Chapter IV: Mechanical Theory of Phase Coexistence
4.4 Motility-Induced Phase Separation
which is an integral condition for no net mass transport between the bulk phases.
By comparing eq.(4.36) and the Maxwell construction (4.7), the integrating fac- tor πΈ(π) = π/π2 up to a nonzero multiplicative constant, which results π = π π(π/π)/ππ.
As a final note, we emphasize that effects ofat leastthe second order density gra- dients need to be considered to properly describe microscopic details of even-order tensorial quantities (e.g. free energy, pressure, stress, etc.). Odd-order density gradients do not affect the even-order tensors due to the reflection symmetry. For example, even if first order gradient terms are present, they do not impart information about microscopic details to even-order tensorial quantities. Thus, for both ther- modynamic and mechanical descriptions of phase interfaces, of which the central physical quantities are free energy and stress, the mass transport condition should include effects of at least the second order gradients in density. Otherwise, we simply obtain macroscopic description of the system as seen in the local density approximation from the preceding section.
drag forces and torques: βππΌπΌ and βππ ππΌ, where π and ππ are translation and rotational drag coefficients. The self-propulsive swim force is ππ π€π ππΌ = π π0ππΌ, whereπ0is the free-space swim velocity andππΌis the direction of swimming. The interparticle collisional force on particleπΌisπππΌ. Our theory applies to any types of interaction potentials but the results presented in this section are obtained with the Weeks-Chandler-Andersen (WCA) potential to highlight a unique property of MIPS
β a liquid-gas phase separation with purely repulsive interactions. The random Brownian force and reorientation torque areππ΅πΌ and π³πΌπ΅ and follow the white noise statistics:
ππΌπ΅(π‘)=0, ππΌπ΅(π‘)ππ΅π½(π‘β²)=2π2π·ππ°πΏπΌ π½πΏ(π‘βπ‘β²), (4.40) π³πΌπ΅(π‘)=0, π³πΌπ΅(π‘)π³π΅π½(π‘β²)=2π2
π /ππ π°πΏπΌ π½πΏ(π‘βπ‘β²), (4.41) where π·π is the translational diffusivity, πΏπΌ π½is the Kronecker delta, and ππ is the time scale of random reorientations.
A dimensional analysis shows that there are four governing dimensionless parame- ters in the problem: π,π0ππ /π·,π2
0ππ /π·π, andπ/(π· π π0), whereπis the volume (or area in 2D) fraction and π· and π are the length and energy parameter of the interaction potential. Here we consider the hard-sphere limit in which the repulsive interactive force is much larger than the Brownian and swim forces: π/(π· π π0) β«1 and π/(π π·π) β« 1. Consequently, the problem is essentially governed by three dimensionless parameters: π,π0ππ /π·, andπ2
0ππ /π·π.
Now we coarse-grain the microscopic dynamics of ABPs using the standard Irving- Kirkwood-Noll procedure [33, 37, 38] (see appendix 4.7 for the derivations of equations in the current section) and obtain a particle balance equation to describe the mass transport in the active suspension:
π π
π π‘
+ β Β· ππ =0, (4.42)
ππ =π0π+ 1 π
πππ₯ π‘π+ β Β·ππ
βπ·πβπ , (4.43)
where π = β¨Γ
πΌπΏ(π βππΌ)β© is the average number density field, ππ is the particle flux, and ππ is the collisional stress of the ABPs. The swimming motion of ABPs brings about a unique advective transport of particles π0π, where π =
β¨Γ
πΌπΏ(πβππΌ)ππΌβ©is the polar order or net-orientation vector field. We note that the particle balance (4.42)-(4.43) is exact.
Equation (4.43) is indeed a mechanical force balance on the ABPs. By multiplying the drag coefficientπto eq. (4.43) and using the Stokes-Einstein-Sutherland relation π π·π = ππ΅π, we obtain
0=βπππ+π π0π+πππ₯π‘π+ β Β· (ππ+ππ π π π), (4.44) whereππ π π π =βπ ππ΅ππ° is the osmotic stress. The suspending medium exerts the drag forceβπππon particles for their relative motion to the medium. The average of the individual swim forces acts as a self-generated βbodyβ force, π π0π, on the particle phase [39] and the divergence of particle stresses balances the forces.
In order to understand how the self-generated body force develops, we obtain an evolution equation for the polar order from the microscopic dynamics:
ππ
π π‘
+ β Β· ππ + πβ1 ππ
π=0, (4.45)
ππ =π0π
πΈ+ 1 π
ππ°
+ 1 π
πππ₯ π‘π+ 1 π
β Β· π π π
βπ·πβπ, (4.46) whereπ(π)is the reduction in the swim velocity from its free- space valueπ0due to collisions with other particles, π is the spatial dimension for reorientation, and πΈ = β¨Γ
πΌπΏ(πβ ππΌ) (ππΌππΌβ π°/π)β© is the nematic order tensor field. The nematic order satisfies an equation analogous to the polar order equation:
ππΈ
π π‘
+ β Β· ππ + 2π ππ
πΈ =0, (4.47)
ππ =π0ππ©+π0πΒ· π
π+2πΆβ 1 π
π° π°
! + 1
π
πππ₯π‘πΈβ 1 ππ
β Β·πππ°βπ·πβπΈ, (4.48) where π© = β¨Γ
πΌπΏ(πβ ππΌ) (ππΌππΌππΌ βπΆ Β· ππΌ/(π +2)β© is the traceless third order orientational moment tensor field andπΆis the fourth order isotropic tensor.
The equations for the polar order (4.45)-(4.46) and nematic order (4.47)-(4.48) show the well-known hierarchical structure of orientational moments due to the coupling between the moments of neighboring orders [39, 40]. Due to the sink terms in the equations, a higher order moment has an order of the gradient of a lower order moment (e.g. π βΌ βπ, πΈ βΌ βπ βΌ ββπ, etc.) for high activities (β2
0/(π·πππ ) β« 1) that are required to achieve MIPS. Thus, for active suspensions the hierarchy of orientational moments naturally introduces microscopic nonlocal effects via orientational moments even without a density-gradient expansion. As discussed at the end of the preceding section, a mechanical description of interfaces
requires effects of at least the second order density gradients. We close the hierarchy by keeping all terms up to the orderπ(π2) in Fourier space assuming reasonably slow variations in the number density and investigate results of the minimal closed set of equations.1
It is notable that at steady state the polar order equation (4.45) allows the self- generated βbodyβ force to be written in a form of divergence: π π0π = β Β·ππ π€π π, whereππ π€π π β‘ π β0ππ/(πβ1)is the swim stress [41] andβ0=π0ππ is the run length in a free space. The swim stress is an effective stress arising from the self-propulsive forces. While the swim stress is not a true stress [36, 42] but a βbodyβ force recast as a form of a stress β analogous to the equivalent (or dynamic) pressure in fluid mechanics [17] or Maxwell stress in electrostatics [43] β2it is useful to determine the force exerted by ABPs on no-flux boundaries at steady state [39]. We show below that this steady-state property of active matter systems is indeed important in determining the phase coexistence criteria of MIPS.
Now let us consider a planar MIPS interface with the normal inπ§direction at steady state when there is no net mass transport ππ,π§ =0. We label the coexisting phases πΌ (π§β ββ) and π½ (π§β +β). Here, we focus on the effects of the self-propelling motion of ABPs to discover the underlying physics of MIPS by considering athermal limit (π·π = 0), but it is straightforward to include the thermal Brownian motion.
Now, the problem is described by two dimensionless parameters: πandβ0/π·. The mechanical force balance and the polar and nematic order equations become
0=π π0ππ§β πΞ π π π§
=0, (4.49)
π π π§
π0π
ππ§ π§+ 1 π
π + πβ1 ππ
ππ§ =0, (4.50)
π π π§
3π π+2 β 1
π
!
π0ππ§ + 1 ππ
πΞ π π π§
! + 2π
ππ
ππ§ π§ =0, (4.51) where Ξ π is the collisional pressure. Combining and integrating eqs. (4.49) and (4.50), we obtain
π π0β0
πβ1π ππ§ π§+Ξ π π€π π +Ξ π =const., (4.52)
1The effects of the correlation between the collisional stress and orientationβ Β· (π π/π)/π (βΌ
π(π2))in the polar order flux (4.46) is investigated in appendix 4.8.
2Treating the swim stress as a true stress leads to extremely negative interfacial tension of MIPS which contradicts a stable interfaces between coexisting phases [36, 44, 45].
whereΞ π π€π π =π π0β0π π/(π(πβ1))is the swim pressure [8, 41]. From the nematic order equation (4.51),ππ§ π§ =0 deep inside phases where spatial gradients vanish due to the homogeneity. The sum of the swim and collisional pressures defines anactive pressure which is identical in the bulk part of coexisting phases. This motivates us to name the undetermined integration constant in eq. (4.49) as the coexistence active pressureΞ ππ ππ₯π ππ‘ :
Ξ ππ ππ₯π ππ‘ = Ξ πΌπ ππ‘ = Ξ π ππ‘π½ . (4.53) The active pressure represents the pressure exerted by ABPs on a no-flux boundary at steady-state and acts as an observable steady-state pressure. For the mechanical stabilityπΞ π ππ‘/π π β₯ 0. Thus, the spinodal is determined byπΞ π ππ‘/π π=0 and the critical point determined byπ2Ξ π ππ‘/π π2=0.
To construct the binodal of MIPS, one more condition is required to determine the coexistence active pressure. Now we need to utilize the microscopic information in the continuum-balance equation. From eqs. (4.49), (4.51) and (4.52) we obtain
Ξ π ππ‘βΞ ππ ππ₯π ππ‘ = 3β2
0
2π(πβ1) (π+2)π π π π§
π
πΞ π π π§
. (4.54)
Using the chain rule eq. (4.54) can be expressed as Ξ π ππ‘βΞ ππ ππ₯π ππ‘ = 3β2
0
2π(πβ1) (π+2)π
"
π
π π
π
πΞ π
π π
ππ π π§
2
+π
πΞ π
π π ππ2 π π§2
#
, (4.55) which has the the same mathematical form with eq. (4.30).3 Thus, Eifantis and Serrinβs analysis [31] applies and there exists an integrating factor (4.35) such that the integration of eq. (4.55) from phase πΌ to π½ makes the second order gradient terms on the right-hand side disappear. Indeed the existence of such integrating factor is more clearly seen by using the collisional pressure gradientπΞ π/π π§for the original equation (4.54):
β« Ξ ππ½
Ξ ππΌ
Ξ π ππ‘ βΞ ππ ππ₯π ππ‘
πΞ π = 3β2
0
2π(πβ1) (π+2)
β« β
ββ
π πΞ π
π π§ π π π§
π
πΞ π π π§
π π§
= 3β2
0
4π(πβ1) (π+2)
"
π
πΞ π π π§
2#π½
πΌ
=0, (4.56) where the last equation is obtained by the homogeneity in each bulk phase.
3We note that the gradient terms are all multiplied byβ2
0and thus are unique to active suspensions.
Gradient terms similar to those present in equilibrium systems could be included in (4.54) but are not necessary. Furthermore, these terms are directly connected to the surface tension of an interface, and MIPS interface has been shown to have essentially zero surface tension [36].
<latexit sha1_base64="FHfErOCEC8wfFgctyvXkeEFGKdM=">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</latexit>
β§
act= β§
c+ β§
swim <latexit sha1_base64="2YXUoUuik0ZJiYKeKexRGDL5Y7A=">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</latexit>v = 1/
2 4 6
8 10 12
8 10 12
2 4 6
Mechanical Theory Maxwell Construction
<latexit sha1_base64="Cxj77y4cF7ds/PaY6rol3mshpSQ=">AAAGeXicddRbT9swFADgwEbH2A027WkvYWUS2kOVwEbHG6Jcxk0UWKES6ZDjuGloHIfELS1Wf8xet1+037KXJafpSelYpKrnO/aRHTu2HfpeLA3j99T0o8czhSezT+eePX/x8tX8wuvzWHQiympU+CKq2yRmvhewmvSkz+phxAi3fXZhtytp+0WXRbEngm+yH7IGJ27gNT1KZJK6mn9riaQ5rVZW1Rt8V1Sw3uBqvmiUTGN9xSjrRmm1vG6srSSBsWZ8Lhu6WTLgKWrZU71amDm0HEE7nAWS+iSOL00jlA1FIulRnw3mrE7MQkLbxGWXSRgQzuKGgvkP9A9JxtGbIkp+gdQhO16hCI/jPreTnpzIVjzZliYfbLP5vZFV2kMK4ccTE5LNLw3lBWFHsoAO59Ps+LoUerpmuuNFjEq/nwSERl7ySjptkYhQmazsnBWwWyo4J4GjLMkGl2ZDWdwWPWXZwnfSMfUlvWjqS4PBZOf44d7K8mRaMfinhHTdsZIlyyeB67O0bwQRFDisaXlBwCJlLVrUEdJazLI2C2PPF0E6trIyjErsbrJdLSbJsHUkbJakM2xJAkzya8hdjxKU+FWl0j+9OparZLkKFm5B3Ra6Bq6ht8Hb6B3wDroKxjHsOriOPgQfovfAe+iJeds9cA89XIX87W2wjd4Eb6L74D66DW7n9QwSLF9tcBc9XNx8ZW/Bt2hgExl7Ls92CkJsmBznDHyGjsARevgx5F8BB3M0fKjKQZ+CT9FH4CP0CfgEfQe+Q++D99EH4IP8xcAx2gW76BAcok2wid4F76JvwDfoCjj/EIkftrKFhDCfOB6Fu/FjIDhzszyE+ULnO3J2b0eO84rj8YpdwpPDCfeAstJjmiSgLLl/R5es/v/gfKVkrpVWTz4VNzazm3hWe6e915Y1UytrG9pXrarVNKop7Yf2U/s186ewWFgufBx2nZ7Kat5o957C6l8vdk+h</latexit>
β§
coex0 5 10
8 10 12
<latexit sha1_base64="Kga4kfu1oWAZDLIMB3QLh72NA9E=">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</latexit>
β§
c8
10 12
<latexit sha1_base64="FHfErOCEC8wfFgctyvXkeEFGKdM=">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</latexit>
β§
act= β§
c+ β§
swim0 4 10
<latexit sha1_base64="Cxj77y4cF7ds/PaY6rol3mshpSQ=">AAAGeXicddRbT9swFADgwEbH2A027WkvYWUS2kOVwEbHG6Jcxk0UWKES6ZDjuGloHIfELS1Wf8xet1+037KXJafpSelYpKrnO/aRHTu2HfpeLA3j99T0o8czhSezT+eePX/x8tX8wuvzWHQiympU+CKq2yRmvhewmvSkz+phxAi3fXZhtytp+0WXRbEngm+yH7IGJ27gNT1KZJK6mn9riaQ5rVZW1Rt8V1Sw3uBqvmiUTGN9xSjrRmm1vG6srSSBsWZ8Lhu6WTLgKWrZU71amDm0HEE7nAWS+iSOL00jlA1FIulRnw3mrE7MQkLbxGWXSRgQzuKGgvkP9A9JxtGbIkp+gdQhO16hCI/jPreTnpzIVjzZliYfbLP5vZFV2kMK4ccTE5LNLw3lBWFHsoAO59Ps+LoUerpmuuNFjEq/nwSERl7ySjptkYhQmazsnBWwWyo4J4GjLMkGl2ZDWdwWPWXZwnfSMfUlvWjqS4PBZOf44d7K8mRaMfinhHTdsZIlyyeB67O0bwQRFDisaXlBwCJlLVrUEdJazLI2C2PPF0E6trIyjErsbrJdLSbJsHUkbJakM2xJAkzya8hdjxKU+FWl0j+9OparZLkKFm5B3Ra6Bq6ht8Hb6B3wDroKxjHsOriOPgQfovfAe+iJeds9cA89XIX87W2wjd4Eb6L74D66DW7n9QwSLF9tcBc9XNx8ZW/Bt2hgExl7Ls92CkJsmBznDHyGjsARevgx5F8BB3M0fKjKQZ+CT9FH4CP0CfgEfQe+Q++D99EH4IP8xcAx2gW76BAcok2wid4F76JvwDfoCjj/EIkftrKFhDCfOB6Fu/FjIDhzszyE+ULnO3J2b0eO84rj8YpdwpPDCfeAstJjmiSgLLl/R5es/v/gfKVkrpVWTz4VNzazm3hWe6e915Y1UytrG9pXrarVNKop7Yf2U/s186ewWFgufBx2nZ7Kat5o957C6l8vdk+h</latexit>
β§
coexMechanical Theory Maxwell Construction
(a)
<latexit sha1_base64="8CqNsSaQ6j/PW+8O/NmFtYncR4c=">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</latexit>
(b)
Figure 4.2: (a) Generalized equal-area construction from the mechanical theory (eqs. (4.53) and (4.56)) in the (Ξ π ππ‘ = [Ξ π +Ξ π π€π π], Ξ π) plane and (b) Maxwell construction forβ0/π· =35Γ2β1/6in 2D. Pressures are normalized byπ π0/(π π·). The shaded areas are visual guides for equal-area constructions.
The equivalence of the active pressures (4.53) and the new integral condition (4.56) allows a new generalized equal-area construction for MIPS rather than the con- ventional Maxwell construction (4.6) for thermodynamic systems. In fig. 4.2(a) we show the equal-area construction of the active pressureΞ π ππ‘ = Ξ π+Ξ π π€π π vs.
Ξ π that determines phase coexistence. From the equal-area construction the colli- sional pressure in the two phases is determined and then the corresponding number densities are found from the equation of state (EOS) for the collisional pressure.
Figure 4.2(b) shows what the corresponding Maxwell construction would predict.
The predicted coexistence pressure is larger and correspondingly the dilute phase
Homogeneous
Phase-separated
0.2 0.4 0.6 0.8
0 0 20 40 60
80
Mechanical TheoryMaxwell Construction
Simulation Critical point (theory) Critical point (simulation)
<latexit sha1_base64="12coR1wtR6xGOseynTwvw4O6xwk=">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</latexit>
Area fraction
<latexit sha1_base64="roeARPautbQuiWchXs/6csVjWdA=">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</latexit>
Ac ti v it y `
0/D
0 0.2 0.4 0.6 0.8 0
20 40 60 80
Figure 4.3: Mechanical theory prediction for the phase behavior of active suspen- sions (solid line) compared with simulations (open circles). Here,β0/π· =π0ππ /π· andπ=π π·2π/4, is the area fraction of active particles of diameter π·. Also shown is the phase boundary from using a Maxwell construction (dotted line).
density is higher.
In order to construct aΞ π ππ‘ vs. Ξ π plot we need expressions for the collisional and swim pressures as a function of the number densityπ. For the equal-area construction the equaion of state (EOS) is required. In the following, we use the EOS of ABPs in 2D from the work by Mallory, Omar, and Brady [8], which includes a careful discussion of how to extrapolate homogeneous system behavior into the two-phase region. The EOS in 3D can be obtained from Brownian dynamics simulations by following the procedures in [8] as
Ξ π
π π0/(π π·2) =6Γ2β7/6 π2
βοΈ1βπ/π0
, (4.57)
Ξ π π€π π
π π0/(π π·2) =π β0
π·
π =π β0
π· 1+
1βexp
β27/6 β0
π·
π 1βπ/π0
β1
,
(4.58) whereπ0 =0.645 is the maximum random packing density achieved in 3D from the simulations. The extrapolation can be challenging, but is made easier here because the collisional pressure is a monotonically increasing function of concentration and, while the swim pressure displays a maximum, the reduction in the swim speed π(π) is a monotonically decreasing function of concentration; their sum displays the van-der Waals-like loop shown in figs. 4.2(a) and 4.2(b).
<latexit sha1_base64="5mOgendgILKSPLXjdlXd98jFHL8=">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</latexit>
z/D <latexit sha1_base64="5mOgendgILKSPLXjdlXd98jFHL8=">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</latexit>
z/D
Figure 4.4: Interfacial profiles of the number density, polar order, and nematic order scaled by the volume of a particleππ =π π·3/6 at the interface forβ0/π· =50 in 3D obtained by the a) Brownian Dynamics (BD) simulation and (b) mechanical theory.
Figure 4.3 shows a comparison of this new mechanical theory for the motility- induced phase separation (MIPS) of ABPs in 2D. Also shown for comparison is the phase coexistence predicted via the a Maxwell construction, which over predicts the concentration in the dilute phase. The same Ξ π are Ξ π π€π π are used in both constructions. There are no adjustable parameters in the comparison.
The mechanical theory not only predicts phase coexistence, but, by its very nature, makes predictions about the structure of the interface that can be compared with simulation. A solution of eq. (4.55) is shown in fig. 4.4, where the interfacial profile, polar order,ππ§and nematic order,ππ§ π§, are shown in comparison to simulation in this case for 3D in the two-phase region at an activityβ0/π· =50. From the particle force balance (4.49), the difference in the collisional pressures between the two phases is balanced by the integral of the swim force: ΞΞ π =β«
π π0ππ§π π§. Particles at the