Skip to content

Chapter 56: Pfaffians

Prerequisites: Determinants (Ch03) · Permanents (Ch40A) · Skew-symmetric Matrices (Ch02)

Chapter Outline: Special Determinants of Skew-symmetric Matrices → Definition and Polynomial Construction of the Pfaffian → Fundamental Property: \(\operatorname{Pf}(A)^2 = \det(A)\) → Recursive Expansion and Index Matching → Algebraic Identity: \(\operatorname{Pf}(M^T A M) = \det(M)\operatorname{Pf}(A)\) → Relationship between Pfaffians and Perfect Matchings in Graphs → Applications: The Ising Model in Statistical Mechanics, Perfect Matching Counting in Planar Graphs (FKT Algorithm), and Supersymmetry in Quantum Mechanics

Extension: The Pfaffian is the "algebraic square root" of the determinant of a skew-symmetric matrix; it not only resolves the sign ambiguity of taking the root of a determinant in skew-symmetric systems but also, through its unique index-matching structure, serves as the core link between matrix algebra and the dimer problem in statistical physics.

When discussing symmetric matrices, we focus on eigenvalues. However, when examining skew-symmetric matrices (\(A^T = -A\)), a surprising phenomenon occurs: their determinant is always a perfect square. The Pfaffian is the algebraic expression of this square root. It is more compact in form than the determinant and plays an irreplaceable role in treating matching problems in graph theory and describing fermion pairs in physical systems.


56.1 Definition of the Pfaffian

Definition 56.1 (Pfaffian)

For a skew-symmetric matrix \(A\) of even order \(2n\), the Pfaffian is defined as: $\(\operatorname{Pf}(A) = \frac{1}{2^n n!} \sum_{\sigma \in S_{2n}} \operatorname{sgn}(\sigma) \prod_{i=1}^n a_{\sigma(2i-1), \sigma(2i)}\)$ It is a homogeneous polynomial of degree \(n\) in the entries of the matrix.

Theorem 56.1 (The Core Identity)

For any skew-symmetric matrix \(A\) of even order: $\(\operatorname{det}(A) = [\operatorname{Pf}(A)]^2\)$ This means the Pfaffian determines the signed root of the skew-symmetric determinant.


56.2 Fundamental Properties

Algebraic Identities

  1. Scaling: \(\operatorname{Pf}(\lambda A) = \lambda^n \operatorname{Pf}(A)\).
  2. Congruence: \(\operatorname{Pf}(M A M^T) = \det(M) \operatorname{Pf}(A)\).
  3. Block Diagonal: \(\operatorname{Pf}(\operatorname{diag}(A, B)) = \operatorname{Pf}(A)\operatorname{Pf}(B)\).

56.3 Graphs and Matchings

Application: Matching in Planar Graphs

The FKT algorithm proves that for a planar graph, by assigning specific orientations to edges (a Pfaffian orientation), the Pfaffian of its associated skew-adjacency matrix is exactly equal to the number of perfect matchings in the graph. This transforms an exponential counting problem into a polynomial determinant-like calculation.


Exercises

1. [Basics] Calculate the Pfaffian of the \(2 \times 2\) skew-symmetric matrix \(A = \begin{pmatrix} 0 & a \\ -a & 0 \end{pmatrix}\).

Solution

Steps: 1. Calculate the determinant: \(\det(A) = 0 \cdot 0 - a(-a) = a^2\). 2. By the identity \(\operatorname{Pf}(A)^2 = \det(A)\), we have \(\operatorname{Pf}(A) = \pm a\). 3. From the explicit degree-1 definition: \(\operatorname{Pf}(A) = a_{12} = a\). Conclusion: \(\operatorname{Pf}(A) = a\).

2. [Dimension] Prove that the Pfaffian of an odd-order skew-symmetric matrix is not well-defined (or equals zero).

Solution

Proof: 1. Let \(A\) be a skew-symmetric matrix of order \(n\) (odd). 2. \(\det(A) = \det(A^T) = \det(-A) = (-1)^n \det(A) = -\det(A)\). 3. Thus, \(\det(A) = 0\). 4. Since \(\operatorname{Pf}(A)^2 = \det(A)\), the value must be zero. Conclusion: Pfaffian theory primarily concerns even-dimensional spaces.

3. [Calculation] Calculate the Pfaffian of the standard symplectic matrix \(J = \begin{pmatrix} 0 & 1 \\ -1 & 0 \end{pmatrix} \oplus \cdots \oplus \begin{pmatrix} 0 & 1 \\ -1 & 0 \end{pmatrix}\).

Solution

Using Properties: 1. The Pfaffian of a block-diagonal matrix is the product of the Pfaffians of the blocks. 2. Each \(2 \times 2\) block \(J_2 = \begin{pmatrix} 0 & 1 \\ -1 & 0 \end{pmatrix}\) has \(\operatorname{Pf}(J_2) = 1\). Conclusion: \(\operatorname{Pf}(J) = 1 \cdot 1 \cdots 1 = 1\). This defines the orientation of the volume form in symplectic geometry.

4. [Property] How does the Pfaffian change if rows 1 and 2 are swapped (and columns 1 and 2 are swapped simultaneously)?

Solution

Conclusion: It changes sign (multiplied by -1). Reasoning: This operation is equivalent to a congruence \(M A M^T\) where \(M\) is a permutation matrix with \(\det(M) = -1\). By the identity \(\operatorname{Pf}(M A M^T) = \det(M) \operatorname{Pf}(A)\), the result flips sign.

5. [Recursion] Write the expansion of a \(4 \times 4\) Pfaffian along its first row.

Solution

Formula: \(\operatorname{Pf}(A) = a_{12} a_{34} - a_{13} a_{24} + a_{14} a_{23}\). This shows how the Pfaffian covers all elements through perfect matchings (dimers) of indices.

6. [Comparison] Compare the computational complexity of \(\operatorname{perm}(A), \det(A)\), and \(\operatorname{Pf}(A)\).

Solution

Comparison: - \(\det(A)\): \(O(n^3)\), polynomial. - \(\operatorname{perm}(A)\): \(O(n 2^n)\), #P-complete (extremely hard). - \(\operatorname{Pf}(A)\): \(O(n^3)\) (via algorithms similar to Gaussian elimination), polynomial. Significance: Remarkably, the Pfaffian maintains combinatorial counting power while possessing the same efficiency as the determinant.

7. [Application] What is the "Pfaffian method" for the Ising Model?

Solution

In statistical mechanics, the partition function of the 2D Ising model can be expressed as the Pfaffian of a large skew-symmetric matrix. This discovery allowed physicists to solve the model exactly, revealing the algebraic nature of phase transitions.

8. [Property] Is \(\operatorname{Pf}(A \otimes J_2) = \det A\) true?

Solution

Yes, for an \(n \times n\) matrix \(A\), the Pfaffian of the \(2n \times 2n\) matrix \(A \otimes J_2\) is equal to \(\det A\). This relates standard linear operators to their skew-symmetric representations in symplectic spaces.

9. [Basics] Determine the Pfaffian of \(\begin{pmatrix} 0 & 1 & 2 & 3 \\ -1 & 0 & 1 & 2 \\ -2 & -1 & 0 & 1 \\ -3 & -2 & -1 & 0 \end{pmatrix}\).

Solution

Calculation: Use the \(4 \times 4\) formula: \(a_{12}a_{34} - a_{13}a_{24} + a_{14}a_{23}\). \(= 1 \cdot 1 - 2 \cdot 2 + 3 \cdot 1 = 1 - 4 + 3 = 0\). Verification: The matrix has rank 2 and determinant 0, so \(\operatorname{Pf}=0\) is correct.

10. [Supersymmetry] Why is the Pfaffian important in describing Supersymmetry?

Solution

Supersymmetry involves the transformation between fermions (antisymmetric) and bosons (symmetric). In calculating fermion path integrals, the result often manifests as the Pfaffian of the skew-symmetric part of an operator. This ensures the consistency of signs for physical amplitudes under various symmetry operations.

Chapter Summary

The Pfaffian is the core operator of skew-symmetric algebra:

  1. Elegance of the Square Root: It proves that the global measure (determinant) of a skew-symmetric system has a more fundamental, half-degree algebraic root, eliminating sign ambiguity.
  2. Shortcut for Counting: By transforming complex combinatorial matching problems into polynomial-time Pfaffian operations, it provides the most powerful algebraic pivot for solving graph-theoretic challenges.
  3. Physical Mapping: As the algebraic carrier for fermion statistics and the Ising model, the Pfaffian reveals the profound connection between microscopic symmetry and macroscopic phase transitions in nature.