A relation between the HOMFLY-PT and Kauffman polynomials via characters

This paper establishes a relationship between HOMFLY-PT and Kauffman polynomials for specific knot classes using Birman-Murakami-Wenzl algebra characters to prove a conjectured correspondence with Harer-Zagier factorisability for 3-strand knots, while demonstrating through 4-strand counterexamples that this correspondence does not hold universally for knots with higher braid indices.

The Gist

Imagine you are trying to describe the shape of a tangled piece of string (a "knot") using a mathematical recipe. In the world of mathematics and physics, there are two famous recipes for this: the HOMFLY–PT polynomial and the Kauffman polynomial.

Think of these recipes as two different languages describing the same knot.

• The HOMFLY–PT language is like a high-tech, digital camera. It sees the knot in a very specific, "oriented" way (like a left-handed screw vs. a right-handed screw).
• The Kauffman language is like a vintage film camera. It sees the knot in a slightly different, "unoriented" way, capturing a broader, more flexible view.

Usually, these two languages are very different. Translating a knot description from one to the other is like trying to translate a poem from English to a language that doesn't exist yet—it's messy and often impossible.

The Big Discovery: When the Languages Match

The authors of this paper, Andreani Petrou and Shinobu Hikami, discovered a special club of knots where these two languages suddenly start speaking the exact same dialect.

They found that for certain knots (specifically those made by twisting strands in very specific patterns called "full twists" and "Jucys-Murphy twists"), the complex Kauffman recipe can be built directly from the simpler HOMFLY–PT recipe. It's as if they found a secret dictionary that perfectly translates between the two for a specific group of knots.

The "Magic Mirror" (The BMW Algebra)

How did they find this dictionary? They used a mathematical tool called the Birman-Murakami-Wenzl (BMW) algebra.

Imagine the BMW algebra as a magic mirror.

• When you look at a knot in this mirror, it breaks the knot down into its fundamental building blocks (called "characters").
• For the HOMFLY–PT language, the mirror shows you blocks made of "SU(N)" bricks.
• For the Kauffman language, the mirror shows you blocks made of "SO(N+1)" bricks.

The authors realized that for the "special club" of knots, the "SO(N+1)" bricks are just a slightly modified version of the "SU(N)" bricks. If you know how to build the knot with SU(N) bricks, you can almost instantly build it with SO(N+1) bricks, provided you add a tiny bit of "correction glue."

The Twist in the Tale: The 4-Strand Problem

Here is where the story gets interesting. The researchers had a hunch (a conjecture) that this perfect translation would work for any knot that has a "clean" mathematical structure (what they call "HZ factorisability").

They tested this on knots made of 3 strands (like a 3-strand braid).

• Result: It worked perfectly! Every 3-strand knot with a clean structure could be translated. The "correction glue" was simple and predictable.

But then, they tried it on knots made of 4 strands.

• Result: The translation broke! They found "counterexamples." These were 4-strand knots that had the "clean structure" but still couldn't be translated from HOMFLY–PT to Kauffman.

The Analogy:
Imagine you have a rule: "If a car has 4 wheels and an engine, it can drive on the highway."

• For 3-wheeled vehicles (tricycles), this rule holds true.
• But for 4-wheeled vehicles, they found a car with 4 wheels and an engine that still couldn't drive on the highway because it was missing a specific type of tire (the "correction glue" was too complicated).

This proved that the rule "Clean Structure = Perfect Translation" is not true for 4 strands or more. The HOMFLY–PT/Kauffman relationship is actually a stricter, more demanding condition than just having a clean structure.

Why Should We Care? (The Physics Connection)

Why do mathematicians and physicists care about translating knot recipes?

1. Topological Strings: In physics, these knots represent the paths of tiny strings in the universe. The HOMFLY–PT recipe counts "oriented" strings, while Kauffman counts "unoriented" ones.
2. BPS States: The difference between the two recipes corresponds to a specific type of particle state (BPS states) that involves "cross-caps" (like a Möbius strip).
3. The Vanishing Act: When the two recipes match (the translation works), it means those specific "cross-cap" particles disappear (their count becomes zero). This is a huge deal for physicists trying to understand the stability of the universe at a quantum level.

Summary

• The Goal: Find a way to translate between two different mathematical descriptions of knots.
• The Method: Use a "magic mirror" (BMW algebra) to break knots into building blocks.
• The Success: They proved the translation works perfectly for a large family of 3-strand knots.
• The Surprise: They found that for 4-strand knots, the translation fails even when the knots look "clean."
• The Lesson: The relationship between these two knot languages is more complex and restrictive than previously thought. It's not just about the shape of the knot; it's about the deep, hidden algebraic structure holding the strands together.

In short, the authors built a better dictionary for knot languages, but they also discovered that the dictionary has a "Page 4" where the rules change, teaching us that the universe of knots is even more intricate than we imagined.

Technical Summary

Here is a detailed technical summary of the paper "A relation between the HOMFLY–PT and Kauffman polynomials via characters" by Andreani Petrou and Shinobu Hikami.

1. Problem Statement

The paper addresses the long-standing problem of establishing a general relationship between two major knot invariants: the HOMFLY–PT polynomial (associated with the Lie group $SU(N)$) and the Kauffman polynomial (associated with $SO(N+1)$).

• Context: While these polynomials coincide at $N=2$ (corresponding to the Jones polynomial due to the isomorphism $su(2) \simeq so(3)$), finding a general relation for arbitrary $N$ is non-trivial.
• Previous Work: A specific relation was known for torus knots (Labastida and P´erez) and recently extended to certain hyperbolic knots via the Harer-Zagier (HZ) factorisability of the HOMFLY–PT polynomial.
• The Conjecture: It was conjectured that the HOMFLY–PT/Kauffman relation holds if and only if the HZ transform of the HOMFLY–PT polynomial is factorisable.
• The Gap: The authors aim to rigorously prove this relation for broader families of knots, understand the representation-theoretic origin of the differences between the polynomials, and test the validity of the conjecture for higher braid indices (specifically 4 strands).

2. Methodology

The authors employ a representation-theoretic approach utilizing character expansions and the Birman-Murakami-Wenzl (BMW) algebra.

• Character Expansions:
  • They revisit the $SU(N)$ character expansion for the HOMFLY–PT polynomial, which expresses the polynomial as a sum over Young diagrams $Q$ weighted by Racah coefficients ($h_Q$) and Schur functions ($S_Q$).
  • They propose a "naive" $SO(N+1)$ expansion ($X_{SO}$) by replacing Schur functions with quantum dimensions ($d_Q$) of $SO(N+1)$ while keeping the Racah coefficients unchanged.
• BMW Algebra Representations:
  • To explain the discrepancies between the naive $X_{SO}$ and the actual Kauffman polynomial, the authors utilize the BMW algebra $C_m(\alpha, \beta)$.
  • They derive the Kauffman character expansion, which involves a sum over Young diagrams with $m, m-2, \dots, 0$ boxes. Crucially, this expansion includes a term $\chi_0$ (associated with the trivial representation of the BMW algebra) that is absent in the standard $SU(N)$ expansion.
• Correction Terms:
  • The difference between the actual Dubrovnik polynomial ($Y$) and the naive expansion ($X_{SO}$) is identified as a correction term $\delta$.
  • The authors prove that $\delta$ is directly proportional to the trace of the BMW algebra representation $\chi_0$.
• Braid Families:
  • The analysis focuses on knots constructed via full twists ($F_k$) and Jucys-Murphy twists ($E_k$), which are central or specific elements in the BMW algebra.

3. Key Contributions and Results

A. Derivation of the Correction Term $\delta$

The paper establishes that the Kauffman polynomial $Y$ can be written as:
$$Y = X_{SO} + \delta$$
where $X_{SO}$ is the naive character expansion using $SO(N+1)$ quantum dimensions.

• 3-Strand Case: The correction term $\delta$ arises solely from the $\chi_0$ component of the BMW algebra representation.
• Explicit Formulas: The authors provide explicit matrix representations for the BMW algebra generators ($\sigma_i, e_i$) for 2, 3, and 4 strands. They derive exact formulas for $\delta$ in terms of the trace $\chi_0(-iA, iq)$.
• Torus Knots: For torus knots $T(m, n)$, the correction term $\delta$ vanishes (or is a simple monomial for even $n$), confirming the known relation for these cases.

B. Proof for 3-Strand Knots

The authors prove that for a specific general family of 3-strand knots ($K^{\pm}_{j,k,l}$) constructed by combining full twists and Jucys-Murphy twists:

1. The HOMFLY–PT/Kauffman relation holds.
2. The HZ transform is factorisable.
   This supports the conjecture that HZ factorisability $\iff$ HOMFLY–PT/Kauffman relation for the 3-strand case.

C. Counterexamples in 4-Strand Knots (The Main Breakthrough)

The paper critically tests the conjecture for 4-strand knots.

• HZ Factorisability Condition: A sufficient condition for HZ factorisability in 4 strands involves the vanishing of the Racah coefficient $h_{[22]}$ and a specific alternating polynomial structure for $h_{[31]}$.
• HOMFLY–PT/Kauffman Condition: The condition for the polynomials to be related requires a specific constraint on the $\chi_0$ term involving the sum of characters for diagrams with 2 boxes.
• Disproof of the Conjecture: The authors identify a family of 4-strand knots ($K^{(4)}_{j,k,l}$) that satisfy the HZ factorisability conditions but fail the HOMFLY–PT/Kauffman relation condition.
  • Result: The HOMFLY–PT/Kauffman relation implies HZ factorisability, but the converse is false for braid index $m \geq 4$.
  • Implication: The HOMFLY–PT/Kauffman relation is a stronger condition than HZ factorisability for higher strand numbers.

D. Physical Interpretation

The paper connects these mathematical results to Topological String Theory:

• The difference between the $SU(N)$ and $SO(N+1)$ polynomials corresponds to the difference between oriented and unoriented surfaces.
• The vanishing of the HOMFLY–PT/Kauffman relation (i.e., when they are related) implies the vanishing of 2-cross cap BPS invariants.
• The failure of the relation in 4-strand knots suggests the presence of non-vanishing 2-cross cap BPS states, even when the HZ transform is factorisable.

4. Significance

1. Refinement of Knot Theory Conjectures: The paper corrects a widely held conjecture regarding the equivalence of HZ factorisability and the HOMFLY–PT/Kauffman relation, establishing a strict hierarchy where the polynomial relation is a subset of factorisability for $m \geq 4$.
2. Representation Theory Application: It demonstrates the power of the BMW algebra and character expansions in solving complex problems in knot invariants, providing explicit matrix representations and trace formulas that simplify computations compared to traditional skein relations.
3. BPS Invariants: The results offer new insights into the structure of BPS states in topological string theory, specifically regarding the conditions under which unoriented surface contributions (cross-caps) vanish.
4. Computational Efficiency: The character expansion method is shown to be computationally efficient, allowing for the analysis of knots with high crossing numbers (e.g., ~70 crossings) in minutes, whereas skein relation methods would be intractable.

Summary Conclusion

Petrou and Hikami successfully articulate the conditions under which the HOMFLY–PT and Kauffman polynomials are related using the character theory of the BMW algebra. While they confirm the relation for a broad class of 3-strand knots, they definitively disprove the "if and only if" nature of the conjecture for 4-strand knots, proving that HZ factorisability is necessary but not sufficient for the polynomial relation in higher dimensions. This work bridges deep algebraic structures with physical invariants in topological string theory.