Notes on Bethe Ansatz

Notes on Bethe Ansatz Mikhail Zvonarev December, 010 Coordinate and Algebraic Bethe Ansatz techniques are reviewed; the former is applied to the Lieb-Liniger model, and the latter to the Heisenberg model. Suggestions for the further reading are given. 1 Coordinate Bethe Ansatz Bethe Ansatz means Bethe s substitution, the method is named after the work by Hans Bethe [1]. Bethe found eigenfunctions and spectrum of the one-dimensional spin-1/ isotropic magnet (often called one-dimensional spin-1/ isotropic Heisenberg model). I will demonstrate the usage of the Algebraic Bethe Ansatz on this model in the second part of my lecture. Next model was solved by coordinate Bethe Ansatz more than 30 years later []. This is a model of one-dimensional gas of bosons interacting via the δ-function potential. Since it was solved by Lieb and Liniger, it is often called the Lieb-Liniger model. I will demonstrate how coordinate Bethe Ansatz works using this model as an example. 1.1 Hamiltonian of the Lieb-Liniger model in the first and secondquantized form The Hamiltonian of the one-dimensional Bose gas with zero-range interaction potential is H = dx [ Ψ (x) xψ(x) + cψ (x)ψ (x)ψ(x)ψ(x)], (1.1) where the units have been chosen such that = m = 1. The interaction constant c 0 has the dimension of inverse length. The dimensionless coupling strength γ is given by γ = c ρ. (1.) The boson fields Ψ and Ψ satisfy canonical equal-time commutation relations: [Ψ(x), Ψ (y)] = δ(x y), [Ψ (x), Ψ (y)] = [Ψ(x), Ψ(y)] = 0. (1.3) 1

The number of particles operator N is N = and the momentum operator P is P = i dx Ψ (x)ψ(x), (1.4) dx Ψ (x) x Ψ(x). (1.5) The number of particles and momentum are integrals of motion: The equation of motion for the Hamiltonian (1.1) is called the quantum nonlinear Schrödinger equation. [H, N] = [H, P ] = 0. (1.6) i t Ψ = xψ + cψ ΨΨ (1.7) Since the Hamiltonian commutes with N, Eq. (1.6), the number of particles in the model is conserved and one can look for the eigenfunctions of the Hamiltonian (1.1) in a sector with the fixed number of particles. Write Ψ N (k 1,..., k N ) = 1 d N x χ N (x 1,..., x N k 1,..., k N )Ψ (x 1 )... Ψ (x N ) 0. (1.8) N! Here Ψ N (k 1,..., k N ) is an eigenfunction of the second-quantized Hamiltonian (1.1): H Ψ N (k 1,..., k N ) = E N (k 1,..., k N ) Ψ N (k 1,..., k N ), (1.9) the symbol 0 denotes the Fock vacuum state Ψ(x) 0 = 0, 0 Ψ (x) = 0, 0 0 = 1, (1.10) the symbol d N x dx 1... dx N, and the parameters k 1,... k N are called quasi-momenta. The coordinate wave function χ N is an eigenfunction of the Hamiltonian (1.1) written in the first-quantized form: H N = + c δ(x i x j ), (1.11) i=1 x i 1 i<j N H N χ N = E N (k 1,..., k N )χ N. (1.1) The Bose statistics of the model is encoded in the symmetry properties of χ N : it should be symmetric with respect to the arbitrary permutations of the coordinates x 1,..., x N. Finally, write the momentum operator (1.5) in the first-quantized form: P N = i x j. (1.13) No boundary conditions were imposed so far. Since it is our purpose to get the gas at a finite density in the limit N, we should compactify the system in the space direction

when working at finite N. This can be done in several different ways: one can put the system on a ring of circumference L and impose the periodic boundary conditions, or put it in the hard-walls box of length L, or, more generally, impose twisted boundary conditions. The influence of the boundary conditions should disappear in the thermodynamic limit, but for finite N they are, of course, affect the results. The exact eigenfunctions of the model are known for the general case of the twisted boundary conditions. The particular case of the periodic boundary conditions is, however, enough for our purposes and will be assumed henceforth. 1. Lieb-Liniger model in the c limit: Tonks-Girardeau gas The limit of infinite repulsion, c, is often called the Tonks-Girardeau limit [3, 4]. We will consider this limit in full details since it will give us a lot of intuition about the behavior of the model at arbitrary c. 1..1 Eigenfunctions and spectrum of the Tonks-Girardeau gas The eigenfunctions of the Hamiltonian (1.11) should go to zero when x i x j for arbitrary i j : χ N (x 1,..., x i,..., x j,..., x N ) 0, x i x j, i j, (1.14) and they should solve the free Schrödinger equation when all x j are different. A possible candidate is the determinant of an N N matrix with the entries exp{ik j z l } : This function should be symmetrized in coordinates: χ N (x 1,..., x N k 1,..., k N ) = det[exp{ik j x l }]. (1.15) C N! det[exp{ik j x l }] 1 l<j N sgn(x j x l ), (1.16) where C is a normalization constant. It is easy to check that (1.16) is a common eigenfunction of the operators H, P and N with the corresponding eigenvalues being equal to E N = kj, P N = k j. (1.17) In getting (1.17) we did not specify the boundary conditions. Let us now put a system on a ring of circumference L and impose the periodic boundary conditions. The wave function (1.16) should thus satisfy χ N (x 1,..., x j = 0,..., x N k 1,..., k N ) = χ N (x 1,..., x j = L,..., x N k 1,..., k N ) (1.18) for all j = 1,..., N. This implies the following conditions on the quasimomenta k j : exp{ik j L} = ( 1) N 1, j = 1,..., N. (1.19) 3

These equations are called Bethe equations; in this particular case they can be solved easily: k j = π L n j, j = 1,..., N. (1.0) Here n j are arbitrary odd numbers for N even and arbitrary even for N odd. Apart from this selection rule, the quantization condition for quasimomenta are the same as for free fermions: the allowed values of k j are equidistant in the momentum space and multiple occupation of the same position in the momentum space is prohibited: the determinant in Eq. (1.19) becomes equal to zero. The fermion-like behavior of this boson system is very typical for the quantum systems in one spacial dimension; this phenomenon is called statistical transmutation. It should be stressed, however, that the Tonks-Girardeau gas is not a system of the non-interacting fermions: the wave function (1.19) is symmetric in coordinates, while the wave functions of the fermion system should be antisymmetric. Thus, correlation properties of the Tonks-Girardeau gas are very different in many aspects from those of the free fermion system. An example illustrating this difference, the one-particle density matrix, will be studied in section 1.5. The norm of the wave function (1.19) with the periodic boundary conditions can be calculated easily: therefore L 0 d N x χ(x 1,..., x N k 1,..., k N ) = C L N, (1.1) C = 1 L N. (1.) It is not difficult to prove that the wave functions (1.16) with the quasi-momenta quantized by the conditions (1.19) form a complete set of eigenfunctions of the Hamiltonian (1.11) in the limit c. 1.. The ground state and thermodynamics of the Tonks-Girardeau gas at zero temperature The ground state of the Tonks-Girardeau gas can be found easily. To minimize E N, Eq. (1.17), one should take the following subset of k j from the set (1.0): k j = π ( j N + 1 ), j = 1,..., N. (1.3) L The momentum P N of the ground state is thus equal to zero. The thermodynamic limit is the limit of infinite number of particles at finite density ρ: ρ = N L = const, N, L. (1.4) One has in this limit L π πρ πρ dk. (1.5) 4

The ground state energy in the canonical ensemble is, therefore, E 0 = L π πρ πρ dk k = L 3 π ρ 3. (1.6) In the grand canonical ensemble the number of particles is not fixed and depends on the chemical potential; if the average number of particles in the system is N then E N = kj µ. (1.7) In the thermodynamic limit the chemical potential is related to the density of particles as follows: µ = (πρ). (1.8) One can see that the thermodynamics of the Tonks-Gigardeau gas at zero temperature is the same as that of the spinless free-fermion gas, with πρ playing a role of the Fermi momentum k F : k F = πρ. (1.9) The pressure of the 1D gas is given by the formula ( ) ( ) E0 F P = =. (1.30) L σ,n L T,N Here σ is the entropy of the gas, and F is the Helmholtz free energy, F = E 0 T σ. The partial derivatives taken at a fixed T coincide with the partial derivatives taken at a fixed σ for the zero temperature case. Because of this fact we will omit the corresponding subscripts in the zero temperature case. For example, we will write P = ( E 0 / L) N. The thermodynamic definition of the sound velocity is v S = L mρ ( ) P. (1.31) L N We set m = 1/ in the definition of the Lieb-Liniger model (1.1). One thus gets for the Tonks-Girardeau gas and P = 3 π ρ 3 (1.3) v S = πρ. (1.33) One can see from the expression for k F that the Fermi velocity v F is v F k F m = πρ (1.34) (recall that we set m = 1/). Comparing Eqs. (1.33) and (1.34) one concludes that the Fermi velocity coinside with the sound velocity in the Tonks-Girardeau gas. This is the expected result since the spectrum of the Tonks-Girardeau gas is the same as that of the free-fermion gas. In our studies of the Lieb-Liniger model with finite coupling c we will use the notion of the Fermi velocity as given by Eq. (1.34); the sound velocity v S at arbitrary c is always less than v F, see Fig. 1. 5

1.3 Lieb-Liniger model at arbitrary repulsion strength We will repeat, for arbitrary repulsion strength c, the steps performed in section 1.: we construct the eigenfunctions and the spectrum of the Lieb-Liniger model (1.11) at arbitrary coupling strength c, identify the ground state of the system and take the thermodynamic limit. We then calculate the velocity of sound v S, the compressibility κ, and the Luttinger parameter K from the thermodynamics of the Lieb-Liniger model. We discuss the weak (Bogoliubov) and strong (Tonks-Girardeau) coupling limits in the solution of the Lieb-Liniger model. 1.3.1 Eigenfunctions and spectrum of the Lieb-Liniger model The structure of the eigenfunctions at general c is much more complicated as compared to the eigenfunctions at c =, considered in section 1..1. That particular case, however, shows a way one should proceed: The eigenfunctions χ N (x 1,..., x N ) in the N-particle sector are symmetric under any interchange of coordinates, therefore it is sufficient to consider the following domain T in the coordinate space: T : x 1 < x < < x N. (1.35) In this domain χ N is an eigenfunction of the free Hamiltonian HN 0 = x j (1.36) H 0 Nχ N = E N (k 1,..., k N )χ N (1.37) The following boundary condition should be satisfied: ( ) c χ N = 0, x j+1 x j x j+1 = x j + 0. (1.38) This boundary condition obviously reduces to Eq. (1.14) in the Tonks-Girardeau limit. To prove Eq. (1.38) use the variables z = x j+1 x j and Z = (x j+1 + x j )/. One has = x x 1 z, + x 1 x = z + 1 Z. (1.39) Integrating Eq. (1.1) over z in the small domain z < ϵ 0 one can see that the eigenfunction of the free-particle Hamiltonian (1.36) obeying the boundary condition (1.38) is indeed an eigenfunction of the Hamiltonian (1.11). To get an explicit expression for the eigenfunctions (1.1) we start from the trial expression (1.15). We suggest the following form of the eigenfunctions: [ ( χ N = const ) ] + c det[exp{ik j x l }]. (1.40) x j x l N j>l 1 One should show that this expression satisfies Eqs. (1.36) and (1.38). Check, for instance, that ( ) c χ N = 0, x = x 1 + 0. (1.41) x x 1 6

Do do this, rewrite Eq. (1.40) as follows: ( χ N = ) + c χ N, (1.4) x x 1 where χ N = const N j=3 ( x j ) ( + c x 1 x j ) + c x N j>l 3 ( ) + c det[exp{ik j x l }]. (1.43) x j x l The function χ N is antisymmetric with respect to the interchange of x 1 and x, χ N (x 1, x ) = χ N (x, x 1 ). (1.44) Writing the left hand side of the Eq. (1.41) in the form [ ( ) ] c χ N (1.45) x x 1 one sees that it is antisymmetric with respect to the interchange of x 1 and x, and, therefore, equal to zero when x x 1. The boundary condition (1.38) for the other x j can be checked similarly. We thus showed that the wave function (1.40) is indeed an eigenfunction of the Hamiltonian (1.11). Using the definition of the determinant det[exp{ik j x l }] = { } ( 1) [P ] exp i x n k Pn (1.46) P n=1 one gets for (1.40) χ N = const { N! j>l[(k j k l ) + c ] } 1/ P { ( 1) [P ] exp i x n k Pn} [k Pj k Pl ic sign(x j x l )]. (1.47) n=1 j>l This is the desired expression for the eigenfunctions of the Hamiltonian (1.11). We wrote Eq. (1.47) in such a form that it is valid for arbitrary values of x 1,..., x N, thus removing the restriction (1.35). One can see that the wave functions given by Eq. (1.47) reduce to the wave functions given by Eq. (1.16) in the c limit. The eigenvalues of the Hamiltonian (1.11) and the momentum operator (1.13) on the eigenfunctions (1.47) are E N = kj, P N = k j, (1.48) 7

respectively. These formulas look very similar to that of a free system and because of this similarity the parameters k j are called quasi-momenta. No boundary conditions were imposed so far. Let us do it now and use the periodic boundary conditions. The condition (1.18) imposed on the function (1.47) gives exp {ik j L} = N l j k j k l + ic, j = 1,..., N. (1.49) k j k l ic These coupled nonlinear equations are called Bethe equations. In the c limit these equations reduce to Eqs. (1.19). It can be shown that all their solutions are real (see, for example, Ref. [5]). It is often more convenient to use the logarithmic form of Bethe equations: Lk j + θ(k j k m ) = π m=1 ( n j N + 1 ), j = 1,..., N. (1.50) Here {n j } is a set of integers, and the function θ is ( ) ic + k θ(k) = i ln, ic k θ(± ) = ±π. (1.51) An important property of the representation (1.50) is that the sets of n j and the sets of quasimomenta k j are in one to one correspondence. It is often more convenient to parameterize a quantum state by the set n j rather than by the set of quasi-momenta k j. Let us sum up Bethe equations (1.50) over j. The function θ(k) defined by Eq. (1.51) is antisymmetric, therefore θ(k j k m ) = 0 (1.5) and one gets L j,m=1 k j = π ( n j N + 1 ). (1.53) Comparing this equation with Eq. (1.48) one gets a relation between the momentum of an arbitrary state and the set of quantum numbers n j : P N = π ( n j N + 1 ). (1.54) L The calculation of the normalization constant in Eq. (1.47) is a nontrivial task (recall that in the Tonks-Girardeau limit this constant is given by Eqs. (1.1) and (1.)). A closed expression for this constant was suggested by Gaudin (see [6] and references therein) and proved in Ref. [7]; a detailed discussion can be found in Ref. [5]. 1.3. The ground state and thermodynamics at zero temperature The ground state of the model in the N-particle sector corresponds to the following choice of the quantum numbers n j in Eq. (1.50): n j = j, j = 1,..., N, (1.55) 8

that is the Bethe equations (1.50) take the form Lk j + m=1 ( θ(k j k m ) = π j N + 1 ), j = 1,..., N. (1.56) in the ground state of the model. The momentum of the ground state is equal to zero. Consider now the thermodynamic limit of the model, defined by Eq. (1.4). Write = k j+1 k j k j+1 k j = L ρ(k j )(k j+1 k j ) (1.57) where we have defined ρ(k j ) as follows ρ(k j ) = 1 L(k j+1 k j ). (1.58) In the thermodynamic limit one has for an arbitrary function f(k j ) : Λ f(k j ) L dk ρ(k)f(k). (1.59) Λ The parameter Λ plays a role of the Fermi momentum: all the states with k < Λ are occupied, and all the states with k > Λ are empty. The value of Λ is given by the normalization condition Λ Λ dk ρ(k) = N L = ρ. (1.60) To get the thermodynamic limit of the Bethe equations (1.56), take the difference of the equations with the indices j + 1 and j : L(k j+1 k j ) + [θ(k j+1 k m ) θ(k j k m )] = π. (1.61) The difference k j+1 k j is small in the thermodynamic limit, therefore Eq. (1.61) can be written as follows L(k j+1 k j ) + (k j+1 k j ) θ (k j k m ) = π, N, L, (1.6) m=1 where θ (k) θ(k)/ k. Using Eqs. (1.58) and (1.59) one gets after some algebra ρ(k) 1 Λ dq ρ(q)k(k, q) = 1, Λ k Λ, (1.63) π Λ π where K(k, q) θ (k q) = c c + (k q). (1.64) The linear integral equation (1.63) is called Lieb equation. Its solution, together with the normalization condition (1.60), gives the quasi-momentum distribution function ρ(k). All 9

thermodynamic parameters can be calculated from ρ(k) and Λ. The ground state energy in the canonical ensemble is Λ E 0 = kj L dk ρ(k)k. (1.65) Λ Following [], change the variables as follows: k = Λz, c = Λα, ρ(λz) = g(z). (1.66) Written in these variables, Eqs. (1.60), (1.63), and (1.65) are, respectively, γ 1 1 g(z) 1 1 π 1 dz g(z) = α, (1.67) αg(y) dy α + (y z) = 1 π, (1.68) and e(γ) E 1 0 Nρ = γ3 dy g(y)y. (1.69) α 3 1 Recall that γ is given by Eq. (1.). The system of equations (1.67) and (1.68) is much more suitable for the numerical analysis than the system (1.60) and (1.63). To solve the system (1.67) and (1.68) the following two steps should be done: (i) Solve Eq. (1.68) for a given α; (ii) use Eq. (1.67) to determine γ as a function of α; Having solved the system (1.67) and (1.68), one can easily go back to the equations in the form (1.60) and (1.63), if necessary. The Fermi momentum Λ can be found by combining Eqs. (1.) and (1.66): Λ = ρ γ [ 1 1 α = ρ dz g(z)]. (1.70) The quasi-momentum density ρ(k) is related to g(z) by Eq. (1.66). 1 1.3.3 Velocity of sound, compressibility, and Luttinger parameter Discuss first the thermodynamic definition of the sound velocity, Eq. (1.31). Taking E 0 from the Eq. (1.69), and using Eq. (1.30) one gets [ v S = ρ 3e(γ) γe (γ) + 1 1/ [ γ e (γ)] = µ(γ) 1 1/ (γ)] γµ, (1.71) where prime denotes differentiation with respect to γ and µ is the chemical potential given by ( ) E0 µ = = ρ [3e(γ) γe (γ)]. (1.7) N L Note an important technical point: since P and µ depend on N and L only via ρ = N/L, the following identities are valid: ( ) ( ) ( ) ( ) P P µ µ =, =. (1.73) ρ N ρ L ρ N ρ L 10

These identities are very useful in getting various thermodynamic relations. Another, microscopic definition of the sound velocity is v S = ϵ(k) P (k), (1.74) k=λ where ϵ(k) is the energy change of the system when a particle with the quasi-momentum k is added; P (k) is the momentum change of the system. The equivalence of the definitions (1.31) and (1.74) in the Lieb-Liniger was proven in Ref. []; for more modern presentation one can consult Ref. [5]. We will omit this proof since it is quite long; we only present here several equivalent representations for v S obtained in course of this proof: v S = ρ µ ρ = µ Λ = ρ πρ(λ). (1.75) Using the compressibility κ 1 L we can write the velocity of sound as ( ) L P N = 1 ( ) ρ ρ µ L (1.76) v S = κρ. (1.77) The Luttinger Liquid Hamiltonian contains two parameters: the velocity of sound, v S, and the Luttinger parameter, K. These parameters are related to the compressibility of the system [8]: K v S = πρ κ any model (1.78) In case the microscopic theory has some extra symmetry, there could exist some other relations. For example, if the system is Galilean-invariant, that is, the kinetic energy of any particle it consists of is p /m, one can show [9] that Kv S = πρ Galilean-invariant model (1.79) One can see that Eqs. (1.78) and (1.79) are consistent with Eq. (1.77), as it should be since the Lieb-Liniger model is Galilean-invariant. Note that πρ = v F, where v F is the sound velocity of the Tonks-Girardeau gas, Eqs. (1.33) and (1.34). 1.3.4 Weak and strong coupling limits and numerics The large γ limit is easy to work out since Eq. (1.68) admits a regular perturbative expansion with 1/α being a small parameter. One has Equation (1.67) then gives g(z) = 1 π + 1 π α + +, α, (1.80) π 3 α γ = πα + 4 3α 3 15α 16 +, α. (1.81) 4 45πα5 11

Inverting this expression, one gets α = γ π + π 4π 3γ + 16π 3γ 3 + ( ) π 16π 15 1 +, γ. (1.8) γ4 For the ground state energy per particle defined by Eq. (1.69) one gets [ e(γ) = π 1 4 3 γ + 1 γ + (π 15) 3 ] 15γ + 3 (1.83) From the exact expression (1.71) for the sound velocity in the Lieb-Liniger model, one gets the following large γ expansion: ( v large = v F 1 4 γ + 1 ) γ +, γ (error < 1% for γ > 10). (1.84) The leading terms in all the expansions written in the present section correspond to the Tonks-Girardeau limit and were obtained in the section 1... The opposite limit, γ 0, is much more difficult to analyze since the kernel in Eq. (1.63) becomes singular when α 0. The leading term is obtained in Ref. []: 1 z g(z), α 0. (1.85) πα For the ground state energy per particle defined by Eq. (1.69) one gets ( e(γ) γ 1 4 ) γ, γ 0. (1.86) 3π Let us substitute the expansion (1.86) into Eq. (1.71): ( 1 v small = v F γ 1 ) 1/ π π γ3/, γ 0 (error < 1% for γ < 10). (1.87) Since only two terms are written explicitly in the expansion (1.86), one should expand Eq. (1.87) and drop all the terms except the leading and the first subleading. However, the numerical analysis shows that Eq. (1.87) gives much better approximation to the exact expression (1.71) if not expanded. We plot v S /v F, v large /v F and v small /v F versus γ in Fig. 1. Recall that v F is the velocity of sound in the Tonks-Girardeau gas, Eq. (1.34), and the ratio v S /v F is the inverse Luttinger parameter, Eq. (1.79), 1.4 Excitations in the Lieb-Liniger model K 1 = v S v F. (1.88) We construct elementary (single quasi-particle and single quasi-hole) excitations in the Lieb-Liniger model and construct more complex excitations out of them. 1

v vf 1 0.8 0.6 0.4 0. 0.01 0.1 1 10 100 Γ Figure 1: Shown is the log-linear plot of the inverse Luttinger parameter in the Lieb-Liniger model, Eq. (1.88), versus γ. Red curve corresponds to the exact (numerical) result for v S, black dotted curve is obtained from Eq. (1.84), and black dashed curve from Eq. (1.87). 1.4.1 One-particle excitations in the Tonks-Girardeau gas The ground state quasimomentum distribution in the N-particle sector of the Tonks-Girargeau gas is given by Eq. (1.3): ( Lk j = π j N + 1 ), j = 1,..., N, (1.89) while in the N + 1-particle sector it takes a form ( L k j = π j N + ), j = 1,..., N + 1. (1.90) Therefore, every k j is shifted by π/l with respect to k j. A way to avoid dealing with this shift is to impose anti-periodic boundary conditions in the N + 1-particle sector. Then the ground state quasimomentum distribution is L k j = π ( j N + 1 ), j = 1,..., N + 1. (1.91) Note that the ground state is -fold degenerate, with another configuration corresponding to j = 0, 1,..., N. While the above choice of the boundary conditions eliminates the π/l shift, its influence onto the ground state energy is negligible small (vanishes as 1/N) in the large N limit. The structure of the one-particle excitations is now evident. They can be decomposed into the topological part resulting in the boundary conditions change and truly one-particle excitation. The corresponding Bethe equations are L k j = π ( j N + 1 13 ), j = 1,..., N. (1.9)

and L k N+1 = πm, M > N/. (1.93) One can see by comparing Eqs. (1.89) and (1.9) that k j = k j for j = 1,..., N. This result is specific to the Tonks-Girardeau gas and does not hold true for the finite repulsion case. However, the separation of the topological part of the one-particle excitation by using the above trick with the boundary conditions will prove useful in the finite repulsion case as well. 1.4. One-particle shift function We consider the N-particle ground state, Eq. (1.56). We then add one article and impose anti-periodic boundary conditions, like in the Tonks-Girardeau limit. The resulting Bethe equations are ( L k j + θ( k j k k ) + θ( k j k N+1 ) = π j N + 1 ), j = 1,..., N, (1.94) and k=1 L k N+1 + θ( k N+1 k k ) = πm, M > N. (1.95) k=1 The difference of Eqs. (1.56) and (1.94) is L(k j k j ) + [θ(k j k k ) θ( k j k k )] θ( k j k N+1 ) = 0. (1.96) k=1 Using that k j+1 k j 1/L in the thermodynamic limit one gets [θ(k j k k ) θ( k j k k )] k=1 = (k j k j ) θ (k j k k ) k=1 and it follows from Eq. (1.56) that [ (k j+1 k j ) L + (k k k k )θ (k j k k ), N, L (1.97) k=1 ] θ (k j k k ) = π, N, L. (1.98) k=1 Let us define the so-called shift function F (k j k N+1 ) as follows (we write k p instead of k N+1 in order to lighten the notations) F (k j k p ) = k j k j k j+1 k j. (1.99) Using Eqs. (1.97) (1.99) one gets from Eq. (1.96) an integral equation (often called dressing equation) onto the shift function: F (k k p ) 1 π Λ Λ dq K(k, q)f (q k p ) = 1 π θ(k p k). (1.100) This dressing equation is of crucial importance for the analysis of one-particle excitations. 14

1.4.3 Momentum of the excited state We introduce the so-called dressed momentum k(λ) by the formula p(k) = k + Λ An important property of k(k), valid for any γ is Λ dq ρ(q)θ(k q). (1.101) p(±λ) = ±k F. (1.10) It can be proven by integrating Eq. (1.63) with respect to k. Note that in the infinite repulsion limit the integral in Eq. (1.101) vanishes: p(k) = k as γ =. (1.103) The importance of the definition (1.101) becomes evident when one calculates the momentum difference P (k p ) between the N-particle ground state and N + 1-particle excited state. One has P (k p ) = ( k k k k ) + k p = F (k j k p )(k j+1 k j ) + k p k=1 Λ dk F (k k p ) + k p, N. (1.104) Λ Multiplying both parts of Eq. (1.100) by ρ(k), integrating over k, and taking into account Eq. (1.63) one gets an identity Λ Λ Λ dk F (k k p ) = dk ρ(k)θ(k p k) (1.105) Λ which implies P (k p ) = p(k p ), k p Λ. (1.106) 1.4.4 Energy of the excited state Let us introduce the integral equation onto the so-called dressed energy ε(λ) : ε(k) 1 Λ dq ε(q)k(k, q) = k µ (1.107) π Λ with an extra condition ε(±λ) = 0. (1.108) It can be shown [5] that ε(k) possesses the following properties ε(k) = ε( k) (1.109) ε (k) > 0, k > 0 (1.110) 15

and ε(k) < 0, k < Λ (1.111) ε(k) > 0, k > Λ (1.11) Differentiating Eq. (1.107) and taking the integral by parts one gets ε (k) 1 Λ dq ε (q)k(k, q) = k. (1.113) π Λ Let us now calculate the energy difference E(k p ) between the N-particle ground state and N + 1-particle excited state. One has E(k p ) = k p µ + ) ( k j kj = k p µ k j (k j k j ) kp µ Λ Λ dk kf (k k p ). (1.114) Multiplying Eq. (1.113) by F (k k p ), integrating over k, and using Eq. (1.100), one arrives at the identity Λ dk kf (k k p ) = 1 Λ dq ε(q)k(q, k p ). (1.115) Λ π Λ Substituting this identity into Eq. (1.114) we finally get E(k p ) = ε(k p ), k p Λ. (1.116) This formula clarifies a physical meaning of ε(k) introduced earlier by Eq. (1.107): it can be interpreted as the energy of an excitation with the momentum p(k). 1.4.5 One-hole excitations One-hole type of excitations can be generated by removing one particle from the N-particle ground state. An analysis similar to the one carried out for the one-particle excitation gives for the dressed momentum of the excitation, and for the dressed energy. P (k h ) = p(k h ), Λ k h Λ (1.117) E(k h ) = ε(k h ), Λ k h Λ (1.118) 16

Figure : Particle-hole excitation corresponding to min E for a given momentum 0 P k F. 1.4.6 Multiple particle-hole excitations An arbitrary excitation in the Lieb-Liniger model can be created from one-particle and onehole excitations. The momentum difference between the excited state and the ground state is P = p(k h ) (1.119) and the energy difference is E = particles p(k p ) holes particles ε(k p ) holes ε(k h ). (1.10) 1.4.7 Dispersion curves and their exact lower bound By the dispersion curve we call the curve of E plotted as a function of P. What is the minimal value of E for a given P lying in the region 0 P k F? The answer to this question (first given in Ref. [10]) is very elegant and intuitively clear: the corresponding excitation is shown in Fig.. It consists of a single particle-hole pair, with the particle lying at the right Fermi point and k F p(k h ) = P k F k F. (1.11) We plot E as a function of P for several values of γ, see Fig. 3. Note that the explicit analytic expression for min E can be found easily for γ = : E = P ( P ), 0 P k kf F as γ =. (1.1) k F k F 1.5 Correlation functions of the Tonks-Girardeau gas The knowledge of the exact eigenfunctions and spectrum for the Lieb-Liniger model makes it possible to develop the microscopic approach to the calculation of its correlation functions. 17

1 0.8 E kf 0.6 0.4 0. 0 0 0.5 1 1.5 P k F Figure 3: Shown are the plots of min E as a function of P for γ =, 100, 10, 1, 0.1 (from upper to lower). The procedure is straightforward: calculate the form-factors of local field operators in the basis formed by the eigenfunctions of the Hamiltonian (1.1) in the sector with given number of particles N, then sum up over all intermediate states, and, finally, take the thermodynamic limit. It is very easy to formulate such a program, and very difficult to implement it because of the complicated structure of the Bethe wave function (1.47) and Bethe equations (1.49). The book [5] is mostly devoted to the development of this program and is, probably, the most comprehensive monograph in this field. Among the results not included in the cited book one should mention the papers [11, 1, 13], where the local field operators were represented via the fundamental operators of the Quantum Inverse Scattering Problem (this approach work for spin chains, where the quantum space is finite-dimensional, but not for the Lieb-Liniger model; there is, however a trick (unpublished) solving the inverse scattering problem for the Lieb-Liniger model as well). Unfortunately, the final expressions for the correlation functions obtained within this straightforward approach are, in general, too complicated to be successfully analyzed either by analytic methods or numerically. There are, however, several important exceptions. One of them, the Tonks-Girardeau gas, is discussed here. The relatively simple form of the eigenfunctions of the Tonks-Girardeau gas make it possible to calculate the correlation functions explicitly for arbitrary number of particles N. The first non-trivial correlation function which was calculated is the one-particle density matrix. The calculations were done by Lenard [14] and the result is now known as the Lenard s formula. We present this formula in section 1.5.1. Lenard s formula contains the determinant of an N N matrix. The N limit of the determinant of an N N matrix is called the Fredholm determinant. It is discussed in section 1.5.. The thermodynamic limit of the Lenard s formula is studied in sections 1.5.3 and 1.5.4. The momentum distribution function for the Tonks-Girardeau gas is calculated in section 1.5.5. 18

1.5.1 One-particle density matrix: Lenard s formula The one-particle density matrix in the N-particle sector is defined as follows ρ N (x) = Ψ (x)ψ(0), (1.13) where the average is taken over the ground state of the system in the N-particle sector. The expression for the ρ N (x) in the first-quantized representation can be obtained easily (remember the commutation relations (1.3) and the representation (1.8)): ρ N (x) = Ψ N (k 1,..., k N ) Ψ (x)ψ(0) Ψ N (k 1,..., k N ) = 1 N! L 0 d N z d N z χ N (z 1,..., z N )χ N (z 1,..., z N) 0 Ψ(z 1 )... Ψ(z N ) Ψ (x)ψ(0) Ψ (z 1)... Ψ (z N ) 0 = N L 0 dz... dz N χ N (x, z,..., z N )χ N (0, z,..., z N ) (1.14) In the Tonks-Girardeau case χ N is given by Eqs. (1.16), the normalization constant C for the periodic boundary conditions is given by Eq. (1.), and the ground state quasi-momentum distribution is given by Eq. (1.3). Therefore ρ N (x) = 1 (N 1)!L N P P L 0 dz... dz N N ( 1) [P ]+[ P ] exp{ ixk P1 } exp{iz a (k Pa k Pa )}sgn(z a x). (1.15) All the integrals in this expression can be taken explicitly and give 1 L where L 0 a= dz a exp{iz a (k Pa k Pa )}sgn(z a x) = f(k Pa k Pa ), a =, 3,..., N, (1.16) One gets, therefore, ρ N (x) = 1 (N 1)!L f(k a k b ) = δ ka k b L P P exp{ix(k a k b )} 1. (1.17) i(k a k b ) ( 1) [P ]+[ P ] exp{ ixk P 1} f(k P1 k P1 ) N f(k Pa k Pa ). (1.18) Represent the permutation P in the internal sum as a sequence of two permutations a=1 P = P P (1.19) where P is the running index of the external sum. Further, introduce the notation k Pa P [k a ]. (1.130) 19

Written in these notations, Eq. (1.18) reads ρ N (x) = 1 (N 1)!L P ( 1) [P ] exp{ ixp [k 1 ]} f(p [P [k P 1 ]] P [k 1 ]) One can easily notice the following identity: therefore ρ N (x) = N f(p [P [k a ]] P [k a ]) = a=1 1 [P ] ( 1) (N 1)!L P = 1 L P ( 1) N N f(p [P [k a ]] P [k a ]). (1.131) a=1 N f(p [k a ] k a ), (1.13) a=1 a=1 f(p [k a ] k a ) P [P ] = 1 L N f(p [k a ] k a ) a=1 exp{ ixp [k 1 ]} f(p [P [k 1 ]] P [k 1 ]) exp{ ixk j } P exp{ ixk j } f(p [k j ] k j ) ( 1) [P ] a j f(p [k a ] k a ). (1.133) Consider the determinant of an N N matrix with the entries f(k j k l ) : f(k 1 k 1 ) f(k 1 k N )..... = N ( 1) [P ] f(p [k a ] k a ). (1.134) f(k N k 1 ) f(k N k N ) P a=1 To reproduce the last line of the Eq. (1.133) one should replace the l-th column of the matrix f(k j k l ) with the expression exp{ iz 1 k l }/L : ρ N (x) = λ det N(I + Ṽ1 + λṽ) = det N (I + Ṽ1 + Ṽ) det N (I + Ṽ1), (1.135) λ=0 where (Ṽ1) ab = exp{ix(k a k b )} 1, L i(k a k b ) (Ṽ) ab = 1 L exp{ ixk b}, a, b = 1,..., N. (1.136) Equation (1.135) can be rewritten in the following form ρ N (x) = λ det N(I + V 1 + λv ) = det N (I + V 1 + V ) det N (I + V 1 ), λ=0 (1.137) where (V 1 ) ab = π L sin[ x(k a k b )], (V ) ab = 1 { i π(k a k b ) L exp x } (k a + k b ), a, b = 1,..., N. (1.138) 0

Equations (1.137) and (1.138) are the desired Lenard s formula for the one-particle density matrix of the Tonks-Girardeau gas at zero temperature. It can be easily generalized to a finite temperature case. It is proper place to stress again that the Tonks-Girardeau gas is not equivalent to the free-fermion gas: though the spectrum of the Tonks-Girardeau gas is the same as that of the free fermion gas, the wave functions have different symmetry. This changes some observables drastically: for example, the one-particle density matrix for the free fermion gas ρ ff N (x) = 1 L exp{ ixk j } (1.139) is totally different from that of the Tonks-Girardeau gas, Eqs. (1.137) and (1.138). 1.5. The Fredholm determinant We have seen that in the N-particle sector the density matrix (1.13) is expressed via the determinants of N N matrices, Eqs. (1.137) and (1.138). In the thermodynamic limit (1.4) these matrices becomes infinite-dimensional. Thus, an important object comes into play: the determinant of an infinite-dimensional matrix (it is called Fredholm determinant). This object appears regularly in the calculations of the correlation functions of the integrable models. We give in this section two alternative definitions of the Fredholm determinant. The first one is: Let V be an N N matrix with the entries V ab = V (k a, k b ). Let ( ) a k a = N 1 1, a = 0, 1,..., N 1. (1.140) Then the Fredholm determinant det(î + ˆV ) is defined as follows: ( det(î + ˆV ) = lim det N I + ) N N 1 V. (1.141) The matrix I on the right hand side of Eq. (1.141) is the N N identity matrix. We see that with increasing number of divisions, N, the possible values of k a, Eq. (1.140), fill the interval 1 k a 1 densely. The objects Î and ˆV on the left hand side of the Eq. (1.141) should thus be some integral operators with the kernels defined on [ 1, 1] [ 1, 1]. This can be readily seen from an alternative definition of the Fredholm determinant: det(î + ˆV ) = N=0 1 1 1 dk 1... N! 1 1 V (k 1, k 1 ) V (k 1, k N ) dk N det....., (1.14) V (k N, k 1 ) V (k N, k N ) where ˆV is a linear integral operator with the kernel V (k, p) defined on [ 1, 1] [ 1, 1]. The equivalence of the definitions (1.141) and (1.14) is proven, for example, in Ref. [15]. These definitions are somehow complementary: the former one is more suitable for a numerical analysis, while the latter one reveals better the analytic structure of the problem. 1

1.5.3 One-particle density matrix: thermodynamic limit Let us set up a convention first. The one-dimensional density ρ is controlled by the chemical potential µ in the infinite γ limit as given by Eq. (1.8): µ = πρ. (1.143) It follows from this equation that µ has a dimension of inverse length. In the rest of section 1.5 we measure distances in units of 1/ µ. This convention implies that the momentum is measured in units of the Fermi momentum k F, Eq. (1.9). The ground state and zero temperature thermodynamics of the Tonks-Girardeau gas are described in section 1... Using formulas from that section, write the thermodynamic limit of the Lenard s formula (1.137) in the form prescribed by the definition (1.141): ρ(x) ρ(0) = det(î + ˆV 1 + ˆV ) det(î + ˆV 1 ). (1.144) The kernels of the operators ˆV 1 and ˆV in Eq. (1.144) are given by the thermodynamic limit of Eq. (1.138): V 1 (k, p) = α sin x (k p) (1.145) π(k p) and V (k, p) = 1 exp{ ix (k + p)}. (1.146) These kernels are defined on [ 1, 1] [ 1, 1], in accordance with our convention formulated below Eq. (1.143). The parameter α is Note that ρ(0) = ρ; one has ρ = 1/π if measured in units of µ. α =. (1.147) 1.5.4 One-particle density matrix: asymptotics and numerics Since ρ(x) = ρ( x), we assume x 0 in this section. For short distances (x 1) one can easily find from Eqs. (1.144) and (1.14) that ρ(x) ρ(0) = 1 x 6 α x 3 π 18 + x4 10 + α 11x 5 +..., x 0. (1.148) π 700 For α = 0 this is simply the expansion of sin x/x. Notice that higher order terms that are not exhibited on the r.h.s. of Eq. (1.148), depend on α nonlinearly. Next, consider the long-distance expansion of Eq. (1.144). Compared to the shortdistance analysis, this is a more sophisticated task. The analysis carried out in Refs. [16, 17, 18] shows the power-law decay of the density matrix: ρ(x) ρ(0) = C [ 1 + 1 ( cos x 1 ) + 3 ] sin x, x (1.149) x 8x 4 16x3

1 0.8 Ρ x Ρ 0 0.6 0.4 0. 0 0 5 10 15 0 5 30 x Figure 4: Shown is the function ρ(x)/ρ(0) versus x. The density matrix ρ(x) is given by Eq. (1.144); in getting the plot the Fredholm determinants entering Eq. (1.144) were approximated by finite-dimensional matrices, Eq. (1.141), with N = 800. with relative corrections of the order of x 1. The constant C is given by C = πe 1/ 1/3 A 6 0.94, (1.150) with A = 1.8471... being Glaisher s constant. The result (1.149) confirms the Luttinger liquid theory predictions. For the curve in Fig. 4 we took N = 800 in using the representation (1.141) for the Fredholm determinants entering Eq. (1.144). For x large than 30 it is safe to use the asymptotic expansion (1.149) instead of the exact expression (1.144): the relative difference (x) between the exact expression and the leading term of the large x expansion is very small for x = 30 : (x) = ρ(x) C/ x ρ(x) (1.151) (30) 6 10 4. (1.15) so the subleading terms in expansion (1.149) are not relevant for most applications. On the other hand, the function 1/ x converges to zero very slowly and the knowledge of constant C entering the expansion (1.149) is required with a good precision. Thus, the exact expression (1.150) is very much appreciated. 1.5.5 The momentum distribution The momentum distribution function n(k) is given by n(k) = dx e ikx ρ(x), (1.153) where ρ(x) is the one-particle density matrix (1.13). Recall that x is measured in units of 1/ µ as defined below Eq. (1.143), and that ρ(x) = ρ( x). The momentum distribution is 3

1.5 n k 1 0.5 0 0 0.5 1 1.5 k Figure 5: Shown is the momentum distribution for the Tonks-Girardeau gas (solid curve) and for the free fermion gas (dashed curve). Both distributions are normalized by the condition (1.154). Dotted curve represents the large k asymptotics of the momentum distribution, Eq. (1.155). also symmetric, n(k) = n( k). It is normalized as follows 0 dk n(k) = 1. (1.154) We plot n(k) in Fig. 5. The density matrix ρ(x) for x 30 was calculated from the Fredholm determinant representation; for x 30 we have used the asymptotic expansion (1.149). This momentum distribution is compared with the momentum distribution of the free fermion gas. For the large momentum tail of n(k) one gets n(k) 4 1, k. (1.155) 3π k4 The 1/k 4 decay originates from the cusp of ρ(x) at x = 0 seen in the third term on the righthand side of Eq. (1.148). The coefficient 3/4π in Eq. (1.155) is taken from the Ref. [19]. It is useful to note that Ref. [19] contains more general result: the large momentum tail of n(k) decays as 1/k 4 at any coupling strength; the coefficient stands at 1/k 4 term was also found in the cited reference. It follows from the large x expansion (1.149) that the momentum distribution is divergent at small momenta, n(k) C, k 0. (1.156) πk where the coefficient C is given by Eq. (1.150). This singularity is predicted by the Luttinger liquid theory for any coupling strength γ, and is the manifestation of quasi-long-range order, that is, a quasicondensate. The singularity is integrable for any γ 0, therefore zero momentum state is not macroscopically occupied. Note, however, that a significant fraction of particles lies in low-momentum states: for the Tongs-Girardeau gas about 15% of particles have a momentum lower than 0.01 (in units of k F, as discussed below Eq. (1.143)). When γ lowers this fraction is increased. 4

Algebraic Bethe Ansatz The Quantum Inverse Scattering Method, and its descendant, Algebraic Bethe Ansatz, were proposed in late 70 s (several crucial works appeared in the years 1978 and 1979). One can consult chapters VI and VII of the book [5] for the list of references. Various relativistic and non-relativistic quantum field theory models in 1 + 1 (one space one time) dimension were analyzed. Given the time limitations of a single lecture, we will learn the method by using it to solve the spin-1/ Heisenberg chain. This is, probably, the simplest model which can be solved by the Algebraic Bethe Ansatz: it is non-relativistic (remember how careful one should treat infinite Dirac sea in the Luttinger model) and has finite-dimensional quantum space..1 Coordinate Bethe Ansatz for the spin-1/ XXZ Heisenberg chain The coordinate Bethe Ansatz for the spin-1/ XXZ Heisenberg spin chain is reviewed. It is similar to the coordinate Bethe Ansatz for the Lieb-Liniger model presented in section 1..1.1 Hamiltonian, eigenfunctions and spectrum Recall some basic notations. The Pauli spin matrices are ( ) ( ) 0 1 0 i σ (x) = σ (y) = 1 0 i 0 σ (z) = ( 1 0 0 1 ). (.1) It is often convenient to use σ ± instead of σ (x) and σ (y) : Useful multiplication formula σ (x) = σ (+) + σ ( ), σ (y) = i(σ (+) σ ( ) ) (.) σ (p) σ (q) = iϵ pqr σ (r) + δ pq I (.3) The tensor product for matrices a and b is defined as follows 1 : ( ) ( ) a 11 b 11 a 11 b 1 a 1 b 11 a 1 b 1 a11 a a b = 1 b11 b 1 = a 11 b 1 a 11 b a 1 b 1 a 1 b a 1 a b 1 b a 1 b 11 a 1 b 1 a b 11 a b 1. (.4) a 1 b 1 a 1 b a b 1 a b We will work with the operators living in the tensor product of M matrices. Periodic boundary conditions are imposed. The Hamiltonian of the XXZ chain is H = H 0 + M hs z, (.5) 1 When the type of the operator s product, ordinary or tensorial, is clear from the context, the symbol is usually omitted to lighten the formulas. Expression (.6) is an example, there σ m σ m+1 means σ m σ m+1. 5

H 0 = M m=1 Eigenfunctions of this Hamiltonian are ψ N (λ 1,..., λ N ) = 1 N! [ ] σ m (x) σ (x) m+1 + σ m (y) σ (y) m+1 + σ m (z) σ (z) m+1, (.6) M m 1,...,m N =1 S z = 1 M m=1 σ (z) m. (.7) χ N (m 1,..., m N λ 1,..., λ N )σ m ( ) 1 σ m ( ) N 0 (.8) where 0 is the ferromagnetic vacuum (all spins up), 0 = M. There exists an operator V performing an important unitary transformation VH(h)V 1 = H( h), V = M m=1 σ (z) m. (.9) Since σ (x) = and σ (x) σ ( ) = σ (+), the wave function V ψ N (λ 1,..., λ N ) is V ψ N (λ 1,..., λ N ) = 1 N! M m 1,...,m N =1 χ N (m 1,..., m N λ 1,..., λ N )σ (+) m 1 σ (+) m N 0, (.10) where 0 is the dual ferromagnetic vacuum (all spins down), 0 = M. Comparing Eqs. (.8) and (.10) we see that the first-quantized wave function is the same in these expressions. Therefore, it is sufficient to consider h > 0, since all the observables will have a definite parity with respect to h in our problem. In particular, the ground state energy is an even function of h, and the magnetization is an odd one. Another useful similarity transformation, which exists for even number of sites only, is UH(, h = 0)U 1 = H(, h = 0), U = M/ m=1 σ (z) m. (.11) Due to the existence of this transformation, some authors choose the + sign in front of the Hamiltonian (.5). We, however, will keep the sign throughout this text. In our further calculations M is assumed to be even. The function χ N is the Bethe ansatz wave function: χ N = [ N b>a 1 ] [ sgn(m b m a ) ( 1) Q exp i Q exp [ i ] m a p 0 (λ Qa ) a=1 N b>a 1 θ(λ Qb λ Qa ) sgn(m b m a ) ] (.1) The summation runs through all the permutations Q of the integers 1,,..., N. The anisotropy parameter is convenient to represent as follows = cos(η). (.13) 6

Since the model is defined on a lattice, the momentum operator should be defined via the shift operator. The eigenfunctions of the shift operator Q in the N-particle sector have the form Q N = p 0 (λ α ) (.14) α=1 Thus, one calls sometimes p 0 by a momentum density. For the parameters λ α the word rapidities us used. Unfortunately, the word quasi-momentum distribution is used for λ α as well, so one should be careful. Momentum density p 0 and scattering phase θ are expressed via rapidities as follows: and p 0 (λ) = i ln cosh(λ iη) cosh(λ + iη), p 0(0) = 0, (.15) θ(λ 1 λ ) = i ln sinh(iη + λ 1 λ ), θ(0) = 0. (.16) sinh(iη λ 1 + λ ) Note that p 0 (λ) and θ(λ) are antisymmetric, and θ(λ) π 4η as λ (.17) Periodic boundary conditions imposed onto the eigenfunctions of the Hamiltonian (.5) imply a set of nonlinear equations onto rapidities. These equations are called Bethe equations: [ ] M cosh(λα iη) N = ( 1) N 1 sinh(iη λ α + λ β ) cosh(λ α + iη) sinh(iη + λ α λ β ). (.18) β α They can be written in the logarithmic form Mp 0 (λ α ) + θ(λ α λ β ) = πn α, α = 1,..., N. (.19) β=1 Here n α are integers (half-integers) when N is odd (even). We stress that these equations do not contain magnetic field. The eigenvalues of the Hamiltonian (.5) are, of course, field-dependent: E N = ε 0 (λ α ) hm, (.0) α=1 where the one-particle bare dispersion ε 0 is the followin function of the quasi-momentum p 0 : ε 0 (λ) = sin(η) dp 0(λ) dλ Here sin (η) + h = cosh(λ + iη) cosh(λ iη) + h dp 0 (λ) dλ = 4 sin (η) = + h. (.1) cosh(λ) + cos(η) sin(η) cosh(λ) + cos(η). (.) We see that the state ψ N contains N particles with quasi-momenta p 0 (λ α ) (α = 1,..., N) and energies ε 0 (λ α ). These excitations are excitations over the ferromagnetic vacuum 0 and they are usually called spin waves. 7

.1. Particular case: XXX model The limiting case = 1 corresponds to the isotropic ferromagnetic. The ground state is the ferromagnetic vacuum, so the magnetization takes maximal possible value. For > 1 the ground state remains to be ferromagnetic. The point = 1 corresponds to the isotropic antiferromagnetic. (I) Consider = 1 (antiferromagnetic). To take this limit we perform the rescaling procedure λ δλ, η π δ, δ 0. (.3) The Bethe equations (.18) become (recall that M is even) ( ) M λα + i/ = λ α i/ The bare dispersion (.1) becomes N β α λ α λ β + i, α = 1,..., N. (.4) λ α λ β i ε 0 (λ) = λ + 1 4 + h. (.5) (II) Consider = 1 (ferromagnetic). To take this limit we perform the rescaling procedure λ δλ + i π, η δ, δ 0. (.6) The Bethe equations (.18) become the same as for the antiferromagnetic case, while the bare dispersion is ε 0 (λ) = λ + 1 + h. (.7) 4. Thermodynamics of the XXZ model (optional reading) The thermodynamic limit of the Bethe equations is taken. The procedure is similar to the one carried out for the Lieb-Liniger model, section 1...1 Lieb equation To begin with, note some important properties of the ground state. It can be proven that the parameters n α fill the symmetric interval around zero and the rapidities λ α are real in the ground state. It will be convenient to work with the density of particles in the rapidity space (we will sometimes call it the distribution function of λ α s): ρ(λ α ) = 1 M(λ α+1 λ α ), (.8) that is f(λ α ) = M (λ α+1 λ α )ρ(λ α )f(λ α ), (.9) α=1 α=1 8

where f(λ α ) is an arbitrary function. We consider now the limit of large number of sites M and large number of particles N : D = N M const, N, M. (.30) This limit will be called thermodynamic limit. One has in this limit (λ α+1 λ α )ρ(λ α )f(λ α ) α=1 Λ The integration limits are defined by the normalization condition D = Λ The Bethe equations (.19) become in the thermodynamic limit where πρ(λ) Λ Λ Λ Λ dλ ρ(λ)f(λ). (.31) dλ ρ(λ). (.3) dµ K(λ, µ)ρ(µ) = dp 0(λ) dλ. (.33) K(λ, µ) = θ(λ µ) λ = sin(4η) sinh(λ µ + iη) sinh(λ µ iη) = sin(4η) cosh[(λ µ)] cos(4η). (.34) The magnetization σ is σ = σ (z) n = 1 D. (.35) Having calculated ρ(λ) one gets all the thermodynamics of the model. The ground state energy density ϵ can be found from Eq. (.0): ϵ = h + Λ Λ ϵ = E M dλ ρ(λ)ε 0 (λ) = h(1 D).. Interaction with magnetic field Λ Λ (.36) 4 sin (η) dλ ρ(λ) cosh(λ) + cos(η). (.37) We need to know how to calculate the ground state density at a given value of a chemical potential. We write ε(λ) 1 Λ dµ K(λ, µ)ε(µ) = ε 0 (λ), (.38) π Λ with the condition ε(±λ) = 0. (.39) 9

Ε0 Η 0 0.5 1 1.5.5 0 0.5 0.5 0.75 1 1.5 1.5 Η Figure 6: Shown is the ground state energy density ε(η) as a function of anisotropy η for h = 0. When h = 0 one has σ = 0 and D = 1/. The system of equations (.3) and (.33) can be solved analytically and gives ρ(λ) = [ ( )] 1 πλ (π η) cosh. (.40) π η The ground state at zero field is antiferromagnetic. The ground state has nontrivial structure in the region 0 h < h c, where h c = 4 sin η = (1 ). (.41) Above the critical field h c the ground state is ferromagnetic. The parameter Λ decays from infinity to zero when h is changing from 0 to h c. At h = 0 one can calculate an energy per site, ϵ, from Eqs. (.37) and (.40). For several values of it is given in the table. η 0 π 6 π 4 π 3 π = cos η 1 1 0 1 1 (.4) ϵ 0 3 3 4 π.77 A result for an arbitrary value of is shown in Fig. 6 An important characteristics of the system is the magnetic susceptibility ( ) σ χ =. (.43) h We obtain the curves of σ and χ at = 1/ solving the system of Eqs. (.3) and (.33) numerically. The technique and the results are discussed in a separate section. D 30