Autonomous navigation of shape-shifting microswimmers

Kyle J. M. Bishop; Yong Dou

REVIEW 3 major objections 3 minor 29 references

Autonomous navigation of shape-shifting microswimmers

T0 review · 3 major / 3 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read This paper establishes a design principle for programming autonomous navigation by making a particle's shape respond to local stimulus and thereby redirect its self-propelled motion.

desk verdict A genuinely new design idea for autonomous colloids, with a real but fixable gap: deformation-induced swimming is dropped, not proven negligible. read the letter →

arxiv 1908.05808 v1 pith:URZSIIDX submitted 2019-08-16 cond-mat.soft

Yong Dou , Kyle J. M. Bishop This is my paper · ORCID

classification cond-mat.soft

keywords autonomousnavigationactivecolloidschemotaxisself-phoresisshape-shiftingmicroswimmersstimulusgradientsCMA-ESoptimization

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper establishes a design principle for programming autonomous navigation in active colloids: let the particle's shape respond to the local magnitude of a scalar stimulus (such as a chemical concentration), and let that shape change redirect the particle's self-propelled motion. The authors show that when the curvature radius of the particle's circular trajectory is independent of the stimulus while the propulsion orientation rotates with stimulus, the particle drifts steadily along the stimulus gradient at a speed set by the slope of that rotation. They demonstrate the idea by simulating shape-shifting clusters of self-phoretic spheres, using numerical optimization to find geometries that satisfy the constant-radius condition and produce positive or negative chemotaxis. If correct, this offers a route to synthetic colloidal robots that navigate heterogeneous environments without onboard sensors, memory, or external control.

What carries the argument

The load-bearing identity is the drift-velocity formula V = −(1/2)GURα′e_x + (1/2)GUR′e_y + O(G²), derived by averaging the particle's circular motion over one orbit; it converts the design problem into two geometric requirements: constant signed orbit radius R (so R′ = 0) and a propulsion orientation that rotates steeply with the stimulus (large α′). The requirements are realized in silico through three components: a far-field method-of-reflections solution for the self-phoretic velocities of rigid sphere clusters, a set of 'standard shapes' (with the particle frame chosen so that shape change itself produces no swimming) that maps a shape parameter s onto cluster geometry, and a CMA-ES optimization (covariance matrix adaptation evolution strategy) over cluster geometry that minimizes the objective ⟨[R(s,d) − R0]²⟩ over all shape states s.

What would settle it

Measure the drift velocity of the optimized three-sphere cluster (a2 = 0.0985, a3 = 0.0958, L2 = 0.677) in a steady linear chemical gradient of known strength G; if the drift does not occur along the gradient at speed −(1/2)GURα′ (with U, R, α′ measured independently), the design criterion fails, and the failure would be even clearer if the shape response time is comparable to the orbital period.

Watch

Extended reading notes

Core claim

The paper's central claim is that shape is a programmable medium for autonomous navigation. For a self-propelled particle whose linear speed U, propulsion orientation α, and angular speed Ω depend only on the local stimulus S, motion in a weak uniform gradient S = Gx produces a drift velocity V = −(1/2)GURα′e_x + (1/2)GUR′e_y + O(G²), where R = U/Ω is the signed radius of the circular trajectory and primes denote derivatives with respect to S. If the geometry is designed so that R′ = 0, the cross-gradient drift vanishes, and a monotone α(S) that rotates the propulsion direction by up to 180° yields steady migration up or down the gradient. The authors implement this design for a three-sphere self-phoretic cluster with one stimulus-responsive bond, optimize the remaining bond length and sphere radii so that R varies by only about 2%, and compute trajectories that climb the gradient with the predicted drift speed. The same framework yields clusters that swim down the gradient or perpendicular to it by choosing different response functions or objectives.

Load-bearing premise

The particle's internal shape state is assumed to be a single-valued, instantaneous function of the stimulus magnitude at the particle center; if real responsive materials have delays, hysteresis, or sensitivity to the gradient across the particle, the derived drift velocities and the designed trajectories will not hold.

Editorial extensions

If this is right

Autonomous chemotaxis can be encoded entirely in the particle's geometry and material response, removing the need for onboard sensing, memory, or external control.
The same design framework produces negative chemotaxis, and with a modified objective function it produces motion perpendicular to the gradient, so a library of behaviors is accessible from one physical mechanism.
Because the arguments depend only on shape-dependence of propulsion, the strategy should transfer to other shape-based propulsion mechanisms such as induced-charge electrophoresis or acoustic actuation.
The noise analysis sets a concrete size scale—roughly a micrometer or larger for typical phoretic speeds in water—below which rotational diffusion erases the navigational signal, matching the scale of chemotactic bacteria.
The comparison of optimization methods indicates that greedy searches fail while evolutionary strategies reliably find cluster geometries with the required constant-radius property, making the design step computationally feasible.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The approach suggests that any stimuli-responsive material (e.g., hydrogels or liquid-crystal elastomers) that changes a cluster's bond lengths could be matched, via the same optimization, to a desired navigation behavior, turning 'shape' into a general programming language for active matter.
The drift formula assumes the stimulus varies slowly across one orbit; an immediate extension would test steep or time-dependent gradients, where the O(G²) terms and finite shape-response time become important, potentially enabling navigation in fluctuating environments.
Because the stimulus only sets an internal shape state, the mechanism is not limited to chemical cues: light, temperature, pH, or magnetic field strength could serve as the scalar stimulus, as long as the shape response is fast and single-valued.
A finite response time in the shape state would introduce memory, which could be exploited to emulate bacterial run-and-tumble behavior; the model's assumption of an instantaneous unique mapping is the main obstacle to such an extension.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Dear [Colleague],

The short version: this paper has a clean, appealing idea—encode chemotaxis in a particle's stimulus-responsive shape, then optimize the geometry so the orbit radius is constant. The drift formula is plausible, the optimization comparison is fair, and the noise estimate gives a concrete size scale (~1 μm). If you work on active colloids or micro-robotics, it's worth a look.

What's genuinely new: prior work aligned particles in gradients or used static shape control. Here the shape itself is the sensor, and they turn the design problem into an optimization over cluster geometry. That's a real step. The CMA-ES comparison is also done honestly (hill climbing gets trapped, random search underperforms), and the in-silico demonstrations are consistent with the model.

The soft spots are mostly about what's not in the paper. The drift derivation and the 'standard shapes' construction are in an unavailable supplemental. The far-field method of reflections is used without validation against full numerics or experiment. No code or data are shipped. For a paper whose whole argument is a calculation, that's thin.

The bigger question is the one the stress-test raises: deformation-induced swimming. The paper says standard shapes are chosen so the particle frame doesn't translate or rotate as s changes, 'effectively removing' swimming due to shape change. That framing is misleading. In low-Reynolds locomotion, the frame is a gauge choice; the physical displacement from a non-reciprocal shape trajectory is not zero just because you choose a nice frame. The paper never proves its shape trajectory is swimming-neutral, and the text suggests they simply dropped the term. That is a genuine gap.

But the stress-test overstates the damage. Scaling it out: a shape change of order Δs ~ f'GR per orbit gives a deformation-induced displacement of order L Δs, versus the designed drift displacement of order R Δs (or G R^2 α'). Since R >> L, the omitted term is smaller by L/R—a few percent in their examples. So the central design principle probably survives, but it needs a proper derivation and the supplemental proof before I'd trust the quantitative claims.

The 'circularity' worry doesn't land for me. The paper is a design procedure: derive a criterion, optimize to meet it, then simulate to check. That is not circular; it's engineering.

Bottom line: worth a serious referee, but revision should require (1) the supplemental with the drift derivation and the standard-shape construction, (2) an explicit accounting of deformation-induced swimming (even if it's an estimate showing it's small), and (3) validation of the far-field model against at least one full boundary-integral solution.

Regards.

Referee Report

3 major / 3 minor

Summary. The paper proposes a design strategy for autonomous navigation of active colloidal particles in which the particle shape responds to a local scalar stimulus and thereby redirects self-propelled motion. For a particle whose speed U, propulsion orientation α, and angular velocity Ω depend on an internal shape state s=f(S(x_p,t)), the authors state that motion in a uniform gradient S=Gx produces a drift velocity V=−(1/2)GURα′e_x+(1/2)GUR′e_y+O(G^2), where R=U/Ω, and that chemotaxis along the gradient requires R to be independent of S. They apply this to rigid self-phoretic clusters of spheres, compute response functions with a far-field method of reflections, and formulate design as minimization of the objective O(d)=⟨[R(s,d)−R0]^2⟩. An optimized three-sphere cluster with one stimulus-responsive bond is shown in simulation to migrate up an applied gradient, and a rotational-noise estimate gives a minimum particle size of order 1 μm for effective navigation.

Significance. If the technical gaps are closed, this is a conceptually attractive design principle: it replaces direct gradient sensing with shape-mediated feedback, is not tied to a single propulsion mechanism, and leads to an explicit, testable prediction about particle size. The optimization framework and the noise analysis are constructive, and the paper clearly identifies the objective function needed to encode chemotaxis. The main weaknesses are that the central drift formula and the standard-shape construction are deferred to an unavailable supplemental, that the shape-change swimming contribution is not shown to be negligible, and that the demonstration is entirely in silico and is entangled with the same model used to produce the design. The strengths are the clean conceptual framing, the falsifiable design objective, and the concrete size estimate, rather than any empirical validation.

major comments (3)

[Particle motion in stimulus gradients] The drift formula V=−(1/2)GURα′e_x+(1/2)GUR′e_y+O(G^2) is the mathematical core of the paper, but it is stated without derivation and deferred to supplemental material [17]. Moreover, the text immediately following the formula claims a maximum drift speed of Vmax=GUR and decay V≈GUR/S^2 for α(S)=−arctan(S); substituting α′=−1/(1+S^2) into the stated formula gives Vmax=GUR/2 and V≈GUR/(2S^2). This factor-of-two discrepancy must be fixed, and the derivation with all underlying assumptions must be included in the main text or an accessible supplement before the design claims can be evaluated.
[Shape-shifting clusters] The claim that the standard shapes are chosen 'such that the particle frame does not translate or rotate within the viscous fluid as the particle changes its shape' is, as written, a statement about the body frame rather than about the physical flow. At zero Reynolds number, a non-closed sequence of shapes generally produces a net rigid-body displacement, and fixing a frame does not eliminate that effect. Because ds/dt = f′(S)(∇S·U) is O(GU), any omitted deformation-induced swimming velocity is of the same order as the designed drift V=−(1/2)GURα′. The paper must specify the deformation kinematics and demonstrate that the shape changes are swimming-neutral; otherwise the trajectories in Figs. 2d and 3d are not established to describe the actual shape-shifting particle.
[Self-phoretic clusters and Design of chemotactic clusters] The response functions R(s) and α(s) that enter the design are computed with a far-field method of reflections [19], but no accuracy estimate or comparison with exact or boundary-element solutions is reported. For the optimized cluster of Fig. 3b, the 2% variation claimed for R(s) around R0=2 is used as evidence of successful design; if the far-field approximation carries errors of that size, the optimized geometry may not be robust. A convergence check or validation for at least the optimal geometry is needed.

minor comments (3)

[References] Reference [8] spells 'viscosicty'; it should be 'viscosity'.
[Figure captions] The phrase 'the shape parameter is assume to vary with the local stimulus' appears in both Fig. 2d and Fig. 3d captions; it should read 'is assumed to vary'.
[Particle motion in stimulus gradients] The statement that setting Sl=0 and Ss=1 is 'without loss of generality' should be reconciled with the fact that these parameters carry physical units; the rescaling applies only after all stimuli are expressed in units of Ss.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the drift formula is a parameter-free expansion, and the optimized-cluster demonstration is a design validation rather than a fitted prediction.

full rationale

The paper's central derivation is the drift velocity V = -1/2 G U R alpha' for a particle whose response functions depend on the local stimulus. This is a small-parameter expansion of the orientation dynamics, and nothing in the main text or equations shows it being assumed into the response functions. The design step minimizes O(d)=<[R(s,d)-R0]^2> to enforce the R=constant condition that the drift theory identifies; this is an explicit design objective, not a hidden fit or a parameter tuned to a subset of data and then renamed as a prediction. The simulated up-gradient trajectory (Fig. 3d) uses the optimized response functions U(s), alpha(s), Omega(s) and the chosen sigmoidal coupling s=(1+e^{-S})^{-1}; the sign and magnitude of the resulting drift are emergent consequences of the optimized geometry (in particular of alpha(s)), so the demonstration is not equivalent to the input by construction. The main caveat is a model-completeness issue rather than a circular one: the paper defers to supplemental [17] the construction of 'standard shapes' that remove shape-change swimming, and the concern that deformation-induced swimming is of the same order as the designed drift is a legitimate physical check that is not addressed in the main text. This belongs to correctness risk, not to circularity. Self-citations [12-14] are contextual examples of shape-directed propulsion and are not load-bearing for the drift derivation or the optimization; the far-field cluster hydrodynamics is credited to external works [18,19]. Accordingly, no circular step can be exhibited with a quote and a specific reduction.

Assumptions & free parameters 4 free parameters · 7 assumptions · 0 invented entities

The central design demonstration rests on several modeling assumptions: instantaneous local sensing, shape-response decoupled from propulsion, and the far-field phoretic model. The free parameters are prescribed or assumed for the example, not fitted to data. No new physical entities are introduced.

free parameters (4)

Prescribed cluster geometry parameters = a1 = 0.1, L3 = 1, Lmin = 0.5, Lmax = 1.5, R0 = 2
Chosen by hand to define the example three-sphere cluster and target radius; the method itself does not depend on these specific values, but the demonstration trajectory does.
Response function location and scale = Sl = 0, Ss = 1
Set by rescaling the stimulus without loss of generality, so not a substantive free parameter.
Stimulus-shape response function = s = (1 + e^{-S})^{-1}
Assumed ad hoc to generate the chemotactic trajectory in Fig. 3d; no specific stimuli-responsive material model is given.
Typical phoretic angular velocity for noise estimate = Omega = 1 rad/s
Assumed to estimate the minimum particle size of about 1 micrometer for navigation; a different value shifts the size estimate.

assumptions (7)

domain assumption The internal shape state s is a unique, instantaneous function of the local stimulus magnitude at the particle center: s = f(S(xp,t)).
Invoked at the start of the model; if shape response has memory or depends on the gradient across the particle, the response-function framework and drift formula fail.
domain assumption The particle's linear and angular velocities depend only on the internal state s, not on orientation or position in the gradient: U=U(s), Omega=Omega(s).
Fundamental to deriving drift from response functions; stated in the model setup.
domain assumption The far-field method of reflections accurately captures self-phoretic cluster velocities.
Used for all computations of U(s), Omega(s), and R(s) without comparison to full numerical solutions or experiments.
domain assumption Shape changes can be implemented kinematically without producing additional swimming (standard shapes chosen so the particle frame does not translate or rotate).
Section 'Shape-shifting clusters'; the paper removes swimming due to shape change by construction, but physical realizability is not demonstrated.
domain assumption Low Péclet number and low Reynolds number limits, so diffusion is quasi-steady and Stokes flow applies.
Standard assumptions for phoretic colloids; invoked in the self-phoretic cluster model.
domain assumption Rotational diffusion can be modeled with a thin-disk friction coefficient for the noise analysis.
Used to estimate the minimum particle size; the approximation may affect the quantitative estimate.
ad hoc to paper The sigmoidal relationship s = (1 + e^{-S})^{-1} between stimulus and shape.
Assumed for the trajectory demonstration; not derived from a specific stimuli-responsive material model.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Autonomous navigation of shape-shifting microswimmers." pith.science (2026). https://pith.science/paper/URZSIIDX

@misc{pith2026190805808,
  author       = {Pith},
  title        = {Pith review of: Autonomous navigation of shape-shifting microswimmers},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/URZSIIDX}},
  note         = {Machine review of arXiv:1908.05808}
}

read the original abstract

We describe a method for programming the autonomous navigation of active colloidal particles in response to spatial gradients in a scalar stimulus. Functional behaviors such as positive or negative chemotaxis are encoded in the particle shape, which responds to the local stimulus and directs self-propelled particle motions. We demonstrate this approach using a physical model of stimuli-responsive clusters of self-phoretic spheres. We show how multiple autonomous behaviors can be achieved by designing the particle geometry and its stimulus response.

Figures

Figures reproduced from arXiv: 1908.05808 by the authors.

**Figure 2.** FIG. 2. Shape-shifting self-phoretic clusters. (a) Spheres con [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗

**Figure 3.** FIG. 3. Design of a chemotactic three-sphere cluster. (a) [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

29 extracted references · 26 canonical work pages

[17]

See supplemental material at [url will be inserted by pub- lisher] for (1) derivation of the drift velocity in a uniform gradient, (2) model details on rigid self-phoretic clusters, (3) deﬁnition of standard shapes for shape-shifting parti- cles, (4) a designed cluster moving down the gradient, (5) a designed cluster moving perpendicular to the gradient, ...

work page
[19]

Varma, T

A. Varma, T. D. Montenegro-Johnson, and S. Miche- lin, Clustering-induced self-propulsion of isotropic au- tophoretic particles, Soft Matter 14, 7155 (2018)

work page 2018
[1]

M. E. Cates, Diﬀusive transport without detailed bal- ance in motile bacteria: does microbiology need statisti- cal physics?, Rep. Prog. Phys. 75, 042601 (2012)

work page 2012
[2]

Bechinger, R

C. Bechinger, R. Di Leonardo, H. L¨ owen, C. Reichhardt, G. Volpe, and G. Volpe, Active particles in complex 5 and crowded environments, Rev. Mod. Phys. 88, 045006 (2016)

work page 2016
[3]

Y. Hong, N. M. Blackman, N. D. Kopp, A. Sen, and D. Velegol, Chemotaxis of nonbiological colloidal rods, Phys. Rev. Lett. 99, 178103 (2007)

work page 2007
[4]

M. N. Popescu, W. E. Uspal, C. Bechinger, and P. Fis- cher, Chemotaxis of active Janus nanoparticles, Nano Lett. 18, 5345 (2018)

work page 2018
[5]

T. R. Kline, W. F. Paxton, T. E. Mallouk, and A. Sen, Catalytic nanomotors: Remote-controlled autonomous movement of striped metallic nanorods, Angew. Chem. Int. Ed. 44, 744 (2005)

work page 2005
[6]

Lozano, B

C. Lozano, B. Ten Hagen, H. L¨ owen, and C. Bechinger, Phototaxis of synthetic microswimmers in optical land- scapes, Nat. Commun. 7, 12828 (2016)

2016

Show all 29 references

[7]

Palacci, S

J. Palacci, S. Sacanna, A. Abramian, J. Barral, K. Han- son, A. Y. Grosberg, D. J. Pine, and P. M. Chaikin, Artiﬁcial rheotaxis, Sci. Adv. 1, e1400214 (2015)

2015
[8]

Liebchen, P

B. Liebchen, P. Monderkamp, B. ten Hagen, and H. L¨ owen, Viscotaxis: microswimmer navigation in vis- cosity gradients, Phys. Rev. Lett. 120, 208002 (2018)

2018
[9]

Golestanian, T

R. Golestanian, T. Liverpool, and A. Ajdari, Designing phoretic micro-and nano-swimmers, New J. Phys. 9, 126 (2007)

2007
[10]

Palagiand P

S. Palagiand P. Fischer, Bioinspired microrobots, Nat. Rev. Mater. 3, 113 (2018)

2018
[11]

K. Han, C. W. Shields IV, and O. D. Velev, Engineering of self-propelling microbots and microdevices powered by magnetic and electric ﬁelds, Advanced Functional Mate- rials 28, 1705953 (2018)

2018
[12]

A. M. Brooks, S. Sabrina, and K. J. M. Bishop, Shape- directed dynamics of active colloids powered by induced- charge electrophoresis, Proc. Natl. Acad. Sci. USA 115, E1090 (2018)

2018
[13]

Sabrina, M

S. Sabrina, M. Tasinkevych, S. Ahmed, A. M. Brooks, M. Olvera de la Cruz, T. E. Mallouk, and K. J. M. Bishop, Shape-directed microspinners powered by ultra- sound, ACS Nano 12, 2939 (2018)

2018
[14]

A. M. Brooks, M. Tasinkevych, S. Sabrina, D. Vele- gol, A. Sen, and K. J. M. Bishop, Shape-directed ro- tation of homogeneous micromotors via catalytic self- electrophoresis, Nat. Commun. 10, 495 (2019)

2019
[15]

Magdanz, G

V. Magdanz, G. Stoychev, L. Ionov, S. Sanchez, and O. G. Schmidt, Stimuli-responsive microjets with recon- ﬁgurable shape, Angew. Chem. Int. Ed. 53, 2673 (2014)

2014
[16]

Palagi, A

S. Palagi, A. G. Mark, S. Y. Reigh, K. Melde, T. Qiu, H. Zeng, C. Parmeggiani, D. Martella, A. Sanchez- Castillo, N. Kapernaum, F. Giesselmann, D. S. Wiersma, E. Lauga, and P. Fischer, Structured light enables biomimetic swimming and versatile locomotion of pho- toresponsive s...

2016
[18]

Sotoand R

R. Sotoand R. Golestanian, Self-assembly of catalytically active colloidal molecules: Tailoring activity through sur- face chemistry, Phys. Rev. Lett. 112, 068301 (2014)

2014
[20]

R. Niu, A. Fischer, T. Palberg, and T. Speck, Dynamics of binary active clusters driven by ion-exchange particles, ACS Nano 12, 10932 (2018)

2018
[21]

Schmidt, B

F. Schmidt, B. Liebchen, H. L¨ owen, and G. Volpe, Light-controlled assembly of active colloidal molecules, J. Chem. Phys. 150, 094905 (2019)

2019
[22]

Shapereand F

A. Shapereand F. Wilczek, Geometry of self-propulsion at low reynolds number, J. Fluid Mech. 198, 557 (1989)

1989
[23]

Hansen, The cma evolution strategy: A tutorial, arXiv:1604.00772 (2016)

N. Hansen, The cma evolution strategy: A tutorial, arXiv:1604.00772 (2016)

2016 arXiv
[24]

M. Z. Miskinand H. M. Jaeger, Adapting granular ma- terials through artiﬁcial evolution, Nat. Mater. 12, 326 (2013)

2013
[25]

Kimand S

S. Kimand S. J. Karrila, Microhydrodynamics (Dover, New York, 2005)

2005
[26]

F. Ma, X. Yang, H. Zhao, and N. Wu, Inducing propul- sion of colloidal dimers by breaking the symmetry in electrohydrodynamic ﬂow, Phys. Rev. Lett. 115, 208302 (2015)

2015
[27]

critters

M. Driscoll, B. Delmotte, M. Youssef, S. Sacanna, A. Donev, and P. Chaikin, Unstable fronts and stable “critters” formed by microrollers, Nat. Phys. 13, 375 (2017)

2017
[28]

Alvarez, M

L. Alvarez, M. A. Fernandez-Rodriguez, and L. Isa, Adaptable active colloids, in preparation (2019)

2019
[29]

Liebchenand H

B. Liebchenand H. L¨ owen, Optimal control strategies for active particle navigation, arXiv:1901.08382 (2019)

2019 arXiv

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [17]

See supplemental material at [url will be inserted by pub- lisher] for (1) derivation of the drift velocity in a uniform gradient, (2) model details on rigid self-phoretic clusters, (3) deﬁnition of standard shapes for shape-shifting parti- cles, (4) a designed cluster moving down the gradient, (5) a designed cluster moving perpendicular to the gradient, ...

work page

[2] [19]

Varma, T

A. Varma, T. D. Montenegro-Johnson, and S. Miche- lin, Clustering-induced self-propulsion of isotropic au- tophoretic particles, Soft Matter 14, 7155 (2018)

work page 2018

[3] [1]

M. E. Cates, Diﬀusive transport without detailed bal- ance in motile bacteria: does microbiology need statisti- cal physics?, Rep. Prog. Phys. 75, 042601 (2012)

work page 2012

[4] [2]

Bechinger, R

C. Bechinger, R. Di Leonardo, H. L¨ owen, C. Reichhardt, G. Volpe, and G. Volpe, Active particles in complex 5 and crowded environments, Rev. Mod. Phys. 88, 045006 (2016)

work page 2016

[5] [3]

Y. Hong, N. M. Blackman, N. D. Kopp, A. Sen, and D. Velegol, Chemotaxis of nonbiological colloidal rods, Phys. Rev. Lett. 99, 178103 (2007)

work page 2007

[6] [4]

M. N. Popescu, W. E. Uspal, C. Bechinger, and P. Fis- cher, Chemotaxis of active Janus nanoparticles, Nano Lett. 18, 5345 (2018)

work page 2018

[7] [5]

T. R. Kline, W. F. Paxton, T. E. Mallouk, and A. Sen, Catalytic nanomotors: Remote-controlled autonomous movement of striped metallic nanorods, Angew. Chem. Int. Ed. 44, 744 (2005)

work page 2005

[8] [6]

Lozano, B

C. Lozano, B. Ten Hagen, H. L¨ owen, and C. Bechinger, Phototaxis of synthetic microswimmers in optical land- scapes, Nat. Commun. 7, 12828 (2016)

2016