This paper was converted on www.awesomepapers.org from LaTeX by an anonymous user.
Want to know more? Visit the Converter page.

Captured molecules could make a Bose star visible

V. V. Flambaum [email protected]    I. B. Samsonov [email protected] School of Physics, University of New South Wales, Sydney 2052, Australia
Abstract

A Bose star passing through cold molecular clouds may capture atoms, molecules and dust particles. The observational signature of such an event would be a relatively small amount of matter that is gravitationally bound. This binding may actually be provided by invisible dark matter forming the Bose star. We may expect a relative excess of heavier atoms, molecules, and solid dust compared to the content of giant cold molecular clouds since the velocity of heavy particles at a given temperature is lower and it may be small compared to the escape velocity, vrms=3kBT/mgasvesc=2GM/Rv_{\mathrm{rms}}=\sqrt{3k_{\mathrm{B}}T/m_{\mathrm{gas}}}\ll v_{\mathrm{esc}}=\sqrt{2GM/R}. Finally, the velocity of this captured matter cloud may correlate with the expected velocity of free dark matter particles (e.g. expected axion wind velocity relative to Earth).

I Introduction

Bose stars are assumed to be formed from cold dark matter (DM) fields (scalars, axions, etc.) which may form Bose condensate and solitons bound mainly by the gravitational attraction [1, 2]. Bose condensation temperature is TBn2/3/mT_{B}\propto n^{2/3}/m, so in thermal equilibrium axions (or scalars) form Bose condensate for sufficiently small mass mm and large density nn. Bose stars may comprise a significant fraction of dark matter which may be potentially detectable, as will be argued in this work.

A simple estimate, similar to that for hydrogen atom, makes relation between mass MNmM\approx Nm and size RR of Bose star, where NN is the number of bosons and mm is their mass. Estimating kinetic energy of the ground state using uncertainty relation for minimal momentum p/rp\sim\hbar/r and estimating gravitational energy we obtain equation for the non-relativistic energy of Bose star:

E=a1N2mr2a2N(N1)2Gm2r,E=a_{1}N\frac{\hbar^{2}}{mr^{2}}-a_{2}\frac{N(N-1)}{2}\frac{Gm^{2}}{r}, (1)

where a1a_{1}, a2a_{2} are constants 1\sim 1. Minimizing energy by varying size rr or using virial theorem (U=2EkU=-2E_{k}, E=EkE=-E_{k}) we obtain size of gravitationally bound Bose star

R=4a12a2(N1)Gm3,R=\frac{4a_{1}\hbar^{2}}{a_{2}(N-1)Gm^{3}}, (2)

binding energy

E=N(N1)2a22G2m516a12,E=N(N-1)^{2}\frac{a_{2}^{2}G^{2}m^{5}}{16a_{1}\hbar^{2}}, (3)

and number density

nNV0.004N(N1)3G3m96,n\sim\frac{N}{V}\sim 0.004N(N-1)^{3}\frac{G^{3}m^{9}}{\hbar^{6}}, (4)

where the star volume is V4πR3/3V\sim 4\pi R^{3}/3. We see very strong dependence on the boson mass mm. We also see that the density nn in the ground state is proportional to N4N^{4} so any, even small non-gravitational self interaction between bosons may become important for large NN. We may add the self-interaction contribution of the self-interaction term λϕ4\lambda\phi^{4} to the Bose star energy

E=a1N2mr2a2N(N1)2Gm2r+a3λN(N1)2m2r3,E=a_{1}N\frac{\hbar^{2}}{mr^{2}}-a_{2}\frac{N(N-1)}{2}\frac{Gm^{2}}{r}+a_{3}\lambda\frac{N(N-1)}{2m^{2}r^{3}}, (5)

where a1a_{1}, a2a_{2} a3a_{3} are constants 1\sim 1. Using dE/dr=0dE/dr=0, this expression gives quadratic equation for rr with two solutions. For the QCD axion, the selfinteraction constant λ=m2/(24fa2)\lambda=-m^{2}/(24f_{a}^{2}) is expressed via the axion decay constant faf_{a} (see e.g. [3, 4]). In the case of λ<0\lambda<0 the energy tends to minus infinity for r0r\to 0, so the Bose star is metastable. Solution with larger rr corresponds to a minimum whereas the smaller rr solution is maximum. More accurate analytical calculations using variational approach will be given in the next section.

Distribution of gas particles of ordinary matter is given by Boltzmann distribution in the gravitational field of Bose star. Velocity distribution is given by Maxwell distribution. We can add molecular matter terms EmE_{m} to the energy EE:

Em=\displaystyle E_{m}= a4Nm52kT(r)a5NMNGmmmr\displaystyle a_{4}N_{m}\frac{5}{2}kT(r)-a_{5}N_{M}N\frac{Gm_{m}m}{r} (6)
a6Nm(Nm1)2Gmm2r.\displaystyle-a_{6}\frac{N_{m}(N_{m}-1)}{2}\frac{Gm_{m}^{2}}{r}\,.

If the total mass of the molecular cloud bound to the Bose star, Mm=NmmmM_{m}=N_{m}m_{m}, is relatively small as compared with the mass of the Bose star, NmmmNmN_{m}m_{m}\ll Nm, we may assume that in the leading approximation molecular distribution does not affect Bose star parameters. Then we may include gravitational effect of molecules using iterations.

Time of formation of Bose stars strongly depends on the DM particle mass mm. For m>1022m>10^{-22} eV the formation time is smaller than the lifetime of the Universe 13.7 billion years (see e.g. [3]).

Total mass of the Bose star MM is strongly model dependent and depends on time. There are estimates of a typical scale for expected mass MM. The QCD axion dark matter field with axion mass m105m\sim 10^{-5} eV makes Bose star with mass M1015MM\sim 10^{-15}M_{\odot} [3, 5, 6, 7]. Fuzzy dark matter with particle mass m1022m\sim 10^{-22} eV may compose a Bose star with mass M108MM\sim 10^{8}M_{\odot} [8, 9]. Bose star accumulating certain critical mass becomes unstable, collapses and decays to relativistic jets [10].

Axions and scalars are neutral, so in zero approximation Bose star is transparent and is hardly observable by telescopes. In principle, Bose stars may be observed using gravitational lensing of passing light and axion decays to photons. We discuss a different possibility: Bose star may capture atoms, molecules and dust. Radiation and absorption of light by this matter could make Bose stars visible.

II Variational approach to the Bose star solution

The variational approach allows one to get an approximate analytic solution to the particle density which minimizes total energy of the system of interacting particles. We consider NN bosonic DM particles with position vectors ri\vec{r}_{i} and momenta pi\vec{p}_{i}. Given that each particle has mass mm, they can interact gravitationally. The corresponding non-relativistic Hamiltonian is

H=22mi=1Nk2i<jNGm2rij,H=-\frac{\hbar^{2}}{2m}\sum_{i=1}^{N}\nabla_{k}^{2}-\sum_{i<j}^{N}\frac{Gm^{2}}{r_{ij}}\,, (7)

where GG is the gravitational constant and rij|rirj|r_{ij}\equiv|\vec{r}_{i}-\vec{r}_{j}|.

In the ground state, all particles are in the same state with the wave function ψ(ri)\psi(\vec{r}_{i}), and the full system is described by a product of these wave functions,

Ψ=i=1Nψ(ri).\Psi=\prod_{i=1}^{N}\psi(\vec{r}_{i})\,. (8)

Below we consider two approximate analytic solutions to the function ψ(r)\psi(\vec{r}).

II.1 Exponential ansatz

The ground state wave function may be searched for within the ansatz

ψ(r)=1πR3/2er/R,\psi(r)=\frac{1}{\sqrt{\pi}}R^{-3/2}e^{-r/R}\,, (9)

where RR is a free parameter. The function ψ(r)\psi(r) is the 1s solution of the stationary Schrödinger equation with the potential V=2mRrV=-\frac{\hbar^{2}}{mRr} and energy E0=22mR2E_{0}=-\frac{\hbar^{2}}{2mR^{2}}.

Making use of the function (8) we find the expectation value of the Hamiltonian (7):

E=Ψ|H|Ψ=N2[2mR258Gm2R(N1)].E=\langle\Psi|H|\Psi\rangle=\frac{N}{2}\left[\frac{\hbar^{2}}{mR^{2}}-\frac{5}{8}\frac{Gm^{2}}{R}(N-1)\right]. (10)

This energy reaches its minimum value

Emin=25512G2m52N(N1)2E_{\text{min}}=-\frac{25}{512}\frac{G^{2}m^{5}}{\hbar^{2}}N(N-1)^{2} (11)

at

R=1652Gm3(N1).R=\frac{16}{5}\frac{\hbar^{2}}{Gm^{3}(N-1)}\,. (12)

The energy of the solution (11) agrees at 90% level with the corresponding numerical result in Ref. [1].111Numerical solution of the Schrödinger-Poisson equation for the gravitating system on NN bosons was first presented in Ref. [1]. The energy of this system in the large NN limit was calculated to E=0.1626N3G2m52E=-0.1626N^{3}G^{2}m^{5}\hbar^{-2} which was factor 3 off the correct result E=0.05426N3G2m52E=-0.05426N^{3}G^{2}m^{5}\hbar^{-2} as was pointed out in Ref. [11]. The 10% discrepancy of our solution (11) from the numerical result is still acceptable accuracy for qualitative study of this system considering the simplicity of the trial function (9). For the Gaussian ansatz (16) the error is 5 times smaller.

The root-mean-squared (RMS) radius of this solution is

rrms=(Ψ|ri2|Ψ)1/2=3R.r_{\text{rms}}=\left(\langle\Psi|r_{i}^{2}|\Psi\rangle\right)^{1/2}=\sqrt{3}R\,. (13)

This quantity characterizes the size of the Bose star. We will use rrmsr_{\mathrm{rms}} as the effective boundary of the Bose star in subsequent sections.

As an example, consider a Bose star with mass of Bennu asteroid, M=7×1010M=7\times 10^{10} kg composed of QCD axions with mass m=105m=10^{-5} eV. It consists of N=4×1051N=4\times 10^{51} particles. Its size is

rrms=4×1010km.r_{\text{rms}}=4\times 10^{10}\,\mbox{km}. (14)

Note that the diameter of the Bennu asteroid is 0.5 km.

The particles in the obtained solution have a distribution of velocities. An important characteristic of this distribution is the root-mean-squared speed, which represents a typical speed of particles in the Bose star,

vrms=1m(Ψ|pi2|Ψ)1/2=mR.v_{\text{rms}}=\frac{1}{m}\left(\langle\Psi|\vec{p}_{i}^{2}|\Psi\rangle\right)^{1/2}=\frac{\hbar}{mR}\,. (15)

For the axion star with mass M=7×1010M=7\times 10^{10} considered above, the RMS speed is vrms8×1016cv_{\mathrm{rms}}\approx 8\times 10^{-16}c. Thus, the non-relativistic approximation is justified. However, for solutions with boson mass m105m\sim 10^{-5} eV and total mass of order the Solar mass, a relativistic treatment is needed.

II.2 Gaussian ansatz

The simple exponential ansatz (9) for the wave function provides a rather low accuracy for the energy. A better accuracy may be achieved within the Gaussian ansatz [12]

ψ(r)=(2πR2)3/4er2/R2,\psi(r)=\left(\frac{2}{\pi R^{2}}\right)^{3/4}e^{-r^{2}/R^{2}}\,, (16)

where RR is the parameter to be found. Making use of this wave function, we find the expectation value of the Hamiltonian (7):

E=H=3N22mR2N(N1)Gm2πR.E=\langle H\rangle=\frac{3N\hbar^{2}}{2mR^{2}}-\frac{N(N-1)Gm^{2}}{\sqrt{\pi}R}\,. (17)

This energy reaches its minimum value

Emin=N(N1)26πG2m52E_{\text{min}}=-\frac{N(N-1)^{2}}{6\pi}\frac{G^{2}m^{5}}{\hbar^{2}} (18)

at

R=3π2(N1)Gm3.R=\frac{3\sqrt{\pi}\hbar^{2}}{(N-1)Gm^{3}}\,. (19)

The energy (18) agrees at 98% level with the numerical solution presented in Ref. [1] (see the footnote on page 2). Therefore, we will consider only the Gaussian ansatz in what follows.

The RMS radius calculated with the wave function (16) is

rrms=3R2=33π22(N1)Gm3,r_{\text{rms}}=\frac{\sqrt{3}R}{2}=\frac{3\sqrt{3\pi}\hbar^{2}}{2(N-1)Gm^{3}}\,, (20)

and the RMS speed is

vrms=3mR.v_{\text{rms}}=\frac{\sqrt{3}\hbar}{mR}\,. (21)

The non-relativistic approximation is applicable for vrms0.1cv_{\mathrm{rms}}\lesssim 0.1c. Making use of equations (19) and (21), this relation translates into the constraint on the particle number and mass,

Nm2<0.1c3πG1.5×1010g2.Nm^{2}<\frac{0.1\hbar c\sqrt{3\pi}}{G}\approx 1.5\times 10^{-10}\,\mathrm{g}^{2}\,. (22)

II.3 Gravitational potential

The Bose star solution is given by the product of NN identical wave functions (16) each describing a particle with mass mm. They create a gravitational potential V(r)V(r) obeying the Poisson equation

2V(r)=4πGρ(r),\nabla^{2}V(r)=-4\pi G\rho(r)\,, (23)

where the mass density is

ρ=Nmψψ=Nm(2πR2)3/2e2r2/R2.\rho=Nm\psi^{\dagger}\psi=Nm\left(\frac{2}{\pi R^{2}}\right)^{3/2}e^{-2r^{2}/R^{2}}\,. (24)

The solution to Eq. (23) reads

V(r)\displaystyle V(r) =GNmerf(2r/R)r\displaystyle=-GNm\frac{\mathrm{erf}(\sqrt{2}r/R)}{r}
GNm{22πR42r23πR3(rR),1r(rR).\displaystyle\approx-GNm\left\{\begin{array}[]{ll}\frac{2\sqrt{2}}{\sqrt{\pi}R}-\frac{4\sqrt{2}r^{2}}{3\sqrt{\pi}R^{3}}&(r\ll R)\,,\\ \frac{1}{r}&(r\gg R)\,.\end{array}\right. (27)

On the boundary r=rrmsr=r_{\mathrm{rms}}, the gravitational potential is

V(rrms)=2GNm3Rerf(3/2)0.2N2G2m42.V(r_{\mathrm{rms}})=-\frac{2GNm}{\sqrt{3}R}\mathrm{erf}\left(\sqrt{3/2}\right)\approx-0.2N^{2}\frac{G^{2}m^{4}}{\hbar^{2}}\,. (28)

Thus, the escape velocity on the boundary is

vesc=2|V(rrms)|0.6NGm2/.v_{\mathrm{esc}}=\sqrt{2|V(r_{\mathrm{rms}})|}\approx 0.6NGm^{2}/\hbar\,. (29)

Consider, for example, the QCD axion with mass m=105m=10^{-5} eV and the total mass comparable with the mass of the Earth, Nm=MNm=M_{\oplus}. The size of this object is about rrms0.4r_{\mathrm{rms}}\approx 0.4 m. The corresponding escape velocity at the boundary is vesc0.14cv_{\mathrm{esc}}\approx 0.14c.

If the DM particle mass is m=109m=10^{-9} eV, the corresponding RMS radius and escape velocity are rrms6.3Rr_{\mathrm{rms}}\approx 6.3R_{\oplus}, vesc4.3v_{\mathrm{esc}}\approx 4.3 km/s.

II.4 Adding λϕ4\lambda\phi^{4} selfinteraction

The scalar field ϕ\phi is related to the wave function (8) as

ϕ=1mΨ.\phi=\frac{1}{\sqrt{m}}\Psi\,. (30)

Making use of the Gaussian ansatz (16), we find

λϕ4d3r=λπ3/2m2R3.\lambda\int\phi^{4}d^{3}r=\frac{\lambda}{\pi^{3/2}m^{2}R^{3}}\,. (31)

In the contribution to the energy, this expression should be multiplied by N(N1)/2N(N-1)/2 which accounts for the pairwise particle interaction. Thus, the total energy of this field configuration reads

E\displaystyle E =Ekin+Egrav+Eλ,\displaystyle=E_{\text{kin}}+E_{\text{grav}}+E_{\lambda}\,, (32a)
Ekin\displaystyle E_{\text{kin}} =3N22mR2,\displaystyle=\frac{3N\hbar^{2}}{2mR^{2}}, (32b)
Egrav\displaystyle E_{\text{grav}} =N(N1)Gm2πR,\displaystyle=-\frac{N(N-1)Gm^{2}}{\sqrt{\pi}R}\,, (32c)
Eλ\displaystyle E_{\lambda} =N(N1)2λ3π3/2cm2R3.\displaystyle=\frac{N(N-1)}{2}\frac{\lambda\hbar^{3}}{\pi^{3/2}cm^{2}R^{3}}\,. (32d)

This expression reaches its extrema at

R±=3π22Gm3(N1)(1±1+2λGm2(N1)23π2c).R_{\pm}=\frac{3\sqrt{\pi}\hbar^{2}}{2Gm^{3}(N-1)}\left(1\pm\sqrt{1+\frac{2\lambda Gm^{2}(N-1)^{2}}{3\pi^{2}\hbar c}}\right). (33)

The case λ<0\lambda<0 corresponds to unstable solution because the energy (32) is unbounded from below in this case. However, there is a potential barrier separating energy minimum at R+R_{+} in Eq. (33) from the collapse area close to r=0r=0, with the barrier maximum at RR_{-}. In the case of small negative λ\lambda, |λ|π2c/[Gm2(N1)2]|\lambda|\ll\pi^{2}\hbar c/[Gm^{2}(N-1)^{2}], we obtain minimum of energy

Emin=N(N1)26πG2m52+O(λ)E_{\text{min}}=-\frac{N(N-1)^{2}}{6\pi}\frac{G^{2}m^{5}}{\hbar^{2}}+O(\lambda) (34)

at

R=3π2Gm3(N1)+λ(N1)2π3/2mc+O(λ2),R=\frac{3\sqrt{\pi}\hbar^{2}}{Gm^{3}(N-1)}+\frac{\lambda\hbar(N-1)}{2\pi^{3/2}mc}+O(\lambda^{2})\,, (35)

and maximum at small RλR\propto\lambda,

R=λ(N1)2π3/2mc+O(λ2).R=-\frac{\lambda\hbar(N-1)}{2\pi^{3/2}mc}+O(\lambda^{2})\,. (36)

In the case of small λ\lambda this barrier will be sufficiently high to make the decay time of the metastable Bose star exponentially long, making this kind of Bose star potentially observable. This result is similar to the gravitational instability of scalar fields discussed in Ref. [13].

For positive selfinteraction constant, λ>0\lambda>0, only “+” solution in Eq. (33) is physical. If λ\lambda is sufficiently small, the leading contribution has the same form as in Eq. (35). For strong selfinteraction, 2λGm2(N1)23π2c1\frac{2\lambda Gm^{2}(N-1)^{2}}{3\pi^{2}\hbar c}\gg 1, the parameter RR is independent of NN,

R=3/2m23λ2πcG.R=\frac{\hbar^{3/2}}{m^{2}}\sqrt{\frac{3\lambda}{2\pi cG}}\,. (37)

For an illustration of the above results, let us introduce an unitless energy of the solution E^=E/E0\hat{E}=E/E_{0}, where E0=mNc2E_{0}=mNc^{2} is rest energy of the Bose star. Making use of Eqs. (32) this unitless energy may be represented as follows

E^=321R^2G^πR^+λ^2π3/2R^3,\hat{E}=\frac{3}{2}\frac{1}{\hat{R}^{2}}-\frac{\hat{G}}{\sqrt{\pi}\hat{R}}+\frac{\hat{\lambda}}{2\pi^{3/2}\hat{R}^{3}}\,, (38)

where we have introduced the following unitless parameters:

R^=mcR,G^=NGm2c,λ^=Nλ.\hat{R}=\frac{mcR}{\hbar}\,,\quad\hat{G}=\frac{NGm^{2}}{\hbar c}\,,\quad\hat{\lambda}=N\lambda\,. (39)

For instance, consider a Bose star solution with total mass M=MM=M_{\oplus} composed of particles with mass m=105m=10^{-5} eV. In this case, G^0.2\hat{G}\approx 0.2, and the typical behaviour of the unitless energy (38) is presented in Fig. 1 for various values of λ^\hat{\lambda} ranging from -40 to 1. This function has a shallow minimum near R^+3π/G^24\hat{R}_{\mathrm{+}}\approx 3\sqrt{\pi}/\hat{G}\approx 24, and a finite potential barrier for negative λ^\hat{\lambda}. For positive λ^\hat{\lambda} this barrier becomes infinitely high.

Note that for the QCD axion, ϕ\phi, the potential reads V(ϕ)=m2fa2[1cos(ϕ/fa)]=12m2ϕ2m24!fa2ϕ4+V(\phi)=m^{2}f_{a}^{2}[1-\cos(\phi/f_{a})]=\frac{1}{2}m^{2}\phi^{2}-\frac{m^{2}}{4!f_{a}^{2}}\phi^{4}+\ldots, where faf_{a} is the axion decay constant. The latter is related to the axion mass, fa/TeV6 keV/mf_{a}/\mbox{TeV}\simeq 6\mbox{\,keV}/m, and the values fa109f_{a}\lesssim 10^{9} GeV are excluded, see, e.g., [14]. Thus, for the QCD axion, the selfinteraction constant is small and negative,

λQCD=m24!fa21053.\lambda_{\mathrm{QCD}}=-\frac{m^{2}}{4!f_{a}^{2}}\simeq-10^{-53}\,. (40)

The full axion potential V(ϕ)V(\phi) is, however, non-negative.

Refer to caption
Figure 1: Plot of the unitless energy E^\hat{E} as a function of the unitless size parameter R^\hat{R} for different values of the selfinteraction constant λ^=Nλ\hat{\lambda}=N\lambda. This function has a shallow minimum near R^=24\hat{R}=24 and a finite potential barrier for negative λ^\hat{\lambda}.

III Ideal gas distribution around the Bose star

Let us consider an ideal gas with particles of mass mgasm_{\mathrm{gas}} in the gravitational potential V(r)<0V(r)<0 at temperature TT. Assuming that the total mass of the gas is sufficiently small so that the gravitational potential is not changed significantly, we can write the density distribution in the gas cloud,

ρ(r)=ρ0exp(mgasV(r)kBT).\rho(r)=\rho_{0}\exp\left(\frac{m_{\mathrm{gas}}V(r)}{k_{\mathrm{B}}T}\right)\,. (41)

The total mass of the gas cloud is

Mgas=4πρ00exp(mgasV(r)kBT)r2𝑑r.M_{\mathrm{gas}}=4\pi\rho_{0}\int_{0}^{\infty}\exp\left(\frac{m_{\mathrm{gas}}V(r)}{k_{\mathrm{B}}T}\right)r^{2}dr\,. (42)

The gas at every point should follow the Maxwell-Boltzmann distribution

f(v)=4πv2(mgas2πkBT)3/2emgasv22kBT.f(v)=4\pi v^{2}\left(\frac{m_{\mathrm{gas}}}{2\pi k_{\mathrm{B}}T}\right)^{3/2}e^{-\frac{m_{\mathrm{gas}}v^{2}}{2k_{\mathrm{B}}T}}\,. (43)

This distribution, however, should be truncated at v=vescv=v_{\mathrm{esc}}. Note that vrms=3kBT/mgasv_{\mathrm{rms}}=\sqrt{3k_{\mathrm{B}}T/m_{\mathrm{gas}}}.

At the Earth’s surface, the temperature is about T=300T=300 K. Empirically, we find the condition for the gas molecules to be captured by the gravitational field of the Earth:

vrms0.1vesc.v_{\mathrm{rms}}\lesssim 0.1v_{\mathrm{esc}}\,. (44)

This corresponds to the escape probability suppressed by the exponential factor in the distribution (43) exp(32vesc2vrms2)exp(150)\exp{(-{\frac{3}{2}\frac{v_{\mathrm{esc}}^{2}}{v_{\mathrm{rms}}^{2}}})}\lesssim\exp{(-150)}. For the Earth, this condition implies that the gas molecules with mass number A<6A<6 escape from the gravitational attraction. This is confirmed by the fact that the fractions of hydrogen and helium gases in the Earth’s atmosphere are very low.

Below we extend the condition (44) to Bose star solutions. We will consider separately the cases of pure gravitational interaction and a strong selfinteraction between DM particles.

III.1 Pure gravitational interaction

In the case of Bose stars, making use of Eq. (29) the condition (44) may be cast in the form

mgas830kBT2N2G2m4,m_{\mathrm{gas}}\gtrsim\frac{830k_{\mathrm{B}}T\hbar^{2}}{N^{2}G^{2}m^{4}}\,, (45)

Representing the gas particle mass as mgas=Ampm_{\mathrm{gas}}=Am_{p}, with mpm_{p} the proton mass, we find the condition on the atomic number AA of gas particles captured near the surface of the Bose star,

A\displaystyle A\gtrsim 830kBT2N2G2m4mp\displaystyle\frac{830k_{\mathrm{B}}T\hbar^{2}}{N^{2}G^{2}m^{4}m_{p}}
=1.351029(MM)2(eV/c2m)2(T10K).\displaystyle=1.35\cdot 10^{-29}\left(\frac{M_{\odot}}{M}\right)^{2}\left(\frac{\mathrm{eV}/c^{2}}{m}\right)^{2}\left(\frac{T}{10K}\right)\,. (46)
Refer to caption Refer to caption
(a) (b)
Figure 2: Relation between the mass and radius of the Bose star MM (in units of solar mass and radius) and the scalar particle mass mm, assuming that this object can gravitationally bound gas particles with atomic mass numbers AA exceeding 1, 10 and 100. The gas is assumed to have temperature T=10T=10 K. Left panel (a): pure gravitational interaction, λ=0\lambda=0; right panel (b): strong repulsive interaction between the DM particles, λ=1053\lambda=10^{-53}.

The temperature in cold molecular clouds in our Galaxy is typically in the range from 10 to 20 K. Assuming T=10T=10 K, we estimate the possibility of capture of gases in the fuzzy dark matter with m=1022m=10^{-22} eV and total mass of the object M=Nm=107MM=Nm=10^{7}M_{\odot}:

A>13.A>13\,. (47)

Thus, this object can accumulate only heavy atoms and molecules, as well as dust particles.

In Fig. 2a, we present plots of mass (and radius) of the Bose star as a function of scalar particle mass mm which can capture molecules with atomic mass numbers A=1A=1, 10, and 100.

If radius RR of Bose star is small, there will be too little matter inside the star and impossible to observe it. However, matter may be captured to distant orbits, but this will require bigger mass MM of the star, since the escape velocity will be smaller; for matter at distance rr, the required star mass is M(r)=M0R/rM(r)=M_{0}R/r, where M0M_{0} is the star mass required to capture matter inside.

Cold molecular gas and dust may also be captured by large axion clusters.

If matter capture conditions are fulfilled inside Bose star for all AA, there will be violation of such conditions outside Bose star at some distance rr from the centre. Firstly, this condition will be violated for A=1A=1, going further, for higher AA. Therefore, there will be separation of heavy and light elements outside the Bose star. This is important for Bose star of large mass and small radius.

The results of this subsection qualitatively apply to the case of attractive self-interaction with small negative λ\lambda. In this case the size parameter RR receives a small correction (35) which may be neglected in the result (46). The case of strong attractive interaction is not interesting within the present work because it leads to collapse of the boson star to the center and raises the questions of survival of spatially non-compact solutions.

III.2 Strong repulsive self-interaction

In the case of dominant self-interaction with positive coupling constant λ>0\lambda>0, the size parameter RR is given by Eq. (37). Near the boundary of the Bose star, r=rrmsr=r_{\mathrm{rms}}, the repulsive self-interaction energy (32d) becomes comparable with the gravitational attraction one (32b), but it quickly drops with distance by the law 1r3\frac{1}{r^{3}}. Therefore, we can still use Eqs. (28) and (29), but with the size parameter RR defined by Eq. (37). As a result, we find the following expression for the escape velocity,

vesc1.8N1/2G3/4m3/2λ1/4c1/43/4.v_{\mathrm{esc}}\simeq 1.8N^{1/2}G^{3/4}m^{3/2}\lambda^{-1/4}c^{1/4}\hbar^{-3/4}\,. (48)

Making use of the empirical relation (44) and the ideal gas root-mean-squared velocity vrms=3kBT/mgasv_{\mathrm{rms}}=\sqrt{3k_{B}T/m_{\mathrm{gas}}}, we find the condition of the capture of gas particles with mass mgasm_{\mathrm{gas}}:

mgas100kBTλ3G3/2m3Nc.m_{\mathrm{gas}}\gtrsim\frac{100k_{B}T\sqrt{\lambda\hbar^{3}}}{G^{3/2}m^{3}N\sqrt{c}}\,. (49)

This relation may be rewritten for the atomic number Amgas/mpA\equiv m_{\mathrm{gas}}/m_{p},

A4.7×1019(T10K)(MM)(eVm)2(λ1053)1/2.A\gtrsim 4.7\times 10^{-19}\left(\frac{T}{10K}\right)\left(\frac{M_{\odot}}{M}\right)\left(\frac{\mathrm{eV}}{m}\right)^{2}\left(\frac{\lambda}{10^{-53}}\right)^{1/2}\,. (50)

Note that in contrast with Eq. (46) this relation has different dependence on the total Bose star mass MM.

In Fig. 2b we present plots of mass of the Bose star as a function of scalar particle mass mm which can capture molecules with atomic mass numbers A=1A=1, 10 and 100 in the case of repulsive scalar particle self-interaction with the coupling λ=1053\lambda=10^{-53}.

IV Summary

We applied the analytic Gaussian solution for Bose star density to study the distribution of interstellar gas particles within the gravitational field of a Bose star, assuming the gas does not significantly perturb the star’s gravitational field. We examined the conditions under which interstellar gas particles with a given atomic number AA could escape the gravitational attraction of a Bose star with a total mass MM composed of scalar dark matter particles of mass mm. Our findings indicate that, for a range of values of mm and MM, the Bose star can capture atoms, molecules, and dust particles.

A small mass cloud of interstellar gas cannot form a gravitationally bound state without an additional gravitational source. Thus, the observation of such a gravitationally bound gas cloud would suggest it is held by the gravitational field of a Bose star or a dark matter cluster. Due to differences in escape velocities between light and heavy molecules, molecular clouds captured by a Bose star may exhibit higher metallicity compared to typical giant molecular clouds. The discovery of small mass, gravitationally bound molecular clouds with high metallicity in the interstellar medium would provide compelling evidence supporting this dark matter model. To the best of our knowledge, such objects have not been observed so far. The aim of this paper is to motivate the search of compact gas clouds with high metallicity in the interstellar medium of our galaxy. Radio telescopes such as ALMA or GBT and infrared telescopes (e.g., James Webb Space Telescope), which are used to study cold molecular clouds, could be employed in the search for such objects based on new measurements and the analysis of collected data.

Acknowledgements.— We are grateful to Marco Gorghetto, Maxim Khlopov, Mikhail Kozlov, Sergei Levshakov and John Webb for the interest in this work, useful discussions and references. The work was supported by the Australian Research Council Grants No. DP230101058 and No. DP200100150.

References