4\), or \(5\le d_H^{5^-}(v)\le 7\) and v is not incident with any bad 3-cycle. If v is of Type I, then \(n_{\mathrm{II}}(v)\le 2\), otherwise \(n_{\mathrm{II}}(v)\le 3\). By R1–R5, \(w'(v)\ge 13-4-\max \{\frac{1}{2}+1\times 2+\frac{1}{2}\times 11,1\times 3+\frac{1}{2}\times 10\}-\frac{1}{2}\times 2=0\).

- \(d_H^{3}(v)\le 1\). By Fact 5, if v is of Type I, then \(n_{\mathrm{II}}(v)\le 4-d_H^{3}(v)\), otherwise \(n_{\mathrm{II}}(v)\le 5-d_H^{3}(v)\). By R1–R5, \(w'(v)\ge 13-4-\max \{\frac{1}{2}+1\times (4-d_H^{3}(v))+\frac{1}{2}\times (13-(4-d_H^{3}(v))),1\times (5-d_H^{3}(v))+\frac{1}{2}\times (13-(5-d_H^{3}(v)))\}-\frac{1}{2}\times d_H^{3}(v)=0\).

\(\bullet \) \(d_{H^\times }(v)=14\). We first give the following fact.

Fact 6

If \(d_{H^\times }(v)=14\), then either \(d_H^{3}(v)=0\), or \(1\le d_H^{3}(v)\le 5\) and \(d_H^{5^-}(v)\le 11-d_H^{3}(v)\). Furthermore, if \( d_H^{3}(v)\ge 4\) and \(d_H^{5^-}(v)\ge 5\), or \(2\le d_H^{3}(v)\le 3\) and \(d_H^{5^-}(v)\ge 6\), then v is not incident with any bad 3-cycle.

Proof

By Table 1, we have \(d_G(v)\ge 14\). (a): \(d_G(v)=14\). If \(d_G^3(v)\ge 1\), by Claim 7(2), \(d_G^{14}(v)\ge 2d_G^{5^-}(v)+1\). Noting that \(d_G^{5^-}(v)+d_G^{14}(v)\le 14\), we have \(d_H^{5^-}(v)\le d_G^{5^-}(v)\le 4\). So, in this case, Fact 6 holds. (b): \(d_G(v)=15\), then \(d_G^{2^-}(v)=1\). By Claim 8(2), we have \(d_H^{3}(v)=d_G^{3^-}(v)-d_G^{2^-}(v)\le 3\). By Claim 8(3), let \(m=2\), we have \(d_G^{15}(v)\ge 2d_G^{{4}^-}(v)+1\). Thus \(d_H^{5^-}(v)\le d_G(v)-d_G^{2^-}(v)-d_G^{15}(v)\le 14-3d_G^{2^-}(v)-2d_G^{3}(v)-2d_G^{4}(v)\le 11-2d_H^{3}(v)\), which implies that \(d_H^{3}(v)\le 3\). Furthermore, if \(d_H^{5^-}(v)\ge 6\), by \(d_G^{2^-}(v)+d_H^{5^-}(v)+d_G^{15}(v)\le 15\) and Claim 8(4), v is not incident with any bad 3-cycle. So, in this case, Fact 6 holds. (c): \(d_G(v)=k\) (\(k\ge 16\)), then \(d_G^{2^-}(v)=k-14\). By Claim 11, \(d_G^{3^-}(v)\le \lceil \frac{k}{2}\rceil -1\) and \(d_G^k(v)\ge d_G^{3^-}(v)+1\). Thus \(d_H^{3}(v)=d_G^{3}(v)\le \lceil \frac{k}{2}\rceil -1-(k-14)\le 5\) and \(d_H^{5^-}(v)\le k-d_G^{2^-}(v)-d_G^{k}(v)\le k-(k-14)-(k-13+d_G^{3}(v))\le 11-d_H^{3}(v)\). Furthermore, suppose that \(d_H^{3}(v)\ge 4\) and \(d_H^{5^-}(v)\ge 5\), or \(2\le d_H^{3}(v)\le 3\) and \(d_H^{5^-}(v)\ge 6\). Assume that v is incident with a bad 3-cycle, by Claim 10, \(d_G^k(v)\ge 2d_G^{4^-}(v)+1\). Noting that \(d_G^{2^-}(v)+d_H^{5^-}(v)+d_G^{k}(v)-k\le 0\), while \(d_G^{2^-}(v)+d_H^{5^-}(v)+d_G^{k}(v)-k\ge d_G^{2^-}(v)+d_H^{5^-}(v)+2d_G^{4^-}(v)+1-k\ge 3d_G^{2^-}(v)+2d_H^{3}(v)+d_H^{5^-}(v)+1-k\ge 3(k-14)+\min \{2\times 4+5,2\times 2+6\}+1-k=2k-31>0\), a contradiction. So, in this case, Fact 6 holds. \(\square \)

By Fact 6, we consider the following cases.

- \(d_H^{3}(v)=5\) and \(d_H^{5^-}(v)\le 6\), or \(d_H^{3}(v)=4\) and \(5\le d_H^{5^-}(v)\le 7\). By Fact 6, v is not incident with any bad 3-cycle. If v is of Type I, by Lemma 1, we have \(n_{\mathrm{II}}(v)\le 5-d_H^{3}(v)\), otherwise \(n_{\mathrm{II}}(v)\le 6-d_H^{3}(v)\). By R1–R5, \(w'(v)\ge 14-4-\max \{\frac{1}{2}+1\times (5-d_H^{3}(v))+\frac{1}{2}\times (14-(5-d_H^{3}(v))),1\times (6-d_H^{3}(v))+\frac{1}{2}\times (14-(6-d_H^{3}(v)))\}-\frac{1}{2}d_H^{3}(v)=0\).

- \(d_H^{3}(v)=d_H^{5^-}(v)=4\). If v is of Type I, by Lemma 1, we have \(n_{\mathrm{II}}(v)\le 1\), otherwise, by Claim 9, we have \(n_{\mathrm{II}}(v)\le 2\). By R1–R5, \(w'(v)\ge 14-4-\max \{\frac{1}{2}+1+\frac{1}{2}\times 13,1\times 2+\frac{1}{2}\times 12\}-\frac{1}{2}\times 4=0\).

- \(2\le d_H^{3}(v)\le 3\) and \(d_H^{5^-}(v)\le 5\), or \(d_H^{3}(v)=3\) and \(6\le d_H^{5^-}(v)\le 8\) and v is not incident with any bad 3-cycle. If v is of Type I, then \(n_{\mathrm{II}}(v)\le 2\), otherwise \(n_{\mathrm{II}}(v)\le 3\). Noting that \(d_H^{3}(v)\le 3\), by R1–R5, we have \(w'(v)\ge 14-4-\max \{\frac{1}{2}+1\times 2+\frac{1}{2}\times 12,1\times 3+\frac{1}{2}\times 11\}-\frac{1}{2}\times 3=0\).

- \(d_H^{3}(v)=2\), \(6\le d_H^{5^-}(v)\le 9\) and v is not incident with any bad 3-cycle. If v is of Type I, then \(n_{\mathrm{II}}(v)\le 3\), otherwise \(n_{\mathrm{II}}(v)\le 4\). By R1–R5, \(w'(v)\ge 14-4-\max \{\frac{1}{2}+1\times 3+\frac{1}{2}\times 11,1\times 4+\frac{1}{2}\times 10\}-\frac{1}{2}\times 2=0\).

- \(d_H^{3}(v)=1\) and \(d_H^{5^-}(v)\le 10\). If v is of Type I, then \(n_{\mathrm{II}}(v)\le 4\), otherwise \(n_{\mathrm{II}}(v)\le 5\). By R1–R5, \(w'(v)\ge 14-4-\max \{\frac{1}{2}+1\times 4+\frac{1}{2}\times 10,1\times 5+\frac{1}{2}\times 9\}-\frac{1}{2}=0\).

- \(d_H^{3}(v)=0\). Then \(n_{\mathrm{II}}(v)\le 5\), or \(n_{\mathrm{II}}(v)=6\) and \(n_{4^+}(v)\ge 1\), or \(n_{\mathrm{II}}(v)=7\) and \(n_{4^+}(v)\ge 5\) by Claims 23. By R1–R4, \(w'(v)\ge 14-4-\frac{1}{2}-\max \{1\times 5+\frac{1}{2}\times 9,1\times 6+\frac{1}{2}\times 7,1\times 7+\frac{1}{2}\times 2\}=0\).

Remark 3

For any \(15^+\)-vertex \(v\in V(H^\times )\), if v is not incident with any bad 3-cycle and \(d_H^{3b}(v)\ge 2\), then \(n_{4^+}(v)\ge 1\).

\(\bullet \) \(d_{H^\times }(v)=15\). We first give the following fact.

Fact 7

If \(d_{H^\times }(v)=15\), then either \(d_H^{3}(v)=0\), or \(1\le d_H^{3}(v)\le 7\) and \(d_H^{6^-}(v)\le 14-d_H^{3}(v)\). Furthermore, if \( d_H^{3}(v)\ge 3\) and \(d_H^{6^-}(v)\ge 7\), or \(d_H^{3}(v)=2\) and \(d_H^{6^-}(v)\ge 9\), then v is not incident with any bad 3-cycle.

Proof

By Table 1, we have \(d_G(v)\ge 15\). (a): \(d_G(v)=15\). If \(d_G^{3}(v)\ge 1\), by Claim 8(3), let \(m=3\), we have \(d_G^{15}(v)\ge d_G^{4^-}(v)+1\). Noting that \(d_H^{3}(v)\le d_G^{4^-}(v)\le d_G^{6^-}(v)\le d_G(v)-d_G^{15}(v)\), we have \(d_H^{3}(v)\le 7\) and \(d_H^{6^-}(v)\le 14-d_H^3(v)\). Furthermore, if \(d_H^{6^-}(v)\ge 7\), by \(d_G^{6^-}(v)+d_G^{15}(v)\le 15\) and Claim 8(4), v is not incident with any bad 3-cycle. So, in this case, Fact 7 holds. (b): \(d_G(v)=k\) (\(k\ge 16\)), then \(d_G^{2^-}(v)=k-15\). By Claim 11, \(d_G^{3^-}(v)\le \lceil \frac{k}{2}\rceil -1\) and \(d_G^k(v)\ge d_G^{3^-}(v)+1\). Thus \(d_H^{3}(v)=d_G^{3}(v)\le \lceil \frac{k}{2}\rceil -1-(k-15)\le 6\) and \(d_H^{6^-}(v)\le k-d_G^{2^-}(v)-d_G^{k}(v)\le k-(k-15)-(k-14+d_G^{3}(v))<14-d_H^{3}(v)\). Furthermore, suppose that \(d_H^{3}(v)\ge 3\) and \(d_H^{6^-}(v)\ge 7\), or \(d_H^{3}(v)=2\) and \(d_H^{6^-}(v)\ge 9\). Assume that v is incident with a bad 3-cycle, by Claim 10, \(d_G^k(v)\ge 2d_G^{4^-}(v)+1\). Noting that \(d_G^{2^-}(v)+d_H^{6^-}(v)+d_G^{k}(v)-k\le 0\), while \(d_G^{2^-}(v)+d_H^{6^-}(v)+d_G^{k}(v)-k\ge d_G^{2^-}(v)+d_H^{6^-}(v)+2d_G^{4^-}(v)+1-k\ge 3d_G^{2^-}(v)+2d_H^{3}(v)+d_H^{6^-}(v)+1-k\ge 3(k-15)+\min \{2\times 3+7,2\times 2+9\}+1-k=2k-31>0\), a contradiction. So, in this case, Fact 7 holds.\(\square \)

By Fact 7, we consider the following cases.

- \(d_H^{3}(v)=d_H^{6^-}(v)=7\), and v is not incident with any bad 3-cycle. If v is of Type I, by Lemma 1, we have \(n_{\mathrm{II}}(v)=0\) and \(n_{4^+}(v)\ge 1\); otherwise we have \(n_{\mathrm{II}}(v)\le 1\) and either \(d_H^{3b}(v)\le 1\) or \(n_{4^+}(v)\ge 1\) by Remark 3. By R1–R5, \(w'(v)\ge 15-4-\max \{\frac{1}{2}+\frac{1}{2}\times 14+\frac{1}{2}\times 7,1+\frac{1}{2}\times 14+\frac{1}{2}+\frac{1}{3}\times 6,1+\frac{1}{2}\times 13+\frac{1}{2}\times 7\}=0\).

- \(d_H^{3}(v)=d_H^{6^-}(v)=6\). By Claim 9, v is incident with at most one bad 3-cycle. If v is of Type I, by Lemma 1, then \(n_{\mathrm{II}}(v)\le 1\) and \(n_{4^+}(v)\ge 1\); otherwise we have either \(n_{\mathrm{II}}(v)\le 1\), or \(n_{\mathrm{II}}(v)=2\) and \(n_{\mathrm{III}}(v)\le 10\). By R1–R5, \(w'(v)\ge 15-4-\max \{\frac{1}{2}+1+\frac{1}{2}\times 13,1+\frac{1}{2}\times 14,1\times 2+\frac{1}{2}\times 10+\frac{1}{3}\times 3\}-\frac{1}{2}\times 6=0\).

- \(d_H^{3}(v)=6\), \(7\le d_H^{6^-}(v)\le 8\), and v is not incident with any bad 3-cycle. If v is of Type I, we have \(n_{\mathrm{II}}(v)\le 1\) and either \(d_H^{3b}(v)\le 1\) or \(n_{4^+}(v)\ge 1\) by Remark 3; otherwise we have \(n_{\mathrm{II}}(v)\le 2\) and either \(d_H^{3b}(v)\le 1\) or \(n_{4^+}(v)\ge 1\) by Remark 3. By R1–R5, \(w'(v)\ge 15-4-\max \{\frac{1}{2}+1+\frac{1}{2}\times 14+\frac{1}{2}+\frac{1}{3}\times 5,\frac{1}{2}+1+\frac{1}{2}\times 13+\frac{1}{2}\times 6,1\times 2+\frac{1}{2}\times 13+\frac{1}{2}+\frac{1}{3}\times 5,1\times 2+\frac{1}{2}\times 12+\frac{1}{2}\times 6\}=0\).

- \(d_H^{3}(v)=5\) and \(d_H^{6^-}(v)\le 6\). By Claim 9, v is incident with at most one bad 3-cycle. If v is of Type I, then \(n_{\mathrm{II}}(v)\le 1\), otherwise \(n_{\mathrm{II}}(v)\le 2\). By R1–R5, \(w'(v)\ge 15-4-\max \{\frac{1}{2}+1+\frac{1}{2}\times 14,1\times 2+\frac{1}{2}\times 13\}-\frac{1}{2}\times 5=0\).

- \(d_H^{3}(v)=5\), \(7\le d_H^{6^-}(v)\le 9\), and v is not incident with any bad 3-cycle. If v is of Type I, then either \(n_{\mathrm{II}}(v)\le 1\), or \(n_{\mathrm{II}}(v)=2\) and \(n_{4^+}(v)\ge 1\) by Claims 23; otherwise we have \(n_{\mathrm{II}}(v)\le 3\) and either \(d_H^{3b}(v)\le 1\) or \(n_{4^+}(v)\ge 1\) by Remark 3. By R1–R5, \(w'(v)\ge 15-4-\max \{\frac{1}{2}+1+\frac{1}{2}\times 14+\frac{1}{2}\times 5,\frac{1}{2}+1\times 2+\frac{1}{2}\times 12+\frac{1}{2}\times 5,1\times 3+\frac{1}{2}\times 12+\frac{1}{2}+\frac{1}{3}\times 4,1\times 3+\frac{1}{2}\times 11+\frac{1}{2}\times 5\}=0\).

- \(3\le d_H^{3}(v)\le 4\) and \(d_H^{6^-}(v)\le 6\). By Claim 9, v is incident with at most one bad 3-cycle. If v is of Type I, then \(n_{\mathrm{II}}(v)\le 2\), otherwise \(n_{\mathrm{II}}(v)\le 3\). Noting that \(d_H^{3}(v)\le 4\), by R1–R5, we have \(w'(v)\ge 15-4-\max \{\frac{1}{2}+1\times 2+\frac{1}{2}\times 13,1\times 3+\frac{1}{2}\times 12\}-\frac{1}{2}\times 4=0\).

- \(3\le d_H^{3}(v)\le 4\), \(7\le d_H^{6^-}(v)\le 14-d_H^{3}(v)\), and v is not incident with any bad 3-cycle. If v is of Type I, then either \(n_{\mathrm{II}}(v)\le 6-d_H^{3}(v)\), or \(n_{\mathrm{II}}(v)=7-d_H^{3}(v)\) and \(n_{4^+}(v)\ge 1\) by Claims 23; otherwise we have either \(n_{\mathrm{II}}(v)\le 7-d_H^{3}(v)\), or \(n_{\mathrm{II}}(v)=8-d_H^{3}(v)\) and \(n_{4^+}(v)\ge 1\) by Claims 23. By R1–R5, \(w'(v)\ge 15-4-\max \{\frac{1}{2}+1\times (6-d_H^{3}(v))+\frac{1}{2}\times (15-(6-d_H^{3}(v))),\frac{1}{2}+1\times (7-d_H^{3}(v))+\frac{1}{2}\times (14-(7-d_H^{3}(v))),1\times (7-d_H^{3}(v))+\frac{1}{2}\times (15-(7-d_H^{3}(v))),1\times (8-d_H^{3}(v))+\frac{1}{2}\times (14-(8-d_H^{3}(v)))\}-\frac{1}{2}d_H^{3}(v)=0\).

- \(d_H^{3}(v)=2\) and \(d_H^{6^-}(v)\le 8\). If v is of Type I, then \(n_{\mathrm{II}}(v)\le 4\), otherwise \(n_{\mathrm{II}}(v)\le 5\). By R1–R5, \(w'(v)\ge 15-4-\max \{\frac{1}{2}+1\times 4+\frac{1}{2}\times 11,1\times 5+\frac{1}{2}\times 10\}-\frac{1}{2}\times 2=0\).

- \(d_H^{3}(v)=2\), \(9\le d_H^{6^-}(v)\le 12\), and v is not incident with any bad 3-cycle. If v is of Type I, then either \(n_{\mathrm{II}}(v)\le 4\), or \(n_{\mathrm{II}}(v)=5\) and \(n_{4^+}(v)\ge 1\) by Claims 23; otherwise we have either \(n_{\mathrm{II}}(v)\le 5\), or \(n_{\mathrm{II}}(v)=6\) and \(n_{4^+}(v)\ge 1\) by Claims 23. By R1–R5, \(w'(v)\ge 15-4-\max \{\frac{1}{2}+1\times 4+\frac{1}{2}\times 11,\frac{1}{2}+1\times 5+\frac{1}{2}\times 9,1\times 5+\frac{1}{2}\times 10,1\times 6+\frac{1}{2}\times 8\}-\frac{1}{2}\times 2=0\).

- \(d_H^{3}(v)\le 1\) and \(d_H^{6^-}(v)\le 15-2d_H^{3}(v)\). If v is of Type I, then either \(n_{\mathrm{II}}(v)\le 6-d_H^{3}(v)\), or \(n_{\mathrm{II}}(v)=7-d_H^{3}(v)\) and \(n_{4^+}(v)\ge 1\) by Claims 23; otherwise we have either \(n_{\mathrm{II}}(v)\le 7-d_H^{3}(v)\), or \(n_{\mathrm{II}}(v)=8-d_H^{3}(v)\) and \(n_{4^+}(v)\ge 1\) by Claims 23. By R1–R5, \(w'(v)\ge 15-4-\max \{\frac{1}{2}+1\times (6-d_H^{3}(v))+\frac{1}{2}\times (15-(6-d_H^{3}(v))),\frac{1}{2}+1\times (7-d_H^{3}(v))+\frac{1}{2}\times (14-(7-d_H^{3}(v))),1\times (7-d_H^{3}(v))+\frac{1}{2}\times (15-(7-d_H^{3}(v))),1\times (8-d_H^{3}(v))+\frac{1}{2}\times (14-(8-d_H^{3}(v)))\}-\frac{1}{2}d_H^{3}(v)=0\).

\(\bullet \) \(d_{H^\times }(v)=k\) (\(k\ge 16\)). By Claim 2 and Table 1, every \(5^-\)-vertex has at most one conflict vertex.

- \(d_H^{3}(v)=0\). (a): \(3n_{\mathrm{II}}(v)\le {k+5}\). By R1–R4, \(w'(v)\ge k-4-\frac{1}{2}-1\times n_{\mathrm{II}}(v)-\frac{1}{2}\times (k-n_{\mathrm{II}}(v))\ge \frac{k-16}{3}\ge 0\). (b): \(3n_{\mathrm{II}}(v)>{k+5}\). Note that a 3-face of Type II is incident with two \(5^-\)-vertices. If v is not adjacent to any false vertex, then \(d_{H}^{9^+}(v)\le k-2n_{\mathrm{II}}(v)\) and \(n_{4^+}(v)\ge n_{\mathrm{II}}(v)-d_{H}^{9^+}(v)\); otherwise we have \(d_{H}^{9^+}(v)\le k-2n_{\mathrm{II}}(v)+2\) and \(n_{4^+}(v)\ge n_{\mathrm{II}}(v)-d_{H}^{9^+}(v)-1\). Thus \(n_{4^+}(v)\ge \min \{n_{\mathrm{II}}(v)-(k-2n_{\mathrm{II}}(v)),n_{\mathrm{II}}(v)-(k-2n_{\mathrm{II}}(v)+2)-1\}=3n_{\mathrm{II}}(v)-k-3\). By R1–R4, \(w'(v)\ge k-4-\frac{1}{2}-1\times n_{\mathrm{II}}(v)-\frac{1}{2}\times (k-n_{\mathrm{II}}(v)-n_{4^+}(v))=\frac{1}{2}(k-n_{\mathrm{II}}(v)+n_{4^+}(v)-9)\ge \frac{1}{2}(k- n_{\mathrm{II}}(v)+(3n_{\mathrm{II}}(v)-k-3)-9)=n_{\mathrm{II}}(v)-6>\frac{k+5}{3}-6>0\).

- \(d_H^{3}(v)\ge 1\) and v is incident with a bad 3-cycle. (a): If v is of Type I, then v is not incident with any \((4,5,16^+)\)-face. By Lemma 1 and Claim 9, we have \(n_{\mathrm{II}}(v)\le \frac{d_H^4(v)}{2}+1\). (b): If v is not of Type I. Noting that v is incident with at most two \((4,5,16^+)\)-faces, we have \(n_{\mathrm{II}}(v)\le \frac{d_H^4(v)}{2}+3\). By Claim 10, \(d_H^3(v)+d_H^4(v)\le d_G^{4^-}(v)\le \frac{k-1}{3}\). By R1–R5, \(w'(v)\ge k-4-\max \{\frac{1}{2}+1\times (\frac{d_H^4(v)}{2}+1)+\frac{1}{2}\times (k-\frac{d_H^4(v)}{2}-1),1\times (\frac{d_H^4(v)}{2}+3)+\frac{1}{2}\times (k-\frac{d_H^4(v)}{2}-3)\}-\frac{1}{2}d_H^3(v)=\frac{k}{2}-\frac{d_H^3(v)}{2}-\frac{d_H^4(v)}{4}-\frac{11}{2}\ge \frac{k-11}{2}-\frac{d_H^3(v)+d_H^4(v)}{2}\ge \frac{k-11}{2}-\frac{k-1}{6}=\frac{k-16}{3}\ge 0\).

- \(d_H^{3}(v)\ge 1\) and v is not incident with any bad 3-cycle.

- - \(n_{4^+}(v)=0\). Then \(3n_{\mathrm{II}}(v)+2d_H^3(v)-4\le k\). By Remark 3, \(d_H^{3b}(v)\le 1\). By R1–R5, \(w'(v)\ge k-4-\frac{1}{2}-1\times n_{\mathrm{II}}(v)-\frac{1}{2}\times (k-n_{\mathrm{II}}(v))-(\frac{1}{2}+\frac{1}{3}(d_H^3(v)-1))=\frac{k}{2}-\frac{1}{6}(3n_{\mathrm{II}}(v)+2d_H^3(v))-\frac{14}{3}\ge \frac{k}{2}-\frac{k+4}{6}-\frac{14}{3}=\frac{k-16}{3}\ge 0\).

- - \(n_{4^+}(v)\ge 1\) and \(3n_{\mathrm{II}}(v)+2d_H^{3g}(v)+3\lfloor \frac{d_H^{3b}(v)}{2}\rfloor \le k+4\). Note that v is not incident with any bad 3-cycle. If v is of Type I, \(n_{4^+}(v)\ge \lceil \frac{d_H^{3b}(v)}{2}\rceil \); otherwise we have \(n_{4^+}(v)\ge \lceil \frac{d_H^{3b}(v)}{2}\rceil -1\). By R1–R5, \(w'(v)\ge k-4-\max \{\frac{1}{2}+1\times n_{\mathrm{II}}(v)+\frac{1}{2}\times (k-n_{ \mathrm {II}}(v)-\lceil \frac{d_H^{3b}(v)}{2}\rceil ),1\times n_{\mathrm{II}}(v)+\frac{1}{2}\times (k-n_{ \mathrm {II}}(v)-(\lceil \frac{d_H^{3b}(v)}{2}\rceil -1))\}-(\frac{1}{3}\times d_H^{3g}(v)+\frac{1}{2}\times d_H^{3b}(v))=\frac{k-9}{2}-\frac{n_{\mathrm{II}}(v)}{2}-\frac{d_H^{3g}(v)}{3}-\frac{d_H^{3b}(v)}{2}+\frac{1}{2}\lceil \frac{d_H^{3b}(v)}{2}\rceil =\frac{k-9}{2}-\frac{1}{6}(3n_{\mathrm{II}}(v)+2d_H^{3g}(v)+3\lfloor \frac{d_H^{3b}(v)}{2}\rfloor )\ge \frac{k-9}{2}-\frac{k+4}{6}=\frac{2k-31}{6}>0\).

- - \(n_{4^+}(v)\ge 1\) and \(3n_{\mathrm{II}}(v)+2d_H^{3g}(v)+3\lfloor \frac{d_H^{3b}(v)}{2}\rfloor \ge k+5\). Note that v is not incident with any bad 3-cycle. If v is of Type I, \(n_{4^+}(v)\ge 3n_{\mathrm{II}}(v)+2d_H^{3g}(v)+3\lfloor \frac{d_H^{3b}(v)}{2}\rfloor -(k+5)+\lceil \frac{d_H^{3b}(v)}{2}\rceil \); otherwise we have \(n_{4^+}(v)\ge 3n_{\mathrm{II}}(v)+2d_H^{3g}(v)+3\lfloor \frac{d_H^{3b}(v)}{2}\rfloor -(k+5)+(\lceil \frac{d_H^{3b}(v)}{2}\rceil -1)\). By R1–R5, \(w'(v)\ge k-4-\max \{\frac{1}{2}+1\times n_{\mathrm{II}}(v)+\frac{1}{2}\times (k-n_\mathrm{{II}}(v)-(3n_{\mathrm{II}}(v)+2d_H^{3g}(v)+3\lfloor \frac{d_H^{3b}(v)}{2}\rfloor -(k+5)+\lceil \frac{d_H^{3b}(v)}{2}\rceil )),1\times n_{\mathrm{II}}(v)+\frac{1}{2}\times (k-n_\mathrm{{II}}(v)-(3n_{\mathrm{II}}(v)+2d_H^{3g}(v)+3\lfloor \frac{d_H^{3b}(v)}{2}\rfloor -(k+5)+\lceil \frac{d_H^{3b}(v)}{2}\rceil -1))\}-(\frac{1}{3}\times d_H^{3g}(v)+\frac{1}{2}\times d_H^{3b}(v)) ={n_{\mathrm{II}}(v)}+\frac{2}{3}d_H^{3g}(v)+\frac{1}{2}\lceil \frac{d_H^{3b}(v)}{2}\rceil +\frac{3}{2}\lfloor \frac{d_H^{3b}(v)}{2}\rfloor -\frac{d_H^{3b}(v)}{2}-7=\frac{1}{3}(3{n_{\mathrm{II}}(v)}+2{d_H^{3g}(v)}+3\lfloor \frac{d_H^{3b}(v)}{2}\rfloor )-7\ge \frac{k+5}{3}-7\ge 0\).

In conclusion, the new charge of \(x\in V({H^\times })\cup F({H^\times })\) is nonnegative, a contradiction. The proof of Theorem 1 is done.