Skip to main content

Full text of "A fermionic superfluid state for many spinful species - III"

See other formats


On the new superfluid state - III 

George Kastrinakis* 

Institute of Electronic Structure and Laser (IESL), 

Foundation for Research and Technology - Hellas (FORTH), P.O. Box 1527, Iraklio, Crete 71110, Greece 

(Dated: July 26, 2010) 

In two previous reports, we introduced a new fermionic variational wavefunction, suitable for 
interacting multi-species systems and sustaining superfluidity. This disentangled wavefunction con- 
tains a new quantum index of "ghostly" character, which does not appear in any observable quantity. 
Here we introduce a general spin triplet version of this wavefunction. 

o. 

In two previous reports [l|, |2| , we introduced a new variational wavefunction, suitable for interacting multi-species 
'— H , systems and sustaining superfluidity. This a generalization of the Bardeen-Cooper-Schrieffer (BCS) wavefunction 3] 
)—* ', I^bcs) — Ilfc(' u 's + v k c \ t-fc ±)|0)- The creation/annihilation operators c k a /ck,a describe fermions with momentum 
i *■* • k and spin a, and |0) is the vacuum state. 

(vq | We introduce a spin triplet version of this wavefunction, which is a generalization of the Balian-Werthamer state 

[J], including all three components of the total spin. In [2j we introduced the "equal spin pairing" (ESP) case with 
parallel pair spins only. 

Let the usual fermionic operators be c x /c x with x = {i, k,a}, where i denotes the fermion species/flavor. 



o 



We introduce a new quantum index, which is related to the internal symmetry of the quantum state. It serves 
»' . to enumerate the otherwise "entangled" components of the quantum state as a function of both momentum and 
CIh' spin. In this way, the treatment of the coherence factors (coefficients entering the wavefunction and then everywhere 

else in the theory) is greatly facilitated. Thereby we introduce the new fermionic operators c x / 'c x ^ obeying the 

anticommutators ({a, b} = ab + ba) 



CO 

3 



X 



\Cx,fj.i Cy,v j — U j \Cx,/i.; Cy } i/J — ®xy LLIS ■ [*-) 



t3 

r^ ■ Then, we write the usual c x /c x as the superposition 

O' 

c * = -/wJ2<6 > C, = -?wY, c ^ ■ ( 2 ) 

CN ■ v^o s=1 v^o s=1 

> 

The usual anticommutation relations of c x /c x are preserved, while 

{Cx,Cy,v} = , {Cx,cl <v } = -^= . (3) 

r-^ ■ We also introduce 

Q,k,6 = Ui ^ k + V hk \ C i,k,f,S c i,-k,l,S + c i,k,l,6 c i,-k,t,8) + w i,k,f c i,fc,f,5 C i,-k,t,S + w i,k,i %k,±,5 C i,-k,i,8 ( 4 ) 

+ s J,fc \ c i,k^,8 c j,-k,l,8 + C i,k,i,5 C j,-k,t,8 "*" C i,-fc,T,<5 C j,k,l,5 + C i,-k,l,8 c j,k.^,8) 
+U,k,f { c i,k,t,8 c j,-k,t;8 + c i,-k,t,S c j,k,t,8) + *».*4 \ C i,k,l,5 C j,-k,\,,S + C i,-k,i,5 C j,k,i,S) ' 

C\ k s is a bosonic operator, creating spin triplet pairs of fermions (for singlet pairs c.f. (l|), and (i,j) = {(1, 2), (2, 1)}. 

Other variants of C ik s can be envisaged as well. 

Henceforth we divide the momentum space into two parts, say k > (sgn(fc) = +) and k < (sgn(fc) = — ). For 
k > wc form the following multiplet of Cj k s 's 

M k = C lkS=1 C 2kS=2 . (5) 

This multiplet creates all states with momenta ±fc, and we take N — 2 (c.f. |l|, l2| also). 

The new index allows for the bookkeeping of a superposition of states of a given particle, i.e. same x = {i, k,a}, 
without the difficulties due to entanglement within the multiplet M k , if the index were removed. In that case, the 
treatment of the coherence factors u, v, w, s, t is prohibitively complicated. 



This index is "ghostly", i.e. it appears only within the state |^) below. It does not appear in any 
observable quantity! (C.f. the Faddeev-Popov ghosts in gauge field theories, where the ghosts disappear from 
the final physical outcome.) We note that there is no change whatsoever implied in the Hamiltonian or in the 
representation of any observable. 

Now we introduce the dis-entanglcd state 

i*)=n M ^i°) ■ ( 6 ) 

k>0 

Note that all C ik ^s in |\l/) commute with each other. 

Y$!) generalizes I^bcs) & n d sustains superfluidity. This wavefunction makes sense for two or more fermion species, 
with an interaction between different species. It can obviously be generalized for three or more fermion species. 
Moreover, a similar wavefunction using the new quantum index can be written in the real space representation 
instead of the momentum space one. At the moment, it is not possible to say how close |\t) is to experimental reality. 
It does represent a very promising avenue though, as can be seen from the discussion which follows. 

It allows for inequivalence between spin up and down fermions. Plus, it allows for the "exact" variational treatment 
of a wider class of Hamiltonians than sheer BCS type, e.g. comprising interaction and hybridization between different 
fermion species, in the well known manner of the BCS-Gorkov theory [3] , [5] . 

The normalization ( 1 $ r | , $ r ) = 1 implies 

Kfcl 2 + 2K fe | 2 + Kfc, t | 2 + K,*al 2 + 4 kfcl 2 + 2fc fe , t | 2 + 2|t life4 | 2 = 1 . (7) 

Fermion statistics yields v^-fc — —Vi, k and Wi t - k ,a = — Wi,fc,<r- 

Two fermion species at zero temperature. Calculations are straightforward for the matrix elements derived 
from |\&). E.g. for 2 fermion species with dispersions ti. k . a — £i,-k,a and for k > we have 

\/N~o Cl,fc, t Ml |0) = (Vl,k 4 ,-fc4,5=l + «Jl,fc,t A,-k,t,6=\ + Sl - k 4.-k,l,5=l + *l,fc,t 4,-fc,t,«5=l) C 2,k,S=2 1°) 

-(S2,k 4,-fe,4,5=2 + f 2,fe,t 4,-fc4,«5=2) C lk.8=l 1°) ■ ( 8 ) 



Then 



Likewise, 



M k 4, fct ci,fe,t M l 1°) = Tr(Kfc| 2 + ki.fc.tl 2 + Wk? + |*iAt| 2 + l s 2,fe| 2 + |t 2 ,fe,t| 2 ) • (9) 



and 



(0| M k 4,fc,t C l,*,t M k 1°) = -Jf{i*,k,r Wl .*.t + f 2,fc,t W 2,k,t + Vl,kS*,k + V 2,k S 2.k) ■ (10) 

(0| M k pa.-fc.tci.k.t Ml |0) = ^-« fc ti,fc,t ~ u$,k *2,fc,t) ■ (11) 

Using the commutativity of C ik a 's and generalizing the previous equations, we obtain ((B) = (^\B\^)) 

rii,k,<7 = (4,k,a C i,k,o) = Jj-(\ v i,k\ 2 + \ w i,k,a\ 2 + \ s i,k\ % + l*»,fe,a| 2 + \sjm\ 2 + \tj.k,a\ 2 ) , 



(12) 



Ck.a = (4,fc )CT Cj.fc.o-) = - — (W* X(7 ti, k ,a + Wj.k.a t* ka + < fe S l ,fe + V^ k S* k ) , (13) 



Ik.a = (Cj,-k,aC i,k,a) = TT«,fc *i,fe,tr ~ u j,fc <J,A,<r) . (I 4 ) 

fffe, CT = (C2,-fe,- CT Ci,fe, CT ) = -J^r(ul tk Sl,fe - U^j. s 2 : fc) , (15) 

&i,ft,<x = (Ci,_fc,-<T c J,fe,ff) = -7J- U*,k v i,k , di,k,<7 = (Ci-k,a Ci,k,<r) = jr U*,k w i,k,a , (16) 



with (i,j) = (1,2), (2,1). 



A general Hamiltonian for two fermion species interacting via intra-species potentials Vi.,2 and via an inter-species 
potential F q , and hybridizing via hk, is 



H = Y^ &X" C \,k,a C ^<j +^h k [c\ k(J C2,k,a + 4,k,<r C l,fc,<r) (17) 

i,k,a k,a 

o 2^i *'? C i,k+q,<r C i,p-q,a /Ci >P^' Ci < k <" + 2-~i q C l,fc+g,CT C 2,j>-q,cr' C 2,P,CT' c l,fc,CT i 



i.k.p.q,a.a' k,p,q,a,<r' 



with i = 1,2, £»,&,«, = £i,fe lCT — ^i, CT and /Zj i(T the chemical potential. 
Considering ^ above, we have for (H) = (^f\H\^f), 



(H) — 2^ £i,fc,<7 n i,k,a + 2_^, ^k [Ck,tr + Cfc,<r) — 9 2_^ K:,fc-p «i,fc,<r 1i,p, CT + « 2_^ ^,9=0 n i,fc,<7 ™i,p,o- (18) 

i,k,<r k,a i,k,p,a i,k,p,a 

+ 9 Z^ ^». fc -P \bi,k,ab*,p,<T + di,k,crdi, p ,<r) ~ 2_^ F^-p (k,a Cp,cr + -^=0 /^ n l,k,<r n 2,p,er 
i,k,p,a k,p,a k,p,a 

+ 51 ^-P (I*,* 7p,o- + 9k.a 9p ta ) , 
A:,p,cr 

with (i,j) = {(1, 2), (2, 1)}. The first term in the second line is exactly the usual BCS-like pairing term, and the last 
term is the equivalent inter-species pairing term due to F q . Allowing for pairs with non-zero total momentum in |W), 
as shown in |l| , yields additional terms in (H) . 

The minimization procedure for (H) and the finite temperature extension proceed as shown in [l|. In general, 
expanding the Hilbert space of |\&) via the inclusion of more pairing correlations than the ones shown, may lead to 
further reduction of the ground state energy. 

* E-mail address : kast@iesl.forth.gr , giwkast@gmail.com 



[1] G. Kastrinakis, arXiv: 0901. 2487 v9 . 
[2] G. Kastrinakis, arXiv:10 07T0745l . 

[3] J. Bardeen, L. N. Cooper, and J. R. Schrieffer, Phys. Rev. 108, 1175 (1957). 
[4] R. Balian and N. R. Werthamer, Phys. Rev. 131, 1553 (1963). 

[5] A. A. Abrikosov, L. P. Gorkov, I. E. Dzyaloshinski, Methods of Quantum Field Theory in Statistical Physics, Prentice-Hall 
(Cliffwoods, NY, 1964).