Jeffery orbits for an object with discrete rotational symmetry

Ishimoto, Kenta

doi:10.1063/5.0015056

We theoretically investigate the motions of an object immersed in a background flow at a low Reynolds number, generalizing the Jeffery equation for the angular dynamics to the case of an object with n-fold rotational symmetry (n ≥ 3). We demonstrate that when n ≥ 4, the dynamics are identical to those of a helicoidal object for which two parameters related to the shape of the object, namely, the Bretherton constant and a chirality parameter, determine the dynamics. When n = 3, however, we find that the equations require a new parameter that is related to the shape and represents the strength of triangularity. On the basis of detailed symmetry arguments, we show theoretically that microscopic objects can be categorized into a small number of classes that exhibit different dynamics in a background flow. We perform further analyses of the angular dynamics in a simple shear flow, and we find that the presence of triangularity can lead to chaotic angular dynamics, although the dynamics typically possess stable periodic orbits, as further demonstrated by an example of a triangular object. Our findings provide a comprehensive viewpoint concerning the dynamics of an object in a flow, emphasizing the notable simplification of the dynamics resulting from the symmetry of the object’s shape, and they will be useful in studies of fluid–structure interactions at a low Reynolds number.

I. INTRODUCTION

Ever since Stokes¹ first derived an expression for the hydrodynamic drag on a moving sphere in a viscous fluid, the so-called Stokes law of drag, microscopic particles in fluids have usually been modeled as spheres. Microparticles in nature, however, exhibit a tremendous diversity in shape, including not only bodies of revolution such as rods and disks but also objects with discrete rotational symmetry such as snow crystals² and planktonic microorganisms, the latter of which are illustrated in Haeckel’s famous book Art Forms in Nature.³

The first extension from a sphere to be considered was a spheroid, the motion of which was analyzed mathematically by Jeffery⁴ who derived the equations for the angular motion of a spheroid and found a nonlinear periodic orbit in a simple shear, which is now known as the “Jeffery orbit.” The concept of periodic Jeffery orbits was later extended to arbitrary bodies of revolution by Bretherton,⁵ with further mathematically rigorous proofs subsequently being provided.^6,7 The angular dynamics of any body of revolution are then described by an equation with a single parameter related to the aspect ratio of the object, which is now known as the “Bretherton constant.” Such a unified expression for a large class of object geometries suggests that symmetry arguments could allow the classification of particle dynamics in a shear flow into a smaller number of categories.

Recently, with attention focused on the time-averaged nature of the hydrodynamics of microswimmers in a flow, the Jeffery equations of angular dynamics have been extended to helicoidal objects, including chiral objects such as helices.⁸ A helicoidal object is mathematically defined as either a passive particle or an active swimmer whose resistance tensors are unchanged under a π/2-rotation around an axis. This is the symmetry referred to as helicoidal symmetry,^6,9 first introduced in the late nineteenth century in the theory of inviscid flow^10,11 (see Ref. 12 for more details).

Although, strictly speaking, a simple helix is not a helicoidal object according to this definition, a good description of the associated hydrodynamics can still be obtained by considering it as if it were such an object.^13,14 Moreover, the rapid rotation of a bacterial flagellar swimmer around its rotation axis allows a description of such a microswimmer as a helicoidal object,^15–17 and the biased locomotion of a bacterial swimmer in a shear flow, known as bacterial rheotaxis, has been successfully formulated.¹⁸ Such a representation is also applicable to a large class of microswimmers, considering their time-average motions over a beat cycle.¹⁹

Under the assumption of helicoidal symmetry, the Jeffery equations can be derived as follows, with an additional constant arising from the chirality of the object:⁸

\dot{d} = Ω^{\infty} \times d + β (I - d d) \cdot E^{\infty} \cdot d + α [(I - d d) \cdot E^{\infty} \cdot d] \times d,

(1)

where the director vector d represents the axis of rotation, Ω^∞ and E^∞ are the background rotational velocity vector (equal to half the background vorticity) and the rate-of-strain tensor of the background flow, respectively, and I is the identity tensor. The shape parameter β is the Bretherton constant reflecting the aspect ratio of the object. The new constant in the last term, α, represents the effects arising from the chirality of the object, and we shall refer to it in this paper as the “chirality parameter.” An equivalent parameter has been introduced in models of chiral objects.^17,20 The angular dynamics described in Eq. (1) have been investigated in detail for an object in a simple shear flow, and it has been found that if the Bretherton constant is nonzero, the chirality parameter α can orient the director vector so that it is aligned parallel or antiparallel to the background vorticity vector.

The motions of nonspherical microscopic particles in a turbulent flow have been intensively studied in recent years,^20–27 and simple equations such as Eq. (1) could be usefully applied not only to passive particles in turbulence but also to active swimmers,^28–31 for example, in studies of the distribution of planktonic microalgae in the turbulent ocean.^32,33

Drift motions in flows are affected by object chirality, and this has been utilized for sorting in microdevices.^34–36 In the context of active swimmers in a moderately strong background flow, the resulting hydrodynamic interactions could orient the swimming direction.^{17,18,37–39} Another possible application of the associated equations is to the study of Brownian motion and the rheology of nonspherical particles.^40–43

In this paper, we extend the Jeffery equations to an object with discrete rotational symmetry, which implies the existence of an n-fold rotational axis, where n ≥ 3 is an integer.

As presented in Haeckel’s book,³ the diatoms, a major group of microalgae producing a significant portion of atmospheric oxygen, are known for their symmetric shell shape.⁴⁴ For example, Triceratium and Trigonium possess n-gonal symmetry, the rod-like species Thalassiothrix and Asterionella form star-shaped colonies, and species with more complicated triangular shapes, such as Hydrosera,⁴⁵ are also known. Some other protozoa, the radiolarians, produce mineral skeletons with discrete rotational symmetry, including a triangular pyramidal shape in the case of Lithochytris pyramidalis, threefold rotational symmetry such as in Thyrsocyrtis and Podocyrtis,^46,47 and a cell with fourfold rotational symmetry in the case of Lithoptera muelleri.⁴⁸ These microorganisms can also be found in fossils and have been used in research on ancient climate and marine environments. Other microorganisms, for example, discoasters, distinct star-shaped marine algae (now extinct),⁴⁹ do not necessarily have a mirror symmetry in the plane perpendicular to the rotation axis. Microorganisms with discrete rotational symmetry include other microalgae such as Dictyochales (also known as Silicoflagellates because of their siliceous skeleton).^50,51

The Jeffery orbits for a rigid particle with discrete rotational symmetry have been investigated by Fries et al.⁵² who studied the angular dynamics of a particle with n-fold rotational symmetry and an additional mirror plane to decouple the angular dynamics from the translational motions. The primary aim of our paper is therefore to generalize the Jeffery equations to an arbitrary microscopic object, which can be either a passive particle or an active swimmer, with n-fold rotational symmetry but without a mirror plane, where n ≥ 3. Following the Schoenflies notation⁵³ for specifying the symmetry of an object, such as C_n for an object with an n-fold rotational symmetry, we call the object by its specified symmetry, for example, a C_n-object. Our secondary aim is to classify the shape of an object based on its dynamics in a flow, showing that the number of parameters in the generalized Jeffery equations is reduced from the original degrees of freedom in the resistance tensors.

Another important type of discrete symmetry can be found in a particle with triaxially mirror-symmetric planes, such as an ellipsoid. The equation for the dynamics of an ellipsoidal particle in a shear flow was also derived by Jeffery,⁴ and numerical studies^54,55 revealed that such a particle can exhibit quasiperiodic and chaotic rotational motion. More recently, Thorp and Lister⁵⁶ derived the general angular dynamics for non-axisymmetric particles with two planes of symmetry, demonstrating that the angular dynamics are identical to those of ellipsoids with three constants of particle shape.

The contents of the remainder of this paper are as follows: In Sec. II, we introduce hydrodynamics at a low Reynolds number to describe the dynamics of an object in a background flow, together with the symmetry arguments regarding resistance tensors under n-fold rotational symmetry, demonstrating that when n ≥ 4, the dynamics coincide with those obtained for a helicoidal object, which is a C₄-object. We also show that when n = 3, we need additional terms in the equations that reflect the triangularity of the object shape. Then, in Sec. III, we derive expressions for the linear and angular velocities of an arbitrary C_n-object, where n ≥ 3. We also discuss the reduction of the Jeffery equations by additional symmetries. In Sec. IV, we examine the angular dynamics in more detail for a C₃-object in a simple shear flow. Since new contributions will appear as the object becomes less symmetric, we will discuss each new term in the Jeffery equations in terms of the symmetry class. We present an example of a triangular object, and we estimate the new constants for the triangularity by detailed calculations in Sec. V. This is followed by our concluding remarks in Sec. VI.

II. HYDRODYNAMICS AND SYMMETRY

A. Stokes hydrodynamics

We consider a neutrally buoyant microscopic object immersed in a background flow field u^∞ and assume that the flow field around the object obeys the Stokes equation,

η Δ u = \nabla p,

(2)

where p is the pressure field, the constant η is the viscosity, and the velocity field u satisfies the incompressibility condition ∇ · u = 0.

On the surface of the object, S, we impose the no-slip boundary condition. To describe the surface of the object, we introduce a laboratory-fixed reference frame {e_i} (i = 1, 2, 3) and a body-fixed reference frame ${{\hat{e}}_{i}}$ whose origin is denoted by x₀, as illustrated in Fig. 1. We allow the object to be self-propelled by its deformation, and we decompose the velocity into rigid body motion and the deformation velocity as follows:^57,58

u (x) = U + Ω \times r + u^{'},

(3)

where r = x − x₀, U and Ω are the translational and rotational velocities, respectively, and u′ represents the surface deformation velocity.

FIG. 1.

$FIG. 1. Schematic of an object with discrete rotational symmetry immersed in a linear background flow u∞. The laboratory-fixed reference frame is {ei}, and the body-fixed frame is {e^i} whose origin is denoted by x0. The vector r represents the relative position with respect to the origin. The translational and rotational velocities of the object are U and Ω, respectively.$

View large Download slide

Schematic of an object with discrete rotational symmetry immersed in a linear background flow u^∞. The laboratory-fixed reference frame is {e_i}, and the body-fixed frame is ${{\hat{e}}_{i}}$ whose origin is denoted by x₀. The vector r represents the relative position with respect to the origin. The translational and rotational velocities of the object are U and Ω, respectively.

In the far field away from the object, the velocity fields should approach the background velocity for which we assume a linear flow in this study. The velocity gradient tensor G = ∇u^∞ is therefore uniform in space. Introducing the background linear velocity U^∞, the background angular velocity Ω^∞, which is obtained from the antisymmetric part of the velocity gradient tensor, and the rate-of-strain tensor E^∞ = (G + G^T)/2, we have the boundary condition in the far field as

u (x) \to u^{\infty} = U^{\infty} + Ω^{\infty} \times r + r \cdot E^{\infty} (| r | \to \infty),

(4)

where superscript T indicates the transpose of a tensor.

From the linearity of the Stokes equations, the hydrodynamic force and torque acting on the object may be written as combinations of drag- and shear-induced effects and propulsion by surface deformation^6–8 via

F = F_{d r a g} + F_{s h e a r} + F_{p r o p},

(5)

M = M_{d r a g} + M_{s h e a r} + M_{p r o p} .

(6)

The hydrodynamic drag force and torques are

F_{d r a g} = K \cdot (U^{\infty} - U) + C^{T} \cdot (Ω^{\infty} - Ω),

(7)

M_{d r a g} = C \cdot (U^{\infty} - U) + Q \cdot (Ω^{\infty} - Ω),

(8)

respectively, where the second-rank resistance tensors K and Q, which are symmetric and negative-definite, are known as the translational and rotational tensors, respectively, and the second-rank (pseudo)tensor C, which is not necessarily symmetric, is known as the coupling tensor. Using the third-rank shear-force tensor $Γ$ and the shear-torque (pseudo)tensor $Λ$ ⁠, we can write the shear-induced force and torque as

F_{s h e a r} = Γ : E^{\infty},

(9)

M_{s h e a r} = Λ : E^{\infty},

(10)

respectively. Here, the double dot products are defined such that ${[Γ : E^{\infty}]}_{i} = Γ_{i j k} E_{k j}^{\infty}$ ⁠, where there is a summation over the repeated indices, following the Einstein convention that will be used throughout this paper. From the symmetric property of the rate-of-strain tensor, without loss of generality, we can define $Γ$ and $Λ$ so that they are symmetric with respect to their second and third indices, i.e., Γ_ijk = Γ_ikj and Λ_ijk = Λ_ikj. We use the transposition symbol T to indicate the exchange of the last two indices in the third-rank tensors, enabling us to rephrase the relations as $Γ$ ^T = $Γ$ and $Λ$ ^T = $Λ$ ⁠. Note that the tensors K, Q, C, $Γ$ ⁠, and $Λ$ are determined only by the shape of the object and the viscosity constant.

In the last terms in Eqs. (5) and (6), the vectors F_prop and M_prop represent the propulsive force and torque generated by the surface deformation of the object,^57–60

F_{p r o p} = \int_{S} u^{'} \cdot Σ d S,

(11)

M_{p r o p} = \int_{S} u^{'} \cdot Π d S,

(12)

where the second-rank tensors $Σ$ and $Π$ are the translational and rotational surface force resistance tensors, defined by the decomposition of the local hydrodynamic force on the surface for an arbitrary rigid motion⁶¹ such that $\hat{f} = Σ \cdot \hat{U} + Π \cdot \hat{Ω}$ ⁠, and these tensors are also dependent only on the shape of the object and the viscosity constant. Henceforth, we focus on their dependence on the object shape.

B. Hydrokinetic symmetry

We now proceed to symmetry arguments that enable us to reduce the degrees of freedom in the resistance and shear-related tensors, following the approach of Brenner.^6,9 As repeatedly noted in Sec. II A, these tensors are determined only by the shape of the object, and therefore, their matrix representations with respect to body-fixed coordinates should inherit the symmetry of the object shape. We consider a transformation of the body-fixed frame ${{\hat{e}}_{i}}$ into another body-fixed frame ${{\hat{e}}_{i}^{'}}$ via rotation and/or reflection. Let us write the matrix representations of an rth-rank tensor X with respect to the body-fixed frames ${{\hat{e}}_{i}}$ and ${{\hat{e}}_{i}^{'}}$ as $X_{i_{i} i_{2}, \dots, i_{n}}$ and $X_{i_{i} i_{2}, \dots, i_{n}}^{'}$ ⁠, respectively. The transformation is then given by

X_{i_{1} i_{2} \dots i_{r}}^{'} = a_{i_{1} j_{1}} a_{i_{2} j_{2}} \dots a_{i_{r} j_{r}} X_{j_{i} j_{2} \dots j_{r}}

(13)

using the orthogonal matrix a_ij with its determinant |a_ij| = σ = ±1. Similarly, we have the following relation between the matrix representations of an rth-rank pseudotensor Y under the transformation of the body-fixed frame:

Y_{i_{1} i_{2} \dots i_{r}}^{'} = σ a_{i_{1} j_{1}} a_{i_{2} j_{2}} \dots a_{i_{r} j_{r}} Y_{j_{i} j_{2} \dots j_{r}} .

(14)

If the shape of an object is unchanged under a transformation, we say that the object possesses the symmetry associated with the transformation matrix. A series of symmetries are discussed by Brenner,^6,9 and in this paper, we focus on the symmetry with n-fold rotation (n ≥ 3), which is referred to as “generalized helicoidal symmetry” in Brenner’s papers.

Adopting the Schoenflies notation that is commonly used to categorize point groups in the studies of the molecular structure,⁵³ let us call an object with n-fold rotational symmetry a C_n-object. We set the axis of rotation to be ${\hat{e}}_{1}$ ⁠, and the transformation matrix is then given by

a_{i j} = (\begin{matrix} 1 & 0 & 0 \\ 0 & \cos Θ & \sin Θ \\ 0 & - \sin Θ & \cos Θ \end{matrix}),

(15)

where the angle Θ = 2π/n. For a C_n-object, the matrix representations in Eqs. (13) and (14) should be unchanged after the transformation in Eq. (15), and we obtain $X_{i_{1} i_{2} \dots i_{r}}^{'} = X_{i_{1} i_{2} \dots i_{r}}$ for the three resistance tensors and two shear-related tensors, since the determinant of a_ij is σ = 1.

For the resistance tensor K, its matrix representation with respect to the body-fixed frame satisfies K_ij = a_ika_jlK_kl, and similar relations should hold for Q_ij and C_ij. When equating the matrices obtained by Θ/2- and −Θ/2-rotations, we may rephrase the relations for each component as

K_{12} = K_{21} = K_{31} = K_{13} = 0,

(16)

K_{23} + K_{32} = 0,

(17)

K_{22} - K_{33} = 0,

(18)

with similar relations holding for Q_ij and C_ij. For the symmetric tensors K and Q, the off-diagonal components should vanish from Eq. (17), and thus, the three resistance tensors for a C_n-object are given by

K = K_{1} {\hat{e}}_{1} {\hat{e}}_{1} + K_{2} (I - {\hat{e}}_{1} {\hat{e}}_{1}),

(19)

Q = Q_{1} {\hat{e}}_{1} {\hat{e}}_{1} + Q_{2} (I - {\hat{e}}_{1} {\hat{e}}_{1}),

(20)

C = C_{1} {\hat{e}}_{1} {\hat{e}}_{1} + C_{2} (I - {\hat{e}}_{1} {\hat{e}}_{1}) + C_{23} ({\hat{e}}_{2} {\hat{e}}_{3} - {\hat{e}}_{3} {\hat{e}}_{2}),

(21)

where I is the second-rank identity tensor, and the constants K₁, K₂, Q₁, and Q₂ are all negative.

For the shear-related tensors, we again consider the matrix representations, which satisfy the relation Γ_ijk = a_ila_jma_knΓ_lmn. When n ≥ 3, this may be expressed as the following relations for each component, after equating the two matrix representations transformed by Θ/2- and −Θ/2-rotations:

Γ_{112} = Γ_{113} = Γ_{121} = Γ_{131} = Γ_{211} = Γ_{311} = 0,

(22)

Γ_{123} = Γ_{132} = 0,

(23)

Γ_{212} = Γ_{211} = Γ_{313} = Γ_{331},

(24)

Γ_{213} = Γ_{231} = - Γ_{312} = - Γ_{321},

(25)

and

Γ_{222} = - Γ_{233} = - Γ_{323} = - Γ_{332},

(26)

Γ_{333} = - Γ_{322} = - Γ_{223} = - Γ_{223},

(27)

with further constraints on Γ₂₂₂ and Γ₃₃₃ such that

Γ_{222} \sin \frac{Θ}{2} (4 \cos^{2} \frac{Θ}{2} - 1) = Γ_{333} \sin \frac{Θ}{2} (4 \cos^{2} \frac{Θ}{2} - 1) = 0 .

(28)

When Θ ≠ 2π/3 or, equivalently, n ≥ 4, the last constraint [Eq. (28)] gives Γ₂₂₂ = Γ₃₃₃ = 0, and thus, the components in Eqs. (26) and (27) all vanish. When n = 3, however, these components generally remain unless the object possesses further symmetries. The components Γ₁₁₁ and Γ₁₂₂ = Γ₁₃₃ are generally left unrestricted from symmetry arguments. The shear-related tensors affect the motion in the form of $Γ$ : E^∞, but the incompressibility condition I : E^∞ = 0 enables us to eliminate one of the components between Γ_j11, Γ_j22, and Γ_j33 for each j = 1, 2, 3, if none of three components are left nonzero after the symmetry arguments. These additional degrees of freedom, referred to as the principle of indeterminacy, were discussed in detail by Brenner.⁶ We then choose Γ₁₂₂ = Γ₁₃₃ = 0 and arrive at the expression for the shear-force tensor $Γ$ ⁠, and the same arguments follow for the shear-torque tensor $Λ$ ⁠, yielding the following expressions for the shear-related tensors:

\begin{matrix} Γ & = & Γ_{1} {\hat{e}}_{1} {\hat{e}}_{1} {\hat{e}}_{1} + \frac{Γ_{2}}{2} [(I - {\hat{e}}_{1} {\hat{e}}_{1}) {\hat{e}}_{1} + {\{(I - {\hat{e}}_{1} {\hat{e}}_{1}) {\hat{e}}_{1}\}}^{T}] \\ + \frac{Γ_{3}}{2} [({\hat{e}}_{2} {\hat{e}}_{3} - {\hat{e}}_{3} {\hat{e}}_{2}) {\hat{e}}_{1} + {\{({\hat{e}}_{2} {\hat{e}}_{3} - {\hat{e}}_{3} {\hat{e}}_{2}) {\hat{e}}_{1}\}}^{T}] \\ + Γ_{4} [{\hat{e}}_{2} {\hat{e}}_{2} {\hat{e}}_{2} - {\hat{e}}_{2} {\hat{e}}_{3} {\hat{e}}_{3} - {\hat{e}}_{3} {\hat{e}}_{2} {\hat{e}}_{3} - {\hat{e}}_{3} {\hat{e}}_{3} {\hat{e}}_{2}] \\ + Γ_{5} [{\hat{e}}_{3} {\hat{e}}_{3} {\hat{e}}_{3} - {\hat{e}}_{3} {\hat{e}}_{2} {\hat{e}}_{2} - {\hat{e}}_{2} {\hat{e}}_{3} {\hat{e}}_{2} - {\hat{e}}_{2} {\hat{e}}_{2} {\hat{e}}_{3}] \end{matrix}

(29)

and

\begin{matrix} Λ & = & Λ_{1} {\hat{e}}_{1} {\hat{e}}_{1} {\hat{e}}_{1} + \frac{Λ_{2}}{2} [(I - {\hat{e}}_{1} {\hat{e}}_{1}) {\hat{e}}_{1} + {\{(I - {\hat{e}}_{1} {\hat{e}}_{1}) {\hat{e}}_{1}\}}^{T}] \\ + \frac{Λ_{3}}{2} [({\hat{e}}_{2} {\hat{e}}_{3} - {\hat{e}}_{3} {\hat{e}}_{2}) {\hat{e}}_{1} + {\{({\hat{e}}_{2} {\hat{e}}_{3} - {\hat{e}}_{3} {\hat{e}}_{2}) {\hat{e}}_{1}\}}^{T}] \\ + Λ_{4} [{\hat{e}}_{2} {\hat{e}}_{2} {\hat{e}}_{2} - {\hat{e}}_{2} {\hat{e}}_{3} {\hat{e}}_{3} - {\hat{e}}_{3} {\hat{e}}_{2} {\hat{e}}_{3} - {\hat{e}}_{3} {\hat{e}}_{3} {\hat{e}}_{2}] \\ + Λ_{5} [{\hat{e}}_{3} {\hat{e}}_{3} {\hat{e}}_{3} - {\hat{e}}_{3} {\hat{e}}_{2} {\hat{e}}_{2} - {\hat{e}}_{2} {\hat{e}}_{3} {\hat{e}}_{2} - {\hat{e}}_{2} {\hat{e}}_{2} {\hat{e}}_{3}] . \end{matrix}

(30)

Note that the 1/2 prefactors for the terms involving Γ₂, Γ₃, Λ₂, and Λ₃ are inherited from the helicoidal object.⁸ With regard to the new terms involving Γ₄, Γ₅, Λ₄, and Λ₅, which did not appear in the case of a helicoidal object, it is important to note that these constants all become zero when n ≥ 4, and we recover the same expressions as for a helicoidal object when n ≥ 4. Thus, the dynamics of a C_n-object with n ≥ 4 coincide with those of a helicoidal object derived in the previous study,⁸ which emphasizes the surprisingly wide applicability of the Jeffery equations derived for a helicoidal object, i.e., a C₄-object.

For simplicity in the following discussions, we assume that the propulsions F_prop and M_prop also satisfy the n-fold rotational symmetry, considering the propulsion to be generated by the surface deformation of the object under the same rotational symmetry. This assumption is valid for a self-deforming swimmer in a moderately strong background shear, if one approximates the swimming by its time average over the periodic deformation.⁸ Thus, these vectors should be aligned with the rotation axis ${\hat{e}}_{1}$ ⁠, i.e., $F_{p r o p} = F_{0} {\hat{e}}_{1}$ and $M_{p r o p} = M_{0} {\hat{e}}_{1}$ ⁠. From the expressions in Eqs. (29) and (30), this condition is satisfied if the surface velocity possesses the same n-fold rotation axis as for a squirmer swimmer with swirling,^62–64 while the shape of the swimmer is not necessarily a body of revolution but is a C_n-object.

III. GENERALIZED JEFFERY EQUATIONS

A. Dynamics of a C_n-object

We proceed to derive the expressions for the translational and angular velocities of the object, generalizing the Jeffery equations to a C_n-body with n ≥ 3, noting that the dynamics for the n ≥ 4 case have already been presented as the helicoidal Jeffery equations (1).

We start by decomposing the shear-induced force and torque in the form

Γ : E^{\infty} = Φ_{1} \cdot E^{\infty} \cdot {\hat{e}}_{1} + Φ_{2} \cdot E^{\infty} \cdot {\hat{e}}_{2} + Φ_{3} \cdot E^{\infty} \cdot {\hat{e}}_{3},

(31)

Λ : E^{\infty} = Ψ_{1} \cdot E^{\infty} \cdot {\hat{e}}_{1} + Ψ_{2} \cdot E^{\infty} \cdot {\hat{e}}_{2} + Ψ_{3} \cdot E^{\infty} \cdot {\hat{e}}_{3},

(32)

where Φ_i (i = 1, 2, 3) are given by

Φ_{1} = Γ_{1} {\hat{e}}_{1} {\hat{e}}_{1} + Γ_{2} (I - {\hat{e}}_{1} {\hat{e}}_{1}) + Γ_{3} ({\hat{e}}_{2} {\hat{e}}_{3} - {\hat{e}}_{3} {\hat{e}}_{2}),

(33)

Φ_{2} = Γ_{4} ({\hat{e}}_{2} {\hat{e}}_{2} - {\hat{e}}_{3} {\hat{e}}_{3}) - Γ_{5} ({\hat{e}}_{2} {\hat{e}}_{3} + {\hat{e}}_{3} {\hat{e}}_{2}),

(34)

Φ_{3} = - Γ_{4} ({\hat{e}}_{2} {\hat{e}}_{3} + {\hat{e}}_{3} {\hat{e}}_{2}) - Γ_{5} ({\hat{e}}_{2} {\hat{e}}_{2} - {\hat{e}}_{3} {\hat{e}}_{3}),

(35)

and $Ψ$ _i (i = 1, 2, 3) have similar forms with Γ₁, …, Γ₅ replaced by Λ₁, …, Λ₅.

The force and torque on the object [Eqs. (5) and (6)] must be balanced owing to the negligible inertia, since no other external force and torque are applied. We introduce the 6 × 6 grand resistance tensor

K = (\begin{matrix} K & C^{T} \\ C & Q \end{matrix}),

(36)

and the balance equations F = M = 0 can then be solved as

(\begin{matrix} U - U^{\infty} \\ Ω - Ω^{\infty} \end{matrix}) = - K^{- 1} \cdot (\begin{matrix} Φ_{i} \cdot E^{\infty} \cdot {\hat{e}}_{i} + F_{p r o p} \\ Ψ_{i} \cdot E^{\infty} \cdot {\hat{e}}_{i} + M_{p r o p} \end{matrix}),

(37)

where it should be noted that the summations must be performed over the repeated label i (i = 1, 2, 3). The inverse of $K$ can be calculated directly, and some straightforward but rather lengthy calculations lead to an explicit form for the velocities,

U = U^{\infty} + α_{μ} d_{μ},

(38)

Ω = Ω^{\infty} + β_{μ} d_{μ},

(39)

where summations are performed over the repeated greek indices μ = 1, 2, …, 5. The 27 degrees of freedom in the resistance and shear-related tensors are now reduced to the 10 constants α_μ and β_μ, which contain the object’s shape factors Γ_μ and Λ_μ, and the constants in the resistance tensors [Eqs. (19)–(21)], together with the propulsions F₀ and M₀, as shown explicitly below.

The vectors d_μ are functions of the background rate-of-strain tensor and the orientation of the object, given by

d_{1} = {\hat{e}}_{1},

(40)

d_{2} = ({\hat{e}}_{2} {\hat{e}}_{3} - {\hat{e}}_{3} {\hat{e}}_{2}) \cdot E^{\infty} \cdot {\hat{e}}_{1},

(41)

d_{3} = (I - {\hat{e}}_{1} {\hat{e}}_{1}) \cdot E^{\infty} \cdot {\hat{e}}_{1},

(42)

d_{4} = ({\hat{e}}_{2} {\hat{e}}_{2} - {\hat{e}}_{3} {\hat{e}}_{3}) \cdot E^{\infty} \cdot {\hat{e}}_{2} - ({\hat{e}}_{2} {\hat{e}}_{3} + {\hat{e}}_{3} {\hat{e}}_{2}) \cdot E^{\infty} \cdot {\hat{e}}_{3},

(43)

d_{5} = ({\hat{e}}_{2} {\hat{e}}_{3} + {\hat{e}}_{3} {\hat{e}}_{2}) \cdot E^{\infty} \cdot {\hat{e}}_{2} + ({\hat{e}}_{2} {\hat{e}}_{2} - {\hat{e}}_{3} {\hat{e}}_{3}) \cdot E^{\infty} \cdot {\hat{e}}_{3} .

(44)

From Eqs. (40)–(44), the vectors d₂, …, d₅ are all perpendicular to d₁, and it is found that the vectors {d₁, d₂, d₃} and {d₁, d₄, d₅}, respectively, form right-handed orthogonal bases, as

d_{2} = d_{3} \times d_{1},

(45)

d_{4} = d_{5} \times d_{1},

(46)

where it should be noted that these vectors are not orthonormal, in general.

With the determinants of the block matrix in the grand resistance matrix $K$ ⁠,

Δ_{1} = K_{1} Q_{1} - C_{1}^{2},

(47)

Δ_{2} = K_{2} Q_{2} - C_{2}^{2} - C_{23}^{2},

(48)

both of which are positive owing to the negative-definiteness of the tensor, the constants α_μ and β_μ may be written as follows:

α_{1} = Δ_{1}^{- 1} [- Q_{1} (Γ_{1} + F_{0}) + C_{1} (Λ_{1} + M_{0})],

(49)

α_{2} = Δ_{2}^{- 1} [- Q_{2} Γ_{3} + C_{2} Λ_{3} - C_{23} Λ_{2}],

(50)

α_{3} = Δ_{2}^{- 1} [- Q_{2} Γ_{2} + C_{2} Λ_{2} + C_{23} Λ_{3}],

(51)

α_{4} = Δ_{2}^{- 1} [- Q_{2} Γ_{4} + C_{2} Λ_{4} + C_{23} Λ_{5}],

(52)

α_{5} = Δ_{2}^{- 1} [+ Q_{2} Γ_{5} - C_{2} Λ_{5} + C_{23} Λ_{4}]

(53)

and

β_{1} = Δ_{1}^{- 1} [C_{1} (Γ_{1} + F_{0}) - K_{1} (Λ_{1} + M_{0})],

(54)

β_{2} = Δ_{2}^{- 1} [C_{2} Γ_{3} + C_{23} Γ_{2} - K_{2} Λ_{3}],

(55)

β_{3} = Δ_{2}^{- 1} [C_{2} Γ_{2} - C_{23} Γ_{3} - K_{2} Λ_{2}],

(56)

β_{4} = Δ_{2}^{- 1} [C_{2} Γ_{4} - C_{23} Γ_{5} - K_{2} Λ_{4}],

(57)

β_{5} = Δ_{2}^{- 1} [- C_{2} Γ_{5} - C_{23} Γ_{4} + K_{2} Λ_{5}] .

(58)

The director vector for the axis of the rotational symmetry, d₁, thus follows the equation for its time evolution, ${\dot{d}}_{1} = Ω \times d_{1}$ ⁠, given by

{\dot{d}}_{1} = Ω^{\infty} \times d_{1} + β_{3} d_{2} - β_{2} d_{3} + β_{5} d_{4} - β_{4} d_{5},

(59)

which generalizes the Jeffery equations for a C_n-object. When n ≥ 4, we readily recover the helicoidal Jeffery equations (1) from β₄ = β₅ = 0, with the Bretherton constant β = −β₂ and the chirality constant α = β₃. However, when n = 3, the angular dynamics of the object can be more complicated owing to the vectors d₄ and d₅, which depend on the instantaneous orientations ${\hat{e}}_{2}$ and ${\hat{e}}_{3}$ ⁠. The new vectors d₄ and d₅ are different from those found in the angular dynamics of an object with triaxial reflection symmetry, such as an ellipsoid.^54,65,66

Note that the vector d₅ is obtained by π/2-rotation in the ${\hat{e}}_{2} - {\hat{e}}_{3}$ plane, and the entire dynamics should be identical on rotation of the orientation of the ${\hat{e}}_{2}$ axis. There is only a single parameter that determines the dynamics, which we refer to as the “triangularity strength,” defined as $γ = \sqrt{β_{4}^{2} + β_{5}^{2}}$ ⁠.

When the object is a deformable self-propelled swimmer with discrete symmetry, the shape-dependent coefficients α_μ and β_μ are, in general, functions of time. One may, however, approximate them by their time-averaged values, if the timescale of the swimmer deformation is sufficiently faster than the shear-induced motions, as has been shown to be the case for bacterial swimmers^15–17 and other flagellate swimmers¹⁹ in a moderate background flow.

The detailed motions of a general C_n-object in the presence of a simple linear shear will be discussed in Sec. IV. Before proceeding to that discussion, however, we shall present comprehensive arguments for C_n-objects with further symmetry.

B. Dynamics of an object with further symmetry

In this subsection, we provide comprehensive reductions of the dynamics from the generalized Jeffery equations (38) and (39) for objects with further symmetry using the Schoenflies notation. The results are summarized in Tables I and II, and more detailed discussions are provided below. It should be noted that the dynamics in a simple shear flow will be discussed in detail in Sec. IV.

TABLE I.

Summary of the zero and nonzero parameters of a C_n-object with n ≥ 3 and classification in terms of the dynamics in a flow when n ≥ 4. An asterisk * indicates that the value is nonzero in general. The parameters α₁ and β₁ are shown for a passive particle, whereas these are, in general, nonzero owing to propulsion for an active particle. Note that the Bretherton constant and the chirality parameter are given by β = −β₂ and α = β₃, respectively.

	α₁	β₁	α₂	α₃	β₂	β₃	Dynamical classification
C_n	*	*	*	*	*	*	Helicoidal object
C_nv	*	0	0	*	*	0	Body of revolution
C_nh	0	*	0	0	*	*	Heterochiral object
D_n	0	0	*	0	*	0	Homochiral object
D_nh	0	0	0	0	*	0	Body of revolution
D_nd	0	0	0	0	*	0	Body of revolution
S_2n	0	*	0	0	*	*	Heterochiral object

	α₁	β₁	α₂	α₃	β₂	β₃	Dynamical classification
C_n	*	*	*	*	*	*	Helicoidal object
C_nv	*	0	0	*	*	0	Body of revolution
C_nh	0	*	0	0	*	*	Heterochiral object
D_n	0	0	*	0	*	0	Homochiral object
D_nh	0	0	0	0	*	0	Body of revolution
D_nd	0	0	0	0	*	0	Body of revolution
S_2n	0	*	0	0	*	*	Heterochiral object

TABLE II.

Summary of zero and nonzero parameters and the classifications of the angular dynamics in a simple shear when n = 3. An asterisk * indicates a nonzero entry, in general. The angular dynamics are discussed in detail in Sec. IV.

	α₄	α₅	β₄	β₅	Angular dynamics
C₃	*	*	*	*	Can be chaotic
C_3v	*	0	0	*	Can be chaotic
C_3h	*	*	0	0	Helicoidal Jeffery orbits
D₃	*	0	*	0	Can be chaotic
D_3h	*	0	0	0	Jeffery orbits
D_3d	*	0	0	0	Jeffery orbits
S₆	0	0	*	*	Can be chaotic

	α₄	α₅	β₄	β₅	Angular dynamics
C₃	*	*	*	*	Can be chaotic
C_3v	*	0	0	*	Can be chaotic
C_3h	*	*	0	0	Helicoidal Jeffery orbits
D₃	*	0	*	0	Can be chaotic
D_3h	*	0	0	0	Jeffery orbits
D_3d	*	0	0	0	Jeffery orbits
S₆	0	0	*	*	Can be chaotic

1. C_nv-object

We start with the dynamics of an object with symmetry C_nv, which possesses n reflection planes containing the axis of n-fold rotation. Examples include a regular pyramid with an n-sided base. We are able to reduce the number of parameters through symmetry arguments on such an object. Setting ${\hat{e}}_{2}$ to coincide with one of the n reflection planes, we have an additional reflection symmetry in the ${\hat{e}}_{1}$ – ${\hat{e}}_{2}$ plane, and we obtain C₁ = C₂ = 0 and Γ₃ = Γ₅ = 0, Λ₁ = Λ₂ = Λ₄ = 0. We, therefore, have α₂ = 0, α₃ ≠ 0, β₂ ≠ 0, and β₃ = α = 0. For a passive particle, we find α₁ ≠ 0 and β₁ = 0. We, thus, recover the angular dynamics, which are the same as those of a body of revolution with zero drift velocity when n ≥ 4. From Eq. (38), the terms α₂, …, α₅ induce a drift velocity, which is perpendicular to the director vector d₁. For a C_nv-object, which can be either a passive particle or an active swimmer, however, the drift velocity from the α₃ term will vanish over the periodic motion described by the Jeffery orbit, as obtained for a C_4v-object.⁸ When n = 3, however, we have nonzero contributions from β₅ ≠ 0 and α₄ ≠ 0, while the α₅ and β₄ terms vanish as α₅ = β₄ = 0.

2. C_nh-object

This class of objects corresponds to the chiral crystals that possess reflection symmetry in the plane perpendicular to the axis of n-fold rotation, as studied by Fries et al.⁵² We consider the additional reflection symmetry in the ${\hat{e}}_{2}$ – ${\hat{e}}_{3}$ plane, and when n ≥ 4, we have C₁ = C₂ = C₂₃ = 0 and Γ₁ = Γ₂ = Γ₃ = 0, but the two parameters in the shear-torque tensor remain nonzero, i.e., Λ₂ ≠ 0 and Λ₃ ≠ 0, yielding the relations α₂ = α₃ = 0, β₂ ≠ 0, and β₃ = α ≠ 0. These indicate that no drift velocity is generated, but the chiral parameter α does not vanish. We refer to the class of objects that exhibit such behavior in a simple shear flow as “heterochiral objects,” following the terminology used for molecules in chemistry,⁶⁷ since this case is a generalization of the dynamics for an object with two helices with opposite chirality connected at one end.^8,20 This result again recovers the dynamics obtained for a C_4h-object.⁸ For a passive particle, α₁ = 0 and β₁ ≠ 0 are obtained. When n = 3, with this symmetry, we obtain Γ₄ ≠ 0 and Γ₅ ≠ 0, while Λ₄ = Λ₅ = 0 holds, and we find that α₄ and α₅ do not vanish, but there is no additional effect on the angular dynamics,⁵² since β₄ = β₅ = 0.

3. D_n-object

The symmetry denoted by D_n is called “dihedral” in which the object possesses n additional twofold axes perpendicular to the n-fold rotation axis. We set ${\hat{e}}_{2}$ as one of the n twofold axes and consider the twofold rotational axis around it, which enables us to derive the relations C₂₃ = 0 and Γ₁ = Γ₂ = Γ₅ = Λ₁ = Λ₂ = Λ₅ = 0. We, thus, obtain α₃ = β₃ = 0, and the relation α₁ = β₁ = 0 also holds for a passive particle. When n ≥ 4, the nonzero constants are α₂ and β₂, which generate a nonzero drift velocity, but the angular dynamics follow the Jeffery orbits. We define the class of objects with this dynamics as “homochiral objects,” following again the terminology from chemistry, since such dynamics can be obtained for a uniform helical object.^14,35 When n = 3, however, we need additional nonzero contributions from α₄ and β₄.

4. D_nh-object

The class of D_nh-objects possesses a further reflection symmetry in the ${\hat{e}}_{2}$ – ${\hat{e}}_{3}$ plane, which is the plane perpendicular to the n-fold rotational axis, in addition to the dihedral symmetry discussed above. Examples of such objects include an n-gonal prism and a regular n-bipyramid. The additional reflection symmetry yields further relations, Γ₃ = Λ₄ = 0, enabling us to set α₂ = β₄ = 0, whereas these remain nonzero for a D_n-object. Therefore, we have angular dynamics that obey the Jeffery orbits for n ≥ 3, and no drift velocities are generated when n ≥ 4. However, when n = 3, a new drift velocity term remains owing to the nonzero α₄ term.

5. D_nd-object

An object denoted by D_nd has further n reflectional planes that pass between the twofold axes, in addition to the symmetry of D_n. An n-gonal antiprism belongs to this symmetry class. As in the discussions for a D_n-object, we set ${\hat{e}}_{2}$ as one of the n twofold axes.

When n is an odd number, n = 3, 5, 7, …, the object possesses reflection symmetry in the ${\hat{e}}_{1}$ – ${\hat{e}}_{2}$ plane, and we have an additional constraint on the resistance tensors, C₁ = C₂ = 0, and on the shear-related tensors, Λ₄ = 0 as for the C_nv-object. We still have nonzero values of Γ₄ and Λ₃.

When n is even but not a multiple of 4, i.e., n = 6, 10, 14, …, the additional n reflection plane for a C_nd-object passes through the ${\hat{e}}_{3}$ axis. We, therefore, have further constraints from the reflection symmetry in the ${\hat{e}}_{1}$ – ${\hat{e}}_{3}$ plane, which results in the same relations as in the odd-n cases, C₁ = C₂ = Λ₄ = 0, and Γ₄ = 0 since n ≠ 3. Only the constant Λ₃ is left nonzero. Similar arguments hold when n is a multiple of 4, since the object then possesses reflection symmetry in both the ${\hat{e}}_{1}$ – ${\hat{e}}_{2}$ and ${\hat{e}}_{1}$ – ${\hat{e}}_{3}$ planes. The constraint relations are found to be the same as in the cases above.

From these symmetry arguments, we finally obtain the same constraints on the coefficients in the tensors as in the D_nh case, yielding angular dynamics coinciding with those for a body of revolution, with an additional drift velocity from the α₄ term that appears only when n = 3.

6. S_2n-object

The symmetry denoted by S_2n involves only a 2n-fold rotation–reflection axis and is not a subgroup of C_2n but of C_n, where it should be noted that S_n = C_nh for odd n. The rotation–reflection symmetry provides further constraints such that C₁ = C₂ = C₂₃ = Γ₁ = Γ₂ = Γ₃ = Γ₄ = Γ₅ = 0 (the detailed calculations are provided in the Appendix). We, therefore, have α_μ = 0 for μ = 1, …, 5, whereas β_μ remain nonzero, in general, and when n ≥ 4, the dynamics are thus identical to those of a C_nv-object from β₄ = β₅ = 0, and the object is therefore classified as a heterochiral object. When n = 3, the additional angular velocity needs to be considered owing to the β₄ and β₅ terms.

IV. GENERALIZED JEFFERY ORBITS: MOTIONS IN A SIMPLE SHEAR

As shown in Fig. 2(a), we consider the background simple shear given by u^∞ = Gye₃, where G > 0 is the shear strength, although we henceforth set G = 1 since this factor may be absorbed into the timescale. The background angular velocity vector is then directed toward the e₁ axis, Ω^∞ = (1/2)e₁, and the rate-of-strain tensor E^∞ is given by E^∞ = (e₂e₃ + e₃e₂)/2 We now consider more detailed dynamics, focusing on motions in a simple shear flow that generalize the Jeffery orbits of a body of revolution. If n ≥ 4, as the angular dynamics are described by the director vector d₁ = e₁, which satisfies the helicoidal Jeffery equations,⁸ we focus on the dynamics for n = 3 in this section. The dynamics of the director then need three angles to describe its orientation for which we introduce the Euler angles (ϕ, θ, ψ) via the zxz convention, where ϕ is the azimuthal angle, θ is the polar angle, and ψ is the roll angle [Fig. 2(b)]. The time evolutions are given by⁶⁸

\frac{d ϕ}{d t} = \frac{1}{\sin θ} ({\hat{Ω}}_{2} \sin ψ + {\hat{Ω}}_{3} \cos ψ),

(60)

\frac{d θ}{d t} = {\hat{Ω}}_{2} \cos ψ - {\hat{Ω}}_{3} \sin ψ,

(61)

\frac{d ψ}{d t} = {\hat{Ω}}_{1} - \frac{\cos θ}{\sin θ} ({\hat{Ω}}_{2} \sin ψ + {\hat{Ω}}_{3} \cos ψ),

(62)

using the values of the angular velocity Ω in the body-fixed coordinates, $Ω = {\hat{Ω}}_{1} {\hat{e}}_{1} + {\hat{Ω}}_{2} {\hat{e}}_{2} + {\hat{Ω}}_{3} {\hat{e}}_{3}$ ⁠. Substituting the representations of Ω^∞ and E^∞ in the body-fixed frame into the expressions for the vectors d_μ [Eqs. (40)–(44)], we can derive a closed form of Eqs. (60)–(62) in terms of the Euler angles by using Eq. (39). After some straightforward calculations, we can write down the equations as

\frac{d ϕ}{d t} = \frac{1}{2} + h_{v} \csc θ,

(63)

\frac{d θ}{d t} = h_{p},

(64)

\frac{d ψ}{d t} = β_{1} - h_{v} \cot θ .

(65)

Here, we have introduced h_p and h_v as functions of the Euler angles and the shape-dependent parameters β₂, …, β₅, given by

h_{p} = β_{2} q_{2} + β_{3} q_{3} + β_{4} q_{4} + β_{5} q_{5},

(66)

h_{v} = β_{3} q_{2} - β_{2} q_{3} + β_{5} q_{4} - β_{4} q_{5},

(67)

where

q_{2} = \frac{1}{4} \sin 2 ϕ \sin 2 θ,

(68)

q_{3} = - \frac{1}{2} \cos 2 ϕ \sin θ,

(69)

q_{4} = \cos 2 ϕ \cos θ \sin 3 ψ + \frac{1}{2} \sin 2 ϕ (1 + \cos^{2} θ) \cos 3 ψ,

(70)

q_{5} = \cos 2 ϕ \cos θ \cos 3 ψ - \frac{1}{2} \sin 2 ϕ (1 + \cos^{2} θ) \sin 3 ψ .

(71)

FIG. 2.

View large Download slide

(a) Schematic of a background simple shear flow u^∞ = ye₃. (b) The Euler angles (ϕ, θ, ψ) via the zxz convention, where ϕ is the azimuthal angle, θ is the polar angle, and ψ is the roll angle.

From the expressions above, we readily recover the helicoidal Jeffery equations, which are decoupled from the angle ψ when β₄ = β₅ = 0. When n = 3, however, these two terms are left nonzero, and the angular dynamics can be different from what has been known. We will discuss how each new term contributes to the angular dynamics in detail below, and the results will be summarized at the end of this section.

A. Motions of D_3h-, D_3d-, and C_3h-objects

We start with the motions of a D_3h-object, a D_3d-object, and a C_3h-object for which the angular dynamics are described by the Jeffery equations of a body of revolution, since β₃ = β₄ = β₅ = 0, although a new drift velocity term appears.

The new drift velocity term for D_3h- and D_3d-objects is given by U₄ = α₄d₄ from Eq. (38) and the results in Sec. III. To clarify the direction of this vector, we consider the dynamics with θ = 0 and π in which the object is parallel/antiparallel to the background vorticity vector, and the angular dynamics follow

\frac{d}{d t} (ϕ \pm ψ) = \frac{1}{2} \pm β_{1},

(72)

\frac{d θ}{d t} = 0,

(73)

where the upper and lower signs indicate the cases θ = 0 and θ = π, respectively, and the object rotates around the e₁ axis with a constant rotational velocity. The vector d₄ is then found to be

d_{4} = \sin (2 ϕ \pm 2 ψ) {\hat{e}}_{2} \mp \cos (2 ϕ \pm 2 ψ) {\hat{e}}_{3},

(74)

which can be rewritten in terms of the laboratory-fixed frame as

d_{4} = \sin (3 ϕ \pm 3 ψ) e_{2} + \cos (3 ϕ \pm 3 ψ) e_{3} .

(75)

From the constant rotation given by Eq. (72), the drift velocity U₄ induces a threefold horizontal velocity during one period of the rotation around e₁, reflecting the triangular geometric features of the object.

Since the angles ϕ and θ are periodic in time following the Jeffery orbit, ψ is also periodic, and the time average of the drift velocity ⟨U₄⟩ is found to vanish, yielding no net drift velocity.

We now proceed to the dynamics of a C_3h-object for which the α₄ and α₅ terms are nonzero, but the angular dynamics are described by the helicoidal Jeffery equation with nonzero chirality parameter α = β₃. The stable orientation of the object is then either parallel or antiparallel to the background vorticity vector, as follows from analysis of the helicoidal Jeffery equation.⁸ At the stable orientation, the drift velocity U₄, which is proportional to the vector d₄, is toward the orientation in Eq. (75). As the drift velocity U₅ = α₅d₅ is perpendicular to both e₁ and d₄, we have a similar threefold periodic drift motion from U₅, which again leads to zero average drift velocity over the period of rotation.

B. Motions of C_3v- and D₃-objects

We now consider C_3v- and D₃-objects for which we need an additional term β₄ or β₅ in the angular dynamics. Since the vector d₅ is obtained by π/2-rotation of the angle ψ, we first consider the dynamics with nonzero β₄ with β₅ being kept zero. From Eqs. (74) and (75), the orientations at θ = 0 and π are not expected to be equilibria owing to the angular velocity generated by the new terms. Substituting θ = π/2 into Eqs. (63)–(65) gives the relation

\frac{d ϕ}{d t} = \frac{1}{2} (1 + β_{2} \cos 2 ϕ + β_{4} \sin 2 ϕ \sin 3 ψ),

(76)

\frac{d θ}{d t} = \frac{β_{4}}{2} \sin 2 ϕ \cos 3 ψ,

(77)

with dψ/dt = 0. The angle θ = π/2 is thus found to be stationary when cos 3ψ = 0, and we then find that the dynamics of the angle ϕ are periodic in time. Nevertheless, the stability of this periodic orbit is nontrivial. We proceed with the numerical solution of Eqs. (63)–(71) for an object with only the parameter β₄ being nonzero.

For the numerical integration, we use different initial values of the angle θ but fix the initial values of the angles ϕ and ψ to be zero, and we perform the numerical integration via the MATLAB ode45 solver from time t = 0 to t = 10⁵, which will be fixed throughout the computations shown below.

We plot the sample dynamics of objects with β₄ = −0.5 [Figs. 3(a) and 3(b)] and β₄ = −0.9 [Figs. 3(c) and 3(d)], with the other parameters, including the Bretherton constant, being set to zero (β₁ = β₂ = β₃ = β₅ = 0). In the left panels of Fig. 3, the Poincaré map at $ϕ = 2 π Z$ ⁠, where $Z = {0, \pm 1, \pm 2, \dots}$ ⁠, is shown for the dynamics with 10 different initial values of the angle θ. The right panels show the corresponding time evolutions of θ for five different simulations, with the same colors as used in the left panels for the same simulation. The horizontal axes in Figs. 3(b) and 3(d) are normalized by the time period of the Jeffery orbit with the corresponding β₂, i.e., T_J = 2π(r + r⁻¹), with $r = \sqrt{(1 - β_{2}) / (1 + β_{2})}$ ⁠.

FIG. 3.

View large Download slide

[(a) and (c)] Poincaré maps at the sections of $ϕ = 2 π Z$ with zero Bretherton constant. [(b) and (d)] Time evolutions of θ normalized by the time period of the corresponding Jeffery orbits T_J. Sample dynamics with different initial angles θ are shown for β₄ = −0.5 in (a) and (b) and for β₄ = −0.9 in (c) and (d), with other parameters β₁ = β₂ = β₃ = β₅ set to zero. The initial angles are fixed as ϕ = ψ = 0, and different simulations are shown in different colors. As the magnitude of β₄ increases, chaotic angular dynamics can be realized.

The Jeffery orbit of a body of revolution with zero Bretherton constant β = −β₂ = 0 is simple rotation with constant θ. As shown in Figs. 3(a) and 3(b), when β₂ = 0 but β₄ ≠ 0, the angle θ can vary in time, although the angular dynamics are still quasiperiodic in time as long as the strength of the triangularity, |β₄|, is small enough. However, as this strength increases, the angular dynamics can become chaotic, in particular, around the angle θ ≈ π/2, as in Figs. 3(c) and 3(d), in contrast to the helicoidal object (i.e., a C_n-object with n ≥ 4), which always exhibits nonchaotic dynamics. It is therefore found that the symmetries indicated by the Schoenflies notations C_3v and D₃ are the minimal symmetries that can exhibit chaotic angular dynamics.

We now consider a nonzero value of β₂, which is the negative of the Bretherton constant β = −β₂ in the range −1 < β < 1 for a standard object,⁵ although ring-like particles can possess values outside this range.⁶⁹ When β₄ is nonzero, the object exhibits a periodic Jeffery orbit, but when |β₄| is small, an orbit starting around θ ≈ π/2 is captured by an attracting periodic orbit at θ = π/2, which is one of the periodic orbits of cos 3ψ = 0. The basin of the attractor becomes wider as the strength of the triangularity increases, accompanied by the localization of the regions of quasiperiodic motions.

The attraction of the dynamics toward the angle θ = π/2, which is perpendicular to the background vorticity vector, can be found in large parameter ranges of β ≠ 0, β₄, and β₅. Figure 4 illustrates Poincaré maps and angular evolutions with (β, β₄, β₅) = (0.1, −0.5, 0) in (a) and (b) and with (0.5, 0, 0.5) in (c) and (d). With a nearly spherical Bretherton constant, β = 0.1, the attraction is fairly weak, but the orbit starting near θ = π/2 is slowly attracted by the stable periodic orbit at θ = π/2, as indicated by the arrow in Fig. 4(a). Noting that the vector d₅ is obtained by π/2-rotation in the plane perpendicular to d₁, we can interpret the exchange between β₄ and β₅ as a π/2-shift of the angle ψ [Fig. 4(c)]. The attractive behavior of the orbits toward the angle θ = π/2 is found also for the parameters in Figs. 4(c) and 4(d), but more complex structures of the quasiperiodic orbits are observed.

FIG. 4.

View large Download slide

Poincaré map and time series of θ with nonzero Bretherton constants and triangularity strengths: [(a) and (b)] (β, β₄, β₅) = (0.1, −0.5, 0) and [(c) and (d)] (0.5, 0, 0.5). The plots are generated as in Fig. 3. The orbits starting near θ ≈ π/2 are captured by one of the periodic orbits at θ = π/2. These attractive dynamics are fairly weak for the nearly spherical case shown in (a) and (b), with the attracting periodic orbit being indicated by the arrow in (a).

C. Motions of S₆- and C₃-objects

We now proceed to more general situations for less-symmetric objects such as S₆- and C₃-objects for which β₁ and β₃ can take nonzero values.

We first consider the effects of β₃, which is equivalent to the chirality parameter α = β₃. When β = 0, there is still a nonlinear periodic orbit for a helicoidal object,⁸ despite the nonzero chirality term α ≠ 0, and strong attractive behavior around θ = π is not found even when both the chirality and triangularity terms are considered. Figure 5 shows the Poincaré maps for the angular dynamics with parameters (β, β₄, β₅) = (0, −0.5, 0) as in Figs. 3(a) and 3(b), but with nonzero-α effect. It can be seen from the figure that as the chirality strength increases, the quasiperiodic orbits collapse into chaotic dynamics.

FIG. 5.

View large Download slide

Poincaré maps for the angular dynamics of an object with zero Bretherton constant β = 0, but with chirality contribution (a) α = 0.4 or (b) α = 0.7 and with the triangularity strength fixed at (β₄, β₅) = (−0.5, 0) as in Fig. 3(a). The quasiperiodic orbits collapse into chaotic motions owing to the chirality effect.

When the Bretherton constant is nonzero, β ≠ 0, however, depending on the sign of the parameter α, the axis of rotational symmetry, ${\hat{e}}_{1}$ ⁠, is attracted to align parallel or antiparallel to the background vorticity vector, if the triangularity is not considered, as for a helicoidal object. In Figs. 6(a)–6(c), the angular dynamics are shown, with the Poincaré maps and time evolutions of the angle θ for the same parameter set (β, β₄, β₅) = (0.5, 0, 0.5) as in Figs. 4(c) and 4(d), but with the nonzero chirality term α = 0.1. The time evolution shows two different attractive periodic orbits [Figs. 6(b) and 6(c)]: one is the stable periodic orbit at θ = π/2 generated by the triangularity term, and the other is located near θ ≈ 0, where the stable orientation is created by the chirality term with α > 0. From the three spiral structures in the Poincaré map in Fig. 6(a), the stable periodic dynamics near θ = 0 are found to have two periods with threefold difference, which reflects the triangular geometric features of the object. As the chirality term is strengthened to α = 0.3, the periodic orbit at θ = π/2 becomes less attractive and the dynamics falls into a doubly periodic stable orbit [Fig. 6(d)].

FIG. 6.

View large Download slide

(a) Poincaré map and [(b)–(d)] angular time evolutions for the angular dynamics of an object with nonzero Bretherton constant and chirality contribution. The dynamics are shown for the parameter set (β, β₄, β₅) = (0.5, 0, 0.5). The chirality term is fixed at α = 0.1 for the Poincaré map in (a) and the shorter and longer time evolutions in (b) and (c) and at α = 0.3 for the time evolutions in (d). Two attractive dynamics are found at θ = π/2 and θ ≈ 0, the latter of which corresponds to the dynamics with two periods with threefold difference and dominates as the chirality term becomes stronger.

We finally consider the effects of the β₁ term, which generates spinning motion around the axis of rotation and is decoupled from the angular dynamics in a helicoidal object. With triangularity, however, this term is coupled with the rotation dynamics. The parameter β₁ can also be generated by self-propulsion, and this effect needs to be considered for a general active particle with self-spinning motion.

When only the β₁ term and the triangularity terms are nonzero, it is found, for β₁ > 0, that the β₁ term can generate chaotic orbits in the region with θ > π/2, although the chaotic motions are suppressed as β₁ increases. These tendencies are also found in larger parameter regions.

When the Bretherton constant is nonzero, β ≠ 0, with the existence of the stable orbits at θ = π/2, it is less likely that chaotic dynamics will be realized. With the β₁ term, however, when the triangularity effect is comparable in size to the β₁ term, chaotic orbits are possible in either hemisphere θ > π/2 or θ > π/2, depending on the sign of β₁. Figure 7 shows the Poincaré maps and time evolutions of the angle θ for dynamics with fixed values of the parameters β = 0.9 and β₁ = 0.2 and different values of β₅: 0.1 in (a) and (b) and 0.3 in (c) and (d). With an intermediate size of the β₁ effect, chaotic behavior can be realized in the region with θ > π/2, as is found in the Poincaré map in Fig. 7(a). The chaotic angular dynamics, however, disappear as β₁ increases, with attraction by a periodic orbit occurring at a certain angle θ > π/2 [Figs. 7(c) and 7(d)].

FIG. 7.

View large Download slide

Poincaré maps and angular evolutions for dynamics with nonzero Bretherton constant β = 0.9, α = 0, and β₁ = 0.1. The triangularity strengths are (β₄, β₅) = (0, 0.1) in (a) and (b) and (β₄, β₅) = (0, 0.3) in (c) and (d).

D. Summary of results for the generalized Jeffery orbits

In this section, we have considered the generalized Jeffery orbits, which represent the angular dynamics in a simple shear, where a body of revolution follows nonlinear periodic orbits, parameterized by the Bretherton constant β. When β = 0, as is the case for a sphere, an object will rotate around a background vorticity vector with a constant rotational angular velocity. Even if a chirality constant α is added, there still exist nonlinear periodic orbits. However, if triangularity is considered for a C₃-object, the nonlinear orbits are found to collapse into chaotic angular dynamics.

Although the equations are not integrable, the chaotic behavior looks very similar to that in Hamiltonian chaos and can be compared to the angular dynamics of objects with triaxial reflection symmetry, such as ellipsoids and cuboids.⁵⁵ These objects are denoted in the Schoenflies notation by D_2h and are known to exhibit chaotic angular dynamics. Since no chaotic angular motions are possible for an object with n-fold rotational symmetry (n ≥ 4), n = 3 is the maximum number for the emergence of chaotic Jeffery orbits.

A Jeffery orbit with a nonzero Bretherton constant exhibits nonuniform rotation in a simple shear flow, accompanied by longer stays along the background velocity vector. The coupling between this nonuniform rotation and the triangularity of the object shape can allow a stable periodic orbit in which the axis vector of the threefold rotation periodically rotates in a plane perpendicular to the background vorticity vector without spinning motion in the plane. When a chirality constant α is added, the object orientation is attracted by a stable orbit that is perpendicular or parallel to the background vorticity vector, depending on the initial orientation and the strengths of the chirality and triangularity.

Finally, we have considered the effects of nonzero spinning rotation along the axis of n-fold rotation. This spinning term can also be generated by self-spinning propulsion, and its magnitude can be changed by controlling the propulsion for biological and artificial microswimmers. From numerical computations, the spinning term is found to yield chaotic angular dynamics when β ≠ 0 if the size of this term is comparable to that of the triangularity terms.

In the absence of the β₁ term, the angle ψ typically increases when θ < π/2 and decreases when θ > π/2. A positive β₁ term is related to the angle ψ according to Eq. (65), and so the angular motion is accelerated when θ < π/2 but decelerated by the β₁ term when θ > π/2. This asymmetric contribution could be key to the emergence of chaotic motion as a result of the β₁ term. Nonetheless, the mechanisms underlying the chaotic motion and its suppression require further theoretical analysis.

We do not discuss in detail other symmetric objects, such as those possessing tetrahedral and icosahedral symmetries: T, T_d, T_h, O, O_h, I, and I_h. These higher symmetry constraints may impose further relations on the Bretherton constant, as in the case of an isotropic helicoid,⁷⁰ which possesses a triaxial helicoidal symmetry.

V. EXAMPLE OF A TRIANGULAR OBJECT

In this section, we discuss an example of an object that possesses threefold rotational symmetry by considering a set of slender rods spatially arranged to form a regular triangular pyramid (Fig. 8). Such a tetrahedral object could be used as a simplified model of radiolarians with threefold rotational symmetry such as Thyrsocyrtis and Podcrytis.^46,47

FIG. 8.

View large Download slide

Schematic of an example of a triangular object consisting of four slender rods arranged to form a regular tetrahedron. Each rod has length L and is separated from the origin by a distance c. One of the rods is oriented toward the ${\hat{e}}_{1}$ axis.

One of the rods of the model object is directed toward the axis of threefold rotational symmetry ${\hat{e}}_{1}$ ⁠, and the remaining rods have a polar angle of θ = 2π/3 from the axis of rotation. We label these three rods by k = 1, 2, 3, with rod k being obtained by a 2πk/3-rotation around the ${\hat{e}}_{1}$ axis after a 2π/3-rotation around the ${\hat{e}}_{3}$ axis. The combined rotation matrix, which maps from the rod parallel to the ${\hat{e}}_{1}$ axis to one of the others, is then given by

R_{i j}^{(k)} = \frac{1}{2} (\begin{matrix} - 1 & - \sqrt{3} & 0 \\ \sqrt{3} c_{k} & - c_{k} & - 2 s_{k} \\ \sqrt{3} s_{k} & - s_{k} & 2 c_{k} \end{matrix}),

(78)

where we have introduced the notations s_k = sin(2πk/3) and c_k = cos(2πk/3). From the method of construction, we could straightforwardly generalize this example to an object consisting of N + 1 rods with C_N-symmetry.¹⁵ As a rod is a body of revolution, this example of an object is a C_3v-object. It is straightforward to extend the model to general C_n-objects by the chirality effect, replacing a simple rod by a chiral object such as a helix. Nonetheless, our current focus is on providing an example of an object with triangular symmetry and estimating the values of the triangularity strength, and therefore, we do not pursue any further generalization in this paper.

From symmetry, we can write the resistance tensors for the slender rod in the ${\hat{e}}_{1}$ axis, labeled by superscript (0), as $K_{11}^{(0)} = K_{1}^{(0)}$ and $K_{22}^{(0)} = K_{33}^{(0)} = K_{2}^{(0)}$ and similarly for the rotational resistance tensor Q and the coupling tensor $C_{23}^{(0)} = - C_{32}^{(0)}$ ⁠, represented in body-fixed coordinates. We can also write the nonzero components of the shear-force and shear-torque matrices as

Γ_{111}^{(0)} = Γ_{1}^{(0)}, Γ_{212}^{(0)} = Γ_{221}^{(0)} = Γ_{313}^{(0)} = Γ_{331}^{(0)} = \frac{Γ_{2}^{(0)}}{2}

(79)

and

Λ_{231}^{(0)} = Λ_{213}^{(0)} = - Λ_{321}^{(0)} = - Λ_{312}^{(0)} = \frac{Λ_{3}^{(0)}}{2} .

(80)

We approximate the resistance tensors of a rod complex by summing the contribution from each rod k, neglecting the hydrodynamic interactions between the rods. Following the transformation of the matrix representation in Eqs. (13) and (14), we obtain

K_{1} = K_{1}^{(0)} + \frac{1}{4} \sum_{k = 1}^{3} (K_{1}^{(0)} + 3 K_{2}^{(0)}),

(81)

K_{2} = K_{2}^{(0)} + \frac{1}{4} \sum_{k = 1}^{3} (3 c_{k}^{2} K_{1}^{(0)} + (c_{k}^{2} + 4 s_{k}^{2}) K_{2}^{(0)}) .

(82)

With the identity $\sum_{k = 1}^{N} \sin^{2} (2 π k / N) = \sum_{k = 1}^{N} \cos^{2} (2 π k / N) = N / 2$ for N ≥ 3, these are calculated as

K_{1} = \frac{1}{4} (7 K_{1}^{(0)} + 9 K_{2}^{(0)}),

(83)

K_{2} = \frac{1}{8} (9 K_{1}^{(0)} + 23 K_{2}^{(0)}),

(84)

and similar relations also follow for Q₁ and Q₂. The representation of the coupling tensor is obtained as

C_{23} = C_{23}^{(0)} - \frac{1}{2} \sum_{k = 1}^{3} C_{23}^{(0)} = - \frac{1}{2} C_{23}^{(0)} .

(85)

We now consider the coefficients in the shear-related tensor. A similar calculation leads to the following expressions for the shear-force tensor components:

Γ_{1} = Γ_{1}^{(0)} + \frac{1}{8} \sum_{k = 1}^{3} (- Γ_{1}^{(0)} - 3 Γ_{2}^{(0)}) = \frac{1}{8} (5 Γ_{1}^{(0)} - 9 Γ_{2}^{(0)})

(86)

and

Γ_{2} = Γ_{2}^{(0)} + \frac{1}{8} \sum_{k = 1}^{3} (6 c_{k} Γ_{1}^{(0)} - (2 c_{k}^{2} - 4 s_{k}^{2}) Γ_{2}^{(0)}) = \frac{5}{8} Γ_{2}^{(0)} .

(87)

The new term from the triangular geometry remains nonzero and is given by

Γ_{4} = \frac{1}{8} \sum_{k = 1}^{3} [- 3 \sqrt{3} c_{k}^{3} Γ_{1}^{(0)} + (\sqrt{3} c_{k}^{3} + 4 \sqrt{3} s_{k}^{2} c_{k}) Γ_{2}^{(0)}],

(88)

which we can further reduce to

Γ_{4} = - \frac{9 \sqrt{3}}{32} (Γ_{1}^{(0)} + Γ_{2}^{(0)}) .

(89)

We note here that the identities $\sum_{k = 1}^{N} \sin^{3} (2 π k / N) = \sum_{k = 1}^{N} \cos^{3} (2 π k / N) = 0$ hold only for N ≥ 4 and that for N = 3, the sums $\sum_{k = 1}^{3} c_{k}^{3} = 3 / 4$ and $\sum_{k = 1}^{3} s_{k}^{2} c_{k} = - 3 / 4$ do not vanish, yielding a particular contribution for a C₃-object while this term does not appear for a helicoidal object.

Similar calculations lead to the following expressions for the shear-torque tensor components for Λ₃:

Λ_{3} = Λ_{3}^{(0)} + \frac{1}{8} \sum_{k = 1}^{3} (c_{k}^{2} - 3 s_{k}^{2}) Λ_{3}^{(0)} = \frac{5}{8} Λ_{3}^{(0)} .

(90)

For Λ₅, however, we have

Λ_{5} = 0,

(91)

which should, in general, be nonzero from symmetry arguments but perhaps vanishes as a result of the neglect of the hydrodynamic interactions between the rods.

If we take L = 1 as the unit of length and C_T = −1 as the unit of force, then the parameters of the object under consideration are c/L and ξ = C_N/C_T, the latter of which typically ranges from 1.5 to 2 for cilia and flagella but can be less than unity for some flagellated protists with mastigonemes.⁷¹ The scalar components of the resistance tensors for a single rod are given^8,72 by $K_{1}^{(0)} = - 1$ ⁠, $K_{2}^{(0)} = - ξ$ ⁠, $Q_{1}^{(0)} = 0$ ⁠, $Q_{2}^{(0)} = - ξ (c^{2} + c + 1 / 3)$ ⁠, and $C_{23}^{(0)} = - ξ (c + 1 / 2)$ ⁠, and those for the shear-related tensors are similarly obtained⁸ as $Γ_{1}^{(0)} = - (c + 1 / 2)$ ⁠, $Γ_{2}^{(0)} = - ξ (c + 1 / 2)$ ⁠, and $Λ_{2}^{(0)} = - ξ (c^{2} + c + 1 / 3)$ ⁠.

As the object under consideration possesses C_3v-symmetry, the nonzero constants in the angular dynamics are the β₂ and β₅ terms. Using the expressions above and substituting these values into Eqs. (55) and (58), we can calculate the values of β₂ and β₅, depending on the geometrical parameters c/L and ξ. The results are shown in Fig. 9 using the Bretherton constant β (= −β₂). The tetrahedral rod complex has a positive Bretherton constant, which is around β ≈ 0.25, although it can be varied slightly by changing the morphology. The effective aspect ratio is then estimated as $\tilde{c} = \sqrt{(1 + β) / (1 - β)} \approx 1.3$ ⁠, which is the aspect ratio of a spheroid with the corresponding Bretherton constant. The value of the triangularity strength is obtained as β₅ ≈ − 0.04 for a typical slender rod with ξ = 1.5–2.

FIG. 9.

View large Download slide

Coefficients of the angular dynamics for the example of a tetrahedral rod complex with different c and ξ.

Taking these values as representing the typical dynamics of the example of a tetrahedral rod complex, we have calculated the angular dynamics as in Sec. IV, and the Poincaré map and time evolution of the angle θ thus obtained are shown in Fig. 10. Even though the triangularity strength is relatively small owing to the nonzero Bretherton constant, there are attracting stable periodic orbits at θ = π/2. For a periodic orbit, the angle ψ satisfies sin 3ψ = 0, and two attracting orbits with ψ = π/3 and 5π/3 can be found in Fig. 10(a). The stable periodic motion is then a rotation around the background vorticity vector, with one of the rods being perpendicular to it, while the projection onto the e₂–e₃ plane is a regular triangle with two of the rods facing away from the background vorticity vector. As the attraction is weak owing to the small value of |β₅|, the object exhibits quasiperiodic motions if the initial configuration is away from the stable point at θ = π/2.

FIG. 10.

View large Download slide

(a) Poincaré map and (b) time series of angular dynamics for the example of a tetrahedral rod complex with typical parameter values β = 0.25 and β₅ = −0.04.

VI. CONCLUDING REMARKS

We have theoretically investigated the dynamics of a microscopic object, which can be either a passive particle or an active swimmer, with a n-fold rotational symmetry (n ≥ 3), which we have referred to in this paper as a C_n-object, using the Schoenflies notation. When n ≥ 4, the angular motions of such an object in a background linear flow are found to be described by the helicoidal Jeffery orbit, which is determined by the two shape parameters: the Bretherton constant and a constant associated with the object’s chirality.

Further symmetries can simplify the helicoidal Jeffery equations, and we have found that C_nh- and S_2n-objects can be categorized into a class of heterochiral objects, while D_n-objects belong to the class of homochiral objects, each of which classes contains a subclass of bodies of revolution. C_nv-, D_nh-, and D_nd-objects follow the Jeffery orbits of a body of revolution. These results illustrate that a wide variety of shapes of microscopic objects can be classified into a small number of hydrodynamic shapes that share the same dynamics in a flow. Even a discrete symmetry can be identical to a continuous symmetry as far as the dynamics in a flow are considered.

The unified description of the dynamics of a C_n-object as those of a helicoidal is only possible for n ≥ 4, and it is found that the motions of a C₃-object require further parameters in the dynamics, which reflect the triangularity of the object’s shape. In turn, this new effect of a C₃-object leads to a mechanical coupling between the direction of the axis of rotation and the other two perpendicular axes, and thus, the orientation dynamics need to be specified by three angles.

Owing to triangularity, the angular dynamics of a C₃-object can be chaotic, whereas the orbits typically possess stable periodic orbits when the Bretherton constant is nonzero. The existence of a stable periodic orbit has been confirmed for the example of an object consisting of four slender rods in a tetrahedral configuration. The basin of attraction is fairly small, and only slow attractions are possible owing to the small values of the triangularity constant for this particular object. The triangularity also generates a mechanical coupling between the self-propelling spin and the entire rotation induced by background shear, and it is found that the spin of the swimmer can trigger chaotic angular dynamics, which emphasizes the quantitative difference emerging from the triangularity of self-propelling swimmers.

As discussed in Sec. I, many microorganisms, such as diatoms and radiolarians, have shape with discrete rotational symmetry. The dynamics in a flow can be simply described by helicoidal Jeffery equations if the object has an n-fold rotational axis with n ≥ 4. Triangularity, however, will affect the dynamics of cells with a threefold rotational axis but without extra mirror symmetry, such as Lithochytris pyramidalis, Thyrsocyrtis, and Podocyrtis,^46,47 for which the example of a tetrahedral rod complex presented here could be useful.

The inertia of particles could lead to chaotic motion for a spheroidal particle,^73,74 although motion is stabilized in the ellipsoidal case.⁷⁵ The notable difference between the angular dynamics of spheroids and ellipsoids suggests that small shape perturbations could dramatically affect the entire dynamics.^55,56 The integrable Jeffery orbits are modulated by such small effects of geometry, inertia, or non-Newtonian rheology^76,77 and could lead to chaotic motion. This is, however, in contrast to the case where the dynamical system possesses an orbit that is robustly stable under perturbations.^38,78 This suggests that the stable orbits induced by triangularity and chirality with a nonzero Bretherton constant would not be collapsed by geometrical perturbations, and the equations of motion derived in this study should be useful for a biological particle or swimmer, which are not often exactly symmetric and can change their shape in time, although further studies will be needed to confirm this.

The classification of objects in terms of symmetries of hydrodynamic resistance tensors has been examined in the presence of gravity,^79,80 which can be understood in terms of the hydrodynamic response of an object to a vectorial external force, while the classification of dynamics in shear is connected with the response to the external second-rank tensors. More recently, such symmetry-based arguments have been made in the context of a magnetic swimmer propelled by an external rotating magnetic field.^81,82 Since symmetry arguments can be systematically applied, it is expected that classifications of objects in terms of their hydrodynamic response to higher-rank tensorial actions can be developed.

In conclusion, we have provided a general Jeffery equation for an object with n-fold rotational symmetry (n ≥ 3) with a new triangularity parameter in addition to the Bretherton constant and chirality parameter, and we have demonstrated that microscopic objects can be categorized into a small number of classes based on their dynamics in a background flow. Triangularity can lead to chaotic Jeffery orbits in a simple shear flow, although the dynamics typically possess stable periodic orbits. These findings emphasize the significance of the symmetry classification of an object for fluid–structure interactions at a low Reynolds number.

DATA AVAILABILITY

Data sharing is not applicable to this article as no new data were created or analyzed in this study.

ACKNOWLEDGMENTS

K.I. acknowledges JSPS-KAKENHI for Young Researchers (Grant No. 18K13456) and JST, PRESTO, Japan (Grant No. JPMJPR1921). This work was partially supported by the Research Institute for Mathematical Sciences, an International Joint Usage/Research Center located at Kyoto University.

APPENDIX: SYMMETRY ARGUMENTS FOR AN S_2n-OBJECT

In this appendix, we provide a derivation of the symmetric reductions of the shape-related hydrodynamic tensors for an S_2n-object. As an S_2n-object contains the symmetry denoted by C_n, we can parameterize the tensors as in Eqs. (19)–(30) in the main text. Let the angle Θ′ = π/n represent the 2n-fold rotation–reflection symmetry. We proceed by equating the representation by a Θ′/2-rotation together with reflection in the ${\hat{e}}_{2}$ – ${\hat{e}}_{3}$ plane and the representation by a −Θ′/2-rotation without reflection. Noting that the coupling tensor C needs an additional minus sign when the reflection operation is applied as in Eq. (14), we obtain the following relations for the matrix representation:

C_{11} = 0,

(A1)

C_{22} + C_{33} = 0,

(A2)

C_{23} - C_{32} = 0

(A3)

from which we have C₁ = C₂ = C₂₃ = 0, whereas the rotation–reflection transformation does not affect the translational and rational resistance tensors K and Q.

We now proceed to the constraint imposed by the rotation–reflection symmetry on the shear-force tensor $Γ$ ⁠. Equating the two representations leads to the following relations:

Γ_{111} = Γ_{122} = Γ_{133} = 0,

(A4)

Γ_{212} = Γ_{211} = Γ_{313} = Γ_{331} = 0,

(A5)

Γ_{213} = Γ_{231} = Γ_{312} = Γ_{321} = 0,

(A6)

and

Γ_{222} \sin \frac{Θ^{'}}{2} (4 \cos^{2} \frac{Θ^{'}}{2} - 1) = Γ_{333} \sin \frac{Θ^{'}}{2} (4 \cos^{2} \frac{Θ^{'}}{2} - 1) = 0,

(A7)

as in Eq. (28). However, the factor 4 cos²(Θ′/2) − 1 does not vanish, since the angle Θ′/2 ≤ π/6 for n ≥ 3, and we obtain Γ₄ = Γ₅ = 0 and Γ₁ = Γ₂ = Γ₃ = 0 from Eqs. (A4)–(A6).

For the shear-torque tensor, on the other hand, noting the additional sign in Eq. (14) from the reflection transformation, we find that no further constraints are obtained from the comparison between the two representations, except that

Λ_{222} \cos \frac{Θ^{'}}{2} (4 \sin^{2} \frac{Θ^{'}}{2} - 1) = Λ_{333} \cos \frac{Θ^{'}}{2} (4 \sin^{2} \frac{Θ^{'}}{2} - 1) = 0 .

(A8)

Nonetheless, 4 sin²(Θ′/2) − 1 = 0 for n = 3, and thus, Λ₄ and Λ₅ are still nonzero for an S₆-object, while these vanish when n ≥ 4.

REFERENCES

1.

G. G.

Stokes

, “

On the effect of internal friction of fluids on the motion of pendulums

,”

Trans. Cambridge Philos. Soc.

9

,

8

–

106

(

1851

).

Google Scholar

2.

U.

Nakaya

,

Snow Crystals: Natural and Artificial

(

Harvard University Press

,

1954

).

Google Scholar

Crossref

3.

E.

Haeckel

,

Art Forms in Nature

(

Dover Publications

,

1974

).

Google Scholar

4.

G. B.

Jeffery

, “

The motion of ellipsoidal particle immersed in a viscous fluid

,”

Proc. R. Soc. London, Ser. A

102

,

161

–

179

(

1922

).

https://doi.org/10.1098/rspa.1922.0078

Google Scholar

Crossref

5.

F. B.

Bretherton

, “

The motion of rigid particles in a shear flow at low Reynolds number

,”

J. Fluid Mech.

14

,

284

–

304

(

1962

).

https://doi.org/10.1017/s002211206200124x

Google Scholar

Crossref

6.

H.

Brenner

, “

The Stokes resistance of an arbitrary particle—III. Shear fields

,”

Chem. Eng. Sci.

19

,

631

–

651

(

1964

).

https://doi.org/10.1016/0009-2509(64)85052-1

Google Scholar

Crossref

7.

S.

Kim

and

S. J.

Karrila

,

Microhydrodynamics: Principles and Selected Applications

(

Dover Publications

,

2005

).

Google Scholar

8.

K.

Ishimoto

, “

Helicoidal particles and swimmers in a flow at low Reynolds number

,”

J. Fluid Mech.

892

,

A11

(

2020

).

https://doi.org/10.1017/jfm.2020.142

Google Scholar

Crossref

9.

H.

Brenner

, “

The Stokes resistance of an arbitrary particle—II. An extension

,”

Chem. Eng. Sci.

19

,

599

–

629

(

1964

).

https://doi.org/10.1016/0009-2509(64)85051-x

Google Scholar

Crossref

10.

G.

Kirchhoff

, “

Uber die benegung eines rotationskörpers in earner flüssigkeit

,”

J. Reine Angew. Math.

71

,

237

–

262

(

1869

).

https://doi.org/10.1515/crll.1870.71.237

Google Scholar

Crossref

11.

L.

Kelvin

, “

Hydrokinetic solutions and observations

,”

Philos. Mag.

42

,

362

–

377

(

1871

).

Google Scholar

12.

H.

Lamb

,

Hydrodynamics

, 6th ed. (

Dover Publications

,

1932

).

Google Scholar

13.

Y. J.

Kim

and

W. J.

Rae

, “

Separation of screw-sensed particles in a homogeneous shear field

,”

Int. J. Multiphase Flow

17

,

717

–

744

(

1991

).

https://doi.org/10.1016/0301-9322(91)90053-6

Google Scholar

Crossref

14.

P.

Chen

and

Q.

Zhang

, “

Dynamical solutions for migration of chiral DNA-type objects in shear flows

,”

Phys. Rev. E

84

,

056309

(

2011

).

https://doi.org/10.1103/physreve.84.056309

Google Scholar

Crossref

15.

K.

Ishimoto

and

E.

Lauga

, “

The N-flagella problem: Elastohydrodynamic motility transition of multi-flagellated bacteria

,”

Proc. R. Soc. A

475

,

20180690

(

2019

).

https://doi.org/10.1098/rspa.2018.0690

Google Scholar

Crossref

16.

K.

Ishimoto

, “

Bacterial spinning top

,”

J. Fluid Mech.

880

,

620

–

652

(

2019

).

https://doi.org/10.1017/jfm.2019.714

Google Scholar

Crossref

17.

A. J. T. M.

Mathijssen

,

N.

Figueroa-Morales

,

G.

Junot

,

E.

Clément

,

A.

Lindner

, and

A.

Zöttl

, “

Oscillatory surface rheotaxis of swimming E. coli bacteria

,”

Nat. Commun.

10

,

3434

(

2019

).

https://doi.org/10.1038/s41467-019-11360-0

Google Scholar

Crossref

PubMed

18.

Marcos

,

H. C.

Fu

,

T. R.Powers

, and

R.

Stocker

, “

Bacterial rheotaxis

,”

Proc. Natl. Acad. Sci. U. S. A.

109

,

4780

–

4785

(

2012

).

https://doi.org/10.1073/pnas.1120955109

Google Scholar

Crossref

PubMed

19.

B. J.

Walker

,

K.

Ishimoto

,

R. J.

Wheeler

, and

E. A.

Gaffney

, “

Response of monoflagellate pullers to a shearing flow: A simulation study of microswimmer guidance

,”

Phys. Rev. E

98

,

063111

(

2018

).

https://doi.org/10.1103/physreve.98.063111

Google Scholar

Crossref

PubMed

20.

S.

Kramel

,

G. A.

Voth

,

S.

Tympel

, and

S.

Tympe

, “

Preferential rotation of chiral dipoles in isotropic turbulence

,”

Phys. Rev. Lett.

117

,

154501

(

2016

).

https://doi.org/10.1103/physrevlett.117.154501

Google Scholar

Crossref

PubMed

21.

G.

Bellani

,

M. L.

Byron

,

A. G.

Collignon

,

C. R.

Meyer

, and

E. A.

Variano

, “

Shape effects on turbulent modulation by large nearly neutrally buoyant particles

,”

J. Fluid Mech.

712

,

41

–

60

(

2012

).

https://doi.org/10.1017/jfm.2012.393

Google Scholar

Crossref

22.

M.

Byron

,

J.

Einarsson

,

K.

Gustavsson

,

G.

Voth

,

B.

Mehlig

, and

E.

Variano

, “

Shape-dependence of particle rotation in isotropic turbulence

,”

Phys. Fluids

27

,

035101

(

2015

).

https://doi.org/10.1063/1.4913501

Google Scholar

Crossref

23.

L.

Zhao

,

N. R.

Challabotla

,

H. I.

Andersson

, and

E. A.

Variano

, “

Rotation of nonspherical particles in turbulent channel flow

,”

Phys. Rev. Lett.

115

,

244501

(

2015

).

https://doi.org/10.1103/physrevlett.115.244501

Google Scholar

Crossref

PubMed

24.

G. A.

Voth

and

A.

Soldati

, “

Anisotropic particles in turbulence

,”

Annu. Rev. Fluid Mech.

49

,

249

–

276

(

2017

).

https://doi.org/10.1146/annurev-fluid-010816-060135

Google Scholar

Crossref

25.

K.

Gustavsson

,

M. Z.

Sheikh

,

D.

Lopez

,

A.

Naso

,

A.

Pumir

, and

B.

Mehlig

, “

Effect of fluid inertia on the orientation of a small prolate spheroid settling in turbulence

,”

New J. Phys.

21

,

083008

(

2019

).

https://doi.org/10.1088/1367-2630/ab3062

Google Scholar

Crossref

26.

M. Z.

Sheikh

,

K.

Gustavsson

,

D.

Lopez

,

E.

Lévêque

,

B.

Mehlig

,

A.

Pumir

, and

A.

Naso

, “

Importance of fluid inertia for the orientation of spheroids settling in turbulent flow

,”

J. Fluid Mech.

886

,

A9

(

2020

).

https://doi.org/10.1017/jfm.2019.1041

Google Scholar

Crossref

27.

M.

Borgnino

,

K.

Gustavsson

,

F. D.

Lillo

,

G.

Boffetta

,

M.

Cencini

, and

B.

Mehlig

, “

Alignment of nonspherical active particles in chaotic flows

,”

Phys. Rev. Lett.

123

,

138003

(

2019

).

https://doi.org/10.1103/physrevlett.123.138003

Google Scholar

Crossref

PubMed

28.

S. I.

Heath-Richardson

,

A. W.

Baggaley

, and

N. A.

Hill

, “

Gyrotactic suppression and emergence of chaotic trajectories of swimming particles in three-dimensional flows

,”

Phys. Rev. Fluids

3

,

023102

(

2018

).

https://doi.org/10.1103/physrevfluids.3.023102

Google Scholar

Crossref

29.

W.

Clifton

,

R. N.

Bearon

, and

M. A.

Bees

, “

Enhanced sedimentation of elongated plankton in simple flows

,”

IMA J. Appl. Math.

83

,

743

–

766

(

2018

).

https://doi.org/10.1093/imamat/hxy024

Google Scholar

Crossref

30.

J.-A.

Arguedas-Leiva

and

M.

Wilczek

, “

Microswimmers in an axisymmetric vortex flow

,”

New J. Phys.

22

,

053051

(

2020

).

https://doi.org/10.1088/1367-2630/ab776f

Google Scholar

Crossref

31.

S. A.

Berman

and

K. A.

Mitchell

, “

Trapping of swimmers in a vortex lattice

,”

Chaos

30

,

063121

(

2020

).

https://doi.org/10.1063/5.0005542

Google Scholar

Crossref

PubMed

32.

J. S.

Guasto

,

R.

Rusconi

, and

R.

Stocker

, “

Fluid mechanics of planktonic microorganisms

,”

Annu. Rev. Fluid Mech.

44

,

373

–

400

(

2012

).

https://doi.org/10.1146/annurev-fluid-120710-101156

Google Scholar

Crossref

33.

T.

Kiørboe

,

A.

Visser

, and

K. H.

Andersen

, “

A trait-based approach to ocean ecology

,”

ICES J. Mar. Sci.

75

,

1849

–

1863

(

2018

).

https://doi.org/10.1093/icesjms/fsy090

Google Scholar

Crossref

34.

M.

Makino

and

M.

Doi

, “

Migration of twisted ribbon-like particles in simple shear flow

,”

Phys. Fluids

17

,

103605

(

2005

).

https://doi.org/10.1063/1.2107867

Google Scholar

Crossref

35.

H. C. F.

Marcos

,

T. R.

Powers

, and

R.

Stocker

, “

Separation of microscale chiral objects by shear flow

,”

Phys. Rev. Lett.

102

,

158103

(

2009

).

https://doi.org/10.1103/physrevlett.102.158103

Google Scholar

Crossref

PubMed

36.

T. M.

Hermans

,

K. J. M.

Bishop

,

P. S.

Stewart

,

S. H.

Davis

, and

B. A.

Grzybowski

, “

Vortex flows impart chirality-specific lift forces

,”

Nat. Commun.

6

,

5640

(

2015

).

https://doi.org/10.1038/ncomms6640

Google Scholar

Crossref

PubMed

37.

A.

Bukatin

,

I.

Kukhtevich

,

N.

Stoop

,

J.

Dunkel

, and

V.

Kantsler

, “

Bimodal rheotactic behavior reflects flagellar beat asymmetry in human sperm cells

,”

Proc. Natl. Acad. Sci. U. S. A.

112

,

15904

(

2015

).

https://doi.org/10.1073/pnas.1515159112

Google Scholar

Crossref

PubMed

38.

K.

Ishimoto

, “

Guidance of microswimmers by wall and flow: Thigmotaxis and rheotaxis of unsteady squirmers in two and three dimensions

,”

Phys. Rev. E

96

,

043103

(

2017

).

https://doi.org/10.1103/physreve.96.043103

Google Scholar

Crossref

PubMed

39.

G.

Jing

,

A.

Zöttl

,

E.

Clément

, and

A.

Lindner

, “

Chirality-induced bacterial rheotaxis in bulk shear flows

,”

Sci. Adv.

6

,

eabb2012

(

2020

).

https://doi.org/10.1126/sciadv.abb2012

Google Scholar

Crossref

PubMed

40.

E. J.

Hinch

and

L. G.

Leal

, “

The effect of Brownian motion on the rheological properties of a suspension of non-spherical particles

,”

J. Fluid Mech.

52

,

683

–

712

(

1972

).

https://doi.org/10.1017/s002211207200271x

Google Scholar

Crossref

41.

H.

Brenner

, “

Rheology of dilute suspension of axisymmetric Brownian particles

,”

Int. J. Multiphase Flow

1

,

195

–

341

(

1974

).

https://doi.org/10.1016/0301-9322(74)90018-4

Google Scholar

Crossref

42.

S.

Mueller

,

E. W.

Llewellin

, and

H. M.

Mader

, “

The rheology of suspensions of solid particles

,”

Proc. R. Soc. A

466

,

1201

–

1228

(

2010

).

https://doi.org/10.1098/rspa.2009.0445

Google Scholar

Crossref

43.

D.

Palanisamy

and

W. K.

den Otter

, “

Intrinsic viscosities of non-spherical colloids by Brownian dynamics simulations

,”

J. Chem. Phys.

151

,

184902

(

2019

).

https://doi.org/10.1063/1.5127001

Google Scholar

Crossref

PubMed

44.

R.

Gordon

,

D.

Losic

,

M. A.

Tiffany

,

S. S.

Nagy

, and

F. A. S.

Sterrenburg

, “

The glass menagerie: Diatoms for novel applications in nanotechnology

,”

Trends Biotechnol.

27

,

116

–

127

(

2009

).

https://doi.org/10.1016/j.tibtech.2008.11.003

Google Scholar

Crossref

PubMed

45.

M.

Idei

,

S.

Sato

,

C.

Nagasato

,

T.

Motomura

,

K.

Toyoda

,

T.

Nagumo

, and

D. G.

Mann

, “

Spermatogenesis and auxospore structure in the multipolar centric diatom Hydrosera

,”

J. Phycol.

51

,

144

–

158

(

2015

).

https://doi.org/10.1111/jpy.12261

Google Scholar

Crossref

PubMed

46.

K.

Ogane

,

N.

Suzuki

,

Y.

Aita

,

T.

Sakai

, and

D.

Lazarus

, “

Ehrenberg’s radiolarian collections from Barbados

,”

Natl. Mus. Nat. Sci. Monogr.

40

,

97

–

106

(

2009

).

Google Scholar

47.

A.

Sanfilippo

and

W. R.

Riedel

, “

Revision of the radiolarian genera Theocotyle, Theocotylissa and Thyrsocyrtis

,”

Micropaleontology

28

,

170

–

188

(

1982

).

https://doi.org/10.2307/1485229

Google Scholar

Crossref

48.

J.

Decelle

and

F.

Not

, “

Acantharia

,” in

eLS

(

Wiley

,

2015

).

Google Scholar

49.

D.

Bukry

, “

Discoaster evolutionary trends

,”

Micropaleontology

17

,

43

–

52

(

1971

).

https://doi.org/10.2307/1485036

Google Scholar

Crossref

50.

H. S.

Poelchau

, “

Distribution of Holocene silicoflagellates in north Pacific sediments

,”

Micropaleontology

22

,

164

–

193

(

1976

).

https://doi.org/10.2307/1485399

Google Scholar

Crossref

51.

D.

Murray

and

H.

Schrader

, “

Distribution of silicoflagellates in plankton and core top samples from the Gulf of California

,”

Mar. Micropaleontology

7

,

517

–

539

(

1983

).

https://doi.org/10.1016/0377-8398(83)90013-0

Google Scholar

Crossref

52.

J.

Fries

,

J.

Einarsson

, and

B.

Mehlig

, “

Angular dynamics of small crystals in viscous flow

,”

Phys. Rev. Fluids

2

,

014302

(

2017

).

https://doi.org/10.1103/physrevfluids.2.014302

Google Scholar

Crossref

53.

P.

Atkins

,

Physical Chemistry

, 6th ed. (

Oxford University Press

,

1998

).

Google Scholar

54.

E. J.

Hinch

and

L. G.

Leal

, “

Rotation of small non-axisymmetric particles in a simple shear flow

,”

J. Fluid Mech.

92

,

591

–

608

(

1979

).

https://doi.org/10.1017/s002211207900077x

Google Scholar

Crossref

55.

A. L.

Yarin

,

O.

Gottlieb

, and

I. V.

Roisman

, “

Chaotic rotation of triaxial ellipsoids in simple shear

,”

J. Fluid Mech.

340

,

83

–

100

(

1997

).

https://doi.org/10.1017/s0022112097005260

Google Scholar

Crossref

56.

I. R.

Thorp

and

J. R.

Lister

, “

Motion of a non-axisymmetric particle in viscous shear flow

,”

J. Fluid Mech.

872

,

532

–

559

(

2019

).

https://doi.org/10.1017/jfm.2019.367

Google Scholar

Crossref

57.

E.

Yariv

, “

Self-propulsion in a viscous fluid: Arbitrary surface deformations

,”

J. Fluid Mech.

550

,

139

–

148

(

2006

).

https://doi.org/10.1017/s002211200500830x

Google Scholar

Crossref

58.

K.

Ishimoto

and

M.

Yamada

, “

A coordinate-based proof of the scallop theorem

,”

SIAM J. Appl. Math.

72

,

1686

–

1694

(

2012

).

https://doi.org/10.1137/110853297

Google Scholar

Crossref

59.

J. W.

Swan

,

J. F.

Brady

,

R. S.

Moore

, and

ChE 174

, “

Modeling hydrodynamic self-propulsion with Stokesian dynamics. Or teaching Stokesian dynamics to swim

,”

Phys. Fluids

23

,

071901

(

2011

).

https://doi.org/10.1063/1.3594790

Google Scholar

Crossref

60.

B.

Nasouri

and

G. J.

Elfring

, “

Higher-order force moments of active particles

,”

Phys. Rev. Fluids

3

,

044101

(

2018

).

https://doi.org/10.1103/physrevfluids.3.044101

Google Scholar

Crossref

61.

C.

Pozrikidis

,

Boundary Integral and Singularity Methods

(

Cambridge University Press

,

1992

).

Google Scholar

62.

K.

Ishimoto

and

E. A.

Gaffney

, “

Squirmer dynamics near a boundary

,”

Phys. Rev. E

88

,

062702

(

2013

).

https://doi.org/10.1103/physreve.88.062702

Google Scholar

Crossref

63.

O. S.

Pak

and

E.

Lauga

, “

Generalized squirming motion of a sphere

,”

J. Eng. Math.

88

,

1

–

28

(

2014

).

https://doi.org/10.1007/s10665-014-9690-9

Google Scholar

Crossref

64.

T. J.

Pedley

,

D. R.

Brumley

, and

R. E.

Goldstein

, “

Squirmers with swirl: A model for Volvox swimming

,”

J. Fluid Mech.

798

,

165

–

186

(

2016

).

https://doi.org/10.1017/jfm.2016.306

Google Scholar

Crossref

PubMed

65.

M.

Junk

and

R.

Illner

, “

A new derivation of Jeffery’s equation

,”

J. Math. Fluid Mech.

9

,

455

–

488

(

2007

).

https://doi.org/10.1007/s00021-005-0208-0

Google Scholar

Crossref

66.

N.

Pujara

and

E. A.

Variano

, “

Rotations of small, inertialess triaxial ellipsoids in isotropic turbulence

,”

J. Fluid Mech.

821

,

517

–

538

(

2017

).

https://doi.org/10.1017/jfm.2017.256

Google Scholar

Crossref

67.

H.

Jędrzejewska

and

A.

Szumna

, “

Making a right or left choice: Chiral self-sorting as a tool for the formation of discrete complex structures

,”

Chem. Rev.

117

,

4863

–

4899

(

2017

).

https://doi.org/10.1021/acs.chemrev.6b00745

Google Scholar

Crossref

PubMed

68.

H.

Goldstein

,

C. P.

Poole

, Jr., and

J. L.

Safko

,

Classical Mechanics

, 3rd ed. (

Pearson

,

2001

).

Google Scholar

69.

N. S.

Borker

,

A. D.

Stroock

, and

D. L.

Koch

, “

Controlling rotation and migration of rings in a simple shear flow through geometric modifications

,”

J. Fluid Mech.

840

,

379

–

407

(

2018

).

https://doi.org/10.1017/jfm.2018.20

Google Scholar

Crossref

70.

K.

Gustavsson

and

L.

Biferale

, “

Preferential sampling of helicity by isotropic helicoids

,”

Phys. Rev. Fluids

1

,

054201

(

2016

).

https://doi.org/10.1103/physrevfluids.1.054201

Google Scholar

Crossref

71.

M. E. J.

Holwill

and

M. A.

Sleigh

, “

Propulsion by hispid flagella

,”

J. Exp. Biol.

47

,

267

–

276

(

1967

).

Google Scholar

Crossref

PubMed

72.

E.

Lauga

,

W. R.

DiLuzio

,

G. M.

Whitesides

, and

H. A.

Stone

, “

Swimming in circles: Motion of bacteria near solid boundaries

,”

Biophys. J.

90

,

400

–

412

(

2006

).

https://doi.org/10.1529/biophysj.105.069401

Google Scholar

Crossref

PubMed

73.

C.

Nilsen

and

H. I.

Andersson

, “

Chaotic rotation of inertial spheroids in oscillating shear flow

,”

Phys. Fluids

25

,

013303

(

2013

).

https://doi.org/10.1063/1.4789376

Google Scholar

Crossref

74.

T.

Rosén

, “

Chaotic rotation of a spheroidal particle in simple shear flow

,”

Chaos

27

,

063112

(

2017

).

https://doi.org/10.1063/1.4985640

Google Scholar

Crossref

PubMed

75.

F.

Lundell

, “

The effect of particle inertia on triaxial ellipsoids in creeping shear: From drift toward chaos to a single periodic solution

,”

Phys. Fluids

23

,

011704

(

2011

).

https://doi.org/10.1063/1.3548864

Google Scholar

Crossref

76.

J.

Einarsson

and

B.

Mehlig

, “

Spherical particle sedimenting in weakly viscoelastic shear flow

,”

Phys. Rev. Fluids

2

,

063301

(

2017

).

https://doi.org/10.1103/physrevfluids.2.063301

Google Scholar

Crossref

77.

S. A.

Abtahi

and

G. J.

Elfring

, “

Jeffery orbits in shear-thinning fluids

,”

Phys. Fluids

31

,

103106

(

2019

).

https://doi.org/10.1063/1.5125468

Google Scholar

Crossref

78.

K.

Ishimoto

and

D. G.

Crowdy

, “

Dynamics of a treadmilling microswimmer near a no-slip wall in simple shear

,”

J. Fluid Mech.

821

,

647

–

667

(

2017

).

https://doi.org/10.1017/jfm.2017.220

Google Scholar

Crossref

79.

J.

Happel

and

H.

Brenner

,

Low Reynolds Number Hydrodynamics with Special Applications to Particular Media

(

Martinus Nijhoff

,

1983

).

Google Scholar

Crossref

80.

O.

Gonzalez

,

A. B. A.

Graf

, and

J. H.

Maddocks

, “

Dynamics of a rigid body in a Stokes fluid

,”

J. Fluid Mech.

519

,

133

–

160

(

2004

).

https://doi.org/10.1017/s0022112004001284

Google Scholar

Crossref

81.

K. I.

Morozov

,

Y.

Mirzae

,

O.

Kenneth

, and

A. M.

Leshansky

, “

Dynamics of arbitrary shaped propellers driven by a rotating magnetic field

,”

Phys. Rev. Fluids

2

,

044202

(

2017

).

https://doi.org/10.1103/physrevfluids.2.044202

Google Scholar

Crossref

82.

J.

Sachs

,

K. I.

Morozov

,

O.

Kenneth

,

T.

Qiu

,

N.

Segreto

,

P.

Fischer

, and

A. M.

Leshansky

, “

Role of symmetry in driven propulsion at low Reynolds number

,”

Phys. Rev. E

98

,

063105

(

2018

).

https://doi.org/10.1103/physreve.98.063105

Google Scholar

Crossref

2020

Author(s)

Jeffery orbits for an object with discrete rotational symmetry

I. INTRODUCTION

II. HYDRODYNAMICS AND SYMMETRY

A. Stokes hydrodynamics

B. Hydrokinetic symmetry

III. GENERALIZED JEFFERY EQUATIONS

A. Dynamics of a C_n-object

B. Dynamics of an object with further symmetry

1. C_nv-object

2. C_nh-object

3. D_n-object

4. D_nh-object

5. D_nd-object

6. S_2n-object

IV. GENERALIZED JEFFERY ORBITS: MOTIONS IN A SIMPLE SHEAR

A. Motions of D_3h-, D_3d-, and C_3h-objects

B. Motions of C_3v- and D₃-objects

C. Motions of S₆- and C₃-objects

D. Summary of results for the generalized Jeffery orbits

V. EXAMPLE OF A TRIANGULAR OBJECT

VI. CONCLUDING REMARKS

DATA AVAILABILITY

ACKNOWLEDGMENTS

APPENDIX: SYMMETRY ARGUMENTS FOR AN S_2n-OBJECT

REFERENCES

Citing articles via

Resources

Explore

pubs.aip.org

Connect with AIP Publishing

Jeffery orbits for an object with discrete rotational symmetry

I. INTRODUCTION

II. HYDRODYNAMICS AND SYMMETRY

A. Stokes hydrodynamics

B. Hydrokinetic symmetry

III. GENERALIZED JEFFERY EQUATIONS

A. Dynamics of a Cn-object

B. Dynamics of an object with further symmetry

1. Cnv-object

2. Cnh-object

3. Dn-object

4. Dnh-object

5. Dnd-object

6. S2n-object

IV. GENERALIZED JEFFERY ORBITS: MOTIONS IN A SIMPLE SHEAR

A. Motions of D3h-, D3d-, and C3h-objects

B. Motions of C3v- and D3-objects

C. Motions of S6- and C3-objects

D. Summary of results for the generalized Jeffery orbits

V. EXAMPLE OF A TRIANGULAR OBJECT

VI. CONCLUDING REMARKS

DATA AVAILABILITY

ACKNOWLEDGMENTS

APPENDIX: SYMMETRY ARGUMENTS FOR AN S2n-OBJECT

REFERENCES

Citing articles via

Resources

Explore

pubs.aip.org

Connect with AIP Publishing

This Feature Is Available To Subscribers Only

A. Dynamics of a C_n-object

1. C_nv-object

2. C_nh-object

3. D_n-object

4. D_nh-object

5. D_nd-object

6. S_2n-object

A. Motions of D_3h-, D_3d-, and C_3h-objects

B. Motions of C_3v- and D₃-objects

C. Motions of S₆- and C₃-objects

APPENDIX: SYMMETRY ARGUMENTS FOR AN S_2n-OBJECT