Abstract
Quark nuggets are a candidate for dark matter consistent with the Standard Model. Previous models of quark nuggets have investigated properties arising from their being composed of strange, up, and down quarks and have not included any effects caused by their selfmagnetic field. However, Tatsumi found that the core of a magnetar star may be a quark nugget in a ferromagnetic state with core magnetic field B_{surface} = 10^{12±1} T. We apply Tatsumi’s result to quarknugget darkmatter and report results on aggregation of magnetized quark nuggets (MQNs) after formation from the quarkgluon plasma until expansion of the universe freezes out the mass distribution to ~ 10^{−24} kg to ~ 10^{14} kg. Aggregation overcomes weakinteraction decay. Computed mass distributions show MQNs are consistent with requirements for dark matter and indicate that geologic detectors (craters in peat bogs) and spacebased detectors (satellites measuring radiofrequency emissions after passage through normal matter) should be able to detect MQN dark matter. Null and positive observations narrow the range of a key parameter B_{o} ~ B_{surface} to 1 × 10^{11} T < B_{o} ≤ 3 × 10^{12} T.
Introduction
Eightyfive percent of the mass in the universe is dark matter^{1,2,3,4,5}. Its nature is unknown. Extensive searches for an explanation beyond the Standard Model^{6} have not detected any statistically significant signals^{4}. Macroscopic quark nuggets are a candidate for dark matter composed approximately equal numbers of strange, up, and down quarks and are consistent with the Standard Model. They are called many names^{7,8,9,10,11,12}. All quark nuggets interact^{13,14,15} through gravitational and strong nuclear forces. A brief summary of quarknugget formation, stability, and compliance with darkmatter requirements^{7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35} has been updated from Ref.^{17} and is provided for completeness as Supplementary Information: Quarknugget research summary.
Since quark nuggets are composed of three quarks, they are baryons. Convention often assumes that all baryons are normal (i.e. not dark) matter and designates the number density and mass density of normal matter as n_{b} and ρ_{b} respectively. However, since quark nuggets are baryons too, we use n_{n} and ρ_{n} for those quantities of normal matter, respectively.
Most previous models of quark nuggets have not considered effects caused by their selfmagnetic field. However, Tatsumi^{16} explored the internal state of quarknugget cores in magnetars and found they may exist as a ferromagnetic liquid with a surface magnetic field B_{surface} = 10^{12±1} T. His theory uses the bag model^{21} because more rigorous lattice and perturbative calculations with chromodynamics are intractable for the relevant energy scale of ~ 90 MeV. Tatsumi find that a stable ferromagnetic state should exist if the coupling constant α_{c} is ~ 4 at this energy scale. A metastable ferromagnetic state may exist for α_{c} > 1.75. Although α_{c} gets larger as the energy scale decreases, it is not known if α_{c} is indeed that large. Even with those considerations, his conclusions have important consequences and are testable through searches for quarknugget dark matter. We apply his ferromagneticliquid model to quarknugget dark matter. In a previous paper, we found the selfmagnetic field strongly enhances the interaction cross section of a magnetized quark nugget (MQN) with a surrounding plasma^{17}. In this paper, we explore the interaction between MQNs, which are magnetically attracted into aggregating collisions. As ferromagnetic liquids, they will combine and remain strongly magnetized after aggregating.
Mass distributions have been computed and are reported in this paper as a function of the key parameter B_{o} that is related to the average value <B_{surface}>:
Throughout this paper, we will use B_{o} as a key parameter. The value of B_{o} equals <B_{surface}> if the mass density of MQNs ρ_{QN} = 10^{18} kg/m^{3} but that number is quite uncertain. Witten’s^{7} estimate of “somewhat greater than nuclear density” and a rough calculation that gives ~ 7.5 × 10^{17} kg/m^{3} is consistent with 6 × 10^{17} to 7 × 10^{17} kg/m^{3} covering the range of uncertainty in the proton radius and the corresponding mass density. Peng et al.’s^{30} more recent work covers a range of 1.7 × 10^{17} to 3.3 × 10^{18} kg/m^{3} for quark matter in quark stars. We use ρ_{QN} = 1 × 10^{18} kg/m^{3} in the calculations below. In addition, the B_{o} parameter depends on the density of dark matter ρ_{DM} = 1.6 × 10^{8} kg/m^{3} at time t ≈ 65 μs, when the temperature T ≈ 100 MeV in accord^{36} with the standard ΛCDM cosmology. If more accurate values of ρ_{DM} when T ~ 100 MeV, ρ_{QN}, or B_{o} are found, then Eq. (1) can give a correspondingly more accurate value of <B_{surface}>.
We assume MQNs are formed with baryon number A = 1 at the beginning of baryogenesis, as are protons and neutrons, and aggregate by binary collisions, similar to nucleogenesis of lowA elements in the standard cosmological model. Since quarknuggets are electrically neutral (or neutralized as discussed below) and quickly aggregate to A \(\gg\) 1 (as shown below), we assume they are decoupled from the thermal environment of the comoving universe. With those assumptions, we simulated the aggregation of MQNs from the time baryons form in the quarkgluon plasma until expansion of the universe reduces MQN density and freezes out the mass distribution.
Although Tatsumi’s theory of ferromagnetism is applicable to quarknuggets with A \(\gg\) 1, we note that the magnetic moments and mass densities of neutrons and protons, which are also baryons with A = 1, correspond to magnetic fields B_{o} = 2.5 × 10^{12} T and 1.5 × 10^{12} T respectively. These fields are in the middle of the range identified by Tatsumi, so the aggregation from A = 1 is not unreasonable.
The results are strongly dependent on surface magnetic field B_{surface} from Tatsumi’s theory. We find the strong magnetic interaction of MQNs and their resulting mass distribution are consistent with requirements for darkmatter candidates, as summarized by Jacobs et al.^{12}: large scale structure constraints (cosmic microwave background and the Bullet Galaxy) from selfinteractions and interactions with other baryons and photons, ancient mica observations^{32}, and results from the Skylab experiment^{36}. Additional requirements may be added in the future. We conclude that MQNs provide a candidate for dark matter that appears to be consistent with the Standard Model of particle physics, without any extensions and merit further investigations.
Null results from a systematic search for MQNs with a 30 km^{2} section of the Great Salt Lake that was monitored with an array of three hydrophones are presented in the “Discussion” section. Results^{37} from investigation of episodic nonmeteorite impacts are consistent with an MQN impact and are also briefly summarized in the “Discussion” section. The null and episodic results consistent with the mass distributions for nonexcluded values of B_{o} motivate additional, systematic investigations^{38}.
We use standard international MKS units except temperature, which is expressed in electron volts (eV), and interaction strengths σ_{x}/m_{x} (i.e. cross sections divided by mass) which is expressed in cm^{2}/g to facilitate comparison with values quoted in the literature.
Results
Direct simulation Monte Carlo computation of MQN formation and aggregation
The mass distribution of magnetized quark nuggets, existing as a ferromagnetic fluid held together with the strong force as described by Tatsumi^{16}, is computed from their aggregation in binary collisions under the influence of their selfmagnetic fields. We assume that singlets with A = 1 are formed when the thermal energy of the early universe is much less than the rest mass of the singlet (kT_{f} ≪ m_{s}c^{2}) for Boltzmann constant k, formation temperature T_{f}, singlet mass m_{s}, and speed of light c. Protons and neutrons are similarly formed in the ΛCDM model^{1}. Since m_{s} is somewhat larger^{7,8,11} than the proton mass, we set m_{s} = 1.7 × 10^{−27} kg and choose T_{f} = 100 MeV, which corresponds to time t ~ 65 μs and is compatible with the formation of particles of mass m_{s} from the quarkgluon plasma^{39}.
MQNs interact with each other through their magnetic fields and aggregate through binary collisions when their magnetic potential energy is greater than their initial kinetic energy. Because the magnetic dipole–dipole interaction scales as r^{−3}, there is no centrifugal force barrier to a direct collision and aggregation. Since MQNs are a ferromagnetic liquid in Tatsumi’s theory^{16}, the domains align after aggregation and the surface magnetic field is preserved, so B_{o} is preserved in aggregations.
MQNs reorient as they approach each other to experience maximum attractive force, which is the orientation with minimum potential energy^{40}. Therefore, our simulation does not track orientation of each MQN.
Direct Simulation Monte Carlo (DSMC) methodology was developed for the aggregation of particles in a collisionless fluid. We adapted the procedure described by Kruis, Maisels, and Fissan^{41} who verified the methodology by comparing their results with analytic solutions of particle aggregation. Our adaptation is presented in the “Methods” section.
With both net electric and magnetic fields, the aggregation cross section σ_{EM} = π r_{EM}^{2} for the collision of particles i and j, and r_{EM} is given implicitly by the sum of the electric and magnetic potentials exceeding the kinetic energy at infinity:
in which q_{i} and q_{j}, m_{m},_{i} and m_{m,j}, m_{i} and m_{j}, and \(\vec{u}_{i}\) and \(\vec{u}_{j}\), are, respectively, the electric charges, magnetic dipole moments, masses, and vector velocities of the i and j particles. As usual, ε_{o} is vacuum permittivity and μ_{o} is vacuum permeability. The middle term is the relative magnetic potential for two particles with different magnetic dipole moments separated by the distance r_{EM}^{41}.
Since Tatsumi finds this ferromagnetic configuration is a liquid held together by the strong nuclear force, MQNs will be highly spherical. Consequently, we approximate the magnetic moment m_{m} of each spherical, uniformly magnetized quark nugget as the magnetic moment of a current loop with the same radius r_{QN} and magnetic field B_{o} as the quark nugget, at a distance r_{QN} on axis and above the center of the loop. We assume each quark nugget is a sphere with uniform mass density ρ_{QN}, so the mass m_{QN} of each quark nugget is (4/3)π ρ_{QN} r_{QN}^{3} and
Substituting Eq. (3) into Eq. (2), multiplying each side by r_{EM}^{3}, and rearranging gives a cubic equation in r_{EM}. This is then calculated for each collision using the usual formula for the root of a cubic equation.
We assume that each singlet is formed with A = 1, so each contains one up, one down, and one strange quark. The assembly is in the theoretically predicted, ultradense, colorflavorlocked (CFL) phase^{18} of quark matter. Steiner et al.^{19} showed that the ground state of the CFL phase is color neutral and that color neutrality forces electric charge neutrality. Models of unmagnetized quark nuggets by Xia et al.^{11} and by Zhitnitsky et al.^{42} predict different small (i.e. the ratio of electric charge per baryon mass is much less than one), nonzero internal electric charges, but both predict charge neutralizing surface layers. In addition, the magnetic field of MQNs significantly changes their internal energy^{29,43}. Including the magnetic field in the equilibrium calculation^{44} reduces the internal electron per baryon ratio to ~ 0.0003 for nuclear density quark nuggets and provides electric charge neutrality. Therefore, we assume the net internal charge plus the electrically neutralizing surface layers produce zero net electric charge.
The aggregation cross section is considerably simplified for zero net electric charge:
The DSMC method for computing aggregations^{41} uses the cross sections σ_{ij} for every particle pair (i, j) to simulate the aggregation process, as explained in more detail in the “Methods” section.
The simulation starts with 100,000 particles, as recommended by Kruis, Maisels, and Fissan^{41} to provide adequate statistics. The initial speeds of the particles were generated to fit a Maxwell–Boltzmann distribution with temperature T = 100 MeV, and their velocity vectors were generated by generating velocity unit vectors with random orientation in space. Approximately 96% of the initial velocities have relativistic γ < 2.0 and 61% have γ < 1.2. Less than 0.1% have γ ≥ 5.
Quark nuggets are decoupled from the temperature of the universe when they are formed. They interact only with other quark nuggets in binary collisions that conserve linear momentum. Net angular momentum from their collisions is radiated away by their rotating magnetic field, so their average energy per unit mass decreases as they aggregate.
Approximately 10^{−14} s into the simulation, which is the beginning of the 10th generation, aggregations have reduced the velocities to 99.6% with γ < 2.0 and 96.5% with γ < 1.2. Less than 0.02% have γ ≥ 5. Even though some of the initial velocities are marginally relativistic, we use nonrelativistic dynamics to do these first calculations to minimize computing time.
Selection and aggregation of iparticle, jparticle (i, j) pairs proceed for 50,000 aggregations, which we call one generation. At the end of each generation, only 50,000 particles remain. Since mass is conserved, the mean mass has doubled. At the completion of each generation, the time within the simulation and all data are saved for analysis and for restart if needed. The volume of the simulation is then doubled by expanding the simulated volume to include the surrounding space. Since the aggregation process is uniform in space, the mass and velocity distributions of particles in that additional volume are quite similar to those of the 50,000 remaining particles. Consequently, we duplicate the mass and speed of each of the 50,000 particles to restore the number of particles to 100,000 and maintain adequate statistics. Since elastic scattering randomizes the directions of the velocity vectors, we randomize the directions of the velocities for the duplicated particles and, therefore, avoid dividing by zero as the next generation is simulated.
As described in the “Methods” section, the DSMC process that calculates each aggregation depends on

1.
aggregating collision rates for each of n particle pairs (i, j), with mass m_{i} and m_{j}, respectively,

2.
velocities \(\vec{u}_{i}\) and \(\vec{u}_{j}\), respectively,

3.
aggregation cross section σ_{ij}, which is given by Eq. (4),

4.
comoving volume V of the simulation,

5.
darkmatter mass density ρ_{DM} at the time of each aggregation, and

6.
parameters B_{o} and ρ_{QN}, which are independent of time.
The aggregation calculation is described in the “Methods” section and produces a mass and velocity distribution at the end of each generation. The process also calculates the time increment δ_{g},_{k} associated with each k = 0 to 49,999 aggregation in each generation g. The time increment δ_{g} associated with the gth generation is the sum over k of δ_{g},_{k} values in the gth generation.
The function \(f_{g,k} (m_{i,} ,m_{j,} ,\vec{u}_{i} ,\vec{u}_{j} )\) is derived in the “Methods” section.
Since the comoving volume V = total darkmatter mass divided by ρ_{DM} is continually changing with time, each aggregation k depends on the timevarying darkmatter mass density ρ_{DM}(t) = ρ_{DM_g,k} in Eq. (5). Although B_{o} is a constant for a simulation, its value changes δ_{g},_{k}, which changes ρ_{DM_g,k}. Therefore, a simulation should be completed for each B_{o}. Each simulation requires several months of computing.
However, we found that ρ_{DM}, is approximately constant within the 50,000 aggregations of a generation until the aggregations and expansion of the universe begin to freeze out the mass distribution. Once the aggregation rate slows, freezeout quickly follows. The result is that assuming ρ_{DM_t} is constant within a generation is sufficient for specifying the mean and maximum mass in a distribution to within one generation, which is within a factor of 2. As shown in Fig. 5, mean mass and maximum mass are such a strong function of B_{o}, an uncertainty of a factor of 2 in mass gives an uncertainty of only 8% in the corresponding B_{o}. Eight percent uncertainty is sufficient to compare predictions with observations to narrow the range of B_{o} from Tatsumi’s 10^{12±1}. We found that representing each mass distribution as a sum of the number of masses in each decade of mass is sufficient for this purpose. So the uncertainty in B_{o} dominates the uncertainty in the mass distribution. Once the value of B_{o} is determined to within ± 20%, a full simulation should be run for representative values of B_{o}.
The assumption of constant ρ_{DM_g,k} within a generation g allows approximate mass distributions to be calculated by post processing the results from one detailed simulation with B_{o} = 10^{12} T. Equation (5) is approximated by
The time t_{G} since the beginning of the universe and associated with the completion of the Gth generation equals the 65 μs start time plus the sum of the generation times δ_{g} for generations up to and including the Gth generation.
In post processing, ρ_{DM}(t) = ρ_{DM_g,k} is updated to be consistent with the standard ΛCDM Model, as explained in the next paragraph. The time t_{G} is also associated with the mass and velocity distributions calculated in the Gth generation. The process continues until t_{G} ≥ 13.8 Gyr, the present age of the Universe.
We use a solution^{45} of the Friedmann equation^{46} in a radiationdominated Universe to obtain the temperature T as a function of time t. The solution includes the time dependent effective number of relativistic degrees of freedom which would have the same entropy density at the same photon temperature^{39}. The solution is appropriate for time t up to ~ 50,000 years, which is well beyond the freezeout time of the MQN mass distribution. Fitting that solution to a simple analytic function lets the simulation run as quickly as possible. The fit is adequate to ± 5% in T for 65 μs ≤ t ≤ 122 ms, by which time the mass distributions have frozen out, and to ± 13% for 122 ms < t ≤ 10^{10} s. For example, for B_{o} = 1.085 × 10^{12} T, time goes from 3 μs generation 89–13.8 Gyr in generation 90; therefore, the error after t = 55,000 years is well within the factor of 2 uncertainty in mean mass and 8% uncertainty in B_{o} described above.
We use the Standard Model and the current values of cosmological variables^{1} (i.e. number density n_{γ} of photons in thermal equilibrium, the ratio of normalmatter number density n_{n} to photon number density n_{γ}, and the ratio of colddarkmatter mass density ρ_{DM} to normalbaryon mass density m_{p} n_{n}) to calculate the time dependent darkmatter density ρ_{DM}(t) from temperature T_{MeV}(t):
in which Rieman Zeta function ξ(3) = 1.202 and the temperature T_{MeV} in MeV as a function of time t in seconds.
The DSMC technique made it possible to simulate the aggregation of quark nuggets from ~ 65 μs after the universe began to the present day in about three months of processing on a personal computer.
Aggregation overcomes decay by weak interaction
Since quark nuggets with baryon number A = 1 are not observed in accelerator experiments, they decay through the weak interaction in ~ 0.1 ns unless their decay is interrupted. We find that the aggregation time is typically 0.003 ns, so aggregation overcomes decay and allows growth of quark nuggets to A \(\gg\) 1, where they are magnetically stabilized.
Aggregated mass distributions of electrically neutral or neutralized MQNs
The aggregation process proceeds quickly, as shown in Fig. 1. By t = 1 ms, the mass distribution of electrically neutral or neutralized MQNs has essentially frozen out and changes less than factor of ~ 2 for the next 10 Gyr.
The simulated mass distribution evolves quickly between the assumed formation of singlets with baryon number A = 1 at t = 65 μs, when temperature T = ~ 100 MeV, and 1 ms, when the largest mass is 8 kg. By t = 10 Gyr, the largest mass is 20,800 kg and the rest of the distribution has changed little.
Computed quarknugget mass distributions for electricchargeneutral collisions and the timedependent darkmatter density ρ_{DM}(t) given by Eq. (8) are shown in Fig. 2 as the cumulative fraction F_{M} of the mass distribution between 0 and mass M:
The maximum quarknugget mass is a strong function of B_{o} and is ~ 10^{−2} kg, ~ 10^{4} kg, and ~ 10^{15} kg for B_{o} = 10^{11} T, 10^{12} T, and 10^{13} T, respectively.
The distribution has the character of aggregations: the great majority of the particles have very little mass and a few have the great majority of the total mass^{41}. MQN mass distribution for the baseline B_{o} = 10^{12} T in Fig. 2 covers nearly 33 orders of magnitude in mass. The < 3.5% of the particles with mass > 10^{−9} kg contain all but ~ 10^{−9}% of the total mass. The maximum mass is ~ 20,000 kg. The average mass is 0.50 kg, and 0.9998 of the total mass is in particles with mass greater than the average. The mass distribution is too extreme to use average mass for comparison with data, as is often done.
The mass distribution for the B_{o} = 10^{12} T is also presented in Fig. 3 as the fraction of particles in each decadal mass increment, i.e. the fraction with m_{decade} = Integer(log_{10}(m_{QN})).
We explored the sensitivity of the results to the darkmatter mass density ρ_{DM}(t) by introducing a multiplier to ρ_{DM}(t) given by Eq. (8). The effect of multipliers 0.1, 0.5, 1.0, 2, and 10 on the final mass distributions is shown in Fig. 4.
Variations with changes to ρ_{DM} seem insignificant on the full scale of Fig. 4a. However, comparison of the detailed view in Fig. 4b illustrates a general finding: In the final stages of aggregation, very large masses grow rapidly at the expense of intermediate masses, while very small masses remain relatively unaffected.
The selected view in Fig. 4b also shows that a factor of 2 change in ρ_{DM} can produce a factor of ~ 100 change in the maximum quarknugget mass. A factor of 10 reduction in ρ_{DM} can cause a factor of ~ 10^{5} reduction in quarknugget mass. Nevertheless, aggregation robustly increases the quarknugget masses from their assumed initial mass of ~ 10^{−27} kg to between 0.1 and 10^{6} kg.
These variations with ρ_{DM} are provided to illustrate the effect of the uncertainty in ρ_{DM}. Since Eqs. (5) and (6) show aggregation is a function of \(\rho_{DM}^{  1} B_{o}^{{  \frac{4}{3}}}\), variation in ρ_{DM} gives insight into the effect of the same variation in \(B_{o}^{\frac{4}{3}}\).
As stated in the introduction, uncertainty in B_{o} drives the uncertainty in the mass distribution. Reducing that uncertainty is the goal of the observations. Then Eq. (1) can be used to adjust mass distribution when better values of ρ_{DM} and ρ_{QN} are obtained from observations, as illustrated in the “Discussion” section.
In the next subsection, we will see if this final range of masses is sufficiently large to assure magnetized quark nuggets satisfy the noninteraction requirements for dark matter.
Effects of MQN mass distributions
Jacobs et al.^{12} systematically compare the interaction cross section σ_{x} of unmagnetized quarknuggets of mass m_{x} with observations and requirements of dark matter, assuming quark nuggets have a single mass. Their analysis covers largescale structure constraints (cosmic microwave background and the Bullet Galaxy) from selfinteractions and interactions with other baryons and photons, ancient mica observations^{32}, and results from the Skylab experiment^{36}. They assume the cross section for interactions is approximately the geometrical cross section. We adapt their singlemass analysis to the case of MQNs with the computed mass distribution in Fig. 2 and interacting through their magnetic field in vacuum and their magnetopause^{17} in a surrounding plasma.
Consider local density of darkmatter ρ_{DM} = 7 × 10^{−22} kg/m^{3}^{4} and quark nuggets with a single mass m_{QN} and the same velocity u_{QN} relative to a target of identical particles with target number density n_{t}. The interaction cross section σ_{x} is the same for all particles and the event rate per unit volume is \(\frac{\Delta n}{{\Delta t}} = n_{QN} u_{QN} \sigma_{x} n_{t}\), in which quark nugget number density \(n_{QN} = \frac{{\rho_{DM} }}{{m_{QN} }}\). Solving for σ_{x}/m_{QN} gives
Similarly, consider a sample of N_{QN} quark nuggets that fill a volume V. The number density of quark nuggets is \(n_{QN} = \frac{{N_{QN} }}{V}\). Let the mass, velocity (relative to a target particle), and interaction cross section of the ith quark nugget be m_{i}, u_{i}, and σ_{i}, respectively. Then \(V = \frac{{\sum\nolimits_{i = 1}^{{N_{qn} }} {m_{i} } }}{{\rho_{DM} }}\) and the event rate per unit volume is
The number density n_{QN} of quark nuggets and mean mass m_{mean} are respectively
The approximate expression for a broad mass distribution corresponding to Eq. (10) is
in which effective velocity u_{eff} ≈ 3 × 10^{5} m/s according to Jacobs et al., and including the ~ 2.5 × 10^{5} m/s velocity^{9} of the solar system about the galactic center for Earthbased observations. Since we have m_{i}, u_{i} for N_{QN} = 10^{5}, we can compare (σ_{eff}, m_{mean}) with the various criteria Jacobs et al. developed for singlemass quarknugget dark matter.
As explained in the “Methods” section, m_{mean} doubles during each generation of 50,000 aggregations of our simulation, which starts each generation with 100,000 particles. Since the selfinteraction cross section of Eq. (4) is proportional to (m_{i} + m_{j})^{2/3}, \(\frac{{\sigma_{eff} }}{{m_{mean} }}\) decreases with each generation and, therefore, with time. Since Fig. 1 shows that the quarknugget mass distribution at time t = 1 ms is very close to its final value, if \(\frac{{\sigma_{eff} }}{{m_{mean} }}\) criteria are satisfied at t = 1 ms, they are sufficiently satisfied at the later times evaluated by Jacobs et al.
Since MQNs are just magnetized Macros, the many interactions of Macros examined by Jacobs et al. that are not affected by the selfmagnetic field of the quark nugget are also appropriate to MQNs. The geometric cross section for these nonmagnetic interactions varies as r_{QN}^{2} and the mass varies as r_{QN}^{−3}, so \(\frac{{\sigma_{eff} }}{{m_{mean} }} \propto \frac{1}{{r_{QN} }} \propto \frac{1}{{m_{QN}^{1/3} }}\) and the largest value of \(\frac{{\sigma_{eff} }}{{m_{mean} }}\) is the case with the smallest masses. As shown in Fig. 2, that worst case is B_{o} = 10^{11} T. Evaluating Eq. (13) for that worst case gives
The nonmagnetic phenomenon with the most stringent scattering requirement evaluated by Jacobs et al. is elastic darkmatter/photon scattering. It requires \(\frac{{\sigma_{eff} }}{{m_{mean} }} < 4.5 \times 10^{  7}\) cm^{2} g^{−1} and, as shown by Eq. (14), is easily satisfied for the worst case MQN distributions.
Equation (13) assumes the cross section σ_{i} depends only on the properties of the ith MQN. Selfinteractions are more complicated. MQNs have different masses, so they have different magnetic dipoles. They also have different velocities. The cross section in Eq. (13) is generalized to σ_{ij} in Eq. (14) and computed from the properties of both the ith and jth MQN as a particle pair. For each aggregation, ith and jth MQNs were randomly chosen from the ensemble of MQNs to evaluate selfinteraction with Eq. (14). The time interval between t = 1 ms and 2 ms was chosen as a relevant case because the mass distributions in that time interval are quite similar to those at t = 10 Gyr, as shown in Fig. 1, so they are relevant to the effect of MQN dark matter on the evolution of the universe.
The calculation is also a worst case because the mean MQN mass continues to increase with time and the ratio \(\frac{{\sigma_{eff} }}{{m_{mean} }}\) decreases with increasing mass. The cross section for magnetic aggregation is used as a surrogate for elastic scattering. The rapidly diminishing magnetic field with increasing distance from a quark nugget assures that the annular cross section for scattering is less than or comparable to the cross section for aggregation. The calculation gives \(\frac{{\sigma_{eff} }}{{m_{mean} }}\) = 4.6 × 10^{−7} cm^{2} g^{−1} and is less than Jacobs et al.’s most conservative 0.04 cm^{2} g^{−1}. Consequently, the calculated mass distributions of MQNs easily satisfy the selfinteraction criterion for dark matter.
Protons can also scatter off the magnetic fields of MQNs. The cross section depends on the mass of the MQN and on the velocity of the proton in the rest frame of the MQN. For a representative velocity of 2.5 × 10^{5} m/s, simulations of protonMQN scattering in the equatorial plane of the MQN with baseline parameters of B_{o} and ρ_{QN}, gave \(\frac{{\sigma_{x} }}{{m_{QN} }}\) = 2.3 × 10^{−4} cm^{2} g^{−1} ± 6% and was inversely proportional to proton velocity. Using these results and Eq. (13) to include the effect of the mass distribution gives \(\frac{{\sigma_{eff} }}{{m_{mean} }}\) = 1.9 × 10^{−4} cm^{2} g^{−1} ± 10% for MQNbaryon scattering at t ≈ 1 ms after the big bang and decreasing with increasing time. Jacobs et al. require \(\frac{{\sigma_{eff} }}{{m_{mean} }}\) < 0.06 cm^{2} g^{−1} to be consistent with the observation that dark matter concentrated near galactic centers does not increase gas temperature by collisional heating. Therefore, MQNs comfortably satisfy the requirement in spite of their large magnetic field.
Effect of escape velocity from galaxy
In the simulations, each quark nugget’s velocity evolved to its final velocity by aggregation with conservation of linear momentum. As discussed with respect to Fig. 4, very large masses grow rapidly at the expense of intermediate masses during aggregation, while very small masses remain relatively unaffected. Therefore, it is not surprising that approximately a third of the quarknuggets keep their initial high velocity, which exceeds galactic escape velocity.
Approximately 35% of the final quark nuggets had total velocity > 600 km/s and would not be gravitationally bound within our galaxy. These quark nuggets do not contribute to local dark matter and are dropped from analysis of quarknugget detections on Earth. However, the 35% of quark nuggets that escape the galaxy have less than 1% of the total MQN mass.
Direct detection of quark nugget events
Jacobs et al. found that the null results from the Skylab experiment^{36} excluded quark nuggets with a single mass of less than 2 × 10^{−10} kg based on inability to penetrate 0.25 cm of Lexan polycarbonate to make tracks in plastic with > 400 MeV cm^{2} g^{−1} stopping power. The calculation does not include Skylab’s 1 g/cm^{2} aluminum wall. Applying the magnetopause interaction model^{17} to quarknugget passage through the Skylab wall and into the plastic, we find quarknuggets with mass ≤ 10^{−16} kg are effectively shielded from the detector, as shown in Fig. 3. The computed quarknugget flux F_{j} in the jth decadal mass increment, which includes all quark nuggets with mass m_{QN} such that j = Integer(log_{10}(m_{QN})), is
in which M_{j} is the total mass in decadal increment j, 3 × 10^{j} is approximately the average quarknugget mass in jth decade, 3.15 × 10^{7} is the number of seconds per year, and 4π sr gives the decadal flux in number m^{−2} y^{−1} sr^{−1}.
Jacobs et al. find that the Skylab detector should have been sensitive to unmagnetized quarknuggets with σ/m_{QN} < 3 cm^{2} g^{−1} and rules out singlemass quark nuggets with m_{QN} less than approximately 2 × 10^{−10} kg. The mass distribution for MQNs gives a very different result. Summing flux F_{j} by decadal mass in Fig. 3 for j > 16, i.e. m_{QN} > 10^{−16} kg, gives 1.5 × 10^{−10} m^{−2} y^{−1} sr^{−1} total expected flux of quark nuggets into the Skylab experiment. The experiment had an exposure of about 2 m^{2} y sr. The expected number of events calculated from the total flux, exposure area and time, and 2π sr equals ~ 2 × 10^{−9} events, which is consistent with the 0 observed. Even if the plastic detector had been outside Skylab, the number of predicted events would have been only 6 × 10^{−9}. Skylablike experiments cannot test the MQN darkmatter hypothesis. Very large areas and long exposure times are essential.
More generally, the mass distribution and number flux are strongly dependent on the surface magnetic field B_{o}, which Tatsumi brackets as 10^{12±1} T. The corresponding number fluxes of highvelocity (> 10^{4} m/s) quark nuggets are shown in Fig. 5 as a function of B_{o} for three environments: (1) in space near Earth, (2) on Earth’s surface (after slowing down through the magnetopause effect in the atmosphere), and (3) on Earth’s surface and depositing sufficient energy/length (i.e. ≥ 100 MJ/m) to make > 3.5 m diameter craters^{37} in a peat bog.
The extremely small number flux of quarknugget dark matter per Eartharea per year in Fig. 5 means detectors must have very largearea geophysical or planetary targets. The Energy Density (in units of kilotons of TNT per km) is too large for B_{o} > 3 × 10^{12} T to have occurred at the indicated flux without having been reported.
The sudden drop in number flux at B_{o} ≈ 4 × 10^{12} T in Fig. 5 occurs because the aggregation process runs away at that value of B_{o}. Aggregation runaway becomes even more extreme at B_{o} ≈ 9 × 10^{12} T and creates quarknuggets with mass ≈ 10^{15} kg, as shown in Fig. 6.
Computed mass distributions, the mean directed velocity of 250 km/s, and the interstellar mass density of dark matter give representative values of quarknugget flux by decadal mass and surface magnetic field parameter B_{o} in Table 1. Results for the entire range of B_{o} are provided in Supplementary Data: Mass distributions and fluxes by B_{o} parameter.xlxs. These detailed results can be used to estimate the event rate for future efforts to detect MQNs.
For B_{o} > 1.5 × 10^{12} T, Table 1 shows gaps in the distribution of flux as a function of mass. The continuous distribution breaks into two or even three isolated distributions. The effect is aggregation runaway. Since the cross section for aggregation in Eq. (4) is larger for two largemass MQNs than it is for one largemass and one smallmass MQN, the large mass aggregations preferentially aggregate and leave gaps in the distribution. Aggregation also reduces MQN velocity, so the velocity difference in the denominator of Eq. (4) also contributes to the runaway effect. Runaway explains why the flux rates decrease dramatically for large B_{o}.
Discussion
Tatsumi^{16} found that quark nuggets could exist as a ferromagnetic fluid held together by the strong nuclear force and have a surface magnetic field between 10^{11} and 10^{13} T. We have explored the consequences of his theory as applied to dark matter. In this paper, the direct simulation Monte Carlo method shows magnetized quark nuggets (MQNs) that formed with baryon number A = 1 when the universe is at ~ 100 MeV, approximately t = 65 μs after the big bang, aggregate under the influence of their selfmagnetic fields. We find that aggregation dominates decay by the weak interaction and produces mass distributions, as a function of the B_{o} parameter, that satisfy requirements^{12} for dark matter as early as 1 ms after the universe forms.
Previous searches for dark matter have been based on singlemass analysis. Mass distributions computed from first principles help design precision tests of the MQN dark matter hypothesis.
The mass distributions reported in this paper and the magnetopause interactions described in Ref.^{17} make the MQN hypothesis for dark matter testable. All quark nuggets have mass density somewhat greater than nuclear density ~ 10^{18} kg/m^{3}, so their geometric cross section σ_{o} is very small. The energy deposited per unit length in passage through matter is correspondingly very small. However, each MQN has a very large magnetic field that is compressed by the particle pressure of ionized matter (plasma) streaming into its rest frame with relative velocity v. The particle pressure is balanced by the compressed magnetic field pressure at its magnetopause, which has a cross section σ_{m} that is much larger than its geometric cross section σ_{o}^{17}.
in which μ_{o} = permeability of free space, K ~ 1, and ρ_{p} = the local plasma density. The relative velocity v is on the order of the velocity of the Earth’s motion about the galactic center and through the darkmatter halo, i.e. ~ 250 km/s. At that velocity, incoming matter is quickly heated to the plasma state upon impact, similar to how a meteor creates a plasma passing through atmosphere. The decelerating force F_{e} from the drag on the magnetopause is
Integrating Eq. (17) twice with the cross section from Eq. (16) gives the distance x_{max} at which velocity v = 0 for an MQN with mass m, mass density ρ_{QN}, and initial velocity v_{o} for a given B_{o}:
By definition, the force F_{e} in Eq. (17) equals the energy deposition per unit length. For example, Fig. 6 shows the maximum mass in a collection of 100,000 MQNs in the simulation as a function of B_{o}. Figure 5b shows the kinetic energy (blue) and energy/km deposited (red) by the maximum mass as a function of B_{o}. The simulation also gives the number flux as a function of mass and B_{o}, as illustrated in Table 1. Combining this information gives the approximate event rate for impacts above a minimum energy deposited per unit length as a function of B_{o}. The fact that MegatonTNT/km impacts are not observed every century or so makes B_{o} > 3 × 10^{12} T very unlikely. We, therefore, exclude B_{o} > 3 × 10^{12} T.
In addition, Fig. 5a shows that etchedplastic targets in space without any shielding (blue) would require between ~ 4 × 10^{9} m^{2} and ~ 3 × 10^{4} m^{2} area to detect one event per year for B_{o} between 4 × 10^{12} T and 10^{11} T, respectively. Such large detectors in space are impractical.
Before the MQN mass distribution was calculated, Ref.^{17} proposed acoustically monitoring MQN impacts in water at the Great Salt Lake, Utah, USA, as a possible sensor for MQNs. Energy absorption in the atmosphere prevents detection of MQN masses m < 10^{−4} kg^{17}. However, energy/length deposited in water of density ~ 1 kg/m^{3} for B_{o} ~ 10^{12} T is ~ 10 kJ/m for m = 10^{−4} kg and is ~ 100 MJ/m for m ~ 10 kg. Such a large energy density would create large amplitude acoustic waves that would be easily detected by an array of hydrophones. Shortrange testing, reported in Ref.^{17}, indicated that the background noise and sound propagation through the water would allow detection over most of the Great Salt Lake. When the full system was deployed, longrange testing included absorption and interference effects in the water and material beneath the lake. These effects are not yet fully understood but apparently limited the range to ~ 3 km. Therefore, the array was sensitive to MQN impacts with mass ≥ 10^{−4} kg within a ~ 3 × 10^{7} m^{2} area, which is about 30 times the detector area of IceCube Neutrino Observatory at the South Pole. Weather limited observations to ~ 90 days per year. The system was calibrated with line explosives extending from the surface to the bottom and at various distances to obtain the signature of an MQN impact. No events were recorded with the distinctive signature in 90 days of good observations in 2019.
Mean mass is often used for darkmatter candidates. The Poisson probability of obtaining k events with the expected mean rate λ is given by
The signature from an MQN impact is so far above background for m ≥ 10^{−4} kg that the background is effectively 0. For the meanmass analysis, the expected mean rate is
in which 2π = solid angle of incident MQNs, interstellar darkmatter density ρ_{DM} = 7 × 10^{−22} kg/m^{3}, mean incident velocity v = 250 km/s, detector Area = 3 × 10^{7} m^{2}, and m_{mean} is the mean MQN mass. Probability P from Eq. (19) for null result k = 0 equals 10^{−144}, 0.01, and 0.25 for m_{mean} = 10^{−4} kg, 7.2 × 10^{−3} kg, and 0.025 kg respectively. Therefore, the meanmass analysis implies, m_{mean} ≤ 7.2 × 10^{−3} kg is excluded at the 99% confidence level. Figure 6a gives m_{mean} versus B_{o}. Excluding m_{mean} ≤ 7.2 × 10^{−3} kg excludes B_{o} ≤ 2.5 × 10^{11} T.
The same analysis for distributed mass has been done from the results in Table 1 and can also be done with less precision from Fig. 5a (red). For B_{o} = 1 × 10^{11} T, 2 × 10^{11} T, and 3 × 10^{11} T, λ = 5.2, 0.17, and 0.036, respectively, and P for k = 0 equals 0.005, 0.84, and 0.96, respectively. Therefore, analysis of the data with the computed mass distributions excludes distributions with B_{o} ≤ 1 × 10^{11} T with a confidence level of 99.5% from Poisson statistics. Comparison with the meanmass analysis illustrates the importance of the mass distribution.
The strong dependence of P with B_{o} implies that a much larger area detector and/or much longer time is required to investigate the remaining parameter space of 1 × 10^{11} T < B_{o} ≤ 3 × 10^{12} T. Asteroids provide some possibility of detection over a much longer time because they should have been accumulating MQNs since the solar system was formed ~ 4 Gyr ago. The MQN mass accumulated in an asteroid is conservatively estimated by assuming an asteroid of radius r_{a} stops incident MQNs with range x_{max} if r_{a} ≥ x_{max}. The maximum MQN mass m_{max} that will be absorbed by an asteroid with radius r_{a} is given by solving Eq. (18) for m = m_{max} with x_{max} = r_{a}.
Table 1 provides the number flux (number m^{−2} y^{−1} sr^{−1}) for MQN mass m for a given B_{o}. Multiplying each number flux by the associated decadal mass gives the mass flux by decadal mass. If all dark matter is composed of MQNs, the accumulated MQN mass in an asteroid is estimated as follows:

1.
Sum decadal mass fluxes for masses less than m_{max},

2.
multiply the sum by cross sectional area \(\pi r_{a}^{2}\),

3.
multiply that by 5.56 sr to include MQNs incident from all directions^{38}, and

4.
multiply that by 4 Gyr for the accumulation time.
For B_{o} = 1.5 × 10^{12} T, the total MQN mass \(m_{QN\_total} \approx 2 \times 10^{  11} r_{a}^{4.33}\) kg for r_{a} ≥ 100 m. For example, m_{QN_total} ~ 0.01 kg for r_{a} = 100 m and m_{QN_total} ~ 200 kg for r_{a} = 1000 m.
MQNs accumulated inside an asteroid would be much more detectable if they aggregate into a large MQN. However, aggregation inside an asteroid is much more complicated than it is in free space. The attractive force (in the pointdipole approximation) between two identical magnetic dipoles^{47} with magnetic moment m_{m} separated by a distance x is
The 1/x^{4} dependence assures the magnetic force dominates gravity only over short distances, so both forces have to be modeled with realistic distributions of MQN impacts as a function of asteroid size and with realistic resistance of MQN diffusion through asteroid material. Such a simulation is beyond the scope of this paper, but it should indicate if asteroid mining may provide a method to collect MQN dark matter or if such an attempt would be another null result.
Null results may be insufficient to motivate a very difficult systematic search for MQNs. In the past, episodic observations have tested new theories before investments were made for systematic study, e.g. General Relativity was first tested with the anomalous perihelion of Mercury and then with the bending of light near the sun in an eclipse. The results motivated more systematic studies with gravitational red shifts and, a century later, with gravitational waves. Consequently, we looked for episodic opportunities.
Terrestrial craters caused by nonmeteorite impacts offer larger areas and longer observation times. In recent years, NASAinvestigated, nonmeteorite impacts have been reported in the press approximately once per year. Ref.^{17} catalogues three such events:
“A12m diameter crater occurred at 11:05 PM, September 6, 2014, near Managua, Nicaragua [Cooke, W. Did a meteorite cause a crater in Nicaragua? https://blogs.nasa.gov/Watch_the_Skies/2014/09/08/didameteoritecauseacraterinnicaragua/ and https://www.cnn.com/2014/09/08/tech/innovation/nicaraguameteorite/, (2014) (Date of access: 24/06/2017)]. An event occurred on July 4, 2015, at the Salty Brine Beach in Rhode Island, USA [Shapiro, E., Cathcart, C. & Donato, C. Bomb squad, ATF investigating mysterious explosion at Rhode Island beach. https://abcnews.go.com/US/explosionreportpromptsevacuationrhodeislandbeach/story?id=32384143, (2015) (Date of access: 24/06/2017)]. Finally, an event occurred on February 6, 2016, in Tamil Nadu, India [Hauser, C. That wasn’t a meteorite that killed a man in India, NASA says. https://www.nytimes.com/2016/02/10/world/asia/thatwasntameteoritethatkilledamaninindianasasays.html?_r=0, (2016) (Date of access: 24/06/2017)].”^{17}. (Each link was also accessed on 15/04/2020.)
These reports occur approximately once per year, as illustrated by these three examples within 3 years. Monitoring peat bogs for nonmeteorite impact craters was also proposed in Ref.^{17} to link nonmeteorite craters to MQN impacts. Irish peat bogs offer up to 3 × 10^{8} m^{2} area witness plates that preserve impact craters for 100 to 1000 years, depending on the size of the impacting MQN and the value of B_{o}. One 3.5m diameter crater from an impact in 1985 has been excavated to the bedrock. The results are consistent with a nearly vertical impact of a 10 ± 7 kg MQN. The error bar arises from the uncertainty in B_{o} for 4 × 10^{11} T ≤ B_{o} ≤ 3 × 10^{12} T in Fig. 5a (black). The investigation confirmed the crater is consistent with an MQN with five points of comparison between theory and observation. The nonmeteorite crater provided the first–of many needed–results supporting the MQN hypothesis for dark matter, and narrowed the range of allowed B_{o}^{37}.
The results from nonmeteorite craters motivate finding a systematic way to test the MQN hypothesis with the largest possible detector area. The largest accessible area for realtime MQN searches is Earth’s magnetosphere. An experiment to systematically explore the MQN darkmatter hypothesis with a threesatellite constellation at 51,000 km altitude is described in Ref.^{38} MQNs experience a net torque as they decelerate during passage through Earth’s magnetosphere, ionosphere, and troposphere. The angular acceleration gives them MHz frequencies before they emerge into the magnetosphere where they can be detected by characteristic Dopplershifted, radiofrequency emissions from their rotating magnetic dipole^{38}. The MQN mass distributions presented in this paper are essential for planning such a systematic study.
Methods
The simulation begins when the temperature T_{MeV} ~ 100 MeV, which occurs at ~ 65 μs after the Big Bang. We start with 10^{5} particles, as recommended by Kruis, Maisels, and Fissan^{41}, to provide adequate statistics. The initial speeds of the 10^{5} particles were generated to fit a Maxwell–Boltzmann distribution with temperature T = 100 MeV. Their velocity vectors were generated by random selection of velocity unit vectors.
A representative comoving (i.e. expanding as the universe expands) volume V contains all the masses m_{i} in the simulation for a generation, with index i varying from 0 to n1 for n particles. Equation (8) provides the dark matter density ρ_{DM} versus time.
The DSMC method for computing aggregations^{41} requires the aggregating collision rate β_{ij}^{V} for each particle pair (i, j), with mass m_{i} and m_{j}, respectively, and velocity \(\vec{u}_{i}\) and \(\vec{u}_{j}\), respectively:
Each β_{ij}^{V} is based on the aggregation cross section σ_{ij}, which is given by Eq. (4), relative velocity \(\vec{u}_{i}  \vec{u}_{j} \), and the volume V of the simulation.
The sum S_{i} = the sum of all possible collision rates for aggregations with the ith particle and is the collision kernel for the ith particle:
S is the sum of all S_{i} for the next aggregation event.
Since S is the sum of all possible collision rates β_{ij}^{V} ordered by i and then for all j associated with that i before proceeding to the next i, every collision pair (i, j) is represented by a value between 0 and S. The numeric interval between (i, j) entries is proportional to the probability of their aggregating. To randomly select a pair (i, j) for collision, a random number R is chosen in the interval 0 to 1. Index i_particle is chosen for aggregation by solving
Then Index j_particle is chosen for aggregation by solving
The process, therefore, produces (i, j) pairs for aggregation in proportion to their probability of aggregating.
To record the aggregation of the i_particle and j_particle, the mass m_{i} is replaced by the mass m_{i} + m_{j}, the velocity components of the new ith particle are calculated from conservation of linear momentum, the jth particle is removed from inventory (reducing the total number of particles by 1), the β_{ij}^{V} terms are subtracted from each S_{i} to update it and the new value of S is calculated.
The process repeats for 50,000 aggregations, which is one generation. At the end of a generation, the aggregation process has reduced the number of particle by half and the mean mass m_{mean} of the 50,000 remaining particles has doubled.
To keep enough particles, we double the volume and “incorporate” the 50,000 particles from the added volume to restore the particle count to 100,000. Each of the original particles is duplicated with a twin that has the same mass, speed, and kinetic energy. Since elastic collisions are also occurring, the direction of the velocity vector of a particle and the direction of the velocity vector of its twin in the added volume are uncorrelated. Therefore, the directions of the velocities are randomized for the new particles. This randomization is also necessary to prevent division by zero in the calculation of elapsed time, which is inversely proportional to Eq. (22).
The mean simulated time δ_{g,k} required for one aggregation is the inverse of half the sum of all collision kernels S in Eq. (25):
Its expanded form is derived by combining Eq. (28) with Eqs. (4) and (22) through (25). The result is
which is the expanded form of Eq. (5).
Data availability
All final analyzed data generated during this study are included in this published article plus the accompanying spreadsheet in Supplementary Data: Mass distributions and fluxes by B_{o} parameter.xlxs.
References
Aghanim, N. et al. (Planck Collaboration), Planck 2018 results. VI. Cosmological parameters (2018). arXiv:1807.06209v2 [astroph.CO].
Iocco, F., Pato, M. & Bertone, G. Evidence for dark matter in the inner Milky Way. Nat. Phys. 11, 245–248 (2015).
Steinmetz, M. & Navarro, J. F. The hierarchical origin of galaxy morphologies. New Astron. 7, 155−160 (2002). (Erratum: New Astron. 8, 557–564, 2003).
Salucci, P. The distribution of dark matter in galaxies. Astron. Astrophys. Rev. 27, 2 (2019).
Ostriker, J. P. & Steinhardt, P. New light on dark matter. Science 300, 1909–1913 (2003).
Oerter, R. The Theory of Almost Everything: The Standard Model, the Unsung Triumph of Modern Physics (Penguin Group, London, 2006).
Witten, E. Cosmic separation of phases. Phys. Rev. D 30, 272–285 (1984).
Farhi, E. & Jaffe, R. L. Strange matter. Phys. Rev. D 30, 2379–2391 (1984).
De Rủjula, A. & Glashow, S. L. Nuclearites: a novel form of cosmic radiation. Nature 312, 734–737 (1984).
Zhitnitsky, A. “Nonbaryonic” dark matter as baryonic color superconductor. JCAP 0310, 010 (2003).
Xia, C. J., Peng, G. X., Zhao, E. G. & Zhou, S. G. From strangelets to strange stars: a unified description. Sci. Bull. 61, 172 (2016).
Jacobs, D. M., Starkman, G. D. & Lynn, B. W. Macro dark matter. MNRAS 450, 3418–3430 (2015).
Wandelt, B. D. et al. Selfinteracting dark matter. in Sources and Detection of Dark Matter and Dark Energy in the Universe, from 4th Int’l. Symp., Marina del Rey, CA, USA, February 23–25, 2000 (ed. Cline, D. B.) Ch. 5, 263–274. (Springer, 2001). https://arxiv.org/abs/astroph/0006344 (2000). Accessed on 24 June 2017.
McCammon, D. et al. A high spectral resolution observation of the soft xray diffuse background with thermal detectors. Astrophys. J. 576, 188 (2002).
Tulin, S. SelfInteracting dark matter. AIP Conf. Proc. 1604, 121 (2014).
Tatsumi, T. Ferromagnetism of quark liquid. Phys. Lett. B 489, 280–286 (2000).
VanDevender, J. P. et al. Detection of magnetized quark nuggets, a candidate for dark matter. Sci. Rep. 7, 8758 (2017).
Lugones, G. & Horvath, J. E. Primordial nuggets survival and QCD pairing. Phys. Rev. D 69, 063509 (2004).
Steiner, W. S., Reddy, S. & Prakash, M. Colorneutral superconducting dark matter. Phys. Rev. D 66, 094007 (2002).
Bhattacharyya, A. et al. Relics of the cosmological QCD phase transition. Phys. Rev. D 61, 083509 (2000).
Chodos, A. et al. New extended model of hadrons. Phys. Rev. D 9(12), 3471–3495 (1974).
Aoki, Y., Endr, G., Fodor, Z., Katz, S. D. & Szabó, K. K. The order of the quantum chromodynamics transition predicted by the standard model of particle physics. Nature 443, 675–678 (2006).
Bhattacharya, T. et al. QCD phase transition with chiral quarks and physical quark masses. Phys. Rev. Lett. 113, 082001 (2014).
Gorham, P. W. & Rotter, B. J. Stringent neutrino flux constraints on antiquark nugget dark matter. Phys. Rev. D 95, 103002 (2017).
Ge, S., Lawson, K. & Zhitnitsky, A. The axion quark nugget dark matter model: size distribution and survival pattern. Phys. Rev. D 99, 116017 (2019).
Atreya, A., Sarkar, A. & Srivastava, A. M. Reviving quark nuggets as a candidate for dark matter. Phys. Rev. D 90, 045010 (2014).
Bazavov, A. et al. Additional strange hadrons from QCD thermodynamics and strangeness freeze out in heavy ion collisions. Phys. Rev. Lett. 113, 072001 (2014).
Burdin, S. et al. Noncollider searches for stable massive particles. Phys. Rep. 582, 1–52 (2015).
Chakrabarty, S. Quark matter in strong magnetic field. Phys. Rev. D 54, 1306–1316 (1996).
Peng, G. X., Xu, J. & Xia, C.J. Magnetized strange quark matter in the equivparticle model with both confinement and perturbative interactions. Nucl. Sci. Tech. 27, 98 (2016).
Patrignani, C. et al. Review of particle properties (Particle Data Group). Chin. Phys. C 40, 100001 (2016).
Price, P. B. & Salamon, M. H. Search for supermassive magnetic monopoles using mica crystals. Phys. Rev. Lett. 56(12), 1226–1229 (1986).
Porter, N. A., Fegan, D. J., MacNeill, G. C. & Weekes, T. C. A search for evidence for nuclearites in astrophysical pulse experiments. Nature 316, 49 (1985).
Porter, N. A., Cawley, M. F., Fegan, D. J., MacNeill, G. C. & Weekes, T. C. A search for evidence for nuclearites in astrophysical pulse experiments. Irish Astron. J. 18, 193–196 (1988).
Bassan, M. et al. Dark matter searches using gravitational wave bar detectors: quark nuggets and nuclearites. Astropart. Phys. 78, 52–64 (2016).
Shirk, E. & Price, P. Charge and energy spectra of cosmic rays with Z ≥ 60: the Skylab experiment. Astrophys. J. 220, 719–733 (1978).
VanDevender, J. P., Schmitt, R. G., McGinley, N., VanDevender, A. P., Wilson, P., Dixon, D., Auer, H. & McRae, J. Results of search of magnetized quarknugget dark matter from radial impacts on Earth. https://arxiv.org/abs/2007.04826 (2020).
VanDevender, J. P., Buchenauer, C. J., Cai, C., VanDevender, A. P. & Ulmen, B. A. Radio frequency emissions from darkmattercandidate magnetized quark nuggets interacting with matter. Sci. Rep. 10, 13756 (2020).
Scherrer, R. J. & Turner, M. S. On the relic, cosmic abundance of stable, weakly interacting massive particles. Phys. Rev. D 33, 1585–1589 (1986).
Thompson, C. Giant primeval magnetic dipoles. Astrophys. J. 844, 65 (2017).
Kruis, F. E., Maisels, A. & Fissan, H. Direct simulation Monte Carlo method for particle coagulation and aggregation. AIChE J. 46(9), 1735–1742 (2000).
Zhitnitsky, A. Solar extreme UV radiation and quark nugget dark matter model. arXiv: 1707.03400v2 [astroph.SR] (2017).
Li, C.F., Yang, L., Wen, X.J. & Peng, G.X. Magnetized quark matter with a magneticfield dependent coupling. Phys. Rev. D 93, 054005 (2016).
Peng, C., Peng, G. X., Xia, C. J., Xu, J. F. & Zhang, S. P. Magnetized strange quark matter in equivparticle model with both confinement and perturbative interactions. Nucl. Sci. Tech. 27, 98 (2016).
Kolb, E. W. & Turner, M. S. The Early Universe 47–59 (CRC Press, Boca Raton, 2018).
Friedmann, A. "Über die Krümmung des Raumes". Z. Phys. (in German). 10(1), 377–386 (1922). English translation: Friedmann, A. On the curvature of space. Gen. Relat. Gravit. 31(12), 1991–2000 (1999).
Edwards, B. F., Riffe, D. M., Ji, J. Y. & Booth, W. A. Interactions between uniformly magnetized spheres. Am. J. Phys. 85(2), 130–134 (2017).
Acknowledgements
We gratefully acknowledge S. V. Greene for first suggesting that quark nuggets might explain the geophysical evidence that initiated this research (she generously declined to be a coauthor) and Jesse Rosen for editing the manuscript and suggesting many improvements for clarity.
Author information
Authors and Affiliations
Contributions
J. P. V. was lead physicist and principal investigator. He developed computer program to calculate the quarknugget mass distribution, analyzed the results, wrote the paper, and prepared the figures, and revised the paper to incorporate improvements from the other author and reviewers. I. M. S. was physicist and lead theorist. He provided the critical function of darkmatter density versus time and temperature, consistent with the standard ΛCDM cosmology model, and many improvements to the paper. T. S. was contributing physicist. He provided insight and analysis about the decay of singlets from the weak interaction, which our results show is overcome by aggregation, and about the selfmagnetic field of the A = 1 singlet, which is ~ 10^{12} T and is consistent with the aggregation process. A. P. V. was lead physicist for the Red Team critique of the paper and provided many improvements to the paper through his reviews and suggestions. B. A. U. was contributing physicist and provided useful suggestions on how to detect MQNs, including looking for them in asteroids. All authors reviewed the manuscript.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing interests.
Additional information
Publisher's note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
VanDevender, J.P., Shoemaker, I.M., Sloan, T. et al. Mass distribution of magnetized quarknugget dark matter and comparison with requirements and observations. Sci Rep 10, 17903 (2020). https://doi.org/10.1038/s4159802074984z
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s4159802074984z
This article is cited by

Magnetised quark nuggets in the atmosphere
Scientific Reports (2021)
Comments
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.