betaFeedback

WorksPhysics and Astronomy

The Thermodynamic Limit and the Landau degeneracy

Abstract

Usually, thermodynamic quantities are calculated by first regularizing the system in a confining box, which may have hard walls or periodic boundaries. One then studies the finite system and takes the thermodynamic limit by letting the box grow indefinitely. Here, we discuss a regularization scheme devised in the late 1980s [1] to study the thermodynamic equations of an anyon gas [2]. Instead of confining the system with walls, the method adds a smooth harmonic potential. Quantities are evaluated in its presence, and the potential is then gradually removed. A suitable prescription for taking this limit is needed to obtain meaningful thermodynamic results.

Introduction

We address the question of taking the thermodynamic limit by confining a many body quantum system in a harmonic well and then letting the strength of the well go to zero. First, as an illustration, we show how to recover the right scaling for the spectral degeneracy of a 2-dimensional particle coupled to a perpendicular magnetic field, the so called Landau degeneracy, without relying on semi-classical arguments. We then generalize these considerations to Riemann Zeta functions regularizations and their interpretation in quantum mechanics.

The thermodynamic limit prescription

Consider in full generality a quantum 1-body spectrum ϵi\epsilon_i, i=1,…,ki = 1,\ldots,k and the associated 1-body partition function

Zβ=∑i=1ke−βϵiZ_\beta= \sum_{i=1}^{k} e^{-\beta\epsilon_i}

The thermodynamical potential (i.e. the logarithm of the grand partition function which has to scale in the thermodynamic limit as the volume of the system) for a gas of bosons in such a 1-body spectrum is

log⁡∏i=1k11−ze−βϵi=∑n=1∞cnzn\log\prod_{i=1}^{k} \frac{1}{1-z e^{-\beta\epsilon_i}} = \sum_{n=1}^{\infty} c_n z^n

where the cluster expansion coefficients cnc_n have the simple form

cn=Znβn.c_n = \frac{Z_{n\beta}}{n}.

This is the case for many-body spectra obtained by filling independently 1-body eigenstates with occupation numbers nin_i, in the particular boson case ni=0,…,∞n_i = 0,\ldots,\infty (in the Fermi case ni=0,1n_i = 0,1 there would be an additional (−1)n−1(-1)^{n-1} factor).

Thermodynamic limit prescription

The thermodynamic limit prescription in dd dimensions arises from comparing the thermodynamical potential of a free gas of bosons1 confined in a dd dimensional harmonic well where ω\omega the strength of the well is understood to be small ω→0\omega\to0 to the thermodynamical potential of the same free gas of bosons now confined in a box of size LdL^d where LL a 1d length is understood to be large L→∞L \to\infty.

On the one hand for a gas of bosons in a dd dimensional harmonic well one has

Zβ(ω)=(eβωeβω−1)d⇒cn(ω)=Znβ(ω)n.Z_{\beta}(\omega) = \left(\frac{e^{\beta\omega}}{e^{\beta\omega}-1}\right)^d \Rightarrow c_n(\omega) = \frac{Z_{n\beta}(\omega)}{n}.

On the other hand for a gas of bosons in a dd dimensional box with in the large L→∞L \to\infty limit a 1-body density of states

ρd(ϵ)=1Γ(d/2)(L2π)dϵd/2−1(1)\rho_d(\epsilon) = \frac{1}{\Gamma(d/2)}\left(\frac{L}{\sqrt{2\pi}}\right)^d \epsilon^{d/2-1} \tag*{(1)}

where LdL^d stands for the dd-dimensional volume2 one has

Zβ(L)=∫0∞ρd(ϵ)e−βϵdϵ=(L2πβ)d⇒cn(L)=Znβ(L)n.Z_{\beta}(L) = \int_0^\infty\rho_d(\epsilon)e^{-\beta\epsilon}d\epsilon= \left(\frac{L}{\sqrt{2\pi\beta}}\right)^d \Rightarrow c_n(L) = \frac{Z_{n\beta}(L)}{n}.

In the thermodynamic limit ω→0\omega\to0 the harmonic well cluster coefficients become

lim⁡ω→0cn(ω)≃1n(1nβω)d.\lim_{\omega\to0} c_n(\omega) \simeq\frac{1}{n}\left(\frac{1}{n\beta\omega}\right)^d.

As already said they should coincide with the large box cluster coefficients cn(L)c_n(L) therefore

1n(1nβω)d⟶1n(L2πnβ)d\frac{1}{n}\left(\frac{1}{n\beta\omega}\right)^d \longrightarrow\frac{1}{n}\left(\frac{L}{\sqrt{2\pi n\beta}}\right)^d

It implies at order nn in the cluster expansion the thermodynamic limit prescription

1nβω⟶L2πβ(2)\frac{1}{\sqrt{n}\beta\omega} \longrightarrow\frac{L}{\sqrt{2\pi\beta}} \tag*{(2)}

which can be used for computing the thermodynamic equations of interacting quantum many body systems in any dimension.

The thermodynamic limit prescription (2) has been extensively used in 2d in the context of the lowest Landau level (LLL) LLL-anyon model [3] allowing for a proper LLL-anyonic gas thermodynamic which does interpolate between the LLL-Bose gas and LLL-Fermi gas limits. It has also been used [4] for the same purpose in 1d in the context of the Calogero model. Further one has been able to show that in the vanishing magnetic field limit the thermodynamic of the LLL-anyon model is mapped on those of the Calogero model [5] (for a review on all these efforts see [6]). As a remark this mapping has also been shown to hold at the level of the exact NN-body eigenstates [7].

This prescription has also been used in other many body quantum contexts [8], [9]. It has also been discussed on more formal grounds in [10].

From now on we will concentrate on the 1-body case, i.e., the n=1n = 1 cluster coefficient prescription, to illustrate the usefulness of the mechanism at hand by showing how it allows to retrieve in 2d the proper LLL degeneracy in a purely quantum context without relying on the usual semi-classical arguments. Also how the priori awkward B→0B \to0 limit in the LLL if properly taken allows for unveiling the above mentioned LLL-anyon/Calogero thermodynamics mapping. We will then generalize these considerations to Zeta functions at negative entries and their possible interpretation in a quantum mechanical context.

Application to the Landau degeneracy in the LLL

Landau problem in 2-dimensional plane

Consider a 2d charged particle ee in a perpendicular magnetic field BB with eB≥0eB \ge0 so that ωc=eB/2\omega_c = eB/2 is half the cyclotron frequency (we work in the units ℏ=m=1\hbar= m = 1). The 1-body Landau Hamiltonian

H=−2(∂−ωc2zˉ)(∂ˉ+ωc2z)+ωcH = -2\left(\partial- \frac{\omega_c}{2}\bar{z}\right)\left(\bar{\partial} + \frac{\omega_c}{2}z\right) + \omega_c

has the LLL eigenstates

ψl=zle−12ωczzˉ,l≥0\psi_l = z^l e^{-\frac{1}{2}\omega_c z\bar{z}}, \qquad l \ge0

where l≥0l \ge0 is the orbital momentum quantum number. The 1-body LLL energy

ωc\omega_c

is ll-independent and therefore infinitely degenerate. One well-known claim is that the infinite number of LLL eigenstates, the so-called Landau degeneracy, scales as the infinite surface SS of the plane

BSϕ0\frac{BS}{\phi_0}

i.e., as the flux of BB through the plane counted in unit of the flux quantum ϕ0=2π/e\phi_0 = 2\pi/e. We note that if in the LLL there is only one quantum number ll the system is still bidimensional as encoded in the surface scaling of the degeneracy.

This claim is traditionally supported by semi-classical/hand-waving arguments: consider the radius3 rl=2l/(eB)r_l = \sqrt{2l/(eB)} where the probability density ψlψˉl\psi_l\bar{\psi}_l is maximum, since πrl2B/ϕ0=l\pi r_l^2 B/\phi_0 = l, one can interpret ll as the flux of BB through the ll-orbit of radius rlr_l counted in unit of ϕ0\phi_0. Then when l→∞l \to\infty the total number of LLL-eigenstates acquires the meaning of being the flux of BB through the entire plane counted in unit of ϕ0\phi_0.

Let us now show how this scaling can be retrieved in a purely quantum mechanical context using solely the thermodynamical limit prescription discussed above. To achieve this goal one focuses on the LLL partition function

ZLLL=∑l=0∞e−βωc(3)Z_{\mathrm{LLL}} = \sum_{l=0}^{\infty} e^{-\beta\omega_c} \tag*{(3)}

What is the meaning of the infinite degenerate sum in (3)? Written as such it is an ambiguous, meaningless quantity. To remove this ambiguity as advocated above add a 2d harmonic well, compute the LLL partition function in the presence of the harmonic well, then take the thermodynamic limit by letting the well vanish.

In the presence of a harmonic well the LLL spectrum becomes the harmonic LLL spectrum

ωc→ωt+l(ωt−ωc)\omega_c \to\omega_t + l(\omega_t-\omega_c)

where ωt=ω2+ωc2\omega_t = \sqrt{\omega^2+\omega_c^2}. The LLL partition function becomes the harmonic LLL partition function

ZLLL→ZLLL,ω=∑l=0∞e−β(ωt+l(ωt−ωc))=e−βωt1−e−β(ωt−ωc)(4)Z_{\mathrm{LLL}} \to Z_{\mathrm{LLL},\omega} = \sum_{l=0}^{\infty} e^{-\beta(\omega_t+l(\omega_t-\omega_c))} = \frac{e^{-\beta\omega_t}}{1-e^{-\beta(\omega_t-\omega_c)}} \tag*{(4)}

Take now the thermodynamic limit ω→0\omega\to0 (i.e., strictly speaking βω→0\beta\omega\to0)

lim⁡ω→0ZLLL,ω≃e−βωcβ(ωt−ωc)≃e−βωcβω2/(2ωc)(5)\lim_{\omega\to0} Z_{\mathrm{LLL},\omega} \simeq\frac{e^{-\beta\omega_c}}{\beta(\omega_t-\omega_c)} \simeq\frac{e^{-\beta\omega_c}}{\beta\omega^2/(2\omega_c)} \tag*{(5)}

In (5) one has to give a meaning to the diverging 1/ω21/\omega^2 term. For this use the thermodynamic limit prescription (2) which says that when ω→0\omega\to0 the n=1n=1 cluster coefficient becomes

1(βω)2∼L22πβ(6)\frac{1}{(\beta\omega)^2} \sim\frac{L^2}{2\pi\beta} \tag*{(6)}

i.e., the 2d harmonic well partition function becomes the 2d free partition function. It follows that

lim⁡ω→0ZLLL,ω≃β(2ωc)L22πβe−βωc=eBL22πe−βωc\lim_{\omega\to0} Z_{\mathrm{LLL},\omega} \simeq\beta(2\omega_c)\frac{L^2}{2\pi\beta}e^{-\beta\omega_c}=eB\frac{L^2}{2\pi}e^{-\beta\omega_c}

If one ignores the harmonic well one ends up with

ZLLL=BSϕ0e−βωc(7)Z_{\mathrm{LLL}}=\frac{BS}{\phi_0}e^{-\beta\omega_c} \tag*{(7)}

and for the associated density of states

ρLLL(ϵ)=BSϕ0δ(ϵ−ωc)(8)\rho_{\mathrm{LLL}}(\epsilon)=\frac{BS}{\phi_0}\delta(\epsilon-\omega_c) \tag*{(8)}

The Landau degeneracy BS/ϕ0BS/\phi_0 where the surface S=L2S=L^2 of the plane indeed materialized, has been here rederived by purely quantum arguments. What happened is that the degenerate sum in (3) became tractable because a magnifying lens, i.e., the harmonic well, was applied, and when the thermodynamic limit was taken, thanks to the thermodynamic limit prescription, the proper expression was recovered with the right surface scaling.

2) B→0B \to0 limit: If now one wants to take the B→0B \to0 limit4 (i.e., strictly speaking βωc→0\beta\omega_c \to0) in (7) again we are facing the same type of difficulties, namely an ambiguous, meaningless expression: what means BS/ϕ0BS/\phi_0 when B→0B \to0 while S≃∞S \simeq\infty? Again apply the same kind of ideas as above: remove the ambiguity by adding a 2d harmonic well as a long distance regulator, i.e., instead of (3) use rather (4) where the B→0B \to0 limit can be done safely to obtain

lim⁡B→0ZLLL,ω=e−βω1−e−βω\lim_{B\to0} Z_{\mathrm{LLL},\omega}=\frac{e^{-\beta\omega}}{1-e^{-\beta\omega}}

Then take the thermodynamic limit ω→0\omega\to0

lim⁡ω→0lim⁡B→0ZLLL,ω≃1βω\lim_{\omega\to0}\lim_{B \to0} Z_{\mathrm{LLL},\omega} \simeq\frac{1}{\beta\omega}

and use the very same thermodynamic limit prescription (2) as already used in (6), here 1/(βω)→L/2πβ1/(\beta\omega) \to L/\sqrt{2\pi\beta}. It follows that

lim⁡ω→0lim⁡B→0ZLLL,ω≃L2πβ\lim_{\omega\to0}\lim_{B \to0} Z_{\mathrm{LLL},\omega} \simeq\frac{L}{\sqrt{2\pi\beta}}

where L/2πβL/\sqrt{2\pi\beta} is now the 1d free partition function Z1Z_1. If one ignores the harmonic well one ends up with

lim⁡B→0ZLLL=Z1⇔lim⁡B→0BSϕ0e−βωc=L2πβ(9)\lim_{B \to0} Z_{\mathrm{LLL}} = Z_1 \Leftrightarrow\lim_{B \to0} \frac{BS}{\phi_0}e^{-\beta\omega_c} = \frac{L}{\sqrt{2\pi\beta}} \tag*{(9)}

or equivalently with the LLL-density of states becoming the free 1d density of states ρ1(ϵ)\rho_1(\epsilon) in (1)

lim⁡B→0ρLLL(ϵ)=ρ1(ϵ)⇔lim⁡B→0BSϕ0δ(ϵ−ωc)=1Γ(1/2)L2πϵ−1/2(10)\lim_{B \to0}\rho_{\mathrm{LLL}}(\epsilon) = \rho_1(\epsilon) \Leftrightarrow\lim_{B \to0}\frac{BS}{\phi_0}\delta(\epsilon-\omega_c) = \frac{1}{\Gamma(1/2)}\frac{L}{\sqrt{2\pi}}\epsilon^{-1/2} \tag*{(10)}

The equations (9) and (10) written as such seem a bit awkward. They become meaningful if one keeps in mind they have been obtained via the harmonic well regularization. It amounts to say that the degenerate spectrum of a 2d particle in the LLL of an external magnetic field becomes in the limit where the field vanishes the continuous spectrum of a particle moving freely on a 1d line. Indeed one has followed the successive steps

ωc→ω≠0ωt+l(ωt−ωc)→B→0(l+1)ω→ω→0k22\omega_c \to_{\omega\ne0} \omega_t + l(\omega_t-\omega_c) \to_{B \to0} (l+1)\omega\to_{\omega\to0} \frac{k^2}{2}

where kk stands for the free momentum of a 1d particle on a line with free spectrum k2/2k^2/2. For ω\omega small the spectral gap is ω2/(2ωc)\omega^2/(2\omega_c) in the former B≠0B \ne0 case whereas it is ω\omega in the latter B→0B \to0 case. As far as the partition functions are concerned, moving from (3) to (7) then from (7) to (9) used the same harmonic well LLL partition function (4) and the same thermodynamic prescription (2) with the end results differing because of the different spectral gaps. Clearly a dimensional reduction 2d →\to 1d took place5. Indeed, coming back to LLL anyons, the claim that in the B→0B \to0 limit the LLL anyon thermodynamic becomes the Calogero thermodynamic is in fact encoded in (10) since the only input which differentiates them is the 1-body density of states of the said quantum models, the actual non trivial statistics due to the anyonic Aharonov-Bohm interactions in the LLL on the one hand and to the Calogero 1/x21/x^2 coupling constant on the other hand being encoded [4], [5], [6] in the same common transcendental equation.

Riemann Zeta function

As a remark let’s focus again on (7) where one has ended up claiming that

ZLLL=∑l=0∞e−βωc=(1+1+1+…)e−βωc=BSϕ0e−βωc(11)Z_{\mathrm{LLL}} = \sum_{l=0}^{\infty}e^{-\beta\omega_c} = (1+1+1+\ldots)e^{-\beta\omega_c} = \frac{BS}{\phi_0}e^{-\beta\omega_c} \tag*{(11)}

One can interpret (11) by saying that quantum mechanics in the LLL has given a meaning to the counting 1+1+1+…1+1+1+\ldots, thanks to the harmonic well regularization and the ensuing thermodynamic limit prescription, by unveiling that the regularized ZLLL,ωZ_{\mathrm{LLL},\omega} diverges in the thermodynamic limit ω→0\omega\to0 as the surface SS of the plane. Now the degenerate sum in (11), up to the factor e−βωce^{-\beta\omega_c} that we will from now on ignore as well as the l=0l=0 summand, is nothing but the Riemann Zeta function at zero entry

∑l=1∞1l0=1+1+1+⋯\sum_{l=1}^{\infty}\frac{1}{l^0}=1+1+1+\cdots

Indeed pushing to the next order in ω\omega the expansion of ZLLL,ωZ_{\mathrm{LLL},\omega}, accordingly redefined as a sum without the energy shift ωt\omega_t and starting at l=1l=1, one gets

ZLLL,ω=∑l=1∞1l0e−βl(ωt−ωc)=2ωcβω2−12+12βωc+O(ω2)(12)Z_{\mathrm{LLL},\omega}=\sum_{l=1}^{\infty}\frac{1}{l^0}e^{-\beta l(\omega_t-\omega_c)}=\frac{2\omega_c}{\beta\omega^2}-\frac{1}{2}+\frac{1}{2\beta\omega_c}+O(\omega^2) \tag*{(12)}

where on the one hand and as already discussed the leading term 2ωc/(βω2)→BS/ϕ02\omega_c/(\beta\omega^2)\to BS/\phi_0 yields the Landau degeneracy and on the other hand the zeroth order term −1/2+1/(2βωc)-1/2+1/(2\beta\omega_c) is, not surprisingly, the analytically continued value

ζ(0)=−1/2\zeta(0)=-1/2

here shifted by 1/(2βωc)1/(2\beta\omega_c). So (12) contains at least two informations of interest: on the one hand the actual diverging counting ∑l=1∞1l0=1+1+1+⋯\sum_{l=1}^{\infty}\frac{1}{l^0}=1+1+1+\cdots which scales6 as the infinite surface of the plane and on the other hand its analytically continued value ζ(0)=−1/2\zeta(0)=-1/2.

Riemann Zeta functions regularizations of this type as well as in other settings (see e.g., [12]) are well-known. What in addition we learn here is that the actual divergence in ∑l=1∞1/l0\sum_{l=1}^{\infty}1/l^0 can be interpreted in a LLL quantum mechanical context as a 2 dimensional space dimension, thanks to a proper thermodynamic limit prescription.

Following this line of reasoning we define now for non positive entries n≤0n\leq0 the generalized harmonic LLL partition functions

Z(n)=∑l=1∞1lne−βl(ωt−ωc)(13)Z(n)=\sum_{l=1}^{\infty}\frac{1}{l^n}e^{-\beta l(\omega_t-\omega_c)} \tag*{(13)}

as yet another quantum mechanics7 inspired way to regularize ∑l=1∞1/ln\sum_{l=1}^{\infty}1/l^n by taking in the harmonic LLL the ω→0\omega\to0 limit along the lines discussed above.

Now in quantum mechanics the factors l−nl^{-n} in (13) can be interpreted as the building blocks of quantum spectrum degeneracies stemming from adding (similarly to what is done to obtain the 2d,

lim⁡B→0ZLLL,ω=∑l=1∞e−βlω=1βω−12+O(ω)\lim_{B\to0}Z_{\mathrm{LLL},\omega}=\sum_{l=1}^{\infty}e^{-\beta l\omega}=\frac{1}{\beta\omega}-\frac{1}{2}+O(\omega)

where again not surprisingly the analytically continued value ζ(0)\zeta(0) materializes as it should but the diverging 1/βω→L/2πβ1/\beta\omega\to L/\sqrt{2\pi\beta} term cannot be interpreted as counting anything anymore. This type of Zeta function regularization has been extensively used in the context of the Casimir effect [11].

Z(n)=2n∑k=0k evenn−1A(k2,−n)cosh⁡((n+1+k)β(ωt−ωc)2)sinh⁡(β(ωt−ωc)2)−n+1Z(n)=2^n\frac{\sum_{\substack{k=0\\ k\ \mathrm{even}}}^{n-1}A\left(\frac{k}{2},-n\right)\cosh\left(\frac{(n+1+k)\beta(\omega_t-\omega_c)}{2}\right)}{\sinh\left(\frac{\beta(\omega_t-\omega_c)}{2}\right)^{-n+1}}

3d, …\ldots harmonic well spectra from the 1d harmonic well spectrum by adding it 2 times, 3 times, …\ldots, i.e., one spectrum for each dimension since the Hamiltonians in each space dimension add up) −n+1-n+1 times the 1-body harmonic LLL spectrum l(ωt−ωc)l(\omega_t-\omega_c), l≥1l\ge1, in which case the relevant partition function ends up being Z(0)−n+1Z^{(0)-n+1} instead of Z(n)Z^{(n)}. For example (l1+l2)(ωt−ωc)(l_1+l_2)(\omega_t-\omega_c), l1,l2≥1l_1,l_2\ge1 is identical to l(ωt−ωc)l(\omega_t-\omega_c), l≥2l\ge2 with degeneracy l−1l-1, (l1+l2+l3)(ωt−ωc)(l_1+l_2+l_3)(\omega_t-\omega_c), l1,l2,l3≥1l_1,l_2,l_3\ge1 is identical to l(ωt−ωc)l(\omega_t-\omega_c), l≥3l\ge3 with degeneracy (l−2)(l−1)/2(l-2)(l-1)/2, (l1+l2+l3+l4)(ωt−ωc)(l_1+l_2+l_3+l_4)(\omega_t-\omega_c), l1,l2,l3,l4≥1l_1,l_2,l_3,l_4\ge1 is identical to l(ωt−ωc)l(\omega_t-\omega_c), l≥4l\ge4 with degeneracy (l−3)(l−2)(l−1)/6(l-3)(l-2)(l-1)/6, etc. These degeneracies in turn materialize in the well-known identities

Z(0)2=Z(−1)−Z(0),Z(0)3=Z(−2)−3Z(−1)+2Z(0)2,Z(0)4=Z(−3)−6Z(−2)+11Z(−1)−6Z(0)6Z^{(0)2}=Z^{(-1)}-Z^{(0)},\quad Z^{(0)3}=\frac{Z^{(-2)}-3Z^{(-1)}+2Z^{(0)}}{2},\quad Z^{(0)4}=\frac{Z^{(-3)}-6Z^{(-2)}+11Z^{(-1)}-6Z^{(0)}}{6}

etc, with at zeroth order the combinations8 ζ(−1)−ζ(0)=5/12\zeta(-1)-\zeta(0)=5/12, (ζ(−2)−3ζ(−1)+2ζ(0))/2=−3/8(\zeta(-2)-3\zeta(-1)+2\zeta(0))/2=-3/8, (ζ(−3)−6ζ(−2)+11ζ(−1)−6ζ(0))/6=251/720(\zeta(-3)-6\zeta(-2)+11\zeta(-1)-6\zeta(0))/6=251/720, …\ldots Even though actual quantum Landau like Hamiltonians in higher dimensions yielding these spectra are missing, we still expect that in the same way as the Landau Hamiltonian leads via Z(0)Z^{(0)} to a Landau degeneracy 2d surface scaling, Z(−1)Z^{(-1)} should lead to a 4d volume scaling, Z(−2)Z^{(-2)} to a 6d volume scaling, etc, endowing the corresponding Zeta functions with a quantum mechanical space dimension interpretation, in addition to their analytical continued values.

Indeed in view of the ω→0\omega\to0 expansion Z(n)Z^{(n)} can easily be rewritten9 as

Z(n)=(−n)!1(β(ωt−ωc))1−n+∑k=0∞(−1)k(β(ωt−ωc))kk!ζ(n−k)Z^{(n)}=(-n)!\frac{1}{\left(\beta(\omega_t-\omega_c)\right)^{1-n}}+\sum_{k=0}^{\infty}(-1)^k\frac{\left(\beta(\omega_t-\omega_c)\right)^k}{k!}\zeta(n-k)

where the first term diverges and the analytically continued value ζ(n)\zeta(n) materializes as expected at zeroth order.

Interestingly enough the most diverging term tells us, using again the thermodynamic limit prescription (2), that in the ω→0\omega\to0 limit Z(n)Z^{(n)} scales like

(−n)!1(βωc22ωc)1−n→(−n)!(BL2ϕ0)1−n(-n)!\frac{1}{\left(\beta\frac{\omega_c^2}{2\omega_c}\right)^{1-n}}\to(-n)!\left(\frac{BL^2}{\phi_0}\right)^{1-n}

i.e., as expected, a d=2(1−n)d=2(1-n) space dimension stemming from the fact that, as discussed above, one has added 1−n1-n times the 2d LLL spectrum. It follows that a generalization of the LLL quantum and for nn odd as

Z(n)=2n∑k=0k even−n−3A(k2,−n)cosh⁡((n+1+k)β(ωt−ωc)2)+A(−n−12,−n)2sinh⁡(β(ωt−ωc)2)−n+1Z^{(n)}=2^n\frac{\sum_{\substack{k=0\\ k\ \mathrm{even}}}^{-n-3}A\left(\frac{k}{2},-n\right)\cosh\left(\frac{(n+1+k)\beta(\omega_t-\omega_c)}{2}\right)+\frac{A\left(\frac{-n-1}{2},-n\right)}{2}}{\sinh\left(\frac{\beta(\omega_t-\omega_c)}{2}\right)^{-n+1}}

where the A(k,n)A(k,n)'s are the Eulerian numbers

A(k,n)=∑j=0k(−1)j(n+1j)(−j+k+1)nA(k,n)=\sum_{j=0}^{k}(-1)^j\binom{n+1}{j}(-j+k+1)^n
Z(n)=ζ(n)+(−1)n(−n)!(β2ωc)1−n+(1−n)!(β2ωc)1−n∑k=11−n(2ωcω)2k(k1−n−k)k+∑k=1∞(−1)k(ω2ωc)2k(∑j=1k(2k−j−1k−j)k(j−1)!ζ(n−j)(β2ωc)j+(−1)n(1−n)!(β2ωc)1−n(−n+2k1−n+k)1−n+k)(14)\begin{aligned} Z^{(n)}={}&\zeta(n)+(-1)^n\frac{(-n)!}{(\beta2\omega_c)^{1-n}}+\frac{(1-n)!}{(\beta2\omega_c)^{1-n}}\sum_{k=1}^{1-n}\left(\frac{2\omega_c}{\omega}\right)^{2k}\frac{\binom{k}{1-n-k}}{k}\\ &+\sum_{k=1}^{\infty}(-1)^k\left(\frac{\omega}{2\omega_c}\right)^{2k}\left(\sum_{j=1}^{k}\frac{\binom{2k-j-1}{k-j}}{k(j-1)!}\zeta(n-j)(\beta2\omega_c)^j+(-1)^n\frac{(1-n)!}{(\beta2\omega_c)^{1-n}}\frac{\binom{-n+2k}{1-n+k}}{1-n+k}\right) \tag*{(14)} \end{aligned}

where at zeroth order ζ(n)\zeta(n) is shifted by (−1)n(−n)!/(β2ωc)1−n(-1)^n(-n)!/(\beta2\omega_c)^{1-n}.

mechanics has provided interpretation to the divergence in ∑l=1∞1/ln\sum_{l=1}^{\infty} 1/l^n in terms of volumes in higher spatial dimensions. It would certainly be interesting to push further the analysis by giving as well a meaning, if any, to the subleading diverging terms in (14).

Conclusions

We have presented a robust thermodynamic limit prescription based on regularizing quantum many-body systems with a smooth harmonic well regulator ω→0\omega\to0 rather than traditional boundary boxes L→∞L \to\infty. By matching the cluster expansion coefficients of both systems, we derived a precise thermodynamic limit prescription that connects the vanishing harmonic well frequency to an effective infinite spatial volume. This framework has allowed us to provide a purely quantum-mechanical derivation of the surface scaling for the lowest Landau level degeneracy without invoking traditional semi-classical or geometric orbital arguments. By modifying the LLL partition function to include generalized degeneracies, we demonstrated that the mathematical divergences of the Riemann Zeta function ∑l=1∞1/ln\sum_{l=1}^{\infty} 1/l^n at non positive integer entries carry direct physical interpretations. Specifically, the leading divergent terms scale as higher spatial volumes d=2(−n+1)d = 2(-n + 1), mapping number-theoretic regularizations to physical dimensions in quantum mechanics.

Future work will focus on understanding the physical meaning of the subleading divergent terms in the omega expansion, as well as unveiling, if any, explicit quantum Hamiltonians in higher spatial dimensions that intrinsically mirror these generalized spectra.

Acknowledgements

The author acknowledges useful discussions with Olivier Giraud and would like to thank Li Gan for his careful reading of the manuscript.

References

  1. [3]A. Dasnières de Veigy and S. Ouvry, “Equation of State of an Anyon gas in a Strong Magnetic Field,” Phys. Rev. Lett. 72 (1994) 600.
  2. [4]A. Dasnières de Veigy and S. Ouvry, “One-dimensional Statistical Mechanics for Identical Particles: the Calogero and Anyon Cases,” Mod. Phys. Lett. A 10 (1995) 1; Mod. Phys. Lett. B 9 (1995) 271.
  3. [5]S. Ouvry, “On the Relation between the Anyon and the Calogero Models,” Phys. Lett. B510 (2001) 335DOI
  4. [6]S. Ouvry, “Anyons and Lowest Landau Level Anyons,” Séminaire Poincaré “Le Spin!” (2007) arXiv:0712.2174, Birkhäuser Verlag AG.arxiv.org/abs/0712.2174
  5. [7]S. Ouvry and A.P. Polychronakos, “Mapping the Calogero model on the Anyon model,” Nucl. Phys. B 936 (2018) 189.DOI
  6. [8]Y. Castin, F. Werner, “Le troisième coefficient du viriel du gaz de Bose unitaire,” Rev. can. phys. 91, 382 (2013); S. Endo, Y. Castin, “The interaction-sensitive states of a trapped two-component ideal Fermi gas and application to the virial expansion of the unitary Fermi gas,” J. Phys. A: Math. Theor. 49, 265301 (2016); Y. Castin, “Questions ouvertes pour les gaz de fermions en interaction forte et de portée nulle/Open questions for strongly interacting Fermi gases with zero-range interactions,” Comptes Rendus Physique 26, 393-461 (2025); Chao Gao, Shimpei Endo, Y. Castin, “The third virial coefficient of a two-component unitary Fermi gas across an Efimov-effect threshold,” EPL 109, 16003 (2015); S. Endo, Y. Castin, “Unitary boson-boson and boson-fermion mixtures: third virial coefficient and three-body parameter on a narrow Feshbach resonance,” Eur. Phys. J. D 70, 238 (2016); S. Endo, Y. Castin, “Quatrièmes coefficients d’amas et du viriel d’un gaz unitaire de fermions pour un rapport de masse quelconque/Fourth cluster and virial coefficients of a unitary Fermi gas for an arbitrary mass ratio,” Comptes Rendus Physique 23, 41-110 (2022).DOI
  7. [9]Xia-Ji Liu, Hui Hu, P. D. Drummond, Phys. Rev. Lett. 102, 160401 (2009); Phys. Rev. A 82, 023619 (2010).
  8. [10]K. Olaussen, “On the Harmonic Oscillator Regularization of Partition Functions,” arXiv:cond-mat/9207005 (1992).DOI
  9. [11]Michael Bordag, U. Mohideen, V.M. Mostepanenko, “New developments in the Casimir effect,” Phys. Rept. 353,1-205 (2001).DOI
  10. [12]T. Tao, “The Euler-Maclaurin formula, Bernoulli numbers, the zeta function, and real-variable analytic continuation,” What’s new (2010).

Paper details

vlnc:W_4b2k5ves
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 11

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents