Lecture 3
Magnetic knots and groundstate energy spectrum:
- Magnetic relaxation
- Topology bounds the energy
- Inflexional instability of magnetic knots
- Constrained minimization of magnetic energy
- Groundstate energy spectra of magnetic knots and links - Bending energy spectra: Magnetic vs. elastic systems
Selected references
Maggioni, F & Ricca, RL 2009 On the groundstate energy of tight knots.
Proc R Soc A 465, 2761.
Ricca, RL 2008 Topology bounds energy of knots and links. Proc R Soc A 464, 293.
Ricca, RL & Maggioni, F 2014 On the groundstate energy spectrum of
magnetic knots and links. J Phys A: Math & Theor 47, 205501.
Magnetic relaxation
Magnetic relaxation (Moffatt 1985): consider in a viscous and perfectly conducting fluid;
Near equilibrium we have:
Thus:
with magnetic energy monotonically decreasing as . t → ∞
dM
dt = B ⋅ ∂ B
∂ t
V B
( ) ∫ d
3X = B ⋅ ∇ × u × B ( )
V B
( ) ∫ d
3X
= − u ⋅ ∇ × B ( ) × B
V B
( ) ∫ d
3X = − u ⋅ J × B [ ]
V B
( ) ∫ d
3X
Du
Dt ∝ u
hence
dM
dt ∝ − u
V
( ) ∫
Ln 2d
3X ⇒ M ( L
n,ϕ) ↓t
Du
Dt = − ∇p + J × B ⇒ u ∝ −∇p + J × B ϕ :L
n→ L
n,ϕ.
.
Magnetic relaxation to groundstate
Lk
i= 0 ∀i
N components same flux
Magnetic helicity:
H = A ⋅ B
V (B)
∫ d
3x = Φ
2Lk
iji≠ j
∑
Φ Φ
1Φ
1Φ
2Φ
2ϕ ϕ
knot tightening link tightening
Magnetic relaxation (Moffatt 1985) under -preserving flow: { V, Φ
i}
Theorem 1 (Freedman 1988; Moffatt 1990).
i) ∃
+l.b.: lim ; ii) ,
t→∞
M t ( ) = M
infM
min= m Φ
2V
1/3where is a topological invariant of . m L
N.
Φ
i= Φ .
Φ
ϕ :L
n→ L
n,ϕLet be a “zero-framed” magnetic link:
Topology bounds the energy
Theorem 2 (Arnold 1974; Freedman & He 1991).
i) ; ii) , where depends on the geometry of .
M t ( ) ≥ q H
q > 0 supp B ( )
M t ( ) ≥ 2
π
1/ 3
Φ
2C V
1/ 3L n
Theorem (Ricca 2008). Let be a zero-framed link. Then, we have:
I) : ,
II) : . M
min= 2 π
1/3
c
minΦ
2V
1/3M t ( ) ≥ 2
π
1/ 3
1
V
1/ 3H
m = 2 π
1/3
c
minq = 2
π
1/3
1 V
1/3Proof:
From and we have Φ
i= Φ Lk
i= 0 ∀i ∈ 1,..., n ( )
from C = ε
rand , we have
r
∑
i≠ j
∑ ∑
i≠ jLk
ij= 1 2 ∑
i≠ j∑
rε
rH = Φ
2Lk
i;
i = j
∑ + Lk
iji≠ j
∑ = Φ
2Lk
iji≠ j
= 0 ∑
. Therefore
. Now, from Theorem 2(ii), we have:
or
;
H = Φ
2Lk
iji≠ j
∑ ≤ Φ
2C C ≥ Φ H
2M ≥ 2 π
1/ 3
Φ
2C
V
1/ 3≥ 2 π
1/ 3
Φ
2c
minV
1/ 3also, by using above, we have:
Hence, from Theorem 2(i) follows (I). Moreover, at , from Theorem 1(ii) we have (II).
.
M
min2 π
1/3
Φ
2C
V
1/3≥ 2 π
1/3
Φ
2V
1/3H
Φ
2= 2 π
1/3
H V
1/3(∗)
(∗)
In general, we have:
Question: do knot types of same c
minrelax to same groundstate energy?
M
min∝ c
minC ≥ ε
rr
∑
i≠ j
∑ = Lk
iji≠ j
∑ ≥ Lk
iji≠ j
∑
Inflexion at I (in isolation): c = 0 generic behaviour in : 3
(Ricca & Moffatt 1992)
Reidemeister type I move in action:
(TRACE mission 2002)
F m F
⊥∝ c ˆn
Lorentz force
inflexion
I X s, t ( ) = s − 2
3 t
2s
3, −ts
2, s
3Inflexional magnetic knots
From inflexion-free knots to magnetic braids
Definition. A knot in an inflexion-free configuration is a spiral knot.
Theorem (Ricca 2005). Let denote a loose magnetic knot in inflexional state. Then is in inflexional disequilibrium. K
t0K
t0Corollary. If is in inflexional disequilibrium, it relaxes to a spiral knot , for K t >t t
00 .
K t
inflexional
instability
From minimal knots to spiral knots
Knots and links tabulation
10-crossing knot table (Tait 1885)
3-component links by KnotPlot
(Scharein 2000)
…
up to arbitrarily large # crossings
O(10
6)
Groundstate energy of zero-framed knots and links
Relaxation of 0-framed, 5-crossing knots:
M
min∝ c
minUnder signature-preserving flows, we have:
V = 1, Φ = 1, 0-framing c
min# knot types
3 1
4 1
5 2
6 3
7 7
8 21
9 49
10 166
… …
ϕ K (
5.1)
ϕ K
(
5.2)
6
136
230
4magnetic field:
B = 0, 1 L
dΦ
Pdr , 1 2 π r
dΦ
Tdr + 0, ∂ψ
∂ s , − ∂ψ
∂ϑ
twist parameter:
fluxes , :
Φ
PΦ
TConstrained minimization of magnetic energy of knots
B = 0, B (
ϑ( ) r , B
s( ) r )
Φ
PΦ
Ttubular knot : ; Mercier (orthogonal) system:
Theorem (Maggioni & Ricca 2009). Let us assume that (i) { V, Φ } invariant ( V = 1, Φ = 1 );
(ii) circular cross-section independent of arc-length;
(iii) independent of arc-length;
(iv) L independent of internal twist.
Then, we have ψ ˜
.
K V K ( ) = π a
2L ( r,ϑ,s )
h = Φ
P/ Φ
T∇ ⋅ B = 0
( )
R
*L
*ropelength: λ = L * /R * M
λ*( ) h = λ
4/3
2 π
2/3+
π
4/3h
2λ
2/3= m( λ , h)
c min
h
3 4 5
6 7
8 9
10 Groundstate energy spectrum: averaging over complexity
m( λ , h)
c min
m
min( ) h = 3
2 π
2/3h
4/3( h ≥ 2 )
(see also Chui & Moffatt 1992)
tightening: V = 1, Φ = 1 , h = 0
SONO ( Pieranski et al. 1998)
RIDGERUNNER (Ashton et al. 2010)
Groundstate energy spectrum of first 250 prime knots
10 crossings
9 crossings
8 crossings
7 crossings
6 crossings
5 crossings 4 crossings 3 crossings
m(#
K) = m ( λ (#
K), 0 )
M
ο*= λ (#
K) 2 π
4/3
m(#
K)
#
KM
ο*= 2 ( π
2)
1/3V = 1, Φ = 1 , h = 0 tight unknot:
(Rolfsen table)
m(#
K) = 4.5ln #
K+10.5
best fit:
linear fit over c
min-families
Tight knots
RIDGERUNNER
m(#
K)
m(#
K) = m ( λ (#
K), 0 )
M
ο*= λ (#
K) 2 π
4/3
M
ο*= 2 ( π
2)
1/3V = 1, Φ = 1 , h = 0 tight unknot:
Knot energy spectrum by increasing ropelength (Ricca & Maggioni 2014)
#
KRIDGERUNNER
(increasing ropelength)
m(#
K)
m(#
K) = m ( λ (#
K), 0 )
M
ο*= λ (#
K) 2 π
4/3
m(#
K) = 4.5ln #
K+ 9.3
best fit:
linear fit over c
min-families
Tight links M
ο*= 2 ( π
2)
1/3V = 1, Φ = 1 , h = 0 tight unknot:
Link energy spectrum by increasing ropelength (Ricca & Maggioni 2014)
#
K(increasing ropelength)
RIDGERUNNER
V = 1, h = 0
Bending energy spectrum of elastic knots
e(#
K) e(#
K)
e(#
K) = 0.2 ln #
K+ 0.6
best fit:
linear fit over c
min-families
Tight knots
RIDGERUNNER
#
KRIDGERUNNER
e = E
bE
o= ∫
K[ c(s) ]
2ds
2
4/3π
5/3E
o= π R
*2
1/3π
5/3tight torus:
(increasing ropelength)
e = E
bE
o= ∫
K[ c(s) ]
2ds
2
4/3π
5/3E
o= π R
*2
1/3π
5/3V = 1, h = 0 tight torus:
Bending energy spectrum of elastic links
e(#
K) e(#
K)
e(#
K) = 0.2 ln #
K+ 0.5
best fit:
linear fit over c
min-families
Tight links
RIDGERUNNER RIDGERUNNER
#
K(increasing ropelength)
Magnetic versus bending energy (Ricca & Maggioni 2017)
#
K(increasing ropelength)
χ (#
K)
χ (#
K)
8_4_3
Tight knots: χ (#
K) = 21.62
Tight links: χ (#
K) = 21.42
λ (#
K)
[ ]
4/3∝ a ln #
K+ b [ λ (#
K) ]
4/3∝ c
minin partial agreement with
O c ( )
min3/4≤ λ (#
K) ≤ O c (
minln
5c
min)
(Buck & Simon 1999; Cantarella et al . 2002; Diao 2003; Diao et al. 2013).
m c (
min) ≡ M
minm
ο= c
minπ
λ ( ) #
K≥ 2 π
1/4c
min3/4≈ 2.66 c
min3/4; then , or
From lower bounds on energy (V = 1, Φ = 1 ), we have:
M
min= 2 π
1/3