betaFeedback

WorksPhysics and Astronomy

3D Simulations of Semiconvection in Spheres: Turbulent Mixing and Layer Formation

Abstract

Semiconvection occurs in regions of stars and planets that are unstable to overturning convection according to the Schwarzschild criterion, yet stable according to the Ledoux criterion. Previous simulations in Cartesian boxes have advanced our understanding of the semiconvective instability, layer formation, and transport properties. However, much less is known about semiconvection in spherical geometry and under the influence of rotation or magnetic fields. We present 3D simulations of semiconvection in the full sphere (including r=0r=0), and accounting for rotation. We find that the formation and evolution of semiconvective layers in nonrotating spheres occurs in a similar way to nonrotating Cartesian boxes, in the sense that the critical density ratio at which layers are expected to form is approximately the same in both geometries. Layers rapidly merge once they form, ultimately leading to a fully mixed convective sphere. The transport properties measured through the Nusselt numbers and the buoyancy flux ratio are also similar to results from previous studies in boxes. When rotation is added to the system, layer formation and evolution proceeds in a similar fashion to the nonrotating runs. However, rotation hampers the radial transport of heat and composition, and, as a result, the time required for the sphere to become fully mixed gets longer as the flow becomes more rotationally constrained. We also find that semiconvective layers exhibit spherical mixing in nonrotating cases, whereas in rotating cases, the mixing becomes more cylindrical. We discuss what is needed for future work to build more realistic models.

Introduction

Whether a fluid is stably stratified or unstable to convective overturn plays a crucial role in its dynamics and transport properties. This is particularly true in the interiors of stars and planets, where heat may be transported by convection or by radiation, and which of the two mechanisms is effective determines the evolution of the object. Of particular interest are regions that are unstably stratified in temperature (i.e. unstable to overturning convection according to the Schwarzschild criterion) but stably stratified by a composition gradient (i.e., the fluid is overall stable to overturning convection according to the Ledoux criterion). In this situation, the complex interaction between advection and diffusion of heat and chemical species can trigger an instability that takes the form of overstable gravity waves, driving the fluid to a regime known as double-diffusive convection or semiconvection (see, e.g., [31, 14, 32]). This process enhances the transport of heat and composition within a fluid with respect to just diffusion, and thus plays a crucial role in shaping the internal structure, cooling history, and magnetism of stars and planets.

For example, in intermediate-mass and high-mass stars, nuclear burning gradually creates a mean molecular weight gradient at the edge of the convective core. When the gradient becomes strong enough to stabilize the fluid, the fully convective core shrinks in size. This process leaves behind a semiconvective region, where double-diffusive processes enhance the transport of energy and chemical species between the core and the envelope, which provides additional fuel for nuclear reactions and prolongs the star’s main-sequence phase (see, e.g., [17, 15, 21] (Tayler 1954)). Another example is in the context of giant planets. The formation and evolution models of Jupiter and Saturn predict extended gradients of heavy elements immersed in a hydrogen-helium envelope, with the abundance of heavy elements decreasing toward the planet’s surface (e.g., [26, 35]). This creates a stabilizing compositional gradient, which may lead to semiconvection. In this context, semiconvection has been invoked to explain the abnormally large radius of some giant exoplanets [4], and the anomalous luminosities of Saturn and Uranus [16, 41].

Advances in high-performance computing have made it feasible to study semiconvection in astrophysical fluids using 3D numerical simulations in Cartesian boxes. The seminal calculations of Rosenblum et al. [29] and Mirouh et al. [22] demonstrated that the double-diffusive instability leads to either a state of “homogeneous oscillatory convection” where diffusive transport is only modestly enhanced by small scale turbulence, or to a state of “layered convection”, with numerous, small convective layers separated by thin, diffusive interfaces corresponding to discontinuities in the density of the fluid. The specific state the fluid falls into depends primarily on three non-dimensional parameters: the Prandtl number, the diffusivity ratio, and the density ratio, defined respectively as

Pr=νκT,τ=κCκT,Rρ=NC2NT2,(1)\mathrm{Pr} = \frac{\nu}{\kappa_T}, \quad\tau= \frac{\kappa_C}{\kappa_T}, \quad R_\rho= \frac{N_C^2}{N_T^2}, \tag*{(1)}

where ν\nu is the kinematic viscosity of the fluid, κC\kappa_C and κT\kappa_T are the compositional and thermal diffusivities, respectively, and NT2N_T^2 and NC2N_C^2 are the Brunt–Väisälä frequencies due to the thermal and compositional stratification, respectively. The instability gets excited when

1<Rρ<Pr+1Pr+τ.(2)1 < R_\rho< \frac{\mathrm{Pr}+1}{\mathrm{Pr}+\tau}. \tag*{(2)}

Within this range, spontaneous formation of multiple convective layers is observed for Rρ∈[1,RL]R_\rho\in[1,R_L], while for Rρ∈[RL,(Pr+1)/(Pr+τ)R_\rho\in[R_L,(\mathrm{Pr}+1)/(\mathrm{Pr}+\tau), the instability leads to a state of homogeneous semiconvection. RLR_L corresponds to the point where the composition-to-heat buoyancy flux ratio reaches a minimum with respect to RρR_\rho (see, e.g., Radko [28]; Mirouh et al. [22]). In stars and planets, both Pr\mathrm{Pr} and τ\tau are ≪1\ll1, allowing the double-diffusive instability to develop across a wide range of stratifications RρR_\rho. However, the exact value of the critical density ratio RLR_L, which determines the onset of layer formation, depends on the fluid and flow properties and is difficult to estimate from first principles.

Although further numerical studies have investigated the transport properties of double-diffusive convection (Wood et al. [44]; Moll et al. [25]) and the layer formation process under different boundary conditions (Zaussinger & Kupka [45]; Fuentes et al. [6]; Tulekeyev et al. (2024)), much less is known about semiconvection under the effects of additional physics such as magnetic fields and rotation. Sanghi et al. [30] investigated double-diffusive convection in the presence of a uniform vertical background magnetic field, finding that a sufficiently strong field delays layer formation and substantially reduces the internal flux across the layers. Similarly, uniform rotation in the anti-parallel direction to vertical gravity has been shown to produce similar effects to those of magnetic fields (Moll & Garaud [23]; Liang et al. [18]; Fuentes et al. [7]).

In order to better model real astrophysical objects, it is essential to adopt spherical geometry. For example, in a rapidly rotating planet or star, the latitude-dependent misalignment between gravity and the rotation vector can result in anisotropic mixing and heat transport (e.g., Gastine & Aurnou [11]), thereby influencing the formation and dynamics of semiconvective layers. To our knowledge, the only attempt to study semiconvection in spherical geometry was done by Blies et al. [2], who conducted simulations in rotating spherical shells (i.e., with the coordinate singularity at r=0r=0 avoided by making a “cutout” in the center of the sphere). Similar to previous studies in Cartesian geometries, they found that strong rotation suppresses the transport of both temperature and composition. However, it remains unclear from their results and displayed flow morphologies whether multiple layers developed during the simulations. The full impact of a core cutout on the global solution is unknown. A central cutout forces the use of additional boundary conditions at the shell’s inner surface, potentially affecting flow morphology and transport properties.

In this study, we compute rotationally constrained solutions for semiconvection in the full sphere, including r=0r=0 without the use of a central cutout. We focus on understanding the formation, morphology, and dynamics of semiconvective layers for different degrees of stratification and rotation. In § 2, we describe the set of equations that model the hydrodynamics of semiconvection and the numerical methods. In § 3 and § 4, we present the main results, first discussing non-rotating simulations, and then addressing the effects of rotation. Finally, we conclude in § 5 with a summary and discussion.

Numerical Simulations

In this section, we present the governing fluid equations and boundary conditions. We also describe the parameter space covered by the simulations and the initial conditions. We also briefly outline the numerical methods used.

Model and Initial Conditions

We consider a sphere of radius RR filled with a stably-stratified fluid comprised of two components, one light and the other relatively heavy (e.g., hydrogen and heavy elements). The concentration of the heavy component is represented by CC. We adopt the Boussinesq approximation (Spiegel & Veronis [33]), i.e., we express the fluid density ρ\rho with a linear relationship between the temperature and composition,

ρ=ρ0(1−αT+βC),\rho= \rho_0(1-\alpha T+\beta C),

where ρ0\rho_0 is a fiducial density and β\beta and α\alpha are the coefficients of compositional and thermal contraction/expansion (both assumed positive constants), respectively. Since the Boussinesq approximation is valid for subsonic flows and spatial scales smaller than a density scale height, our simulations should not be interpreted as models of entire stars or planets. Instead, they are more appropriate for the inner core of a gas giant, where the density varies by a factor of 2–4 within 20%–25% of the radius (see, e.g., Hellerd & Howard (2024)). In gas giant interiors, the density scale height is comparable to the planetary radius, making the Boussinesq approximation reasonable for a small core. For stars, this assumption is more questionable, and we discuss its limitations later in the paper.

We initialize the temperature and composition profiles to increase quadratically with depth, i.e., T0(r)=ΔT0(1−r2/R2)T_0(r)=\Delta T_0(1-r^2/R^2) and C0(r)=ΔC0(1−r2/R2)C_0(r)=\Delta C_0(1-r^2/R^2). By this choice the initial density ratio is constant across the sphere (Rρ=βΔC0/αΔT0R_\rho=\beta\Delta C_0/\alpha\Delta T_0).

We present the fluid equations in nondimensional form, using the radius of the sphere RR as the unit of length, and the thermal diffusion time R2/κTR^2/\kappa_T as the unit of time. For the units of temperature and composition, we use [T]=ΔT0[T]=\Delta T_0 and [C]=ΔC0[C] = \Delta C_0, respectively. A unit of pressure corresponds to [P]=ρ0(κT/R)2[P] = \rho_0(\kappa_T/R)^2. Under this normalization, the surface of the sphere is located at r=1r = 1. The dimensionless equations are

∇⋅u=0,(3)\nabla\cdot\boldsymbol{u} = 0, \tag*{(3)}
∂u∂t+u⋅∇u+Pr⁡Ek⁡z^×u=−∇P+Ra⁡Pr⁡(RρC−T)g(r)r^+Pr⁡∇2u,(4)\frac{\partial\boldsymbol{u}}{\partial t} + \boldsymbol{u} \cdot\nabla\boldsymbol{u} + \frac{\operatorname{Pr}}{\operatorname{Ek}} \hat{\boldsymbol{z}} \times\boldsymbol{u} = -\nabla P + \operatorname{Ra}\operatorname{Pr}(R_{\rho}C - T)g(r)\hat{\boldsymbol{r}} + \operatorname{Pr}\nabla^2\boldsymbol{u}, \tag*{(4)}
∂C∂t+u⋅∇C=τ∇2C,(5)\frac{\partial C}{\partial t} + \boldsymbol{u} \cdot\nabla C = \tau\nabla^2 C, \tag*{(5)}
∂T∂t+u⋅∇T=∇2T,(6)\frac{\partial T}{\partial t} + \boldsymbol{u} \cdot\nabla T = \nabla^2 T, \tag*{(6)}

where u\boldsymbol{u} is the velocity field, and g(r)=rg(r) = r is the dimensionless gravity. To be consistent with the initial profiles, and to ensure a stationary solution, we fix the gradient of temperature and composition at the surface of the sphere to match the initial conditions, so that ∂rT∣r=1=∂rC∣r=1=−2\left.\partial_r T\right|_{r=1} = \left.\partial_r C\right|_{r=1} = -2. The boundary conditions for the velocity are impenetrable and stress free, ur=∂r(uθ/r)=∂r(uϕ/r)=0u_r = \partial_r(u_\theta/r) = \partial_r(u_\phi/r) = 0, where uru_r, uθu_\theta and uϕu_\phi, the radial, polar, and azimuthal components of the velocity, respectively.

The set of equations (3)–(6) is governed by five dimensionless numbers: the Prandtl number, the diffusivity ratio, and density ratio (defined in Equation (1)), as well as the Rayleigh number and Ekman number, which are defined respectively as

Ra⁡=αg0ΔT0R3κTν,Ek⁡=ν2ΩR2,(7)\operatorname{Ra} = \frac{\alpha g_0\Delta T_0R^3}{\kappa_T\nu}, \qquad\operatorname{Ek} = \frac{\nu}{2\Omega R^2}, \tag*{(7)}

where g0g_0 is the gravity at r=1r = 1, and Ω\Omega is the rotation rate.

All the simulations were run with the same Rayleigh number, Prandtl number, and diffusivity ratio (Ra⁡=109\operatorname{Ra} = 10^9, Pr⁡=0.1\operatorname{Pr} = 0.1, and τ=0.1\tau= 0.1). Under this choice, double-diffusive instabilities are expected to arise for 1<Rρ<5.51 < R_{\rho} < 5.5. Since we are interested in the layered regime, we varied the stability of the fluid by changing the density ratio RρR_{\rho} from 1.25 up to 2. We set the Coriolis force to be identically zero in the non-rotating cases (Ek⁡=∞\operatorname{Ek} = \infty), while for the rotating cases we vary the Ekman number between Ek⁡≈2.6×10−6\operatorname{Ek} \approx2.6 \times10^{-6}–10−510^{-5}. Real astrophysical conditions, where Ra⁡∼1030\operatorname{Ra} \sim10^{30}, Ek⁡∼10−20\operatorname{Ek} \sim10^{-20}, Pr⁡∼τ∼10−7\operatorname{Pr} \sim\tau\sim10^{-7}–10−210^{-2} [13] are inaccessible for current computational capabilities. However, by fixing Ra⁡≫1\operatorname{Ra} \gg1, Ek⁡≪1\operatorname{Ek} \ll1, Pr⁡<1\operatorname{Pr} < 1, and τ<1\tau< 1, we ensure that the simulations are qualitatively in the same dynamical regime as stars and giant planets.

Simulation Details

We time-evolve equations (3)–(6) using the Dedalus pseudospectral solver [3] version 3. The variables are represented in spherical harmonics for the angular directions and Jacobi polynomials for the radial direction. The number of radial, latitudinal, and longitudinal coefficients in all the simulations are (Nr,Nθ,Nϕ)=(256,384,768)(N_r, N_\theta, N_\phi) = (256, 384, 768), respectively. For time-stepping, we use a second order semi-implicit BDF scheme (SBDF2, [42]), where the linear and nonlinear terms are treated implicitly and explicitly, respectively. To ensure numerical stability, the size of the time steps is set by the Courant–Friedrichs–Lewy (CFL) condition, using a safety factor of 0.2 (based on trial and error). To prevent aliasing errors, I apply the “3/2 rule” in all directions when evaluating nonlinear terms. To start the simulations, we add random-noise temperature perturbations sampled from a normal distribution with a magnitude of 10−510^{-5} compared to the initial temperature field.

Results for Non-rotating Simulations

In this section, we first present qualitative results of the evolution of the simulations for different density ratios. Then, we compare the transport properties with analytical predictions from previous studies.

Qualitative Evolution of the Fluid

We begin with a reference case with stability parameter Rρ=1.25R_{\rho} = 1.25. This case is the closest to overturning convection. We find that the fluid becomes rapidly unstable to semiconvection everywhere. There is a short phase where the system is in the regime of homogeneous oscillatory convection, with weak deviations from the initial density profile. Between t≈0.03t \approx0.03–0.040.04, a few convective layers form spontaneously, with separating interfaces that are highly distorted due to overshooting motions. Since gravity (and in turn, the buoyancy fluxes) increases towards the surface, the system is dominated by a vigorous outer convection zone that propagates inwards, engulfing and mixing everything on its way (Figure 1a–Figure 1d). To clearly see the step-like structures of the layered regime of semiconvection (convective staircases), we show radial profiles of the density in Figure 1e. Instead of performing a full spherical average, we measure the profiles at particular angles θ\theta and ϕ\phi because fluctuations due to overshooting motions cause an artificial interface broadening when horizontally-averaging over all θ\theta and ϕ\phi. Although the profiles are noisy, they are consistent with the snapshot in panel (b), where 3 convective layers are clearly seen at t≈0.036t \approx0.036. Finally, at t≈0.12t \approx0.12, the outer convective layer reaches the center of the sphere and the entire fluid becomes fully-mixed.

Numerical results for the case $R_{\rho}=1.25$, including density-field snapshots and radial density profiles

Figure 1. Numerical results for the case Rρ=1.25R_{\rho}=1.25. Panels (a)–(d): Snapshots of the density field at different times during the evolution of the system. Darker and lighter colors represent denser and less dense fluid, respectively. Panel (e): Radial density profiles for t≈0.03t \approx0.03–0.040.04. As explained in the text, the superposition of different curves is due to radial profiles measured at different angles (ϕ=0\phi= 0, θ=0\theta= 0, π/8\pi/8, π/4\pi/4, π/2\pi/2, π\pi). The different stacks of density profiles are horizontally offset (shifted) from earlier time steps for ease of visualization.

Simulations with larger density ratios (Rρ=1.5,1.75,2R_{\rho} = 1.5, 1.75, 2) exhibit qualitatively similar behavior: multiple convective layers form in the fluid (Figure 2), which subsequently merge and mix until the sphere becomes fully convective. However, the stability of the diffusive interfaces between the layers increases with larger RρR_{\rho}, as the stabilizing effect of the compositional gradient becomes stronger. Consequently, the timescale for the sphere to become fully mixed increases with RρR_{\rho}, with tmix≈0.12,0.175,0.23,0.32t_{\mathrm{mix}} \approx0.12, 0.175, 0.23, 0.32 for Rρ=1.25,1.5,1.75,2R_{\rho} = 1.25, 1.5, 1.75, 2, respectively.

Snapshots of the density field for simulations at different density ratios $R_{\rho}$

Figure 2. Snapshots of the density field for simulations at different density ratios RρR_{\rho}. We selected snapshots at times where multiple convective layers were present in the fluid. Darker and lighter colors represent denser and less dense fluid, respectively.

Heat and Composition Transport

We are interested in the radial fluxes of heat and composition through semiconvection. These fluxes are usually measured in terms of thermal and compositional Nusselt numbers, which are defined as the ratio of the total flux to the diffusive flux

NuT=1−⟨urT⟩⟨dT/dr⟩,(8)\mathrm{Nu}_{T}=1-\frac{\langle u_{r}T\rangle}{\langle\mathrm{d}T/\mathrm{d}r\rangle}, \tag*{(8)}
NuC=1−⟨urC⟩τ⟨dC/dr⟩,(9)\mathrm{Nu}_{C}=1-\frac{\langle u_{r}C\rangle}{\tau\langle\mathrm{d}C/\mathrm{d}r\rangle}, \tag*{(9)}

where the notation ⟨⋅⟩\langle\cdot\rangle represents a volume average over the entire sphere.

Figure 3(a) shows time series of the thermal and compositional Nusselt numbers during the evolution of the simulations. In all cases, the Nusselt numbers increase during different phases characterized by different slopes of the curves with time. First, there is an exponential increase where the semiconvective instability grows and saturates once nonlinear effects become important. Once semiconvection fully estab-

lishes and layers form and evolve, both numbers increase over time but a smaller rate. Later on (starting from the times denoted by the vertical lines in Figure 3a), the increase in the turbulent transport becomes steeper with time as the fluxes become dominated by the more vigorous outer convection zone. Note that due to the reduced overall stability of the flow at lower RρR_{\rho}, the contribution from convective fluxes and consequently the magnitude of Nu, is higher for smaller RρR_{\rho}. Finally, as the entire sphere transitions to a fully convective and well-mixed state, the Nusselt numbers reach a peak and settle to a constant value, independent of RρR_{\rho}, but controlled by the imposed fluxes at the surface of the sphere which are the same for all the simulations.

Figure 3. Panel (a): volume averages of the compositional Nusselt number NuC\mathrm{Nu}_{C} and thermal Nusselt number NuT\mathrm{Nu}_{T}, as a function of time. The vertical lines correspond to the times at which the dynamics of the flow is dominated by the outer convection zone. Results are shown for simulations at different density ratios RρR_{\rho}. Panel (b): (NuC−1)/(NuT−1)(\mathrm{Nu}_{C}-1)/(\mathrm{Nu}_{T}-1) averaged over the semiconvective phase (see text), as a function of density ratio RρR_{\rho}. The lines correspond to theoretical predictions by Spruit [34] (Equation (10)), Rosenblum et al. [29] (Equation (11)), and Wood et al. [44] (Equation (12)).

There exist different models for the relationship between the thermal and compositional Nusselt numbers. For example, in the analytical model of Spruit [34], the Nusselt numbers are related via

NuC−1NuT−1=qRρτ1/2,(10)\frac{\mathrm{Nu}_{C}-1}{\mathrm{Nu}_{T}-1}=\frac{q}{R_{\rho}\tau^{1/2}}, \tag*{(10)}

where qq is a constant of order unity.

On the other hand, the numerical simulations of Rosenblum et al. [29] and Wood et al. [44] propose

NuC−1NuT−1=1Rρτ,(11)\frac{\mathrm{Nu}_{C}-1}{\mathrm{Nu}_{T}-1}=\frac{1}{R_{\rho}\tau}, \tag*{(11)}
NuC−1NuT−1=BARa0.37−1/3Pr1/12τ,(12)\frac{\mathrm{Nu}_{C}-1}{\mathrm{Nu}_{T}-1}=\frac{B}{A}\frac{\mathrm{Ra}^{0.37-1/3}}{\mathrm{Pr}^{1/12}\tau}, \tag*{(12)}

respectively, where the prefactors A≈0.1A \approx0.1 and B≈0.03B \approx0.03 are weakly dependent on RρR_{\rho}, Pr, τ\tau, or Ra.

We tested the three models against measurements of the Nusselt numbers from the simulation results. When averaging over the semiconvective phase (i.e. the time span between the end of the initial exponential grow and the vertical lines in Figure 3(a)), we find that none of the models give the exact measured values from the simulations. However, Spruit’s theoretical prediction (Equation (10)) gives a better fit to the data (Figure 3(b)). Note that Rosenblum’s prediction (Equation (11)) cannot be ruled out, as it provides the correct scaling with the density ratio RρR_{\rho}, and adjusting the prefactor to ≈0.33\approx0.33 also gives a good fit.

Another important quantity in recent studies of semiconvection is the “buoyancy flux ratio” γ−1\gamma^{-1}, defined as the ratio of the buoyancy flux associated with compositional transport to the equivalent buoyancy flux associated with heat transport (see, e.g., Garaud [8] (2021))

γ−1=Rρ⟨urC−τ dC/dr⟩⟨urT−dT/dr⟩.(13)\gamma^{-1}=R_{\rho}\frac{\langle u_r C-\tau\,\mathrm{d}C/\mathrm{d}r\rangle}{\langle u_r T-\mathrm{d}T/\mathrm{d}r\rangle}. \tag*{(13)}

As the sphere becomes mixed over time, the gravitational potential energy released from the thermal stratification must exceed the potential energy required to mix the compositional stratification, implying γ−1<1\gamma^{-1}<1. Figure 4 shows that γ−1\gamma^{-1} remains below unity these simulations. Initially, during the homogeneous phase (before layer formation), γ−1\gamma^{-1} decreases as a function of RρR_{\rho}. Over time, γ−1\gamma^{-1} increases and reaches a peak, although no clear trend emerges for how its maximum value varies with RρR_{\rho}. Previous studies have proposed that γ−1=τ1/2\gamma^{-1}=\tau^{1/2} in experiments of Turner [40]; Linden [19], or γ−1=τ(Pr+1)/(Pr+τ)\gamma^{-1}=\tau(\mathrm{Pr}+1)/(\mathrm{Pr}+\tau) in numerical simulations of semiconvection in triply periodic boxes (e.g., Moll et al. [24]). For τ=Pr=0.1\tau=\mathrm{Pr}=0.1 these predictions yield γ−1≈0.32\gamma^{-1}\approx0.32 and 0.550.55, respectively. In comparison, our results range from 0.3 to 0.4 during the semiconvective regime. Once the sphere becomes fully mixed, both the composition and heat flux reach a statistically steady state, varying linearly with radius. At this stage, the volume-averaged value of γ−1\gamma^{-1} approaches the steady-state solution, τRρ\tau R_{\rho} (horizontal lines in Figure 4). On average, we find that our results are broadly consistent with those from non-rotating simulations in Cartesian boxes (see, e.g., Garaud (2021), for a review).

Figure 4. Time series of the buoyancy flux ratio γ−1\gamma^{-1} (Equation (13)) for different density ratios RρR_{\rho}. The dashed line corresponds to τRρ\tau R_{\rho}, i.e., the expected volume-averaged γ−1\gamma^{-1} when the entire sphere becomes fully mixed and its temperature and composition decrease everywhere at a constant rate.

Results for Rotating Simulations

We analyze the rotating runs using the Rossby number, Ro, of the flow. This nondimensional number expresses the ratio of rotational to convective timescales, so that a system subject to significant Coriolis force possesses low Ro. Convection that is relatively insensitive to rotation is characterized by a high value of Ro. While the Rossby number can be computed using the characteristic speed and length scale of the resultant flow, it can also be estimated a priori by system control parameters, effectively using the freefall time across the domain tff=R/Ufft_{\mathrm{ff}}=R/U_{\mathrm{ff}}, where Uff=(αgΔT0R)1/2U_{\mathrm{ff}}=(\alpha g\Delta T_0R)^{1/2}, as a proxy for the convective timescale

Roc=Uff2ΩR=(αgΔT0R4Ω2R2)1/2=(RaPr)1/2Ek.(14)\mathrm{Ro}_{c}=\frac{U_{\mathrm{ff}}}{2\Omega R}=\left(\frac{\alpha g\Delta T_0R}{4\Omega^{2}R^{2}}\right)^{1/2}=\left(\frac{\mathrm{Ra}}{\mathrm{Pr}}\right)^{1/2}\mathrm{Ek}. \tag*{(14)}

This definition is known as the convective Rossby number. Our rotating simulations have Rρ=1.25R_{\rho}=1.25 and Roc=0.25\mathrm{Ro}_{c}=0.25, 0.5, 1.

Qualitative Evolution and Flow Morphology

The evolution of the rotating runs is qualitatively similar to that of the nonrotating cases, i.e., once the semiconvective instability is triggered, a few convective layers form, which grow and merge until the entire sphere is well-mixed and fully convective. However, there are some significant differences with respect to the nonrotating counterpart. First, the thermal and compositional transport measured through the Nusselt numbers and the buoyancy flux ratio are much smaller than for the nonrotating case (see Figure 5). Consequently, the sphere becomes fully mixed at later times, with the difference respect to the nonrotating case being greater as the Rossby number decreases (tmix≈0.4t_{\mathrm{mix}}\approx0.4, 0.3, 0.26 for Roc=0.25Ro_{\mathrm{c}}=0.25, 0.5, 1, respectively). Further, unlike the nonrotating cases, once the sphere becomes fully convective, the Nusselt numbers are not the same between the different simulations, despite that RρR_{\rho} and the surface gradients are equal in all the simulations. The explanation for this discrepancy is the weaker convective transport and the well-known enhancement of the thermal gradient in rotating convection, which becomes more pronounced as the Rossby number decreases (see, e.g., Barker et al. [1]). This implies that, compared to the non-rotating case, a larger fraction of the heat flux is carried by conduction rather than convection as rotation increases. Overall, rotation exhibits a stabilizing effect on convective fluxes, similar to that of increasing the stability ratio RρR_{\rho} in nonrotating simulations. Note since we run simulations at fixed Rρ=1.25R_{\rho}=1.25, we cannot properly test the analytical predictions in Equations (11)–(12). However, we found that the ratio between the Nusselt numbers varies between 2.5 and 3.5, i.e., close to that of the nonrotating simulation at the same density ratio.

Figure 5. Panel (a): volume averages of the compositional Nusselt number NuC\mathrm{Nu}_{C} and thermal Nusselt number NuT\mathrm{Nu}_{T}, as a function of time. Results are shown for simulations at different convective Rossby numbers (0.25, 0.5, and 1) but fixed Rρ=1.25R_{\rho}=1.25. For comparison, we include the nonrotating case. Panel (b): time series of the buoyancy flux ratio γ−1\gamma^{-1} (Equation 13) for the same cases in panel (a). The dashed line correspond to τRρ\tau_{R_{\rho}}, i.e., the expected volume averaged γ−1\gamma^{-1} when the entire sphere becomes fully mixed and its temperature and composition changes at a constant rate.

Another difference is that rotation significantly affects the spatial scales of the convective flow. In the nonrotating simulations, the flow exhibits approximately isotropic length scales, whereas in the rotating cases, the horizontal length scale (perpendicular to the rotation vector) becomes much smaller than the vertical scale as rotation increases. This behavior is consistent with the Taylor–Proudman theorem, which predicts the formation of vortices and thin columns aligned with the rotation axis (e.g., Proudman [27]; Taylor [38]). This difference is clearly seen for the case at Roc=0.25Ro_{\mathrm{c}}=0.25, at mid and high latitudes (see Figure 6c).

Meridional snapshots of the density field

Figure 6. Meridional snapshots of the density field for simulations at fixed Rρ=1.25R_{\rho}=1.25 but different convective Rossby numbers (1, 0.5 and 0.25 in panels a, b, and c, respectively). We selected snapshots at times where a few convective layers were present in the fluid. Panel (d): Meridional snapshots of the density field at different times for the case at Roc=0.5\mathrm{Ro}_{c}=0.5. In all panels, darker and lighter colors represent denser and less dense fluid, respectively.

To illustrate how the fluid mixes over time, we show in Figure 6d meridional planes of the density field for Roc=0.5Ro_{\mathrm{c}}=0.5. We selected this case because it has an intermediate level of rotational influence, and evolves in a similar way to Roc=0.25Ro_{\mathrm{c}}=0.25, whereas the case Roc=1Ro_{\mathrm{c}}=1 behaves more similar to the nonrotating cases. Initially, the flow is dominated by elongated structures aligned with the rotation axis, which subsequently merge to form 3–4 convective layers. The lack of isotropic density redistribution, which would be characteristic of spherical mixing, further emphasizes the role of rotational effects in organizing the flow. For example, the extent of mixing in the equatorial regions is clearly larger than at the poles. This indicates that mixing is governed by cylindrical geometry, shaped by the interaction between buoyancy and Coriolis forces.

We also observe the formation of prograde zonal flows at the surface, which alternate and reverse to retrograde and then back to prograde as a function of radius. The radial shear between the flows significantly distorts the interfaces between the layers. A comprehensive study exploring the fluid’s transport properties, flow morphology, and zonal flow characteristics as a function of RaRa, RρR_{\rho}, PrPr, and τ\tau will be presented in a future publication.

Latitudinal Dependence of the Heat and Composition Transport

Given the significant differences between convective flows in the polar regions and those at lower latitudes near the equator, we quantify these variations by measuring local Nusselt numbers as a function of latitude. Since we aim to compare with results from previous studies of thermal convection in rotating shells, we show azimuthally averaged profiles of the Nusselt numbers at r≈0.9r\approx0.9, i.e., in the outer convective layer where the fluxes and the velocities are larger. To ensure the flux measurements are captured early in the semiconvective phase and to further reduce noise, we time-average the profiles over Δt≈0.01\Delta t\approx0.01 centered around t≈0.05t\approx0.05.

Figure 7 shows that both compositional and heat transport in the radial direction are more effective near the equator while becoming less efficient at higher latitudes. This aligns well with results from previous studies focusing on thermal convection alone (see, e.g., Figure 3c in Wang et al. [43]). Recently, Gastine & Aurnou [11] showed that differences in heat transport as a function of latitude depend strongly on the supercriticality of the flow, defined as R=RaEk4/3R=\mathrm{Ra}\mathrm{Ek}^{4/3}. For R∼1R\sim1–1010, convective transport is expected to be more efficient at the equator than at the poles, leading to larger Nusselt numbers at the equator. However, for larger supercriticalities (R>10R>10), heat transport at the poles becomes increasingly efficient as RR increases, reaching a point where the equator and poles transport heat equally efficiently for R>100R>100, when rotational effects become less influential. The Nusselt numbers in Figure 7 are measured at times when the fluid is not fully convective, leading to R∼10R\sim10–100100. This means that our results align well with the regime of more efficient convection at the equator.

Image not extracted

Figure 7. Time-averaged local Nusselt numbers normalized by their maximum values (so that they are 1 at the equator) as a function of the colatitude for simulations at Roc=0.25\mathrm{Ro}_{\mathrm{c}} = 0.25, 0.5, and 1. The results are shown within the outer convection zone, at r=0.9r = 0.9, where the fluxes are stronger. The solid and dotted lines distinguish between the compositional and thermal Nusselt numbers, respectively. Similarly to thermal convection in rotating spherical shells, the radial transport of both heat and composition is more efficient at the equator.

Discussion

Summary of the Main Results

In this work, we have studied the onset and evolution of semiconvection in astrophysical fluids, i.e., in fluids of small Prandtl number Pr=ν/κT\mathrm{Pr}=\nu/\kappa_T, and small diffusivity ratio τ=κC/κT\tau=\kappa_C/\kappa_T. Unlike previous work which was limited to conducting simulations on Cartesian boxes, or spherical shells imposing artificial boundary conditions on the inner surface and potentially altering the dynamics, we have solved the Boussinesq fluid equations for the entire sphere, including r=0r=0, for the first time, and including the Coriolis force. For our choice of Pr=τ=0.1\mathrm{Pr}=\tau=0.1, we varied the density ratio RρR_{\rho} between 1.25 and 2 for the non-rotating runs, and we fixed Rρ=1.25R_{\rho}=1.25 while varying the convective Rossby number Roc\mathrm{Ro}_{c} between 0.25 and 1 in the rotating cases.

In all cases, we observed an initial phase of weak, instability-driven turbulence, followed by the formation of a few convective layers in the fluid (see Figure 1, Figure 2, and Figure 6). These layers quickly begin to merge until the sphere becomes fully mixed, with the mixing time increasing with RρR_{\rho} due to the enhanced stability of the composition gradient, and also increasing for smaller convective Rossby numbers Roc\mathrm{Ro}_{\mathrm{c}} due to the much smaller heat and compositional transport hampered by rotation (see measurements of the radial Nusselt numbers in Figure 3(a) and Figure 5(a)). We observe a similar trend during the semiconvective phase (i.e. before the sphere becomes fully convective and well mixed) when comparing the buoyancy flux ratio γ−1\gamma^{-1} for different RρR_{\rho} and Roc\mathrm{Ro}_{\mathrm{c}} (see Figure 4 and Figure 5b). Interestingly, when testing our measurements of the radial Nusselt numbers during the semiconvective phase against theoretical predictions from previous work ([29, 34, 44]), we found that Spruit’s theory (Equation (10)) gives a better agreement with the data. However, Rosenblum’s prediction (Equation (11)) gives the right scaling with the density ratio [NuC−1]/[NuT−1]∝1/Rρ[\mathrm{Nu}_{\mathrm{C}} - 1]/[\mathrm{Nu}_{\mathrm{T}} - 1] \propto1/R_{\rho}, matching the data with a different prefactor equal to 0.33. Therefore, more work is needed to better distinguish between the two models, especially if the prefactor depends on the boundary conditions or geometry of the system.

A noticeable difference between the nonrotating and rotating runs is the morphology of the convective flow. In the nonrotating case, convection exhibits nearly isotropic spatial scales, whereas in the rapidly rotating cases, the flows become anisotropic, with longer axial scales compared to horizontal scales (see Figure 2 and Figure 6). These differences are also reflected in the radial transport as a function of colatitude (Figure 7). Near the poles, the radial component of the convective flow is weaker due to the columnar morphology of the fluid, which favors horizontal fluid motion. In contrast, radial transport is more efficient at the equator because the Coriolis force does not significant affect radial motions.

Comparison to Previous Work

Semiconvection in our non-rotating simulations in the full sphere behaves qualitatively similar to previous work on non-rotating Cartesian boxes, with only minor differences with respect to the critical density ratio at which layers form. For example, in the suite of simulations by Mirouh et al. [22], for Pr=τ=0.1\mathrm{Pr} = \tau= 0.1, layer formation occurs when Rρ≈1.5R_{\rho} \approx1.5. In these simulations, we find layer formation even for Rρ=2R_{\rho} = 2. The fluxes measured through the buoyancy flux ratio are similar. For example, Mirouh et al. [22] measured γ−1∼0.3\gamma^{-1} \sim0.3–0.5 for the same values of RρR_{\rho}, Pr\mathrm{Pr}, and τ\tau considered in this work (see their Table 5). We measured γ−1≈0.3\gamma^{-1} \approx0.3–0.4 during the semiconvective phase. A more significant difference is in the Nusselt numbers, where Mirouh et al. [22] measured NuC∼10\mathrm{Nu}_{\mathrm{C}} \sim10, NuT∼1\mathrm{Nu}_{\mathrm{T}} \sim1–5, while we measured NuC∼100\mathrm{Nu}_{\mathrm{C}} \sim100, NuT∼10\mathrm{Nu}_{\mathrm{T}} \sim10. These differences are not surprising given the significant differences in simulation setups. Our simulations were conducted on a full spherical domain, with gravity and fluxes increasing toward the surface. In contrast, the simulations in Mirouh et al. [22] were performed in Cartesian boxes with constant gravity, triply periodic boundary conditions, and no external flux forcing beyond the initial profiles. However, we emphasize that the formation and evolution of the system are qualitatively similar, with no significant differences.

Comparison with previous work that includes rotation is more difficult because very little is known about rotating semiconvection. Blies et al. [2] investigated semiconvection in rotating spherical shells, but it remains unclear whether they observed layer formation. The Rayleigh number used in their simulations was ∼107\sim10^{7}, which corresponds to a domain size of about 50d50d, where d≈(κν/αg∣dT/dz∣)1/4d \approx(\kappa\nu/ \alpha g |dT/dz|)^{1/4} is the characteristic scale of double-diffusive eddies. Typically, in simulations of semiconvection, the first layers that form are never thinner than 30d30d–50d50d (Garaud 2021). So, it is likely that Blies et al. [2] did not find layers because of the smaller Rayleigh number (our simulations have Ra=109\mathrm{Ra} = 10^{9}, which corresponds to ∼180d\sim180d, so that a few layers are expected to form). Nevertheless, consistent with the findings of this study, they reported that rotation reduces both heat and compositional fluxes. When testing the theoretical predictions described earlier, they also found that Spruit’s theory provides a better approximation to the data.

An intriguing suite of simulations was conducted by Moll & Garaud (2017) on Cartesian boxes. Similar to the results presented here, they demonstrated that rotation consistently diminishes thermal and compositional transport across the system. Interestingly, in some of their simulations, where the flow was strongly constrained by rotation (low Rossby numbers), layer formation was suppressed, and large-scale vortices formed instead, enhancing compositional transport. However, their results appear to be highly sensitive to the size of the boxes adopted in the simulations, likely due to their use of triply periodic boundary conditions. More recently, Fuentes et al. [7] conducted 3D simulations on Cartesian boxes but using fixed flux boundary conditions. They demonstrated that rotation reduces the kinetic energy flux available for mixing, and, as consequence, rotation prolongs the lifetime of semiconvective layers.

Finally, the latitudinal dependence of the fluxes in our simulations aligns well with previous studies of thermal convection in rapidly rotating shells (e.g., Wang et al. [43]; Gastine & Aurnou [11]).

Limitations and Future Work

We have made a number of approximations that must be relaxed before making conclusions for stars and planets. First, we limited the simulations to fluids of Pr=τ=0.1Pr = \tau= 0.1. These values are at the middle of values expected to occur in planetary interiors, where the Prandtl number and diffusivity ratio may extend down to ∼0.01\sim0.01 (e.g., Stevenson & Salpeter [36]; French et al. [5]). In stellar interiors, even lower values are expected. For example, in non-degenerate regions of MS and RGB stars, Pr∼10−6Pr \sim10^{-6} and τ∼10−7\tau\sim10^{-7} (e.g., Garaud et al. [10]). These parameter values are beyond current computational capabilities.

Second, we adopted the Boussinesq approximation, which restricts the vertical scale to be much smaller than a density scale height. Incorporating density stratification would change the mixing rate of the layers through a combination of two effects. First, the potential energy needed for mixing would decrease. While composition is uniformly mixed within a convective layer in both cases, stratified convection adjusts the density to the adiabatic gradient, whereas Boussinesq convection maintains a constant density, which requires more energy. Second, stratification enhances the asymmetry between upflows and downflows (Meakin & Arnett [20]), thereby increasing the net kinetic energy flux available for mixing. At present, we cannot speculate on the net impact of these effects, since, to the best of our knowledge, there are no simulations of semiconvection incorporating density stratification.

Third, we imposed fixed temperature and composition gradients at the surface of the sphere, which, despite that do not affect the layer formation process, are certainly not realistic boundary conditions for stars or planets. For example, in giant planets, the surface flux decreases over time, and there are no composition fluxes at the surface. In stars, internal heat generation and compositional changes from nuclear burning play a significant role, both of which are not included in these simulations.

Improvements in numerical modeling, focusing on adding density stratification, and the use of more suitable boundary conditions, would be of great interest to inform 1D evolution models and interpret the rich set of observations of stars and planets.

I am grateful to Andrew Cumming, Bradley Hindman, and Pascale Garaud for many interesting conversations and valuable suggestions. I am also grateful to the Institute for Pure and Applied Mathematics (IPAM) at UCLA, for support and hospitality during the workshop “Rotating Turbulence: Interplay and Separability of Bulk and Boundary Dynamics” supported by NSF grant DMS-1925919. This research was supported by NASA Solar System Workings grant 80NSSC24K0927.

References

  1. [1]Barker, A. J., Dempsey, A. M., & Lithwick, Y. 2014, ApJ, 791, 13
  2. [2]Blies, P., Kupka, F., Zaussinger, F., & Hollerbach, R. 2014, arXiv e-prints, arXiv:1404.6086arxiv.org/abs/1404.6086
  3. [3]Burns, K. J., Vis, G. M., Oishi, J. S., Lecoeanet, D., & Brown, B. P. 2020, PhRvR, 2, 023068
  4. [4]Chabrier, G., & Baraffe, I. 2007, ApJL, 661, L81
  5. [5]French, M., Becker, A., Lorenz, W., & et al. 2012, ApJS, 202, 5
  6. [6]Fuentes, J. R., Cumming, A., & Anders, E. H. 2022, PhRvF, 7, 124501
  7. [7]Fuentes, J. R., Hindman, B. W., Fraser, A. E., & Anders, E. H. 2024, ApJL, 975, L1
  8. [8]Garaud, P. 2018, AnRFM, 50, 275
  9. [10]Garaud, P., Medrano, M., Brown, J. M., Mankovich, C., & Moore, K. 2015, ApJ, 808, 89
  10. [11]Gastine, T., & Aurnou, J. M. 2023, JFM, 954, R1
  11. [13]Jermyn, A. S., Anders, E. H., Lecoeanet, D., & Cantiello, M. 2022, ApJS, 262, 19
  12. [14]Kato, S. 1966, PASJ, 18, 374
  13. [15]Langer, N. 1991, A&A, 252, 669
  14. [16]Leconte, J., & Chabrier, G. 2013, NatGe, 6, 347
  15. [17]Ledoux, P. 1947, ApJ, 105, 305
  16. [18]Liang, Y., Carpenter, J. R., & Timmermans, M.-L. 2021, JPO, 51, 3335
  17. [19]Linden, P. F. 1975, JFM, 71, 385
  18. [20]Meakin, C. A., & Arnett, D. 2007, ApJ, 667, 448
  19. [21]Merryfield, W. J. 1995, ApJ, 444, 318
  20. [22]Mirouh, G. M., Garaud, P., Stellmach, S., Traxler, A. L., & Wood, T. S. 2012, ApJ, 750, 61
  21. [23]Moll, R., & Garaud, P. 2017, ApJ, 834, 44
  22. [24]Moll, R., Garaud, P., Mankovich, C., & Fortney, J. J. 2017, ApJ, 849, 24
  23. [25]Moll, R., Garaud, P., & Stellmach, S. 2016, ApJ, 823, 33
  24. [26]Müller, S., Helled, R., & Cumming, A. 2020, A&A, 638, A121
  25. [27]Proudman, J. 1916, RSPRSA, 92, 408
  26. [28]Radko, T. 2003, JFM, 497, 365
  27. [29]Rosenblum, E., Garaud, P., Traxler, A., & Stellmach, S. 2011, ApJ, 731, 66
  28. [30]Sanghi, A., Fraser, A. E., Tian, E. W., & Garaud, P. 2022, ApJ, 935, 33
  29. [31]Schwarzschild, M., & Härm, R. 1958, ApJ, 128, 348
  30. [32]Spiegel, E. A. 1969, CoASP, 1, 57
  31. [33]Spiegel, E. A., & Veronis, G. 1960, ApJ, 131, 442
  32. [34]Spruit, H. C. 2013, A&A, 552, A76
  33. [35]Stevenson, D. J., Bodenheimer, P., Lissauer, J. J., & D’Angelo, G. 2022, PSJ, 3, 74
  34. [36]Stevenson, D. J., & Salpeter, E. E. 1977, ApJS, 35, 221
  35. [38]Taylor, G. I. 1917, RSPSA, 93, 99
  36. [40]Turner, J. 1965, IJHMT, 8, 759
  37. [41]Vazan, A., & Helled, R. 2020, A&A, 633, A50
  38. [42]Wang, D., & Ruuth, S. J. 2008, JCM, 26, 838
  39. [43]Wang, G., Santelli, L., Lohse, D., Verzicco, R., & Stevens, R. J. A. M. 2021, GeoRL, 48, e95017
  40. [44]Wood, T. S., Garaud, P., & Stellmach, S. 2013, ApJ, 768, 157
  41. [45]Zaussinger, F., & Kupka, F. 2019, ThCFD, 33, 383

Paper details

vlnc:W_p8pc23me
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

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