Closure of knitted surfaces and surface-links
Extracted source text · arXiv 2412.19432
\documentclass[a4paper,11pt,reqno]{amsart}
\usepackage{amsmath,amsfonts,amsthm,amssymb,enumerate}
% \usepackage[dvipdfmx]{graphicx}
\usepackage{graphicx}
\usepackage{comment}
%\usepackage{svg}
%\usepackage{mediabb}
%\usepackage{here}
%\graphicspath{ {./picture1/}}
\numberwithin{equation}{section}
%\numberwithin{figure}{section}
% Theorem environment
\newtheorem{theorem}{Theorem}[section]
\newtheorem{corollary}[theorem]{Corollary}
\newtheorem{proposition}[theorem]{Proposition}
\newtheorem{conjecture}[theorem]{Conjecture}
%\newtheorem{claim}[theorem]{Claim}
\newtheorem{lemma}[theorem]{Lemma}
%\newtheorem{remark}[theorem]{Remark}
\newtheorem{question}[theorem]{Question}
\newtheorem{fact}[theorem]{Fact}
\newtheorem*{fact*}{Fact}
{
\theoremstyle{definition}
\newtheorem{definition}[theorem]{Definition}
\newtheorem{example}[theorem]{Example}
\newtheorem{remark}[theorem]{Remark}
\newtheorem{claim}[theorem]{Claim}
}
% Definition with Roman Letters.
%\newenvironment{df}{\begin{definition}\rm}{\end{definition}}
\newenvironment{rem}{\begin{remark}\rm}{\end{remark}}
%\newenvironment{ex}{\begin{example}\rm}{\end{example}}
\title{Closure of knitted surfaces and surface-links}
\author{Inasa Nakamura}
\address{Department of Mathematics, Information Science and Engineering,
Saga University,
1 Honjomachi, Saga, 840-8502, Japan.
TEL: +81-952-28-8113}
\email{inasa@cc.saga-u.ac.jp}
\author{Jumpei Yasuda}
\address{Department of Mathematics, Graduate School of Science, Osaka University, 1-1, Machikaneyama, Toyonaka, Osaka 560-0043, Japan}
\email{u444951d@ecs.osaka-u.ac.jp}
\subjclass[2020]{Primary: 57K45, Secondary: 57Q35 57K10}
\keywords{surfaces in 4-space; surface-links; surface braids; tangles; knits; closure; plat closure}
\pagestyle{plain}
\begin{document}
\begin{abstract}
A knitted surface is a surface with boundary properly embedded in $D^2 \times B^2$, which is a generalization of a braided surface.
A knitted surface is called a 2-dimensional knit if its boundary is the closure of a trivial braid. From a 2-dimensional knit $S$, we obtain a surface-link in $\mathbb{R}^4$ by taking the closure of $S$. We show that any surface-link is ambiently isotopic to the closure of some 2-dimensional knit.
Further, we consider another type of the closure of a knitted surface, called the plat closure. It is known that any trivial surface-knot is ambiently isotopic to the plat closure of a knitted surface of degree 2. We show that the plat closure of any knitted surface of degree 2 is a trivial surface-link, and any trivial surface-link is ambiently isotopic to the plat closure of a knitted surface of degree $2$. We also show the same result for the closure of 2-dimensional knits of degree 2.
\end{abstract}
\maketitle
\section{Introduction}\label{sec1}
Let $n$ be a positive integer.
An {\it $n$-knit} \cite{Birman-Wenzl, Murakami} is a tangle obtained from an $n$-braid in a cylinder $D^2 \times I$ by splicing at some crossings. Splicing a crossing results in a \lq\lq hook pair'', and we describe the information of each hook pair
by a segment called a pairing \cite{NY}. We give an equivalence relation to the set of $n$-knits so that two $n$-knits are equivalent if they are related by an isotopy of $n$-knits.
The set of equivalence classes of $n$-knits forms a monoid $D_n$
by the multiplication induced from the composition of knits, which is called the \textit{$n$-knit monoid}.
The $n$-knit monoid $D_n$ is generated by $3(n-1)$ elements, $\sigma_1, \ldots, \sigma_{n-1}$, $\sigma_1^{-1}, \ldots, \sigma_{n-1}^{-1}$, $\tau_1,\ldots, \tau_{n-1}$, where $\sigma_i$ and $\sigma_i^{-1}$ are the standard generator and its inverse of the $n$-braid group $B_n$, and $\tau_i$ is the hook pair between the $i$th and $(i+1)$th strings as in Figure \ref{figure1} $(i=1, \ldots, n-1)$.
Let $e$ be the unit element of $D_n$, which is represented by the trivial braid.
The $n$-knit monoid $D_n$ has the monoid presentation
\begin{align*}
\left\langle
\begin{array}{c|}
\sigma_1, \ldots, \sigma_{n-1},\\ \sigma_1^{-1}, \ldots, \sigma_{n-1}^{-1},\\ \tau_1,\ldots, \tau_{n-1}
\end{array}
\begin{array}{l}
\sigma_i \sigma_{i}^{-1} = \sigma_i^{-1} \sigma_i = e,~
\sigma_i \tau_i = \tau_i \sigma_i = \tau_i,\\
\sigma_i\sigma_k = \sigma_k \sigma_i,~
\sigma_i \tau_k = \tau_k \sigma_i,~
\tau_i \tau_k = \tau_k \tau_i \ \ (|i-k|>1), \\
\sigma_i\sigma_j\sigma_i = \sigma_j\sigma_i\sigma_j,~
\sigma_i \sigma_j \tau_i = \tau_j \sigma_i \sigma_j \ \ (|i-j|=1)
\end{array}
\right\rangle_{\mathrm{monoid}}.
\end{align*}
\begin{figure}[ht]
\includegraphics*[height=3.5cm]{20240615-1a}
\caption{Generators $\sigma_i$, $\sigma^{-1}$ and $\tau_i$ of the monoid $D_n$ of $n$-knits.
%file: 20240615-1a. label:figure1
}\label{figure1}
\end{figure}
A knitted surface \cite{NY} in a bidisk $D^2 \times B^2$ is defined
as an analogue to the notion of an $n$-knit in $D^2 \times I$.
A knitted surface is an extended notion of a braided surface \cite{Rudolph}, which is a surface in $D^2 \times B^2$ in the form of a branched covering over $B^2$. A knitted surface is constructed as the trace of deformations of knits, as follows. Let $\beta_0 \to \beta_1 \to \beta_2 \to \cdots \beta_m$ be a sequence of elements of the free monoid consisting of $\sigma_i, \sigma_i^{-1}$ and $\tau_i$ $(i=1. \ldots, n-1)$
such that $\beta_{j-1}$ and $\beta_j$ are related by one of the following for each $j$ $(j=1, \ldots, m)$:
\begin{enumerate}[(1)]
\item
One of the relations of the $n$-knit monoid $D_n$.
\item
Replacement of $e$ and $\sigma_i^\epsilon$, or replacement of $e$ and $\tau_i$, for some $i$ and $\epsilon \in \{+1, -1\}$.
\item
Replacement of $\tau_i$ and $\tau_i \tau_i$ for some $i$.
\end{enumerate}
We decompose $I=[0,1]$ into $m$ subintervals: $I=I_1 \cup \cdots \cup I_m$, $I_j=[(j-1)/m, j/m]$ $(j=1,\ldots, m)$.
A knitted surface $S$ in $D^2 \times I \times I$ is constructed for each $S \cap (D^2 \times I \times I_j)$ as follows. We denote by the same notation $\beta_j$ a geometric $n$-knit given by the presentation $\beta_j$.
For case (1), let $\{S_{[t]}\}_{t\in I_j}$ be an isotopy of $n$-knits in $D^2\times I$ relating $\beta_{j-1}$ and $\beta_j$ which describes the knit monoid relation. Then we construct $S \cap (D^2 \times I \times I_j)$ by $\cup_{t \in I_j} S_{[t]} \times \{t\}$. This construction is called the {\it trace} of $\{S_{[t]}\}_{t\in I_j}$.
The construction is determined by giving $S_{[t]}$, which is called a {\it slice} of $S$ at $t$.
For case (2), we construct $S \cap (D^2 \times I \times I_j)$ by giving slices as follows: $S_{[t]}=\beta_{j-1}$ $(t \in [(j-1)/m, (2j-1)/2m))$, $S_{[t]}=\beta_j$ $(t \in (2j-1)/2m, j/m])$ and $S_{[t]}$ $(t=(2j-1)/2m)$ is the union of $\beta_{j-1}$ and a twisted band (respectively, an untwisted band) corresponding to the $\sigma_i^\epsilon$ (respectively $\tau_i$). See Figure \ref{figure-branch}.
The $n$-knit $\beta_j$ is the result of band surgery along a band attaching to $\beta_{j-1}$, and the band corresponds to a branch point of the knitted surface.
For case (3), let $S_{[t]}=\beta_{j-1}$ $(t \in [(j-1)/m, (2j-1)/2m))$, $S_{[t]}=\beta_j$ $(t \in ((2j-1)/2m, j/m])$ and let $S_{[t]}$ $(t=(2j-1)/2m)$ be the union of $\beta_{j-1}$ and a disk whose boundary is the simple closed curve formed by the $\tau_i \tau_i$. See Figure \ref{figure-tau}.
The knitted surface is the trace of deformations of knits where a simple closed curve is born or dies.
\\
A {\it surface-link} is a closed surface smoothly embedded in $\mathbb{R}^4$. In particular, a surface-link with one component is called a {\it surface-knot}. A braided/knitted surface is called a {\it 2-dimensional braid/knit} if its boundary is the closure of the trivial braid.
From a 2-dimensional braid $S$, we obtain an orientable surface-link in $\mathbb{R}^4$ by taking the closure of $S$, and Kamada \cite{Kamada94} showed that every orientable surface-link is ambiently isotopic to the closure of some 2-dimensional braid.
For a 2-dimensional knit $S$, we can also define the closure of $S$ by a similar way, which provides a surface-link.
In this paper, we show Claim 9.1 in \cite{NY} as follows.
\begin{theorem}\label{Theorem: Alexander theorem for surface-links}
Every surface-link in $\mathbb{R}^4$ is ambiently isotopic to the closure of some 2-dimensional knit.
\end{theorem}
Further, we consider another type of the closure of a knitted surface, called the plat closure. In \cite[Theorem 8.1]{NY}, we showed that any trivial surface-knot is ambiently isotopic to the plat closure of some knitted surface of degree 2. In this paper, we show the following.
\begin{theorem}\label{thm20241213}
The plat closure of any knitted surface of degree 2 is a trivial surface-link, and any trivial surface-link is ambiently isotopic to the plat closure of a knitted surface of degree 2.
\end{theorem}
Theorem \ref{thm20241213} follows from Corollary \ref{cor5-5} and Theorem~\ref{thm6-7}, where Theorem \ref{thm6-7} is \cite[Claim 8.3]{NY}.
We show this by using a chart description of a knitted surface, which is a generalization of a chart description of a simple braided surface \cite{Kamada96, Kamada92}. As a corollary of Theorem \ref{thm20241213}, we have the following.
\begin{corollary}\label{thm20241217}
The closure of any 2-dimensional knit degree 2 is a trivial surface-link, and any trivial surface-link is ambiently isotopic to the closure of some 2-dimensional knit of degree 2.
\end{corollary}
The paper is organized as follows.
In Section \ref{sec-knitted-surface}, we review knitted surfaces.
In Section \ref{sec3}, we review the closure of knitted surfaces, and we show Theorem \ref{Theorem: Alexander theorem for surface-links}.
In Section \ref{sec5}, we review a chart description of knitted surfaces of degree 2.
In Section \ref{sec8}, we show Theorems \ref{thm6-7} and \ref{thm20241213} and Corollary \ref{thm20241217}.
We give references \cite{CKS, Kamada02, Kawauchi} for the basics of 1-dimensional and 2-dimensional knot theory.
In this paper, we assume that surfaces and tangles are smooth.
\section{Knitted surfaces}\label{sec-knitted-surface}
In this section, we review knitted surfaces \cite{NY}.
\subsection{Motion picture method}\label{sec2}
In order to describe surfaces in 4-space, we use motion pictures, regarding surfaces as the trace of slices of the surfaces by 3-spaces along the 4th dimensional axis.
In this section, we review motion picture method.
Let $S$ be a surface properly embedded in $D^2 \times B^2$, and we denote $B^2=I_3 \times I_4$, where $I_3=I_4=[0,1]$.
We modify the saddle points and minimal/maximal points of $S$ into saddle bands and minimal/maximal disks, repsectively;
see \cite[Section 8.4]{Kamada02}.
Let $p: D^2 \times B^2 \to D^2 \times I_3$ be the projection.
For each $t \in I_4$, the \textit{slice} of $S$ at $t$, denoted by $S_{[t]}$, is defined as the image by $p$ of the intersection $S \cap (D^2\times I_3 \times \{t\})$: that is, that is, $S_{[t]}=p(S \cap (D^2 \times B^2))$.
When each slice $S_{[t]}$ is either (1) a tangle (or a link) or (2) a union of a tangle and several disks (and bands), we describe $S$ by a finite sequence of diagrams of $S_{[t_0]}, \ldots, S_{[t_m]}$ consisting unions of (2) and tangles before and after the time of (2). This sequence of diagrams is called a {\it motion picture} of $S$.
\subsection{Band surgery of a tangle along a band set}
Let $\beta$ be a tangle in $D^2 \times I$.
A {\it band} attaching to $\beta$ is a 2-disk $B$ in $D^2 \times I$ such that $B \cap \beta$
consists of a pair of intervals in $\partial B$. The result of {\it band surgery along $B$}, denoted by $h(\beta; B)$, is the tangle
$\mathrm{Cl}((\beta\cup \partial B)\backslash \beta \cap B)$. We remark that by a slight modification we assume that $h(\beta; B)$ is also smoothly embedded.
Let $\mathcal{B}=\{B_1, \ldots, B_m\}$ be a set of mutually disjoint bands attaching to $\beta$. Put $\cup \mathcal{B}=B_1 \cup B_2 \cup \cdots \cup B_m$. The result of {\it band surgery along bands of $\mathcal{B}$} is given by
\[
h(\beta; \mathcal{B}):=\mathrm{Cl}((\beta\cup \partial (\cup \mathcal{B}))\backslash \beta \cap (\cup \mathcal{B})).
\]
\subsection{Knitted surface presented by a transformation of knits}
For an $n$-knit $\beta$, we regard $\beta$ as a geometric knit and we identify $\beta$ with its presentation $s_1\cdots s_m$ where $s_i\in \{e, \sigma_1^{\pm 1}, \ldots, \sigma_{n-1}^{\pm 1}, \tau_1, \ldots, \tau_{n-1}\}$ for each $i$.
We call $s_1 \cdots s_m$, $s_i$ ($i=1, \ldots, m$), and $w = s_{n_1}s_{n_1 +1} \cdots s_{n'_1}$ ($1 \leq n_1 \leq n_1' \leq n$) a \textit{word} of $\beta$, a \textit{letter} of $\beta$, and a \textit{subword} of $\beta$, respectively.
For subwords $w$ and $w'$ of two $n$-knits $\beta$ and $\beta'$, respectively, we call a sequence $w \to w'$ a \textit{transformation} if $\beta'$ is obtained from $\beta$ by changing $w$ to $w'$.
In this situation, we call a sequence $\beta \to \beta'$ a \textit{transformation} of an $n$-knit.
For $l$ subwords $w_1 \cdots w_l$ of $\beta$ and $l$ transformations $w_1 \to w_1'$, \ldots, $w_l \to w_l'$, we say that the transformations are {\it mutually disjoint} if $w_i$ and $w_j$ do not overlap each other for $i \neq j$, that is, when $w_i=s_{n_i} \cdots s_{n_i'}$ and $w_j=s_{n_j} \cdots s_{n_j'}$, the intersection of intervals $[n_i, n_i']\cap [n_j, n_j']$ is the empty set.
When we have a transformation of $n$-knits
$\beta \to \beta'$ as follow, we construct a knitted surface $S$ of degree $n$ in $D^2 \times B^2$ in such a way as follows; since the construction is unique up to equivalence, we call $S$ the {\it knitted surface presented by $\beta \to \beta'$}. In the following (Case 2) and (Case 3), $\beta \leftrightarrow \beta'$ denotes the sequences $\beta \to \beta'$ and $\beta' \to \beta$.
\noindent
(Case 1) A transformation $\beta \to \beta'$, where $\beta$ and $\beta'$ is related by an isotopy of $n$-knits $\{\beta_t\}_{t \in [0,1]}$: $\beta_0=\beta$ and $\beta'=\beta_1$.
In this case, we define $S$ by
\[
S_{[t]}=\beta_t
\]
for any $t \in [0,1]$.
\noindent
(Case 2) A transformation $e \leftrightarrow \sigma_i^\epsilon$ or $e \leftrightarrow \tau_i$. This is a transformation $\beta \to \beta'$ such that
$\beta'$ is obtained from $\beta$ by band surgery along a band as in Figure \ref{figure-branch}.
\begin{figure}[ht]
\includegraphics*[height=3cm]{Branch-pt-etc}
\caption{Motion pictures of knitted surfaces presented by (1) $e \leftrightarrow \sigma_i$, and (2) $e \to \tau_i$, where we omit the $j$th strings for $j \neq i, i+1$. This figure is \cite[Figure 2]{NY}.
%file: Branch-pt-etc \label: figure-branch
}
\label{figure-branch}
\end{figure}
In this case, we define $S$ by the motion picture as given in Figure \ref{figure-branch}; we define $S$ by slices as follow: for $e \to s_i$ $(s_i \in \{\sigma^\epsilon, \tau_i\})$, we denote by $B$ a band attaching to $e$ so that the result of the band surgery along $B$ is $s_i$. Then we define $S$ by
\[
S_{[t]}=\begin{cases}
e & (0 \leq t <\frac{1}{2})\\
e \cup B & (t=\frac{1}{2})\\
s_i & (\frac{1}{2}<t\leq 1),
\end{cases}
\]
and for $s_i \to e$, the presented surface $S'$ is given by the mirror image of $S$ with respect to $D^2 \times \{1/2\}$, that is, the surface $S'$ given by the slice $S'_{[t]} =S_{[1-t]}$ for each $t \in [0,1]$.
\noindent
(Case 3) A transformation $\tau_i \leftrightarrow \tau_i\tau_i$.
In this case, we define $S$ by the motion picture as given in Figure \ref{figure-tau}; we define $S$ by slices as follow: we define $S$ for $\tau_i \to \tau_i\tau_i$; the other case is given by the same method as in (Case 2). We denote by $D$ a disk whose boundary is the simple closed curve in $\tau_i \tau_i$. Further, we assume that $\tau_i=\tau_i \tau_i \backslash \partial D$. Then we define $S$ by
\[
S_{[t]}=\begin{cases}
\tau_i & (0 \leq t <\frac{1}{2})\\
\tau_i \cup D & (t=\frac{1}{2})\\
\tau_i \tau_i & (\frac{1}{2}<t\leq 1).
\end{cases}
\]
\begin{figure}[ht]
\includegraphics*[height=3cm]{T2=tr}
\caption{Motion picture of the knitted surfaces presented by $\tau_i \leftrightarrow \tau_i\tau_i$, where we omit the $j$th strings for $j \neq i, i+1$. This figure is similar to \cite[Figure 3]{NY}.
%file: T2=tr \label: figure-tau
}\label{figure-tau}
\end{figure}
For a transformation $\beta \to \beta'$ such that $\beta'$ is obtained from $\beta$ by a finite number of mutually disjoint transformations as given in (Case 1)--(Case 3), the knitted surface presented by $\beta \to \beta'$ is defined as the surface determined from those of (Case 1)--(Case 3).
We remark that the construction is called the \lq\lq vertical product'' \cite[Section 15.1]{Kamada02}.
\subsection{Knitted surfaces}\label{subsec-knitted-surface}
Knitted surfaces are defined by using \lq\lq pairings'' and \lq\lq knit structure'' \cite{NY}. Two knitted surfaces are said to be equivalent if they are related by an \lq\lq isotopy of knitted surfaces''. Every knitted surface is equivalent to a knitted surface in a normal form.
In this paper, we give a knitted surface as a knitted surface derived from a normal form, by using deformations of knits.
\begin{definition}\label{Definition: knitted surface in normal form}
Let $S$ be a surface properly embedded in $D^2 \times B^2$.
Then, $S$ is called a {\it knitted surface} of degree $n$ if there exists a partition $0=t_0<t_1< \ldots <t_{m}=1$ for some $m$ satisfying the following conditions, where $\beta_j=S_{[t_j]}$ $(j=0,\ldots, m)$.
\begin{enumerate}
\item Each slice $\beta_j$ is an $n$-knit $(j=0, \ldots, m)$.
\item The surface $S \cap (D^2 \times I_3 \times [t_{j-1}, t_j])$ is the knitted surface presented by $\beta_{j-1} \to \beta_j$ for $j \in \{0, \dots, m\}$ such that the knit $\beta_{j}$ is obtained from $\beta_{j-1}$ by a finite number of mutually disjoint transformations associated with the relations of the $n$-knit monoid $D_n$ as follows $(j=1, \ldots, m)$, where $\beta \leftrightarrow \beta'$ denotes the sequences $\beta \to \beta'$ and $\beta' \to \beta$:
\begin{itemize}
\item \ $e \leftrightarrow \sigma_i^\epsilon$,
\item \ $e \leftrightarrow \tau_i$,
\item $\tau_i \tau_i \leftrightarrow \tau_i, $
\item $\tau_i \leftrightarrow \sigma_i^\epsilon \tau_i,$
\ $\tau_i \leftrightarrow \tau_i \sigma_i^\epsilon,$
\item $\sigma_i^\epsilon \sigma_i^{-\epsilon} \leftrightarrow e$,
\item
$\sigma_i^\epsilon \sigma_j^\delta \leftrightarrow \sigma_j^\delta \sigma_i^\epsilon \quad (|i-j|>1), $
\item
$\sigma_i^\epsilon \tau_j \leftrightarrow \tau_j \sigma_i^\epsilon \quad (|i-j|>1)$, \item $\tau_i \tau_j \leftrightarrow \tau_j \tau_i \quad (|i-j|>1)$, \
\item $\sigma_i^\epsilon \sigma_j^\epsilon \sigma_i^\epsilon \leftrightarrow \sigma_j^\epsilon \sigma_i^\epsilon \sigma_j^\epsilon,\ \sigma_i^\epsilon \sigma_j^\epsilon \sigma_i^{-\epsilon} \leftrightarrow \sigma_j^{-\epsilon} \sigma_i^{\epsilon} \sigma_j^\epsilon \quad (|i-j|=1),$
\item $\sigma_i^\epsilon \sigma_j^\epsilon \tau_i \leftrightarrow \tau_j \sigma_i^\epsilon \sigma_j^\epsilon \quad (|i-j|=1)$,
%\end{align*}
\end{itemize}
where $\epsilon \in \{+1, -1\}$, $i,j \in \{1,\ldots, n-1\}$.
\end{enumerate}
We denote by $Q_n$ the starting point set of $n$-knits, that is a fixed set of $n$ interior points of $D^2$ such that $\beta \cap (D^2 \times \partial I_3)=Q_n \times \partial I_3$ for any $n$-knit $\beta$.
We call the knitted surface $Q_n \times B^2$ the {\it trivial} knitted surface.
A knitted surface of degree $n$ is called a {\it braided surface} of degree $n$ if the transformations $\beta_0 \to \cdots \to \beta_m$ are restricted to the relations of the $n$-braid group $B_n$; that is, the relations do not contain $\tau_i$ $(i=1, \ldots, n-1)$.
A knitted surface $S$ of degree $n$ is called a \textit{2-dimensional knit} if $\partial S=Q_n \times \partial B^2$.
A 2-dimensional knit is called a \textit{2-dimensional braid} if it is a braided surface.
\end{definition}
\section{The closure of 2-dimensional knits}\label{sec3}
In this section, we show Theorem \ref{Theorem: Alexander theorem for surface-links}.
\subsection{Closure of knits and 2-dimensional knits}
Let $m=1,2$. Let $B^m_i$ be copies of $m$-disks $B^m$. Let $S$ be an $n$-knit in $D^2 \times B_1^2$ or a 2-dimensional knit in $D^2 \times B_2^2$, and let $S^m_0 = Q_n \times B^m_2 \subset D^2 \times B_2^m$ be the trivial knits/2-dimensional knits. Then, $\mathrm{cl}(S):= S \cup S_0$ is a closed 1-manifold or a closed surface in $D^2 \times S^m$, where $S^m = B_1^m \cup_\partial B_2^m$.
Let $F_0^m$ be a circle (respectively a 2-sphere) standardly embedded in $\mathbb{R}^{m+2}$ when $m=1$ (respectively $m=2$), and let $N_m$ be a regular neighborhood of $F_0^m$.
We identify $N_m$ with $D^2 \times S^m$ $(m=1,2)$.
The \textit{closure} of $S$ is a link or a surface-link in $\mathbb{R}^{m+2}$ which is given by $\mathrm{cl}(S) \subset N_m \subset \mathbb{R}^{m+2}$.
As an analogue of Alexander's theorem for oriented links, it is known that:
\begin{theorem}[\cite{Kamada94-2}]\label{Theorem: Alexander theorem for orientable surface-links}
Every orientable surface-link is ambiently isotopic to the closure of some 2-dimensional braid.
\end{theorem}
\subsection{The plat closure of braids}
For a set $w=\{w^1, \ldots, w^l\}$, we denote by $\cup w$ the union $w^1 \cup \cdots \cup w^l$.
A \textit{wicket} is a semi-circle properly embedded in $D^2 \times I$ which meets $D^2 \times \{0\}$ orthogonally.
A \textit{configuration of $n$ wickets} is a set of mutually disjoint $n$ wickets.
$\mathcal{W}_n$ denotes the space consisting of configurations of $n$ wickets.
Let $Q_{2n}$ be a set of $2n$ interior points $\{q_1 ,\dots, q_{2n}\}$ of $D^2$ such that $q_1, \dots, q_{2n}$ lie in $\{1/2\} \times I \subset I \times I=D^2$ in this order. Let $w_0$ denotes the configuration of $n$ wickets such that $\partial w_0 = Q_{2n}$ and $q_i$ and $q_{i+1}$ are boundaries of the same component of $w_0$ for each $i \in \{1, \dots, n\}$.
We denote by $w_0^*$ the mirror image of $w_0$ with respect to $D^2 \times \{1/2\}$; remark that $\cup w_0$ and $\cup w_0^*$ are disjoint.
The union $ (\cup w_0) \cup (\cup w_0^*)$ represents the $2n$-knit $\tau := \tau_1 \tau_3 \dots \tau_{2n-1}$.
For a $2n$-braid $\beta$, the \textit{plat closure} of $\beta$, denoted by $\widetilde{\beta}$, is a link in $D^2\times [-1,2] \subset \mathbb{R}^3$ obtained by taking the union of $w_0 \subset D^2\times [-1,0]$, $\beta \subset D^2\times [0,1]$, and $w_0^* \subset D^2\times [1,2]$.
See Figure~\ref{20240707-1}.
The plat closure $\widetilde{\beta}$ can be regarded as the closure of the knit $\tau\beta$, that is, $\widetilde{\beta}=\mathrm{cl}(\tau \beta)$.
\begin{figure}[h]
\centering
\includegraphics[width=0.7\hsize]{20240707-1.pdf}
\caption{The plat closure of a braid. }%file: 20240707-1}
\label{20240707-1}
\end{figure}
\subsection{The plat closure of braided surfaces}\label{sec:pcl-braided-surface}
We recall the definition of the plat closure of a braided surface \cite{Yasuda21}. To take the plat closure, a braided surface is required to be \lq\lq adequate''.
For a loop $f: (I, \partial I) \to (\mathcal{W}_n, w_0)$, we define the $2n$-braid $\beta_f$ by $\beta_f = \bigcup_{t \in I} \partial f(t) \times \{t\} \subset D^2\times \{0\} \times I$.
A braid $\beta$ is called \textit{adequate} if there exists a loop $f: (I, \partial I) \to (\mathcal{W}_n, w_0)$ such that $\beta = \beta_f$.
The set of adequate $2n$-braids forms a group called the {\it Hilden subgroup} of the $2n$-braid group $B_{2n}$, denoted by $K_{2n}$, and $K_{2n}$ is generated by $\sigma_{2i-1}$, $\sigma_{2j}\sigma_{2j-1}\sigma_{2j+1}\sigma_{2j}$, and $\sigma_{2j}\sigma_{2j-1}\sigma^{-1}_{2j+1}\sigma^{-1}_{2j}$ for $i \in \{1, \dots, m\}$ and $j \in \{ 1, \dots, m-1\}$ \cite{Brendle-Hatcher2008} (see Figure \ref{20240707-2} for the case $n=2$).
\begin{figure}[h]
\centering
\includegraphics[width = 0.6\hsize]{20240707-2-2.pdf}
\caption{Adequate braids which generate the Hilden subgroup $K_4$: from the left, $\sigma_{1}$, $\sigma_3$, $\sigma_{2}\sigma_{1}\sigma_{3}\sigma_{2}$, and $\sigma_{2}\sigma_{1}\sigma^{-1}_{3}\sigma^{-1}_{2}$. This figure is \cite[Figure 16]{NY}.
}
\label{20240707-2}
\end{figure}
Let $\beta$ be an adequate $2n$-braid in $D^2 \times I$.
There exists a unique loop $f: (I, \partial I) \to (\mathcal{W}_n, w_0)$ satisfying $\beta_f = \beta$. Let $S^1 = [0,1]/\sim$ with $0 \sim 1$. Let $N=I \times S^1$, an annulus.
We define the surface $A_\beta$ in $D^2 \times N$ by
\begin{equation}\label{eq1218}
A_\beta = \bigcup_{t \in [0,1]} ((\cup f(t))\times \{t\}) \subset (D^2 \times I) \times S^1=D^2 \times N.
\end{equation}
A braided surface $S$ is called \textit{adequate} if $\partial S$ is the closure of an adequate braid. We remark that an adequate braided surface is of degree $2n$ for some $n$, and every 2-dimensional braid of degree $2n$ is adequate.
For an adequate braided surface $S$ in $D^2 \times B^2$, we denote by $\beta$ the adequate braid whose closure is $\partial S$. We take the surface $A_\beta$ given by (\ref{eq1218}).
Recall that $B^2=[0,1] \times [0,1]$. We identify $N$ with $[-1,2] \times [-1,2] \backslash B^2$, and we identify $\partial B^2$ with $\{0\} \times S^1\subset I \times S^1=N$, and we embed $D^2 \times N \subset \mathbb{R}^4$ by $D^2 \subset \mathbb{R}^2$ and $N \subset \mathbb{R}^2$.
We remark that $\partial S = \partial A_\beta = S \cap A_\beta$.
Then the \textit{plat closure} of $S$ is a surface-link in $\mathbb{R}^4$ defined by the union of $S$ and $A_\beta$, denoted by $\widetilde{S}$.
%A surface-link $F$ is said to be in a \textit{plat form} if $F$ is the plat closure of some adequate braided surface.
\begin{theorem}[\cite{Yasuda21}]\label{TheoremY}
Every surface-link is ambiently isotopic to the plat closure of some adequate braided surface.
\end{theorem}
\subsection{Proof of Theorem~\ref{Theorem: Alexander theorem for surface-links}}
Let $F$ be a surface-link. By Theorem \ref{TheoremY}, we take an adequate braided surface $S$ of degree $2n$ such that the plat closure $\widetilde{S}$ is ambiently isotopic to $F$.
\noindent
(Case 1)
First, we consider the case that $S$ is a 2-dimensional braid, that is, $\widetilde{S}$ has the motion picture as in Figure~\ref{20240629-1}. Put $\tau=\tau_1 \tau_3 \cdots \tau_{2n-1}$.
%
\begin{figure}[h]
\centering
\includegraphics{20240629-1.pdf}
\caption{Motion picture of the plat closure of $S$. }%file: 20240629-1}
\label{20240629-1}
\end{figure}
%
We take a surface-link $F_1$ given by
\[
(F_1)_{[t]}=\begin{cases}
\mathcal{D} & (t=-1)\\
\mathrm{cl}(\tau) & (-1 < t<0)\\
\mathrm{cl}(\tau S_{[t]}) & (0 \leq t\leq 1)\\
\mathrm{cl}(\tau) & (1<t<2)\\
\mathcal{D} & (t=2),
\end{cases}
\]
where $\mathcal{D}$ is a union of $n$ copies of disks whose boundaries are $\mathrm{cl}(\tau)$; see Figure \ref{20240629-2}.
The surface $F_1$ is ambiently isotopic to
the plat closure $\widetilde{S}$.
\begin{figure}[h]
\centering
\includegraphics{20240629-2-2r.pdf}
\caption{Motion picture of $F_1$. }%ie: 20240629-2-2r}
\label{20240629-2}
\end{figure}
Let $\Delta^-$ be the surface properly embedded in $\mathbb{R}^3 \times (-\infty, 0]$ given by
\[
(\Delta^-)_{[t]}=\begin{cases}
\mathcal{D'} & (t=-2)\\
\mathrm{cl}(e) & (-2 < t<-1)\\
\mathrm{cl}(e) \cup (\cup\mathcal{B}) & (t=-1)\\
\mathrm{cl}(\tau) & (-1<t\leq 0),
\end{cases}
\]
where $\mathcal{D}'$ is a union of $2n$ disks whose boundaries are $\mathrm{cl}(e)$ and $\mathcal{B}$ is a band set attaching to $e$ such that $h(e; \mathcal{B})=\tau$, and $\cup \mathcal{B}$ is the union of bands of $\mathcal{B}$; see Figure \ref{20240629-3}. The surface $\Delta^-$ consists of $n$ copies of disks.
We denote by $F^-_1$ the surface $F_1 \cap (\mathbb{R}^3 \times (-\infty, 0])$.
Then both $\Delta^-$ and $F_1^-$ consist of $n$ copies of disks with $\Delta^-\cap (\mathbb{R}^3 \times (-\infty,0])=\partial \Delta^-=\partial F_1^-=F_1^- \cap (\mathbb{R}^3 \times (-\infty,0]) \subset \mathbb{R}^3\times \{0\}$. These set of disks are called {\it trivial disk systems}, and it is known \cite{KSS} that they are ambiently isotopic rel $\mathbb{R}^3 \times [0, \infty)$.
\begin{figure}[h]
\centering
\includegraphics{20240629-3r.pdf}
\caption{Motion picture of $\Delta^-$. }%file: 20240629-3r}
\label{20240629-3}
\end{figure}
Similarly, we take the surface $\Delta^+$ properly embedded in $\mathbb{R}^3 \times [1, \infty)$ given by the mirror image of $\Delta^-$ with respect to $\mathbb{R}^3 \times \{1/2\}$, and we denote by $F_1^+$ the surface $F_1 \cap (\mathbb{R}^3 \times [1, \infty))$.
Then $\Delta^+$ and $F_1^+$ are ambiently isotopic rel $\mathbb{R}^3 \times (-\infty, 1]$.
Let $F_2$ be a surface obtained from $F_1$ by replacing $F_1^+$ and $F_1^-$ with $\Delta^+$ and $\Delta^-$, respectively; that is,
\[
F_2=\Delta^- \cup (F_1 \cap (\mathbb{R}^3 \times [0,1])) \cup \Delta^+. \]
By construction, $F_2$ is ambiently isotopic to $F_1$.
Let $D^2 \times I$ be the cylinder associated with the knits in slices.
Put $S_1=F_2 \cap (D^2 \times I \times [-3/2, 5/2])$.
Then $S_1$ is a 2-dimensional knit given by
\[
(S_1)_{[t]}=\begin{cases}
e & (-3/2 \leq t \leq -1)\\
e \cup (\cup \mathcal{B}) & (t=-1)\\
\tau & (-1< t <0)\\
\tau \cdot S_{[t]} & (0 \leq t \leq 1)\\
\tau & (1 < t< 2)\\
\tau \cup (\cup \mathcal{B}) & (t=2)\\
e & (2 <t\leq 5/2).
\end{cases}
\]
Further, we see that the closure of $S_1$ is $F_2$. Thus, the closure of the 2-dimensional knit $S_1$ is ambiently isotopic to the given surface-link $F$.
\noindent
(Case 2)
Next, we consider the case when $F$ is ambiently isotopic to the plat closure $\tilde{S}$ of an adequate braided surface $S\subset D^2 \times B^2$ of degree $2n$.
By considering a self-homeomorphism $f$ of $B^2$ and the induced self-homeomorphism of $D^2 \times B^2$ given by $\mathrm{id} \times f$, we assume that $\partial S \cap (D^2 \times I \times \{1\})$ has a presentation of an adequate braid whose closure is $\partial S$. We denote the adequate braid $\partial S \cap (D^2 \times I \times \{1\})$ by $\gamma$.
By the same isotopy used to obtain $F_1$ from $\tilde{S}$ in (Case 1), we have a surface-link $F_1$.
Let $S'$ be a knitted surface in $D^2 \times B^2$ associated with the adequate braid $\gamma$ given in \cite[Section 6.4.2]{NY}. Then, the surface $S'$ satisfies
\[
S'_{[t]}=\begin{cases}
\tau \cdot \gamma& (t=0)\\
\tau & (t=1)\\
\end{cases}.
\]
Thus, let $S_1$ be the knitted surface given by
\[
(S_1)_{[t]}=\begin{cases}
\tau \cdot S_{[2t]} & (0 \leq t \leq 1/2)\\
S'_{[2t-1]} & (1/2 \leq t \leq 1).
\end{cases}
\]
The
Let $F_2$ be the
surface given by
\[
(F_2)_{[t]}=\begin{cases}
\mathcal{D} & (t=-1)\\
\mathrm{cl}(\tau) & (-1<t<0)\\
\mathrm{cl}((S_1)_{[t]}) & (0 \leq t\leq 1)\\
\mathrm{cl}(\tau) & (1 <t<2))\\
\mathcal{D} & (t=2),
\end{cases}
\]
where $\mathcal{D}$ is a union of $n$ copies of disks whose boundaries are $\mathrm{cl}(\tau)$.
By construction of $S'$, $F_2 \cap (\mathbb{R}^3 \times [1/2, \infty)$ is a trivial disk system (see \cite[Proposition 6.6]{NY}) and $F_2$ is ambiently isotopic to the surface-link $F_1$.
Then, by the same process with Case 1, we obtain a new surface-link $F_3$ from $F_2$ by replacing $F_2^+ = F_2 \cap \mathbb{R}^3 \times [1, \infty)$ and $F_2^- = F_2 \cap \mathbb{R}^3 \times (-\infty, 0]$ with $\Delta^+$ and $\Delta^-$, respectively. The surface-link $F_3$ is ambiently isotopic to $F_2$.
Then, $S_2:=F_3 \cap (D^2 \times I \times [-3/2, 5/2])$ is a 2-dimensional knit and the closure of $S_2$ is the surface-link $F_3$.
Hence, the closure of the 2-dimensional knit $S_2$ is ambiently isotopic to the given surface-link $F$.
\qed
\section{Chart description}\label{sec5}
A knitted surface of degree $n$ is presented by a finite graph on $B^2$ called a \lq\lq chart'' of degree $n$ \cite{NY}, which is a generalization of a chart for braided surfaces \cite{Kamada96, Kamada94, Kamada92}.
Here, we give the definition of charts of degree 2. See \cite{NY} for the definition of degree $n$.
\begin{definition}\label{def5-5}
Let $\Gamma$ be a finite graph in a 2-disk $B^2$.
Then, $\Gamma$ is a {\it knit chart} of degree $2$, or simply a {\it 2-chart}, if it satisfies the following conditions.
\begin{enumerate}
\item
The intersection $\Gamma \cap \partial B^2$ consists of a finite number of endpoints of edges of $\Gamma$ meeting $\partial B^2$ orthogonally. Though they are vertices of degree one, we call elements of $\Gamma \cap \partial B^2$ {\it boundary points}, and we call only a vertex in $\mathrm{Int}(B^2)$ {\it a vertex of $\Gamma$}.
\item
Each edge is equipped with the label $1$ and moreover, each edge is either oriented or unoriented. We call an oriented/unoriented edge a {\it $\sigma$-/$\tau$-edge}, respectively.
\item
Each vertex is of degree 1 or 3 as depicted in Figure \ref{BMWchart}.
A vertex of degree 1 connected with a $\sigma$-edge (respectively $\tau$-edge) is depicted by a small black (respectively red) disk, called a {\it black $\sigma$-vertex} (respectively {\it black $\tau$-vertex}). A vertex of degree $3$ is called a {\it mixed trivalent vertex} (respectively a {\it trivalent $\tau$-vertex}) if it is connected with a $\sigma$-edge and a pair of $\tau$-edges (respectively three $\tau$-edges).
\end{enumerate}
\end{definition}
\begin{figure}[ht]
\includegraphics*[height=4cm]{BMWchart-deg2}
\caption{Vertices of a 2-chart.}%
%file: BMWchart-deg2. label:BMWchart}
\label{BMWchart}
\end{figure}
A knitted surface has a chart description.
From a knitted surface of degree $n$, we obtain a chart of degree $n$, and
from a given chart $\Gamma$,
we can construct a knitted surface whose chart is $\Gamma$ (\cite{NY}).
In the rest of this paper, we focus on knitted surfaces of degree $2$ and their $2$-charts. Thus we only
review the knitted surface $S(\Gamma)$ of degree 2 constructed from a $2$-chart $\Gamma$ as follows.
We decompose $B^2$ into a union of a finite number of 2-disks such that for each disk $E$, by modification, $\Gamma \cap E$ is either an empty graph (called an {\it empty chart}) or one of the $2$-charts as in Figures \ref{BMWchart} and \ref{chart-edge2}, where the dotted square in Figure \ref{chart-edge2} denotes $\partial E$.
We identify $E$ with $I_3 \times I_4$, where the direction of $I_3$ is the vertical direction from up to down and the direction of $I_4$ is the horizontal direction from left to right in the figures.
When $\Gamma \cap E$ is an empty chart, we define $S(\Gamma \cap E)$ by the trivial knitted surface $Q_2 \times E$.
When $\Gamma \cap E$ is one of the $2$-charts as in Figure \ref{chart-edge2}, we define $S(\Gamma \cap E)$ by the knitted surface $\sigma_1 \times I_4$ for Figure \ref{chart-edge2}(1), $\sigma_1^{-1} \times I_4$ for Figure \ref{chart-edge2}(2), and $\tau_1 \times I_4$ for Figure \ref{chart-edge2}(3).
When $\Gamma \cap E$ is one of the $2$-charts as in Figure \ref{BMWchart}, we define $S(\Gamma \cap E)$ by the knitted surface presented by $s \to e$ for the $2$-chart containing one black $\sigma$/$\tau$-vertex for appropriate $s \in \{\sigma_1, \sigma_1^{-1}, \tau_1\}$, and the knitted surface presented by $\tau_1 \to \sigma_1 \tau_1$, $\tau_1 \to \sigma_1^{-1}\tau_1$, or $\tau_1\tau_1 \to \tau_1$ for the $2$-chart containing a trivalent vertex.
We construct the knitted surface presented by $\Gamma$ as the union of the knitted surfaces presented by $\Gamma \cap E$.
\begin{figure}[ht]
\includegraphics*[height=3cm]{chart-edge2}
\caption{The $2$-chart $\Gamma \cap E$ consisting of one edge.}%
%file: label:chart-edge2}
\label{chart-edge2}
\end{figure}
\subsection{Chart moves}\label{sec7-1}
In this section, we give the definition of chart moves for $2$-charts, that are local moves of charts. This includes C-moves of charts of braided surfaces \cite{Kamada02, Kamada96, Kamada92}.
\begin{definition}\label{def6-2}
Let $\Gamma$ and $\Gamma'$ be $2$-charts in $B^2$.
Then, a {\it knit chart move}, or a {\it C-move}, is a local move in a 2-disk $E \subset B^2$ such that $\Gamma$ (respectively $\Gamma'$) is changed to $\Gamma'$ (respectively $\Gamma$), satisfying the following.
\begin{enumerate}
\item
The boundary $\partial E$ does not contain any vertices of
$\Gamma$ and $\Gamma'$ except for boundary points, and $\Gamma \cap \partial E$ (respectively $\Gamma' \cap \partial E$) consists of transverse intersection points of edges of $\Gamma$ (respectively $\Gamma'$) and $\partial E$.
\item
The charts are identical in the complement of $E$:
$\Gamma \cap (B^2\backslash E)=\Gamma' \cap (B^2\backslash E)$.
\item
The charts in $E$, $\Gamma \cap E$ and $\Gamma' \cap N$,
are as in Figure \ref{20241211-1a} and the orientation-reversed figures of Figure \ref{20241211-1a}.
\end{enumerate}
We say that $\Gamma$ and $\Gamma'$ are {\it C-move equivalent} if they are related by a finite sequence of C-moves and ambient isotopies of $B^2$.
\end{definition}
\begin{figure}[ht]
\includegraphics*[height=8cm]{20241211-1a}
\caption{C-moves for 2-charts, where we omit labels, which are all $1$. C-moves include the orientation-reversed version. }%file: label:20241211-1a}
\label{20241211-1a}
\end{figure}
Let $\Gamma$ and $\Gamma'$ be $2$-charts and let $S$ and $S'$ be the knitted surfaces presented by $\Gamma$ and $\Gamma'$, respectively.
Theorem 7.5 in \cite{NY} implies that $S$ and $S'$ are ambiently isotopic if $\Gamma$ and $\Gamma'$ are $C$-move equivalent. Thus, the plat closures of $S$ and $S'$ are ambiently isotopic if $\Gamma$ and $\Gamma'$ are $C$-move equivalent.
\section{Plat closure of knitted surfaces of degree $2$}\label{sec8}
In this section, we consider knitted surfaces of degree $2$.
%
We give a normal form of a $2$-chart (Definition \ref{normal-form-chart} and Proposition \ref{thm6-3}).
We show that any trivial surface-link is ambiently isotopic to the plat closure of knitted surfaces of degree 2 (Corollary \ref{cor5-5}).
We show that the plat closure of any knitted surface of degree 2 is a trivial surface-link (Theorem \ref{thm6-7}), and we show Theorem \ref{thm20241213} and Corollary \ref{thm20241217}.
\\
\subsection{Knitted surfaces of degree $2$}
A $\sigma$-/$\tau$-edge of a chart is called a {\it free $\sigma$-/$\tau$-edge} if its endpoints are black $\sigma$-/$\tau$-vertices. A $\sigma$-edge of a chart
is called a {\it half $\sigma$-edge} if its endpoints consists of a black $\sigma$-vertex and either a boundary point or a mixed
trivalent vertex. A $\tau$-edge of a chart is called a {\it half $\tau$-edge} if its endpoints consists of a black $\tau$-vertex and either a boundary point or a trivalent $\tau$--vertex.
For a chart $\Gamma$, the {\it $\tau$-chart} of $\Gamma$, denoted by $\Gamma_\tau$, is defined as the subgraph of $\Gamma$ consisting of $\tau$-edges and the connecting vertices. The $\tau$-chart $\Gamma_\tau$ may contain $2$-valent vertices which are obtained from mixed
trivalent vertices. Then, we merge the two edges connected by each 2-valent vertex after removing the $2$-valent
vertex. Thus, $\Gamma_\tau$ consists of intervals, circles, and graphs consisting of $1$-valent vertices and $3$-valent
vertices. We call a connected component of $\Gamma_\tau$ {\it a $\tau$-component}. When $\Gamma_\tau$ is contained in $B^2$, a {\it region} of $\Gamma_\tau$ is the closure of a connected component of $B^2\backslash \Gamma_\tau$.
\begin{definition}\label{normal-form-chart}
Let $\Gamma$ be a $2$-chart, let $R_1, \ldots,R_m$ be regions of the $\tau$-chart $\Gamma_\tau$, and put $\Gamma^j := \Gamma \cap R_j$, the subchart of $\Gamma$ $(j=1, \ldots,m)$. Then, we say that $\Gamma$ is in a {\it normal form} if for each $j = 1,\ldots,m$, there exists a $\tau$-component $G_j$ of $\Gamma^j$ such that every $\sigma$-edge
of $\Gamma_j$ is either a free $\sigma$-edge or connected with $G_j$, and half $\sigma$-edges connected with $G_j$ have coherent orientations. We call $G_j$ the {\it initial $\tau$-component} of $\Gamma_j$.
\end{definition}
See Figure \ref{20240330-1} for an example.
\begin{figure}[ht]
\includegraphics*[height=5cm]{20240330-1}
\caption{An example of a $2$-chart in a normal form, where we omit labels, which are all $1$. The $\tau$-edge in the form of the circle in the center is the initial $\tau$-component associated with the region whose closure contains the larger part of $\partial B^2$.}
%file: 20240330-1. label:20240330-1}
\label{20240330-1}
\end{figure}
\begin{proposition}\label{thm6-3}
Any 2-chart is C-move equivalent to a $2$-chart in a normal form given in Definition \ref{normal-form-chart}.
\end{proposition}
\begin{proof}
Let $\Gamma$ be a $2$-chart. The labels of edges are all one. We denote by the same notation $\Gamma$ the charts obtained from $\Gamma$ by C-moves.
From now on, we consider each subchart $\Gamma^j=\Gamma \cap R_j$ of $\Gamma$ as given in Definition \ref{normal-form-chart}. We fix a $\tau$-component $G_j$ of $\Gamma^j$.
By the definition of $2$-charts, $\sigma$-edges are classified by combinations of their endpoints; a $\sigma$-edge is either
a simple closed curve (called a circle component), a free $\sigma$-edge, a half $\sigma$-edge, a $\sigma$-edge connecting with $\partial B^2$/$\tau$-components.
Let $\mathbf{e}$ be a $\sigma$-edge of $\Gamma^j$ disjoint with $G_j$. Since $R_j$ is connected, we take a simple path from a point of $\mathbf{e}$ to $G_j$. Applying (A1), (A2) and (D1)-moves, each circle component of $\sigma$-edges is deformed into
a $\sigma$-edge connecting $\tau$-components.
By applying (A2) and (D1)-moves, we deform half $\sigma$-edges and $\sigma$-edges connecting
$\tau$-components into $\sigma$-edges connected with $G_j$, see Figure \ref{20241222}; the detail is left to the reader as exercise.
Further, by (A2)-moves, we deform half $\sigma$-edges connected with $G_j$ to a set of free $\sigma$-edges and half $\sigma$-edges with coherent orientations.
Thus we arrange $\Gamma$ into a normal form.
\end{proof}
\begin{figure}[ht]
\includegraphics*[height=2.5cm]{20241222-normal_form}
\caption{Deforming $\sigma$-edges in $\Gamma^j$ to $\sigma$-edges connecting with the initial $\tau$-component $G_j$. We omit the labels of the $2$-charts. }
%file: 20240330-1. label:20240330-1}
\label{20241222}
\end{figure}
\begin{remark}
We can deform a 2-chart $\Gamma$ to a more regulated form, also denoted by $\Gamma$, satisfying that (1) every black $\tau$-vertex is an endpoint of a free $\tau$-edge, (2) free $\sigma$-/$\tau$-edges are in a $2$-disk which contains no other parts of $\Gamma$, (3) $\sigma$-edges connecting a $\tau$-component and the initial $\tau$-component have coherent orientations, and (4) there are no $\sigma$-edges both of whose endpoints are connected with the initial $\tau$-component.
\end{remark}
\subsection{The plat closure of knitted surfaces}\label{sec4-3-pcl}
Let $S$ be a knitted surface of degree $2$. We define the {\it plat closure} of $S$ in such a way as follows. If $S$ is a braided
surface, then $S$ is adequate; thus we define the plat closure $\tilde{S}$ of $S$ by that of a braided surface, given by $\tilde{S}=S \cup A_\beta \subset \mathbb{R}^4$, where $A_\beta$ is the surface of (\ref{eq1218}) in Section \ref{sec:pcl-braided-surface}. Further, if $\partial S$ is the closure of a $2$-braid, then we define the plat
closure of $S$ by the same way, given by $\tilde{S}=S \cup A_\beta$. We consider the case that $\partial S$ is the closure of a
$2$-knit having hook pairs. Then, $\partial S$ consists of m simple closed curves, where m is the number of hook pairs in $\partial S$. We define the plat closure of $S$ by the closed surface obtained from $S$ by attaching $m$ copies of $2$-disks trivially along $\partial S$.
For knitted surfaces of degree $n$, we do not give the exact definition in terms of wickets in this paper, but we briefly give the idea of the plat closure, as follows.
We call an $2n$-knit {\it adequate} if it is an element of the submonoid of the $2n$-knit monoid $D_{2n}$ generated by adequate $2n$-braids and $\tau_{2k-1}$ $(k=1, \ldots, n)$.
We call a knitted surface $S$ {\it adequate} if $\partial S$ is the closure of an adequate knit.
For an adequate knitted surface $S$, we construct the {\it plat closure} of $S$ as follows.
Let $\beta$ be the adequate knit whose closure is $\partial S$.
We denote by $\beta'$ the braid obtained from the adequate knit $\beta$ by changing each letter in the form of $\tau_i$ for some $i$ to $\sigma_i$. We denote by $s_1, \ldots, s_m$ the letters of $\beta'$ coming from the letters of $\beta$ in the form of $\tau_i$ $(i=1,\ldots, n-1)$. The $2n$-braid $\beta'$ is adequate. Let $A_{\beta'}$ be the surface associated with $\beta'$ given in (\ref{eq1218}) in Section \ref{sec:pcl-braided-surface}. The part of $A_{\beta'}$ corresponding to $s_j$, that is $A_{\beta'} \cap (D^2 \times I) \times [t_j, t_j']$ for some $0\leq t_j<t_j'\leq 1$ with $\partial S \cap (D^2 \times [t_j, t_j'])=s_j$, is a band with one twist. We delete the band from $A_{\beta'}$ and take the closure for each $s_j$ $(j=1, \ldots, m)$. Then, slightly modifying $A_{\beta'}$, we obtain a surface $A'$ in $D^2 \times N$ with $\partial A'=\partial S=S \cap A'$.
We define the {\it plat closure} of $S$ by the surface-link $S \cup A'$ in $\mathbb{R}^4$.
\begin{remark}\label{remark1223}
The plat closure of a knitted surface $S$ of degree $2$ is a special case of the closure of a $2$-dimensional knit of degree $2$ by the following meaning. Let $\Gamma$ be a $2$-chart of $S$, $y \in \partial B^2$ a point disjoint with $\Gamma$, and let $N$ be a small neighborhood of
$y$ disjoint from $\Gamma$. Let $\tilde{B}^2$ be a $2$-disk such that $B^2$ is contained in the interior of $\tilde{B}^2$. Let ${\Gamma}'$ be the $2$-chart in $\tilde{B}^2$ obtained from $\Gamma$ by adding to $\Gamma$ a union of $\tau$-edges in the form $\partial B^2\backslash (N\cup \Gamma)$. Since ${\Gamma}'$ does not have boundary points,
it represents a $2$-dimensional knit of degree $2$, denoted by $S'$. By applying the argument in the proof of Theorem \ref{Theorem: Alexander theorem for surface-links},
the plat closure of S and the closure of $S'$ is ambiently isotopic as surface-links. In
particular, when $\Gamma$ has no boundary points, then ${\Gamma}'$ is obtained from $\Gamma$ by an addition of a single free $\tau$-edge.
\end{remark}
We can calculate the Euler characteristics of knitted surfaces of degree $2$ by using the information derived from charts. The proof is left to the reader; see \cite[Exercise 9.15]{Kamada02}.
For simplicity, we consider a knitted surface $S \subset D^2 \times S^2$, presented by a $2$-chart in a 2-sphere $S^2$.
\begin{proposition}
Let $F \subset D^2 \times S^2$ be a knitted surface of degree $2$, and let $\Gamma \subset S^2$ be its $2$-chart. Let $R_1, \ldots,R_m$ be regions of the $\tau$-chart $\Gamma_\tau$, and put $\Gamma^j := \Gamma \cap R_j$, the subchart of $\Gamma$ $(j=1, \ldots,m)$. Then $F$ is a split union of $F_1, \ldots, F_m$, where $F_j$ is the plat closure of a knitted surface $S(\Gamma^j)$, $j=1, \ldots, m$, and the Euler characteristic $\chi(F_j)$ is calculated by
\[
\chi(F_j)=2\chi(R_j')-b(\Gamma^j),
\]
where $b(\Gamma^j)$ is the number of black $\sigma$-vertices
of $\Gamma^j$, and $\chi(R_j')$ is
the Euler characteristic of
the closure of $R_j \cap (B^2 \backslash N(\Gamma_\tau))$, where $N(\Gamma_\tau)$ is a regular neighborhood of $\Gamma_\tau$ $(j=1, \ldots, m)$.
\end{proposition}
\subsection{Plat closure of knitted surfaces of degree $2$ presenting trivial surface-links}
A {\it standard 2-sphere} (respectively a {\it standard torus}) in $\mathbb{R}^4$ is the boundary of a 3-ball (respectively an unknotted solid torus) in $\mathbb{R}^3 \times \{0\}$, and a standard positive projective plane (respectively a {\it standard negative projective plane}) is the surface given by Figure \ref{20240330-2}(1) (respectively (2)).
These surfaces are called {\it standard surfaces}.
A surface-knot in $\mathbb{R}^4$ is called {\it trivial} if it is a connected sum of a finite number of standard surfaces in $\mathbb{R}^4$, and a surface-link is {\it trivial} if it is a split union of a finite number of trivial surface-knots.
\begin{figure}[ht]
\includegraphics*[height=3cm]{20240330-2r}
\caption{(1) A standard positive projective plane $P_+$ and (2) a standard negative projective plane $P_-$.}
%file: 20240330-2r. label:20240330-2}
\label{20240330-2}
\end{figure}
When a half $\sigma$-edge has an orientation toward (respectively from) the black $\sigma$-vertex, we call it {\it positive} (respectively {\it negative}).
We denote by $T$ a standard torus, and by $P_+$ (respectively $P_-$) a standard positive (respectively negative) projective plane.
\begin{theorem}[\cite{NY}]\label{prop6-5}
Let $F$ be a trivial surface-knot, which is a connected sum of $g$ copies of $T$ and $m$ copies of $P_+$ and $n$ copies of $P_-$.
Then $F$ is the plat closure of the knitted surface presented by a $2$-chart $\Gamma$ such that $\Gamma$ consists of $\min\{m,n\}$ copies of free $\sigma$-edges, $g$ copies of free $\tau$-edges,
and $|m-n|$ copies of positive (respectively negative) half $\sigma$-edges if $m \geq n$ (respectively $m<n$).
\end{theorem}
\begin{corollary}\label{cor5-5}
Any trivial surface-link is ambiently isotopic to the plat closure of some knitted surface of degree $2$.
\end{corollary}
\begin{proof}
Let $F$ be a trivial surface-link, and let $F_1, \ldots, F_m$ denote connected components of $F$. Let $l_j$ be the segment
$I \times \{j/m\}$ in $B^2 = I \times I$ $ (j= 1,\ldots,m-1)$. Then, $l_1\ldots,l_{m-1}$ divide $B^2$ into $m$ $2$-disks
$E_j=I \times [(j-1)/m, j/m]$,
$j = 1,\ldots,m$. Applying Theorem \ref{prop6-5}, we obtain a $2$-chart $\Gamma_j$ in $E_j$ such that the plat closure of
a knitted surface $S(\Gamma_j)$ is ambiently isotopic to $F_j$. Let $\Gamma$ be the union of $\Gamma_j$ $(j=1, \ldots, m)$ and $\tau$-edges in the form of $\cup_{j=1}^{m-1} l_j$. Then, the plat closure of $S(\Gamma)$ is a split union of the plat closures of $S(\Gamma_j)$, $j=1, \ldots, m$, which is ambiently isotopic to $F$.
\end{proof}
\subsection{Knit plat index and the knit index}
The {\it plat index} \cite{Yasuda21} of a surface-link $F$, denoted by $\mathrm{Plat}(F)$, is the half of the minimum of the degrees of braided surfaces whose plat closures are ambiently isotopic to $F$. We define the {\it knit plat index}, denoted by $\mathrm{knit.Plat}(F)$, as the half of the minimum of the degrees of knitted surfaces whose plat closure is ambiently isotopic to $F$.
By definition, for a surface-link $F$,
\[
\mathrm{knit.Plat}(F)\leq \mathrm{Plat}(F).
\]
It is known \cite{Yasuda21} that the plat index of surface-link is one if and only if it is a trivial 2-sphere or nonorientable trivial surface-knot, and the plat index is two if the surface-link is a trivial orientable surface-knot with positive genus. %
Corollary \ref{cor5-5} implies the following:
\begin{corollary}\label{cor5-7}
Let $F$ be any trivial surface-link. Then, the knit plat index of $F$ is one:
\[
\mathrm{knit.Plat}(F)=1.
\]
\end{corollary}
The {\it braid index} \cite{Kamada02} of an oriented surface-link $F$, denoted by $\mathrm{Braid}(F)$, is the minimum of the degrees of 2-dimensional braids whose closures are ambiently isotopic to $F$. We define the {\it knit index} of a surface-link $F$, denoted by $\mathrm{Knit}(F)$, by the minimum of the degrees of 2-dimensional knits whose closures are ambiently isotopic to $F$. By definition, for an oriented surface-link $F$,
\[\mathrm{Knit}(F) \leq \mathrm{Braid}(F).
\]
By the argument in the proof of Theorem \ref{Theorem: Alexander theorem for surface-links}, we have the following.
\begin{corollary}\label{cor:knit-index}
Let $F$ be any surface-link. Then, the knit index of $F$ is bounded from above by twice the plat index of $F$:
\[
\mathrm{Knit}(F)\leq 2\mathrm{Plat}(F).
\]
\end{corollary}
By Corollary \ref{thm20241217}, we have the following.
\begin{corollary}
Let $F$ be a trivial surface-link.
Then, if $F$ is not a trivial 2-sphere, then,
\[
\mathrm{Knit}(F)=2.
\]
\end{corollary}
\subsection{Remark on 2-charts presenting trivial surface-knots} The trivial surface-knot $P_+ \# P_-$ is presented by a 2-chart other than that given in Theorem \ref{prop6-5}.
\begin{proposition}\label{prop6-7}
The trivial surface-knot $P_+ \# P_-$ is ambiently isotopic to the plat closure of the knitted surfaces of degree $2$ presented by the 2-charts as in Figure \ref{20241212-2}.
\end{proposition}
\begin{figure}[ht]
\includegraphics*[height=3cm]{20241212-2}
\caption{A pair of $2$-charts whose plat closures are ambiently isotopic as surface-knots to $P_+\# P_-$, where we omit labels. The left chart is the chart given in Theorem \ref{prop6-5}.}
%file:20241212-2}
\label{20241212-2}
\end{figure}
\begin{proof}
The trivial surface-knot $P_+\# P_-$ has the motion picture as in Figure \ref{20241212-3}(1). The surface is ambiently isotopic to the surface as in Figure \ref{20241212-3}(2). This surface is ambiently isotopic to the surface as in Figure \ref{20241212-3}(3), which is the surface given by the gray regions of the 2-charts in Figure \ref{20241212-1}; so it is the plat closure of the knitted surface presented by the $2$-chart in the left figure of Figure \ref{20241212-2}.
By Theorem \ref{prop6-5}, $P_+\# P_-$ is ambiently isotopic to the plat closure of the knitted surface presented by the $2$-chart in the right figure of Figure \ref{20241212-2}.
Thus the $2$-charts as in Figure \ref{20241212-2} both present $P_+\# P_-$.
\end{proof}
\begin{figure}[ht]
\includegraphics*[height=5cm]{20241212-3r}
\caption{Motion pictures of $P_+ \# P_-$.}
%file:20241212-1}
\label{20241212-3}
\end{figure}
\begin{figure}[ht]
\includegraphics*[height=3cm]{20241212-1}
\caption{The gray regions of the $2$-charts present the same surface of Figure \ref{20241212-3}(3). }
%file:20241212-1}
\label{20241212-1}
\end{figure}
Let $F$ be a trivial surface-knot, which is a connected sum of $g$ copies of $T$ and $m$ copies of $P_+$ and $n$ copies of $P_-$.
If $m \neq n$, then $F$ is a connected sum of $\min\{m,n\}+g$ copies of $T$ and $|m-n|$ copies of $P_+$ (respectively $P_-$) if $m>n$ (respectively $m<n$).
We can show this by using $2$-charts, applying Theorem \ref{prop6-5}, Proposition \ref{prop6-7} and C-moves. The proof is left to the reader as exercise. In the following, we give the result in terms of $2$-charts.
\begin{corollary}
Let $\Gamma$ be a $2$-chart consisting of $\mu$ copies of free $\sigma$-edges, $\nu$ copies of free $\tau$-edges and $\lambda$ copies of positive (respectively negative) half $\sigma$-edges. If $\lambda \neq 0$, then the plat closure of the knitted surface presented by $\Gamma$ is ambiently isotopic to the plat closure of the knitted surface presented by the $2$-chart $\Gamma'$ consisting of $\mu+\nu$ copies of free $\tau$-edges and $\lambda$ copies of positive (respectively negative) half $\sigma$-edges.
\end{corollary}
\subsection{Proof of Theorem \ref{thm20241213} and Corollary \ref{thm20241217}}
Let $F$ be a surface-link in $\mathbb{R}^4$.
A {\it 1-handle attaching to $F$} is a 3-ball $b=D^2 \times I$ embedded in $\mathbb{R}^4$ such that $b \cap F=(D^2 \times \partial I) \cap F$. The arc $\{0\} \times I \subset D^2 \times I$ is called the {\it core} of the 1-handle. The {\it 1-handle surgery} is the operation which changes $F$ to a new surface-link $F'$ given by
$F'=\mathrm{Cl}(F \backslash (D^2 \times \partial I) \cup (\partial D^2 \times I))$.
A 1-handle $b$ attaching to $F$ is called {\it trivial} if we can move the 1-handle $b$ by ambient isotopy so that $b$ is contained in a 4-ball $U$ in $\mathbb{R}^4$ such that the pair $(U, F \cap U)$ is homeomorphic to the standard $(4,2)$-ball pair $(B^4, B^2)$.
\begin{theorem}\label{thm6-7}
Let $F$ be the plat closure of an arbitrary knitted surface of degree 2. Then, $F$ is a trivial surface-link.
\end{theorem}
\begin{proof}
Let $\Gamma$ be an arbitrary $2$-chart.
It suffices to show the result for the subchart of each region divided by the $\tau$-chart $\Gamma_\tau$ of $\Gamma$.
So we assume that $\Gamma \subset B^2$ is
in a normal form such that $B^2\backslash \Gamma_\tau$ is connected.
We call a $\sigma$-edge which is not a half $\sigma$-edge a {\it non-half $\sigma$-edge}.
Let $l_1, \ldots, l_s$ be segments in $B^2$ such that each $l_j$ coincides with each non-half $\sigma$-edge.
Since we take the plat closure, we regard $\partial B^2$ as the initial $\tau$-component.
Let $\Gamma'$ be the 2-chart obtained from $\Gamma$ by changing each non-half $\sigma$-edge into a $\tau$-edge. Let $F$ (respectively $F'$) be the plat closure of the knitted surface presented by $\Gamma$ (respectively $\Gamma'$).
We see that $F'$ is a split union of the plat closures of knitted surfaces of $2$-charts of regions divided by the $\tau$-chart of $\Gamma'$, and $F$ is obtained from $F'$ by 1-handle surgery along 1-handles $b_1, \ldots, b_s$ such that the projection to $B^2$ of the core of $b_j$ transversely intersects with the segment $l_j$ $(j=1, \ldots, s)$. We remark that since $\Gamma$ is in a normal form, each region is a 2-disk. We consider the case $s=1$, when the number of non-half $\sigma$-edges is one. We put $b:=b_1$, the 1-handle along which $F$ is obtained from $F'$ by 1-handle surgery.
The plat closure $F$ is a surface-knot, and $F'$ is either a surface-knot or a surface-link with two components.
By Theorem \ref{prop6-5}, using (T2)-moves (see Figure \ref{20241211-1a}) if necessary, we see that $F$ and $F'$ are trivial surface-knots/links. If $F'$ is a surface-link with two components, denoted by $F_1'$ and $F_2'$, then, $F$ is a connected sum of $F_1'$ and $F_2'$. If $F'$ is a surface-knot, then, since $F'$ is trivial, the 1-handle $b$ is trivial \cite{Kamada2014}.
Hence, the result of 1-handle surgery, $F$, is a connected sum of $F'$ and either a standard torus $T$ or a connected sum $P_+ \# P_-$ (see Figures \ref{20241212-3}(3) and \ref{20241212-1}). For the general case, since $F'$ is a trivial surface-link by Theorem \ref{prop6-5}, by induction on the number $s$ of non-half $\sigma$-edges, we see that $F$ is obtained from components of $F'$ by taking connected sum with several copies of $T$ and $P_+ \# P_-$. Hence we see that $F$ is a trivial surface-knot. Thus we have the required result.
\end{proof}
\begin{proof}[Proof of Theorem \ref{thm20241213}]
By Corollary \ref{cor5-5} and Theorem~\ref{thm6-7}, we have the required result.
\end{proof}
\begin{proof}[Proof of Corollary \ref{thm20241217}]
By Remark \ref{remark1223},
Corollary \ref{cor5-5} implies that any trivial surface-link is ambiently isotopic to the closure of some 2-dimensional knit.
Let $\Gamma$ be a 2-chart in $B^2$. Theorem \ref{Theorem: Alexander theorem for surface-links} implies that if $\Gamma$ contains at least one $\tau$-edge, then the closure of the 2-dimensional knit presented by $\Gamma$ is a trivial surface-link. If $\Gamma$ does not contain $\tau$-edges, then it $\Gamma$ presents a 2-dimensional braid, the closure of which is an orientable trivial surface-link \cite{Kamada02}.
\end{proof}
\section*{Acknowledgements}
The first author was partially supported by JST FOREST Program, Grant Number JPMJFR202U. The second author was partially supported by JSPS KAKENHI Grant Number 22J20494.
\begin{thebibliography}{0}
\bibitem{Birman-Wenzl}
J. S. Birman, H. Wenzl, {\it Braids, link polynomial and a new algebra}, Trans. Amer. Math. Soc. {\bf 313} (1989) 249--273.
\bibitem{Brendle-Hatcher2008}
T. Brendle, A. Hatcher, {\it Configuration spaces of rings and wickets}, Comment. Math. Helv., {\bf 88} (2008) 131-162.
\bibitem{CKS}
J. S. Carter, S. Kamada, M. Saito,
{\it Surfaces in 4-space}, Encyclopaedia of Mathematical Sciences 142, Low-Dimensional Topology III, Berlin, Springer-Verlag, 2004.
\bibitem{Kamada2014}
S. Kamada, {\it Cords and 1-handles attached to surface-knots}, Bol. Soc. Mat. Mex. {\bf 20} (2014) 595-609.
\bibitem {Kamada02}
S. Kamada, {\it Braid and Knot Theory in Dimension Four}, Math. Surveys and Monographs 95, Amer. Math. Soc., 2002.
\bibitem{Kamada96}
S. Kamada, {\it An observation of surface braids via chart description}, J. Knot Theory Ramifications {\bf 5} (1996), no. 4, 517--529.
\bibitem{Kamada94}
S. Kamada, {\it A characterization of groups of closed orientable surfaces in 4-space}, Topology {\bf 33} (1994) 113--122.
\bibitem{Kamada94-2}
S. Kamada, {\it Alexander’s and Markov’s theorems in dimension four}, Bull. Amer. Math. Soc. (N.S.), {\bf 31} (1994), no. 1, 64–67.
\bibitem{Kamada92}
S. Kamada, {\it Surfaces in $\mathbb{R}^4$ of braid index three are ribbon}, J. Knot Theory Ramifications {\bf 1} (1992), no. 2, 137--160.
\bibitem{Kawauchi}
A. Kawauchi, {\it A Survey of Knot Theory},
Birkh\"{a}user Verlag, Basel, 1996.
\bibitem{KSS}
A. Kawauchi, T. Shibuya, S. Suzuki, {\it Descriptions on surfaces in four-space. I. Normal forms}, Math. Sem. Notes Kobe Univ., {\bf 10} (1982), no. 1, 75–125.
\bibitem{Hilden1975}
H. M. Hilden, {\it Generators for two groups related to the braid group}, Pacific J. Math., {\bf 59} (1975), no. 2, 475–486.
\bibitem{Murakami}
J. Murakami, {\it The Kauffman polynomial of links and representation theory}, Osaka J. Math. {\bf 24} (1987) 745--758.
\bibitem{NY}
I. Nakamura and J. Yasuda, {\it Surfaces in the 4-ball constructed using generators of knits and their graphical description}, arXiv: 2306.10479v2.
\bibitem{Rudolph}
L. Rudolph, {\it Braided surfaces and Seifert ribbons for closed braids}, Comment. Math. Helv. {\bf 58} (1983) 1--37.
\bibitem{Yasuda21}
J. Yasuda, {\it A plat form presentation for surface-links}, arXiv:2105.08634, to appear in Osaka J. Math.
\end{thebibliography}
\end{document}