Bonded Braids: Theory and Applications

• Bonded braids are classical braids enriched with embedded bonds that join strands to model both backbone entanglement and internal molecular connections.
• They feature specialized isotopy moves and generators, distinguishing between topological and rigid settings to accurately capture bond interactions.
• The framework extends classical results like the Alexander and Markov theorems and fosters applications in biomolecular modeling and topological protein analysis.

Bonded braids are classical braids enriched by embedded bond arcs joining points on strands. In the formulation developed for bonded knots and links, a bonded braid is the algebraic counterpart of a bonded link in the same sense that an ordinary braid is the algebraic counterpart of an ordinary link: the braid records both classical crossings and additional bonded connections. The subject is motivated by structures in proteins, RNA, and other molecular systems, where one wants to model backbone entanglement together with internal attachments such as disulfide bridges or other noncovalent connections. Recent work develops geometric definitions, monoid and group structures, closure operations, bonded analogues of the Alexander and Markov theorems, linear representations, and extensions to enhanced bonds and open-ended braidoid formalisms (Diamantis et al., 20 Jul 2025, Cavicchioli et al., 6 Jul 2025).

1. Geometric definition and ambient categories

A bonded knot is described as a pair $(L,\mathbf b)$, where $K$ is an oriented knot embedded in $S^3$ (or $\mathbb R^3$), $\mathbf b=\{b_1,\dots,b_n\}$ is a collection of pairwise disjoint embedded intervals, and each bond has endpoints on the knot, so that

$K\cap b_i=K\partial b_i.$

The endpoints of bonds create trivalent vertices, so bonded knots may also be viewed as edge-colored spatial graphs. Within this framework, a bonded braid is introduced as the braid-theoretic version of the same structure (Cavicchioli et al., 6 Jul 2025, Diamantis et al., 20 Jul 2025).

A bonded braid on $n$ strands is defined as a pair

$(\beta,B),$

where $\beta$ is a classical braid on $n$ strands and $K$0 is a set of disjoint embedded horizontal simple arcs called bonds. Each bond has endpoints called nodes. A bond connecting the $K$1 and $K$2 strands is denoted

$K$3

and when the bond connects consecutive strands $K$4 and $K$5, it is an elementary bond,

$K$6

The braid strands run monotonically downward, but the bonds may thread through strands with specified over/under data. A bond $K$7 can therefore be encoded by a sequence of $K$8's and $K$9's describing whether it passes over or under the strands in between (Diamantis et al., 20 Jul 2025).

The literature distinguishes several ambient categories. Bonded knots are considered in three categories—long, standard, and tight—according to the type of bonds, and in two categories—topological vertex and rigid vertex—according to the allowed isotopy moves. The braid theory is developed first for standard and tight bonded braids, and a parallel rigid theory introduces local rigidity at the vertices (Diamantis et al., 20 Jul 2025, Cavicchioli et al., 6 Jul 2025).

A recurrent technical simplification is the use of isolated bonds. In that convention, each bond can be contracted into a tiny disk without crossings, and any bonded knot diagram can be transformed, by Reidemeister-type moves, into one with isolated bonds. This allows the subsequent braiding constructions to be stated cleanly (Cavicchioli et al., 6 Jul 2025).

2. Isotopy, local moves, and rigid versus topological settings

Bonded braid isotopy extends classical braid isotopy by adding bond-specific moves. The move system is organized into planar isotopy moves for bonds and braid strands, commuting rules for far-apart bonds, interaction rules between bonds and strands, interaction rules between bonds and crossings, and forbidden moves. The geometric point is that bonds are embedded arcs, so they cannot be manipulated as if they were purely symbolic decorations (Diamantis et al., 20 Jul 2025).

Several commuting phenomena are explicitly allowed. If two bonds are sufficiently separated, they commute. If one bond is a uniform over/under bond, it can commute past an inner bond. If the crossing sequence of one bond matches a subsequence of another, they commute as well. The move set also includes braided vertex slide moves for the interaction of a bond with a braid strand, and moves such as the bonded flype and bonded $S^3$0 move for interaction with crossings. At the same time, some configurations are forbidden; this marks a sharp distinction from tied-link formalisms in which ties are allowed to move more freely (Diamantis et al., 20 Jul 2025).

The distinction between topological and rigid vertices is fundamental. In the topological setting, bonds may twist locally at the vertices, and the relevant diagrammatic move set includes the bonded Reidemeister moves $S^3$1. In the rigid setting, vertex twisting is disallowed; the move $S^3$2 is replaced by rigid versions $S^3$3 or $S^3$4, and the isotopy relation is generated by Reidemeister moves I–IV together with $S^3$5. Precisely, two rigid-vertex bonded knots are rigid-vertex isotopic if their diagrams are related by a finite sequence of Reidemeister moves I–IV and $S^3$6 (Cavicchioli et al., 6 Jul 2025).

This separation of move sets has direct algebraic consequences. The topological theory requires bond generators, whereas the rigid theory requires additional generators encoding local rigidity defects. A plausible implication is that the rigid and topological braid theories should be viewed not as minor variants but as distinct algebraic envelopes of different isotopy relations.

3. Monoids, groups, generators, and notation

The algebraic structures used for bonded braids are not uniform across the recent literature, and the notation is potentially confusing. One paper denotes the bonded braid monoid by $S^3$7, while another denotes the topological bonded braid monoid by $S^3$8 and reserves $S^3$9 for the corresponding group obtained after adjoining bond inverses. Because the same symbol $\mathbb R^3$0 is also used for blocked-braid groups in an unrelated quotient construction, keeping the source-dependent notation explicit is essential (Diamantis et al., 20 Jul 2025, Cavicchioli et al., 6 Jul 2025, Maglia et al., 2013).

Source	Notation	Role
(Diamantis et al., 20 Jul 2025)	$\mathbb R^3$1	bonded braid monoid
(Cavicchioli et al., 6 Jul 2025)	$\mathbb R^3$2, $\mathbb R^3$3	topological and rigid bonded braid monoids
(Cavicchioli et al., 6 Jul 2025)	$\mathbb R^3$4, $\mathbb R^3$5	topological and rigid bonded braid groups
(Maglia et al., 2013)	$\mathbb R^3$6	blocked-braid groups

In the topological case of (Cavicchioli et al., 6 Jul 2025), the bonded braid monoid $\mathbb R^3$7 on $\mathbb R^3$8 strands is generated by the classical braid generators and inverses

$\mathbb R^3$9

together with bond generators

$\mathbf b=\{b_1,\dots,b_n\}$0

The defining relations are the braid relations

$\mathbf b=\{b_1,\dots,b_n\}$1

the bond relation

$\mathbf b=\{b_1,\dots,b_n\}$2

and the mixed relations

$\mathbf b=\{b_1,\dots,b_n\}$3

$\mathbf b=\{b_1,\dots,b_n\}$4

$\mathbf b=\{b_1,\dots,b_n\}$5

Restricting to $\mathbf b=\{b_1,\dots,b_n\}$6 and the braid relations recovers the usual Artin braid group $\mathbf b=\{b_1,\dots,b_n\}$7, so $\mathbf b=\{b_1,\dots,b_n\}$8 (Cavicchioli et al., 6 Jul 2025).

In the rigid setting, the rigid bonded braid monoid $\mathbf b=\{b_1,\dots,b_n\}$9 adds kink generators $K\cap b_i=K\partial b_i.$0. Besides the braid relations, it has bond and kink commuting relations

$K\cap b_i=K\partial b_i.$1

and mixed relations with both bonds and kinks: $K\cap b_i=K\partial b_i.$2

$K\cap b_i=K\partial b_i.$3

The inclusions

$K\cap b_i=K\partial b_i.$4

are stated explicitly (Cavicchioli et al., 6 Jul 2025).

The presentation in (Diamantis et al., 20 Jul 2025) begins with generators

$K\cap b_i=K\partial b_i.$5

and then reduces to a tight bonded braid monoid presentation using only

$K\cap b_i=K\partial b_i.$6

subject to

$K\cap b_i=K\partial b_i.$7

$K\cap b_i=K\partial b_i.$8

$K\cap b_i=K\partial b_i.$9

$n$0

$n$1

$n$2

$n$3

That paper further reduces the presentation to one using only the single bond generator $n$4, with a key derived relation

$n$5

It also defines an enhanced bonded braid group $n$6 by adjoining inverses $n$7, interpreted as a second bond type inverse to $n$8, with

$n$9

By contrast, in (Cavicchioli et al., 6 Jul 2025) the topological bonded braid group $(\beta,B),$0 is obtained from $(\beta,B),$1 by adding inverses of the bond generators, and $(\beta,B),$2 is interpreted as an anti-bond that annihilates a bond (Diamantis et al., 20 Jul 2025, Cavicchioli et al., 6 Jul 2025).

4. Closure, Alexander-type theorems, and Markov-type equivalence

The closure of a bonded braid is formed exactly as in classical braid theory by connecting the top endpoints to the bottom endpoints in the standard way. The result is a bonded link rather than a plain link. This closure operation is the bridge between the diagrammatic theory of bonded knots and the algebraic theory of bonded braids (Diamantis et al., 20 Jul 2025, Cavicchioli et al., 6 Jul 2025).

A bonded Alexander theorem is one of the central structural results. In the language of (Diamantis et al., 20 Jul 2025),

$(\beta,B),$3

The proof adapts the classical Alexander/Lambropoulou–Rourke braiding procedure. Up-arcs in a link diagram are replaced by braid strands, bonds are straightened horizontally and arranged so that braiding does not interfere with them, and if a bond lies on an up-arc, topological vertex twists or rigid vertex twists are used to reconfigure the local picture before braiding. After the classical braiding algorithm is applied to the link part, the braid is adjusted to tight form using braided vertex slide moves (Diamantis et al., 20 Jul 2025).

The companion results in (Cavicchioli et al., 6 Jul 2025) state separately that every topological bonded knot $(\beta,B),$4 can be represented as the closure $(\beta,B),$5 of an element $(\beta,B),$6 of $(\beta,B),$7, and every rigid bonded knot $(\beta,B),$8 can be represented as the closure $(\beta,B),$9 of an element $\beta$0 of $\beta$1. The proof is described by the steps: put the bonded knot diagram into PL form, isolate the bonds and make them parallel when needed, choose a braiding point $\beta$2, ensure each segment is braided around $\beta$3, apply $\beta$4-moves to convert non-braided segments into braided ones, and then read off the braid word. An explicit example is

$\beta$5

On the equivalence side, (Diamantis et al., 20 Jul 2025) formulates an $\beta$6-equivalence theorem. Two oriented topological standard bonded links are isotopic if and only if any corresponding bonded braids are related by a finite sequence of $\beta$7-moves and bond commuting

$\beta$8

For tight bonded links, one replaces $\beta$9-moves by resolved $n$0-moves and uses elementary bond commuting

$n$1

The same paper then reformulates the result as a Markov theorem: two bonded braids have isotopic closures if and only if they are related by braid isotopy and a finite sequence of Markov conjugation,

$n$2

elementary bond commuting,

$n$3

and Markov stabilization,

$n$4

It explicitly excludes stabilization involving a bond, because such a move would create a vertical bond outside the bonded braid category (Diamantis et al., 20 Jul 2025).

In (Cavicchioli et al., 6 Jul 2025), the topological Markov theorem is expressed as bonded Markov equivalence: closures $n$5 and $n$6 are equivalent if and only if the braid representatives are related by a finite sequence of conjugation,

$n$7

cyclic permutation of bonds,

$n$8

and stabilization,

$n$9

The rigid analogue additionally allows cyclic permutation of kink generators,

$K$00

5. Singular-braid correspondence and bonded Burau representations

A striking algebraic observation is that the bonded braid monoid of (Diamantis et al., 20 Jul 2025) is isomorphic to the singular braid monoid: $K$01 The identification is explicit: $K$02 This transfers algebraic information between bonded braids and singular braids. It also explains why several low-dimensional representation-theoretic properties of bonded braid monoids parallel known results for singular braid monoids (Diamantis et al., 20 Jul 2025, Cavicchioli et al., 6 Jul 2025).

The paper "Bonded braids and the Markov theorem" extends the classical Burau representation to bonded braid monoids and groups. The classical braid generators are sent to the usual Burau blocks

$K$03

while the bond generators are sent to

$K$04

For the rigid theory, the kink generators are sent to

$K$05

The explicit inverses include

$K$06

and

$K$07

so the group representation extends over

$K$08

A reduced bonded Burau representation is obtained by conjugating with

$K$09

which produces block upper-triangular forms and yields a representation on $K$10. For $K$11, this reduced representation is one-dimensional: $K$12

The representation-theoretic conclusions are explicit. The bonded Burau representation is reducible for all $K$13. The reduced bonded Burau representation is faithful for $K$14, faithful for $K$15, faithful for $K$16 by comparison with the singular braid monoid case and a known result of Dasbach, unknown for $K$17, and not faithful for $K$18 because the classical Burau representation is already non-faithful for $K$19 and all $K$20 (Cavicchioli et al., 6 Jul 2025).

6. Enhanced bonds, bonded braidoids, applications, and neighboring theories

The bonded braid framework sits inside a broader program. The 2025 synthesis develops enhanced bonded knots and braids by introducing two types of bonds, attracting and repelling, which are inverse to each other. In braid form this yields the enhanced bonded braid group $K$21, generated by

$K$22

with the added inverse relations

$K$23

The bonded braid monoid embeds into this group. The same paper also introduces bonded knotoids and their algebraic counterpart, bonded braidoids, to model open chains with inter and intra-chain bonds; closure in that setting is called the bonded closure (Diamantis et al., 20 Jul 2025).

The motivating applications remain geometric and biological. Bonded knots arise naturally in topological protein modeling, where intramolecular interactions such as disulfide bridges stabilize folded configurations. The formalism is presented as a model for proteins, RNA, and other biological macromolecules, and as a way to encode not only entanglement but also internal attachments. This suggests a division of labor: braid generators record crossing data, while bond generators record embedded connections that are invisible to ordinary braid theory (Cavicchioli et al., 6 Jul 2025, Diamantis et al., 20 Jul 2025).

Several neighboring braid theories are conceptually related but should not be conflated with bonded braids. Blocked-braid groups are quotients of Artin braid groups defined by inserting a braid between fixed boundary pieces $K$24 and $K$25, in composites of the form $K$26. They are relevant to the broader intuition of “braids with attachments,” but their defining mechanism is quotienting by blocked-braid equivalence rather than adding bond arcs to a braid diagram (Maglia et al., 2013). “Magic” leatherworking braids are classified as a kernel

$K$27

inside a quotient of a spherical framed braid group; the essential constraints there are fixed coupon order and ribbon framing rather than embedded bonds (Hobkirk et al., 8 Jun 2026). Braids-and-ties algebras provide a further algebraic analogue: they decorate braid generators $K$28 with tie idempotents $K$29, and the associated paper states explicitly that ties are analogous rather than literal bonded braids (Fasano et al., 25 Nov 2025).

A common source of confusion is therefore terminological rather than mathematical. Bonded braids are not merely blocked braids, leatherworking braids, or braids-and-ties under a different name. In the recent topological literature, they are classical braids together with embedded bonds, equipped with their own isotopy calculus, closure theory, monoid and group structures, and Markov-type equivalence theorems. Their significance lies in extending braid theory from pure crossing data to braid diagrams carrying additional embedded linkage data, with both topological and biomolecular applications (Diamantis et al., 20 Jul 2025, Cavicchioli et al., 6 Jul 2025).