on the twisted alexander polynomial for metabelian sl 2 c
play

On the twisted Alexander polynomial for metabelian SL 2 ( C ) - PowerPoint PPT Presentation

On the twisted Alexander polynomial for metabelian SL 2 ( C ) representations with the adjoint action Yoshikazu Yamaguchi JSPS Research fellow (PD) Tokyo Institute of Technology RIMS Seminar Twisted topological invariants and topology of


  1. On the twisted Alexander polynomial for metabelian SL 2 ( C ) –representations with the adjoint action Yoshikazu Yamaguchi JSPS Research fellow (PD) Tokyo Institute of Technology RIMS Seminar “Twisted topological invariants and topology of low-dimensional manifolds” Yoshikazu Yamaguchi The twisted Alexander poly. for metabelian reps.

  2. Basic notion and notation The twisted Alexander polynomial The twisted A refinement of ∆ K ( t ) Alexander = (the Alexander polynomial) with ρ : π 1 → GL ( V ) polynomial Notation E K := S 3 \ N ( K ) a knot exterior, ρ Ad SL 2 ( C ) Ad ◦ ρ : π 1 ( E K ) − → − − → Aut ( sl 2 ( C )) Ad ρ ( γ ) : v �→ ρ ( γ ) v ρ ( γ ) − 1 γ �→ ρ ( γ ) �→ � 0 1 � 1 0 � 0 0 � � � sl 2 ( C ) = C ⊕ C ⊕ C 0 − 1 0 0 1 0 The adjoint action Ad gives a connection with the character variety Hom ( π 1 ( E K ) , SL 2 ( C )) // SL 2 ( C ) . Yoshikazu Yamaguchi The twisted Alexander poly. for metabelian reps.

  3. Metabelian representations Definition of metabelian reps. ρ : π 1 ( E K ) → SL 2 ( C ) is metabelian ρ ([ π 1 ( E K ) , π 1 ( E K )]) ⊂ SL 2 ( C ) an abelian subgroup. ⇐ ⇒ Remark ρ : π 1 ( E K ) → SL 2 ( C ) is abelian ρ ( π 1 ( E K )) ⊂ SL 2 ( C ) ⇐ ⇒ an abelian subgroup, ⇐ ⇒ ρ ([ π 1 ( E K ) , π 1 ( E K )]) = { 1 } , ρ ⇐ ⇒ SL 2 ( C ) π 1 ( E K ) H 1 ( E K ; Z ) π 1 ( E K ) / [ π 1 ( E K ) , π 1 ( E K )] ≃ � µ � ρ is determined by only ρ ( µ ) . We focus on non–abelian representations. Yoshikazu Yamaguchi The twisted Alexander poly. for metabelian reps.

  4. Definitions of reducible and irreducible reps. ρ : π 1 ( E K ) → SL 2 ( C ) ρ : reducible Def ⇒ ∃ L ⊂ C 2 ⇐ s.t. ρ ( g )( L ) ⊂ L ( ∀ g ∈ π 1 ( E K )) by taking conjugation, �� �� a � ∗ ⊂ SL 2 ( C ) � ρ : π 1 ( E K ) → � a ∈ C \ { 0 } a − 1 � 0 ρ : irreducible Def ⇐ ⇒ ρ : not reducible. Yoshikazu Yamaguchi The twisted Alexander poly. for metabelian reps.

  5. Reducible and irreducible reps. in metabelian ones ρ : π 1 ( E K ) → SL 2 ( C ) metabelian ρ : reducible �� � � 1 � ω ⊂ SL 2 ( C ) � ρ : [ π 1 ( E K ) , π 1 ( E K )] → ± � ω ∈ C � 0 1 �� a �� � ∗ since Im ρ ⊂ � � a ∈ C \ { 0 } and a − 1 � 0 � 1 � ∗ A , B ∈ SL 2 ( C ) : upper triangular ⇒ ABA − 1 B − 1 = 0 1 ρ : irreducible �� �� a � 0 ⊂ SL 2 ( C ) � ρ : [ π 1 ( E K ) , π 1 ( E K )] → � a ∈ C \ { 0 } a − 1 � 0 � 0 � 1 ρ ( µ ) = by F. Nagasato. − 1 0 Yoshikazu Yamaguchi The twisted Alexander poly. for metabelian reps.

  6. Background ρ : π 1 ( E K ) → SL 2 ( C ) metabelian ρ : reducible ∆ K ( t ) appears in the twisted Alexander polynomial. → Hyperbolic torsion at “bifurcation points” − in Hom ( π 1 ( E K , SL 2 ( C ))) // SL 2 ( C ) . ρ : irreducible “Does ∆ K ( t ) appear in the twisted Alexander?” Yoshikazu Yamaguchi The twisted Alexander poly. for metabelian reps.

  7. Main Theorem Theorem Suppose that � an irred. metabelian and; ρ : π 1 ( E K ) → SL 2 ( C ) s.t. “ longitude–regular ” . (the twisted Alexander poly.) · ∆ α ⊗ Ad ◦ ρ Then ( t ) = ( t − 1 )∆ K ( − t ) P ( t ) , E K where ∆ K ( t ) is the Alexander polynomial of K. Yoshikazu Yamaguchi The twisted Alexander poly. for metabelian reps.

  8. Hierarchy of metabelian representations Remark ρ : π 1 ( E K ) → SL 2 ( C ) reducible = ⇒ ρ : metabelian �� a � 1 ∗ ( ∵ ) Im ρ ⊂ ∗ � a � = 0 � � � � �� ⇒ ρ ([ π 1 ( E K ) , π 1 ( E K )]) ⊂ ± 0 a − 1 0 1 We have the following hierarchy of metabelian representations: abelian ⊂ reducible ⊂ metabelian difference difference roots of ∆ K ( t ) ∆ K ( − 1 ) Yoshikazu Yamaguchi The twisted Alexander poly. for metabelian reps.

  9. The details of reducible metabelian representations Existence of reducible representations (G. Burde, G. de Rham) ∃ ρ : π 1 ( E K ) → SL 2 ( C ) : reducible and non–abelian if and only if ∆ K ( λ 2 ) = 0 where λ is an eigenvalue of ρ ( µ ) . { the characters of non–abelian reducible representations } Hom ( π 1 ( E K ) , SL 2 ( C )) // SL 2 ( C ) , = C abel ∩ C non–abel in where C abel is the component of abelian characters and C non–abel is the components of non–abelian ones. Yoshikazu Yamaguchi The twisted Alexander poly. for metabelian reps.

  10. The twisted Alexander polynomial of reducible SL 2 ( C ) -representations Theorem ρ : π 1 ( E K ) → SL 2 ( C ) reducible and non–abelian, � λ is an eigenvalue of ρ ( µ ) and corresponding to λ ∈ C s.t. , ∆ K ( λ 2 ) = 0 then ( t ) = ∆ K ( λ 2 t ) ( t − 1 ) · ∆ K ( λ − 2 t ) ( λ 2 t − 1 ) · ∆ K ( t ) ∆ α ⊗ Ad ◦ ρ ( λ − 2 t − 1 ) . E K Yoshikazu Yamaguchi The twisted Alexander poly. for metabelian reps.

  11. The details of irreducible metabelian representations Existence of irreducible metabelian reps. (Nagasato) ♯ { the characters of irereducible metabelian reps. } = | ∆ K ( − 1 ) | − 1 . 2 ι ∈ H 1 ( E K ; Z 2 ) induces ι on Hom ( π 1 ( E K ) , SL 2 ( C )) // SL 2 ( C ) . an involution ˆ Remark (Nagasato & Y.) { the characters of irereducible metabelian reps. } in Hom ( π 1 ( E K ) , SL 2 ( C )) // SL 2 ( C ) . = { the fixed point set of ˆ ι } Remark A higher rank analogy was given by H. Boden and S. Friedl. Yoshikazu Yamaguchi The twisted Alexander poly. for metabelian reps.

  12. Main Theorem (again) Theorem Suppose that � an irred. metabelian and; ρ : π 1 ( E K ) → SL 2 ( C ) s.t. “ longitude–regular ” . (the twisted Alexander poly.) · ∆ α ⊗ Ad ◦ ρ Then ( t ) = ( t − 1 )∆ K ( − t ) P ( t ) , E K where ∆ K ( t ) is the Alexander polynomial of K. Yoshikazu Yamaguchi The twisted Alexander poly. for metabelian reps.

  13. The details of ∆ α ⊗ Ad ◦ ρ ( t ) E K Homomorphisms ρ : π 1 ( E K ) → SL 2 ( C ) is metabelian ρ ([ π 1 ( E K ) , π 1 ( E K )])( ⊂ SL 2 ( C )) is abelian, ⇐ ⇒ Suppose that ρ ([ π 1 ( E K ) , π 1 ( E K )]) � = { 1 } . α : π 1 ( E K ) → π 1 ( E K ) / [ π 1 ( E K ) , π 1 ( E K )] ≃ H 1 ( E K ) = � t � s.t. α ( µ ) = t the twisted Alexander poly. � � � � ∂ r i det α ⊗ Ad ◦ ρ ∂ g j i , j � = 1 ∆ α ⊗ Ad ◦ ρ ( t ) = E K det ( α ⊗ Ad ◦ ρ ( g 1 − 1 )) from a presentation π 1 ( E K ) = � g 1 , g 2 , . . . , g k | r 1 , . . . , r k − 1 � . Yoshikazu Yamaguchi The twisted Alexander poly. for metabelian reps.

  14. Main tools We need a “good” presentation of π 1 ( E K ) for metabelian reps. X-S. Lin introduced a suitable presentation of π 1 ( E K ) by using a free Seifert surface of K . a Seifert surface S is free ⇔ S 3 = S × [ − 1 , 1 ] ∪ S 3 \ S × [ − 1 , 1 ] : a Heegaard splitting. Figure: a free Seifert surface of the trefoil knot Yoshikazu Yamaguchi The twisted Alexander poly. for metabelian reps.

  15. Main tools We need a “good” presentation of π 1 ( E K ) for metabelian reps. X-S. Lin introduced a suitable presentation of π 1 ( E K ) by using a free Seifert surface of K . a Seifert surface S is free ⇔ S 3 = S × [ − 1 , 1 ] ∪ S 3 \ S × [ − 1 , 1 ] : a Heegaard splitting. Figure: a free Seifert surface of the trefoil knot Yoshikazu Yamaguchi The twisted Alexander poly. for metabelian reps.

  16. Main tools We need a “good” presentation of π 1 ( E K ) for metabelian reps. X-S. Lin introduced a suitable presentation of π 1 ( E K ) by using a free Seifert surface of K . a Seifert surface S is free ⇔ S 3 = S × [ − 1 , 1 ] ∪ S 3 \ S × [ − 1 , 1 ] : a Heegaard splitting. Figure: a free Seifert surface of the trefoil knot Yoshikazu Yamaguchi The twisted Alexander poly. for metabelian reps.

  17. Main tools We need a “good” presentation of π 1 ( E K ) for metabelian reps. X-S. Lin introduced a suitable presentation of π 1 ( E K ) by using a free Seifert surface of K . a Seifert surface S is free ⇔ S 3 = S × [ − 1 , 1 ] ∪ S 3 \ S × [ − 1 , 1 ] : a Heegaard splitting. Figure: a free Seifert surface of the trefoil knot Yoshikazu Yamaguchi The twisted Alexander poly. for metabelian reps.

  18. Lin’s presentation By using a free Seifert surface S with genus 2 g , i µ − 1 = a − π 1 ( E K ) = � µ, x 1 , . . . , x 2 g | µ a + i ( i = 1 , . . . 2 g ) � where x i is a closed loop corresponding to 1–handle in S 3 \ S × [ − 1 , 1 ] , a ± i is a word in x 1 , . . . , x 2 g , corresponding to closed loops in the spine of S . Figure: a free Seifert surface of the trefoil knot Yoshikazu Yamaguchi The twisted Alexander poly. for metabelian reps.

  19. Lin’s presentation By using a free Seifert surface S with genus 2 g , i µ − 1 = a − π 1 ( E K ) = � µ, x 1 , . . . , x 2 g | µ a + i ( i = 1 , . . . 2 g ) � where x i is a closed loop corresponding to 1–handle in S 3 \ S × [ − 1 , 1 ] , a ± i is a word in x 1 , . . . , x 2 g , corresponding to closed loops in the spine of S . Figure: a free Seifert surface of the trefoil knot Yoshikazu Yamaguchi The twisted Alexander poly. for metabelian reps.

Recommend


More recommend